Over time, the cholinergic deficiency hypothesis culminating in cognitive and memory disruptions has proven realistic when memory-enhancing medications relied on increasing acetylcholine (ACh) (Baxter and Crimins, 2018). Unlike the early reduction in ACh levels, the involvement of acetylcholinesterase (AChE), the ACh metabolizing enzyme, seemed to occur relatively late (Perini et al., 2002). Among neurotransmitters implicated as modifiers of cholinergic signaling is melatonin, which was able to prevent the inhibition of ACh synthesizing enzyme in several neuronal proteins (Guermonprez et al., 2001). Durand-de Cuttoli et al. (2018) discussed dopamine interactions with ACh in modifying multiple brain functions, including cognitive performance. Furthermore, dopamine type 5 receptors (D5) were claimed to assist the role of ACh in maintaining healthy learning and memory processes (Rizzi and Tan, 2017).

The link between some antipsychotic medications, sleep issues, and dementia, with subsequent progressive neurodegeneration and cognitive impairment, were previously explored (Pillai and Leverenz, 2017; Xu et al., 2020). Regular monthly injections of the depot form of haloperidol, haloperidol decanoate (HD), results in the slow release of free haloperidol into the systemic circulation. Warnings have been issued regarding the use of HD in cases of neurologic and cardiac disorders, including some types of dementia, cerebrovascular stroke, arrhythmia, and heart failure. In rodent studies, HD precipitated QTc prolongation, and increasing the dose of HD to 1 mg/kg or above precipitated ventricular arrhythmia. The induced cardiac adverse effects seemed of interest based on the reported link between cardiovascular issues and cognitive defects (Liang et al., 2021). However, off-label use of haloperidol in patients with dementia-associated agitation and psychotic features is still adopted (Mühlbauer et al., 2021).

In clinical practice, HD can be coadministered with hypnotics and/or antidepressants to alleviate some of the antipsychotics-associated adverse effects such as agitation, depression, and insomnia (Haldol decanoate Package Leaflet 2020). Agomelatine (AGO), a hypnotic and antidepressant drug, acts as a nonselective melatonin [melatonin type 1 receptor and melatonin type 2 receptor (MT2)] agonist and a serotonin 5-HT2C receptor antagonist (Gobbi and Comai, 2019). Unlike other antidepressants, AGO exerts an additional anxiolytic activity (Yohn et al., 2017) and was considered a favorable hypnotic in terms of its claimed enhancing effect over learning and memory, in addition to its neuroprotective properties (Gupta et al., 2015). Hence, AGO has been suggested as a potential treatment for the cognitive defects and psychotic features associated with low muscarinic signaling in human psychoses and animal models of schizophrenia (Barak and Weiner, 2009).

As cerebrovascular disorders were implicated in cognitive deficits (Hedges et al., 2019), recently, angiopoietin-like 4 (ANGPTL4), a member of the angiogenic-regulating secreted protein superfamily (Zhu et al., 2012), which plays a key role in angiogenesis, was found to serve as a diagnostic biomarker in patients with clinically assessed vascular-related cognitive defects (Chakraborty et al., 2018). Currently, research on the implication of ANGPTL4 in neurologic disorders is ongoing.

Identified as a required regulator of central nervous system and neuronal maturation (Pollak et al., 2018), Krüppel-like factor 9 (KLF9) is a member of the novel Krüppel-like factor family of evolutionarily conserved zinc finger transcription factors (Wang et al., 2008). Targeting KLF9 was able to improve stress-provoked depression, as well as other related behavioral issues (Besnard et al., 2018). KLF9 was also linked to cardiac ischemia (Yan et al., 2019). Up until now, the involvement of KLF9 in neuronal dysfunctions remains an obscure area of research interest.

Therefore, the diversity of factors affecting cognition drew our attention to conduct our experiment, where we aimed at comparing the effects of long-term combined HD + AGO on cognition. Conducting behavioral tests, while analyzing the brain’s electrical signals, with special emphasis on beta and delta wave activities, being closely linked to cognition as well as to each other (Harmony, 2013; Jang et al., 2019), together with histopathologic examination were integral parts of our assessments. The electrical activity of the heart was recorded as well to identify any relevant cardiovascular issues. The cholinergic deficiency hypothesis of cognition was verified against the melatonergic/dopaminergic hypothesis. ANGPTL4 and KLF9 were investigated as respective novel mediators of brain vascularization and neuroregeneration. Secondarily, we aimed at discriminating between two commonly used AGO doses in research. The presumptive cognitive, neuropathological, biochemical, and cardiac outcomes of our experiment would have important clinical implications by offering novel cognitive enhancers and/or raising awareness about combined HD + AGO use, taking into consideration the balance between benefit and risk potentialities.

To our knowledge, this is the first study to address the cognitive impact of chronic HD combined with AGO, compared with either drug. Our work discriminated between two AGO doses (40 mg/kg and 80 mg/kg) commonly employed in previous research work. As HD releases active haloperidol, our discussion included the studies on both HD and haloperidol, the latter being the more extensively studied. The selection of male rats matched the reports highlighting a better performance of males when assessing visual-spatial tasks (McCarrey et al., 2016) employed herein. In clinical settings, one of the major indications for the use of HD is schizophrenia, which is more prevalent in males (X. Li et al., 2022). Avoiding the effects of fluctuating hormones during estrus cycles and potential effects on some of the assessed signaling mechanisms (Gunn et al., 2016; Spencer et al., 2008; Zachry et al., 2021), the male gender was chosen.

In agreement with the lack of cognitive alteration with prolonged corn oil intake in control rats, a randomized controlled trial reported a similar lack of additional benefit after 6-month corn oil intake in subjects with normal cognition (Maltais et al., 2022). However, other studies demonstrated improved cognition following alternate-day intraperitoneal injection of corn oil to a mouse model of Down syndrome after 30 days (Giacomini et al., 2018) or when corn oil, supplemented by other vitamins, was consumed by humans (Gutierrez et al., 2021).

In this study, the cognitive improvement observed with HD + high-dose AGO could be attributed to enhanced dopaminergic signaling (Speranza et al., 2021). The involvement of the MT2/D5 interplay is plausible, given the concomitant reduction in brain MT2 against D5 upregulation, further supported by the negative correlation between MT2 and D5, corroborating the inhibitory effect of melatonin over dopamine signaling through D1 and D2 receptors (Sweis, 2005). Furthermore, the positive correlation between brain ACh and MT2 versus the negative correlation linking ACh to D5 was in partial agreement with a reduced brain ACh, relieving the disinhibition of dopaminergic neurons and subsequently enhancing dopamine release (Exley & Cragg, 2008). Based on the assumption that a “cholinergic-melatonergic-dopaminergic trialogue” exists in the context of cognition, we hypothesized that the concomitantly reduced brain ACh and downregulated MT2, subsequently upregulating brain D5, might enhance some aspects of cognition, complementary to the cholinergic dysfunction hypothesis, which could involve other aspects of cognition, depending on the disordered function in context. The involvement of noncholinergic pathways was supported herein when opposite neurochemical changes observed with both AGO doses precipitated some cognitive impairment. In the presence of similar ACh, MT2, and D5 changes as HD + high-dose AGO, such cognitive impairment was not evident with HD, despite showing the most extensive neuropathologic affection. Thus, the cholinergic-melatonergic-dopaminergic trialogue we hypothesized could be the basis to explore novel pharmacologic targets, offering diverse therapeutic alternatives that might not only fill major cognitive gaps between normal individuals but also ameliorate resistant cases to the currently available cognitive enhancers.

We extended our cholinergic-melatonergic-dopaminergic trialogue to ANGPTL4 and KLF9, being essential targets in the issue of distorted brain microstructure (Zheng et al., 2021). To our knowledge, no previous studies have assessed the effects of prolonged treatment with HD, AGO, and their combination on brain ANGPTL4 and KLF9 in the context of cognition. We found that the HD + high-dose AGO was the least to induce ANGPTL4 expression but, more importantly, without altering KLF9 expression, unlike the other treated groups, in which KLF9 was downregulated. Thus, our study proposed the implication of KLF9 deficiency in cognitive disorders, as was with KLF2 in the mouse model of AD and in humans (Fang et al., 2017; Wu et al., 2013). The negative correlation between ANGPTL4 and KLF9 highlighted their inverse relationship in the context of cognition (Shuff et al., 2021), after previous studies detected elevated ANGPTL4 in postmortem specimens of patients with AD (Yan et al., 2019) and its association with poor prognosis of ischemic stroke (Zheng et al., 2021).

The relatively less cognitive improvement observed with HD + low-dose AGO could be attributed to worse neuropathologic features and different neurochemical alterations, lacking similar MT2 and D5 changes as HD + high-dose AGO, added to a more extensive ANGPTL4 upregulation against a downregulated KLF9, reinforcing the cholinergic-melatonergic-dopaminergic trialogue and the feasibility to explore noncholinergic targets for cognitive dysfunction (Doraiswamy, 2002; Petrikis et al., 2004). The exclusive increase in body weight with HD + low-dose AGO might have also contributed, based on previous studies linking obesity to reduced cognition (Buie et al., 2019; T. Li et al., 2022). Though not the focus of this work, it is interesting to mention that HD + low-dose AGO was the only treatment regimen associated with a downregulated brain AChE expression and reduced catalytic activity, emphasizing the negative correlation between AChE and body mass index reported in a recent study involving female university students (Hamouda et al., 2019). It is noteworthy to highlight that reduced ACh occurred, despite a downregulated AChE. So whether this is some sort of protective compensatory mechanism to arrest the progression of cholinergic deficiency (Brown, 2019) requires further elaborative research work and, if proven, could rationalize the lack of efficacy of anticholinesterases in some cases of cognitive impairment. In our study, the positive correlation linking AChE activity and the brain ACh level could validate such a compensatory relationship.

In our model, the slight cognitive amelioration following prolonged HD administration was associated with increased mean power of source wave frequency, so that the power spectral density of the source wave was maximum at a frequency corresponding to the fast gamma range (60–100 Hz) (Colgin, 2016), linked to enhanced memory and learning in previous mice models (Sharpe et al., 2020; Zheng et al., 2016). Although HD induced almost similar neurochemical alteration as HD + high-dose AGO, one of the major differences was KLF9 downregulation; the other was the extensive AD-like features. In a previous rodent model, daily intramuscular injection of haloperidol (1.5 mg/kg) for 28 days induced neurodegenerative changes in the striatum and hippocampus (Polydoro et al., 2004), justifying the less cognitive impairment seen with HD, relative to acute intraperitoneal haloperidol injection (0.05 mg/kg) in rotenone-treated neonatal male Wistar rats (Varga et al., 2021). Nonetheless, the effect of HD on cognition seems controversial (Lalonde & Strazielle, 2017).

Both AGO doses improved some cognitive features and worsened others. The improvement of some cognitive indices using AGO agreed with previous studies Navarro-Francés and Arenas, 2014; Rao Barkur and Bairy, 2015; Meseguer Henarejos et al., 2020). The associated increased brain ACh, if supporting the cognitive improvement, could not justify the featured cognitive dysfunction, except after applying the hypothetical cholinergic-noncholinergic crosstalk, in which the concomitantly increased ACh and upregulated MT2, subsequently downregulating D5, added to the inclusion of ANGPTL4 upregulation, against KLF9 downregulation are introduced as novel co-players in the cognitive issue paradigm. Considering the inverse relationship between AChE activity on one side, enhanced herein, and brain D5 and KLF9 on the other side, as well as its positive link to ANGPTL4, AChE activity should be considered an independent factor in the cognitive dilemma. Elevated ACh activates presynaptic muscarinic receptors, thus suppressing the excitatory inputs needed for memory retrieval, leading to cognitive dysfunction (Hasselmo, 2006). Nonetheless, the concomitantly upregulated AChE expression along with enhanced catalytic activity highlighted its plausible protective role to maintain an optimum ACh level (Brown, 2019), thus justifying some concurrent cognitive improvement. The excess melatonergic precipitating cognitive dysfunction was previously supported by a phase IV clinical trial involving women above 60 years old who experienced memory impairment following 1 to 6 months of melatonin intake (Melatonin and Memory loss, a phase IV clinical study of FDA data - eHealthMe). In our experiment, such worsening did not occur with either HD, when MT2 was downregulated, or HD + low-dose AGO, when MT2 was redeemed to control levels. The associated reduced weight of the right and left brain hemispheres harmonized with the cognitive dysfunction and was in partial agreement to a previous correlation between brain size and both cognitive decline and dementia risk (Tan et al., 2017), possibly supporting the use of brain size as a surrogate marker to predict behavioral aspects of AD (Boublay et al., 2020). The inverse relationship linking the brain-hemispheric weights and AChE catalytic activity supports, in part, the therapeutic merit of anticholinesterases in neurodegenerative disorders, regardless of ACh level. It is worth mentioning that the brain weight might not be an accurate indicator though, based on the discrepancies between the changes of the weights of the brain hemispheres and the extent of neuropathology, as observed with HD. In contrast to our work, neither brain weights nor brain microstructure was changed when Yang et al. (2019) employed toxic AGO doses, but for a shorter duration of >28 days. The duration might be a crucial determinator, as in various disease models, short-term AGO did not provoke AD-like features (Yucel et al., 2016; Chanmanee et al., 2022).

The discrepancies between the treated groups could be also attributed to differences recovered on EEG recordings, impacting sleep (Zavecz et al., 2023), albeit this was not our focus. With low-dose AGO, a reduced mean power frequency delta wave was noticeable, seemingly reflecting an enhanced non-rapid eye movement previously detected in mice receiving the same AGO dose intraperitoneally for 21 days (Tchekalarova et al., 2020). The sleep-promoting ability of high-dose AGO was evident when total beta power was increased, being linked to a tranquilizing activity (Walker et al., 2001) and a sleep-like state (Fernandez‐Mendoza et al., 2019), substantiated by concurrently increased delta amplitude (Kawai et al., 2016). With HD + low-dose AGO, the increased delta wave amplitude exceeded that observed with HD, indicating different sleep-promoting potentialities (Jones et al., 2014) and subsequent differential effects on cognition.

It is noteworthy to add some findings involving different epileptogenic activities that could not be overlooked, not only as part of the pharmacovigilance evaluation of HD-AGO combination but also due to their relevance, based on the link between epileptogenesis and both working memory impairment (Arski et al., 2022) and the occurrence of dementia-like cognitive decline (Chin and Scharfman, 2013; Imfeld et al., 2013). In this context, the reduced mean power frequency of the source wave added to the undisturbed polarity of EEG waves detected with HD + high-dose AGO, exhibiting cognitive enhancement, highlighted a lower risk for epileptogenic activity (Kara et al., 2020), against the increased mean power frequency of source wave seen with HD, showing minimal cognitive improvement, reflected discrepant effects on cortical excitability (Barbiero et al., 2007; Ly et al., 2016) and varied potentialities for proconvulsant activity (Zhou and Li, 2020). Knowing that the EEG montage was uniform among all studied groups and that the source wave is likely to be negative as corresponding to superficial depolarizations using scalp electrodes, the inversion of negative to positive polarity observed in source and delta waves seen with HD, high-dose AGO and HD + low-dose AGO, might increase the likelihood of proconvulsive activity, with a possible link to reduced cognitive performance (Janati et al., 2013) or improvement, especially in the presence of AD-like neuropathologic features. Therefore, such epileptogenic activity contributed, in part, to the cognitive deterioration observed with high-dose AGO and the lack of substantial cognitive improvement with HD and HD + low-dose AGO. The implication of melatonin in provoking epileptogenesis was encountered previously in human cases (Sandyk et al., 1992; Sheldon, 1998) and was corroborated herein by detecting a positive correlation between brain MT2 and total EEG power, added to high-dose AGO displaying the highest source EEG wave amplitude in the presence of MT2 upregulation.

The brain-cardiac crosstalk was recently reported (Yamaguchi et al., 2020; Lin et al., 2022) and was demonstrated in patients who experienced myocardial injury following cerebral stroke (Krause et al., 2017). Similarly, manipulation of central neurotransmission can impact cardiorespiratory function (Díaz et al., 2021). In turn, cardiac ischemia might alter brain neurotransmission and trigger neuroinflammatory changes (Díaz et al., 2020), also impacting cognitive performance (Amorim et al., 2022; van Nieuwkerk et al., 2023). Therefore, ECG was recorded.

Our work is one of few studies investigating the impact of prolonged AGO administration on cardiac electrophysiology in a normal setting, while other studies and reviews evaluated the use of melatonin in managing cardiovascular diseases, employing animals and patients with disordered cardiac functions (Dominguez-Rodriguez, 2012; Sun et al., 2016; Nduhirabandi and Maarman, 2018).

In this study, while all treatments triggered changes indicative of cardiac ischemia and arrhythmia, it is worth mentioning that the most extensive ECG alterations were observed with both AGO doses, possibly rationalizing the concurrent cognitive dysfunction. Exclusively, reduced TpTe interval, indicating reduced time for transmural dispersion of ventricular repolarization (Opthof et al., 2009), was evident with both AGO doses. Interestingly, the catalytic activity of brain AChE, enhanced with both AGO doses, was negatively correlated to TpTe. When such AChE activity was reduced or unaltered, TpTe was unaffected. A thorough literature search did not reveal studies outlining the significance of short TpTe nor its correlation to central AChE activity. Most studies focused on TpTe prolongation as a risk of ventricular tachycardia and fibrillation (Zumhagen et al., 2016; Tse et al., 2017). As late sodium channels or sodium-calcium exchange are important determinants of TpTe interval (Antzelevitch, 2007), the occurrence of channelopathies or electrolyte imbalances could have provoked such an acceleration. Recently, a case report published in 2022 (Siao et al., 2022) suggested a causal relationship between AGO and cardiac arrhythmias, supporting the AGO-triggered cardiac electrical events found in our work. In male and female rats administered 28 days of oral daily AGO at doses of 200, 400, and 800 mg/kg, exceeding the doses used in clinical settings, there was no evidence for microstructural cardiac abnormalities (Yang et al., 2019). The ECG disturbances detected herein should be further weighed against some of the recently claimed vascular benefits of AGO when administered to female Wistar rat models of lipopolysaccharide-induced endothelial injury (Asci et al., 2022).

Contrasting low- to high-dose AGO, the former reduced QRS duration and was associated with elevated ST; the latter reduced P wave amplitude, prolonged RR interval, and slowed HR. The slowed HR associated with prolonged RR interval suggests a sinus node or autonomic dysfunction, triggering bradycardia and recurrence of atrial fibrillation (AF) (Amasyali et al., 2014). RR was adopted also as a prognostic marker in patients with AF (Zyśko et al., 2021); in turn, AF can provoke bradycardia (Barrett et al., 2012). The reduced P wave amplitude, denoting reduced atrial contractility, can also be associated with a higher risk for AF, added to heart failure (Park et al., 2016; Ostrowska et al., 2022; Zhang et al., 2023). ST elevation reflected the possibility of cardiac ischemia (Coppola et al., 2013). Reduced QRS interval, corresponding to accelerated inter- or intraventricular conduction, together with ST elevation, albeit denoting apparently favorable enhanced ventricular conduction and early repolarization, as previously published in case reports by (Wolpert et al., 2008), the clinicians were advised to be cautious as these EEG features might denote abnormal sodium channels disorders or increased Purkinjie fibers penetrating deep into the myocardium.

The ECG changes observed with HD were, seemingly, akin to those seen with high-dose AGO; however, inverted P wave, elevated ST, and prolonged QRS duration relative to high-dose AGO were superimposed. The inverted P wave could indicate a sinus node dysfunction or myocardial infarction (Khanna et al., 2022). The extensive ECG changes could have some clue to the worst neuropathology featured with HD (Shuff et al., 2021).

HD + AGO, regardless of AGO dose, triggered de novo ECG features, distinct from either HD or AGO monotherapy. HD + low-dose AGO caused a positive upstroke of S wave, while HD + high-dose AGO triggered a tall, inverted T wave, indicative of cardiac ischemia (Kashou et al., 2019).

Despite that combined HD + AGO redeemed many of the HD- and AGO-induced ECG changes, inverted P wave, observed with HD, and reduced P wave duration as well as tall R wave, seen with both HD and AGO, persisted. HD + low-but not high-dose AGO redeemed the reduced PR interval, as seen in the other treated groups. In turn, HD + high-but not low-dose AGO redeemed the reduced QTc seen in the other treated groups. The reduced P wave duration, denoting reduced atrial conductivity, can be also associated with a higher risk for AF and heart failure. The reduced PR interval, signifying an accelerated atrio-ventricular conduction, heightened AF risk, predisposing also to the potentially lethal ventricular fibrillation (Neshiewat et al., 2023). The increased R wave amplitude could highlight an increased blood mass inside the cardiac ventricles during early ventricular depolarization, predisposing to ventricular hypertrophy (Bonoris et al., 1978). A persistent tall R wave could denote myocardial ischemia and/or bundle branch block (Sinno et al., 2008), both being previously documented as adverse events of haloperidol (Ter Bekke and Volders, 2020). Reduced QTc, reflecting accelerated cardiac repolarization, could occur secondary to channelopathies or electrolyte imbalances (https://www.acc.org/latest-in-cardiology/articles/2016/10/05/08/06/short-qt-syndrome), predisposing to arrhythmia (Anttonen et al., 2009).

Although no mortalities occurred, such adverse ECG changes should be weighed against the cognitive benefit when transitioning to clinical settings, given the higher risk of AF-related hospitalization in patients with heart failure and the substantial risk for morbidity and sudden cardiac death reported in patients with heart failure in case of arrhythmia (Masarone et al., 2017; Yuyun et al., 2023).

Biochemical analysis did not include cardiac specimens. Liver function tests and electrolytes were not assessed.