Stiffness memory of EA.hy926 endothelial cells in response to chronic hyperglycemia

EA.hy 926 is a permanent cell line derived by fusing human umbilical vein endothelial cells with cell line A549 [23]. The culture was maintained according to the protocol provided by the distributor of the cell line (ATCC cat. CRL 2922, Manassas, Virginia, USA). EA.hy926 cells were cultured in DMEM medium with 25 mM glucose (HG) supplemented with 10% fetal bovine serum (cat. No 10082-147, Invitrogen), 2 mM L-glutamine, and 2% HAT supplement (cat. 21060-017, Invitrogen) [22]. After 14 passages (P14 HG), the whole set of cells was divided into two groups cultured in parallel: the first one was still kept in the conditions described above; the second one was cultured in DMEM with 5 mM glucose (NG), supplemented as above.

The glass slides were cleaved into 1 cm x 1 cm plates, rinsed both sides with ethanol and exposed to UV light in laminar chamber for 30 min in order to sterilize. Next, EA.hy926 cells were seeded on the glass slides at density of 10⁴ cells/ml, and cultured to 80% confluence. Afterwards, the sample was immediately taken for measurements.

AFM measurements were performed with a commercial instrument Nanoscope IIIa Multimode SPM (Veeco Instruments, Santa Barbara, CA, USA). All experiments were performed on unfixed ECs in DMEM solution using a commercial fluid chamber (Veeco Instruments). V-shape gold-coated cantilevers (MLCT multilever, Veeco Probes, Camarillo, CA, USA) with nominal spring constants of 0.01 N/m were used for nanoindentation measurements. The spring constant of each cantilever was calculated basing on the thermal oscillations spectrum analyzed using NanoScope 6.1 software, while the tip radius was estimated basing on a scan of a test gird (TGT-1, NT-MDT).

Figure 1A shows the optical image of the monolayer formed by cells during the experiment. For a given passage, two samples were examined, about 12-14 cells for each sample, chosen randomly. The time of characterization of each sample was limited to 1.5 h. In the nanoindentation spectroscopy, the AFM tip is attached to a flexible cantilever and acts as an indenter. Using a piezo-scanner, a cell placed on a stiff substrate is pushed towards the tip. This leads to a gentle indentation (deformation) of the cell. The resulting elastic force (F) applied to the cantilever (related to its bending) is measured as a function of the scanner position (z) - this dependence is visualized on a so called force-distance curve. Later, force-indentation curve [F(δ z)] is obtained from two separately measured force-distance curves [F(z)]. The first force-distance curve, measured on the cell, contains the information about both the elastic properties of the cell and the cantilever (see Figure 1B). Since the AFM hardware only controls the relative scanner position, the indentation depth δ z cannot be directly determined from a single measurement. Therefore, for a reference, a second force-distance curve is measured on a significantly stiffer substrate (glass), for which the elastic deformation is negligible and the cantilever bending is solely determined by the position of the piezo-scanner (Figure 1B). Comparison of curves acquired for the cell and for the stiff substrate allows to obtain the indentation depth δ z and plot the final force-indentation curve (see Figure 1C). Figure 1D presents histograms of the elasticity parameter value calculated basing on force-indentation curves acquired for 8 cells. In order to evaluate the most probable value of the elasticity parameter, a Log-normal function was fitted to the histogram values.

For fluorescence microscopy, cells were seeded on 24-well plates in the same concentration as described above. Before staining, the cells were rinsed with pre-warmed PBS and then fixed with 2.5% formaldehyde for 10 minutes. After rinsing them again, cells were permeabilized with 0.1% Triton X for 4 minutes. Next, cells were incubated in PBS with 1% bovine serum albumin (BSA) for 30 minutes and immediately exposed to phalloidyne conjugated with Alexa Fluor (Invitrogen) for 20 minutes. Prior to the measurement, wells were rinsed twice with PBS. Images were obtained by Olympus IX71 through 40x objective recorded and processed with Olympus Cell Sense software. Fluorescence was excited with Olympus X-Cite Q120 lamp and detected by Olympus U-MWIB2 (Phallodoxine) emission filters.

Main motivation of our experiment was to investigate the elastic properties of EA.hy926 ECs cultured in HG conditions and after the transfer of the cells from HG to NG. The base culturing medium for this cell line is the high glucose medium, hence the first part of the experiment was carried out for cells cultured in HG. Measurements were continued until the cells incubated in HG conditions lost the ability to proliferate, which occurred at passage P26-P27. At the time of half achievable passages - after P14, we intentionally changed the concentration of HG to NG for half of the cells, while continuing with HG for the other half.

The changes of elastic properties for consecutive passages of EC are presented in Figure 2. Each point on the graph corresponds to the mean elasticity parameter value obtained from the analysis of c.a. 1200 indentation curves (12-14 cells) collected for each passage. Black dots represent results obtained for cells cultured in HG medium (glucose concentration 25 mM). There is a distinct stiffening trend up to P20. It can be observed, that the elasticity parameter does not increase smoothly, but rather in a stepwise manner - after each relevant growth in stiffness (i.e. P4-P7, P10-P13, P14-P16), the value remains similar in the few subsequent passages. In late passages (higher than P21), the elasticity parameter decreases with time.

After P14, a subpopulation of the cells was transferred to medium with normal glucose level (NG) (5 mM). The results of elasticity measurements for those cells are marked in Figure 2 with red asterisks. Before the transfer to NG conditions, the elasticity parameter value was (1.24±0.02) kPa. Initially, the exchange of culturing conditions results in a distinct drop, to (0.88±0.04) kPa, in elasticity parameter, i.e., cells become softer. However, maintaining the culture in NG for 4 additional passages results in the return of the elastic properties to the value observed before the change of glucose concentration and very close to the value characteristic for cells cultured constantly in HG.

The latter observation indicates that culturing the cells in hyperglycemic environment develops persistent modification of elastic properties, which cannot be reverted by a simple decrease in glucose concentration, even to the physiological level.

Elasticity measurements were supplemented with fluorescent cytoskeleton staining of actin fibers (F-actin). Representative images for selected HG passages are presented in Figure 3: P5 (A), P10 (B), P14 (C). A normal (unchanged) cytoskeleton structure could be observed in cells cultured up to 10 subsequent passages in hyperglycemic conditions (Figure 3A). After 14 passages, an increase in the F-actin content is observed. This indicates that hyperglycemia stimulates both polymerization of F-actin and stress fibers formation. Moreover, it is visible, that after P14 HG the modification in the structure of the cortical actin cytoskeleton leads to the intercellular gap formation and thus to the increase in the permeability of the endothelium.

After P20 the ECs layer becomes less cohesive and the integrity of endothelial barrier decreases significantly: cells are shrunk and numerous gaps between them are present. As depicted in Figure 2D and 2E, the shape of cells changes: they become elongated, F-actin fibers are formed along the cell, across its central part. These observations are typical for dysfunctional endothelium - ROS production and activation of intracellular pathways contributes to the intercellular gaps formation [25]. Lowering the glucose concentration to 5 mM after initial 14 passages in HG, caused significant changes in the organization of cytoskeleton, as presented in Figures 3G-I. Initial depolymerization of F-actin in (P14+)P2 (Figure 3G) is reverted in (P14+)P6 (Figure 3F) to the typical HG pattern observed in the passage P20 HG-P24 HG. Such changes correlate with the changes of elasticity parameter presented in Figure 2.