Inhibition mechanism of the intracellular transporter Ca2+-pump from sarco-endoplasmic reticulum by the antitumor agent dimethyl-celecoxib

Abstract

Dimethyl-celecoxib is a celecoxib analog that lacks the capacity as cyclo-oxygenase-2 inhibitor and therefore the life-threatening effects but retains the antineoplastic properties. The action mechanism at the molecular level is unclear. Our in vitro assays using a sarcoplasmic reticulum preparation from rabbit skeletal muscle demonstrate that dimethyl-celecoxib inhibits Ca2+-ATPase activity and ATP-dependent Ca2+ transport in a concentration-dependent manner. Celecoxib was a more potent inhibitor of Ca2+-ATPase activity than dimethyl-celecoxib, as deduced from the half-maximum effect but dimethyl-celecoxib exhibited higher inhibition potency when Ca2+ transport was evaluated. Since Ca2+ transport was more sensitive to inhibition than Ca2+-ATPase activity the drugs under study caused Ca2+/Pi uncoupling. Dimethyl-celecoxib provoked greater uncoupling and the effect was dependent on drug concentration but independent of Ca2+-pump functioning. Dimethyl-celecoxib prevented Ca2+ binding by stabilizing the inactive Ca2+-free conformation of the pump. The effect on the kinetics of phosphoenzyme accumulation and the dependence of the phosphoenzyme level on dimethyl-celecoxib concentration were independent of whether or not the Ca2+-pump was exposed to the drug in the presence of Ca2+ before phosphorylation. This provided evidence of non-preferential interaction with the Ca2+-free conformation. Likewise, the decreased phosphoenzyme level in the presence of dimethyl-celecoxib that was partially relieved by increasing Ca2+ was consistent with the mentioned effect on Ca2+ binding. The kinetics of phosphoenzyme decomposition under turnover conditions was not altered by dimethyl-celecoxib. The dual effect of the drug involves Ca2+-pump inhibition and membrane permeabilization activity. The reported data can explain the cytotoxic and anti-proliferative effects that have been attributed to the celecoxib analog. Ligand docking simulation predicts interaction of celecoxib and dimethyl-celecoxib with the intracellular Ca2+ transporter at the inhibition site of hydroquinones.

Figure 1. Depedence of Ca²⁺ –ATPase activity on drug concentration in leaky vesicles.

The catalytic activity was measured at 25°C in a reaction medium containing 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl₂, 1 mM EGTA, 0.686 mM CaCl₂, 4 µM A23187, 0.05 mg/ml SR vesicles and 1 mM ATP. A given concentration of DMC (blue circles) or CLX (red circles) was included when indicated. The rate of P_i release was measured by a discontinuous colorimetric method.

Figure 2. Drug effect on Ca²⁺-ATPase activity, active Ca²⁺ transport and Ca²⁺/P_i coupling.

Reaction mixtures included native SR vesicles and 5²⁺-ATPase activity (A) or Ca²⁺ transport (B) on DMC (blue circles) or CLX (red circles). Inset, coupling ratios in the presence of DMC (blue bars) or CLX (red bars).

Figure 3. Drug effect on Ca²⁺-loaded vesicles.

The Ca²⁺ loading medium was 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl₂, 0.2 mM EGTA, 0.138 mM [⁴⁵Ca]CaCl₂, 0.05 mg/ml native SR vesicles and 1 mM ATP. Reaction medium aliquots were filtered at different times and processed to evaluate ⁴⁵Ca²⁺ retained by the vesicles. (A) Samples were left untreated (green squares) or supplemented with 50 µM DMC (blue circles) or 50 µM CLX (red circles) after 2 min of reaction. (B) The addition after 2 min was 5 mM EGTA (green squares), 5 mM EGTA plus 50 µM DMC (blue circles) or 5 mM EGTA plus 50 µM CLX (red circles). Times of addition are marked by an arrow. Inset, rapid component of Ca²⁺ release in the presence of DMC (blue bars) or CLX (red bars) when the Ca²⁺-pump was arrested. Data in the inset correspond to the first time point after drug addition once the EGTA-induced component was subtracted.

Figure 4. Effect of DMC on Ca²⁺ binding to the transport sites.

The incubation medium consisted of 20₂, 0.1 mM EGTA, 1 mM glucose, 0.2 mg/ml native SR vesicles in the absence (green squares) or presence (blue circles) of 50 µM DMC. Aliquots of 0.6 ml were mixed with a certain volume of medium containing 20 mM Mops, pH 7.0, 80 mM KCl, 3 mM MgCl₂, 0.1 mM EGTA, 3 mM ⁴⁵CaCl₂ and 1 mM [³H]glucose. Samples were processed to determine specific ⁴⁵Ca²⁺ bound.

Figure 5. Effect of DMC on the steady-state EP level.

The experiments were performed at ice-water temperature. The final reaction mixture was 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl₂, 1 mM EGTA, 0.686 mM CaCl₂, 0.2 mg/ml SR vesicles, 8 µM A23187, 40 µM [γ-³²P]ATP and a given DMC concentration when indicated. SR vesicles in the presence of Ca²⁺ were exposed to DMC and then mixed with radioactive ATP (orange circles) or SR vesicles in the absence of Ca²⁺ were exposed to DMC and then mixed with Ca²⁺ plus radioactive ATP (purple circles). The reaction was acid quenched after 5 s. Inset, time course of EP accumulation when the vesicles were exposed to 30 µM DMC in the presence (orange circles) or absence (purple circles) of Ca²⁺ before phosphorylation.

Figure 6. EP levels in the presence of DMC at various Ca²⁺ concentrations.

The final reaction medium at ice-water temperature was 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl₂, 1 mM EGTA, CaCl₂ to give a certain pCa, 0.2 mg/ml SR vesicles, 8 µM A23187, in absence or presence of 50 µM DMC and 40 µM [γ-³²P]ATP. Phosphorylation was initiated by adding radioactive ATP to SR vesicles in the presence of Ca²⁺ and stopped 5 s later by acid quenching. The EP level at each pCa in the absence of DMC gaves the corresponding 100% value in the ordinate axis.

Figure 7. Kinetics of EP decay under turnover conditions.

The experiments were performed at ice-water temperature. The phosphorylation medium was 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl₂, 1 mM EGTA, 0.686 mM CaCl₂, 0.4 mg/ml SR vesicles, 16 µM A23187 and 40 µM [γ-³²P]ATP in the absence (green circles) or presence (blue circles) of 50 µM DMC. After 5 s, the reaction mixture was 10-fold diluted with a medium containing non-radioactive ATP plus Ca²⁺ and EP was quenched by acid at different times.

Figure 8. Computational docking of CLX and DMC compared to BHQ in the SERCA protein.

Electrostatic potential surface of the BHQ binding cavity with CLX (A) or DMC (C) bound. Protein residues polarity, from non-polar to polar, is shown in a scale from red to white and blue. CLX (brown) and DMC (magenta) are shown in stick representation. BHQ (yellow) is included as a reference. Detail of SERCA residues at the BHQ site involved in interaction with CLX (B) or DMC (D). Interacting residues are denoted by grey sticks and protein structure was removed for the sake of clarity. Docking-predicted poses correspond to binding conformations with the lowest free energy change.