Evidence for macromolecular crowding as a direct apoptotic stimulus

Abstract

Potassium loss and persistent shrinkage have both been implicated in apoptosis but their relationship and respective roles remain controversial. We approached this problem by clamping intracellular sodium and potassium in HeLa or MDCK cells using a combination of ionophores. Although ionophore treatment caused significant cell swelling, the initial volume could be restored and further reduced by application of sucrose. The swollen cells treated with ionophores remained viable for at least 8 h without any signs of apoptosis. Application of sucrose and the resulting shrinkage caused volume-dependent intrinsic apoptosis with all its classical features: inversion of phosphatidylserine, caspase activation and Bcl-2-dependent release of cytochrome c from mitochondria. In other experiments, apoptosis was induced by addition of the protein kinase inhibitor staurosporine at various degrees of swelling. Our results show that: (1) persistent shrinkage can cause apoptosis regardless of intracellular sodium or potassium composition or of the state of actin cytoskeleton; (2) strong potassium dependence of caspase activation is only observed in swollen cells with a reduced density of cytosolic proteins. We conclude that macromolecular crowding can be an important factor in determining the transition of cells to apoptosis.

Keywords: Apoptosis; Cell volume; Hyperosmotic stress; K+; Macromolecular crowding; Volume sensing.

Fig. 1.

Changes in volume and macromolecular density in VNGO-treated HeLa cells. (A) Changes in the volumes of adherent HeLa cells after a 60 min incubation in HPB containing VNGO or in HSB containing VNGO, with the addition of different concentrations of sucrose. The volumes are shown relative to pretreatment volumes in DMEM (horizontal dotted line). The results represent 10–27 cells per treatment in a single experiment. Similar data have been published previously (Rana et al., 2019). Longer incubation times did not cause major changes in cell volumes. (B) Changes in macromolecular density after 1 h and 6 h in HPB containing VNGO with the addition of different concentrations of sucrose. The range (mean±s.e.m.) for the same cells in DMEM before treatment (92 cells) is indicated with horizontal dotted lines. For every treatment condition, 20–26 cells were analyzed. (C,D) TTD images of cells before (C) and after (D) 6 h incubation in HPB containing VNGO with 100 mM sucrose. The intensities are displayed on a logarithmic scale, where local brightness is proportional to cell thickness. Scale bar: 20 µm.

Fig. 2.

Caspase-3 activation in cells treated with VNGO in HPB or HSB in the presence of various sucrose concentrations. (A,B) Percentage of caspase-positive HeLa and MDCK cells post-treatment. Bars represent mean±s.e.m. of n=2 (for 75 mM sucrose) or n=5 (for all other treatments) experiments. (C) Overlay of brightfield images of HeLa cells and NucView 488 fluorescence (green), indicating caspase-3 activation. Scale bar: 100 μm.

Fig. 3.

Phosphatidylserine translocation in MDCK cells treated with VNGO in HPB or HSB in the presence of various sucrose concentrations. (A) Percentage of annexin-positive MDCK cells (n=3). (B) An overlay of fluorescence and brightfield images of MDCK cells stained with Annexin V-FITC (green channel). Scale bar: 100 µm.

Fig. 4.

Cytochrome c release in HeLa cells treated with VNGO in HPB or HSB in the presence of various sucrose concentrations. (A) Percentage of apoptotic HeLa cells, as reported by cytochrome c release. Cells with uniform and nuclear staining were counted as positive (n=4). (B) Cytochrome c fluorescence immunostaining. Scale bar: 50 µm.

Fig. 5.

Expression of anti-apoptotic Bcl-2 inhibits cytochrome c release in HeLa cells exposed to HPB containing VNGO and 100 mM sucrose. GFP–Bcl-2 is shown in green, cytochrome c staining is shown in red and Hoechst staining is shown in blue. The two arrows in the GFP–Bcl-2 transfected sample point to cells that failed to express Bcl-2; those cells underwent apoptosis. Images on the right show the boundary of the nucleus overlaid on the cytochrome c image. The boundaries were created by applying a median filter to images of Hoechst staining, segmenting the nuclei by intensity threshold and making outlines in ImageJ. Scale bar: 25 µm.

Fig. 6.

Caspase activity during ST-induced apoptosis in HeLa cells was sensitive to potassium concentration only when the cells were swollen in VNGO. Apoptotic sensitivity to potassium was lost when the initial volume was restored by addition of 50 mM sucrose. The experiment was repeated in triplicate, and a total of 636–1464 cells were analyzed for each condition.

Fig. 7.

Addition of cytochalasin D has no effect on caspase activation in ST- and shrinkage-induced apoptosis. (A) Percentage of caspase-positive cells in the presence and absence of cytochalasin D during treatment with ST or VNGO and 100 mM sucrose (between 98 and 621 cells analyzed per treatment; n=3). (B) The lack of effect of cytochalasin D on caspase activation in ST- and shrinkage-induced apoptosis in HeLa cells. Overlays of brightfield images and NucView 488 fluorescence (green; indicating caspase-3 activation) are shown. Scale bar: 50 µm.

Fig. 8.

The hypothesized signaling junctions in shrinkage-induced apoptosis. An increase in macromolecular crowding is the only inevitable consequence of hyperosmotic shrinkage. Changes in the cytoskeleton are possible but linked to shrinkage less directly. The loss of potassium can only be a more remote response. Of these factors, macromolecular crowding had the strongest impact on apoptosis in our experiments.