Dioxane as a scaffold for potent and selective 5-HT 1 AR agonist with in-vivo anxiolytic , anti-depressant and anti-nociceptive activity

Please cite this article as: S. Franchini, C. Sorbi, P. Linciano, G. Carnevale, A. Tait, S. Ronsisvalle, M. Buccioni, F. Del Bello, A. Cilia, L. Pirona, N. De Nora, R.M. Iacobazzi, L. Brasili, 1,3-Dioxane as a scaffold for potent and selective 5-HT1AR agonist with in-vivo anxiolytic, anti-depressant and antinociceptive activity, European Journal of Medicinal Chemistry (2019), doi: https://doi.org/10.1016/ j.ejmech.2019.05.024.


M A N U S C R I P T
A C C E P T E D ACCEPTED MANUSCRIPT 5 In this work, we explored the distance between the basic centre and the 2,2-diphenyl portion ( Figure   2a, b), the expansion (Figure 2c, d), the opening and the simplification of the 1,3-dioxolane ring (Figure 2e, f), as a way to increase the intrinsic activity and selectivity for 5-HT 1A R. A detailed SAR study was carried out, and a potent and selective 5-HT 1A R agonist with in-vivo anxiolytic, antidepressant and anti-nociceptive activity was discovered.

Structure-affinity and structure-activity relationship studies
Compounds 3-16 were tested for binding affinity (pK i ) and activity (pK b ) at human α 1 and 5-HT 1A receptors . The most active and selective compounds were chosen for their functional characterization (pD 2 and %E max ).
In a previous paper, we showed that compounds 1 and 2 bind to both α 1 and 5-HT 1A R receptors [12]. Binding studies in human cloned receptors have shown that at all the three α 1 subtypes (α 1a, α 1b and α 1d ), the affinities of both compounds are quite similar, with compound 2 showing a small, but scarcely significant, preference with respect to 1. In the case of 5-HT 1A R, compound 2 has a higher affinity than 1, of about one order of magnitude (9.22 vs 8.45). During the functional studies, both behaved as antagonists at α 1 adrenoceptors and partial agonists at 5-HT 1A R. Compound 1 showed a selective profile towards α 1D (more than 100-fold), with respect to the α 1A and α 1B subtypes. The functional data for the 5-HT 1A R indicated that the agonist potency decreased by about 28-fold, going from 1 to 2 (pD 2 of 8.8 and 7.36 respectively). This is contrary to the trend observed in the binding experiment, where the affinity of 1 was 6-fold lower than that of compound 2 (pK i M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT 10 8.45 and 9.22 respectively). Therefore, the presence of the methoxy group in the ortho position had a positive effect on binding and a negative one on agonist potency. Furthermore, during the functional studies at α 1 adrenoceptors , the methoxy group increased potency but decreased selectivity towards the α 1D subtype, while leaving the affinity at the α 1 subtypes in the binding studies almost unchanged.
away by the insertion of a methylene in the lateral chain, with respect to 2, showed a small increase in binding affinity (pK i ) at all α 1 subtypes, while leaving unchanged the affinity at 5-HT 1A R.
During the functional studies, the antagonist potency of 4 at α 1 subtypes was slightly decreased, while the selectivity towards the α 1D subtype was almost the same (about 10-fold). At 5-HT 1A R, the agonist potency was also decreased by about 17-fold. The same modification made on the desmethoxy derivative 1 to give 3 produced different results. In fact, the binding affinities were decreased and the decrease was much larger for 5-HT 1A R. During the functional studies at α 1, the most significant variation was a complete loss of α 1D selectivity, due to a more than 10-fold decrease in potency at this subtype and a concomitant, although small, increase at the α 1A and α 1B subtypes. In compound 6 the diphenyldioxolane was replaced with a benzhydryldioxolane to increase the distance between the diphenyl moiety and the central amine and two disatereomers were obtained (t and c). The most important change was in the selectivity 5-HT 1A /α 1, which was greatly reduced, as a result of a significant decrease in 5-HT 1A R affinity. Combining the two variations, as in 8, was generally negative for both affinity and potency for α 1 and 5-HT 1A receptor systems. The exception was the increase in efficacy at 5-HT 1A R for both diastereomers, which was accompanied by a decrease in potency. A similar trend was observed with the desmethoxy derivatives 5c,t and 7c,t. As far as the stereochemistry is concerned, no clear difference emerged for the two pairs of diastereomers.
The study of the expansion of the dioxolane ring was achieved in two ways: (i) by the insertion of a methylene unit adjacent to the oxygen atom in position 1 of the 1,3-dioxolane to give the asymmetric 1,3-dioxanes 9,10 ( Compound 10 showed a significant enhancement of the affinity at α 1b and α 1d receptors, with an α 1d /α 1a selectivity ratio of about 12-fold higher than that of compound 2. The affinity at 5-HT 1A R M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT 13 was practically unchanged, with a consequential loss of 5-HT 1A /α 1 selectivity. During the functional studies, compound 10 showed an increase in α 1D selectivity (31-fold) and, at 5-HT 1A R, a significant increase in agonist potency (about 10-fold) and efficacy (doubled).
With respect to 2, compound 12 showed a decrease, although limited, in affinity at both receptor systems, the exception being the affinity at the α 1b receptor subtype. Also the antagonist potencies at the three α 1 receptor subtypes were decreased. At 5-HT 1A R the agonist potency was enhanced by about 72-fold, while the efficacy was three times the one observed with compound 2. These results are in agreement with the antagonist potency trend at the α 1 subtypes. The agonist potency at 5-HT 1A R was retained, while the efficacy was increased by about 3-fold.
Molecular simplification of 14, by removing the hydroxymethyl moiety, to give 16, interestingly gave an increase in binding affinity and antagonist potency at the α 1 receptor subtype, the exception being the potency at the α 1b subtype. At 5-HT 1A R, the affinity and potency remained unchanged, with a small variation in efficacy.
Compounds 9, 11, 13 and 15 were synthesized in order to confirm the effects on the activity of the previously described ortho-methoxy group (compound 2 vs 1). The methoxy group improved, with some exceptions, the pharmacological parameters at both receptor systems. In particular, the most significant variation is the potency of compound 12, which shows a pD2 of 9.22, 2240-fold higher than the desmethoxy derivative 11, showing a pD2 of 5.87.
Overall, the above described structural modifications allowed the identification of compound 12, which is the most interesting in the series, due to its high potency at 5-HT 1A R. In direct comparison with the starting point 2, compound 12 clearly showed enhancement of the pharmacological profile M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT 14 at 5-HT 1A : the selectivity ratio 5-HT 1A /α 1 was maintained and, despite a limited reduction in affinity, an increase of about two orders of magnitude in agonist potency and three in efficacy was observed. These results allowed us to consider 12 as one of the most potent 5-HT 1A R full agonists.  Table 2, showed a dosedependent cytotoxicity for the above-mentioned compounds at the tested concentrations.  [20]. These neurotoxins were used at a concentration equal to their IC 50 . As reported in Table 3, compound 12, at 1 µΜ, showed neuroprotective activity against H 2 0 2 and oligomycin A damage, while a minimal effect was observed against rotenone. According to [18] The data are expressed as percentages of neuroprotection ±SD of three independent experiments.

M A N U S C R I P T
A C C E P T E D ACCEPTED MANUSCRIPT 16 permeability for the control (Fluorescein isothiocyanate-dextran, FD-4) was within the expected values.

Pharmacokinetic studies
Before proceeding to the in vivo studies, preliminary pharmacokinetic analysis was performed in rats. Compound 12 was administered per os at a dose of 10 mg/Kg and brain and plasma concentrations were quantified following the previously reported and validated bioanalytical method [23]. The concentration vs. time curves are reported in Figure 3 and the pharmacokinetic parameters for brain and plasma are summarized in Table 5. Figure 3. Cerebral (blue circles) and plasmatic (red squares) concentrations (nmol/g and nmol/mL, respectively) of 12 after an oral dose of 10 mg/kg. Table 5. Pharmacokinetic parameters in rat brain and plasma after an oral dose of 12 (10mg/kg), calculated using "PK Solutions" software. [ 6.2 (15-45 min) 142 (45-480 min) [a] concentrations are expressed as nmol/mL in the plasma and in nmol/g in the brain.
Comparing brain and plasma curves, it is possible to observe a different profile of 12 in the two compartments. The areas under the brain or plasma concentration vs. time (AUC0-t) were 7167 and 17 nmol/g·min, respectively. This difference is reflected in the high brain/plasma ratio (B/P), which was calculated as the ratio between the brain and the plasma AUC(0-t). Compound 12 showed a B/P ratio value of 420, demonstrating its elevate capability to permeate the blood-brain barrier (BBB).
These data were supported by in silico BBB-passage prediction for non-active transport. Predicted brain/blood partition coefficient (QPlogBB) and predicted apparent MDCK cell permeability (QPPMDCK) were calculated with QiKProp [25]. Compound 12 showed a QPlogBB of 0.467 (for CNS penetration -3< QPlogBB <1.2) and a QPPMDCK of 1250 (for CNS penetration QPPMDCK >500), suggesting that is able to cross the BBB by passive diffusion. Moreover, the predicted QPPMDCK is in accordance with the bi-directional transport studies on MDCK-MDR1 monolayers.
As shown in Table 5, the concentration of 12 in the rat brain increased rapidly, reaching a maximum of 157.9 ± 8.7 nmol/g (C max ) in the first 60 minutes (t max ). The effect of 12 in the behavioral studies was evaluated at this time point. The initial increased concentrations of 12 in the brain occurred during stable plasma levels, as observed for (R)-8-OH-DPAT, dipropylaminotetraline (DPAT) derivatives and (S)-UH-301. The concentration then decreased, following a biphasic trend: rapidly during the first phase (60-120 minutes) and more slowly during the second phase (120-480 min) with two half-lives (t ½ ) of 6.16 and 142 min, respectively. This trend follows the one reported for (R)-8-OH-DPAT [26]. The rapid decrease of 12 in the brain does not seems to be linked to a drug efflux transport by P-gp, as suggested by the in vitro transport studies.
In the plasma, the concentration of 12 was lower than in the brain, with a C max of 0.095 nmol/mL during the first 30 minutes, but with a monophasic and slow elimination rate, resulting in a t 1/2 of 102.6 min, 3-times higher than the reference drug 8-OH-DPAT (t 1/2 = 27 min) [26].
According to the high bio-distribution of 12 in brain, this compound was tested in vivo for its activity on the central nervous system.

In vivo behavioural studies
Compound 12 was assessed in adult male Sprague-Dawley rats for anxiolytic, locomotor and antidepressant activity.

Anxiolytic effect
The anxiolytic effect of 12 was evaluated in rats using the Elevated Plus Maze test (EPM) [27].  Dunnett's test).

Locomotor activity and anxiolytic effect
The potentiality of compound 12 to possess excitatory activity was evaluated using the Open Field test [27]. The test measures the total distance tracked by the rats as an index of the locomotory activity following CNS excitation. Compound 12 was administered per os at 10 mg/Kg. 8-OH-DPAT (at 0.5 mg/Kg) was used as a reference. As reported in Figure 5A, the rats administered both

Anti-depressant activity
To assay the anti-depressant activity of compound 12, the Forced Swim test (Porsolt) was used in rats [28][29][30]. The administration of 12 at the doses of 10 and 20 mg/kg was able to significantly reduce the time that the rats spent immobile and to increase the time spent swimming, with an effect comparable to that of 8-OH-DPAT ( Figures 6A and B, respectively). In contrast, the administration of 12 at all doses did not significantly influence the time spent climbing ( Figure 6C). The same results were obtained using the reference drug 8-OH-DPAT. Taken together, these data are in accordance with the effect of 5-HT 1A R agonists, whereas catecholaminergic agents cause a decrease in the time spent immobile, together with an increase in the time spent climbing [31,32]. This behaviour indicates that the anti-depressant action of 12 is strictly due to its interaction with the serotonergic system. These data are in agreement with the higher affinity of 12 for the 5-HT 1A R rather than for the α-adrenergic receptor (5-HT1a/α1D selectivity = 16), as shown by the binding data.

Anti-nociceptive activity
The formalin test was used to assess the potential analgesic activity of compound 12 in vivo. [33] Indeed, at the dose of 10mg/kg, i.p., 12 was able to significantly decrease perception of the II phase

M A N U S C R I P T
A C C E P T E D ACCEPTED MANUSCRIPT 29

General procedure for the synthesis of amines 17-30
To a solution of 2-phenoxy-ethylamine (5 eq.) or 2-(2-methoxyphenoxy-)ethylamine (5 eq.) in 2methoxyethanol (25 mL per mmol of amine) the appropriate aliphatic chloride 33-39 (1 eq.) and KI (cat.) was added. The mixture was refluxed for 18-48 h and concentrated. The residue was suspended in CHCl 3 and washed with 1M NaOH, brine, dried over anhydrous Na 2 SO 4 and concentrated. The crude was purified by flash chromatography to give the titled compound.

Radioligand Binding Assay
Binding assays for recombinant human α 1 adrenoceptors and 5-HT 1A R were performed following published procedures [18].

Functional studies
Functional studies on isolated tissue (vas deferens prostatic portion, spleen and aorta) were used to assess antagonism toward α 1A, α 1B and α 1D adrenoceptors subtypes, respectively, as already reported [34]. Compound potency and efficacy were measured by [ 35 S]GTPγS binding in cells expressing recombinant human 5-HT 1A R, according to Stanton and Beer [35], with minor modifications [18].

Data Analysis
During the functional studies on isolated tissue, the concentration-response curves were analyzed as described earlier [13]. The [ 35 S]GTPγS binding data were analysed using GraphPad as reported [18].

Cytotoxicity Assays
The human neuroblastoma cell line SH-SY5Y was used for assessing the cytotoxicity of the compounds, as previously described [36]. The cells were grown in a DMEM medium (EuroClone), supplemented with 10% heat inactivated FBS, 2 mM L-glutamine, 100 U/mL penicillin and 100 M A N U S C R I P T A C C E P T E D

Bi-directional Transport Studies
To evaluate the ability of compound 12 to permeate the blood brain barrier (BBB), MDCKII-MDR1 cells were employed as an in vitro model of BBB, following the previously described protocol [37,38]. Diazepam and FD4 (fluorescein isothiocyanate-dextran) were used as internal controls for the transcellular and paracellular pathways, respectively. The apparent permeabilities (P app AP and BL in cm/sec) and the efflux ratio (ER) were calculated according to the equations described in Franchini et al. [18].

In vivo study Animals
For the pharmacokinetic and behavioral studies, the experiments were performed on male Sprague The rats and mice were housed two and six per cage, respectively, in a temperature-controlled (22°C ± 1°C) colony room under a 12/12h light-dark schedule. Food and water were freely available. All animals were handled daily for a week before behavioral testing. Experimental procedures were approved by the Local Ethical Committee (IACUC) and conducted in accordance with international guidelines as well as European Communities Council Directive and National Regulations (CEE Council 86/609 and DL 116/92). All the tests were performed blind to treatment.

Pharmacokinetic studies
Twenty-five rats were treated orally with 10 mg/Kg of 12, solubilized in 5% Tween 80 in distilled water, and administered per os by gastric tube. Five rats, used as control animals, received an equivalent volume of the above mentioned solvent. Five animals were sacrificed by decapitation at the following time points : 0, 30, 60, 120, 240 and 480 minutes after treatment. Compound 12 was quantified in rat brain and plasma according to Franchini et al. [23] .

Behavioral studies
Twenty-five rats were divided equally into five groups and treated as follows: (1) Group 1: 0.

Open field test
The open field test was carried out on rats to evaluate the exploratory activity and emotional response of the animals, as previously described by Carnevale et al. [40]. Briefly, the apparatus consisted of a black-painted wooden arena (100 cm × 100 cm) with 50 cm high walls, placed in a dimly lit soundproof room. The arena was sub-divided into two areas: the central area corresponding to 25% of the total area, and the peripheral one, corresponding to the remaining area.
At the beginning of the test, each rat was placed in the center of the arena and its activity was recorded for 10 min using a video tracking system (SMART 2.5 version, PanLab, Barcellona, Spain). The activities in the central zone including the percentage of time spent in the central zone and the number of entries were measured automatically.

Forced Swim Test (Porsolt)
The Forced Swim test was used to assess the anti-depressant activity of the compound [28]. The rats were placed into a glass cylinder (21 cm diameter) filled with water (23-25 ± 1°C; 30 cm depth) for M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT home cage. A 5 min swim test was conducted 24 h after the 15 min session. This test was videotaped and scored for the duration of climbing, swimming, and immobility behavior.

Anti-nociceptive activity
For the assessment of anti-nociceptive activity, the mice were subjected to the formalin test, in accordance with a previous publication [18].

Statistical Analysis
The data obtained from the tests which are reported in the tables and graphs in this study are the mean ± standard error (SEM) obtained from groups of 5 animals each. The statistical analysis was performed using the ANOVA test followed by the post-hoc Dunnett's test using the program GraphPad Prism version 5 for Windows (GraphPad Software, San Diego, California, USA). In all cases, p<0.05 was considered as a minimum level of significance.

Conflicts of interest
The authors state no conflict of interest.

Highlights
• A new series of 1,3-dioxolane analogues were prepared and tested in vitro for binding affinity, potency, efficacy at 5-HT 1A R and α 1 adrenoceptors.
• Compound 12 emerged as a potent and selective 5-HT 1A R agonist with an high biodistribution in the brain compartment as assessed by pharmacokinetic data in rats.
• Compound 12 exhibited anxiolytic (Elevated Plus Maze and Open Field test) and antidepressant (Forced Swim test) effect.
• Compound 12 showed anti-nociceptive activity in the formalin test.