Investigation of Pharmacokinetic and Pharmacodynamic Interactions between Citalopram and Duloxetine: An Integrated Analytical, Computational, Behavioral, and Biochemical Approach

Abstract

Despite the prevalent utilization of antidepressant combinations in clinical settings, concerns persist regarding heightened side effects and potential drug–drug interactions (DDI). In response, this study investigates the interaction between citalopram (CIT) and duloxetine (DUL) using a multifaceted approach encompassing analytical, computational, behavioral, and biochemical techniques. Notably, the absence of published analytical methods tailored for studying antidepressant interactions underscores the novelty of our endeavor. We present the development and validation of a robust and sensitive assay, coupling liquid chromatography-tandem mass spectrometry. This method facilitates the simultaneous determination of DUL, a serotonin-norepinephrine reuptake inhibitor (SNRI), and CIT, a selective serotonin reuptake inhibitor (SSRI), in rat plasma following oral administration. Successful pharmacokinetic and DDI monitoring of DUL and CIT in rat plasma post a single oral dose of 120 mg/kg is achieved using this method. Our findings reveal DUL’s influence on CIT’s pharmacokinetic parameters, resulting in an increased area under the concentration–time curve (AUC) by 4-fold, peak plasma concentrations (C_max) by 20-fold, maximum plasma concentration–time (T_max) by 4-fold, and oral clearance (Cl/F) of CIT by 1.3-fold upon coadministration. Furthermore, our investigation explores the behavioral and biochemical ramifications of coadministering CIT and DUL through the sucrose preference test (SPT), forced swimming test (FST), and enzyme-linked immunosorbent assay (ELISA). We observe potential exacerbation of serotonin concentration and serotonin syndrome in rat models. Molecular modeling studies indicate that DUL may competitively inhibit CYP2D6, the principal enzyme responsible for CIT metabolism, as well as P-glycoprotein (P-gp), which extrudes CIT back to the intestinal lumen. These findings emphasize the imperative of further research into potential DDIs in psychiatric patients undergoing chronic treatment with DUL and CIT to mitigate adverse effects and serotonin syndrome.

Depression is a complex neurological disorder characterized by persistent feelings of sadness, anhedonia, sleep disturbances, cognitive impairments, and suicidal ideation, affecting approximately 280 million people globally, according to the World Health Organization (WHO).^1,2 It remains a leading cause of disability, with suicide ranking as the fourth leading cause of death among individuals aged 15 to 29.³ Treatment typically involves cognitive-behavioral therapy and pharmacotherapy. Among antidepressants, selective serotonin reuptake inhibitors (SSRIs) like citalopram (CIT) are widely prescribed due to their favorable side effect profile, as recommended by the National Institute for Health and Care Excellence (NICE) guidelines.^4,5

Despite the effectiveness of monotherapy, a significant number of patients experience suboptimal responses. To enhance therapeutic outcomes, clinicians frequently combine antidepressants, even though the neuropharmacological rationale behind such combinations is often weak.⁴ This practice raises the risk of increased side effects without necessarily providing added benefits. Therefore, understanding the pharmacokinetic (PK) and pharmacodynamic (PD) interactions of combined antidepressant therapy is critical for minimizing adverse effects.

CIT (Figure 1), an SSRI approved by the FDA in 1998,^6,7 is primarily used to treat depression but is also effective for off-label conditions like generalized anxiety disorder and obsessive-compulsive disorder.⁸⁻¹¹ Pharmacokinetically, CIT undergoes hepatic metabolism mainly via CYP3A4, CYP2C19, and CYP2D6 enzymes.¹² Duloxetine (DUL) (Figure 1), an serotonin-norepinephrine reuptake inhibitor (SNRI) approved in 2004 for major depressive disorder, also treats conditions like diabetic neuropathy and stress urinary incontinence.^13,14 DUL is metabolized predominantly by CYP1A2 and CYP2D6, with renal excretion of its metabolites.¹³

Figure 1.

Chemical structures and chemical names of the studied antidepressants.

Despite the availability of numerous antidepressants, a substantial proportion of patients exhibit poor response to initial monotherapy.¹⁵ Consequently, antidepressant combinations have emerged as a viable strategy to enhance treatment efficacy by targeting multiple pathways. However, the rationale behind certain combinations lacks neuropharmacological justification, potentially increasing patients’ side effect burden without commensurate benefits.^4,15 Currently, clinical practice involves combining SNRIs like venlafaxine with SSRIs, yielding improved outcomes compared to monotherapy.^16,17 However, the lack of studies on DUL in conjunction with SSRIs warrants cautious consideration due to its CYP2D6 inhibitory effects and the potential for serotonin syndrome.¹⁸

Combining SSRIs and SNRIs, such as CIT and DUL, is becoming more common, despite limited data on their interactions.¹⁵ DUL, a known inhibitor of CYP2D6,¹⁹ may impact CIT metabolism, potentially leading to serotonin syndrome. This highlights the importance of studying DDIs, particularly because CIT is also a substrate of P-glycoprotein (P-gp),²⁰ which mediates drug efflux in the intestines. Drug–drug interactions (DDIs) can be categorized as pharmacokinetic (PK-DDI) or pharmacodynamic (PD-DDI). PK-DDIs occur when one drug alters the absorption, metabolism, or excretion of another, such as through CYP2D6 inhibition. PD-DDIs involve additive or synergistic effects at shared drug targets, potentially exacerbating serotonin levels.^19,20 Notably, although antidepressant combinations like venlafaxine (an SNRI) with SSRIs are well-documented,²¹ studies specifically on the combination of DUL with SSRIs are limited.²²

In our previous research,²³ we examined the toxicity and pharmacokinetics of CIT in combination with another SSRI, sertraline (SER) (Figure 1), in rat models. Building on these findings, we aim to investigate the interaction between CIT and DUL, expanding our scope to an SSRI-SNRI combination. The lack of published analytical methods tailored for monitoring DDIs between CIT and DUL represents a critical gap. This study addresses that gap by developing a robust assay for the simultaneous quantification of CIT and DUL in biological fluids, providing insights into both pharmacokinetic and pharmacodynamic interactions.

In addition to analytical methods, our research includes molecular modeling to explore competitive inhibition between CIT and DUL at CYP2D6 and P-gp, key sites for DDIs. Behavioral tests, including the sucrose preference test (SPT) and forced swimming test (FST), along with enzyme-linked immunosorbent assays (ELISA),²⁴ were employed to assess serotonin concentration and the potential onset of serotonin syndrome in rat models. Our findings underscore the importance of thoroughly investigating antidepressant combinations to optimize therapeutic interventions and minimize adverse effects in depression treatment.

Results

In this investigation, we explored the potential DDI between CIT and DUL utilizing a multifaceted approach encompassing both pharmacokinetic and pharmacodynamic analyses. To our knowledge, this study represents the first utilization of molecular modeling techniques coupled with pharmacokinetic modeling to delve into the DDI between CIT and DUL. Initially, we developed a robust PBPK model in rodents to forecast the potential interactions between these antidepressants. Our simulations revealed that CIT and DUL exhibit complex metabolic interactions, with DUL displaying inhibitory effects on CIT metabolism at clinically relevant doses. Subsequently, we ventured into elucidating the pharmacodynamic interactions between CIT and DUL by examining their impact on neurotransmitter systems implicated in depression. Molecular modeling studies unveiled intriguing insights into the competitive binding of CIT and DUL at crucial binding sites within CYP2D6 enzyme and P-gp, shedding light on the intricate interplay between these drugs at the molecular level. Notably, our findings suggest a potential synergistic or additive effect when CIT and DUL are coadministered, highlighting the significance of pharmacodynamic mechanisms in mediating their DDI.

LC–MS/MS Method Development

Optimizing the Separating Conditions

The hydrophobic nature of these compounds makes it challenging to distinguish between SSRIs and SNRIs in intricate matrices and drug formulations. Both analytical and toxicological chemistry encounter this issue. The preferred technique for analyzing target analytes in various matrices is still reversed-phase chromatography. C18 (octadecyl silane), C8 (octyl silane), and cyanide (CN) were taken into consideration while selecting which reversed-phase columns to use in order to isolate the target SSRIs and SNRIs in rat plasma. With the Agilent ZORBAX Eclipse Plus C18 column, ideal outcomes were attained, including symmetric peak form, enhanced resolution, and acceptable separation time. Because of its unique architecture, this column is ideal for rapidly separating basic chemicals, such as DUL, CIT, and internal standard (IS) SER, which have pK_a values of 9.48, 10.02, and 9.5, respectively.²³ Furthermore, a variety of mobile phases were assessed, including buffers with varying pH levels and other chemical modifiers like methanol and acetonitrile (ACN). Because of its simplicity, isocratic elution was first used to separate the SSRIs; and SNRIs but, because of their identical hydrophobic qualities, this method proved insufficient. Gradient elution was therefore used to obtain the best possible resolution and separation. The target analytes and the IS were efficiently eluted using a gradient elution scheme consisting of formic acid (0.1%) and ACN. ACN outperformed other organic modifiers in terms of peak intensity and resolution. Under these optimized chromatographic conditions DUL, CIT, and the IS were eluted at 10.64 ± 0.04, 10.02 ± 0.03, and 10.31 ± 0.02 minute (min), respectively (Figures S1 and S2).

Optimizing the Tandem Mass (MS/MS) Conditions

Currently, therapeutic drug monitoring (TDM), drug analysis, metabolite analysis, and bioequivalence investigations in biological contexts are best done by mass spectrometry. By directly infusing 500 ng/mL of DUL, CIT, and SER, multiple reaction monitoring (MRM) conditions were set up. This made it possible to use electrospray ionization (ESI) sources in positive ion modes to identify the protonated precursor ion and the most intense and stable product ion within a mass range of 100–500 amu for DUL, SER, and 100–1000 amu for CIT. MRM was chosen over selected ion monitoring due to its improved reproducibility, decreased matrix interference, and increased selectivity. Additionally, by utilizing only ESI in positive ion mode as a mass spectrometry interface, the best signal-to-noise ratio (S/N), specificity, and accuracy were ensured. The most stable and intense precursor ions (Q1) [M + H]⁺¹ for CIT, DUL, and SER were detected at m/z 325.2, 298.1, and 306.1, respectively. The most significant and stable product ions for antidepressant quantification and procedure validation were also discovered for CIT, DUL, and SER at m/z 325.2 to 262.0 and 109.0, 298.1 to 154.1, and 306.1 to 158.9, respectively (Figure S3). These product ion peaks were attributed to 1-fluoro-4-methylbenzene radical and N-methyl-3-(thiophen-2-yl) propan-1-amine, formed from the respective precursors of CIT and DUL, respectively (Figure S3).

Sample Cleanup and Analytes Extraction

Three purification techniques were used to remove endogenous compounds that interfered with rat plasma samples: protein precipitation (PP), liquid–liquid extraction (LLE), and solid-phase extraction (SPE). At first, PP was selected because it worked well with the target analytes’ hydrophobic properties. However, PP showed limited recovery and poor selectivity, especially at the lower limit of quantification (LLOQ), which led to the investigation of substitute techniques. Although ethyl acetate LLE demonstrated encouraging specificity, its intricacy and time-consuming nature made it unfeasible. Afterward, SPE was assessed as an extraction and purification method for the intended antidepressants. Oasis HLB (hydrophilic–lipophilic balance) sorbent was chosen because of its exceptionally consistent recovery. An HLB cartridge was cleaned and conditioned, and the sample was loaded, cleaned, and eluted as part of the plasma sample preparation procedure. Since the target analytes and IS were in salt form and might have eluted during the washing step, alkalizing the sample with 50 μL of 1 N NaOH was essential to releasing the parent medicines. Alkalization also improved the release of medicines from plasma, especially those that bound to plasma proteins strongly. Typical chromatograms showed that the suggested SPE approach had great selectivity and specificity. In an effort to maximize recovery, many elution systems were evaluated; the system that used 300 μL of neat ACN produced the highest recovery rates. The technique recovery varied from 87 to 103% under these ideal circumstances, guaranteeing precise quantitative detection of CIT and DUL in rat plasma (Table S1).

Bioanalytical Method Validation

The LC–MS/MS system validation was conducted in accordance with the guidelines provided in the Experimental Section and the U.S. Industry Advice on bioanalytical process validation. Plotting relative ion intensity against target analytes concentration revealed a linear relationship under optimal conditions. Based on peak-area ratios (peak area of DUL or CIT/peak area of SER) as a function of concentration, the calibration curve showed linearity in the ranges of 1–2000 ng/mL for CIT and 1–500 ng/mL for DUL, with correlation coefficients (r) greater than 0.9997. The mean regression equations derived from seven concentrations over 3 days were Y = 156.672764 × X – 559.074572 for CIT and Y = 1.568520 × X + 13.917036 for DUL, where Y represents the ratio of CIT or DUL peak area to that of SER, and X represents the plasma concentration of the selected drugs in ng/mL. The limits of quantification (LOQ) and detection (LOD) for CIT and DUL were determined to be 0.37 and 0.51 and 0.11 and 0.15 ng/mL, respectively, with relative standard deviations (RSD) <20% for S/N of 3. The impact of the matrix was evaluated by examining possible ion enhancement/suppression arising from endogenous plasma components. The IS-normalized matrix factors ranged from 1.02 to 1.13 for DUL and from 1.05 to 1.12 for CIT. Accuracy and precision of the procedure were assessed at low, moderate, and high concentrations of DUL and CIT, within and between days. The accuracy of the approach was assessed using the relative standard error (RE%) for CIT and DUL, which ranged from −13 to 2.3% and −4.6 to 3%, respectively. The recommended method showed strong reproducibility, with relative standard deviations (RSD%) ranging from 1.06 to 11.07% for DUL and from 0.27 to 20.8% for CIT across four quality control (QC) samples (Table S1). Compared to previous methods for determining CIT and DUL, the proposed method offers enhanced sensitivity, simplicity, plasma volume, and sample processing efficiency.^{25 −31}

Stability of CIT and DUL in Rat Plasma

In order to evaluate the stability of CIT and DUL when endogenous biological plasma components were present, multiple storage conditions were applied using the novel LC-MS/MS technology. These comprised 1 h room temperature (RT) stability, one-month-long storage stability at −80 °C, one-day refrigeration stability at 4 °C, and stability following three thawing cycles. The stability of DUL and CIT in rat plasma is summarized in Table S2. The target analytes showed no discernible degradation after 1 h at room temperature, according to the data, which showed relative standard deviation (RSD%) values ranging from 0.27 to 14.1%. This length of time worked well for the pretreatment of the sample. Additionally, both medications showed stability after a day of refrigeration at 4 °C, with RSD% values ranging from 0.15 to 13%—the amount of time needed for sample injection. Furthermore, after a month of freezing at −80 °C, both CIT and DUL showed respectable stability, with RSD% values ranging from 0.34 to 10.6%. Furthermore, during three thaw-freeze cycles, the medications showed good stability, with RSD% values ranging from 0.4 to 13.3% (Table S2).

Pharmacokinetic Studies in Rats

The current work completely validates the effectiveness of the LC-MS/MS technique in measuring the pharmacokinetic characteristics of DUL and CIT in rat plasma when the two drugs are taken orally in doses of 120 mg/kg, either separately or in combination. The primary PKSolver-calculated pharmacokinetic parameters for CIT and DUL are shown in Table 1, and the mean plasma concentration–time curves are shown in Figure 2. CIT showed a peak plasma concentration (C_max) of 71.6 ng/mL at a relatively short time (T_max) of 1 h, with a significant area under the curve (AUC_0–12) of 203.9 ng/mL·h. It demonstrated a brief terminal half-life (t_1/2) of 2.05 h and a mean residence time of 3.41 h, owing to its wide volume of distribution (V/F) of 71.7 L and high plasma clearance of 48.9 L/h. Conversely, DUL demonstrated an AUC_0–12 of 220.93 ng/mL h, and after 2 h (T_max), it reached a C_max of 39.79 ng/mL, suggesting accumulation in tissues. Table 1 shows that DUL has a moderate plasma clearance of 28.3 L/h and a broad distribution of 63.4 L, along with a moderate terminal half-life (t_1/2) of 4.43 h. The obtained pharmacokinetic parameters closely align with those reported in previously published methods.^25,27−29

Table 1. Pharmacokinetic Parameters of CIT and DUL in Rat Plasma after Oral Administration of Single Dose and Combined.

		DUL (120 mg/kg)		CIT (120 mg/kg)
Parameter	Unit	Alone	With CIT (120 mg/kg)	Alone	With DUL (120 mg/kg)
λz	1/h	0.31 ± 0.22	0.23 ± 0.09	0.33 ± 0.12	0.76 ± 0.15
Ka	1/h	0.474 ± 0.18	0.39 ± 0.13	0.7 ± 0.23	0.4 ± 0.21
t1/2	h	4.43 ± 0.36	5.18 ± 0.88	2.05 ± 0.24	0.91 ± 0.12
Tmax	h	2 ± 0.27	2 ± 0.51	1 ± 0.16	4 ± 0.86
Cmax	ng/mL	39.79 ± 2.13	78.34 ± 7.3	71.6 ± 3.1	1431.8 ± 77.4
AUC0–t	ng/mL·h	220.93 ± 12.1	493.37 ± 16.3	203.92 ± 17.4	8782.26 ± 133.5
AUC0–inf	ng/mL·h	276.27 ± 6.4	652.8 ± 12.4	210.75 ± 11.5	8794.95 ± 210.3
MRT	h	7.58 ± 0.75	8.6 ± 0.78	3.41 ± 1.5	4.97 ± 1.1
Kel	1/h	0.45 ± 0.18	0.36 ± 0.11	0.4 ± 0.11	0.68 ± 0.16
V/F	L	63.4 ± 4.5	23.9 ± 3.2	71.7 ± 12.4	1 ± 0.22
Cl/F	L/h	28.3 ± 2.43	8.7 ± 1.3	48.9 ± 4.2	64.7 ± 3.2

Figure 2.

Mean plasma concentration time curve of (A) CIT after oral administration of CIT alone and coadministration of CIT and DUL; (B) DUL after oral administration of DUL alone and coadministration of DUL and CIT.

Pharmacokinetic Drug–Drug Interaction (PK-DDI)

The prevalence of depression and other mental illnesses is rising globally at the moment, which frequently calls for changing medication dosages or implementing combination therapy in order to achieve effective treatment. The narrow therapeutic index of many antidepressants, however, poses a significant obstacle to depression therapy and raises the possibility of drug toxicity.³²

As a result, tracking these medications’ plasma concentrations is essential for assessing and modifying treatment plans. In order to better understand potential toxicity, we have developed a novel analytical method in this study that concurrently determines the plasma levels of DUL and CIT in rats after oral coadministration, utilizing tandem mass spectrometry in conjunction with liquid chromatography. Our results revealed significant increases in the C_max, area under the concentration–time curve (AUC_0–12), and the time to reach maximum plasma concentration (T_max) of CIT, with percentage increases of 1890%, 4210%, and 300%, respectively. Similarly, both C_max and AUC_0–12 of DUL showed considerable increases of 96% and 124%, respectively, in rats administered both DUL and CIT. The prolonged T_max of CIT may be attributed to a decrease in the absorption rate constant (K_a) by 43%. Furthermore, the increased C_max and AUC of both drugs may result from a decrease in the metabolism rate. Additionally, the decreased half-life (t_1/2) of CIT by 125% could be attributed to an increase in the elimination rate constant (K_el) and oral clearance (Cl/F) by 70% and 32%, respectively. Given the established inhibitory effects of SER, paroxetine, fluoxetine, and DUL on CYP2D6,³² which is the predominant cytochrome P450 enzyme responsible for metabolizing, we anticipated that the primary site of PK-DDI would occur during the metabolic process. To validate this hypothesis, molecular modeling simulations were employed, as elaborated in the subsequent section.

Molecular Modeling Study

To predict the binding affinity between CIT and DUL and their respective targets, as well as to analyze the characteristics of the binding sites, docking analysis was conducted targeting CYP2D6 and P-gp, which are the main anticipated sites for DDI. The docking results indicated a docking score of −8.74 and glide emodel of −71.7 Kcal/mol for DUL on CYP2D6, whereas, for CIT, the docking score was −7.73 and glide emodel was −64.7 Kcal/mol for the same target. A lower docking score is indicative of better docking performance. The binding modes of both drugs demonstrated significant accommodation within the substrate-binding pocket of CYP2D6, aligning closely with the crystallized ligand (prinomastat) (Figure 3A). Various ligand-target interactions were observed, including an ionic bond between the secondary amine of DUL and ASP 301, and a π–π stacking interaction between the thiophene ring of DUL and the heme, as well as between the aromatic ring of CIT and PHE120 (Figure 3B,C). The presence of a strong ionic bond in DUL conferred superior binding compared to CIT, confirming DUL’s inhibitory activity on CYP2D6, which prevents the metabolism of CIT, thus explaining the observed increase in CIT’s Cmax and AUC when both drugs are administered together. Conversely, the docking score and binding energy of CIT toward CYP2D6 were comparable to those of DUL, suggesting potential competition between CIT and DUL for interaction with CYP2D6, albeit to a lesser extent, which may account for the moderate increase in DUL’s C_max and AUC upon coadministration with CIT.

Figure 3.

(A) 3D superimposition of CIT and DUL in the crystal of cytochrome P450 2D6 (PDB ID 3TDA); (B) 2D ligand interaction of DUL within substrate binding site of CYP2D6; (C) 2D of ligand interaction CIT within substrate binding site of CYP2D6.

As CIT is known to be a substrate of P-gp, the concurrent administration of P-gp inducers or inhibitors may impact the levels of P-gp substrates in the bloodstream. P-gp functions by actively transporting substrates back into the intestinal lumen from the cytosol of enterocytes. Drugs such as fluoxetine, fluvoxamine, and paroxetine, acting as P-gp inhibitors, can potentially lead to significant increases in blood levels of these agents, thereby elevating the risk of adverse outcomes. This risk is particularly heightened when these medications are administered alongside P-gp substrates.³³ In our study, coadministration of DUL with CIT resulted in increased Cmax and AUC of CIT, suggesting a possible P-gp inhibitory effect of DUL. Molecular modeling of paroxetine, well-known P-gp inhibitors, and DUL on P-gp revealed significant accommodation within P-gp substrate-binding pocket (Figure 4A), with docking scores and binding energies of DUL larger than those of paroxetine suggesting superior inhibitory power of DUL than that of paroxetine. Paroxetine exhibited a docking score of −8.06 and a glide emodel of −53.19 kcal/mol, while DUL demonstrated a docking score of −8.48 and a glide emodel of −55.44 kcal/mol.

Figure 4.

(A) 3D superimposition of paroxetine and DUL within the binding pocket of P-glycoprotein (P-gp) (PDB ID 4XWK); (B) 2D ligand interaction of paroxetine within binding pocket of P-gp; (C) 2D ligand interaction of DUL within binding pocket P-gp.

Several ligand-target interactions were observed, including an ionic-π interaction between the secondary amine of paroxetine and TYR303, and a π–π stacking interaction between the phenyl group and PHE728. Similarly, DUL demonstrated similar interactions, forming an ionic-π interaction between the secondary amine of DUL and PHE 728, as well as π–π stacking interactions between the naphthyl group and PHE728, and the thiophene ring and TYR303 (Figure 4B,C).

Based on the docking results, we can tentatively infer that DUL could potentially compete with CIT for the same binding pockets of P-glycoprotein (P-gp), while CIT might similarly compete with DUL for the same binding pocket of cytochrome P450 2D6 (CYP2D6). However, it is essential to note that docking experiments provide preliminary insights, and further, in vivo studies are warranted to confirm these interactions conclusively, as docking results are limited by their static nature and dependence on the resemblance of the binding pocket to the native ligand’s.³⁴

Pharmacodynamic Studies in Rats

Effects of CIT, DUL, and the Combination of Both Drugs on the Dexamethasone-Induced Depressive-Like Behavior

To investigate the potential induction of depressive-like behavior in rats following extended administration of dexamethasone (DEX) (1.5 mg/kg, i.p.), FST was employed. The results depicted in Figure 5A demonstrated significant treatment effects based on one-way ANOVA [F (4, 25) = 152.7, p < 0.0001]. Posthoc analysis using Tukey’s multiple comparisons test revealed a 4.5-fold increase in immobility time in the DEX-treated group compared to the control group (p < 0.0001), suggesting that DEX injections triggered depressive-like behavior. Conversely, treatment with the antidepressant CIT and DUL significantly reduced immobility time by 54% and 46%, respectively, compared to the DEX group (p < 0.0001). Notably, there was no distinction in immobility between the DEX-treated and combination therapy groups, implying that the combined treatment led to a sharp increase in serotonin levels, consequently causing serotonin syndrome, which paradoxically worsened depressive behavior in the rats.

Figure 5.

Effects of CIT, DUL, and the combination of both drugs on the dexamethasone-induced depressive-like behavior as assessed in the forced swimming test (A) and anhedonia as evaluated by the sucrose preference test (B). Effect of repeated dexamethasone administration on serotonin level (C). Data are represented as mean ± SD (n = 6) using one-way ANOVA followed by Tukey’s multiple comparison test at p < 0.0001. CIT, citalopram; CTRL, control; DEX, dexamethasone; DUL, duloxetine.

On the other hand, anhedonia is recognized as a fundamental symptom associated with depressive-like behavior. An evaluation of anhedonia in rats was conducted using the SPT. As depicted in Figure 5B, one-way ANOVA revealed significant treatment effects [F (4, 25) = 118.4, p < 0.0001]. Subsequent posthoc analysis via Tukey’s multiple comparisons test indicated that DEX treatment led to a 45% reduction in sucrose intake compared to the control group (p < 0.0001). Conversely, treatment with the antidepressants CIT and DUL resulted in an improvement in the hedonic state, increasing sucrose intake by 1.5 and 1.4-fold, respectively, compared to the DEX group (p < 0.0001). Notably, there was no discernible difference in the level of anhedonia between the DEX-treated group and the combination therapy group.

Effects of CIT, DUL, and the Combination of Both Drugs on the Dexamethasone-Induced Changes in Serotonin Level

As depicted in Figure 5C the DEX group exhibited a substantial decrease in serotonin levels [F (4, 25) = 122.7, p < 0.0001], approximately 3.2 times (69%) lower than the control group. However, both CIT and DUL treatments restored serotonin levels by approximately 2.17 and 2.29 times (p < 0.0001), respectively, when compared to the DEX group. Interestingly, the combination group displayed a significant rise in serotonin levels compared to both CIT and DUL individually (p < 0.0001), suggesting a potential occurrence of serotonin syndrome.

Discussion

The majority of second-generation antidepressants interact with drugs that either activate or inhibit the enzymes involved in their biotransformation pathways because they undergo a substantial degree of oxidative metabolism in the liver.³⁵ Therefore, concurrent use of both drugs may alter the rate at which second-generation antidepressants are excreted, thereby influencing plasma concentrations and clinical effects. Notably, the risk of pharmaceutical interactions is increased since general practitioners are increasingly prescribing certain second-generation antidepressants, like venlafaxine and SSRIs, for mental illnesses other than depression, like anxiety disorders.²¹ Despite the widespread use of these agents, the prevalence of clinically relevant drug interactions with antidepressants remains inadequately characterized. Thus, our study aims to elucidate the impact of coadministering two second-generation antidepressants on each other’s pharmacokinetic and pharmacodynamic profiles, filling a critical gap in current knowledge regarding antidepressant combination therapy. The findings of our study shed light on the intricate pharmacokinetic interactions between CIT and DUL, two commonly prescribed antidepressants, and underscore the necessity for comprehensive evaluation when considering combined therapy in depression management. Through rigorous pharmacokinetic analysis, we observed significant alterations in the plasma concentrations of both CIT and DUL when administered in combination, indicative of a PK-DDI. Specifically, the considerable increases in C_max and AUC of CIT suggest a potentiation of its pharmacological effects, possibly due to inhibition of its metabolism by DUL, as evidenced by molecular modeling simulations targeting CYP2D6. Conversely, while DUL also exhibited increased C_max and AUC_0–12 when coadministered with CIT, the effect was less pronounced, possibly due to competitive inhibition between the two drugs at the same binding site of CYP2D6. On the other hand, molecular docking studies validated DUL’s strong affinity for inhibiting P-gp even more effectively than the established P-gp inhibitor, paroxetine. Our findings are consistent with previous reports demonstrating DUL’s ability to inhibit P-gp function both in vitro and in vivo.³⁶ Consequently, caution is warranted when coadministering DUL with drugs that are substrates of P-gp. These findings emphasize the importance of monitoring plasma drug concentrations and considering potential PK-DDIs when designing antidepressant treatment regimens, particularly in cases of combination therapy.

In addition to pharmacokinetic interactions, our study elucidates the behavioral and biochemical consequences of coadministering CIT and DUL, highlighting the potential risks associated with serotonin syndrome. Our behavioral assessments using FST and SPT revealed paradoxical exacerbation of depressive-like behavior in rats receiving combination therapy, despite individual antidepressant treatments showing efficacy in alleviating such symptoms. This observation suggests a possible synergistic effect leading to serotonin syndrome, characterized by heightened serotonin levels, which may manifest as exacerbated depressive symptoms. Notably, the occurrence of serotonin syndrome underscores the need for cautious monitoring and individualized treatment strategies, particularly in vulnerable patient populations.

Furthermore, our study underscores the significance of molecular modeling simulations in elucidating the underlying mechanisms of drug interactions and predicting potential adverse outcomes. By employing docking analysis targeting key enzymes and transporters involved in drug metabolism and disposition, such as CYP2D6 and P-gp, we gained insights into the binding affinity and interaction patterns of CIT and DUL. The observed competitive binding between CIT and DUL at the same binding sites of CYP2D6 and P-gp corroborates our pharmacokinetic findings and provides a mechanistic understanding of the observed PK-DDIs. However, it is imperative to acknowledge the limitations of docking simulations, which represent a static model and may not fully capture the dynamic nature of drug interactions in vivo.³⁴ Hence, further experimental validation, including in vivo studies, is warranted to corroborate our findings and guide clinical decision-making regarding antidepressant combination therapy.

Conclusion

Our study sheds light on the complex interactions between CIT and DUL, two commonly prescribed antidepressants with distinct pharmacological profiles. Through a multifaceted approach encompassing analytical, computational, behavioral, and biochemical techniques, we have unraveled significant insights into the pharmacokinetic and pharmacodynamic interplay of these medications. Our findings underscore the critical need for further research into antidepressant combinations, particularly the potential for DDIs, to optimize treatment efficacy and minimize adverse effects. The development and validation of a robust assay for simultaneous quantification of CIT and DUL in biological fluids represent a significant contribution to the field, offering a valuable tool for future investigations into antidepressant DDIs. Additionally, our exploration of the behavioral and biochemical consequences of coadministering CIT and DUL highlights the importance of considering both clinical and preclinical evidence when evaluating antidepressant therapies. Moving forward, it is imperative to continue exploring the intricate mechanisms underlying antidepressant interactions to guide clinical decision-making and improve patient outcomes. Future research efforts should focus on elucidating the specific pathways and receptors involved in CIT–DUL interactions, as well as investigating potential strategies to mitigate adverse effects, such as serotonin syndrome. Ultimately, our study underscores the ongoing need for collaborative interdisciplinary research to advance our understanding of antidepressant pharmacology and enhance depression management strategies.

Methods

Pharmacokinetic and DDI Study in Rats

Three sets of three male Wistar Albino rats, each weighing between 200 and 250 g, were randomly assigned at the age of two months. Before the studies, the rats were kept in separate metabolic cages for 3 days under a 12-h light/dark cycle, with free access to food and drink. The studied antidepressants were dissolved in the proper volume of saline. Rats in the first group (Group I) were given 120 mg/kg of DUL orally after an overnight fast; rats in the second group (Group II) were given 120 mg/kg of CIT; and rats in the third group (Group III) were given a combination of 120 mg/kg of DUL and 120 mg/kg of CIT. Under anesthesia produced by ethyl carbamate (1.5 g/kg, i.p.), blood samples of approximately 300 μL each were taken via the femoral artery using indwelling arterial catheters. At 0.5, 1, 2, 3, 4, 6, 8, and 12 h after injection, blood samples were taken. Rats were instantly rewarded with an equivalent dose of saline injection after each sample collection. After quickly centrifuging the blood samples at 10,000g for 10 min, the plasma layer was separated, moved to clean tubes, and kept cold until LC analysis. Rats were given blank plasma samples, which were used for technique validation and optimization. No drugs were administered to the rats. The Suez Canal University Faculty of Pharmacy’s Research Ethics Committee gave its approval to all experimental techniques used in this study (Approval number. 201608RH1).

Pharmacokinetic Analysis

The freeware menu-driven add-in tool PKSolver was used to do noncompartmental pharmacokinetic analysis for Microsoft Excel. The area under the plasma concentration–time curve (AUC_0–t) was calculated using the linear trapezoidal method. Regression analysis was used to estimate the elimination rate constant (K_el) based on the slope of the best-fit line, and the equation 0.693 divided by K_el was used to get the half-life (t_1/2) of the antidepressant of interest. Furthermore, PKSolver made it easier to calculate the volume of distribution following nonintravenous dosing (Vd/F), mean residence time (MRT), and complete clearance of the drug from plasma (CL/F).³⁷

Plasma Sample Preparation

Following the addition of 50 μL of 1 N NaOH solution, 50 μL of IS (500 ng/mL), and 100 μL of either calibration solutions or rat plasma, the mixture was spun at room temperature for 20 s. Rat plasma was used to separate the target analytes and IS using Oasis hydrophilic–lipophilic balance (HLB) 1 cm³/30 mg cartridges. These cartridges were condition-treated with 0.5 mL of methanol and equilibrated with 0.5 mL of clean water. After that, each sample was run through a separate cartridge, which was then cleaned twice with 500 μL of distilled water. Upon eluting the target analytes with 300 μL of ACN, 5 μL of the eluent was subsequently introduced into the UPLC-MS/MS equipment.²³

Molecular Modeling

The molecular modeling approach utilized the atomic-resolution crystallographic structures of Homo sapiens CYP2D6 bound to its inhibitor Prinomastat (PDB ID: 3TDA; resolution: 2.67 Å) and Mus musculus (house mouse) P-gp bound to its inhibitor PBDE-100 (PDB ID: 4XWK; resolution 3.5 Å). The preparation and refinement of the CYP2D6/Prinomastat and P-gp/PBDE-100 binary complexes were conducted using the OPLS-2005 force field, ensuring appropriate ionization states for acidic and basic residues at physiological pH (7.0) during 3D-protonation. Energy minimization was carried out with the aim of resolving steric clashes, employing the OPLS-2005 force field with a preset cutoff at an average root-mean-square deviation (RMSD) of 0.3 Å for heterogeneous atoms. Additionally, the construction and optimization of CIT and DUL in three dimensions were accomplished using Maestro’s build panel, with subsequent refinement of ligand structures involving partial atomic charges via the OPLS-2005 force field. Characterization of the CYP2D6 and P-gp active sites was performed using the “grid generation panel” in Glide, and the molecular modeling protocol employed rigid-docking algorithms with optimized protein–ligand geometries.³⁸

Animals for Pharmacodynamic Studies

Adult male Wistar albino rats, weighing between 200 and 250 g and aged 8 to 10 weeks, were obtained from the animal facility at the Faculty of Pharmacy, Cairo University, Cairo, Egypt. The animals were given 2 weeks to acclimate to laboratory conditions prior to the experiment. The rats were housed in metabolic cages with controlled temperature (23 ± 1 °C) and humidity (40–60%) settings and were subjected to a 12-h light-dark cycle. They had unrestricted access to water and were fed a standard rat pellet diet.

Experimental/Study Design

The study timeline is illustrated in Figure S4. Rats were randomly assigned to five groups (n = 8) as follows:

• Group I (Control): rats received intraperitoneal (i.p.) administration of normal saline for 21 consecutive days, followed by oral saline administration for an additional 21 days.

• Group II (DEX): rats were treated with dexamethasone (DEX) at a dose of 1.5 mg/kg (i.p.) for 21 days, followed by oral saline administration for 21 days.³⁹

• Group III (DEX + CIT): rats received DEX (1.5 mg/kg, i.p.) for 21 days, followed by oral administration of citalopram (CIT) at 20 mg/kg for 21 days.⁴⁰

• Group IV (DEX + DUL): rats were treated with DEX (1.5 mg/kg, i.p.) for 21 days, followed by duloxetine (DUL) at 20 mg/kg (p.o.) for 21 days.⁴¹

• Group V (DEX + CIT + DUL): rats received DEX (1.5 mg/kg, i.p.) for 21 days, followed by combined oral administration of CIT (20 mg/kg) and DUL (20 mg/kg) for 21 days.

Following the final injection, the animals underwent behavioral assessments, including the forced swimming test (FST) and the sucrose preference test (SPT). Afterward, rats from each group were divided into two sets and humanely euthanized by cervical dislocation under light anesthesia using thiopental (50 mg/kg).⁴² The brains were immediately dissected (sagittal sectioning) and rinsed with ice-cold saline. In one set (n = 2), the brains were fixed in 10% formalin for 24 h and prepared for histopathological examination using hematoxylin and eosin (H&E) staining (data not shown). In the second set, the hippocampi were promptly isolated and stored at −80 °C for subsequent serotonin level measurement via enzyme-linked immunosorbent assay (ELISA). All experimental procedures complied with the Guide for the Care and Use of Laboratory Animals (NIH publication No. 85-23, revised 2011), and were approved by the Research Ethics Committee of the Faculty of Pharmacy, Suez Canal University (Approval number: 201608RH1)

Behavioral Tests in Rats

Sucrose Preference Test (SPT)

The sucrose preference test (SPT) was conducted following the induction of the depressive-like state. Initially, the rats were acclimatized to a 1% sucrose solution for 48 h. After this period, they were deprived of water for 4 h. Subsequently, the rats were presented with two identical bottles—one containing the sucrose solution and the other containing water—for a duration of 1 h. Sucrose preference was determined by calculating the ratio of sucrose consumed to the total volume of liquid (sucrose plus water) consumed during the test period.^43,44 The test was conducted between 7:00 AM and 7:00 PM.

Forced Swimming Test (FST)

During the forced swim test (FST), rats were placed in a plastic cylinder with a diameter of 20 cm and a depth of 40 cm. The water temperature was maintained between 23 and 25 °C, with a water level of 25 cm from the base. In the first training session, the rats were forced to swim for 15 min. After 24 h, the rats were placed back in the same cylinder for a 5 min test session, during which the duration of their immobility was recorded. Following each session, the rats were removed from the cylinder, gently dried with paper towels, and returned to their home cages. The water was replaced for each new rat. Immobility time, defined as the period during which the rats remained floating without struggling, was used as an index of depressive-like behavior.⁴² The test was conducted between 7:00 AM and 7:00 PM.

Serotonin Level Measurement

Enzyme-Linked Immunosorbent Assay (ELISA)

Hippocampal serotonin levels were measured using rat-specific sandwich ELISA kits (Cat#: MBS166089, My BioSource, San Diego, CA, USA) according to the manufacturer’s protocol. Absorbance values were recorded using a microplate reader (Biotech, USA) and quantified by comparison to a standard curve generated during the same experiment.

Statistical Analysis

Data are expressed as mean ± standard deviation (S.D.). Normality was assessed using the Shapiro–Wilks test. Subsequently, a one-way analysis of variance (ANOVA) was performed, followed by Tukey’s multiple comparison test for posthoc analysis. Statistical analyses were conducted using GraphPad Prism software (version 8, San Diego, CA, USA), with a significance threshold set at p < 0.05 for all tests.

Data Availability Statement

Data will be available on request.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsptsci.4c00506.

• Containing chemicals, stocks preparation solution, instrumentation, method validation (PDF)

Author Contributions

M.S.E. contributed to methodology, conceptualization, resources, and writing-original draft and editing. A.M.A. contributed to methodology, conceptualization, and writing-original draft and editing. S.M.H. contributed to resources, supervision, and writing-review and editing. S.S.A. contributed to methodology and writing–original draft. A.M.A.M. contributed to methodology and writing-original draft. M.A.M. contributed to resources, supervision, and writing-review and editing. M.A.A. contributed to resources, supervision, and writing-review and editing.

The authors declare no competing financial interest.