Multi-Target Directed Ligands (MTDLs) Binding the σ₁ Receptor as Promising Therapeutics: State of the Art and Perspectives

Abstract

The sigma-1 (σ₁) receptor is a ‘pluripotent chaperone’ protein mainly expressed at the mitochondria–endoplasmic reticulum membrane interfaces where it interacts with several client proteins. This feature renders the σ₁ receptor an ideal target for the development of multifunctional ligands, whose benefits are now recognized because several pathologies are multifactorial. Indeed, the current therapeutic regimens are based on the administration of different classes of drugs in order to counteract the diverse unbalanced physiological pathways associated with the pathology. Thus, the multi-targeted directed ligand (MTDL) approach, with one molecule that exerts poly-pharmacological actions, may be a winning strategy that overcomes the pharmacokinetic issues linked to the administration of diverse drugs. This review aims to point out the progress in the development of MTDLs directed toward σ₁ receptors for the treatment of central nervous system (CNS) and cancer diseases, with a focus on the perspectives that are proper for this strategy. The evidence that some drugs in clinical use unintentionally bind the σ₁ protein (as off-target) provides a proof of concept of the potential of this strategy, and it strongly supports the promise that the σ₁ receptor holds as a target to be hit in the context of MTDLs for the therapy of multifactorial pathologies.

1. Introduction

The most exploited paradigm in medicinal chemistry is the “one-target, one-disease”, according to which ligands are developed to act toward a single target in order to exert a beneficial effect. This kind of approach, even if largely applied, may lead to unsuccessful results. Indeed, pathologies are mostly based on tangled mechanisms that involve different kinds of targets. Thus, a single target approach could be inappropriate because of its impossibility to produce an effect on the different pathways associated with the pathology.

On the basis of this idea, a multi-target strategy has been proposed in order to hit at the same time the diverse targets involved and modulate the diverse mechanisms that underlie the pathology.

The first step of this multi-target strategy has been the co-administration of more drugs all together, with the aim of obtaining a synergic effect. Some examples of this approach are represented by current therapeutic programs. Pathologies such as hypertension, cancer, neurodegeneration, AIDS, and depression can require the administration of up to three different classes of drugs.

Although this approach has therapeutic beneficial effects, the higher the number of drugs administered, the higher the number of metabolites that could produce side effects. Moreover, the co-administration of different drugs can produce cross-resistance to the therapeutics and inconvenient pharmacokinetic profiles. With the aim to overcome such problems, the multi-target directed ligand (MTDL) approach has been introduced. This strategy is based on the administration of a single drug with multiple effects because such a drug incorporates two or more pharmacophores in a single molecule. This incorporation can produce hybrid drugs or chimeric drugs. Hybrid drugs are obtained when two or more pharmacophores are connected using a stable or metabolizable linker. These structures not only have a high molecular weight, which is detrimental for oral bioavailability, but they are also very flexible, so that binding to the targets can be reduced. On the other hand, chimeric drugs are obtained by fusing or merging two or more different pharmacophores into one. In contrast to the former type, chimeric drugs have better pharmacokinetic properties and better interaction with targets because of their rigidity [1,2].

The synthesis of MTDLs can follow a knowledge-based approach or a screening approach. The former approach is based on the use of diverse and focused libraries of molecules and relies on well-known drugs and compounds. In the screening approach instead, ligands that demonstrate a solid affinity for a certain target are explored in their activities/affinities for other targets. However, in both cases, a fine work of optimization is needed to obtain the best compromise that guarantees the interaction with all targets [3].

In this review, the authors want to discuss the use of the MTDL approach that exploits the interaction with sigma (σ) receptors. In fact, even if these receptors are still not completely understood, their involvement in several pathologies has been widely demonstrated, so that the synergy obtained by a simultaneous interaction with σ receptors and other biological targets appears as a beneficial strategy to face multifactorial pathologies.

σ Receptors were discovered in 1976, when studies performed on opioid receptors revealed the presence of proteins that do not bind naloxone and etorphine but display high affinity for (+)-pentazocine, which is still used in binding studies [4]. A milestone in the history of σ receptors has been attained when two different protein subtypes were discovered, namely σ₁ and σ₂ receptors [5].

2. σ1 Receptors

σ₁ Receptors have been found in the central nervous system (CNS), particularly in the granular layer of the olfactory bulb, in many cortical layers, and in the dentate gyrus [6], where they exert their most important activities. A slightly lower expression has been also found in some pyramidal layers of the hippocampus, various hypothalamic nuclei, the septum, the central gray, the motor nuclei of the hindbrain, and the dorsal horn of the spinal cord [6]. At the cellular level, σ₁ receptors have been found in ependymocytes, which border the ventricular compartments, and in neurons located within the CNS parenchyma [6].

In peripheral organs, σ₁ receptors have been found in the gastrointestinal tract [7], vas deferens [8], in liver and kidney [9], heart [10], adrenal medulla, pituitary, testis, and ovaries [11].

The σ₁ receptor is localized at the mitochondrial-associated endoplasmic reticulum (ER) membranes (MAM) in association with the binding immunoglobulin protein (BiP) in a resting state. Upon agonist stimulation, or Ca²⁺ depletion, the σ₁ receptor dissociates from BiP and stabilizes inositol 1,4,5-trisphosphate receptor type 3 (IP3R3) at MAM, increasing Ca²⁺ transfer from ER into mitochondria where adenosine triphosphate (ATP) production is facilitated [12]. Recently, it has been proposed that σ₁ receptors affect store-operated Ca²⁺ Entry (SOCE), with agonists, such as (+)-pentazocine, increasing the SOCE activity [13].

The σ₁ subtype is also able to protect against ER stress associated with reactive oxygen species (ROS), which can easily activate the inositol-requiring enzyme 1 (IRE1) at MAM. In this context, the σ₁ receptor stabilizes IRE1 and enhances cellular survival, prolonging the activation of the IRE1–XBP1 signaling pathway [14]. Moreover, the interaction of σ₁ with B-cell lymphoma 2 (Bcl-2) or nuclear factor erythroid 2-related factor 2 (Nrf2) pathways is able to protect against apoptosis [15]. It is now clear that the σ₁ receptor, which is defined as a ‘pluripotent chaperone’ [16], interacts with a number of proteins, as recently reviewed by Schmidt and colleagues [17], although more evidence is required to ascertain a functional interaction.

This pluripotent nature of the σ₁ receptor is responsible for a series of complex effects with the cell state, such as stress conditions, which can produce modifications in the σ₁-mediated response to stimuli. As recently reviewed, it is likely through these protein–protein interactions that the σ₁ receptor exerts its action in viral infections and also against SARS-CoV-2 [18]. It has been demonstrated that the σ₁ receptor interacts with the NSP6 viral protein, and its knock-out (KO) and knock-down (KD) abate the cells’ infection by both SARS-CoV-1 and 2, prompting focused research in this direction. By contrast, the KO and KD of the σ₂ subtype, which interact with the viral protein ORf9c, do not impact on cells’ infection [19].

The neuroprotective action of σ₁ receptor agonists is well established, and it has been proved that this activity is exerted through different mechanisms such as intracellular Ca²⁺ regulation, the prevention of oxidative stress, and anti-apoptotic effects. The σ₁ subtype contributes to protein homeostasis: it can stimulate neurotrophin receptor signaling and reduce protein aggregation responsible for neurodegenerative disease, and it can also activate autophagy as a protective mechanism against damage arisen by misfolded proteins. Recently, many studies proved the ability of σ₁ receptors to directly or indirectly interact with receptors or enzymes with key roles in neurodegeneration, particularly in Alzheimer’s disease (AD). Importantly, in patients with AD, a reduction of σ₁ receptor density [20] has been demonstrated. Some studies attributed such reduction to the E4 variant of the apolipoprotein E gene (APOE 4) [21], although some controversies still exist [22]. All these mechanisms can promote cell survival and consolidate the role of the σ₁ receptor as a target for therapies against neurodegeneration [23].

From a structural point of view, the σ₁ protein has a 223 amino acid sequence, conserved among vertebrates, but devoid of similarity with any other protein. The closest protein seems to be ERG2p, which is a C8-C7 sterol isomerase expressed in the yeast. The two proteins share the same steroid binding domain-like (SBDL) regions, although no isomerase activity has been shown for the σ₁ receptor [24]. A breakthrough milestone in the σ₁ receptor-related research is the crystal structure of the protein, which was published in 2016 [25]. Later, also the crystal structures of the σ₁ protein with agonists and antagonists bound to it were disclosed [26], and important insights into the binding modes of the σ₁ ligands were provided with evidence for the so far ligands-based generated models. The most important interaction between ligands and the receptor is represented by an electrostatic bond between a basic amine (present in all the high-affinity σ₁ ligands) and Glu172, while the rest of the ligand is accommodated in two different hydrophobic areas, which is in line with Glennon’s pharmacophore model [27].

According to this model, the σ₁ receptor pharmacophore consists of a basic amine site tolerating a little substituent (which could be H or small alkyl groups) and two hydrophobic regions: the Primary Hydrophobic Region is 6–10 Å far from the amine site, while the Secondary Hydrophobic one is 2.5–3.9 Å far from the amine. Such a model very well fits into the crystal structure, whose availability has rationalized the binding of several molecules to σ₁ receptors [28].

Importantly, crystallization revealed a trimeric complex, suggesting that the protein acts through polymerization states. Each subunit is formed by a single transmembrane domain (TMD) and a large cytosolic region that contains the binding pocket for each monomer. The 223 amino acid sequence is organized into five α-helices and ten β-strands: the N-terminus passes through the membrane, forms the only TMD, and protrudes in the ER lumen. The C-terminus is a flat hydrophobic sequence associated with the cytosolic surface of the ER. Within the binding pocket, besides the ionic interaction with Glu172, σ₁ receptor ligands interact with Tyr103 (through hydrogen bonds with hydrophilic residues), which provides a favorable environment for the hydrophobic bonds within the primary hydrophobic region (formed by Val84, Ala86, Trp89, Met93, Leu95, Ala98, Tyr103, Leu105, Phe107, Tyr120, Ile124, Trp164, Met170, Ile178, Thr181, Leu182, Phe184, Ala185, Thr202, and Tyr206). The secondary hydrophobic region (made of Trp89, Ser117, Tyr120, Ile124, Asp126, Phe133, Val152, His154, Thr160, Val162, Trp164, and Met170, with Ser117 and His154 exhibiting their hydrophobic side chain to the cavity) tolerates smaller groups compared to the primary one [27,28]. Interestingly, the crystal structure of the receptor bound to its ligands revealed how the agonists, in contrast to the antagonists, induce a structural shift of the α 4 helix, likely preventing the oligomerization state, supporting that the action of the σ₁ receptor occurs through protein–protein interaction (PPI). Therefore, while σ₁ ligands may function as allosteric modulators of PPIs, the different oligomeric structures could be responsible for the many activities performed by these versatile receptors, with agonists and antagonists differently influencing the association among protomers: agonists produce lower oligomeric structures such as monomers and dimers, while antagonists produce higher ones [29].

Overview on σ₁ Receptors Ligands: Functional Activity, Proposed Endogenous Ligands, and Reference Compounds

Functional activity. Although the terms agonist and antagonist have been used to classify σ₁ receptor ligands, an unambiguous definition of the functional activity (agonist versus antagonist) is often hard. As a matter of fact, both agonists (i.e., fluvoxamine and other selective serotonin reuptake inhibitors (SSRIs) [30]) and antagonists (i.e., rimcazole [31,32]) are effective in treating some pathologies such as psychosis, and there are some ligands (i.e., BMY 14802 [33,34]) that show both agonist and antagonist activities in different conditions [35].

With the aim to unequivocally ascertain agonist vs. antagonist activity, many studies have been carried out. Among them, an interesting approach was developed by Gomez-Soler and colleagues, who developed a biosensor exploiting the Fluorescence Resonance Energy Transfer (FRET) technique in living cells. The model was based on the genetic decoration of the σ₁ receptor with cyan and yellow fluorescent proteins. The distance between these two portions depends on the receptor conformation, which is influenced by the nature of the ligand. In fact, while agonists make the two fluorescent portions more distant, antagonists produce a conformational change that makes the portions closer, resulting in a fluorescence emission due to the energy transfer between them [36]. Therefore, the claimed σ₁ receptor agonists and antagonists studied produced decreased and increased FRET signals, respectively. Another step in the same direction was taken by Yano et al., who produced a novel receptor homomer assay based on the Bioluminescence Resonance Energy Transfer (BRET) technique. The σ₁ receptor was fused in its C-terminal and N-terminal sites with NanoLuciferace (NL) (BRET donor) and the yellow fluorescent protein Venus (VN), which works as an acceptor. The most suitable fusion construct pairs were identified to evaluate the ligand-induced BRET signal, and antagonists were shown to promote higher-order homomerization [35], with results from this assay in agreement with biochemical assays. The same BRET approach was also used to evaluate the heteromerization between the σ₁ receptor and BiP, with the former protein bearing NL and the latter bearing the VN portions. Again, haloperidol and (+)-pentazocine displayed an opposite trend. Nevertheless, the same authors did not use the definition ‘agonist’ or ‘antagonist’ but preferred to discuss the phenotypes of the studied molecules, as some unexpected discrepancies emerged in the ligands studied (e.g., NE100, cocaine). All in all, even if a clear distinction between the two classes is still missing, these fluorescent-based assays provide useful information about the conformational changes that underlie σ₁ receptors’ mechanisms of action.

Proposed endogenous ligands. The σ₁ protein has been considered an orphan receptor for a long time because of the lack of an endogenous ligand. Nevertheless, some years ago, the endogenous hallucinogen compound N,N-dimethyltryptamine (DMT) showed to protect σ₁ receptors by photolabeling performed using two radioactive photoaffinity labels: 3-[¹²⁵I]iodo-4-azidococaine ([¹²⁵I]-IACoc), more specific for σ₁ receptors and 1-N-(2′,6′-dimethyl-morpholino)-3-(4-azido-3-[¹²⁵I]iodo-phenyl)propane ([¹²⁵I]IAF), with affinity for both σ receptor subtypes. DMT provided a dose-dependent protection when either [¹²⁵I]-IACoc or [¹²⁵I]IAF was used. Moreover, DMT showed to inhibit voltage-gated Na⁺ ion channels in both native cardiac myocytes and heterologous cells that express σ₁ receptors. This pharmacological profile suggested DMT as an endogenous agonist for the σ₁ receptor (K_d = 14.75 μM) [37]. In addition to DMT, Progesterone, which binds both σ subtypes (σ₁ K_i = 239 nM, σ₂ K_i = 441 nM [38]), has been also proposed as an endogenous σ₁ receptor antagonist able to coordinate endocrine, immune, and CNS [38,39]. This is not surprising if we consider that σ₁ receptors share homology with yeast sterol isomerase [24]. However, these ligands display low σ₁ receptor affinity, and their blood concentration is too low for these ligands to be considered endogenously active. Plenty of other ligands can bind σ₁ receptors with high affinity. Molecular dynamic simulations at the σ₁ receptor binding pocket have shown how the diverse chemical structures to which high-affinity σ₁ receptor ligands belong may easily be accommodated producing different associations among protomers that lead to the different activities observed.

Reference compounds. The most important classes of ligands are briefly summarized below (Table 1):

• (+)-Benzomorphans as agonists: This class of ligands has a great historical value, as (+)-N-allylnormetazocine ((+)-SKF-10047) led to distinguishing a new group of receptors different from opioid receptors [40] that were named σ, after the first letter of the compound code [41]. Moreover, (+)-Pentazocine is still a σ₁ reference compound, and its tritium radiolabeled form, which shed light on the existence of two classes of σ receptors [4], is still used for σ₁ receptor radioligand binding assays.
• Antipsychotics as antagonists: Haloperidol has high affinity for σ₁ receptors (σ₁ K_i = 8 nM [9]), but the involvement of the σ₁ protein in schizophrenia has not completely been demonstrated. Several investigations on σ₁ receptor gene polymorphs in this context are underway [42,43].
• Antidepressants as agonists: Imipramine, Fluoxetine, Fluvoxamine, and Sertraline showed moderate to high affinities for σ₁ receptors (K_i values ranging from 30 to 300 nM) [44], and several σ₁ receptor agonists have shown antidepressant-like effects in animal models of depression, proving the involvement of the protein in depression [45,46].
• Neurosteroids as agonists/antagonists: This class contains both agonists and antagonists. Progesterone in fact is a σ₁ receptor antagonist, while deoxycorticosterone, testosterone, pregnenolone–sulfate, and dehydroepiandrosterone (DHEA) can produce agonist activities. All these ligands have unconventional structures because of the absence of a basic nitrogen that determines weak affinities for σ₁ receptors. Neurosteroids have been proposed as endogenous ligands for σ₁ receptors, as their effect seems also to be mediated by σ₁ receptor binding [39,47].
• Others: A number of other ligands have been synthesized with the aim to better understand the role of these receptors in different pathologies. The most representative examples are as follows:

  -

  SA4503: a highly selective agonist proved to exert effects on memory, depression, cardiac hypertrophy, and hearing [48,49,50,51,52];

  -

  PB190: a selective agonist that showed neuroprotective and antidepressant activities [53,54,55];

  -

  PRE-084: a selective agonist with beneficial effects on learning impairment [56,57,58];

  -

  NE-100: a σ₁ reference antagonist often used to counteract the action of claimed σ₁ agonists [59,60];

  -

  PB212: a subnanomolar affinity and selective σ₁ antagonist [53,61];

  -

  AC915: a highly selective ligand with an antagonist profile [62];

  -

  Pridopidine: initially considered a dopaminergic D₂ receptor antagonist, it was later repositioned as a selective σ₁ receptor agonist (see Section 3.5) [63];

  -

  FTC-146: the ligand with the highest affinity and selectivity for σ₁ receptor. It has been radiofluorinated and employed in Positron Emission Tomography (PET) in mice, rats, squirrels, monkeys, and humans [64,65,66];

  -

  S1RA: a σ₁ receptor antagonist developed by Esteve company for the treatment of neuropathic pain and the strengthening of opioid analgesia, which successfully completed Phase I clinical trial showing safety and tolerability [67,68];

  -

  BD1063: a σ₁ receptor antagonist with anti-cocaine effects in mice [69,70].

Table 1. Representative σ₁ receptor (σ₁ R) ligands and their affinities for σ₁ R and σ₂ receptor (σ₂ R).

Name	Structure	σ1 R Ki nM	σ2 R Ki nM	Reference for Binding Values
(+)-Benzomorphans:
(+)-SKF-10047		153	154335	[9]
(+)-pentazocine		15.4	3475	[9]
Antipsychotics:
Haloperidol		1.09	41.9	(σ1 R) [71] (σ2 R) [9]
Antidepressants:
Imipramine		343	2107	[44]
Fluoxetine		240	16100	[44]
Fluvoxamine		36	8439	[44]
Sertraline		57	5297	[44]
Neurosteroids:
Progesterone		239	441	[38]
Deoxycorticosterone		938	n.d.	[39]
Testosterone		1014	n.d.	[39]
Pregnenolone-sulphate		3196	n.d.	[39]
DHEA		2.959	n.d.	[47]
Others:
SA4503		4.63	63.09	[72]
PB190		0.42	36.3	[53]
PRE-084		53.2	32.100	[73]
NE-100		1.1	170	[74]
PB212		0.030	17.9	[53]
AC915		4.89	>10000	[62]
Pridopidine		57	5450	[75]
FTC-146		0.0025	364	[64]
S1RA		23.5	>1000	[76]
BD1063		9.15	449	[70]

Table 1. Representative σ₁ receptor (σ₁ R) ligands and their affinities for σ₁ R and σ₂ receptor (σ₂ R).

Name	Structure	σ1 R Ki nM	σ2 R Ki nM	Reference for Binding Values
(+)-Benzomorphans:
(+)-SKF-10047		153	154335	[9]
(+)-pentazocine		15.4	3475	[9]
Antipsychotics:
Haloperidol		1.09	41.9	(σ1 R) [71] (σ2 R) [9]
Antidepressants:
Imipramine		343	2107	[44]
Fluoxetine		240	16100	[44]
Fluvoxamine		36	8439	[44]
Sertraline		57	5297	[44]
Neurosteroids:
Progesterone		239	441	[38]
Deoxycorticosterone		938	n.d.	[39]
Testosterone		1014	n.d.	[39]
Pregnenolone-sulphate		3196	n.d.	[39]
DHEA		2.959	n.d.	[47]
Others:
SA4503		4.63	63.09	[72]
PB190		0.42	36.3	[53]
PRE-084		53.2	32.100	[73]
NE-100		1.1	170	[74]
PB212		0.030	17.9	[53]
AC915		4.89	>10000	[62]
Pridopidine		57	5450	[75]
FTC-146		0.0025	364	[64]
S1RA		23.5	>1000	[76]
BD1063		9.15	449	[70]

3. σ₁ Receptor Ligands in a MTDLs Approach

In addition to the direct neuroprotective action of σ₁ receptors, the additional modulation of targets such as opioid, N-methyl-D-aspartate (NMDA), dopaminergic, and cholinergic receptors renders the σ₁ protein an intriguing target for the modulation of processes that involve these systems, prompting the development of σ₁-based MTDLs for the treatment of complex and multifactorial pathologies such as AD.

3.1. MTDLs Acting at σ₁ Receptor and Cholinergic System

Drug metabolism leads to metabolites that can be more potent or less potent than the parent drug. In the former case, the drug can be considered a sort of prodrug. Thus, the simultaneous presence of the two chemical entities may produce a synergistic powerful effect.

Starting from this idea, in 2010, Lecanu et al. developed pairs of drugs and the corresponding prodrugs, which act on different targets such as acetylcholinesterase (AChE) whose inhibition is a consolidated therapeutic strategy in AD, and σ₁ receptors. With this aim, two series of molecules were built: hydroxyphenyl-butan-1-ones (drug) and their corresponding carbamates (prodrugs). Upon inhibition of the AChE, the prodrug releases the drug. Thus, the prodrug and drug bind and act through the σ₁ receptor and AChE, while the drug exerts its antioxidant action.

Among the studied ligands, two compounds (i.e., 2,3,4-trihydroxy-phenyl)butan-1-one 1 and the corresponding N,N-dimethylcarbamoyloxy prodrug 2; Figure 1) showed the best compromise in terms of activity on the different targets. Compound 2 was able to (i) inhibit AChE with a potency comparable to reference compounds; (ii) bind σ₁ receptors (high selectivity, sub-micromolar affinity, and antagonist activity); and (iii) prevent mitochondrial toxicity by inhibiting mitochondrial complex IV and V. On the other hand, compound 1, even if much less active than its bio-precursor toward σ₁ receptors, was also able to attenuate reactive oxygen species (ROS), Aβ_1-42-induced neurotoxicity, and mitochondrial toxicity by inhibiting complexes I, II, IV, and V.

Moreover, 2 was demonstrated to cross the blood–brain barrier (BBB) very shortly after its administration and be transformed into 1 very quickly.

Behavioral studies on animal models are needed, but preliminary data suggest that this multi-target profile can represent a novel therapeutic strategy to treat AD [77].

Many pharmacological and genetic data have proven that activation of muscarinic M1 receptors (mAChRs) attenuates symptoms of neurodegenerative pathologies [78,79], as in the case of MTDLs able to bind both M1 and σ₁ receptors. The most investigated compound representative of this class is ANAVEX 2-73 (also known as Blarcamesine) (Figure 2). This compound is in advanced clinical phases for several CNS diseases such as AD, Parkinson’s disease (PD), Rett, and Fragile X Syndromes (anavex.com/#!/pipeline, accessed Apr 3, 2021). In addition to binding M1 and σ₁ receptors, Blarcamesine also binds M2–M4 receptors (with micromolar affinity), Na⁺ channel site 2, and NMDA receptor (NMDAR) [80,81]. In this context, Fisher and co-workers, who previously developed orthosteric M1 receptors agonists (e.g., AF102B, AF267B, and AF292), extended their approach in order to target also σ₁ receptors. With this aim, compound AF710B (also known as ANAVEX 3-71, Figure 2) that can activate both M1 and σ₁ receptors with high potency and selectivity was identified [66]. AF710B is a positive allosteric modulator (PAM) of M1 receptor, as it improves the efficacy of carbachol, and this activity, together with a comparable agonism at the σ₁ receptor can preserve synaptic elements in vitro. In vivo studies performed on trihexyphenidyl-treated rats and 3xTg-AD mice showed that AF710B can restore cognitive deficits and attenuate signs of AD phenotypes by the reduction of β-secretase 1 (BACE1) levels, GSK3β and CDK5/p25 activity (which contribute to the hyperphosphorylation of tau protein), neuroinflammation, soluble and insoluble Aβ₄₀ and Aβ₄₂ plaques, and tau pathology [82]. In fact, AF710B reduces the expression of the putative BACE1, so that proteolytic fragments produced by β-secretase were considerably lower in 3xTg-AD treated than in untreated mice. Moreover, tau kinases GSK3β and CDK5 take part in the mechanism of hyperphosphorylation of tau protein. In particular, in AD patients, CDK5 activator p35 is cleaved to produce the protein p25, which binds with high affinity and activates GSK3β [83]. The activation of M1 receptors produces a reduction of GSK3β expression, while the activation of σ₁ prevents the formation of CDK5/p25. Therefore, the inhibition of GSK3β and CDK/p25 by AF710B, upon interaction with both M1 and σ₁ receptors, results as a promising approach in the treatment of tauopathies [82].

The effect of σ₁ receptors in inflammation through microglia modulation has been reported [54,86,87], and AF710B was shown to reduce reactive astrocytes and activated microglia in the animals, as detected by the low levels of glial fibrillary acidic protein (GFAP) and ionized calcium-binding adapter molecule 1 (Iba-1). Notably, astrocytes and microglia are increased in number and size in AD patients.

Another study performed using AF710B on McGill-R-Thyl-APP transgenic (tg) rats revealed that this MTDL can reduce amyloid pathology and markers of neuroinflammation while increasing amyloid cerebrospinal fluid clearance and synaptic marker. The most important achievement is represented by the prolonged duration of these effects, which are maintained five weeks after the treatment is interrupted [88].

In addition to AF710B and ANAVEX 2-73, Anavex Life Sciences Corp portfolio comprehends an isomer of ANAVEX 2-73, named ANAVEX 1-41 (Figure 2) that next to the activity toward σ₁ and M1 receptors, also displays activity for α₁, 5-HT₂, and D₃ receptors [81], with an indication for the treatment of depression, stroke, and neurodegenerative diseases (anavex.com/#!/pipeline).

3.2. MTDLs Acting at σ₁, 5-HT₄ Receptors and AChE

In 2015, Rochais and co-workers developed the MTDL Donecopride from the combination of the AChE inhibitor Donepezil and the serotoninergic 5-HT₄ receptor (5-HT₄R) agonist RS67333 (Figure 3). Indeed, Donecopride inhibits AChE and exerts a partial 5-HT₄R agonist activity on 5-HT₄R.

Subsequent studies revealed that Donecopride has high σ₁ receptor affinity (K_i = 51 nM [90]), which is in agreement with the nanomolar affinity of Donepezil for the same target [91].

This finding inspired many studies about the possible therapeutic exploitation of such combinations of activities [92,93,94,95]. Starting from these pieces of evidence, Lalut et al. postulated that the replacement of the benzene ring by an indole group, together with the insertion of a piperidine-chained spacer, could improve affinity for the σ₁ receptor. Thus, N-substituted and N-unsubstituted indoles were developed, with diverse decoration on the indole and piperidine rings.

The N-unsubstituted indoles displayed a profile comparable to their lead compound in terms of affinity at 5-HT₄ receptors and inhibitory activity at AChE. The most promising ligands were tested in vitro on Jukart cell membranes for the determination of the σ₁ affinity, according to Ganapathy [96].

The N-benzyl derivative 3 (Figure 4) displayed the highest affinity for the σ₁ receptor and the most potent inhibition toward AChE but a moderate affinity for 5-HT₄ receptors. The best compromise among affinities at σ₁ and 5-HT₄ receptors and activity at AChE was reached by the cyclohexylmethyl derivative 4 (Figure 4), which was thus considered the lead compound of the study.

This compound was tested in vivo using the dizocilpine-induced amnesia assay in mice. In the 0.1–1 mg/kg dose-range, 4 did not attenuate the dizocilpine-induced spontaneous alternation performance deficit, while at lower doses, the passive avoidance deficit was attenuated with a 40% protection. The authors concluded that the involvement of the σ₁ receptor needed further explorations by including challenges with the σ₁ receptor antagonist NE-100 [90].

3.3. MTDLs Acting at σ₁ Receptor, AChE, BuChE, BACE1, MAO-A, MAO-B, 5-LOX, ROS, and Stem Cells

Another important class of MTDLs able to treat AD is represented by ligands developed by Estrada and co-workers in 2016. Their strategy was based on the synthesis of MTDLs able to inhibit AChE/Butyrylcholinesterase (BuChE), BACE1, and bind σ₁ receptors.

For the study, the authors combined lipoic acid (LA), whose antioxidant properties [97] are exploited in AD [98], together with either the N-benzylpiperidine (NBP) moiety (featured by Donepezil, as discussed above) or the N,N-dibenzyl-N-methylamine (DBMA) moiety of the BACE1 inhibitor AP2238. This approach led to the MTDLs LA-NBPs and LA-DBMA, respectively.

More than fourteen ligands were developed, and some of them were tested for σ receptors affinity. K_i values in the sub-micromolar and sub-nanomolar range were found for σ₁ receptors, and radical scavenging properties were evaluated using the Oxygen Radical Absorbance Capacity (ORAC) assay, with LA-based ligands showing good antioxidant capacities, comparable to vitamin E.

In the LA-NBP class, the lowest K_i value at the σ₁ receptor was displayed by ligand (R)-5; in the LA-DBMA series, the best σ₁ receptor affinity values were displayed by enantiomeric mixtures of compounds 6 and 7, which differ for the position of the amidic substituent on the central aromatic ring (3-position for compound 6 and 4-position for compound 7) (Figure 5). Both compounds showed K_i values in the nanomolar range, with no enantioselectivity.

In addition, compounds 5 and 7 showed promising in vitro absorption, distribution, metabolism, excretion and toxicity (ADMET) properties, as they are able to cross the BBB by in vitro parallel artificial membrane assays (PAMPA-BBB). They also displayed low or moderate cytotoxicity in the SH-SY5Y neuroblastoma cell line (supporting their s₁ agonist or partial agonist rather than antagonist activity) [99].

In order to develop a more holistic strategy to hit AD, the same group designed novel MTDLs to target monoamine oxidase A (MAO-A), monoamine oxidase B (MAO-B), lipoxygenase-5 (5-LOX), AChE/BuChE, and σ₁ receptors. Additionally, ROS attenuation and mechanisms of differentiation of neural stem cells were taken into account for these novel compounds. Thus, twenty-nine new Donepezil-flavonoid hybrids (DFHs) were synthesized. The N-benzylpiperidine fragment was connected through an amide or ester linker to the 4-oxo-4H-chromene flavonoid scaffold, which is endowed with MAO-A, MAO-B, and 5-LOX inhibitory activity, as well as antioxidant, anti-inflammatory, and neurogenic properties. Moreover, the flavonoid scaffold underwent some modifications such as substitution of the 4-oxo-4H-chromene with a quinoline-4(1H)-one or quinoline and the introduction of hydroxyl/methoxy groups on the aromatic ring.

Three classes of compounds (i.e., 4-chromenone, 4-quinolone, and quinoline-2-carboxamide) were tested for their activity at AChE/BuChE, 5-LOX, and MAOs as well as for their BBB permeability and antioxidant capacities. A restricted group of ligands belonging to the 4-chromenone and 4-quinolone classes (general structure in Figure 6) was selected and tested at σ receptors, showing low nanomolar K_i values (0.98 nM > K_i > 47.4 nM) and selectivity versus the σ₂ subtype. The best affinities were obtained when a single hydrogen-bond acceptor group (NO₂ > NH₂ > OH > OCH₃) was in the 6-position of either the 4-oxo-chromene or 4-oxo-quinoline system, while the length of the linker between the amide and the N-benzylpiperidine groups had no influence on the affinity for σ₁ receptors.

The best compromise among all studied compounds is represented by compound 8 (R = 6,7-diOMe; Z = O; X = NH and n = 2, Figure 6), which showed good neuroprotective properties: hAChE IC₅₀ = 46 nM; hBuChE IC₅₀ >10000 nM; PAMPA-BBB = 5.9 × 10⁻⁶ cm s⁻¹; 5-LOX IC₅₀ = 74.3 µM; MAO-A IC₅₀ = 15.3 µM; MAO-B IC₅₀ = 5.2 µM; ORAC = n.a.; σ₁ R K_i = 37.4 nM; σ₂ R K_i = 239 nM.

Moreover, compound 8 was shown to promote the differentiation of adult mice neural stem cells (NSC), and its predicted metabolites have been analyzed using the Derek Nexus program and shown to be not toxic. Additionally, docking by Molecular Dynamics (MD) simulation into AChE, 5-LOX, and the σ₁ receptor showed low ΔG_bind values in agreement with the experimentally detected interactions with the three targets. Based on these data, compound 8 was proposed as a promising candidate against AD and neurodegenerative diseases in general [100].

Based on these encouraging data, the same group developed other MTDLs targeting AChE/BuChE, BACE1, MAOs, 5-LOX, the differentiation of neural stem cells, and σ₁ receptors. With this aim, the already exploited flavonoid cores (i.e., 4-chromenone or 4-quinolone) were conjugated with the N,N-dibenzyl-N-methylamine (DBMA) fragment, which was chosen for its proven interaction with the catalytic anionic site (CAS) of AChE. The two classes of 4-chromenone and 4-quinolone derivatives were tested, and the best compromise was obtained with ligands belonging to the former class. The most promising ligands (i.e., 9, 10, 11, and 12, Figure 7) were also tested at σ₁ and σ₂ receptors, showing comparable sub-micromolar affinities for σ₁ and micromolar affinities for σ₂ receptors. These results showed that functionalization in 3′ or 4′ of the central benzene ring as well as changes in the position of the methoxy/hydroxy groups on the 4-chromenone scaffold poorly influence the σ receptors’ affinity.

Despite its low affinity and selectivity for σ₁ receptors (σ₁ R K_i = 0.53 µM; σ₂ R K_i = 1.30 µM), compound 12 represented the best compromise in terms of activity toward different targets (hAChE IC₅₀ = 4.5 µM; hMAO-A IC₅₀ > 100 µM; hMAO-B IC₅₀ > 100 µM; 5-LOX IC₅₀ = 30.4 µM; and PAMPA-BBB = 18.8 × 10⁻⁶ cm s⁻¹). ΔG_bind values obtained through docking to AChE, 5-LOX, BACE1, and σ₁ were in agreement with the experimental ones, while a relatively safe profile was predicted in terms of toxicity. However, a low-confidence alert (below 67%) for human ether-a-go-go-related gene (hERG) channel inhibition was recorded, so that cardiac toxicity might be expected with this compound, which needs to be ascertained with further experimental studies. Neurogenic evaluation was performed on a primary culture of neural stem cells from the SGZ of adult rats, and compound 12 stimulated differentiation. Altogether, these data suggest compound 12 as a possible therapeutic agent that promotes brain auto-repair processes and blocks early steps of neurodegenerative cascades [101].

3.4. MTDLs Acting at σ₁, NMDA Receptors, and VGCC

NMDA receptors are glutamate-gated Ca²⁺ permeable ion channels associated with σ₁ receptors. In 1987, Quirion et al. identified the σ₁ receptor as the phencyclidine (PCP) binding site [102], but further studies performed by the same group refuted the thesis and classified σ receptors as a separate class [103]. However, a strong connection between these two receptors has been found in the following years. σ₁ Receptors can enhance Ca²⁺ influx via NMDAR, but they can also prevent toxicity induced by the excessive flux of this cation. This activity occurs through the modulation of NMDAR-associated intracellular signaling component, i.e., the neuronal Nitric Oxide Synthase (nNOS), which produces nitric oxide (NO) [104].

Low doses (100 nM) of σ₁ receptor agonists enhance NMDAR function, producing a potentiation of the NMDAR response. Curiously, high doses (≥10 μM) do not exert a stronger effect. Indeed, a bell-shaped dose–response trend, which is typical of the σ₁ receptor agonists-mediated action, can be observed accounting for the majority of the effects of σ₁ receptor agonists on NMDAR response [105,106]. A potential explanation of this effect is that σ₁ receptor agonists can directly bind NMDARs and block channel conductance [107], but further studies are needed to solve the question. Moreover, the NMDAR increased activity is fast and prolonged [108], so that interactions also with other targets may be involved [104]. Even if these mechanisms have not been completely clarified, an interaction between the σ₁ receptor and NMDAR exists. In particular, as recently reviewed, σ₁ receptors can associate together with GluN1, Glun2a, and Glun2b subunits [17].

Taking into consideration all these aspects, dual drugs acting on σ₁ and NMDA receptors may be a promising strategy to target neurodegenerative disorders.

With the aim to discover potentialities of the aza-cage compound of the pentacyclo[5.4.0.0^2,6.0^3,10.0^5,9]undecane ring system (PCU), Van der Schyf et al. engaged physiological studies using sheep Purkinje fibers. These studies classified NGP1-01 (Figure 8) as the lead compound of a novel class of Ca²⁺ channel antagonist having a “cage” molecular structure [109].

Subsequent studies demonstrated the multi-target profile of NGP1-01 evidenced by its ability to reduce Ca²⁺ influx through interaction with L-type voltage-gated calcium channel (VGCC) [113] and NMDAR, with an activity comparable to the NMDAR antagonist memantine [114].

The discovery that amantadine had micromolar K_i values for σ receptors [115] led to investigate whether cage compounds could bind σ receptors. The micromolar affinity for σ₁ receptors of NGP1-01 inspired the synthesis of new trishomocubanes pentacyclo[5.4.0.0^2,6.0^3,10.0^5,9]undecylamines and 4-azahexacyclo[5.4.1.0^2,6.0^3,10.0^5,9.0^8.11]dodecanes, such as ANSTO-1 and ANSTO-6 (Figure 8). These compounds showed good affinity for σ receptors but low affinity for NMDAR (i.e., ANSTO-1 and ANSTO-6, Figure 9) [110,111].

The subsequent SAR, 3D-QSAR, and docking studies clarified the following: (i) aza-trishomocubane derivatives have high affinity for both σ receptors; (ii) affinity for σ₁ receptors is higher when the chain is longer; (iii) para- or meta-substitution with halogens is well tolerated, but the absence of substituents produces the highest selectivity and the lowest K_i value for σ₁ receptors (ANSTO-14, ANSTO-5, and ANSTO-7, Figure 8) [103,104].

Nevertheless, the improved σ₁ receptor affinity was accompanied by a dramatic loss in the affinity for NMDARs. Consequently, NGP1-01 can still be considered the best MTDL in the class of the cage compounds.

Deeper studies able to clarify the mechanism of the σ₁ receptors-mediated antineurodegenerative activity of NGP1-01 are not available, to the best of our knowledge, but the affinity of the compound for the σ₁ receptor and the antineurodegenerative properties mediated by the reduction of Ca²⁺ influx have been demonstrated, qualifying NGP1-01 as a potential MTDL that is useful in the treatment of neurodegenerative pathologies.

Among cage compounds, amantadine and memantine (Figure 9) deserve to be mentioned because of their well-known antagonistic activity at the phencyclidine (PCP) binding site of NMDARs [116] together with the already mentioned affinity for σ₁ receptors [115]. However, while the affinity of memantine at NMDARs is around 40-fold higher than the affinity at σ₁, amantadine interacts to the same extent with σ₁ and NMDARs. Additionally, Peeters et al. confirmed the agonist σ₁ receptor profile and shed the light on amantadine ability to positively increase dopaminergic transmission through the activation of σ₁ receptors [117]. Therefore, amantadine can be considered a full-fledged MTDL.

3.5. MTDLs Acting at σ₁ Receptor/Dopaminergic Transmission

σ₁ Receptors are also associated with the dopaminergic transmission. BRET, coimmunoprecipitation (Co-IP), and proximity ligation assays studies demonstrated that σ₁ receptors interact with dopamine (DA) D₁ and D₂ receptors, as well as with dopamine transporter (DAT) [17]. Beneficial effects of the σ₁ chaperone function on dopaminergic transmission may be exploited in pathologies such as AD, PD, Huntington’s disease (HD), and schizophrenia.

Interactions with D₁ receptors were studied by Moreno et al. that performed in vitro and ex vivo assays, which led to identify, in the presence of cocaine, the formation of a heterotrimer composed of σ₁, histamine H₃, and D₁ receptors. Upon the formation of this trimer, the inhibition of D₁ receptors was abolished, and G_s proteins were activated with an increase of cAMP, recruitment of β-arrestin, and increase in p-ERK1/2 levels [118]. As a result, σ₁ antagonists have the potential to reduce cocaine effects and consequent cocaine addiction.

σ₁ Receptors can also form heteromers with the D₂ receptor. Biophysical, biochemical, and cell biology assays performed on mouse striatum demonstrated that the administration of cocaine can induce the formation of σ₁–D₂ dimers, inhibiting the downstream signaling [119].

Finally, σ₁ receptors are also able to associate with DAT and reduce DAT-mediated dopamine efflux produced by methamphetamine exposure [120].

A large number of dopaminergic agents have been examined to better define their pharmacological profiles, and some of them were found to bind σ₁ receptors, in accordance with a partially overlapping pharmacophore.

For a long time, Pridopidine (Figure 10) has been considered a dopaminergic antagonist because of its micromolar affinity for D₂ receptors, but some in vivo data and its structural resemblance to the σ₁ receptor ligands inspired Sahlholm and co-workers to perform studies that defined Pridopidine as a highly selective σ₁ receptor agonist [63,121].

To better understand the neuroprotective and neurorestorative role of Pridopidine, Francardo et al. assayed the drug in a mouse model of Parkinson’s disease (unilateral 6-hydroxydopamine (6-OHDA) lesion). After 5 weeks of daily administration of a low dose (0.3 mg/kg), Pridopidine produced an improvement in forelimb use and abolished the ipsilateral rotational bias typical of hemiparkinsonian animals. Moreover, the protection of nigral cell bodies, increased dopaminergic fiber density in the striatum, striatal upregulation of glial cell line-derived neurotrophic factor (GDNF), and brain-derived neurotrophic factor (BDNF) and phosphorylated extracellular signal-regulated kinases 1/2 (ERK1/2) were also observed [122]. At higher doses, Pridopidine was able to interact with other targets such as alpha-2 adrenergic α_2C and serotonin 5-HT_1A receptors, providing a solution to reduce L-dopa-induced dyskinesia (LID), both in PD and HD [75]. Of note, a phase III trial is in progress to evaluate the efficacy and safety of Pridopidine in HD patients (PRidopidine’s Outcome on Function in Huntington Disease, PROOF-HD, ClinicalTrials.gov Identifier: NCT04556656). Importantly, promising data of Pridopidine for the treatment of AD are emerging as the compound prevents mushroom spine loss in hippocampal cultures from APP knock-in (APP-KI) and presenilin-1-M146 V knock-in (PS1-KI) mice [123].

The ability to functionally interact with different receptors identifies Pridopidine as a MTDL, which is useful in the treatment of neurodegenerative disease.

Afobazole (Figure 11) is an anxiolytic drug designed and pharmacologically studied by FSBI “Research Zakusov Institute of Pharmacology”, Russia. This drug binds melatonin MT₁ (K_i = 16 µM) and MT₃ (K_i = 0.97 µM) receptors. MT₃ is a regulatory site of quinone reductase 2 enzymes (NQO2), which is responsible for ROS production. For this reason, the reversible inhibition performed by Afobazole produces a cytoprotective effect.

Some studies performed on Afobazole metabolites revealed that one of them, M-11, can bind the MT₃ receptor (K_i = 0.39 µM) with a slightly higher affinity, while a lower affinity is associated to the MT₁ receptor (K_i = 44 µM). Consequently, Afobazole activity can be considered empowered by the additive activity of its metabolite M-11 (Figure 11) [124].

Afobazole can also inhibit MAO-A, thus increasing adrenaline, noradrenaline, and dopamine levels in CNS. Additionally, it also behaves as a σ₁ receptor regulator, producing a cytoprotective effect through the inhibition of ROS production [125].

The need to clarify the involvement of σ₁ receptors in neurodegenerative diseases inspired some studies on the 6-OHDA-induced parkinsonism model in mice. The study revealed that the administration of Afobazole (2.5 mg/kg, i.p.) over 14 days was able to restore motor dysfunction as well as prevent decreases in dopamine in the 6-OHDA-lesioned striatum and the loss of tyrosine hydroxylase positive (TH+) neurons in the substantia nigra. Importantly, these activities were abolished by the administration of the σ₁ receptor antagonist BD-1047 (3.0 mg/kg, i.p.), demonstrating that the effects of Afobazole are mediated by the σ₁ receptor [126].

These pieces of evidence prove a strict connection between the dopaminergic transmission and the σ₁ receptor and encourage more in-depth investigation to clarify the mechanisms involved.

With high affinity and antagonist properties at D₂ receptors, Haloperidol (Figure 12) equally binds the σ₁ receptor, behaving as an antagonist [127].

After administration, Haloperidol undergoes a complex metabolism that leads to toxic metabolites such as 4-(4-chlorophenyl)-1-[4-(4-fluorophenyl)-4-oxybutyl]pyridinium ion (HPP+) and 4-(4-(chlorophenyl)-1–4-(fluorophenyl)-4-hydroxybutyl-pyridinium (RHPP+) (Figure 12). In addition, also non-toxic metabolites have been observed (HP metabolites), such as HP metabolite II [130] and HP metabolites I and III [131] (Figure 12). Biological activities of the non-toxic HP metabolites (I, II, and III) were studied, and in particular, their affinities for σ₁ receptors were evaluated. While HP metabolite III does not bind σ₁ receptors, HP metabolite I exerts a reversible inhibition, and HP metabolite II exerts an irreversible inhibition at the σ₁ receptor. Thus, haloperidol metabolites can reinforce its action [128].

Further studies are needed to better understand the exploitability of this pharmacological activity, but preliminary data suggest that σ₁ receptor antagonists, such as Haloperidol, could be useful in blocking the acute toxicity and the rewarding effects of cocaine because of an influence on the DA release (pre-synaptic activity) and reduction of Ca²⁺ mobilization produced by the activation of D₁ receptors (post-synaptic activity) [132]. The additional activity of Haloperidol and its active metabolites could be foreseen in the antinociceptive action occurring through the modulation of opioid receptors [133].

Moreover, studies performed by Marrazzo and colleagues revealed that HP metabolite II is also able to exert anticancer activity because of its antagonism at the σ₁ receptor and agonism at the σ₂ receptor [134]. A more detailed discussion of HP metabolite II anticancer potentialities is reported in Section 3.7.

3.6. MTDLs Acting at σ₁ and μ-Opioid Receptors

One of the most difficult challenges in medicinal chemistry is represented by the synthesis of drugs able to manage pain for a prolonged period. In particular, chronic pain therapies are often associated with the adaptation phenomenon that is responsible for most of the side effects of these drugs, because it leads to periodic increases of the doses to be administered in order to keep the analgesia. Opioids, with their generally low therapeutic index, are still the most effective class of drugs able to relieve pain, making the need to increase the dose extremely life-threatening.

With this in mind, Garcia et al. developed a heterogeneous class of ligands with the aim to reinforce the antinociceptive activity mediated by μ-opioid receptor (MOR) agonists with an antagonist activity at the σ₁ receptor that has a well-established role in the treatment of pain. Some pieces of evidence have demonstrated that the antagonist S1RA [124,127] is able to potentiate morphine antinociceptive activity without increasing side effects, so that it entered clinical trials for treating different kinds of pain [135,136].

These premises paved the way to the development of a pharmacophore model obtained by overlapping the MOR pharmacophore generated with the “Auto Pharmacophore generation” protocol (that enumerates all the pharmacophores compatible with an input molecule considering hydrogen bond acceptors and donors, hydrophobic portions, ionizable groups, and aromatic rings) within the Discovery Studio [137] with the σ₁ pharmacophoric model developed by Langer [138]. This model was based on the presence of two hydrogen bond acceptors (HBA1 and HBA2), three hydrophobic portions (HYD1, HYD2, and HYD34, obtained by the fusion of two different hydrophobic portions HBA3 and HBA4 of Laggner’s pharmacophore), an aromatic ring (AR) and two positive ionizable domains (PIs).

Based on this model, several compounds were synthesized, and the 4-aryl-1-oxa-4,9-diazaspiro[5.5]undecane derivative 13 (Figure 13) provided the best results on both targets.

Starting from compound 13, SAR studies were performed to optimize the structure. In particular, the 2-position on the spiro[5.5]undecane was explored through the introduction of small substituents such as methyl groups (compound (R)-14, Figure 13), which increased the activity mediated by MOR and σ₁ receptors, while reducing the activity at the hERG channel, which was responsible for its cardiac toxicity. Moreover, pure enantiomers were tested demonstrating that (R)-14 is 4-fold less active toward hERG and displays a 1-order of magnitude higher affinity for MOR than its (S)-counterpart. Elongation of the same alkyl group from methyl to ethyl or isopropyl produced an increase in the activity toward MOR but decreased the activity toward σ₁ receptors, thus spoiling the beneficial synergy. Substitution with more polar groups produced compounds characterized by a lower affinity than compound (R)-14, but the activity at the two targets reached a more balanced compromise. Nevertheless, the reduction in lipophilicity did not improve the profile at hERG, which remained quite similar to the inhibition exerted by the more hydrophobic compound (R)-14.

Another structural exploration was conducted by introducing substituents on the phenyl ring on the 4-position of the 1-oxa-4,9-diazaspiro[5.5]undecane system. In particular, halogens and trifluoromethyl groups in ortho positions provided the best results toward the two targets, which was probably because of an increase in steric bulk. On the other hand, no improvement toward hERG inhibition was observed. Position 4 was also explored by introducing heteroaryl groups, and the best results were obtained when 2- or 3-pyridyl systems were inserted. As for the 9-position on the 1-oxa-4,9-diazaspiro[5.5]undecane system, the best results were obtained with the same groups introduced in position 4.

Among all of these compounds, 15 (Figure 13) resulted as the best compromise because of its activity toward the two targets and appreciable selectivity, with the most important achievement represented by the absence of hERG inhibition.

The promising pharmacodynamic profile, together with good solubility, moderate basicity and lipophilicity, and in vitro metabolic stability of compound 15 led the group to study its antinociceptive properties in vivo. Results from the preclinical model, based on the mechanical pressure test on a hindpaw, demonstrated a dose-dependent antinociceptive effect with an ED₅₀ of 15 mg/kg after i.p. administration. Moreover, local administration of the compound provided promising results [139]. Further optimization of the compound was conducted with the production of new dual ligands with enhanced activity. Replacement of the aromatic ring in the 4-position of the 1-oxa-4,9-diazaspiro[5.5]undecane system with an ethyl group provided encouraging data. In fact, small alkyl groups produced a slightly reduced affinity toward MOR and σ₁ receptors but considerably improved the profile toward alpha-1 adrenergic receptor (α₁AR) and hERG, which was likely because of a reduced lipophilicity. As already shown by structural changes leading to compound 14, different small alkyl substituents were well tolerated at the 2-position (spirocyclopropyl, ethyl, and (R)-methyl groups), improving the affinity toward the two targets and the hERG profile. These optimization studies led to compound 16 (Figure 13), which showed good physicochemical properties, high BBB permeability, and no cytotoxic activity in the hepatocarcinoma HepG2 cells. For these reasons, compound 16 was successfully tested in the paw pressure test and in the partial sciatic nerve ligation (PSNL) model in CD1 male mice. Thus, compound 16 was selected as a Phase 1 clinical candidate in the treatment of pain [140].

3.7. σ1. Receptor and Cancer in a MTDLs Approach

The σ₁ receptor is a well-established target in the neurodegenerative field. However, several cancer cell lines (e.g., lung: H69, H2019, H510; breast: MDA-MB-361, MDA-MB-435, BT20, MCF-7, T47D, MDA-MB-231, MDA-MB-468, SKBR3, BT474; prostate: PC3, LNCaP, LAPC4, C4-2, 22Rv1, VCaP, PC3, DU145; human esophageal squamous cell carcinoma: KYSE150, KYSE180, EC109; pancreas: Panc1; liver: HepG2; neuroblastoma: SK-N-BE(2)C) show high levels of σ₁ proteins or SIGMAR1 transcripts. While the receptor is not an oncogene [141], σ₁ receptor ligands with claimed antagonist activity exert antiproliferative and growth-inhibiting effects [142].

The involvement of the σ₁ receptor in cancer prompted Riganas et al. to develop (1-adamantyl)diarylalkylamines as σ₁ receptor ligands whose scaffold was based on two aromatic rings and a hydrophobic adamantane moiety, which were linked to an amine site through alkyl chains of different length. The activity of these ligands was also tested on Na⁺ channels, whose involvement in cancer has been proven [141,143]. Compounds from the butyl series displayed the best profile with good σ₁ receptor antagonist and weak σ₂ receptor agonist activities, while all the other compounds showed a lower selectivity for σ₁ vs. σ₂. Moreover, all of them showed from micromolar to sub-micromolar affinities for the site 2 of Na⁺ channels. Compound 17 (Figure 14) emerged for its cytotoxic effect in ovarian cancer cells (IGROV-1) and in vitro antiangiogenic activity in normal cell lines such as Human Umbilical Vein Endothelial Cells (HUVEC).

Compound 17 was characterized by high-affinity interaction with the σ₂ receptor as agonist (i.e., inducer of cytotoxicity) and with the σ₁ receptor as antagonist, representing the optimal compromise if anticancer agents based on σ receptors-mediated action want to be produced. Additionally, 17 was associated with an analgesic effect against neuropathic pain, which was probably due to the block of Na⁺ channels, thus highlighting the therapeutic potentials of the combination of these actions in one molecule [145].

Another ligand conveniently endowed with antagonism at σ₁ receptors and agonism at σ₂ receptors is PB28, which has been found as a potential anti SARS-CoV-2 agent, and for this reason, researchers have recently reviewed its structure–affinity/activity relationships [144,146]. This compound emerged from structure–affinity relationship (SAfiR) studies that clearly indicated how N-alkyl-piperazine connected through a methylene linker to a hydrophobic nucleus (i.e., tetralin), rather than the corresponding N-Aryl ones, addressed the σ receptors affinity, and weakened the interaction with dopamine D₂ and serotonin 5-HT_1A receptors [71,147]. With its sub-nanomolar affinity toward both σ receptor subtypes, and the combination of functional activities (σ₁ antagonism and σ₂ agonism), PB28 was tested in several cancer cells (e.g., breast, neuroblastoma, pancreas tumor cell lines) [144]. Moreover, studies performed in pancreatic cancer cell lines (i.e., Panc02, KP02, KCKO, MIA PaCa-2, BxPC-1, and Panc-1) highlighted the increase in ROS production, lysosomal membrane permeabilization (LMP), and mitochondrial superoxide production as PB28 mechanisms of action that lead to oxidative stress [148]. However, the PB28 anticancer effect both in vitro and in a preclinical model was not striking despite the promising σ binding profile. With the aim of increasing the cytotoxic activity, structural changes were made and several analogues of PB28 were synthesized, some of which were endowed with promising anticancer action. As an example, the piperidinyl analogue PB221 (Figure 15), with reduced σ receptor affinity and a more σ₂-oriented profile, was thoroughly investigated in CNS tumors and displayed promising results in preclinical models [149]. Additionally, the cytotoxic activity of PB28 in resistant tumors was increased through the generation of multi-target ligands directed toward σ receptors (the σ₂ receptor in particular) and P-glycoprotein, with the aim to evade the P-gp efflux whose overexpression is responsible for the drug-resistance phenomenon. While most of the compounds were active both in resistant and wild-type (wt) cancer cells [150], some MTDLs were endowed with the collateral sensitivity phenomenon that is characterized by more potent cytotoxicity in resistant rather than in the wt cells. The promising data obtained prompted to push the MTDL concept further in this context by merging the σ/P-gp scaffold with a metals chelator portion in order to stress the collateral sensitivity property thanks to the more important impact that the chelation of metals, such as iron and copper ions, has on the energy state of cells [151]. Excellent results were obtained with this class of compounds with a thiosemicarbazone structure that showed advantages in preclinical pancreatic tumor models, where the σ-mediated delivery to pancreatic tumors appeared devoid of side effects [152]. However, the σ₁ involvement was not investigated in the overall effect, as focus on the σ₂ subtype was given.

PB28, also Haloperidol metabolite II (Figure 12), is characterized by a σ₁ receptor antagonism/σ₂ receptor agonism profile, and it is able to exert modest antiproliferative activity because of the increase in the Intracellular-Free Calcium Levels [Ca²⁺]_i and apoptosis induction [155].

This evidence supported the synthesis of (R,S)-MRJF4 (Figure 16), which combined the Haloperidol metabolite II profile together with 4-phenylbutyric acid, which inhibits Histone Deacetylase (HDAC) and is involved in the transcription of genes that regulate the cell cycle, with a clear role in cancerogenesis and cancer progression.

The esterification of Haloperidol metabolite II with 4-phenylbutyric acid appeared to be detrimental for the affinity toward both classes of receptors, with a drop from nanomolar to low micromolar values. Nevertheless, antiproliferative studies demonstrated that (R,S)-MRJF4 is able to produce a more potent antiproliferative effect in LNCaP and PC3 than 4-phenylbutyric acid or Haloperidol metabolite II alone or in combination [134].

In the following years, the pure enantiomers (R)-MRJF4 and (S)-MRJF4 were studied, showing that (R)-enantiomer has a higher affinity for both σ receptors and higher anticancer activity (Figure 16) [129].

Noteworthy are some studies performed by Nieto et al. about cancer-associated pain in which paclitaxel was used to induce allodynia. This effect was observed in wild-type mice but not in the SIGMAR1 KO mice. Moreover, σ₁ receptors antagonists, such as BD1063 and S1RA (Table 1), reverted paclitaxel-induced neuropathic pain in wild-type mice [156,157].

Altogether, these studies show that the antiproliferative and growth-inhibiting activity of σ₁ receptor antagonists can be empowered and likely better oriented when associated to other activities. Importantly, these same antagonists can block cancer-associated pain, which is an effect probably strengthened by the block of Na⁺ channels, thus again supporting the MTDLs approach involving σ₁ receptors also in the treatment of pain associated to cancer.

4. Conclusions and Perspectives

This review aims to point out the state of the art of MTDLs interacting with the σ₁ receptor. As a result of its pluripotent chaperone activity, the σ₁ receptor has the potential to interact and modulate several targets involved in a number of diseases and syndromes. The evidence that several compounds already in clinical use (e.g., Donepezil, HP, etc.) unintentionally bind the σ₁ receptor has shed light on the already exploited interaction with this protein as a successful strategy in treating illnesses or associated symptoms through a MTDLs approach. Donepezil, whose interaction with σ₁ receptors in vivo has been ascertained by receptor occupancy studies in the human brain [94,158], is co-administered together with Memantine through prescription medicines (i.e., Namzaric). On this basis, the rational development of MTDLs with a σ₁ intentional targeting profile may lead to more important therapeutic actions. The amount of herein listed proteins, which are the object of the MTLDs development in association with the σ₁ receptor, is not intended to be exhaustive. Other proteins may be taken into consideration in the MTDLs development, considering the wide range of σ₁ receptor direct and indirect interactors. Cannabinoid type-2 (CB2) receptors are an example, with recent support from computational methods that point out a partial overlap of the σ₁ and CB2 receptors’ pharmacophores [159]. These data strongly suggest the adoption of a merging approach for the development of dual σ₁/CB2 receptors ligands for a synergistic exploitation of the pathways activated by the two targets in oncology and neurodegenerative diseases. A number of σ ligands, besides interacting at the two subtypes (see Section 3.7), are able to modulate the efflux pump P-glycoprotein (P-gp) [150,160]. This double action is worthy of being more intentionally addressed, as it may serve as a strategy to face drug-resistant tumors: the compound overcomes P-gp, which is overexpressed in many resistant tumors and exerts its cytotoxic activity (as a σ₁ antagonist). Importantly, this same double action may also have a role in the treatment of CNS diseases through the bypass of the P-gp at the BBB. The waving interest in σ receptors research increased during the COVD-19 pandemic, when the proteins were found as important key host dependency factors for coronavirus infections. However, while the exploitability of the σ₁ receptor as a druggable target against coronavirus still needs to be fully investigated and validated, it appears clear that this protein is involved in a plethora of pathways hampered in multifactorial CNS and cancer diseases, rendering the σ₁ receptor a full-fledged target for the development of multifunctional therapeutics.