HA14-1, but not the BH3 mimetic ABT-737, causes Ca²⁺ dysregulation in platelets and human cell lines

Many Bcl-2-dependent cancer cells are primed to death due to their upregulation of BH3-only proteins in response to ongoing oncogenic stress. BH3-mimetic drugs like ABT-737 compete with Bim for binding to Bcl-2, releasing Bim and triggering Bax/Bak-mediated apoptosis.¹ While ABT-737 causes regression of established tumors,² it also limits platelet survival.³ The mechanisms underlying the observed thrombocytopenia have been the subject of debate. Since Bcl-Xl is essential for platelet life span,⁴ it was proposed that ABT-737, which does not discriminate between Bcl-2 and Bcl-Xl,⁵ kills platelets by antagonizing Bcl-Xl.³ However, these devastating effects of ABT-737 on platelet function were also associated with disturbed intracellular Ca²⁺ homeostasis and dynamics,³ but this remains poorly understood and controversial.^6,7

We, therefore, explored the effect of ABT-737 on apoptosis in platelets and on intracellular Ca²⁺ signaling in platelets and human cell lines. We compared its effects on Ca²⁺ homeostasis with another Bcl-2 antagonist, the HA14-1 compound, which was reported to exert inhibitory properties on sarco/endoplasmic reticulum Ca²⁺-ATPases (SERCA).⁸ Washed platelets were prepared as described.⁹ For this, venous blood was collected from healthy donors. Permission was given by the Ethical Committee of the Leuven University Hospital to use blood from healthy individuals for further analysis. For apoptosis measurements, platelets were incubated with Annexin-V-FITC and analyzed with an Attune® Acoustic Focusing Flow Cytometer (Applied Biosystems). For the Ca²⁺ measurements in intact cells, platelets were seeded the day of measurement in poly-L-lysine-coated 96-well plates (Greiner) at a density of approximately 3 × 10⁸ platelets/mL. Ca²⁺ measurements in intact Hela cells and unidirectional ⁴⁵Ca²⁺-flux experiments in permeabilized cells were basically performed as previously described.¹⁰ SERCA2b ATPase activity was determined by colorimetric monitoring¹¹ in a Ca²⁺-buffered solution containing microsomes (10 μg of protein) from HEK-293T cells ectopically expressing SERCA2b. Results are expressed as average ± standard deviation (SD). Significance was determined using two-tailed paired Student's t-test. P<0.05 was considered significant.

We confirmed that both HA14-1 (3 and 10 μM; 2 h) and ABT-737 (0.03, 0.1 and 0.3 μM; 2 h) provoked apoptosis in blood platelets (Figure 1A and B). However, only HA14-1, but not ABT-737, triggered a slow and steady increase in the cytosolic [Ca²⁺] originating from the intracellular Ca²⁺ stores in Fura2-loaded platelets exposed to extracellular EGTA (Figure 1C). Moreover, pre-treatment (10 μM; 30 min) of HA14-1, but not of ABT-737, reduced the total Ca²⁺ released from the ER in response to thapsigargin, an irreversible SERCA inhibitor, while not affecting store-operated Ca²⁺ influx (Figure 1D). To confirm our result in another cell model, we used the human cell line HeLa. Again, HA14-1, but not ABT-737, affected intracellular Ca²⁺ homeostasis in these cells in a concentration-dependent manner (Figure 2A). To investigate the underlying mechanism, we applied a highly quantitative ⁴⁵Ca²⁺-flux assay¹⁰ in permeabilized HeLa cells to specifically assess ER 45Ca²⁺-uptake activity in the absence of plasmalemmal and mitochondrial Ca²⁺ fluxes and of IP3R-mediated Ca²⁺-release. Application of HA14-1 during the ER ⁴⁵Ca²⁺-loading phase caused a strong decrease in the steady-state 45Ca²⁺ loading levels, while ABT-737 was much less effective (Figure 2B). The steady-state ER Ca²⁺ levels are determined by the balance between the ER Ca²⁺-uptake and ER Ca²⁺-leak activities. Importantly, application of either ABT-737 or HA14-1 (up to 30 μM) during the efflux phase did not affect the ER Ca²⁺-leak rate (Figure 2C), suggesting an inhibition of the ER Ca²⁺-uptake activity. Next, we directly assessed the effect of ABT-737 and HA14-1 on SERCA2b activity, the housekeeping isoform in human platelets,¹² which was ectopically expressed in HEK-293T cells (Figure 2D). HA14-1 effectively inhibited SERCA2b Ca²⁺ ATPase activity (IC50 = 14 μM) while ABT-737 (up to 30 μM) had no effect. Finally, we assessed whether ABT-737 or HA14-1 directly affected IP3R-mediated Ca²⁺ release (e.g. which occurs in response to thrombin exposure of platelets). We used the ⁴⁵Ca²⁺-flux assay to accurately measure IP₃-induced Ca²⁺ release in permeabilized HeLa cells. We found that 30 μM HA14-1, but not ABT-737, inhibited the IP3R (Figure 2E). Performing a dose-response curve, we found that ABT-737 up to 100 μM did not inhibit IP₃Rs. In contrast, HA14-1 potently inhibited IP₃Rs at concentrations higher than 10 μM with an IC₅₀ of approximately 50 mM and with an IP₃R inhibition of approximately 80% at 100 μM (Figure 2F).

Figure 1.

The effect of HA14-1 and ABT-737 on apoptosis and intracellular Ca²⁺ signaling in platelets. (A-B) Flow-cytometry analysis of HA14-1 and ABT-737-induced apoptosis in platelets. (A) Histogram overlay representation of Annexin V-FITC staining of platelets treated for 2 h without (black) or with 1 μM (gray), 3 μM (blue) and 10 μM HA14-1 (red). (B) Histogram overlay representation of Annexin V-FITC staining of platelets treated for 3 h without (black) or with 0.03 μM (gray), 0.1 μM (blue) and 0.3 μM ABT-737 (red). (C-D) Fluorimetric analysis of the ABT-737- and HA14-1-induced Ca²⁺ responses in Fura2-loaded platelets. (C) The ratio of emitted fluorescence of Fura2 (excitation wavelength 340 nm/ 380 nm) was monitored and the different treatments were applied 60 s after the addition of EGTA (3 mM) which prohibits Ca²⁺ entry. The black, green and cyan lines represent Ca²⁺ signals obtained for platelets treated with vehicle, 10 μM ABT-737 or 10 μM HA14-1, respectively. 0.5 U/mL thrombin was added after 760 s as positive control. (D) Analysis of the thapsigargin (TG)-induced Ca²⁺ responses in untreated platelets (black) or following 30 min of pre-treatment without or with 10 μM HA14-1 (cyan) or ABT-737 (green). 10 μM thapsigargin was added 60 s after EGTA (3 mM) addition. Ca²⁺ (3 mM) was added at 400 s in order to show that the Ca²⁺-influx capacity of the platelets is not affected by any of the different conditions.

Figure 2.

The effect of HA14-1 and ABT-737 on ER Ca²⁺-uptake and –release mechanisms. (A) Fluorimetric analysis of the HA14-1 and ABT-737-induced Ca²⁺ responses in HeLa cells. The ratio of emitted fluorescence of Fura2 (excitation wavelength 340 nm/380 nm) was monitored and the five different treatments, vehicle, 10 and 30 μM ABT-737, and 10 and 30 μM HA14-1, were added 60 s after the addition of 3 mM EGTA. All curves are representative for 3 independent experiments. (B) Normalized ER ⁴⁵Ca²⁺ uptake values (in %) obtained from permeabilized cells in which steady state ER ⁴⁵Ca²⁺ loading was performed in the presence of different ABT-737 and HA14-1 concentrations. Values for vehicle-treated cells were taken as 100%. Data represent mean ± SEM of 3-5 independent experiments. (C) Effect of vehicle, ABT-737, HA14-1 and A23187 on the unidirectional ⁴⁵Ca²⁺ efflux in the absence of SERCA Ca²⁺-uptake activity. Data represent the ER Ca²⁺ content (in cpm) as a function of time (in min) and are plotted as mean ± SD. A typical experiment out of 3 independent experiments is shown. (D) Dose response curve of ABT-737 and HA14-1 on normalized SERCA2b activity. SERCA2b ATPase activity was measured at 1 μM Ca²⁺. ABT-737 does not have any significant effect on the ATPase activity whereas HA14-1 effectively inhibited SERCA2b activity (IC50 ∼14 μM). (E) Unidirectional ⁴⁵Ca²⁺-flux experiments in HeLa cells demonstrating IP3-induced Ca²⁺ release in cells exposed to vehicle (closed symbols), 30 μM ABT-737 (open symbols, left panel) and 30 μM HA14-1 (open symbols, right panel), from 2 min before till 2 min after the addition of IP3, as indicated by the light gray bar. The addition of IP₃ is indicated by the shorter dark gray bar. Representative traces are shown as fractional ⁴⁵Ca²⁺ release (%/2 min) as a function of time (in min). Data represent mean ± SD of duplicate samples. A typical result out of 4 independent experiments is shown. (F) Normalized IP₃-induced ⁴⁵Ca²⁺ release (IICR) from permeabilized cells exposed to 0.3, 1, 3, 10, 30 and 100 μM ABT-737 or HA14-1. Values for vehicle-treated cells were taken as 100%. Data represent mean ± SEM. of 4 independent experiments from 2 to 4 replicates per experiment. HA14-1 but not ABT-737 inhibited IP3R-mediated Ca²⁺ release (IC₅₀ ∼50 μM).

Our experiments clearly indicate that disrupted intracellular Ca²⁺ homeostasis is not a proximal event in ABT-737-induced thrombocytopenia. Thus, earlier studies indicating depleted intracellular Ca²⁺ homeostasis in platelets exposed for 10 μM ABT-737 for prolonged periods (e.g. 2 h) may reflect a late event that is the consequence of Bcl-Xl inhibition and ongoing cell death in platelets.⁶

We conclude that the dysregulation of intracellular Ca²⁺ signaling in platelets and human cell lines by HA14-1 is an off-target effect on SERCA2b and on IP₃Rs, since on-target inhibition of Bcl-2/Bcl-Xl by ABT-737 does not disrupt intracellular Ca²⁺ signaling. Thus, ABT-737 has a safe Ca²⁺-signaling profile, since ABT-737, at therapeutically relevant concentrations (i.e. below 1 μM), does not affect intracellular Ca²⁺-transport mechanisms essential for cellular homeostasis.