Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

Intracellular hydrogelation preserves fluid and functional cell membrane interfaces for biological interactions

Intracellular hydrogelation preserves fluid and functional cell membrane interfaces for... ARTICLE https://doi.org/10.1038/s41467-019-09049-5 OPEN Intracellular hydrogelation preserves fluid and functional cell membrane interfaces for biological interactions 1 1 1 1 1 1 1,2 Jung-Chen Lin , Chen-Ying Chien , Chi-Long Lin , Bing-Yu Yao , Yuan-I Chen , Yu-Han Liu , Zih-Syun Fang , 1 1 1,2 2 1 Jui-Yi Chen , Wei-ya Chen , No-No Lee , Hui-Wen Chen & Che-Ming J. Hu Cell membranes are an intricate yet fragile interface that requires substrate support for stabilization. Upon cell death, disassembly of the cytoskeletal network deprives plasma membranes of mechanical support and leads to membrane rupture and disintegration. By assembling a network of synthetic hydrogel polymers inside the intracellular compartment using photo-activated crosslinking chemistry, we show that the fluid cell membrane can be preserved, resulting in intracellularly gelated cells with robust stability. Upon assessing several types of adherent and suspension cells over a range of hydrogel crosslinking densities, we validate retention of surface properties, membrane lipid fluidity, lipid order, and protein mobility on the gelated cells. Preservation of cell surface functions is further demonstrated with gelated antigen presenting cells, which engage with antigen-specific T lymphocytes and effectively promote cell expansion ex vivo and in vivo. The intracellular hydrogelation technique presents a versatile cell fixation approach adaptable for biomembrane studies and biomedical device construction. 1 2 Institute of Biomedical Sciences, Academia Sinica, Taipei 11574, Taiwan. Department of Veterinary Medicine, National Taiwan University, Taipei 10617, Taiwan. These authors contributed equally: Jung-Chen Lin, Chen-Ying Chien, Chi-Long Lin. Correspondence and requests for materials should be addressed to C.-M.H. (email: chu@ibms.sinica.edu.tw) NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications 1 1234567890():,; ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 he cell membrane is a fluid substrate that harbors a milieu morphology was observed (Supplementary Fig. 2). An evaluation of phospholipids, proteins, and glycans, which dynamically by atomic force microscopy, however, showed that the gelated Tchoreograph numerous biological interactions. The long- cells (GCs) exhibited increasing Young’s moduli that correlated standing fascination with the various biological functions of cell with the PEG-DA concentrations (Fig. 1c). Assessment of GC membranes has inspired model systems and cell-mimetic devices stability by microscopy showed no observable structural alter- 1–3 4,5 6–8 for biological studies , tissue engineering , drug delivery , nation over a 30-day observation period, whereas control cells 9–12 and immunoengineering . Toward replicating the cell mem- and non-crosslinked cells exhibited noticeable disintegration brane interface, synthetic bilayer lipid membranes and bio- within 3 days (Fig. 1d and Supplementary Fig. 3). To further conjugation strategies are commonly adopted in bottom-up confirm the assembly of hydrogel networks in the intracellular engineering of cell membrane mimics . Alternatively, top-down domain, fluorescein-diacrylate was added to the cross-linker approaches based on extraction and reconstitution of plasma mixture to covalently imbue the hydrogel network with green membranes of living cells are frequently applied to capture the fluorescence (Supplementary Fig. 1). Following membrane intricate cell-surface chemistries for biomimetic functionaliza- staining with a lipophilic DiD fluorophore, GCs showed dis- 6–8 tion . As antigen presentation, membrane fluidity, and mem- tinctive membranous and hydrogel components (Fig. 1e and brane sidedness are critical factors behind biomembrane Supplementary Fig. 4), displaying a structure reminiscent of functions and can be influenced by membrane translocation substrate-supported lipid membranes . Solubilization treatment processes, methods for harnessing this membranous component with sodium dodecyl sulfate was applied to examine the integrity continue to emerge with the aim to better study and utilize this of the gelated cytoplasm, and the fluorescent hydrogel matrices in 14–16 complex and delicate biological interface . GCs remained intact following membrane dissolution (Supple- To stabilize the fluid and functional plasma membranes and mentary Fig. 4). In a dye-exclusion study, 4 wt% GCs effectively decouple it from the dynamic state of living cells, we envision excluded a water-soluble fluorescein isothiocyanate (FITC) dye that a synthetic polymeric network can be constructed in the from entering the cytoplasm (Fig. 1f and Supplementary Fig. 4), cytoplasm to replace the cytoskeletal support for stabilizing cel- thereby confirming the plasma membrane integrity on GCs. We lular structures. Unlike endogenous cytoskeletons that are sus- also demonstrated that GCs could be stored by freezing and by ceptible to reorganization and disintegration upon perturbation lyophilization (Supplementary Fig. 5A). In addition, the intra- and cell death , a synthetic substrate scaffold can stably support cellular gelation process was applied to adherent HeLa cells, the cell membrane interface for subsequent applications. As effectively preserving the cells’ adherent property and elongated the mechanical property of cytoskeletons has drawn comparisons structures (Supplementary Fig. 5B). 17,18 to hydrogels , a cellular fixation approach mediated by intracellular assembly of hydrogel monomers is herein developed. We demonstrate that the intracellular hydrogelation technique Intracellular gelation preserves cellular features. Examination effectively preserves cellular morphology, lipid order, membrane of the cell membrane interface and the cytoplasmic hydrogel protein mobility, and biological functions of the plasma mem- matrix on GCs was performed by transmission electron micro- brane, giving rise to cell-like constructs with extraordinary sta- scopy (TEM). In comparison to control cells, GCs possessed a bility. In addition, a highly functional artificial antigen presenting perforated, hydrogel-filled interior. Treatment by detergent cell (APC) is prepared with the gelated system to highlight the stripped GCs of their membranous exterior, leaving behind platform’s utility for biomedical applications. nondissolvable hydrogel matrices (Fig. 2a). To better visualize the membrane interface on GCs, intracellular hydrogelation was applied to avian red blood cells (aRBCs), which are nucleated cells Results devoid of organelles. As hemoglobins were removed during the Intracellular hydrogelation by photoactivated cross-linking. gelation process, gelated aRBCs (G-aRBCs) exhibited a clear Three criteria were considered to establish the intracellular membrane boundary encircling a perforated nucleus (Fig. 2b and hydrogelation technique: (i) Hydrophilic cross-linking monomers Supplementary Fig. 6). Notably, the addition of a hemaggluti- with a low-molecular weight were used to facilitate cytoplasmic nating influenza virus to the G-aRBCs induced direct agglutina- permeation and minimize membrane partitioning. (ii) Cross- tion (Supplementary Fig. 6), and TEM cryosections showed linking chemistry with low-protein reactivity was adopted to similar binding patterns between nongelated aRBCs and G- facilitate nondisruptive cellular fixation. (iii) Extracellular cross- aRBCs (Fig. 2c and Supplementary Fig. 6). These results highlight linking was minimized to prevent cell-surface masking. Based on the intracellular hydrogelation technique enables facile prepara- these considerations, a photoactivated hydrogel system consisting tion of stable, cell-like constructs without masking cellular sur- of poly(ethylene glycol) diacrylate monomer (PEG-DA; M 700) faces. Using adherent HeLa cells, we further assessed the and 2-hydroxyl-4′-(2-hydroxyethoxy)-2-methylpropiophenone influence of intracellular hydrogelation on periplasma compo- photoinitiator (I2959) was employed. The materials are broadly nents and cellular cytoskeletons. In contrast to live cells that lost used in biomedical applications and have little reactivity with membrane ruffles upon incubation in phosphate-buffered saline 19,20 biological components . These hydrogel components were (PBS) for 4 h, both 4 wt% and 20 wt% GCs in PBS retained the introduced into cells through membrane poration with a single ruffled membrane features (Fig. 2d). Confocal microscopy of freeze–thaw cycle. Following a centrifugal wash to remove actin-GFP-transfected HeLa cells also showed that filamentous extracellular monomers and photoinitiators, the cells were irra- actin structures were observable in the GCs (Fig. 2e and Sup- diated with ultraviolet (UV) light for intracellular hydrogelation plementary Fig. 7A). Interestingly, in 20 and 40 wt% GCs, actin (Fig. 1a and Supplementary Fig. 1). To assess the feasibility of filaments could be observed 24 h after the gelation process, which intracellular gelation for cellular fixation, HeLa cells were first suggests that the denser hydrogel matrices could entrap these processed with different PEG-DA cross-linker densities ranging intracellular components and retard their depolymerization and from 4 to 40 wt%. The freeze–thaw treatment allowed PEG-DA dissipation (Supplementary Fig. 7A). We also observed that GCs monomers to penetrate into the intracellular domain efficiently, could retain their ruffled exterior over a long period of time and the collected cells had PEG-DA contents equivalent to the (Supplementary Fig. 7B, C), further illustrating that the synthetic input PEG-DA concentrations (Fig. 1b). Following UV irradia- hydrogel networks can substitute cytoskeleton in supporting these tion to the PEG-DA infused cells, no alteration to the cellular nanoscale membrane features. 2 NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 ARTICLE Intracellularly Cell Crosslinker infusion gelated cell (GC) Fluid membrane Photoactivated Membrane hydrogel crosslinking poration b c d Day 0 Day 30 Without freeze-thaw After freeze-thaw 0% 4% 10% 20% 40% Input PEG-DA (w/v %) 50 µm e f FITC Bright field Cell membrane Hydrogel Merged 10 µm 5 µm5 µm5 µm Fig. 1 Preparation and characterization of intracellularly gelated cells (GCs). a Hydrogel monomers and photoinitiators are infused into the intracellular domain of cells following transient membrane poration. UV-activated hydrogel cross-linking is then performed to stabilize the cell membrane interface. b Intracellular concentrations of PEG in cells before and after freeze–thaw treatments in gelation buffers containing different PEG-DA content. Error bars represent mean ± standard deviation, n = 3. c The Young’s moduli of the GCs prepared with different concentrations of hydrogel monomers were measured by atomic force microscopy. Error bars represent mean ± standard deviation, n = 64. d Bright-field microscopy of 4 wt% GCs and control cells suspended in PBS for 0 and 30 days. Scale bar = 50 μm. e Structure of 20 wt% GCs was visualized with fluorescein-diacrylate (green) for hydrogel labeling and DiD dye (red) for membrane staining. Scale bars = 5 μm. f 4 wt% GCs suspended in fluorescein solution showed that the GCs were impermeable to the dye. Scale bar = 10 μm Intracellular gelation preserves lipid order and fluidity.We noticeable alteration in membrane order and GP values, mem- next examined the influence of intracellular hydrogelation and brane order in GCs of different hydrogel densities was similar to hydrogel densities on plasma membrane fluidity and membrane that of live cells (Fig. 3e, f). These results demonstrate that the lipid order on GCs (Fig. 3a). Assessment of membrane fluidity gelation process has little influence on the phospholipid bilayer, by fluorescence recovery after photobleaching (FRAP) using a effectively retaining the membrane fluidity and membrane order lipophilic DiD dye showed that fluorescence recovery halftimes in the stabilized GCs. were similar among live cells and GCs of different cross-linking densities (Fig. 3b, c and Supplementary Fig. 8), indicating that Membrane proteins retain lateral mobility on GCs. To evaluate the intracellular hydrogel matrices did not influence membrane the mobility of membrane proteins on GCs, we first chose CD80 lipid fluidity regardless of the hydrogel content. Given that as the protein of interest given that CD80’s putative T-cell sti- membrane order is a critical biophysical parameter that influ- mulating functionality is highly dependent on its lateral mobility . ences the dynamics of membrane proteins, we adopted a Laurdan CD80-GFP mobility on GCs was first assessed with total internal dye staining approach to quantify membrane order on GCs . reflection fluorescence (TIRF) microscopy, which revealed Through multiphoton microscopy followed by image processing rapid, random movements of fluorescent punctates. In contrast, to analyze the polarity of the plasma membrane, we were able to fluorescent signals in glutaraldehyde-fixed cells appeared static distinguish the ordered plasma membrane in live HeLa cells and (Fig. 4a, Supplementary Fig. 9, and Supplementary Movies 1–5). derive the corresponding generalized polarization (GP) values by As live cells showed prominent cytoskeleton-directed protein 22,23 tracing the pixel intensities at the cellular periphery (Fig. 3d). movements owing to actin-mediated CD80 localization , cells Upon applying the technique to GCs and control cells subject to treated with a single freeze–thaw cycle were freshly prepared as a freeze–thaw and UV treatments in the absence of hydrogel, control (Supplementary Movies 6 and 7). Calculation of protein 2 −1 preservation of plasma membrane order by intracellular gelation diffusivity showed a mean value of 0.118 μm s for the control was confirmed. Whereas nongelated control cells showed cells, which is in accordance with prior studies on the passive NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications 3 Control cells 4% GC 10% GC 20% GC 40% GC PEG content (mg/mL) Young’s modulus (kPa) Control cells GCs ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 a b Gelated HeLa Avian Gelated Gelated HeLa (20%) HeLa cell erythrocyte avian erythrocyte (20%) (20%) in 1% SDS 2 µm 2 µm 5 µm 5 µm 5 µm 200 nm 200 nm 100 nm 100 nm 100 nm 100 nm 100 nm 100 nm 2 µm 200 nm HeLa Cell HeLa Cell in PBS Gelated HeLa (4%) Gelated HeLa (20%) 10 µm 10 µm 10 µm 10 µm 3 µm 3 µm 3 µm 3 µm Bright field GFP-beta-actin Bright field GFP-beta-actin 10 µm 10 µm 10 µm 10 µm Fig. 2 Structural examination of GCs. a Transmission electron microscopy (TEM) images show the cross-sectional structure of normal HeLa cells (left), 20 wt% gelated HeLa cells (middle), and the hydrogel matrix of 20 wt% gelated HeLa cells following solubilization by 1% SDS (right). (scale bars = 5 μmor 100 nm). Red arrows indicate cell membranes. b TEM cryosection images show the structure of avian erythrocytes and gelated avian erythrocytes (20 wt% PEG-DA) and c binding of hemagglutinating influenza viruses on the surfaces of gelated avian erythrocytes (scale bars= 2 μm or 200 nm). Red arrows indicate cell membranes. Yellow arrows indicate influenza viruses. d Cryogenic scanning electron microscopy (Cryo-SEM) images show the surface features of live HeLa cells, HeLa cells in PBS for 4 h, 4 wt% gelated HeLa cells in PBS, and 20 wt% gelated HeLa cells in PBS. Membrane ruffles were observed on the gelated cells. Scale bars = 10 μm (top row) and 3 μm (bottom row). e HeLa cells transfected with plasmids carrying actin-GFP and the corresponding 4 wt% GCs were imaged under bright-field and fluorescence microscopy. Preservation of actin filament was observed in the GCs. Red arrows indicate membrane ruffles, scale bars= 10 μm 4 NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications HeLa cell Gelated HeLa (4%) NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 ARTICLE a d Ordered channel Membrane lipid Ordered Disordered fluidity membrane membrane GP image Intensity histogram Laurdan Laurdan DiD lmax = 450 nm lmax = 500 nm Disordered channel b e f Photobleach Live cells Non-gelated cells Live cells Live cell Glut-fixed cell 60 4% GC Non-gelated cells 10% GC 20% GC 40% GC 4% GC 10% GC 4% GC 0 50 100 150 Time (s) 10% GC 20% GC 20% GC 40% GC 40 1.0 0.5 40% GC 0.0 –0.5 20 µm –1.0 GP value Fig. 3 Examination of lipid fluidity and membrane order on the plasma membrane of GCs. a A schematic illustration of the fluorescent probes used for assessing lipid fluidity (DiD) and membrane order (Laurdan). b Representative fluorescence recovery curves after photobleaching of DiD dye on HeLa cells, glutaraldehyde-fixed cells, and GCs of different hydrogel densities. c Halftimes of DiD dye fluorescence recovery following photobleaching for live HeLa cells, glutaraldehyde-fixed HeLa cells, and gelated HeLa cells with different hydrogel densities. No significant difference was observed between each of the GCs and live cells. Error bars represent mean ± standard deviation, (n = 10–15). d A flowchart illustrating the derivation of generalized polarization (GP) values from confocal images in the ordered and disordered channels. e Representative pseudo-colored GP-intensity-merged images of live HeLa cells, nongelated control cells, and GCs of different hydrogel densities. Ordered membrane domains are shown in orange. Scale bars= 20 μm. Color scale corresponds to GP value. f Histograms of GP values comparing the membrane order of the plasma membrane on live cells, nongelated control cells and GCs diffusion of transmembrane proteins . For the 4, 10, 20, and on the kinetics analysis, it was confirmed that GCs of varying 40 wt% GCs, the mean diffusivities were 0.118, 0.0966, 0.0967, densities possessed similar CD80-GFP recovery rates to those 2 −1 and 0.0711 μm s , respectively (Fig. 4b). No statistical sig- of live cells (Fig. 4e; Supplementary Fig. 10B, C). In contrast, nificance, however, was observed among the control cell and glutaraldehyde-fixed cells showed significantly reduced recovery the GCs. kinetics, highlighting intracellular hydrogelation as a unique Further assessment of protein mobility was performed with approach for preserving mobile membrane proteins. FRAP on adherent GCs. After photobleaching, rapid fluorescence To evaluate how the hydrogel density may influence the recovery further validated the fluidity of CD80-GFP on GCs mobility of different membrane proteins, HeLa cells transfected (Fig. 4c; Supplementary Fig. 10). Curiously, close examination with glycosylphosphatidylinositol (GPI)-anchored enhanced of the GCs by Z-stacked fluorescence microscopy revealed green fluorescence protein (EGFP-GPI), transferrin receptor distinctive fluorescent filaments consistent with the patterns of (TfR), GFP-tagged tyrosine protein kinase Lyn (Lyn-GFP), and actin cytoskeleton (Supplementary Fig. 10D). Given CD80’s GFP-tagged epidermal growth factor receptor (EGFR-GFP) 22,23 tendency to complex with actin , the observed filamentous were separately prepared for FRAP analysis (Fig. 4d). These patterns can be attributed to the actin/CD80 complexes. To proteins cover a broad range of differently sized cytoplasmic minimize interference by cytoskeleton-directed protein move- domains, with EGFP-GPI and TfR possessing no and small ments and intracellular protein trafficking in comparing protein cytoplasmic segments and Lyn-GFP and EGFR-GFP having mobility between GCs and live cells, a first-order kinetics large cytoplasmic regions. Examination with FRAP showed that equation was applied to the early time points of the fluorescence all the assessed proteins had significantly higher mobile fractions recovery curves to derive fluorescence recovery kinetics. Based in GCs as compared to glutaraldehyde-fixed cells (Fig. 4e and NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications 5 Live cell Glut-fixed 4% GC 10% GC 20% GC 40% GC –0.6 –0.4 –0.2 0.2 0.4 0.6 0.8 Intensity (%) Recovery halftime (s) Add Subtract Divide ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 a b c ns ns ns Live Cell 4% GC 10% GC ns **** –1 200 nm –2 Glut-fixed 20% GC 40% GC –3 –4 10 2 µm 2 µm 2 µm –5 Pre-bleach: –2s Post-bleach: 0s Post-bleach: 120s Time d ™ e Transferrin-Alexa Fluor 488 CD80-GFP EGFP-GPI –1 –1 GFP –2 –2 GFP EGFP-GPI –3 –3 TfR CD80-GFP GFP –4 –4 EGFR-GFP Lyn-GFP GFP 1.5 TfR Lyn-GFP EGFR-GFP –1 –1 –1 10 10 10 1.0 –2 –2 –2 10 10 10 0.5 –3 –3 –3 10 10 0.0 –4 –4 –4 10 10 Fig. 4 Examination of membrane protein lateral mobility on GCs. a Representative trajectories from CD80-GFP fluorescence tracking on different GCs and control cells examined by TIRF microscopy. b Diffusion coefficients of CD80-GFP were calculated from the TIRF fluorescence tracking data. Error bars represent geometric mean with 95% confidence interval. c Representative images showing recovery of CD80-GFP fluorescence in 4 wt% GC following photobleaching. Red rectangles indicate the photobleached area of interest. Scale bars = 2 μm. d A schematic illustration of membrane proteins with different sizes of intracellular domains, including CD80-GFP, EGFP-GPI, TfR, Lyn-GFP, and EGFR-GFP, which were assessed for their lateral mobilityon GCs. e Recovery kinetics of CD80-GFP, EGFP-GPI, TfR, Lyn-GFP and EGFR-GFP on GCs and control cells assessed by FRAP. For GCs with 4 to 20 wt% hydrogel densities, all examined membrane proteins had similar lateral mobility as on live cells. Error bars represent geometric mean with 95% confidence interval. f Membrane protein mobility on 40 wt% GCs relative to their corresponding mean mobility on live cells. Error bars represent geometric mean with 95% confidence interval (n = 7–12). Statistical analysis was performed using a two-tail Student t test, ****p<0.001 Supplementary Figs. 11–14). Notably, Lyn-GFP and EGFR-GFP multiple, mobile, membrane-bound lymphocyte activation sig- exhibited reduced relative mobility and mobile fractions at 40 wt nals , and replicating these biological features remains a primary % hydrogel cross-linking (Fig. 4f and Supplementary Figs. 13 and engineering objective in the development of artificial APC sys- 14). It can be reasoned that the dense hydrogel core at 40 wt% tems. In our system, we hypothesized that G-DCs could trigger cross-linking imposes higher drag to transmembrane proteins T-cell expansion through MHC class I-TCR and CD80-CD28 and may also entrap proteins with large cytoplasmic domains, interactions (Fig. 5a). To prepare G-DCs, DCs were first activated resulting in the reduced lateral mobility of Lyn-GFP and EGFR- by subjecting JAWSII murine DCs to SIINFEKL peptide pulsing GFP. For CD80-GFP, EGFP-GPI, and TfR, however, 40% and lipopolysaccharide stimulation. G-DCs were then prepared hydrogel core had little influence on their relative mobility. In using 4 wt% PEG-DA with both activated and nonactivated addition, at gelation densities between 4 and 20 wt%, recovery DCs (Supplementary Fig. 16A, B). DCs fixed with glutaraldehyde kinetics for all the examined membrane proteins were mostly were prepared as a control. Flow cytometric analysis of DC similar to those of live cells, indicating that sufficiently porous surface markers showed the expression of H-2K /SIINFEKL and hydrogel cores can be prepared for cellular stabilization with CD80 were largely similar before and after the gelation process minimal impact on membrane protein mobility. Upon long-term (Fig. 5b–e). Among the activated DC samples, expression of storage of GCs at 4 °C, the presence of mobile membrane proteins H-2K /SIINFEKL complexes in live and gelated DCs were 98.2% remained readily detectable (Supplementary Fig. 15), further and 89.5%, respectively (p > 0.05). Likewise, the live and gelated demonstrating the construct’s robust stability. DCs showed comparable CD80 levels at 79.3% and 63.4% respectively (p > 0.05). The nonactivated live and gelated DCs Gelated dendritic cells for antigen presentation. To highlight also shared similar basal levels of H-2K /SIINFEKL and CD80 the potential utility of GCs, we prepared gelated dendritic cells expression, demonstrating that the cell-surface protein signature (G-DCs) and assessed its antigen presenting capability. Effective was effectively preserved following the gelation process. The T-lymphocyte expansion by DCs hinges on the presence of interaction between G-DCs and T-lymphocytes was further 6 NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications Live cell Glut-fixed 4% GC 10% GC CD80-GFP 20% GC EGFP-GPI 40% GC TfR Lyn-GFP RGFR-GFP Live cell Glut-fixed 4% GC 10% GC Live cell 20% GC Glut-fixed 40% GC 4% GC 10% GC 20% GC 40% GC Live cell Glut-fixed 4% GC Live cell 10% GC Glut-fixed 20% GC 4% GC 40% GC 10% GC 20% GC 40% GC Live cell Glut-fixed 4% GC 10% GC 20% GC 40% GC Relative mobility at 40 wt% Diffusion coefficient 2 –1 (µm s ) Recovery kinetic (1/s) Recovery kinetic (1/s) Recovery kinetic (1/s) Recovery kinetic (1/s) Recovery kinetic (1/s) NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 ARTICLE CD28 TCR T cells CD80 MHC-I Gelated Immune stimulation T cell expansion antigen-presenting cell bc Glut-fixed Glut-fixed Live DC G-DC Live DC G-DC DC DC 29.3 % 25.4 % 0.23 % 0.9 % 0.21 % 0.1 % 98.2 % 89.5 % 0.23 % 79.3 % 63.4 % 0.15 % H-2K /SIINFEKL-APC CD80-APC d e Non-activated Non-activated Activated Activated 0 0 Live DC G-DC Glut-fixed Live DC G-DC Glut-fixed DC DC Day 1 Day 3 Day 5 Day 1 Day 3 Day 5 Proliferation Proliferation Fig. 5 Gelated dendritic cells (G-DCs) as artificial antigen presenting cells. a A schematic illustration showing the interaction between T cells and G-DCs. JAWSII cells were treated with LPS and pulsed with SIINFEKL peptides for activation. Both nonactivated and activated cells were gelated for surface marker comparison. No significant difference was observed between the expression of b, d H-2K /SIINFEKL complexes and c, e CD80 on the surfaces of live DCs and G-DCs. Unstained T-cell groups (gray) were used as a negative control. Error bars represent mean ± SEM, n = 3. f Co-culture of activated G-DCs and glutaraldehyde-fixed DCs with CFSE-stained OT-I-specific CD8+ T cells showed time-dependent T-cell expansion by the G-DCs but not by the glutaraldehyde-fixed cells. Each culture condition contained a fixed number of DCs at 8 × 10 per well with a T-cell/G-DC ratio of 3:1. Unstained control T cells are plotted in black as a reference examined by incubating CD8+ T cells derived from OT-I Gelated APCs dynamically interact with T cells. T-cell inter- transgenic mice with activated G-DCs. The G-DCs effectively action with activated G-DCs was visually examined through expanded the target T lymphocytes in a time- and cell ratio- confocal microscopy. Upon incubation, multiple T cells were dependent fashion (Fig. 5f and Supplementary Fig. 16C, D), observed to interact with G-DCs actively, forming cell–cell con- validating preservation of functional cell membrane interface jugates (Fig. 6a and Supplementary Movie 8). Time-lapse imaging by intracellular hydrogelation. revealed continuous morphological changes and movements NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications 7 Count b + G-DC % H-2K /SIINFEKL Non- cells Activated activated Count Glut-fixed DC % CD80 cells Non- Activated activated ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 a b G-DC Glut-fixed DC Control G-DC G-DC w/ T cells T cell T cell G-DC G-DC 10 µm 10 µm 10 µm 10 µm 10 µm 10 µm 10 µm 10 µm Fig. 6 Examination of G-DC/T-cell interaction and CD80 clustering. a Incubation of CSFE-labeled CD8+ T cells (green) derived from OT-I transgenic mice and SIINFEKL-pulsed G-DCs (red) showed dynamic interaction between the T cells and G-DCs under bright-field and confocal fluorescence microscopy. Red and white arrows indicate cell-cell conjugates. Scale bars = 10 μm. b Incubation of CD80-GFP-expressing, SIINFEKL-pulsed G-DCs (green) and OT-I CD8 T cells (red dashed-line circles) showed formation of CD80 clusters at the G-DC/T-cell junctions. White arrows indicate sites of CD80 clusters. Scale bars = 10 μm among the T cells in apposition to the G-DCs, and each cell membrane functionality, activated G-DCs and nongelated engagement event lasted approximately 2–3 h (Supplementary DCs were kept in PBS for 21 days at 4 °C. On day 21, the G-DCs, Movie 8). The interaction dynamic between the G-DCs and the which retained their spherical morphology (Supplementary T cells was similar to the many reports on live APCs and Fig. 16B), ably triggered antigen-specific T-cell expansion in vivo 26,27 T cells , and this long-lasting yet nonstatic engagement has (Fig. 7e, f). In contrast, nongelated DCs showed prominent dis- been deemed an essential process for immunological synapse integration and had significantly reduced capacity in expanding formation and T-cell activation. Between glutaraldehyde-fixed T cells. In addition, G-DCs mechanically disrupted by ultra- DCs and T cells, no interaction was observed. We also examined sonication resulted in reduced T-cell expansion, further corro- CD80 distribution on G-DCs upon T-cell engagement by incu- borating hydrogel’s function in maintaining plasma membrane bating CD80-GFP-transfected G-DCs with antigen-specific functionality. T cells. Following 3 h of incubation, fluorescent CD80 clusters were observed at G-DC/T-cell junctions (Fig. 6b and Supple- mentary Fig. 16E). Presence of these CD80 clusters attests to the Discussion fluid membrane interface on GCs and indicates that the T cells In summary, intracellular assembly of hydrogel polymers was were able to sample and recruit these co-stimulatory signals on made possible through photoactivated cross-linking, presenting the G-DC surfaces. With CD80 clusters being a hallmark of a unique cellular fixation strategy that seamlessly bridges the immunological synapse , their presence helps justify the promi- robustness of synthetic materials with the biochemical complexity nent T-cell expansion triggered by the G-DCs. of natural cells. In contrast to common fixation techniques based on chemical fixatives, intracellular hydrogelation avoids Gelated APCs for ex vivo and in vivo T-cell expansion.We cross-linking of membrane-bound components, preserving fluid then compared the antigen presentation capability of G-DCs and functional plasma membrane interfaces for biological to that of live DCs. An ex vivo T-cell proliferation assay using interactions. Several studies have previously examined de novo CFSE-labeled CD8+ OT-I T cells was first performed with generation of globular and filamentous hydrogels in cells to G-DCs, live DCs, and glutaraldehyde-fixed DCs derived from mimic RNA granules , stimulate the phase transition of RNA/ 30 31 the same cell source. Compared to live DCs that expanded the protein bodies , and induce cellular apoptosis . However, T cells by 78.7%, the G-DCs induced 56.2% of T-cell proliferation cellular fixation and cell membrane preservation were not while lacking active T-cell stimulating processes such as cytokine observed in these works. The present work differs from the secretion and microvilli formation . In contrast, negligible T-cell aforementioned approaches in that the rapid, photoactivated expansion was observed for the glutaraldehyde-fixed DCs and assembly of covalently bonded hydrogel networks permits fast nonactivated DC samples (Fig. 7a, b). The ability of G-DCs to cytosolic immobilization, thereby enabling cellular fixation while stimulate T cells was further assessed in vivo with mice adoptively obviating cellular reorganization and other cellular responses. transferred with CFSE-labeled CD8+ OT-I T cells. Twenty-four Future development of the gelated cellular systems could further hour following the adoptive transfer, mice were administered benefit from the growing arsenal of membrane manipulation 6 32 with 10 G-DCs, live DCs, or glutaraldehyde-fixed DCs. Notably, strategies for intracellular hydrogel delivery , which may be less glutaraldehyde-fixed DCs induced asphyxiation and mortality disruptive as compared to the freeze–thaw approach adopted in shortly after injection, whereas G-DCs and live DCs were well the present work. Toward biomimetic materials engineering, tolerated. Flow cytometric analysis of the splenocytes 3 days intracellular hydrogelation permits facile preparation of stable, following the DC injections revealed that both the activated cell-like constructs, offering a robust platform for device devel- G-DCs and live DCs induced substantial expansion of CD8+ opment. As cell membranes are a widely present interface with T cells (Fig. 7c, d). This T-cell stimulation was antigen-specific broad biological implications, the present system sees broad as nonactivated DCs and G-DCs failed to induce T-cell division. potential applications in biomembrane research and biomaterials To further highlight the role of hydrogel support in preserving the engineering. 8 NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 ARTICLE ac e G-DC Live DC Glut-fixed DC G-DC Live DC Day 21 Day 21 Homogenized G-DC DC G-DC 89.1 % 79.3 % 56.2 % 88.8 % 1.8 % 51.3 % 8.9 % 8.34 % Activated Activated 2.39 % 2.38 % 1.42 % 1.03 % 1.75 % Non- Non- activated activated Proliferation Proliferation Proliferation bd f Day 21 G-DC 80 G-DC 80 Day 21 DC Live DC Homogenized G-DC 60 60 60 40 40 20 20 Activated Non-activated Activated Non-activated Fig. 7 Expansion of antigen-specific T cells ex vivo and in vivo by G-DCs. a, b Antigen-specific CD8+ T-cell stimulation by G-DCs, live DCs, and glutaraldehyde-fixed DCs ex vivo. Error bars represent mean ± SEM. n = 3. c, d Antigen-specific CD8+ T-cell stimulation by G-DCs and live DCs in vivo. Error bars represent mean ± SEM. (n = 4 to 8). e, f G-DCs and non-gelated DCs were stored in PBS at 4 °C for 21 days (Day 21 G-DC and Day 21 DCs). In vivo antigen-specific CD8+ T-cell stimulation was examined with day 21 G-DC, day 21 DCs, and day 21 G-DC structurally disrupted by ultrasonication (homogenized G-DCs). T-cell expansion was monitored on day 6. CFSE-labeled T-cell groups (gray) are shown as negative controls. Error bars represent mean ± SEM, n = 3 Methods curve was prepared with serially diluted PEG-DA. The measured PEG-DA content Ethics statement. All animal experiments were carried out in strict accordance was then divided by the total volume of HeLa cells to determine the intracellular with the recommendations from the Guidebook for the Care and Use of Laboratory PEG-DA concentration. Animals (published by The Chinese Taipei Society of Laboratory Animal Sciences). The experiment protocol was approved by the Academia Sinica Institutional Atomic force microscopy and elastic moduli assessment. Elastic moduli were Animal Care & Utilization Committee, Academia Sinica, Taipei, Taiwan. assessed using a Zeiss axiovert microscope and analyzed using JPK NanoWizard 3 (JPK instrument, Berlin, Germany). Glass slides with different samples were Cell culture. HeLa cells, a human epithelial cell line (ATCC, CCL-2), were −1 attached to the tip of AFM cantilever with force constant of 0.08 N m and a grown in complete media (Eagle’s Minimum Essential Medium and 10% fetal ™ resonance frequency of 20 kHz (NANOSENSORS ). Elastic moduli of different bovine serum (FBS)). JAWSII cells, an immature murine dendritic cell line (ATCC, cellular samples were quantified by contact mode. The force scanning technique CRL-11904), were grown in complete media (alpha minimum essential medium was also used to generate high-resolution (64 × 64 points) topographical/elastic with ribonucleosides, deoxyribonucleosides, 4 mM L-glutamine, 1 mM sodium maps of the cells. −1 pyruvate, 5 ng ml murine granulocyte-macrophage colony-stimulating factor, and 20% FBS). Cell roundedness and fluorescence quantification. Cell circularity was measured and calculated by ImageJ software. A built-in option for analyzing roundness is Intracellular gelation of suspension and adherent cells. Gelation buffers were available. Briefly, after image files were imported into ImageJ software, and we TM first prepared by mixing protease inhibitors (Pierce Protease Inhibitor Mini chose the built-in options to enhance contrast, to analyze circularity, and to export Tablets; ThermoFisher), 1 wt% of 2-hydroxy-4′-(2-hydroxyethoxy)−2-methylpro- results for cell roundness. For quantification of fluorescence intensity, fluorescent piophenone (Irgacure D-2959; Sigma-Aldrich), and poly(ethylene glycol) diacrylate image files were processed in ZEN Imaging Software (Carl Zeiss). (PEG-DA; Mn = 700 Da; Sigma-Aldrich) ranging from 4 to 40 wt% in 10 mM phosphate buffer. For fluorescent labeling of the hydrogel network, the gelation buffers were supplemented with 0.05 wt% of fluorescein O,O’-diacrylate (Sigma- FITC-dye exclusion assay. Totally, 1×10 gelated HeLa cells (4 wt%) were sus- Aldrich). For cross-linking cells in suspension, adherent cells (i.e., HeLa and −1 pended in 500 μL of PBS solution containing 10 μgmL of FITC for 2 h. Totally, JAWSII cells) were detached using an enzyme-free cell dissociation buffer (Ther- 50 μL of samples were then added to a confocal dish and observed by a confocal moFisher). Cells in suspension were pelleted at 200×g and resuspended in desig- microscope. nated gelation buffers. For cross-linking adherent cells, cells grown on a tissue culture plate were washed with PBS and immersed in the designated gelation buffers. Immediately following the addition of gelation buffers, the cells were flash Hemagglutination of avian erythrocytes. A/PuertoRico/8/34(H1N1) was frozen in methanol precooled in a −80 °C freezer. After 10 min of freezing, the cells propagated in 10-day-old specific-pathogen-free (SPF) chicken embryos (JD-SPF were thawed in a 37 °C water bath. The suspension cells were pelleted at 200×g and Biotech, Miaoli, Taiwan) via the allantoic route. Native virions were then derived resuspended in PBS on a tissue culture plate, whereas adherent cells were washed by purifying the virus-containing allantoic fluid (AF) through 20–50% sucrose with PBS twice. The tissue culture plates were then placed in an ice bath, and the gradient solution. Avian erythrocytes were prepared from chicken whole blood cells were crosslinked with 365 nm UV wavelength for 10 min using a UV lamp upon removal of plasma and buffy coat following centrifugation at 200 × g. Gelated (UVP UVLMS-38 EL Series) placed 2 in. above the tissue culture plate. The avian erythrocytes were prepared using 20 wt% PEG-DA. Hemagglutination study resulting GCs were washed twice in PBS for further experiments. was performed by adding 10 virions to 1 mL of PBS solution containing 2% of avian erythrocytes. Presence of hemagglutination was monitored following 30 min Quantification of intracellular PEG-DA concentrations. Quantification of of incubation at room temperature. intracellular PEG-DA concentrations was performed using an iodine-based 33 6 quantification method . Briefly, following PEG-DA infusion, 1 × 10 HeLa cells were washed and suspended in PBS to 1 mL. The collected cells were then Transmission electron microscopy. Cellular samples were fixed using 2% sonicated in a bath sonicator for 1 min to release the entrapped PEG-DA, and glutaraldehyde in 0.1 M cacodylate buffer at pH 7.4 overnight at 4 °C. After the cellular debris was spun down via centrifugation at 3000×g for 5 min. The postfixation in 1% osmium tetroxide and pre-embedding staining with 1% uranyl supernatants were collected and mixed with BaCl and iodine solutions in an 8:2:1 acetate, tissue samples were dehydrated and embedded in Agar 100. Sections ratio. Following color development for 15 min, PEG-DA concentrations in the measuring 80 nm were then examined using an FEI Tecnai G2 TF20 Super TWIN samples were determined by measuring the light absorbance at 535 nm. A standard microscope equipped with a field emission gun. NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications 9 % Proliferation Count % Proliferation Count Count % Proliferation ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 −1 Cryogenic scanning electron microscopy. For cryogenic scanning electron days after seeding, the cells were treated with 1 μgmL of LPS for 16 h at 37 °C −1 microscopy (cryo-SEM) imaging, an FEI Quanta 200/Quorum PP2000TR FEI, and then pulsed with 10 μgml SIINFEKL OT-I peptide for 4 h in complete 2007 high-resolution SEM was used. Briefly, HeLa cells were seeded on Aclar media. Activated or nonactivated DCs were gelated and then resuspended in PBS embedding films for 24 h prior to PBS or gelation treatments. Before imaging, the with 10% FBS. FITC-conjugated hamster anti-mouse CD3 (BioLegend, #100306, samples were washed with PBS and suspended in RO water for freezing by liquid clone 145-2C11, 1:100), CD80 (eBioscience, #11–0801–81, clone Ly-53, 1:100), or nitrogen. The samples were then etched under vacuum and imaged at an accel- allophycocyanin-conjugated anti-MHC class I-SIINFEKL antibodies (eBioscience, eration voltage of 3 kV by cryo-SEM. #17-5743-80, clone25-D1.16, 1:100) were added and incubated with the cells at room temperature for 30 min in the dark. The cells were then washed twice, and the expression of surface markers was acquired by FACSCanto (BD Biosciences) Examination of membrane order in GCs. Examination of membrane lipid order and analyzed by FlowJo software (Tree Star). Gating strategies for all flow cyto- in GCs was carried out according to a previously described protocol . Briefly, metric analyses are shown in Supplementary Figure 17. Statistical analysis was −1 GCs and control cells were stained in media containing 100 μgml of Laurdan performed based on a two-tailed, unpaired t test using GraphPad Prism. dye for 1 h. The samples were subsequently washed with PBS, and the images for membrane order analysis were acquired by confocal microscopy. For the imaging setup, the excitation wavelength was set at 405 nm, and the detection T-cell isolation and fluorescence labeling. OT-I cells (CD8+ T cells specific for wavelengths were set at 440–460 nm for the ordered channel and 490–510 nm OVA257–264 peptide in the H2-K context) were isolated from OT-I transgenic for the disordered channel. All images were exported in the TIFF format and mice, which were a gift from Dr. Nan-Shih Liao from the Institute of Molecular saved as 32-bit grayscale image files with ImageJ. The custom-written macro Biology, Academia Sinica. After mice were sacrificed, their spleens were removed provided by Owen et al. was loaded into ImageJ to calculate GP values as well and placed into RPMI1640 complete medium with 10% FBS. In order to harvest as to create pseudo-colored GP-intensity-merged images and intensity histograms. single splenocytes, the spleens were tamped and strained with the tip of a 5 ml GP values were calculated according to the equation: GP value = (I – syringe against a sterile 40 μm nylon cell strainer (BD Biosciences Falcon, 440–460 nm I )/ (I + I ), where I indicate intensity of pixels. #352340). Splenocytes were incubated with BD Pharm Lyse lysing buffer (BD 490–510 nm 440–460 nm 490–510 nm Biosciences, # 555899) for 3 min to remove RBCs. OT-I cells were subsequently isolated from the splenocytes using a Mouse CD8a T Cell Isolation Kit (BD Fluorescence microscopy and FRAP analysis. Cell membrane was stained Biosciences, #19853 A). OT-I cells were stained with carboxyfluorescein diacetate by adding 10 μL of DiD dye solution (1,1′-Dioctadecyl-3,3,3′,3′-tetra- succinimidyl ester (CFSE) by incubating the cells with PBS containing 5 μMof −1 methylindodicarbocyanine; ThermoFisher Scientific) containing 5 μgmL of DiD CFSE (Sigma-Aldrich, #21888) at 37 °C for 5 min. The cells were washed three dye and 0.5% of DMSO to 200 μL of cell suspension. HeLa cells expressing EGFP- times with complete medium. CFSE-labeled cells were harvested for further GPI (Addgene, pCAG: GPI-GFP, #32601), CD80-GFP (Sino Biological Inc., experimental studies. pCMV3-mCD80-C-GFPSpark, MG50446-ACG), Transferrin Receptor (TfR) (Sino Biological Inc., pCMV3-hTfR-C-DDK (flag) tag, HG11020-CF), Lyn-GFP (Sino Biological Inc., pCMV3-hLyn-C-GFPSpark, HG10829-ACG), EGFR-GFP (Sino Examination of G-DC/T-cell interaction. For observation of G-DC/T-cell inter- Biological Inc., pCMV3-mEGFR-C-GFPSpark, MG51091-ACG), and GFP-beta- actions, adherent JAWSII DCs were gelated using 4 wt% PEG-DA and subse- TM actin (Sino Biological Inc., pCMV3-hbeta-actin-N-GFPSpark, HG10962-ANG) quently stained with CellTracker Deep Red dye (Molecular Probes) at 37 °C were prepared via transfection. Plasmids were transfected into cells with Lipo- for 30 min. Stained G-DCs were washed twice using PBS and resuspended in fectamine 3000 (Invitrogen) according to the manufacturer’s instruction. After RPMI1640 complete media supplemented with 10% FBS. CFSE-labeled OT-I cells 48 h, the cells were either gelated or treated with 2.5% glutaraldehyde for 10 min were subsequently added to the G-DCs. The interaction between G-DCs and OT-I prior to fluorescence microscopy or FRAP analysis. Fluorescence microscopy CD8 T cells was subsequently imaged using a confocal microscope (Zeiss LSM780 and FRAP analysis were carried out on a Zeiss LSM780 confocal microscope confocal microscope system, Zeiss) and analyzed using LSM Image Browser soft- (Carl Zeiss, Oberkochen, Germany) equipped with Plan-Apochromat 100×/1.4 oil ware (Zeiss). To examine CD80 clustering on G-DCs, JAWSII DCs were trans- objective. For FRAP analysis, adherent cells and adherent GCs were used rather fected with CD80-GFP plasmids using the TransIT-TKO transfection reagent than suspension cells to minimize artifacts due to random movements. An (Mirus, #2154) following a previously described protocol protocol . Briefly, objective heater was used to maintain samples at 37 °C. Images were collected with plasmids were prepared using a Qiagen Plasmid Midi kit (QIAGEN, #21243). a pinhole of 1.52 AU (1.1 μm section) for optimal signal intensity. The sample was Transfection mixtures consisting of 5 mL of serum-free DMEM, 20 µg of plasmids, first scanned three times with 5% of laser power to measure the fluorescence and 40 µL of transfection reagent were prepared and transfected into JAWSII DCs. intensity before photobleaching, followed by 500 iterative laser pulses at full power Following 4 h of incubation, an additional 10 mL of complete medium was added to photobleach a 27 nm × 6 nm rectangular area at the plasma membrane. Fluor- to the cells. 48 h after transfection, CD80-GFP-expressing JAWSII cells were escence recovery was monitored every 2 s for at least 2 min at 60 frames per second gelated with 4 wt% PEG-DA with and used for examining CD80 clustering upon until a plateau is reached. Fluorescence intensity vs. time was plotted for analyzing incubation with antigen-specific T cells. the fluorescence recovery. The mobile fraction was calculated based on the equa- tion (I − I )/(I − I ) × 100%, where I is the end value of the recovered fluores- E 0 I 0 E T proliferation assay ex vivo. CFSE-labeled OT-I cells were co-cultured with cence intensity, I is the first post-bleach fluorescence intensity, and I is the initial 0 I live DCs, G-DCs or glutaraldehyde-fixed DCs at different ratios. Co-cultured cells (prebleach) fluorescence intensity. The halftime of recovery (t ) is derived as the 1/2 in 96-well v-bottomed plates were cultured at 37 °C for indicated time periods. After time from the bleach to the time point where the fluorescence recovery reaches harvesting, cells were stained with allophycocyanin-conjugated rat anti-mouse 50% of the final recovery intensity. For the recovery rate of CD80-GFP, the rate CD8a antibodies (eBioscience, # 100712, Clone 53-6.7, 1:100) and analyzed by flow constant k was derived by converting the fluorescence recovery to a first-order cytometry. Proliferation analysis platform in FlowJo was used to analyze cell divi- elimination kinetics curve in which concentration A is calculated as I − I with (t) I (t) sion. For experiments involving stored G-DCs and DCs, G-DCs, and DCs were I being set as the first postbleach fluorescence intensity. The first 20 time points (0) stored in PBS at 4 °C for 21 days. Homogenized G-DCs were prepared by sonicating were used for calculating the recovery kinetics. Recovery kinetics were calculated 21-day-old G-DCs using a Fisher Scientific 150E Sonic Dismembrator at 80% power based on the equation ln[A ] = −kt + ln[A ], in which k = −(ln[A ]-ln[A ])/t. (t) (0) (t) (0) pulsed (3 s on/1 s off) for 1 min. The G-DC, DC, and glutaraldehyde-fixed DC samples were derived from the same cell source for each separate experiment. Protein tracking by TIRF microscopy. Movements of CD80-GFP were observed by TIRF microscopy using Leica TIRF MC inverted fluorescence microscope T proliferation assay in vivo. CFSE-labeled splenocytes were adoptively trans- equipped with HCX PL-APO 100× NA 1.46 Oil objective lens (Leica Microsystems, ferred via tail vein injections to 8-week-old C57BL/6 J mice at a cell number of Germany). Cell samples were loaded onto a glass bottom culture plate, and the 3.3 × 10 . Twenty-four hours after the adoptive transfer of OT-I cells, the mice samples were exposed to a 488-nm wavelength laser. The fluorescence image was were challenged with live DCs, G-DCs or Glut-fixed DCs at a cell number of 10 acquired using a hamamatsu EM-CCD camera (C9100-13) at a temporal resolution via tail vein injections. 3 days after the DC injections, splenocytes were harvested of 63 ms. All single-molecular experiments were performed at 37 °C. Protein from the mice and stained with allophycocyanin-conjugated rat anti-mouse CD8a movements were then analyzed using two-dimensional trajectories of CD80-GFP antibodies, followed by flow cytometry and FlowJo analysis. The animal protocol molecules in the plane of the basal membrane and were reconstructed by Imaris was approved by the Institutional Animal Care and Use Committee (IACUC) Image Analysis Software (Bitplane, Switzerland). The diffusion constants were at Academia Sinica. The G-DCs and DCs were derived from the same source of evaluated based on a previously described method . Briefly, the mean square activated or nonactivated DCs. displacement (MSD) was plotted from each trajectory against time (t). For each molecule, the slope of the first three time points in the MSD t plot was used to Reporting summary. Further information on experimental design is available in calculate the diffusion coefficient, D, according to the equation MSD = 4Dt. t→0 Statistical analysis was performed using one-way ANOVA with GraphPad Prism. the Nature Research Reporting Summary linked to this article. The F value is 5.158, and the degrees of freedom is 5. Data availability Dendritic cell preparation and analysis. For activation, JAWSII cells were seeded All relevant data are available from the authors and/or are included within the onto a 100 mm petri dish with 10 mL of media at a density of 10 per dish. Three manuscript and Supplementary Information. 10 NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 ARTICLE Received: 20 June 2018 Accepted: 18 February 2019 27. Bousso, P. & Robey, E. Dynamics of CD8+ T cell priming by dendritic cells in intact lymph nodes. Nat. Immunol. 4, 579–585 (2003). 28. Fisher, P. J., Bulur, P. A., Vuk-Pavlovic, S., Prendergast, F. G. & Dietz, A. B. Dendritic cell microvilli: a novel membrane structure associated with the multifocal synapse and T-cell clustering. Blood 112, 5037–5045 (2008). 29. Nakamura, H. et al. Intracellular production of hydrogels and synthetic RNA granules by multivalent molecular interactions. Nat. Mater. 17,79–89 References (2018). 1. Zhang, L. & Granick, S. Slaved diffusion in phospholipid bilayers. Proc. Natl 30. Shin, Y. et al. Spatiotemporal control of intracellular phase transitions using Acad. Sci. USA 102, 9118–9121 (2005). light-activated optodroplets. Cell 168, 159–171 (2017). 2. Grakoui, A. et al. The immunological synapse: a molecular machine 31. Yang, Z. M., Xu, K. M., Guo, Z. F., Guo, Z. H. & Xu, B. Intracellular enzymatic controlling T cell activation. Science 285, 221–227 (1999). formation of nanofibers results in hydrogelation and regulated cell death. 3. Bromley, S. K. et al. The immunological synapse and CD28–CD80 Adv. Mater. 19, 3152–3156 (2007). interactions. Nat. Immunol. 2, 1159–1166 (2001). 32. Stewart, M. P. et al. In vitro and ex vivo strategies for intracellular delivery. 4. Lin, C. C. & Anseth, K. S. Cell–cell communication mimicry with poly Nature 538, 183–192 (2016). (ethylene glycol) hydrogels for enhancing beta-cell function. Proc. Natl Acad. 33. Cheng, T. L., Chuang, K. H., Chen, B. M. & Roffler, S. R. Analytical Sci. USA 108, 6380–6385 (2011). measurement of PEGylated molecules. Bioconjugate Chem. 23, 881–899 5. Tang, J. A. et al. Therapeutic microparticles functionalized with biomimetic (2012). cardiac stem cell membranes and secretome. Nat. Commun. 8, 13724 (2017). 34. Awasthi, S. & Cox, R. A. Transfection of murine dendritic cell line (JAWS II) 6. Hu, C. M. J. et al. Nanoparticle biointerfacing by platelet membrane cloaking. by a nonviral transfection reagent. Biotechniques 35, 600–602 (2003). Nature 526, 118–121 (2015). 7. Hu, C. M. et al. Erythrocyte membrane-camouflaged polymeric nanoparticles as a biomimetic delivery platform. Proc. Natl Acad. Sci. USA 108, Acknowledgements 10980–10985 (2011). The authors acknowledge technical support from Common Equipment Core, Institute 8. Parodi, A. et al. Synthetic nanoparticles functionalized with biomimetic leukocyte of Biomedical Science, Academia Sinica, for the confocal microscopy image acquisition membranes possess cell-like functions. Nat. Nanotechnol. 8,61–68 (2013). and FRAP analysis, Yi-Ru Li for the technical support on atomic force microscopy, 9. Oelke, M. et al. Ex vivo induction and expansion of antigen-specific cytotoxic Kung-Hsuan Lin and Tzu-Ling Wu for the TIRF image acquisition, Yao-Kuan Huang for T cells by HLA-Ig-coated artificial antigen-presenting cells. Nat. Med. 9, the technical support on transmission electron microscopy. The authors acknowledge 619–624 (2003). funding support from the Academia Sinica Career Development Award (CDA-105-L06) 10. Kosmides, A. K. et al. Biomimetic biodegradable artificial antigen presenting and by the Ministry of Science and Technology, Taiwan (106-2119-M-001-010). with PD-1 blockade to treat melanoma cells synergize. Biomaterials 118, 16–26 (2017). 11. Cheung, A. S., Zhang, D. K. Y., Koshy, S. T. & Mooney, D. J. Scaffolds that Author contributions mimic antigen-presenting cells enable ex vivo expansion of primary T cells. J.C.L., C.Y.C., Y.I.C., H.W.C. and C.M.J.H. conceived the experimental designs. J.C.L., Nat. Biotechnol. 36, 160–169 (2018). C.Y.C., Y.I.C., J.Y.C., B.Y.Y., N.N.L., Z.S.F. and W.Y.C. performed the optimization and 12. Fadel, T. R. et al. A carbon nanotube-polymer composite for T-cell therapy. characterization of the intracellular hydrogelation protocol. C.Y.C., Y.I.C., C.L.L., B.Y.Y. Nat. Nanotechnol. 9, 639–647 (2014). and W.Y.C. performed the membrane fluidity analysis. J.C.L. and Y.H.L. performed the 13. Tanaka, M. & Sackmann, E. Polymer-supported membranes as models of immunological assays. J.C.L., C.Y.C., C.L.L. and C.M.J.H. prepared the paper. All authors the cell surface. Nature 437, 656–663 (2005). have read and approved the paper. 14. Pace, H. et al. Preserved transmembrane protein mobility in polymer- supported lipid bilayers derived from cell membranes. Anal. Chem. 87, 9194–9203 (2015). Additional information 15. Hardy, G. J., Nayak, R. & Zauscher, S. Model cell membranes: techniques Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- to form complex biomimetic supported lipid bilayers via vesicle fusion. 019-09049-5. Curr. Opin. Colloid Interface Sci. 18, 448–458 (2013). 16. Chiang, P. C., Tanady, K., Huang, L. T. & Chao, L. Rupturing giant plasma Competing interests: The authors declare no competing interests. membrane vesicles to form micron-sized supported cell plasma membranes with native transmembrane proteins. Sci. Rep. 7, 15139 (2017). Reprints and permission information is available online at http://npg.nature.com/ 17. Fletcher, D. A. & Mullins, R. D. Cell mechanics and the cytoskeleton. reprintsandpermissions/ Nature 463, 485–492 (2010). 18. Brizard, A. M. & Van Esch, J. H. Self-assembly approaches for the Journal peer review information: Nature Communications thanks the anonymous construction of cell architecture mimics. Soft Matter 5, 1320–1327 (2009). reviewers for their contribution to the peer review of this work. Peer reviewer reports are 19. Park, J. et al. Combination delivery of TGF-beta inhibitor and IL-2 by available. nanoscale liposomal polymeric gels enhances tumour immunotherapy. Nat. Mater. 11, 895–905 (2012). Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in 20. Underhill, G. H., Chen, A. A., Albrecht, D. R. & Bhatia, S. N. Assessment of published maps and institutional affiliations. hepatocellular function within PEG hydrogels. Biomaterials 28,256–270 (2007). 21. Owen, D. M., Rentero, C., Magenau, A., Abu-Siniyeh, A. & Gaus, K. Quantitative imaging of membrane lipid order in cells and organisms. Open Access This article is licensed under a Creative Commons Nat. Protoc. 7,24–35 (2011). Attribution 4.0 International License, which permits use, sharing, 22. Doty, R. T. & Clark, E. A. Two regions in the CD80 cytoplasmic tail regulate adaptation, distribution and reproduction in any medium or format, as long as you give CD80 redistribution and T cell costimulation. J. Immunol. 161, 2700–2707 appropriate credit to the original author(s) and the source, provide a link to the Creative (1998). Commons license, and indicate if changes were made. The images or other third party 23. Tseng, S. Y., Liu, M. L. & Dustin, M. L. CD80 cytoplasmic domain controls material in this article are included in the article’s Creative Commons license, unless localization of CD28, CTLA-4, and protein kinase C theta in the indicated otherwise in a credit line to the material. If material is not included in the immunological synapse. J. Immunol. 175, 7829–7836 (2005). article’s Creative Commons license and your intended use is not permitted by statutory 24. Tanaka, K. A. K. et al. Membrane molecules mobile even after chemical regulation or exceeds the permitted use, you will need to obtain permission directly from fixation. Nat. Methods 7, 865–866 (2010). 25. Chen, L. P. & Flies, D. B. Molecular mechanisms of T cell co-stimulation the copyright holder. To view a copy of this license, visit http://creativecommons.org/ and co-inhibition. Nat. Rev. Immunol. 13, 227–242 (2013). licenses/by/4.0/. 26. Underhill, D. M., Bassetti, M., Rudensky, A. & Aderem, A. Dynamic interactions of macrophages with T cells during antigen presentation. © The Author(s) 2019 J. Exp. Med. 190, 1909–1914 (1999). NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications 11 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nature Communications Springer Journals

Intracellular hydrogelation preserves fluid and functional cell membrane interfaces for biological interactions

Loading next page...
 
/lp/springer-journals/intracellular-hydrogelation-preserves-fluid-and-functional-cell-ZFyi9PQAmp

References (37)

Publisher
Springer Journals
Copyright
Copyright © 2019 by The Author(s)
Subject
Science, Humanities and Social Sciences, multidisciplinary; Science, Humanities and Social Sciences, multidisciplinary; Science, multidisciplinary
eISSN
2041-1723
DOI
10.1038/s41467-019-09049-5
Publisher site
See Article on Publisher Site

Abstract

ARTICLE https://doi.org/10.1038/s41467-019-09049-5 OPEN Intracellular hydrogelation preserves fluid and functional cell membrane interfaces for biological interactions 1 1 1 1 1 1 1,2 Jung-Chen Lin , Chen-Ying Chien , Chi-Long Lin , Bing-Yu Yao , Yuan-I Chen , Yu-Han Liu , Zih-Syun Fang , 1 1 1,2 2 1 Jui-Yi Chen , Wei-ya Chen , No-No Lee , Hui-Wen Chen & Che-Ming J. Hu Cell membranes are an intricate yet fragile interface that requires substrate support for stabilization. Upon cell death, disassembly of the cytoskeletal network deprives plasma membranes of mechanical support and leads to membrane rupture and disintegration. By assembling a network of synthetic hydrogel polymers inside the intracellular compartment using photo-activated crosslinking chemistry, we show that the fluid cell membrane can be preserved, resulting in intracellularly gelated cells with robust stability. Upon assessing several types of adherent and suspension cells over a range of hydrogel crosslinking densities, we validate retention of surface properties, membrane lipid fluidity, lipid order, and protein mobility on the gelated cells. Preservation of cell surface functions is further demonstrated with gelated antigen presenting cells, which engage with antigen-specific T lymphocytes and effectively promote cell expansion ex vivo and in vivo. The intracellular hydrogelation technique presents a versatile cell fixation approach adaptable for biomembrane studies and biomedical device construction. 1 2 Institute of Biomedical Sciences, Academia Sinica, Taipei 11574, Taiwan. Department of Veterinary Medicine, National Taiwan University, Taipei 10617, Taiwan. These authors contributed equally: Jung-Chen Lin, Chen-Ying Chien, Chi-Long Lin. Correspondence and requests for materials should be addressed to C.-M.H. (email: chu@ibms.sinica.edu.tw) NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications 1 1234567890():,; ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 he cell membrane is a fluid substrate that harbors a milieu morphology was observed (Supplementary Fig. 2). An evaluation of phospholipids, proteins, and glycans, which dynamically by atomic force microscopy, however, showed that the gelated Tchoreograph numerous biological interactions. The long- cells (GCs) exhibited increasing Young’s moduli that correlated standing fascination with the various biological functions of cell with the PEG-DA concentrations (Fig. 1c). Assessment of GC membranes has inspired model systems and cell-mimetic devices stability by microscopy showed no observable structural alter- 1–3 4,5 6–8 for biological studies , tissue engineering , drug delivery , nation over a 30-day observation period, whereas control cells 9–12 and immunoengineering . Toward replicating the cell mem- and non-crosslinked cells exhibited noticeable disintegration brane interface, synthetic bilayer lipid membranes and bio- within 3 days (Fig. 1d and Supplementary Fig. 3). To further conjugation strategies are commonly adopted in bottom-up confirm the assembly of hydrogel networks in the intracellular engineering of cell membrane mimics . Alternatively, top-down domain, fluorescein-diacrylate was added to the cross-linker approaches based on extraction and reconstitution of plasma mixture to covalently imbue the hydrogel network with green membranes of living cells are frequently applied to capture the fluorescence (Supplementary Fig. 1). Following membrane intricate cell-surface chemistries for biomimetic functionaliza- staining with a lipophilic DiD fluorophore, GCs showed dis- 6–8 tion . As antigen presentation, membrane fluidity, and mem- tinctive membranous and hydrogel components (Fig. 1e and brane sidedness are critical factors behind biomembrane Supplementary Fig. 4), displaying a structure reminiscent of functions and can be influenced by membrane translocation substrate-supported lipid membranes . Solubilization treatment processes, methods for harnessing this membranous component with sodium dodecyl sulfate was applied to examine the integrity continue to emerge with the aim to better study and utilize this of the gelated cytoplasm, and the fluorescent hydrogel matrices in 14–16 complex and delicate biological interface . GCs remained intact following membrane dissolution (Supple- To stabilize the fluid and functional plasma membranes and mentary Fig. 4). In a dye-exclusion study, 4 wt% GCs effectively decouple it from the dynamic state of living cells, we envision excluded a water-soluble fluorescein isothiocyanate (FITC) dye that a synthetic polymeric network can be constructed in the from entering the cytoplasm (Fig. 1f and Supplementary Fig. 4), cytoplasm to replace the cytoskeletal support for stabilizing cel- thereby confirming the plasma membrane integrity on GCs. We lular structures. Unlike endogenous cytoskeletons that are sus- also demonstrated that GCs could be stored by freezing and by ceptible to reorganization and disintegration upon perturbation lyophilization (Supplementary Fig. 5A). In addition, the intra- and cell death , a synthetic substrate scaffold can stably support cellular gelation process was applied to adherent HeLa cells, the cell membrane interface for subsequent applications. As effectively preserving the cells’ adherent property and elongated the mechanical property of cytoskeletons has drawn comparisons structures (Supplementary Fig. 5B). 17,18 to hydrogels , a cellular fixation approach mediated by intracellular assembly of hydrogel monomers is herein developed. We demonstrate that the intracellular hydrogelation technique Intracellular gelation preserves cellular features. Examination effectively preserves cellular morphology, lipid order, membrane of the cell membrane interface and the cytoplasmic hydrogel protein mobility, and biological functions of the plasma mem- matrix on GCs was performed by transmission electron micro- brane, giving rise to cell-like constructs with extraordinary sta- scopy (TEM). In comparison to control cells, GCs possessed a bility. In addition, a highly functional artificial antigen presenting perforated, hydrogel-filled interior. Treatment by detergent cell (APC) is prepared with the gelated system to highlight the stripped GCs of their membranous exterior, leaving behind platform’s utility for biomedical applications. nondissolvable hydrogel matrices (Fig. 2a). To better visualize the membrane interface on GCs, intracellular hydrogelation was applied to avian red blood cells (aRBCs), which are nucleated cells Results devoid of organelles. As hemoglobins were removed during the Intracellular hydrogelation by photoactivated cross-linking. gelation process, gelated aRBCs (G-aRBCs) exhibited a clear Three criteria were considered to establish the intracellular membrane boundary encircling a perforated nucleus (Fig. 2b and hydrogelation technique: (i) Hydrophilic cross-linking monomers Supplementary Fig. 6). Notably, the addition of a hemaggluti- with a low-molecular weight were used to facilitate cytoplasmic nating influenza virus to the G-aRBCs induced direct agglutina- permeation and minimize membrane partitioning. (ii) Cross- tion (Supplementary Fig. 6), and TEM cryosections showed linking chemistry with low-protein reactivity was adopted to similar binding patterns between nongelated aRBCs and G- facilitate nondisruptive cellular fixation. (iii) Extracellular cross- aRBCs (Fig. 2c and Supplementary Fig. 6). These results highlight linking was minimized to prevent cell-surface masking. Based on the intracellular hydrogelation technique enables facile prepara- these considerations, a photoactivated hydrogel system consisting tion of stable, cell-like constructs without masking cellular sur- of poly(ethylene glycol) diacrylate monomer (PEG-DA; M 700) faces. Using adherent HeLa cells, we further assessed the and 2-hydroxyl-4′-(2-hydroxyethoxy)-2-methylpropiophenone influence of intracellular hydrogelation on periplasma compo- photoinitiator (I2959) was employed. The materials are broadly nents and cellular cytoskeletons. In contrast to live cells that lost used in biomedical applications and have little reactivity with membrane ruffles upon incubation in phosphate-buffered saline 19,20 biological components . These hydrogel components were (PBS) for 4 h, both 4 wt% and 20 wt% GCs in PBS retained the introduced into cells through membrane poration with a single ruffled membrane features (Fig. 2d). Confocal microscopy of freeze–thaw cycle. Following a centrifugal wash to remove actin-GFP-transfected HeLa cells also showed that filamentous extracellular monomers and photoinitiators, the cells were irra- actin structures were observable in the GCs (Fig. 2e and Sup- diated with ultraviolet (UV) light for intracellular hydrogelation plementary Fig. 7A). Interestingly, in 20 and 40 wt% GCs, actin (Fig. 1a and Supplementary Fig. 1). To assess the feasibility of filaments could be observed 24 h after the gelation process, which intracellular gelation for cellular fixation, HeLa cells were first suggests that the denser hydrogel matrices could entrap these processed with different PEG-DA cross-linker densities ranging intracellular components and retard their depolymerization and from 4 to 40 wt%. The freeze–thaw treatment allowed PEG-DA dissipation (Supplementary Fig. 7A). We also observed that GCs monomers to penetrate into the intracellular domain efficiently, could retain their ruffled exterior over a long period of time and the collected cells had PEG-DA contents equivalent to the (Supplementary Fig. 7B, C), further illustrating that the synthetic input PEG-DA concentrations (Fig. 1b). Following UV irradia- hydrogel networks can substitute cytoskeleton in supporting these tion to the PEG-DA infused cells, no alteration to the cellular nanoscale membrane features. 2 NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 ARTICLE Intracellularly Cell Crosslinker infusion gelated cell (GC) Fluid membrane Photoactivated Membrane hydrogel crosslinking poration b c d Day 0 Day 30 Without freeze-thaw After freeze-thaw 0% 4% 10% 20% 40% Input PEG-DA (w/v %) 50 µm e f FITC Bright field Cell membrane Hydrogel Merged 10 µm 5 µm5 µm5 µm Fig. 1 Preparation and characterization of intracellularly gelated cells (GCs). a Hydrogel monomers and photoinitiators are infused into the intracellular domain of cells following transient membrane poration. UV-activated hydrogel cross-linking is then performed to stabilize the cell membrane interface. b Intracellular concentrations of PEG in cells before and after freeze–thaw treatments in gelation buffers containing different PEG-DA content. Error bars represent mean ± standard deviation, n = 3. c The Young’s moduli of the GCs prepared with different concentrations of hydrogel monomers were measured by atomic force microscopy. Error bars represent mean ± standard deviation, n = 64. d Bright-field microscopy of 4 wt% GCs and control cells suspended in PBS for 0 and 30 days. Scale bar = 50 μm. e Structure of 20 wt% GCs was visualized with fluorescein-diacrylate (green) for hydrogel labeling and DiD dye (red) for membrane staining. Scale bars = 5 μm. f 4 wt% GCs suspended in fluorescein solution showed that the GCs were impermeable to the dye. Scale bar = 10 μm Intracellular gelation preserves lipid order and fluidity.We noticeable alteration in membrane order and GP values, mem- next examined the influence of intracellular hydrogelation and brane order in GCs of different hydrogel densities was similar to hydrogel densities on plasma membrane fluidity and membrane that of live cells (Fig. 3e, f). These results demonstrate that the lipid order on GCs (Fig. 3a). Assessment of membrane fluidity gelation process has little influence on the phospholipid bilayer, by fluorescence recovery after photobleaching (FRAP) using a effectively retaining the membrane fluidity and membrane order lipophilic DiD dye showed that fluorescence recovery halftimes in the stabilized GCs. were similar among live cells and GCs of different cross-linking densities (Fig. 3b, c and Supplementary Fig. 8), indicating that Membrane proteins retain lateral mobility on GCs. To evaluate the intracellular hydrogel matrices did not influence membrane the mobility of membrane proteins on GCs, we first chose CD80 lipid fluidity regardless of the hydrogel content. Given that as the protein of interest given that CD80’s putative T-cell sti- membrane order is a critical biophysical parameter that influ- mulating functionality is highly dependent on its lateral mobility . ences the dynamics of membrane proteins, we adopted a Laurdan CD80-GFP mobility on GCs was first assessed with total internal dye staining approach to quantify membrane order on GCs . reflection fluorescence (TIRF) microscopy, which revealed Through multiphoton microscopy followed by image processing rapid, random movements of fluorescent punctates. In contrast, to analyze the polarity of the plasma membrane, we were able to fluorescent signals in glutaraldehyde-fixed cells appeared static distinguish the ordered plasma membrane in live HeLa cells and (Fig. 4a, Supplementary Fig. 9, and Supplementary Movies 1–5). derive the corresponding generalized polarization (GP) values by As live cells showed prominent cytoskeleton-directed protein 22,23 tracing the pixel intensities at the cellular periphery (Fig. 3d). movements owing to actin-mediated CD80 localization , cells Upon applying the technique to GCs and control cells subject to treated with a single freeze–thaw cycle were freshly prepared as a freeze–thaw and UV treatments in the absence of hydrogel, control (Supplementary Movies 6 and 7). Calculation of protein 2 −1 preservation of plasma membrane order by intracellular gelation diffusivity showed a mean value of 0.118 μm s for the control was confirmed. Whereas nongelated control cells showed cells, which is in accordance with prior studies on the passive NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications 3 Control cells 4% GC 10% GC 20% GC 40% GC PEG content (mg/mL) Young’s modulus (kPa) Control cells GCs ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 a b Gelated HeLa Avian Gelated Gelated HeLa (20%) HeLa cell erythrocyte avian erythrocyte (20%) (20%) in 1% SDS 2 µm 2 µm 5 µm 5 µm 5 µm 200 nm 200 nm 100 nm 100 nm 100 nm 100 nm 100 nm 100 nm 2 µm 200 nm HeLa Cell HeLa Cell in PBS Gelated HeLa (4%) Gelated HeLa (20%) 10 µm 10 µm 10 µm 10 µm 3 µm 3 µm 3 µm 3 µm Bright field GFP-beta-actin Bright field GFP-beta-actin 10 µm 10 µm 10 µm 10 µm Fig. 2 Structural examination of GCs. a Transmission electron microscopy (TEM) images show the cross-sectional structure of normal HeLa cells (left), 20 wt% gelated HeLa cells (middle), and the hydrogel matrix of 20 wt% gelated HeLa cells following solubilization by 1% SDS (right). (scale bars = 5 μmor 100 nm). Red arrows indicate cell membranes. b TEM cryosection images show the structure of avian erythrocytes and gelated avian erythrocytes (20 wt% PEG-DA) and c binding of hemagglutinating influenza viruses on the surfaces of gelated avian erythrocytes (scale bars= 2 μm or 200 nm). Red arrows indicate cell membranes. Yellow arrows indicate influenza viruses. d Cryogenic scanning electron microscopy (Cryo-SEM) images show the surface features of live HeLa cells, HeLa cells in PBS for 4 h, 4 wt% gelated HeLa cells in PBS, and 20 wt% gelated HeLa cells in PBS. Membrane ruffles were observed on the gelated cells. Scale bars = 10 μm (top row) and 3 μm (bottom row). e HeLa cells transfected with plasmids carrying actin-GFP and the corresponding 4 wt% GCs were imaged under bright-field and fluorescence microscopy. Preservation of actin filament was observed in the GCs. Red arrows indicate membrane ruffles, scale bars= 10 μm 4 NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications HeLa cell Gelated HeLa (4%) NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 ARTICLE a d Ordered channel Membrane lipid Ordered Disordered fluidity membrane membrane GP image Intensity histogram Laurdan Laurdan DiD lmax = 450 nm lmax = 500 nm Disordered channel b e f Photobleach Live cells Non-gelated cells Live cells Live cell Glut-fixed cell 60 4% GC Non-gelated cells 10% GC 20% GC 40% GC 4% GC 10% GC 4% GC 0 50 100 150 Time (s) 10% GC 20% GC 20% GC 40% GC 40 1.0 0.5 40% GC 0.0 –0.5 20 µm –1.0 GP value Fig. 3 Examination of lipid fluidity and membrane order on the plasma membrane of GCs. a A schematic illustration of the fluorescent probes used for assessing lipid fluidity (DiD) and membrane order (Laurdan). b Representative fluorescence recovery curves after photobleaching of DiD dye on HeLa cells, glutaraldehyde-fixed cells, and GCs of different hydrogel densities. c Halftimes of DiD dye fluorescence recovery following photobleaching for live HeLa cells, glutaraldehyde-fixed HeLa cells, and gelated HeLa cells with different hydrogel densities. No significant difference was observed between each of the GCs and live cells. Error bars represent mean ± standard deviation, (n = 10–15). d A flowchart illustrating the derivation of generalized polarization (GP) values from confocal images in the ordered and disordered channels. e Representative pseudo-colored GP-intensity-merged images of live HeLa cells, nongelated control cells, and GCs of different hydrogel densities. Ordered membrane domains are shown in orange. Scale bars= 20 μm. Color scale corresponds to GP value. f Histograms of GP values comparing the membrane order of the plasma membrane on live cells, nongelated control cells and GCs diffusion of transmembrane proteins . For the 4, 10, 20, and on the kinetics analysis, it was confirmed that GCs of varying 40 wt% GCs, the mean diffusivities were 0.118, 0.0966, 0.0967, densities possessed similar CD80-GFP recovery rates to those 2 −1 and 0.0711 μm s , respectively (Fig. 4b). No statistical sig- of live cells (Fig. 4e; Supplementary Fig. 10B, C). In contrast, nificance, however, was observed among the control cell and glutaraldehyde-fixed cells showed significantly reduced recovery the GCs. kinetics, highlighting intracellular hydrogelation as a unique Further assessment of protein mobility was performed with approach for preserving mobile membrane proteins. FRAP on adherent GCs. After photobleaching, rapid fluorescence To evaluate how the hydrogel density may influence the recovery further validated the fluidity of CD80-GFP on GCs mobility of different membrane proteins, HeLa cells transfected (Fig. 4c; Supplementary Fig. 10). Curiously, close examination with glycosylphosphatidylinositol (GPI)-anchored enhanced of the GCs by Z-stacked fluorescence microscopy revealed green fluorescence protein (EGFP-GPI), transferrin receptor distinctive fluorescent filaments consistent with the patterns of (TfR), GFP-tagged tyrosine protein kinase Lyn (Lyn-GFP), and actin cytoskeleton (Supplementary Fig. 10D). Given CD80’s GFP-tagged epidermal growth factor receptor (EGFR-GFP) 22,23 tendency to complex with actin , the observed filamentous were separately prepared for FRAP analysis (Fig. 4d). These patterns can be attributed to the actin/CD80 complexes. To proteins cover a broad range of differently sized cytoplasmic minimize interference by cytoskeleton-directed protein move- domains, with EGFP-GPI and TfR possessing no and small ments and intracellular protein trafficking in comparing protein cytoplasmic segments and Lyn-GFP and EGFR-GFP having mobility between GCs and live cells, a first-order kinetics large cytoplasmic regions. Examination with FRAP showed that equation was applied to the early time points of the fluorescence all the assessed proteins had significantly higher mobile fractions recovery curves to derive fluorescence recovery kinetics. Based in GCs as compared to glutaraldehyde-fixed cells (Fig. 4e and NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications 5 Live cell Glut-fixed 4% GC 10% GC 20% GC 40% GC –0.6 –0.4 –0.2 0.2 0.4 0.6 0.8 Intensity (%) Recovery halftime (s) Add Subtract Divide ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 a b c ns ns ns Live Cell 4% GC 10% GC ns **** –1 200 nm –2 Glut-fixed 20% GC 40% GC –3 –4 10 2 µm 2 µm 2 µm –5 Pre-bleach: –2s Post-bleach: 0s Post-bleach: 120s Time d ™ e Transferrin-Alexa Fluor 488 CD80-GFP EGFP-GPI –1 –1 GFP –2 –2 GFP EGFP-GPI –3 –3 TfR CD80-GFP GFP –4 –4 EGFR-GFP Lyn-GFP GFP 1.5 TfR Lyn-GFP EGFR-GFP –1 –1 –1 10 10 10 1.0 –2 –2 –2 10 10 10 0.5 –3 –3 –3 10 10 0.0 –4 –4 –4 10 10 Fig. 4 Examination of membrane protein lateral mobility on GCs. a Representative trajectories from CD80-GFP fluorescence tracking on different GCs and control cells examined by TIRF microscopy. b Diffusion coefficients of CD80-GFP were calculated from the TIRF fluorescence tracking data. Error bars represent geometric mean with 95% confidence interval. c Representative images showing recovery of CD80-GFP fluorescence in 4 wt% GC following photobleaching. Red rectangles indicate the photobleached area of interest. Scale bars = 2 μm. d A schematic illustration of membrane proteins with different sizes of intracellular domains, including CD80-GFP, EGFP-GPI, TfR, Lyn-GFP, and EGFR-GFP, which were assessed for their lateral mobilityon GCs. e Recovery kinetics of CD80-GFP, EGFP-GPI, TfR, Lyn-GFP and EGFR-GFP on GCs and control cells assessed by FRAP. For GCs with 4 to 20 wt% hydrogel densities, all examined membrane proteins had similar lateral mobility as on live cells. Error bars represent geometric mean with 95% confidence interval. f Membrane protein mobility on 40 wt% GCs relative to their corresponding mean mobility on live cells. Error bars represent geometric mean with 95% confidence interval (n = 7–12). Statistical analysis was performed using a two-tail Student t test, ****p<0.001 Supplementary Figs. 11–14). Notably, Lyn-GFP and EGFR-GFP multiple, mobile, membrane-bound lymphocyte activation sig- exhibited reduced relative mobility and mobile fractions at 40 wt nals , and replicating these biological features remains a primary % hydrogel cross-linking (Fig. 4f and Supplementary Figs. 13 and engineering objective in the development of artificial APC sys- 14). It can be reasoned that the dense hydrogel core at 40 wt% tems. In our system, we hypothesized that G-DCs could trigger cross-linking imposes higher drag to transmembrane proteins T-cell expansion through MHC class I-TCR and CD80-CD28 and may also entrap proteins with large cytoplasmic domains, interactions (Fig. 5a). To prepare G-DCs, DCs were first activated resulting in the reduced lateral mobility of Lyn-GFP and EGFR- by subjecting JAWSII murine DCs to SIINFEKL peptide pulsing GFP. For CD80-GFP, EGFP-GPI, and TfR, however, 40% and lipopolysaccharide stimulation. G-DCs were then prepared hydrogel core had little influence on their relative mobility. In using 4 wt% PEG-DA with both activated and nonactivated addition, at gelation densities between 4 and 20 wt%, recovery DCs (Supplementary Fig. 16A, B). DCs fixed with glutaraldehyde kinetics for all the examined membrane proteins were mostly were prepared as a control. Flow cytometric analysis of DC similar to those of live cells, indicating that sufficiently porous surface markers showed the expression of H-2K /SIINFEKL and hydrogel cores can be prepared for cellular stabilization with CD80 were largely similar before and after the gelation process minimal impact on membrane protein mobility. Upon long-term (Fig. 5b–e). Among the activated DC samples, expression of storage of GCs at 4 °C, the presence of mobile membrane proteins H-2K /SIINFEKL complexes in live and gelated DCs were 98.2% remained readily detectable (Supplementary Fig. 15), further and 89.5%, respectively (p > 0.05). Likewise, the live and gelated demonstrating the construct’s robust stability. DCs showed comparable CD80 levels at 79.3% and 63.4% respectively (p > 0.05). The nonactivated live and gelated DCs Gelated dendritic cells for antigen presentation. To highlight also shared similar basal levels of H-2K /SIINFEKL and CD80 the potential utility of GCs, we prepared gelated dendritic cells expression, demonstrating that the cell-surface protein signature (G-DCs) and assessed its antigen presenting capability. Effective was effectively preserved following the gelation process. The T-lymphocyte expansion by DCs hinges on the presence of interaction between G-DCs and T-lymphocytes was further 6 NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications Live cell Glut-fixed 4% GC 10% GC CD80-GFP 20% GC EGFP-GPI 40% GC TfR Lyn-GFP RGFR-GFP Live cell Glut-fixed 4% GC 10% GC Live cell 20% GC Glut-fixed 40% GC 4% GC 10% GC 20% GC 40% GC Live cell Glut-fixed 4% GC Live cell 10% GC Glut-fixed 20% GC 4% GC 40% GC 10% GC 20% GC 40% GC Live cell Glut-fixed 4% GC 10% GC 20% GC 40% GC Relative mobility at 40 wt% Diffusion coefficient 2 –1 (µm s ) Recovery kinetic (1/s) Recovery kinetic (1/s) Recovery kinetic (1/s) Recovery kinetic (1/s) Recovery kinetic (1/s) NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 ARTICLE CD28 TCR T cells CD80 MHC-I Gelated Immune stimulation T cell expansion antigen-presenting cell bc Glut-fixed Glut-fixed Live DC G-DC Live DC G-DC DC DC 29.3 % 25.4 % 0.23 % 0.9 % 0.21 % 0.1 % 98.2 % 89.5 % 0.23 % 79.3 % 63.4 % 0.15 % H-2K /SIINFEKL-APC CD80-APC d e Non-activated Non-activated Activated Activated 0 0 Live DC G-DC Glut-fixed Live DC G-DC Glut-fixed DC DC Day 1 Day 3 Day 5 Day 1 Day 3 Day 5 Proliferation Proliferation Fig. 5 Gelated dendritic cells (G-DCs) as artificial antigen presenting cells. a A schematic illustration showing the interaction between T cells and G-DCs. JAWSII cells were treated with LPS and pulsed with SIINFEKL peptides for activation. Both nonactivated and activated cells were gelated for surface marker comparison. No significant difference was observed between the expression of b, d H-2K /SIINFEKL complexes and c, e CD80 on the surfaces of live DCs and G-DCs. Unstained T-cell groups (gray) were used as a negative control. Error bars represent mean ± SEM, n = 3. f Co-culture of activated G-DCs and glutaraldehyde-fixed DCs with CFSE-stained OT-I-specific CD8+ T cells showed time-dependent T-cell expansion by the G-DCs but not by the glutaraldehyde-fixed cells. Each culture condition contained a fixed number of DCs at 8 × 10 per well with a T-cell/G-DC ratio of 3:1. Unstained control T cells are plotted in black as a reference examined by incubating CD8+ T cells derived from OT-I Gelated APCs dynamically interact with T cells. T-cell inter- transgenic mice with activated G-DCs. The G-DCs effectively action with activated G-DCs was visually examined through expanded the target T lymphocytes in a time- and cell ratio- confocal microscopy. Upon incubation, multiple T cells were dependent fashion (Fig. 5f and Supplementary Fig. 16C, D), observed to interact with G-DCs actively, forming cell–cell con- validating preservation of functional cell membrane interface jugates (Fig. 6a and Supplementary Movie 8). Time-lapse imaging by intracellular hydrogelation. revealed continuous morphological changes and movements NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications 7 Count b + G-DC % H-2K /SIINFEKL Non- cells Activated activated Count Glut-fixed DC % CD80 cells Non- Activated activated ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 a b G-DC Glut-fixed DC Control G-DC G-DC w/ T cells T cell T cell G-DC G-DC 10 µm 10 µm 10 µm 10 µm 10 µm 10 µm 10 µm 10 µm Fig. 6 Examination of G-DC/T-cell interaction and CD80 clustering. a Incubation of CSFE-labeled CD8+ T cells (green) derived from OT-I transgenic mice and SIINFEKL-pulsed G-DCs (red) showed dynamic interaction between the T cells and G-DCs under bright-field and confocal fluorescence microscopy. Red and white arrows indicate cell-cell conjugates. Scale bars = 10 μm. b Incubation of CD80-GFP-expressing, SIINFEKL-pulsed G-DCs (green) and OT-I CD8 T cells (red dashed-line circles) showed formation of CD80 clusters at the G-DC/T-cell junctions. White arrows indicate sites of CD80 clusters. Scale bars = 10 μm among the T cells in apposition to the G-DCs, and each cell membrane functionality, activated G-DCs and nongelated engagement event lasted approximately 2–3 h (Supplementary DCs were kept in PBS for 21 days at 4 °C. On day 21, the G-DCs, Movie 8). The interaction dynamic between the G-DCs and the which retained their spherical morphology (Supplementary T cells was similar to the many reports on live APCs and Fig. 16B), ably triggered antigen-specific T-cell expansion in vivo 26,27 T cells , and this long-lasting yet nonstatic engagement has (Fig. 7e, f). In contrast, nongelated DCs showed prominent dis- been deemed an essential process for immunological synapse integration and had significantly reduced capacity in expanding formation and T-cell activation. Between glutaraldehyde-fixed T cells. In addition, G-DCs mechanically disrupted by ultra- DCs and T cells, no interaction was observed. We also examined sonication resulted in reduced T-cell expansion, further corro- CD80 distribution on G-DCs upon T-cell engagement by incu- borating hydrogel’s function in maintaining plasma membrane bating CD80-GFP-transfected G-DCs with antigen-specific functionality. T cells. Following 3 h of incubation, fluorescent CD80 clusters were observed at G-DC/T-cell junctions (Fig. 6b and Supple- mentary Fig. 16E). Presence of these CD80 clusters attests to the Discussion fluid membrane interface on GCs and indicates that the T cells In summary, intracellular assembly of hydrogel polymers was were able to sample and recruit these co-stimulatory signals on made possible through photoactivated cross-linking, presenting the G-DC surfaces. With CD80 clusters being a hallmark of a unique cellular fixation strategy that seamlessly bridges the immunological synapse , their presence helps justify the promi- robustness of synthetic materials with the biochemical complexity nent T-cell expansion triggered by the G-DCs. of natural cells. In contrast to common fixation techniques based on chemical fixatives, intracellular hydrogelation avoids Gelated APCs for ex vivo and in vivo T-cell expansion.We cross-linking of membrane-bound components, preserving fluid then compared the antigen presentation capability of G-DCs and functional plasma membrane interfaces for biological to that of live DCs. An ex vivo T-cell proliferation assay using interactions. Several studies have previously examined de novo CFSE-labeled CD8+ OT-I T cells was first performed with generation of globular and filamentous hydrogels in cells to G-DCs, live DCs, and glutaraldehyde-fixed DCs derived from mimic RNA granules , stimulate the phase transition of RNA/ 30 31 the same cell source. Compared to live DCs that expanded the protein bodies , and induce cellular apoptosis . However, T cells by 78.7%, the G-DCs induced 56.2% of T-cell proliferation cellular fixation and cell membrane preservation were not while lacking active T-cell stimulating processes such as cytokine observed in these works. The present work differs from the secretion and microvilli formation . In contrast, negligible T-cell aforementioned approaches in that the rapid, photoactivated expansion was observed for the glutaraldehyde-fixed DCs and assembly of covalently bonded hydrogel networks permits fast nonactivated DC samples (Fig. 7a, b). The ability of G-DCs to cytosolic immobilization, thereby enabling cellular fixation while stimulate T cells was further assessed in vivo with mice adoptively obviating cellular reorganization and other cellular responses. transferred with CFSE-labeled CD8+ OT-I T cells. Twenty-four Future development of the gelated cellular systems could further hour following the adoptive transfer, mice were administered benefit from the growing arsenal of membrane manipulation 6 32 with 10 G-DCs, live DCs, or glutaraldehyde-fixed DCs. Notably, strategies for intracellular hydrogel delivery , which may be less glutaraldehyde-fixed DCs induced asphyxiation and mortality disruptive as compared to the freeze–thaw approach adopted in shortly after injection, whereas G-DCs and live DCs were well the present work. Toward biomimetic materials engineering, tolerated. Flow cytometric analysis of the splenocytes 3 days intracellular hydrogelation permits facile preparation of stable, following the DC injections revealed that both the activated cell-like constructs, offering a robust platform for device devel- G-DCs and live DCs induced substantial expansion of CD8+ opment. As cell membranes are a widely present interface with T cells (Fig. 7c, d). This T-cell stimulation was antigen-specific broad biological implications, the present system sees broad as nonactivated DCs and G-DCs failed to induce T-cell division. potential applications in biomembrane research and biomaterials To further highlight the role of hydrogel support in preserving the engineering. 8 NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 ARTICLE ac e G-DC Live DC Glut-fixed DC G-DC Live DC Day 21 Day 21 Homogenized G-DC DC G-DC 89.1 % 79.3 % 56.2 % 88.8 % 1.8 % 51.3 % 8.9 % 8.34 % Activated Activated 2.39 % 2.38 % 1.42 % 1.03 % 1.75 % Non- Non- activated activated Proliferation Proliferation Proliferation bd f Day 21 G-DC 80 G-DC 80 Day 21 DC Live DC Homogenized G-DC 60 60 60 40 40 20 20 Activated Non-activated Activated Non-activated Fig. 7 Expansion of antigen-specific T cells ex vivo and in vivo by G-DCs. a, b Antigen-specific CD8+ T-cell stimulation by G-DCs, live DCs, and glutaraldehyde-fixed DCs ex vivo. Error bars represent mean ± SEM. n = 3. c, d Antigen-specific CD8+ T-cell stimulation by G-DCs and live DCs in vivo. Error bars represent mean ± SEM. (n = 4 to 8). e, f G-DCs and non-gelated DCs were stored in PBS at 4 °C for 21 days (Day 21 G-DC and Day 21 DCs). In vivo antigen-specific CD8+ T-cell stimulation was examined with day 21 G-DC, day 21 DCs, and day 21 G-DC structurally disrupted by ultrasonication (homogenized G-DCs). T-cell expansion was monitored on day 6. CFSE-labeled T-cell groups (gray) are shown as negative controls. Error bars represent mean ± SEM, n = 3 Methods curve was prepared with serially diluted PEG-DA. The measured PEG-DA content Ethics statement. All animal experiments were carried out in strict accordance was then divided by the total volume of HeLa cells to determine the intracellular with the recommendations from the Guidebook for the Care and Use of Laboratory PEG-DA concentration. Animals (published by The Chinese Taipei Society of Laboratory Animal Sciences). The experiment protocol was approved by the Academia Sinica Institutional Atomic force microscopy and elastic moduli assessment. Elastic moduli were Animal Care & Utilization Committee, Academia Sinica, Taipei, Taiwan. assessed using a Zeiss axiovert microscope and analyzed using JPK NanoWizard 3 (JPK instrument, Berlin, Germany). Glass slides with different samples were Cell culture. HeLa cells, a human epithelial cell line (ATCC, CCL-2), were −1 attached to the tip of AFM cantilever with force constant of 0.08 N m and a grown in complete media (Eagle’s Minimum Essential Medium and 10% fetal ™ resonance frequency of 20 kHz (NANOSENSORS ). Elastic moduli of different bovine serum (FBS)). JAWSII cells, an immature murine dendritic cell line (ATCC, cellular samples were quantified by contact mode. The force scanning technique CRL-11904), were grown in complete media (alpha minimum essential medium was also used to generate high-resolution (64 × 64 points) topographical/elastic with ribonucleosides, deoxyribonucleosides, 4 mM L-glutamine, 1 mM sodium maps of the cells. −1 pyruvate, 5 ng ml murine granulocyte-macrophage colony-stimulating factor, and 20% FBS). Cell roundedness and fluorescence quantification. Cell circularity was measured and calculated by ImageJ software. A built-in option for analyzing roundness is Intracellular gelation of suspension and adherent cells. Gelation buffers were available. Briefly, after image files were imported into ImageJ software, and we TM first prepared by mixing protease inhibitors (Pierce Protease Inhibitor Mini chose the built-in options to enhance contrast, to analyze circularity, and to export Tablets; ThermoFisher), 1 wt% of 2-hydroxy-4′-(2-hydroxyethoxy)−2-methylpro- results for cell roundness. For quantification of fluorescence intensity, fluorescent piophenone (Irgacure D-2959; Sigma-Aldrich), and poly(ethylene glycol) diacrylate image files were processed in ZEN Imaging Software (Carl Zeiss). (PEG-DA; Mn = 700 Da; Sigma-Aldrich) ranging from 4 to 40 wt% in 10 mM phosphate buffer. For fluorescent labeling of the hydrogel network, the gelation buffers were supplemented with 0.05 wt% of fluorescein O,O’-diacrylate (Sigma- FITC-dye exclusion assay. Totally, 1×10 gelated HeLa cells (4 wt%) were sus- Aldrich). For cross-linking cells in suspension, adherent cells (i.e., HeLa and −1 pended in 500 μL of PBS solution containing 10 μgmL of FITC for 2 h. Totally, JAWSII cells) were detached using an enzyme-free cell dissociation buffer (Ther- 50 μL of samples were then added to a confocal dish and observed by a confocal moFisher). Cells in suspension were pelleted at 200×g and resuspended in desig- microscope. nated gelation buffers. For cross-linking adherent cells, cells grown on a tissue culture plate were washed with PBS and immersed in the designated gelation buffers. Immediately following the addition of gelation buffers, the cells were flash Hemagglutination of avian erythrocytes. A/PuertoRico/8/34(H1N1) was frozen in methanol precooled in a −80 °C freezer. After 10 min of freezing, the cells propagated in 10-day-old specific-pathogen-free (SPF) chicken embryos (JD-SPF were thawed in a 37 °C water bath. The suspension cells were pelleted at 200×g and Biotech, Miaoli, Taiwan) via the allantoic route. Native virions were then derived resuspended in PBS on a tissue culture plate, whereas adherent cells were washed by purifying the virus-containing allantoic fluid (AF) through 20–50% sucrose with PBS twice. The tissue culture plates were then placed in an ice bath, and the gradient solution. Avian erythrocytes were prepared from chicken whole blood cells were crosslinked with 365 nm UV wavelength for 10 min using a UV lamp upon removal of plasma and buffy coat following centrifugation at 200 × g. Gelated (UVP UVLMS-38 EL Series) placed 2 in. above the tissue culture plate. The avian erythrocytes were prepared using 20 wt% PEG-DA. Hemagglutination study resulting GCs were washed twice in PBS for further experiments. was performed by adding 10 virions to 1 mL of PBS solution containing 2% of avian erythrocytes. Presence of hemagglutination was monitored following 30 min Quantification of intracellular PEG-DA concentrations. Quantification of of incubation at room temperature. intracellular PEG-DA concentrations was performed using an iodine-based 33 6 quantification method . Briefly, following PEG-DA infusion, 1 × 10 HeLa cells were washed and suspended in PBS to 1 mL. The collected cells were then Transmission electron microscopy. Cellular samples were fixed using 2% sonicated in a bath sonicator for 1 min to release the entrapped PEG-DA, and glutaraldehyde in 0.1 M cacodylate buffer at pH 7.4 overnight at 4 °C. After the cellular debris was spun down via centrifugation at 3000×g for 5 min. The postfixation in 1% osmium tetroxide and pre-embedding staining with 1% uranyl supernatants were collected and mixed with BaCl and iodine solutions in an 8:2:1 acetate, tissue samples were dehydrated and embedded in Agar 100. Sections ratio. Following color development for 15 min, PEG-DA concentrations in the measuring 80 nm were then examined using an FEI Tecnai G2 TF20 Super TWIN samples were determined by measuring the light absorbance at 535 nm. A standard microscope equipped with a field emission gun. NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications 9 % Proliferation Count % Proliferation Count Count % Proliferation ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 −1 Cryogenic scanning electron microscopy. For cryogenic scanning electron days after seeding, the cells were treated with 1 μgmL of LPS for 16 h at 37 °C −1 microscopy (cryo-SEM) imaging, an FEI Quanta 200/Quorum PP2000TR FEI, and then pulsed with 10 μgml SIINFEKL OT-I peptide for 4 h in complete 2007 high-resolution SEM was used. Briefly, HeLa cells were seeded on Aclar media. Activated or nonactivated DCs were gelated and then resuspended in PBS embedding films for 24 h prior to PBS or gelation treatments. Before imaging, the with 10% FBS. FITC-conjugated hamster anti-mouse CD3 (BioLegend, #100306, samples were washed with PBS and suspended in RO water for freezing by liquid clone 145-2C11, 1:100), CD80 (eBioscience, #11–0801–81, clone Ly-53, 1:100), or nitrogen. The samples were then etched under vacuum and imaged at an accel- allophycocyanin-conjugated anti-MHC class I-SIINFEKL antibodies (eBioscience, eration voltage of 3 kV by cryo-SEM. #17-5743-80, clone25-D1.16, 1:100) were added and incubated with the cells at room temperature for 30 min in the dark. The cells were then washed twice, and the expression of surface markers was acquired by FACSCanto (BD Biosciences) Examination of membrane order in GCs. Examination of membrane lipid order and analyzed by FlowJo software (Tree Star). Gating strategies for all flow cyto- in GCs was carried out according to a previously described protocol . Briefly, metric analyses are shown in Supplementary Figure 17. Statistical analysis was −1 GCs and control cells were stained in media containing 100 μgml of Laurdan performed based on a two-tailed, unpaired t test using GraphPad Prism. dye for 1 h. The samples were subsequently washed with PBS, and the images for membrane order analysis were acquired by confocal microscopy. For the imaging setup, the excitation wavelength was set at 405 nm, and the detection T-cell isolation and fluorescence labeling. OT-I cells (CD8+ T cells specific for wavelengths were set at 440–460 nm for the ordered channel and 490–510 nm OVA257–264 peptide in the H2-K context) were isolated from OT-I transgenic for the disordered channel. All images were exported in the TIFF format and mice, which were a gift from Dr. Nan-Shih Liao from the Institute of Molecular saved as 32-bit grayscale image files with ImageJ. The custom-written macro Biology, Academia Sinica. After mice were sacrificed, their spleens were removed provided by Owen et al. was loaded into ImageJ to calculate GP values as well and placed into RPMI1640 complete medium with 10% FBS. In order to harvest as to create pseudo-colored GP-intensity-merged images and intensity histograms. single splenocytes, the spleens were tamped and strained with the tip of a 5 ml GP values were calculated according to the equation: GP value = (I – syringe against a sterile 40 μm nylon cell strainer (BD Biosciences Falcon, 440–460 nm I )/ (I + I ), where I indicate intensity of pixels. #352340). Splenocytes were incubated with BD Pharm Lyse lysing buffer (BD 490–510 nm 440–460 nm 490–510 nm Biosciences, # 555899) for 3 min to remove RBCs. OT-I cells were subsequently isolated from the splenocytes using a Mouse CD8a T Cell Isolation Kit (BD Fluorescence microscopy and FRAP analysis. Cell membrane was stained Biosciences, #19853 A). OT-I cells were stained with carboxyfluorescein diacetate by adding 10 μL of DiD dye solution (1,1′-Dioctadecyl-3,3,3′,3′-tetra- succinimidyl ester (CFSE) by incubating the cells with PBS containing 5 μMof −1 methylindodicarbocyanine; ThermoFisher Scientific) containing 5 μgmL of DiD CFSE (Sigma-Aldrich, #21888) at 37 °C for 5 min. The cells were washed three dye and 0.5% of DMSO to 200 μL of cell suspension. HeLa cells expressing EGFP- times with complete medium. CFSE-labeled cells were harvested for further GPI (Addgene, pCAG: GPI-GFP, #32601), CD80-GFP (Sino Biological Inc., experimental studies. pCMV3-mCD80-C-GFPSpark, MG50446-ACG), Transferrin Receptor (TfR) (Sino Biological Inc., pCMV3-hTfR-C-DDK (flag) tag, HG11020-CF), Lyn-GFP (Sino Biological Inc., pCMV3-hLyn-C-GFPSpark, HG10829-ACG), EGFR-GFP (Sino Examination of G-DC/T-cell interaction. For observation of G-DC/T-cell inter- Biological Inc., pCMV3-mEGFR-C-GFPSpark, MG51091-ACG), and GFP-beta- actions, adherent JAWSII DCs were gelated using 4 wt% PEG-DA and subse- TM actin (Sino Biological Inc., pCMV3-hbeta-actin-N-GFPSpark, HG10962-ANG) quently stained with CellTracker Deep Red dye (Molecular Probes) at 37 °C were prepared via transfection. Plasmids were transfected into cells with Lipo- for 30 min. Stained G-DCs were washed twice using PBS and resuspended in fectamine 3000 (Invitrogen) according to the manufacturer’s instruction. After RPMI1640 complete media supplemented with 10% FBS. CFSE-labeled OT-I cells 48 h, the cells were either gelated or treated with 2.5% glutaraldehyde for 10 min were subsequently added to the G-DCs. The interaction between G-DCs and OT-I prior to fluorescence microscopy or FRAP analysis. Fluorescence microscopy CD8 T cells was subsequently imaged using a confocal microscope (Zeiss LSM780 and FRAP analysis were carried out on a Zeiss LSM780 confocal microscope confocal microscope system, Zeiss) and analyzed using LSM Image Browser soft- (Carl Zeiss, Oberkochen, Germany) equipped with Plan-Apochromat 100×/1.4 oil ware (Zeiss). To examine CD80 clustering on G-DCs, JAWSII DCs were trans- objective. For FRAP analysis, adherent cells and adherent GCs were used rather fected with CD80-GFP plasmids using the TransIT-TKO transfection reagent than suspension cells to minimize artifacts due to random movements. An (Mirus, #2154) following a previously described protocol protocol . Briefly, objective heater was used to maintain samples at 37 °C. Images were collected with plasmids were prepared using a Qiagen Plasmid Midi kit (QIAGEN, #21243). a pinhole of 1.52 AU (1.1 μm section) for optimal signal intensity. The sample was Transfection mixtures consisting of 5 mL of serum-free DMEM, 20 µg of plasmids, first scanned three times with 5% of laser power to measure the fluorescence and 40 µL of transfection reagent were prepared and transfected into JAWSII DCs. intensity before photobleaching, followed by 500 iterative laser pulses at full power Following 4 h of incubation, an additional 10 mL of complete medium was added to photobleach a 27 nm × 6 nm rectangular area at the plasma membrane. Fluor- to the cells. 48 h after transfection, CD80-GFP-expressing JAWSII cells were escence recovery was monitored every 2 s for at least 2 min at 60 frames per second gelated with 4 wt% PEG-DA with and used for examining CD80 clustering upon until a plateau is reached. Fluorescence intensity vs. time was plotted for analyzing incubation with antigen-specific T cells. the fluorescence recovery. The mobile fraction was calculated based on the equa- tion (I − I )/(I − I ) × 100%, where I is the end value of the recovered fluores- E 0 I 0 E T proliferation assay ex vivo. CFSE-labeled OT-I cells were co-cultured with cence intensity, I is the first post-bleach fluorescence intensity, and I is the initial 0 I live DCs, G-DCs or glutaraldehyde-fixed DCs at different ratios. Co-cultured cells (prebleach) fluorescence intensity. The halftime of recovery (t ) is derived as the 1/2 in 96-well v-bottomed plates were cultured at 37 °C for indicated time periods. After time from the bleach to the time point where the fluorescence recovery reaches harvesting, cells were stained with allophycocyanin-conjugated rat anti-mouse 50% of the final recovery intensity. For the recovery rate of CD80-GFP, the rate CD8a antibodies (eBioscience, # 100712, Clone 53-6.7, 1:100) and analyzed by flow constant k was derived by converting the fluorescence recovery to a first-order cytometry. Proliferation analysis platform in FlowJo was used to analyze cell divi- elimination kinetics curve in which concentration A is calculated as I − I with (t) I (t) sion. For experiments involving stored G-DCs and DCs, G-DCs, and DCs were I being set as the first postbleach fluorescence intensity. The first 20 time points (0) stored in PBS at 4 °C for 21 days. Homogenized G-DCs were prepared by sonicating were used for calculating the recovery kinetics. Recovery kinetics were calculated 21-day-old G-DCs using a Fisher Scientific 150E Sonic Dismembrator at 80% power based on the equation ln[A ] = −kt + ln[A ], in which k = −(ln[A ]-ln[A ])/t. (t) (0) (t) (0) pulsed (3 s on/1 s off) for 1 min. The G-DC, DC, and glutaraldehyde-fixed DC samples were derived from the same cell source for each separate experiment. Protein tracking by TIRF microscopy. Movements of CD80-GFP were observed by TIRF microscopy using Leica TIRF MC inverted fluorescence microscope T proliferation assay in vivo. CFSE-labeled splenocytes were adoptively trans- equipped with HCX PL-APO 100× NA 1.46 Oil objective lens (Leica Microsystems, ferred via tail vein injections to 8-week-old C57BL/6 J mice at a cell number of Germany). Cell samples were loaded onto a glass bottom culture plate, and the 3.3 × 10 . Twenty-four hours after the adoptive transfer of OT-I cells, the mice samples were exposed to a 488-nm wavelength laser. The fluorescence image was were challenged with live DCs, G-DCs or Glut-fixed DCs at a cell number of 10 acquired using a hamamatsu EM-CCD camera (C9100-13) at a temporal resolution via tail vein injections. 3 days after the DC injections, splenocytes were harvested of 63 ms. All single-molecular experiments were performed at 37 °C. Protein from the mice and stained with allophycocyanin-conjugated rat anti-mouse CD8a movements were then analyzed using two-dimensional trajectories of CD80-GFP antibodies, followed by flow cytometry and FlowJo analysis. The animal protocol molecules in the plane of the basal membrane and were reconstructed by Imaris was approved by the Institutional Animal Care and Use Committee (IACUC) Image Analysis Software (Bitplane, Switzerland). The diffusion constants were at Academia Sinica. The G-DCs and DCs were derived from the same source of evaluated based on a previously described method . Briefly, the mean square activated or nonactivated DCs. displacement (MSD) was plotted from each trajectory against time (t). For each molecule, the slope of the first three time points in the MSD t plot was used to Reporting summary. Further information on experimental design is available in calculate the diffusion coefficient, D, according to the equation MSD = 4Dt. t→0 Statistical analysis was performed using one-way ANOVA with GraphPad Prism. the Nature Research Reporting Summary linked to this article. The F value is 5.158, and the degrees of freedom is 5. Data availability Dendritic cell preparation and analysis. For activation, JAWSII cells were seeded All relevant data are available from the authors and/or are included within the onto a 100 mm petri dish with 10 mL of media at a density of 10 per dish. Three manuscript and Supplementary Information. 10 NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-09049-5 ARTICLE Received: 20 June 2018 Accepted: 18 February 2019 27. Bousso, P. & Robey, E. Dynamics of CD8+ T cell priming by dendritic cells in intact lymph nodes. Nat. Immunol. 4, 579–585 (2003). 28. Fisher, P. J., Bulur, P. A., Vuk-Pavlovic, S., Prendergast, F. G. & Dietz, A. B. Dendritic cell microvilli: a novel membrane structure associated with the multifocal synapse and T-cell clustering. Blood 112, 5037–5045 (2008). 29. Nakamura, H. et al. Intracellular production of hydrogels and synthetic RNA granules by multivalent molecular interactions. Nat. Mater. 17,79–89 References (2018). 1. Zhang, L. & Granick, S. Slaved diffusion in phospholipid bilayers. Proc. Natl 30. Shin, Y. et al. Spatiotemporal control of intracellular phase transitions using Acad. Sci. USA 102, 9118–9121 (2005). light-activated optodroplets. Cell 168, 159–171 (2017). 2. Grakoui, A. et al. The immunological synapse: a molecular machine 31. Yang, Z. M., Xu, K. M., Guo, Z. F., Guo, Z. H. & Xu, B. Intracellular enzymatic controlling T cell activation. Science 285, 221–227 (1999). formation of nanofibers results in hydrogelation and regulated cell death. 3. Bromley, S. K. et al. The immunological synapse and CD28–CD80 Adv. Mater. 19, 3152–3156 (2007). interactions. Nat. Immunol. 2, 1159–1166 (2001). 32. Stewart, M. P. et al. In vitro and ex vivo strategies for intracellular delivery. 4. Lin, C. C. & Anseth, K. S. Cell–cell communication mimicry with poly Nature 538, 183–192 (2016). (ethylene glycol) hydrogels for enhancing beta-cell function. Proc. Natl Acad. 33. Cheng, T. L., Chuang, K. H., Chen, B. M. & Roffler, S. R. Analytical Sci. USA 108, 6380–6385 (2011). measurement of PEGylated molecules. Bioconjugate Chem. 23, 881–899 5. Tang, J. A. et al. Therapeutic microparticles functionalized with biomimetic (2012). cardiac stem cell membranes and secretome. Nat. Commun. 8, 13724 (2017). 34. Awasthi, S. & Cox, R. A. Transfection of murine dendritic cell line (JAWS II) 6. Hu, C. M. J. et al. Nanoparticle biointerfacing by platelet membrane cloaking. by a nonviral transfection reagent. Biotechniques 35, 600–602 (2003). Nature 526, 118–121 (2015). 7. Hu, C. M. et al. Erythrocyte membrane-camouflaged polymeric nanoparticles as a biomimetic delivery platform. Proc. Natl Acad. Sci. USA 108, Acknowledgements 10980–10985 (2011). The authors acknowledge technical support from Common Equipment Core, Institute 8. Parodi, A. et al. Synthetic nanoparticles functionalized with biomimetic leukocyte of Biomedical Science, Academia Sinica, for the confocal microscopy image acquisition membranes possess cell-like functions. Nat. Nanotechnol. 8,61–68 (2013). and FRAP analysis, Yi-Ru Li for the technical support on atomic force microscopy, 9. Oelke, M. et al. Ex vivo induction and expansion of antigen-specific cytotoxic Kung-Hsuan Lin and Tzu-Ling Wu for the TIRF image acquisition, Yao-Kuan Huang for T cells by HLA-Ig-coated artificial antigen-presenting cells. Nat. Med. 9, the technical support on transmission electron microscopy. The authors acknowledge 619–624 (2003). funding support from the Academia Sinica Career Development Award (CDA-105-L06) 10. Kosmides, A. K. et al. Biomimetic biodegradable artificial antigen presenting and by the Ministry of Science and Technology, Taiwan (106-2119-M-001-010). with PD-1 blockade to treat melanoma cells synergize. Biomaterials 118, 16–26 (2017). 11. Cheung, A. S., Zhang, D. K. Y., Koshy, S. T. & Mooney, D. J. Scaffolds that Author contributions mimic antigen-presenting cells enable ex vivo expansion of primary T cells. J.C.L., C.Y.C., Y.I.C., H.W.C. and C.M.J.H. conceived the experimental designs. J.C.L., Nat. Biotechnol. 36, 160–169 (2018). C.Y.C., Y.I.C., J.Y.C., B.Y.Y., N.N.L., Z.S.F. and W.Y.C. performed the optimization and 12. Fadel, T. R. et al. A carbon nanotube-polymer composite for T-cell therapy. characterization of the intracellular hydrogelation protocol. C.Y.C., Y.I.C., C.L.L., B.Y.Y. Nat. Nanotechnol. 9, 639–647 (2014). and W.Y.C. performed the membrane fluidity analysis. J.C.L. and Y.H.L. performed the 13. Tanaka, M. & Sackmann, E. Polymer-supported membranes as models of immunological assays. J.C.L., C.Y.C., C.L.L. and C.M.J.H. prepared the paper. All authors the cell surface. Nature 437, 656–663 (2005). have read and approved the paper. 14. Pace, H. et al. Preserved transmembrane protein mobility in polymer- supported lipid bilayers derived from cell membranes. Anal. Chem. 87, 9194–9203 (2015). Additional information 15. Hardy, G. J., Nayak, R. & Zauscher, S. Model cell membranes: techniques Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- to form complex biomimetic supported lipid bilayers via vesicle fusion. 019-09049-5. Curr. Opin. Colloid Interface Sci. 18, 448–458 (2013). 16. Chiang, P. C., Tanady, K., Huang, L. T. & Chao, L. Rupturing giant plasma Competing interests: The authors declare no competing interests. membrane vesicles to form micron-sized supported cell plasma membranes with native transmembrane proteins. Sci. Rep. 7, 15139 (2017). Reprints and permission information is available online at http://npg.nature.com/ 17. Fletcher, D. A. & Mullins, R. D. Cell mechanics and the cytoskeleton. reprintsandpermissions/ Nature 463, 485–492 (2010). 18. Brizard, A. M. & Van Esch, J. H. Self-assembly approaches for the Journal peer review information: Nature Communications thanks the anonymous construction of cell architecture mimics. Soft Matter 5, 1320–1327 (2009). reviewers for their contribution to the peer review of this work. Peer reviewer reports are 19. Park, J. et al. Combination delivery of TGF-beta inhibitor and IL-2 by available. nanoscale liposomal polymeric gels enhances tumour immunotherapy. Nat. Mater. 11, 895–905 (2012). Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in 20. Underhill, G. H., Chen, A. A., Albrecht, D. R. & Bhatia, S. N. Assessment of published maps and institutional affiliations. hepatocellular function within PEG hydrogels. Biomaterials 28,256–270 (2007). 21. Owen, D. M., Rentero, C., Magenau, A., Abu-Siniyeh, A. & Gaus, K. Quantitative imaging of membrane lipid order in cells and organisms. Open Access This article is licensed under a Creative Commons Nat. Protoc. 7,24–35 (2011). Attribution 4.0 International License, which permits use, sharing, 22. Doty, R. T. & Clark, E. A. Two regions in the CD80 cytoplasmic tail regulate adaptation, distribution and reproduction in any medium or format, as long as you give CD80 redistribution and T cell costimulation. J. Immunol. 161, 2700–2707 appropriate credit to the original author(s) and the source, provide a link to the Creative (1998). Commons license, and indicate if changes were made. The images or other third party 23. Tseng, S. Y., Liu, M. L. & Dustin, M. L. CD80 cytoplasmic domain controls material in this article are included in the article’s Creative Commons license, unless localization of CD28, CTLA-4, and protein kinase C theta in the indicated otherwise in a credit line to the material. If material is not included in the immunological synapse. J. Immunol. 175, 7829–7836 (2005). article’s Creative Commons license and your intended use is not permitted by statutory 24. Tanaka, K. A. K. et al. Membrane molecules mobile even after chemical regulation or exceeds the permitted use, you will need to obtain permission directly from fixation. Nat. Methods 7, 865–866 (2010). 25. Chen, L. P. & Flies, D. B. Molecular mechanisms of T cell co-stimulation the copyright holder. To view a copy of this license, visit http://creativecommons.org/ and co-inhibition. Nat. Rev. Immunol. 13, 227–242 (2013). licenses/by/4.0/. 26. Underhill, D. M., Bassetti, M., Rudensky, A. & Aderem, A. Dynamic interactions of macrophages with T cells during antigen presentation. © The Author(s) 2019 J. Exp. Med. 190, 1909–1914 (1999). NATURE COMMUNICATIONS | (2019) 10:1057 | https://doi.org/10.1038/s41467-019-09049-5 | www.nature.com/naturecommunications 11

Journal

Nature CommunicationsSpringer Journals

Published: Mar 5, 2019

There are no references for this article.