Reagents

Molecular biology reagents, including Q5® High-Fidelity 2X Master Mix, 1 kb Plus DNA ladder (100 bp-10 kb), and agarose gel electrophoresis loading dye were purchased from New England Biolabs (Ipswich, MA, USA). iScript™ cDNA synthesis reagents were purchased from Bio-Rad (Hercules, CA, USA). PowerUp™ SYBR™ Green Master Mix for quantitative RT-PCR was purchased from Thermo Fisher Scientific (Waltham, MA, USA). Agarose and SYBR™ Safe DNA stain were respectively purchased from GoldBio (St. Louis, MO, USA) and APExBIO (Houston, TX, USA). Reagents for plasmid extraction (mini- and midi-prep) and DNA purification were purchased from Promega (Madison, WI, USA) and New England Biolabs. Reagents for RNA extraction were purchased from Qiagen (Hilden, Germany). Chemically competent E. coli was obtained from New England Biolabs. Oligonucleotide primers for PCR amplification were designed using NEBuilder and ordered from Integrated DNA Technologies (Carlsbad, CA, USA). Reagents for Western blotting, including 4–20% Mini-PROTEAN® TGX Stain-Free™ gels and Immun-Blot® PVDF membranes were purchased from Bio-Rad (Hercules, CA, USA). Reagents for the chemiluminescent detection of immunoreactive bands were obtained from Thermo Fisher Scientific. Antibodies were purchased from Sigma-Aldrich (St. Louis, MO, USA).

HT-22 cell lines were purchased from Sigma-Aldrich. MIN6 cells were purchased from Thermo Fisher Scientific. All other cell lines were obtained from the American Type Culture Collection (ATCC). Dulbecco’s Modified Eagle Media (DMEM), Roswell Park Memorial Institute media (RPMI 1640), sodium pyruvate, and casein were purchased from Sigma-Aldrich (St. Louis, MO, USA). GlutaMAX™ (100X), penicillin-streptomycin (104 units/mL penicillin, 10 mg/mL streptomycin), sterile phosphate buffered saline (PBS), sodium butyrate, CellTracker™ Green dye, TrypLE, and anti-human CD3/CD28 magnetic Dynabeads™ were purchased from Thermo Fisher Scientific. Fetal bovine serum (FBS) was purchased from R&D Systems (Minneapolis, MN, USA) and Thermo Fisher Scientific. Linear polyethyleneimine (25 kDa) transfection reagent was purchased from Polysciences Inc. (Warrington, PA, USA). Lenti-X™ viral concentrator was purchased from Takara Bio (San Jose, CA, USA). RIPA cell lysis buffer and polybrene were purchased from Santa Cruz Biotechnology (Dallas, TX, USA).

Reagents for performing cytotoxicity and viability assays, namely CellTiter-Glo®, CellTiter 96® AQueous, Caspase-Glo® 3/7, and CellTiter-Blue® were purchased from Promega (Madison, WI, USA). Lipids were purchased from Avanti (Alabaster, AL, USA) and ATTO-TEC (Siegen, Germany). Silica beads were purchased from Bangs labs (Fishers, IN, USA). Monoclonal mouse anti-biotin IgG1κ was purchased from Jackson ImmunoResearch Labs (West Grove, PA, USA). All other monoclonal antibodies were purchased from Thermo Fisher Scientific. Matrigel® was purchased from Corning (NY, USA). Pierce™ BCA Protein Assay reagent were purchased from Thermo Fisher Scientific. Ultra-Sensitive Mouse Insulin ELISA kit was purchased from Crystal Chem (Elk Grove Village, IL).

Molecular biology

The construction of the Aqp1-expressing lentiviral plasmid (pJY22) has been described in our previous study. An enhanced GFP reporter was co-expressed with Aqp1 using an internal ribosome entry site (IRES) to allow selection of stably transduced cells by fluorescence-activated cell sorting (FACS). To generate a control plasmid (pADM04), the GFP sequence was amplified using Q5® High-Fidelity 2X Master Mix and cloned by Gibson assembly in the same lentiviral vector backbone used for Aqp1 expression. All constructs were verified using Sanger DNA sequencing (Genewiz, San Diego, CA, USA).

Cell culture and engineering

Cells were routinely cultured at 37 °C in a humidified incubator containing 5% CO2. HT22, J774A.1, MIN6, and MDA-MB-231 cells were grown as adherent cultures in DMEM, while Jurkat cells were grown in suspension in RPMI. The growth medium was supplemented with 100 U/mL penicillin, 100 µg/mL streptomycin, and FBS (10–15%). GlutaMAX™ and sodium pyruvate were used as additional supplements in all cell lines except HT22. HT22, J774A.1, MDA-MB-231, and MIN6 cells were cultured in 4.5 g/L glucose. Jurkat cells were cultured in 1 g/L glucose. For passaging, adherent cells were detached from the plate using trypsin or a cell scraper (J774A.1 cells).

Lentivirus was produced using a combination of three plasmid vectors: a transfer plasmid encoding Aqp1 or GFP, a packaging plasmid, and an envelope plasmid encoding the VSV-G protein to confer broad cell-type tropism. To produce lentivirus, 22 µg of the transfer plasmid, 22 µg of the packaging plasmid, and 4.5 µg of the envelope plasmid were mixed and delivered to 293T cells by transient transfection with polyethyleneimine. Approximately 24 h after transfection, cells were treated with sodium butyrate (10 mM) to enhance the expression of viral genes. Viral production was allowed to continue for another 72 h before collecting the spent media and precipitating the lentiviral particles using Lenti-X™ concentrator. The concentrated lentivirus was resuspended in 200 µL PBS and stored as aliquots at -80 °C.

For lentiviral transduction, cells were grown in individual wells of a six-well plate to 70% confluence, aspirated to remove spent media, and incubated with lentiviral particles resuspended in ∼ 1 mL of media containing 8 µg/mL polybrene. Cells were spinfected by centrifuging the six-well plates at 1048 × g at 30 °C for 90 min. Following spinfection, the plates were returned to the incubator for two days. Next, the cells were transferred from to a 10-cm plate and grown to 70% confluence. Stably transduced green fluorescent cells were enriched using FACS (Sony SH800 or Sony MA900 sorter), and the enriched populations were grown back out and stored as cryo-stocks until further use.

Assaying Aqp1 transgene expression using qRT-PCR

Aqp1- and GFP-expressing cells were cultured in six-well plates for 24 h. Next, cells were lysed in Buffer RLT (Qiagen) supplemented with 1% β-mercaptoethanol and the lysate was centrifuged in QIAshredder columns to shear genomic DNA. Total RNA was extracted according to the manufacturer’s (Qiagen) protocol. 1 µg of RNA was then reverse-transcribed using iScript™ cDNA synthesis reagents according to the manufacturer’s protocol. The cDNA was diluted 10-fold in nuclease-free water and gene-specific quantitative PCR (qRT-PCR) was carried out using Power Up™ SYBR™ Green Master Mix and 20 ng of cDNA template, according to the manufacturer’s protocol. qRT-PCR was performed using a CFX96 Touch Real-Time PCR Detection System (Bio-Rad). Cycle threshold values (\( _\)) were determined by regression fitting using CFX Maestro Software (Bio-Rad). Primers were designed using Primer3 with the forward primer binding in the N-terminal FLAG-tag of the Aqp1 transgene (Table S2). Primers were designed to have a melting temperature (\( _\)) of 60 °C, which was experimentally verified by thermal gradient PCR. To ensure that the amplification was specific Aqp1, melt curve analysis was performed, and the expected amplicon size was further confirmed using agarose gel electrophoresis. Additionally, we performed qPCR on serial dilutions of a cDNA sample to assess primer efficiency based on the slope of \( _\) vs. logarithmic dilution. Aqp1 expression was quantified using the \( ^_}\) method [73] relative to GFP-only controls and with GAPDH or actin serving as housekeeping genes for mouse and human cell lines respectively.

Western blotting

Whole-cell lysates were prepared by treating a confluent 10 cm plate of cells with 2x Laemmli sample buffer supplemented with 5% β-mercaptoethanol and 4% sodium dodecyl sulfate. The cell solution was briefly sonicated using a probe sonicator, boiled at 95 °C for 3 min and resolved by denaturing polyacrylamide gel electrophoresis. The proteins were then transferred to a PVDF membrane using a Trans-Blot® Turbo™ Transfer system (Bio-Rad). Membranes were blocked for 1 h at room temperature by incubating with blocking buffer containing 3% bovine serum albumin (BSA) in 1X Tris buffered saline and subsequently incubated at 4 °C overnight with anti-ActB (Sigma Aldrich #A2228) and anti-FLAG (Sigma-Aldrich #F1804) primary antibodies diluted at 1:20,000 and 1:1000 respectively each in blocking buffer. Following overnight incubation, the membranes were washed and incubated for 1 h at room temperature with horse radish peroxidase-conjugated anti-mouse secondary antibody diluted 1:3000 in 5% non-fat dry milk in 1X Tris buffered saline. Immunoreactive bands were detected using SuperSignal™ West Femto Maximum Sensitivity Substrate and a ChemiDoc MP Imaging System (Bio-Rad).

Magnetic resonance imaging

Cells were seeded 24–48 h prior to MRI and grown to full confluence in 10-cm tissue culture plates. In preparation for imaging, cells were harvested from the plate, centrifuged at 350 × g for 5 min, and resuspended in 200 µL of PBS. The resuspended cells were transferred to a microcentrifuge tube, centrifuged at 500 × g for 5 min, and the supernatant was carefully aspirated. The wash step was repeated one more time before centrifuging the cells (500 × g, 5 min) to form a pellet. Tubes containing cell pellets were placed in a water-filled agarose mold (1% w/v) that was housed in a custom 3D-printed MRI phantom.

All MRI experiments were performed at ambient temperature using a 7 T vertical-bore MRI scanner (Bruker) equipped with a 66 mm diameter volume coil. Stimulated echo diffusion-weighted images of cell pellets were acquired in the axial plane using the following parameters: echo time, TE = 18 ms, repetition time, TR = 1000 ms, gradient duration, δ = 5 ms, gradient separation, Δ = 300 ms, matrix size = 128 × 128, field of view (FOV) = 5.08 × 5.08 cm2, slice thickness = 1–2 mm, number of averages = 5, and four effective b-values in the range of 1–3 ms/µm2. The diffusion-weighted intensity at a given b-value was estimated by computing the mean intensity inside a manually drawn region of interest (ROI) encompassing the axial cross-section of a cell pellet. The slope of the logarithmic decay in the signal intensity versus the effective b-value was used to calculate the apparent diffusion coefficient. To generate diffusion maps, apparent diffusivity was computed for each voxel in the ROI. The ensuing image was smoothed using a median filter and pseudo-colored according to an 8-bit color scale. Least-squares regression fitting was performed by using the “fitnlm” function in Matlab (R2022b).

Cell viability assays

Aqp1-and GFP-expressing cells were seeded in 96-well plates at approximately 10,000 cells per well. Cell counting was performed manually using a hemocytometer or an automated cell counter (Countess® II). Cell viability and toxicity assays were performed by measuring intracellular ATP content, MTT bioreduction, and caspase 3/7 activation, using commercially available reagents, namely CellTiter-Glo®, CellTiter 96® AQueous, CellTiter-Blue®, and Caspase-Glo® 3/7, following the manufacturer’s instructions. Absorbance and luminescence were measured using a Tecan Spark® Microplate Reader. The integration time for luminescence was set at 1 s. The caspase 3/7 activation measurements were normalized to the viable cell count based on the total cellular ATP or MTT reduction.

Quantitative phase imaging

Quantitative phase imaging was performed using a Nanolive 3D cell explorer to determine the dry mass, sphericity, and volume of Aqp1- and GFP-expressing cells. For phase imaging, cells were seeded at a 1:24 split ratio and grown for ∼ 24 h in 35 mm glass bottom plates (Ibidi). This split ratio was chosen to ensure that the cells did not become too confluent at the time of imaging, which would hinder the accurate segmentation of individual cells. Tomographic phase images (0.202 μm x 0.202 μm lateral resolution, 0.363 μm z-resolution) were obtained from multiple regions of the plate, and cells were rendered using the “Surfaces” model in Imaris 9.0 (Oxford Instruments). Cells close to the image boundaries and those that could not be clearly differentiated from neighboring cells were excluded from the analysis. Imaris was used to compute cell volume (\( V\)) and surface area (\( S\)), which were then used to estimate sphericity (\( \)) as \( =\frac^}\left(6V\right)}^}}\). The cell mass was calculated from the refractive index tomogram using an established approach. Briefly, the refractive index difference is related to the phase shift or change in optical path length (\( OPL\)) as follows: \( OPL(x,y)= _^[n\left(x,y\right)-_]dz\), where \( z\) is the coordinate along the optical axis, \( h\) is the thickness of the cell,\( n(x,y) \)is the refractive index of the cellular material at planar coordinates \( (x,y)\), and \( _\) is the refractive index of the surrounding medium. The integral of the optical path difference over the plane of a segmented cell is related to dry mass (\( m\)) by the equation: \( m= \frac\), where \( \alpha = 0.18 \mu ^/pg\) is a known constant for eukaryotic cells [74].

Unfolded protein response (UPR) assays

Aqp1-, GFP-expressing, and wild-type cells were cultured in six-well plates for 24 h prior to experimentation. To induce ER stress, the culture medium was supplemented with tunicamycin at a final concentration of 2.5 µg/mL for 4 h. Next, qRT-PCR experiments were performed as described above for the Aqp1 transgene expression analyses. Primers (Table S2) were designed using Primer3 ensuring that the sequences were intron-flanking, produced an amplicon between 50 and 250 bp, and had a melting temperature (\( _\)) of 60 °C. As before, changes in expression of key UPR-associated genes were quantified using the \( ^_}\) method with GAPDH and actin serving as housekeeping genes for mouse and human cell lines respectively.

Phagocytosis assay

Supported lipid bilayer-coated silica beads were prepared as described before [75]. Briefly, chloroform-dissolved lipids were mixed in the following molar ratios: 96.8% POPC (Avanti, Catalog # 850,457), 2.5% Biotinyl Cap PE (Avanti, Catalog # 870,273), 0.5% PEG5000 PE (Avanti, Catalog # 880,230), and 0.2% atto390-DOPE (ATTO-TEC GmbH, Catalog # AD 390–161), and dried under argon gas overnight to remove the chloroform. The dried lipids were resuspended to 10 mM in PBS (pH 7.2) and stored under argon gas. To form small unilamellar vesicles (SUVs), the lipid mixture was subjected to 30 freeze-thaw cycles and stored at -80 °C under argon gas. Immediately before use, the SUVs were filtered through 0.22 μm PTFE syringe filters to remove lipid aggregates. To form supported lipid bilayers, 8.6 × 108 silica beads (10 µL of 10% solids, 4.89 μm mean diameter, Bangs Laboratories) were washed twice with water and twice with PBS by centrifugation at 300 × g and decanting. The beads were then mixed with 1 mM SUVs in PBS, vortexed for 10 s at medium speed, covered with foil, and incubated in an end-over-end rotator at room temperature for 0.5–2 h to allow lipid bilayers to form on the beads. The beads were washed three times with PBS to remove excess SUVs and resuspended in 100 µL of 0.2% casein in PBS for 15 min at room temperature to block nonspecific binding. Anti-biotin monoclonal mouse IgG1κ (Jackson Immuno Labs, Cat#200-602-211) was added to the beads at concentrations ranging from 0 to 10 nM and incubated for 15 min at room temperature with end-over-end mixing. The beads were washed three times to remove unbound IgG and resuspended in 100 µL PBS containing 0.2% casein.

For the bead engulfment assay, ∼ 50,000 Aqp1- or GFP-expressing J774 cells were plated in individual wells of a 96-well glass bottom plate (MatriPlate™) approximately 12–24 h prior to the experiment. Next, the cells were washed four times with engulfment imaging media (20 mM HEPEs, 135 mM NaCI, 4 mM KCl, 10 mM glucose, 1 mM CaCl2, 0.5 mM MgCl2), leaving ∼ 100 µL media between washes, and finally leaving the cells in 300 µL media. Approximately 8 × 105 beads were added to each well and engulfment was allowed to proceed for 45 min in a humidified incubator (37 °C, 5% CO2). A minimum of 36 Images were collected using a spinning disk confocal microscope (Nikon Ti2-E inverted microscope with a Yokagawa CSU-W1 spinning disk unit and an ORCA-Fusion BT scientific CMOS camera) with a 40 × 0.95 NA Plan Apo air objective. The microscope was operated using the NIS-Elements software platform (Nikon). Images were analyzed using CellProfiler [76]. All images were cropped to remove 150 pixels around the edges and background was subtracted. Macrophage cell bodies were identified from the GFP channel using a global minimum cross-entropy threshold. The phagocytic index was estimated from average integrated fluorescence intensity of atto390-labeled lipids per macrophage. Results were validated by a blinded analyzer manually counting the number of beads in 100 macrophages per condition.

Insulin secretion assay

Aqp1- and GFP-expressing MIN6 cells were seeded in individual wells of a 24-well plate at approximately 120,000 cells per well (manually counted using a hemocytometer). Four days after seeding, the culture media was aspirated and cells were washed twice with low-glucose Krebbs-Ringer bicarbonate/HEPES medium (KRBH) comprising 135 mM NaCl, 3.6 mM KCl, 2 mM NaHCO3, 0.5 mM NaH2PO4, 0.4 mM MgCl2, 1.5 mM CaCl2, 10 mM HEPES, 2 mM glucose, and 1% BSA (pH 7.4). The cells were subsequently incubated at 37 °C for 2 h in low-glucose KRBH medium, washed twice as before, and incubated at 37 °C in KRBH media containing either 2 mM (low) or 20 mM (high) glucose. One hour later, 450 µL of supernatant was collected from each well, centrifuged at 1500 × g for 5 min at 4 °C, and stored at -80 °C. Next, the cells were lysed using RIPA Lysis Buffer, and the lysates were stored at -80 °C. The insulin concentration in the thawed supernatant was determined using a mouse insulin ELISA Kit following manufacturer’s recommendations to obtain a dynamic range of 0.1–12.8 ng/mL insulin. A standard curve was generated by measuring known quantities of purified insulin and fitting the curve to a quadratic model. Insulin concentrations were normalized to total cellular protein, which was measured in thawed lysates using the bicinchoninic acid (BCA) assay.

Flow cytometry assay for immune stimulation

We first confirmed the expression of various immune receptors, including CD3, CD4, and CD45 in native Jurkat cells by staining with the following antibodies, namely anti-human CD3-PE (UCHT1, 0.024 µg), anti-human CD4-PE-Cy7 (OKT4, 0.024 µg), and anti-human CD45 eFluor® 450 (HI30, 0.2 µg),.Staining was performed by pelleting Jurkat cells at 500 × g, washing with staining buffer (PBS containing 1% FBS)and incubating with the respective antibodies for 20 min (200 µL staining buffer, 2 µL respective antibody). To avoid cross-staining, each receptor was independently stained. Next, Aqp1- and GFP-expressing Jurkat cells (∼ 500,000 cells/mL) were stimulated using anti-human CD3/CD28 magnetic Dynabeads® at a 4:1 bead-to-cell ratio (2 uL). After 24 h of stimulation, Jurkat cells were collected and de-beaded using a magnetic rack for 1–3 min. The de-beaded cells were pelleted at 500 × g by centrifugation, washed with PBS containing 1% FBS, and stained with mouse anti-human CD3-PE antibody for 20 min. After CD3 staining, the cells were washed as before and stained with anti-mouse CD25 eFluor® 450 (PC61.5, 0.1 µg). Stimulation was assessed by using flow cytometry (Attune NxT Flow Cytometer) to quantify the percentage of CD3 + cells that were also CD25+. Flow cytometry data were analyzed using FlowJo. The gating strategy is illustrated in Fig. 4d and Supplementary Fig. 2. Percentage of CD25 + cells was calculated relative to the total CD3 + population as Stimulation%= CD3+CD25+/[CD3+CD25++CD3+CD25-].

Matrigel invasion assay

MDA-MB-231 cells were grown to 70% confluence and ∼ 50,000 cells were centrifuged at 350 × g for 5 min, decanted, and resuspended in 10 µL Matrigel containing Phenol Red. The resuspended cells were deposited in the form of single drops in individual wells of a six-well plate and incubated at 37 °C for 20 min to allow the Matrigel to solidify. Next, 2 mL of medium was added to each well, and the plate was returned to the incubator. The cells were imaged every 24 h using a scanning confocal microscope (Leica SP8). Both transmitted light and fluorescent images were acquired using a 10 × 10 grid of 1024 × 1024 pixel tiles at 10X magnification to capture the Matrigel drop and the surrounding medium. Fluorescence imaging was performed using 488 nm light for excitation and emission was measured between 498 and 592 nm. Imaging was continued for up to six days, beyond which the Matrigel drops began to detach from the plate. For image analysis, Fiji was first used to manually draw boundaries around the Matrigel border and zero-fill the pixels located inside the ROI, thereby eliminating cells that did not migrate out of the droplet. Next, ilastic was used to automatically segment cells located outside the Matrigel border, using images obtained at intermediate times (e.g., day 4) as the training set. Finally, the percentage of all pixels in the image that corresponded to escaped cells was computed using Fiji and converted to area units by multiplying by the total imaged area (113.5 mm2).

Statistical analysis

Experimental data are summarized by their mean and standard error of mean obtained from multiple (n ≥ 3) biological replicates. All tests are 2-sided with the exception of the tunicamycin-based UPR assay (Fig. 3b), which was 1-tailed given the directionality of tunicamycin’s effect on UPR was known a priori. The qRT-PCR data were analyzed by performing a Student’s t-test on the \( \varDelta \varDelta _\) values. A P-value of less than 0.05 taken to indicate statistical significance.