Amitriptyline may have a supportive role in cancer treatment by inhibiting glutathione S-transferase pi (GST-π) and alpha (GST-α)

Introduction

Depression is an important health problem with a high lifetime prevalence of 2–15%^Citation1. Its increased prevalence with chronic medical and oncological disorders are well documented and antidepressant drugs are widely prescribed into these patients^{Citation1,Citation2}. Various studies have been performed to elucidate the neurobiological mechanism of depression. “Monoamine hypothesis”, the main hypothesis in the depression etiology, states that depression is caused by the deficiency or underactivity of the brain monoamine neurotransmitters (serotonin, dopamine and norepinephrine) and major classes of antidepressants are mediated by increasing cathecholamine concentrations^Citation3. Tricyclic antidepressants (TCAs), the first used drugs in the depression treatment, act through blockage of neurotransmitter reuptake and increase neurotransmission at synaptic clefts^Citation4. Although newest class of antidepressants such as selective serotonin reuptake inhibitors (SSRIs) have been introduced, TCAs (amitriptyline, clomipramine and such) are still used extensively and its efficacy is equal to SSRIs^Citation5 but there are few reports on their pharmacokinetics. TCAs are taken orally, due to their lipophilic structures, are well absorbed from small intestine and they can easily cross the blood-brain barrier. Baranczyk-Kuzma et al. have shown that TCAs are widely distributed in the brain and it was suggested that their intracellular transporter might be glutathione S-transferase-pi (GST-π)6.

Glutathione-S-transferases [GSTs, E.C.2.5.1.18] constitute a Phase II enzyme family involving in the detoxification of xeno- and endobiotics by conjugating them with glutathione (the final product is a glutathione-S-conjugate) and modulates the response to carcinogen exposure. They serve maintaining normal redox homeostasis^{Citation7,Citation8}. Apart from these functions, GSTs also involve as nonenzymatic carrier proteins (ligandins) for several chemical compounds such as bilirubin, heme and are implicated in the synthesis of steroids and eicosanoids, degradation of eicosanoids and aromatic amino acids^{Citation7,Citation8}. GST isozymes were classified based on their substrate preferences, different catalytic efficiency and they are found in cytosol, microsome and mitochondria^Citation8. Altered (most often increased) expression of some cytosolic GST isozymes were implicated in the cancer etiology. Glutathione S-transferase alpha (GST-α) and glutathione S-transferase-pi (GST-π) were suggested as a potential diagnostic and/or prognostic marker for various cancer types^{Citation9,Citation10}. One of the GST cytosolic isozymes, GST-π, is a marker protein in many solid tumors and its overexpression in drug-resistant tumors has been reported whether the drug is a substrate for GST-π or not^{Citation11,Citation12}. Although the overexpression of GST-π in tumor tissues frequently linked to enhancement of drug detoxification but in most instances the selected drugs are not substrate for GST-π. This contradiction made the researchers think that this protein might have other cellular functions beside its role in catalytic detoxification^{Citation11–13}. GSTs are involved in signaling cascades as a regulator via protein-protein interaction. In Caco-2 cells, a subtype of GST-α, GSTA1-1, suppressed activation of c-Jun-N terminal kinase 1 (JNK-1) suggesting a protective role for GSTA1-1 in JNK-1 associated apoptosis^Citation14. It was also shown that GST-π inhibits JNK-1, a protein implicated in apoptosis and cellular survival^Citation15.

Oxidative and nitrosative stress are linked to various diseases and triggers the detoxification system (GSTs, glutathione peroxidase (GPx) etc.) to counteracts reactive oxygen species. There is growing evidence that tumor growth is associated with abnormal redox balance. Thiol buffer system, especially glutathione, have been therapeutically targeted in cancer treatment and a lot of redox modulating drugs are under development^Citation16. In a study performed in breast cancer patients, redox protein expressions (GST isozymes, glutathione reductase, GPx) were investigated and to predict radiotherapeutic response some of those proteins were found promising^Citation17. Nuclear transcription factor-erythroid 2 p45-related factor 2 (Nrf2), an essential transcription factor for the induction of Phase II enzymes, regulates antioxidant enzymes and proteins. Induction of GST expression can be mediated through Nrf2 expression^Citation18. Nrf2 contributes to the induction of alpha-class GST isozymes in liver of acute cadmium or manganese intoxicated rats^Citation19.

It was shown that in MCF7 breast cancer cells GST-π conjugation activity and multidrug resistance protein 1 (MRP1)-mediated conjugate efflux operate together to confer resistance to 1-chloro-2,4-dinitrobenzene (CDNB) cytotoxicity^Citation20. Interaction of antidepressants with GST-π and GST-α may reduce their availability and therapeutic activity. On the other hand, they may decrease the efficiency of the enzymatic barrier formed by GST-π and GST-α, thus increase the exposure time to toxic reactive electrophiles which may contribute to the adverse effects of antidepressant drugs.

Amitriptyline is in the group of drugs called TCAs and is used to treat symptoms of depression^{Citation4,Citation21}. Amitriptyline, like clomipramine, acts primarily as a serotonin-norepinephrine re-uptake inhibitor; its inhibitory action on the serotonin, norepinephrine and dopamine might be classified as strong, moderate and weak, respectively^{Citation4,Citation22}. Some of the cytochrome P450 isozymes are inhibited by TCAs in vitro with K_i values from micro to millimolar ranges but some of them are not affected at all^Citation23. The distribution of GST-π in brain and the binding of TCAs have been shown but there are limited kinetic studies for the clarification of the inhibitory behavior of amitriptyline on GST-π and GST-α^{Citation6,Citation24}.

The highly prescribed tricyclic antidepressant drug amitriptyline like several chemotherapeutics is taken orally, and the small intestine is the intersection organ for the absorption of both drugs. Thus, in this study, rat small intestine is used as the enzyme source. It was shown that increased/altered expression of GST-π and GST-α is associated with the resistance of the chemotherapy. The aim of this study is to elucidate the impact of the inhibitory behavior of amitriptyline which is widely used in the clinical depression treatment in the cancer patients on GST-π and GST-α. The inhibition of GSTs by amitriptyline results in the decreased cellular protective enzymatic barrier and increased exposure time to toxic electrophilic substances but amitriptyline may also alter the effectivity of chemotherapy. Based on our findings, it was also aimed to discuss the possible controversial role of amitriptyline in cancer therapy.

Materials and methods

Results and discussion

GST-π and GST-α was purified from rat small intestine and characterized according to the published procedure^Citation25.

Substrate kinetic studies were performed using the artificial substrate CDNB and GSH and all the results are given in Table 1. When CDNB used as fixed and GSH as the varied substrate the K_m values obtained for GST-π and GST-α were 0.17 ± 0.02 and 0.16 ± 0.01 mM, respectively. On the other hand, when GSH used as fixed and CDNB as the varied substrate the K_m values were 0.20 ± 0.05 mM for GST-π and 0.33 ± 0.03 mM for GST-α. The V_m value for GST-α is less affected by variable substrate than GST-π. The V_m values obtained for GST-α were 4.20 ± 0.12 and 3.85 ± 0.14 U/mg protein for [GSH] and [CDNB] as the varied substrates, respectively. On the other hand, when the varied substrate is CDNB, the V_m value for GST-π was almost one third of the V_m obtained when the varied substrate is GSH. The V_m values at varied [GSH] and at varied [CDNB] were 4.50 ± 0.04 and 1.57 ± 0.51 U/mg protein, respectively. This difference in V_m values might arise either as a result of monomer dimer conversion or by controversial effects of substrates. It seems that high concentration of GSH may cause the conversion of GST-π to a less active conformation because in experiments where [CDNB] is varied, the isozymes first meet with high concentration of GSH. This speculation obviously requires confirmation.

Amitriptyline inhibited both isozymes in a concentration dependent manner and from the % remaining activity vs. [amitriptyline] plot the IC₅₀ values of 5.54 mM (GST-π) and 8.32 mM (GST-α) were calculated (Figure 1A). Using the Scatchard plot (Log{v_i/(V_o-v_i)} vs. Log[Amitriptyline]), the same IC₅₀ values for GST-π and GST-α were obtained (Figure 1B and Table 1). Although no detailed kinetic study was performed, it seems that the inhibition of both isozymes by amitriptyline is reversible and the inhibition does not go zero at the concentration range studied. The IC₅₀ values obtained are in good correlation with the previously reported studies^{Citation6,Citation24}.

Figure 1.  Concentration-dependent inhibition of GST-π ([Amitriptyline]: 0, 0.5, 1, 2, 2.5, 5, 7.5 and 10 mM) and GST-α ([Amitriptyline]: 0, 0.5, 1, 2, 2.5, 3, 5, 7.5 and 10 mM) by amitriptyline. (A) % Remaining activity vs. [Amitriptyline]; (B) Scatchard plot Log {vi / (Vo-vi)} vs. Log [Amitriptyline]; GST-π (• ) and GST-α. (◯).

Display full size

Both isozymes, at fixed 1 mM [CDNB] and at varied [GSH], exhibited competitive inhibition pattern with amitriptyline and the 1/v vs. [Amitriptyline] plots (Dixon plot) gave straight lines intersecting on the left of the 1/v axis (Figure 2A and 2B). The competitive inhibition is further verified by the slopes vs. 1/[GSH] plot which gave straight lines passing through the origin (data not shown). The competitive type inhibition behavior implies that amitriptyline binds to the G-sites on both isozymes: GST-α and GST-π. From the statistical analysis and from the Figure 2A and 2B, the K_i values of 1.61 ± 0.17 mM and 1.45 ± 0.20 for GST-π and GST-α were obtained, respectively (Figure 2A and 2B and Table 1.) These values are in good correlation with the K_i values reported for human placental GST-π^Citation24.

Figure 2.  Inhibition of GST-π (A) and GST-α (B) by amitriptyline. Dixon plot. 1/v vs. [Amitriptyline]. [CDNB] = 1 mM, [GSH] variable. [Amitriptyline]: 0, 0.5, 1, 2 mM. [GSH]: (▪), 25; (□), 50; (▴), 100; (△), 200; (• ), 400 and (◯), 1000 µM.

Display full size

On the other hand, at fixed 1 mM [GSH] and the varied [CDNB], GST-π exhibited a noncompetitive type of inhibition pattern, whereas GST-α exhibited competitive type inhibition pattern with amitriptyline (Figure 3A and 3B). The K_i values obtained for GST-π and GST-α were 1.98 ± 0.31 and 1.57 ± 0.16 mM, respectively (Figure 3A and 3B and Table 1). The noncompetitive and competitive inhibition pattern of GST-π and GST-α are further supported by the replots of slopes vs. 1/[CDNB] obtained from Figure 3A and 3B (data not shown).

Figure 3.  Inhibition of GST-π (A) and GST-α (B) by amitriptyline. Dixon plot: 1/v vs. [Amitriptyline]. [GSH] = 1 mM, [CDNB] variable. [Amitriptyline]: 0, 0.5, 1, 2 mM. [CDNB]: (▪), 25; (□), 50; (▴), 100; (△), 200; (• ), 400 and (◯), 1000 µM.

Display full size

It was shown that all GST isozymes have at least two binding sites; one of them is the G-site, for the binding of GSH, the natural substrate and the other is the H-site(s), for binding of electrophilic xenobiotics, the artificial substrates. The competitive inhibition pattern observed with both isozymes at constant [CDNB] and at varied [GSH] implicates that amitriptyline binds to the G-site on both isozymes and creates competitive inhibition^{CitationCitation28}. On the other hand, when [GSH] is fixed and [CDNB] is the varied substrate, GST-α showed competitive and GST-π noncompetitive inhibition pattern, indicating that amitriptyline binds to H-site on GST-α, whereas on GST-π it binds to a site in H-site other than CDNB binding site. Indeed, recently it was shown that at the H-site there are at least three discrete binding sites for electrophilic xenobiotics within the xenobiotic substrate binding region and only one of them is for the binding of CDNB^Citation30. In the presence of high [CDNB], the other xenobiotic binding sites or ligandin binding site on GST-π are still available for the binding of amitriptyline (Figure 3A) but in case of GST-α it seems that amitriptyline directly binds to CDNB binding site (Figure 3B). This argument is strengthened by the slope vs. 1/[CDNB] plots; in GST-α the straight line obtained passes through the origin, whereas in GST-π it intersects with the slope axis indicating noncompetitive type of inhibition (data not shown). The difference in inhibition pattern of two isozymes could be explained by the structural differences; although GST-α has C-terminal extension in which the presence of GSH folds to give the helix α9 and partly occludes the H-site but it seems that this occlusion does not affect CDNB binding^Citation31.

It was shown by Baranczyk-Kuzma et al. that at fixed [CDNB]-varied [GSH] and at fixed [GSH]-varied [CDNB] the inhibition pattern with amitriptyline was noncompetitive for all GST-π’s isolated from different parts of the brain^Citation6. Dalmizrak et al. reported that at fixed CDNB-varied GSH and at fixed GSH-varied CDNB concentrations the inhibition patterns of human placental GST-π by amitriptyline were linear mixed-type and noncompetitive, respectively^Citation24. In our study, both isozymes were inhibited competitively when the varied substrate is GSH (Figure 2A and 2B), whereas the varied substrate CDNB, GST-π was inhibited noncompetitively but GST-α was inhibited competitively (Figure 3A and 3B). The kinetic differences between human GST-π and rat GST-π, GST-α may arise from being different isozymes or/and could be explained by different isoforms of the GSTs^{Citation6,Citation24}.

It is well known that oxidative and nitrosative stress constitute the basis of various psychological stress-induced disorders such as depression and antidepressants are used to treat these conditions^{Citation32,Citation33}. GSTs, like other antioxidant enzymes are important in the detoxification of xenobiotics. The increase in antioxidant enzyme levels by antidepressants was interpreted by the neuroprotective pharmacological effects of these drugs^Citation33. Indeed, the induction of their mRNAs by very low concentrations (10⁻⁶–10⁻⁷ M) of antidepressants was reported^Citation34. The serum concentration of amitriptyline was found to be 168 nM much higher than their inducer concentrations^Citation35.

We conclude that, amitriptyline in addition to its antidepressant effect, might also have a minor supportive role on the effectiveness of the anticancer drugs by decreasing their elimination through the inhibition of GST-π and GST-α but this also means extended exposure of the patient to toxic electrophiles. The elucidation of all these controversial findings require further studies.

Declaration of interest

This study was supported by H.U.B.A.B. Grant (06D061011001).