Design, synthesis, and in vitro and in vivo characterization of new memantine analogs for Alzheimer’s Disease

Abstract

Currently, of the few accessible symptomatic therapies for Alzheimer’s disease (AD), memantine is the only N-methyl-D-aspartate receptor (NMDAR) blocker approved by the FDA. This work further explores a series of memantine analogs featuring a benzohomoadamantane scaffold. Most of the newly synthesized compounds block NMDARs in the micromolar range, but with lower potency than previously reported hit IIc, results that were supported by molecular dynamics simulations. Subsequently, electrophysiological studies with the more potent compounds allowed classification of IIc, a low micromolar, uncompetitive, voltage-dependent, NMDAR blocker, as a memantine-like compound. The excellent in vitro DMPK properties of IIc made it a promising candidate for in vivo studies in Caenorhabditis elegans (C. elegans) and in the 5XFAD mouse model of AD. Administration of IIc or memantine improved locomotion and rescues chemotaxis behavior in C. elegans. Furthermore, both compounds enhanced working memory in 5XFAD mice and modified NMDAR and CREB signaling, which may prevent synaptic dysfunction and modulate neurodegenerative progression.

Graphical Abstract

1. Introduction

Alzheimer’s disease (AD) is a major and rising public health concern due to not only its increasing global prevalence, but also the lack of effective treatment options.^1–3 To date, only four drugs have approval for treating AD. Unfortunately, these drugs only improve cognitive and behavioral symptoms without modifying the course of the disease. Despite intensive efforts to develop new AD pharmacotherapeutics,⁴ progress has been slow and difficult. For example, a new drug, aducanumab,⁵ has recently been approved by the Food and Drug Administration (FDA). However, its approval is mired in controversy due to risk of severe side effects and conflicting reports of efficacy.^6–9

The first drugs developed for AD, the acetylcholinesterase inhibitors (AChEIs), aim to revert the cholinergic deficit responsible for the cognitive impairment of patients.¹⁰ In addition to currently available AChEIs (donepezil, rivastigmine and galantamine), multiple AChEIs have been developed over the last few years. However, few novel AChEIs have reached clinical trials and none have arrived at the market.^11–16

Another drug approved by the FDA is memantine, a N-methyl-D-aspartate receptors (NMDAR) uncompetitive antagonist.^17–18 Memantine acts as an open channel blocker that is thought to prevent NMDAR overactivation and glutamate-mediated neurotoxicity in the pathogenesis of AD.^19–20 It has better therapeutic tolerability and favorable safety relative to other NMDAR channel blockers because of its ability to preferentially block extrasynaptic NMDARs, which have been associated with activation of cell death pathways.²¹ Memantine is characterized by faster blocking/unblocking kinetics, strong voltage dependency, and moderate affinity for NMDAR.²² Due to its excellent drug profile, numerous companies and academic groups have developed new analogs of memantine. Although only neramexane has reached clinical trials,²³ much effort is still ongoing for developing new NMDAR open channel blockers with pharmacological profiles similar to memantine for use as potential therapies for AD and other neurodegenerative diseases.^24–34 In addition, other NMDAR open channel blockers are also currently being developed for the treatment of tinnitus and as a new class of drugs with fast-onset antidepressant effects.^35–41

Recently, our research group has synthesized new memantine analogs based on a novel benzopolycyclic ring (Figure 1). First, compounds with the general structure I were synthesized, but they were found to be less potent than memantine (e.g., Ia and Ib).^42–43 Interestingly, the replacement of the oxygen atom at the C8 position of Ia (IC₅₀ = 98 ± 26 μM) by a methylene unit led to the seven-fold more potent amine IIa (IC₅₀ = 13.6 ± 3.4 μM). Thereafter, further variations at the C9 position in the general structure II generated compounds with promising potency, in the low-moderate micromolar range. Indeed, three of these derivatives displayed potencies very similar to (IIc and IIg) or even stronger than (IIb) memantine (Figure 1).⁴⁴

Figure 1.

Structures of memantine, previously studied polycycles Ia-b and IIa-g, and the new series III. IC₅₀ values for memantine and polycycles Ia-b and IIa-g are provided under each structure.

In our endeavor to find novel analogs of memantine, herein we report further structure-activity relationships (SAR) on the benzohomoadamantane scaffold, i.e., derivatives of general structure III, in order to select a proper candidate for additional in vitro pharmacokinetic studies, and subsequently perform an in vivo study in a 5XFAD murine model of AD.

2. Results and discussion

2.1. Design and synthesis of new NMDAR channel blockers

First, to explore alternative substituents at the C9 position, we synthesized 2 (Scheme 1). It is known that H/D bioisosteric replacement may lead to compounds with the same activity and selectivity.^45–47 Furthermore, considering that the C9 position is susceptible to being metabolized, the presence of deuterium may be a good strategy to avoid rapid metabolization and prolong the pharmacokinetic stability of the compound.⁴⁸ The approach followed for the synthesis of 2 was similar to that previously applied to the preparation of IIb.⁴⁴ Thus, the reaction of known alcohol 1⁴⁴ with thionyl bromide followed by radical dehalogenation of the unstable bromo intermediate, generated the amine 2 in moderate overall yield (Scheme 1). We have previously reported that the NMDAR antagonist activities of amantadine and its ring-contracted analogs are of the same order of magnitude.⁴⁹ Consequently, we also designed the ring-contracted analog 6, by removing a methylene group from IIb and bonding positions C7 and C9 (Scheme 1). The synthesis of amine 6 involved four steps: a) reaction of known alcohol 3⁵⁰ with mesyl chloride; b) reaction of the corresponding mesylate with NaI in the presence of H₃PO₃ to furnish iodide 4; c) radical photochemical carboxylation of 4 to carboxylic acid 5; and, d) Curtius rearrangement of 5 to generate primary amine 6.

Scheme 1. Synthesis of series III amines 2 (R₁ = D, X = CH₂) and 6 (R₁ = H, X = bond).^a.

^aReagent and conditions: (i) 1) SOBr₂, toluene, rt, 1.5 h; 2) Bu₃SnD, AIBN, anh toluene, 95 °C, 4 h, 40% overall yield. (ii) 1) Methanesulfonyl chloride, pyridine, 120 °C, 5 h; 2) H₃PO₃, NaI, 150 °C, 6 h, 76% overall yield. (iii) 1) bis(tributyltin), methyl oxalyl chloride, anh toluene, hν, 20 h; 2) MeOH, Et₃N; 3) KOH / MeOH 40%, 2 h; 4) H₂O, reflux, 3 h, 37% overall yield. (iv) DPPA, Et₃N, toluene, reflux, 3 h; 2) HCl 6 N, reflux, 24 h; 3) HCl / MeOH, 39% overall yield.

We have previously synthesized and described several derivatives of I and II, but, so far, we have not explored the impact of introducing substituents on the aromatic ring of the benzohomoadamantane scaffold. Although IIb (IC₅₀ = 0.70 ± 0.12 μM), IIc (IC₅₀ = 1.93 ± 0.21 μM) and 2 (IC₅₀ = 1.2 ± 0.1 μM) are more potent NMDAR antagonists than IIa (IC₅₀ = 13.6 ± 3.40 μM, see below), we first decided to screen the NMDAR blocking activity of benzene-substituted derivatives based on IIa, because this amine shows an excellent compromise between acceptable potency as NMDAR channel blocker and better synthetic availability. Therefore, we prepared a series of ten derivatives of general structure III (R₁ = Me, X = CH₂) which are analogs of IIa, by introducing a variety of electron-withdrawing and electron-donating group on the aromatic ring. The key introduction of the substituents in the aromatic ring was accomplished using two alternative procedures (Schemes 2–4).

Scheme 2. Synthesis of amines 9, 10, 12, 14 and 15 (compounds from series III with R₁ = Me and X = CH₂).^a.

^aReagent and conditions: (i) fuming HNO₃, Ac₂O, AcOH, 0 °C to rt, overnight, 85% yield. (ii) 1) thiourea, glacial AcOH, abs. ethanol, reflux, overnight; 2) HCl / Et₂O, 74% overall yield for 9 and 52% overall yield for 12. (iii) 1) H₂, Pd / C, MeOH, 1 atm, rt, 48 h; 2) HCl / MeOH, quant. overall yield. (iv) CH₃COCl, AlCl₃, DCM, rt, 1 h, 54% yield for 11 and 40% yield for 14 as hydrochloride salt obtained with HCl / Et₂O. (v) 1) formalin 37%, NaBH₃CN, glacial AcOH, MeOH, rt, overnight; 2) HCl / Et₂O, 74% overall yield.

Scheme 4. Synthesis of amines 37–40 (compounds from series III with R₁ = Me and X = CH₂).^a.

^aReagent and conditions: (i) 1) dimethyl-1,3-acetonedicarboxylate, Et₂NH, MeOH, reflux, 1.5 h; 2) glacial AcOH, conc. HCl, reflux, 12 h; 3) Toluene, reflux, 13% overall yield for 25, 30% for 26, 51% for 27, 65% for 28. (ii) Ph₃PCH₃I, NaH, anh DMSO, 90 °C, overnight, 60% yield for 29, 23% for 30, 69% for 31, 38% for 32. (iii) ClCH₂CN, conc. H₂SO₄, AcOH, 0 °C to rt, overnight, 48% yield for 33, 75% yield for 34, 68% yield for 35, 57% for 36. (iv) 1) thiourea, glacial AcOH, abs. ethanol, reflux, overnight; 2) HCl / Et₂O, 36% overall yield for 37, 81% for 38, 84% for 39, 81% for 40.

In the first approach, electrophilic aromatic substitution reactions on known chloroacetamide 7⁴³ or acetamide 16⁵¹ were carried out for the synthesis of 9, 10, 12, 14, 15 and 20. Thus, from the known chloroacetamide 7, following a classical approach that involved aromatic nitration and cleavage of the chloroacetyl group, amine 9 was obtained. Catalytic hydrogenation of 9 furnished aniline 10. Additionally, Friedel-Craft acylation of 7, followed by deprotection afforded amine 12 in good yield. A Friedel-Craft reaction of the known 1343 led to amine 14 and a reductive alkylation of 9 provided amine 15 (Scheme 2).

A first attempt to obtain amine 20, through catalytic hydrogenation of 8 followed by a classical Sandmeyer reaction and removal of the chloroacetamide group, met with failure. Thus, 20 was prepared from the known acetamide 16⁵¹ through a four-step process that involved aromatic nitration, catalytic hydrogenation, Sandmeyer reaction, and hydrolysis (Scheme 3).

Scheme 3. Synthesis of amine 20 (compound from series III with R₁ = Me and X = CH₂).^a.

^aReagent and conditions: (i) fuming HNO₃, Ac₂O, AcOH, 0 °C to rt, overnight, 60% yield. (ii) H₂, PtO₂, EtOH, 1 atm, rt, 4 h, 68% yield. (iii) 1) NaNO₂ / conc. HCl, <5 °C; 2) CuCl / conc. HCl, 60 °C, 90 min, 21% overall yield. (iv) 1) conc. HCl / 2-propanol, 6 days, reflux; 2) HCl / Et₂O, 19% overall yield.

A different synthetic strategy was undertaken for the synthesis of the aromatic substituted compounds 37–40, which involved the construction of the benzohomoadamantane scaffold starting from a suitably mono- or disubstituted o-phthaldialdehyde (Scheme 4). Thus, once the substituted o-phthaldialdehydes 21,⁵² 22,⁵³ 23,⁵⁴ and 24⁵⁵ were obtained via procedures previously reported, a synthetic route that included a Weiss-Cook condensation,⁵⁶ hydrolysis, decarboxylation, and dehydration led to the corresponding diketones 25–28. Subsequently, the treatment of these ketones under Wittig olefination and Ritter-type transannular cyclization conditions furnished the chloroacetamides 33–36 that after deprotection provided the desired amines 37–40.

In order to determine the effects of different aromatic substituents on NMDAR block, the potency of the new benzo-substituted amines 9–10, 12, 14–15, 20 and 37–40 was tested (see below) and compared with that of the unsubstituted IIa (Table 1). Briefly, the introduction of substituents in the aromatic ring resulted in a decrease in the potency as NMDAR blockers for the majority of compounds regardless of the position, size or electronic character of the substituent (9, IC₅₀ = 164 μM; 10, IC₅₀ = not active; 12, IC₅₀ = 323 μM; 14, IC₅₀ = 566 μM; 15, IC₅₀ = 309 μM; 20, IC₅₀ = 34.7 μM; 37, IC₅₀ = 30.9 μM; 38, IC₅₀ = 517 μM; 39, IC₅₀ = 13.2 μM; 40, IC₅₀ = 77.4 μM) (see below and Table 1). An exception to this trend is amine 39, with a fluorine atom at the C1 position, the potency of which (IC₅₀ = 13.2 μM) was similar to that of IIa (IC₅₀ = 13.6 μM). Thus, once this screening was completed with derivatives of IIa, and considering the electrophysiological studies carried out with IIb, IIc and 2 (see below), amine 44 was envisaged as a promising compound. 44 contains two fluorine atoms at C1 and C9 and was synthesized starting from the diketone 27, through a four-step sequence that involved a selective mono-Wittig olefination, followed by a Ritter-type transannular cyclization, substitution of the hydroxyl group by a fluorine atom and final removal of the chloroacetamide group (Scheme 5).

Table 1.

IC₅₀ values (μM) of the new series III of benzohomoadamantane derivatives as NMDAR channel blockers and of references IIa-c and memantine.^a

Comp.	IC50 (μM)
2	1.16 ± 0.07
6	111 ± 18.8
9	164 ± 4.48
10	NAb
12	323 ± 89.2
14	566 ± 186
15	309 ± 120
20	34.7 ± 5.98
37	30.9 ± 4.70
38	517 ± 54.7
39	13.2 ± 1.07
40	77.4 ± 13.5
44	29.0 ± 11.9
IIa c	13.6 ± 3.40
IIb d	0.70 ± 0.12
IIc d	1.93 ± 0.21
Memantine	1.5 ± 0.1

^aFunctional data were obtained from primary cultures of cerebellar granule neurons using the method described in the experimental section by measuring the intracellular calcium concentration. Cells were challenged with NMDA as indicated. Data shown are means ± SEM of at least three separate experiments carried out on three different batches of cultured cells.

^bNA, non-active.

^cSee ref. 43.

^dSee ref. 44.

Scheme 5. Synthesis of amine 44 (compound from series III with R₁ = F and X = CH₂).^a.

^aReagent and conditions: (i) Ph₃PCH₃I, NaH, anh DMSO, 90 °C, overnight, 46% yield. (ii) ClCH₂CN, conc. H₂SO₄, DCM, 0 °C to rt, overnight, 32% yield. (iii) diethylaminosulfur trifluoride (DAST), DCM, −30 °C to rt, overnight, 69% yield. (iv) 1) thiourea, glacial AcOH, abs. ethanol, reflux, overnight; 2) HCl / Et₂O, 79% yield.

The new amines, 2, 6, 9–10, 12, 14–15, 20, 37–40, and 44, were fully characterized as hydrochlorides through their spectroscopic data and elemental analyses or HPLC/UV (see Experimental Section and supporting information for further details).

2.2. NMDAR blockers activities and SAR

To evaluate whether the synthesized compounds were able to block NMDARs, we measured their effect on increases in intracellular calcium evoked by 100 μM NMDA (in the presence of 10 μM of glycine) on rat cultured cerebellar granule neurons using a previously reported Fura 2 assay (Table 1).⁵⁷

As expected, the substitution of the hydrogen atom at C9 in IIb by a deuterium atom did not affect the biological potency of 2 (IC₅₀ = 0.70 ± 0.12 μM for IIb vs IC₅₀ = 1.16 ± 0.07 μM for 2). However, the ring contraction from IIb to 6 resulted in a dramatic drop in potency (IC₅₀ = 0.70 ± 0.12 μM for IIb vs IC₅₀ = 111 ± 18.8 μM for 6).

Regarding the aromatic substitution, it can be concluded that the functionalization of the aromatic ring is highly deleterious for the compound’s NMDAR channel blocking potency (Table 1), regardless of the electron donor or acceptor character of the substituent (compare IIa vs 9, 10, 12, 20, 37–40). However, size and substitution position exerted a certain influence on potency. Small size substituents such as chlorine (20) and fluorine (37) at C2 were less deleterious for potency than bigger substituents such as a nitro (9) or an acetyl (12) group. More importantly, the introduction of a fluorine atom at C1, as in 39 (IC₅₀ = 13.2 ± 1.07 μM) does not alter the IC₅₀ value in respect to the unsubstituted IIa (IC₅₀ = 13.6 ± 3.4 μM), unlike 37 (IC₅₀ = 30.9 ± 4.70 μM) which possesses the fluorine atom at C2. Overall, within this series of new amines featuring a methyl group at C9, 39, with its fluorine atom at C1, emerged as the most interesting derivative. Due to the effect of the C1 fluorine atom and our finding that the replacement of IIa’s methyl group at C9 with a fluorine atom, as in IIc, resulted in a remarkable increase of potency, we synthesized amine 44. Unexpectedly, 44 (IC₅₀ = 29.0 ± 11.9 μM) was a less potent NMDAR channel blocker than IIc (IC₅₀ = 1.93 ± 0.21 μM).

Finally, within this new series of benzohomoadamantane derivatives, tertiary amines were clearly less potent than the corresponding primary amines (9 vs 15 and 12 vs 14), in contrast to the trend observed in reference compounds IIa and IIb.

Overall, with the sole exception of the deuterated amine 2 (IC₅₀ = 1.16 ± 0.07 μM), all the amines herein reported were less potent as NMDAR channel blockers than the previously described amines IIb (IC₅₀ = 0.70 ± 0.12 μM) and IIc (IC₅₀ = 1.93 ± 0.21 μM). Particularly striking is the marked decrease in potency when going from IIb to its ring-contracted analog 6 (IC₅₀ = 111 ± 18.8 μM).

2.3. Molecular dynamics (MD) simulations

Electrophysiological studies (see discussion below in electrophysiology section) confirmed NMDAR channel blocking ability for compounds investigated in this study. However, to better understand molecular mechanisms of how these channel blockers occlude ion permeation and to rationalize the differential binding affinities of these benzohomoadamantane derivatives caused by ring contraction, the exchange of substituent at the C9 position, or introduction of different groups on the aromatic ring, we used molecular dynamic (MD) simulations. We have developed a computational structural model of the NMDAR (GluN1/2A) transmembrane domain with a closed gate. We selected four compounds, IIa, IIb, IIc and 6, to generate initial docking-based protein-ligand complex structures for NMDAR-IIa, NMDAR-IIb, NMDAR-IIc and NMDAR-6. The complexes then were embedded in a POPC lipid bilayer and solvated in water and 0.15 M NaCl solution. 200 ns molecular dynamics simulations were carried out for each of the structures to allow for equilibration of the ligand in the channel in native-like conditions. From these simulations we obtained atomic level insights on binding modes and their stabilities to analyze how their structural and dynamic features correlate with the reported difference in the binding affinities of these compounds.

During MD simulations, all compounds showed stable binding in a “flipped down” conformation (Figure 2) with the NH₃⁺ groups facing the intracellular side of the membrane. In this flipped down conformation the NH₃⁺ groups formed H-bond interactions with the asparagine (ASN) clusters at the tips of the M2 helices of GluN1 (N616) and GluN2A (N614) subunits. The hydrophobic aromatic core of the compounds stabilized between the protein channel sub-pockets created by the hydrophobic side chains of L642 and A643 of GluN2A, and V644, A645 and M641 of GluN1 inside the channel (Figures 2A–C). Similar flipped down binding modes were reported for memantine and MK-801 in the GluN1/2B NMDAR channel.⁵⁸

Figure 2.

Representative conformational orientation of IIa (A), IIb (B), IIc (C), and 6 (D). A. IIa shown with hydrophobic channel residues and ASN cluster residues also the methyl of IIa showed a minor clash (dotted circle) with threonine side chains from the cavity that may decrease its potency. B. IIb shows similar interactions as IIa but no clash with threonine. C. IIc shows similar interactions as IIa D. 6 is shown surrounded by water molecules that may prevent it from interacting with the receptor biding pocket in the NMDA channel during MD simulations. In all panels GluN1 subunit is depicted as blue ribbons; GluN2A subunit is depicted as red ribbons; side chains of channel residues within 4Å of compounds are shown in stick representations.

For all compounds that stabilized in flipped down conformations, we observed mobility of the blocker’s hydrophobic core in the plane of its phenyl ring. This mobility allowed for different binding pose occupancies within the sub-pockets created between M3 helices of GluN1 and GluN2A (Figure 3). Similar flexible orientations were reported in MD simulations of MK-801 in the GluN1/2B NMDAR channel.⁵⁸ Orientationally different poses are indeed stable in flipped down binding conformations (Figure S1) and could easily block ion conduction by physical occlusion of the permeation pathway.

Figure 3.

Top view of the hydrophobic core orientation flexibility for IIa, IIb, IIc and 6 in XY plane with respect to phenyl ring inside the channel. Center of mass occupancies of phenyl rings are shown as purple balls.

Comparing MD simulation results provided further insights into binding differences among IIa, IIb and IIc. For example, very close contact of the methyl at C9 in IIa (Figure 2A) with the side chain hydroxyl groups of the threonines (T648 – GluN1 and T647 – GluN2A) inside the channel cavity may cause an energetic penalty on binding, which could explain the reduced binding affinity of IIa in comparison to IIb (Figure 2B). In contrast, due to the smaller size of the fluorine at this position in IIc, there is no steric clash with threonine side chains, which may explain the more similar potencies of IIb and IIc (Figure 2C). These observations also suggest that any substitution larger than hydrogen or fluorine will result in a steric clash with the hydrophobic wall and would require a wider channel opening to accommodate such a compound. Furthermore, the simulations showed that heavy atoms of the protein side chains are within 4Å of the phenyl ring (Figure 2). This suggests that any substitution on the aromatic ring will negatively affect binding since it will clash with the hydrophobic wall. Indeed, consistent with simulations, only unsubstituted derivatives of the phenyl ring (IIa-g and 2) and their fluorinated analogs (37, 39 and 44) showed acceptable potency (Figure 1 and Table 1).

Finally, MD simulations performed with 6 showed a flipped down binding mode similar to IIa, IIb and IIc where the NH₃⁺ of 6 maintains H-bonding interaction to ASN cluster through the majority of the MD simulation time (Figure 2D). However, we observed a slight instability of the flipped down conformation. This could be explained by the presence of water molecules in the channel cavity between the hydrophobic core of compound 6 and the hydrophobic residues of the channel, which is likely due to the smaller size of the ring-contracted 6 (Figure 2D). Considering that hydrophobic residues in the NMDAR channel cavity play an important role in the binding of channel blockers,⁵⁹ the loss of the interaction between 6 and hydrophobic channel residues provides an explanation for the reduction of the potency of this amine.

2.4. Functional block of NMDAR by benzohomoadamantane derivatives IIb, IIc, and 2 compared with memantine

We next selected compounds with similar potency to memantine (IIb, IIc and 2) and measured functional block of NMDAR. To measure the properties of the newly synthesized amines, we performed whole-cell patch-clamp recordings from tsA201 cells transfected with expression plasmids codifying for rat GluN1 and GluN2A subunits.^28,60

Figure 4A shows representative current traces obtained using a protocol that allowed us to quickly screen NMDAR inhibition by the compounds of interest at two membrane potentials. While clamping the cell at −60 mV, NMDAR currents were elicited with 100 μM NMDA plus 10 μM glycine. After NMDA-mediated currents had reached a steady state, a blocking compound (either memantine, IIb, IIc or 2) was applied for 40 seconds in the presence of agonist. During blocker application, the membrane voltage was jumped to +60 mV for 5 seconds and returned to −60 mV. Blocker then was removed to allow blocker unbinding and recovery of the current from inhibition, during which another 5 second jump to +60 mV was performed before returning to −60 mV and subsequent removal of agonists.

Figure 4. Block, unbinding, voltage dependence, and blocking and unblocking kinetics of memantine (Mem), IIb, IIc and 2.

A. Examples of whole-cell currents evoked by bath application of 100 μM NMDA plus 10 μM glycine onto tsA201 cells expressing GluN1/2A NMDARs. Compounds were rapidly applied with a piezoelectric device at 10 μM in the presence of the agonists. B. Blocking percentage at the holding potential of −60 mV. Numbers inside bars denote the number of experiments. C. Blocking kinetics expressed in time constant (s). Numbers inside bars denote the number of experiments. D. Percentage of block at +60 mV (n = 10, 8, 9 and 9 for memantine, IIb, IIc and 2, respectively. E. Degree of unbinding, measured as the percentage of current recovery after removal of the blocker. Numbers inside bars denote the number of experiments. F. Unblocking kinetics expressed in time constant (s). Numbers inside bars denote the number of experiments. Asterisks identify mean values with statistically significant differences from the value for memantine: ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05.

Although memantine, IIb, and 2 show a similar degree of NMDAR inhibition, inhibition by IIc was noticeably weaker (69.1 ± 1.2 % for IIc vs. 86.2 ± 1.4 % for memantine p < 0.0001; one-way ANOVA with Bonferroni’s multiple comparison test; Figure 4B). While no differences were observed in percentage of block for 2 in comparison to memantine (89.3 ± 3.1 %; p > 0.05), IIb appeared to block slightly more efficiently than memantine (94.1 ± 1.4 %; p < 0.05). Regarding blocking kinetics (measured as the time needed to reach approximately 2/3 of the maximal block; τ_weighted (weighted time constant) in seconds; see experimental section), all compounds tested presented values similar to memantine (around 1 second), although development of block by IIc was faintly slower (Figure 4C).

Inhibition by all known NMDAR channel blockers depends on membrane voltage, with potency increasing as membrane voltage is hyperpolarized. We assessed the voltage dependence of the compounds tested by applying two positive voltage steps to +60 mV in the presence of the blocker or after blocker washout (Equation 1). We observed that IIc showed strong voltage dependent block similar to that of memantine (Figure 4D). IIb and 2 showed weaker apparent voltage dependence of block: % block at +60 mV was 26.9 ± 4.7 % for IIb and 20.6 ± 5.9 % for 2 compared to −3.9 ± 2.5 % for memantine (p < 0.001 for both compounds vs. memantine). It is likely that these results reveal that IIb and 2 exhibit both weaker voltage dependence, and slower unblocking kinetics (Figure 4F), than memantine or IIc.

Importantly, a potential memantine-like compound should have similar unbinding kinetics. Therefore, we calculated the percentage of current recovery 40 seconds after removal of the blocker for the selected compounds and we compared it with that of memantine (93.6 ± 3.3% recovery of current). Note that, because of the interposed depolarization, current recovery will depend on the drugs’ unblocking kinetics both at −60 and at +60 mV. IIb showed less current recovery than memantine (69.0 ± 3.6 % recovery; p < 0.0001 vs. memantine), whereas current recovery for IIc and 2 was similar to memantine (Figure 4E). The lower current recovery after 40 seconds observed for IIb suggested that its unbinding kinetics are slower than memantine’s, a conclusion supported by the observed unblocking kinetics (quantified as τ_weighted; see experimental section; τ_weighted = 18.1 ± 4.2 s for IIb vs. 6.4 ± 1.4 s for memantine; p > 0.001; Figure 4F).

Since our main aim was to select the best memantine analog for an in vivo study, IIb and 2 were discarded as candidates due to their relatively weak voltage dependence, despite their high potency. Although IIc did not block to the same degree as memantine, it showed a similar electrophysiological profile and IC₅₀ value, making it a suitable candidate for additional in vitro experiments and a subsequent in vivo study.

2.5. Characteristics of NMDAR inhibition by IIc

We next further quantified the potency and voltage dependence of NMDAR inhibition by IIc using whole-cell patch-clamp recordings from tsA201 cells expressing GluN1/2A receptors. Inhibition by IIc was measured as a function of drug concentration (Figure 5A) and used to calculate the IIc IC₅₀ in cells held at −65 mV. The IC₅₀ of IIc was found to be 4.40 ± 0.15 μM. (Figures 5B, C), well within the range of therapeutically beneficial NMDAR antagonists (e.g. ketamine IC₅₀ ~ 1 μM; memantine IC₅₀ ~ 1 – 2 μM; amantadine IC₅₀ ~ 40 – 75 μM).^28,61–68 This IC₅₀ value slightly differs from the IIc IC₅₀ measured using intracellular Ca²⁺ measurements from cerebellar granule neurons (1.93 ± 0.21 μM; Table 1), consistent with previous similar comparisons of NMDAR channel blocker IC₅₀s.²⁸ The discrepancy may result from differences in recording technique and/or from expression of GluN2 subunits other than GluN2A by cerebellar granule neurons.^69–71

Figure 5. Characteristics of GluN1/2A receptor channel block by IIc. A.

Representative current trace from one cell depicting inhibition of GluN1/2A receptors by IIc at −65 mV. Current evoked by application of 1 mM glutamate (Glu; black bars) was reduced as concentration of IIc (red bars) increased. B. Concentration-inhibition relation for IIc in 0 Mg²⁺ (black line, filled symbols) and in 0.2 mM Mg²⁺ (purple line, open symbols). Symbols represent means, error bar are smaller than symbols. Lines show best fits of Equation 2 to data. C. Comparison of IC₅₀ values in 0 and 0.2 mM Mg²⁺. 0.2 mM Mg²⁺ greatly reduces IIc potency (p < 0.0001, two-sample Student t-test, n = 10 and 6, respectively). Data shown as individual values, line and error bars depict mean ± SEM. D. Representative current traces from one cell depicting the voltage dependence of inhibition by 10 μM IIc. For clarity, traces from only five of the tested membrane potentials are displayed. E. Current-voltage relation for GluN1/2A receptors in 1 mM Glu (filled symbols) and 1 mM Glu + 10 μM IIc (open symbols). Current at each voltage (I_X) was normalized to current at 55 mV (I₅₅ mV) for each cell. Symbols represent means, error bars represent SEM; some error bars are smaller than symbols. F. Mean current-voltage relation data replotted as fractional current in the presence of 10 μM IIc. Solid line shows best fit of Equation 3 to estimate V₀ (25.43 ± 0.54 mV; n = 7). V₀ was subsequently used to calculate δ (1.01 ± 0.02; n = 7) with Equation 4. Error bars in F are smaller than symbols. Colored symbols in E and F correspond to example traces in D; black symbols represent measurements at voltages not shown in D.

NMDAR channel blockers share overlapping binding sites in the NMDAR channel.^72–75 Thus, many organic NMDAR channel blockers (e.g. amantadine, memantine, and ketamine) show competitive binding with the endogenous blocker Mg²⁺.^{68,74,76–78} To ascertain whether IIc inhibits NMDARs via channel block, we tested the effect of extracellular Mg²⁺ on IIc potency. As predicted, the inclusion of 0.2 mM Mg²⁺ in the recording solution led to a substantial rightward shift in the IIc IC₅₀ curve and a significant increase in IIc IC₅₀ (IC₅₀ = 4.40 ± 0.15 μM in 0 Mg ²⁺ vs 21.41 ± 0.56 μM in 0.2 mM Mg²⁺; p < 0.0001, two-sample Student t-test, n = 10 and 6, respectively; Figures 5B, C). This roughly 5-fold decrease in IIc potency suggests a competitive interaction between Mg²⁺ and IIc in the NMDAR channel. We next determined the voltage dependence of inhibition by 10 μM (~2-fold IC₅₀ at −65 mV) IIc. As predicted of a positively charged channel blocker, inhibition by IIc was markedly weaker at depolarized potentials (Figures 5D–F). Equations 3 and 4 were used to quantify V_0, the change in voltage (in mV) that results in an e-fold change in the IC₅₀ of a drug, and δ, an estimate of the fraction of the total transmembrane voltage field felt by the blocker at its binding site.⁷⁹ Inhibition of GluN1/2A receptors by IIc is strongly voltage-dependent, as reflected by the drug’s small V₀ (25.43 ± 0.54 mV) and large δ (1.01 ± 0.02). The IIc V₀ and δ values are similar to previously reported values for monovalent organic channel blockers (e.g. V₀ ~ 26 – 30 mV, δ ~ 0.8 – 1.0 for memantine, ketamine, and recently synthesized polycyclic amines).^{28,63,68,80–81} The profound dependence of IIc inhibition on [Mg²⁺] and on membrane potential strongly supports the conclusion that IIc inhibits NMDARs by binding in and blocking the NMDAR channel.

2.6. In vitro DMPK profile, hERG safety and cytotoxicity of IIc

This compound was characterized in terms of microsomal stability (human and mouse species), hERG (human ether-a-go-go-related gene) inhibition, cytochrome P450 (CYP) inhibition, Caco-2 and predicted brain permeability (PAMPA-BBB assay), to confirm its suitability as a candidate for in vivo studies in Caenorhabditis elegans (C. elegans) and in the 5XFAD mouse model. Results from these measurements are given in Table 2.

Table 2:

Microsomal stability, cytochrome and hERG inhibition, Caco-2 and PAMPA-BBB of IIc.

Microsomal stabiilitya		CYP inhibitionb						hERG inhibitionb	Caco-2			PAMPA-BBB
									Papp (nm / s)c
h	m	1A2	2C19	2C9	2D6	3A4d	3A4e		A-B	B-A	ER
96	77	9 ± 2	21 ± 2	9 ± 3	15 ± 1	4 ± 2	10 ± 2	13 ± 1	310 ± 41	208 ± 12	0.7 ± 0.1	CNS+

^aPercentage of remaining compound after 60 min of incubation with human (h) and mouse (m) microsomes in the presence of NADPH at 37°C.

^bInhibition at 10 μM.

^cThe efflux ratio (ER) was calculated as ER = (Papp B→A) / (Papp A→B).

^dUsing benzyloxytrifluoromethylcoumarin (BFC) as substrate.

^eUsing dibenzylfluorescein (DBF) as substrate.

See methods for further details.

IIc showed high metabolic stability in human and mice liver microsomes, which are widely used to determine the likely degree of primary metabolic clearance in the liver. Next, cytochrome P450 (CYP) inhibition was evaluated through a fluorescence-detection method using human recombinant cytochrome P450 enzymes CYP1A2, CYP2C19, CYP2C9, CYP2D6, and CYP3A4. Pleasantly, the compound did not significantly inhibit the evaluated cytochromes. hERG channel is an ion channel which is commonly used as an important toxicology screen because of its known association with cardiotoxicity. Interestingly, IIc, at 10 μM, did not significantly inhibit hERG. The Caco-2 cell permeability model was used to evaluate the permeability of the compound. Apparent permeability values (P_app) were determined from the amount permeated through the Caco-2 cell membranes in both apical-basolateral (A-B) and basolateral-apical (B-A) directions. Satisfactorily, IIc showed good permeability. Any AD drug should cross the blood-brain-barrier (BBB), so we used the PAMPA-BBB assay to predict the ability of IIc to permeate BBB. IIc had a positive predicted BBB permeability. Next, the cytotoxicity of IIc was tested using the MTT assay in Neuro2a cells. Interestingly, the selected compounds did not show cytotoxic at the highest concentration tested (100 μM) (see section 4.7. for further details).

Overall, IIc has low micromolar activity, an excellent electrophysiological profile as a NMDA receptor blocker, favorable DMPK properties, and non-cytotoxicity. Thus, IIc was deemed a suitable candidate for the in vivo studies.⁸²

2.7. In vivo efficacy studies

To evaluate the effectivity of IIc to ameliorate cognitive impairment through the modulation of NMDAR-mediated neurodegenerative pathways, we performed a series of in vivo studies in C. elegans and in the 5XFAD mouse model.

It has been reported that memory in C. elegans is mediated by NMDA-type ionotropic glutamate receptors.⁸³ This nematode expresses two NMDA-type subunits, NMR-1 and NMR-2, homologous to mammalian GluN1 and GluN2A subunits, respectively.⁸⁴ Thus, C. elegans has been described as a suitable model for the study of NMDAR-mediated neurotoxicity.^85–86 The 5XFAD mouse model, which displays altered NMDAR function,⁸⁷ is one of the most used AD mouse models to evaluate pharmacological interventions.⁸⁸

Considering that IIc has proven to be a memantine analog in terms of potency and electrophysiological behavior, we incorporated memantine as a standard to analyze if these analogies persist in the in vivo efficacy studies. Of note, just a single study has been published on the effects of memantine in C. elegans, using the transgenic strain CL2006,⁸⁴ and only two studies have been published describing effects of memantine on 5XFAD mice.^89–90

2.7.1. Concentration-response effect of memantine and IIc on paralysis induced by Aβ expression in CL2006

The analysis of locomotor deficits has been traditionally used to discriminate between N2 wild-type (WT) and CL2006 mutant C. elegans. Herein, the N2 WT strain was used as a positive control and CL2006 worms, which constitutively expresses human Aβ_1–42, served as an AD model.⁹¹ As expected, CL2006 worms presented paralysis (measured as defective locomotion) compared to N2 WT, which exhibited a significantly lower percentage of defective locomotion (Figure 6A). After treatments, we found that memantine and IIc reduced defective locomotion over a range of concentrations (0.1 to 10 μM) in CL2006 nematodes (Figure 6A–B). These initial results in nematodes showed the capability of IIc, as well as memantine, to rescue the motor deficits in an AD model.

Figure 6.

A. Locomotion defective (%) of N2 WT vs CL2006 strains. B. Motor index (%) calculated in CL2006 worms treated with the indicated memantine or IIc concentrations. C. Chemotaxis index (CI) for CL2122 and CL2355 worms treated with vehicle (Veh), memantine (Mem) and IIc. Values are depicted as mean ± Standard error of the mean (SEM). ****p < 0.0001 for N2 vs CL2006; *p < 0.05, **p < 0.01 for memantine 0.5 μM, 1.5 μM and 10 μM vs memantine 0 μM; #p < 0.05, ##p < 0.01, ####p < 0.0001 for IIc 0.5 μM, 1.5 μM and 10 μM vs IIc 0 μM; *p < 0.05 for CL2355 Veh vs other groups.

2.7.2. Treatment with NMDAR channel blockers rescues chemotaxis behavior of CL2355 nematodes disrupted by Aβ expression

The impressive chemotaxis behavior of C. elegans is well known as a suitable model to evaluate the neuroprotective effects of new compounds. To evaluate the effect of NMDAR antagonists on the chemotaxis behavior of C. elegans, the transgenic strain CL2355, which shows Aβ aggregates as well as deficits in chemotaxis, and CL2122, a positive control strain, were used.⁹¹ Therefore, we performed a chemotaxis assay as previously described.⁹² As expected, CL2355 showed disrupted chemotaxis behavior and a reduced chemotaxis index (CI) in comparison to CL2122. Interestingly, both memantine and IIc treatments significantly reversed the disrupted chemotaxis behavior, suggesting that IIc might protect against Aβ toxicity-related neuronal dysfunction in a fashion similar to memantine (Figure 6C).

2.7.3. NMDA channel blockers memantine and IIc rescue working memory impairment in 5XFAD

To determine the effects of the administration of memantine and IIc on AD-like symptoms, 5XFAD mice, a model of familial AD, were treated with either memantine or IIc at 5 mg / day for 4 weeks. A novel object recognition test (NORT) was used to evaluate short- and long-term working memories in 5XFAD mice. As described previously,^88,93 6-month-old 5XFAD mice presented robust cognitive deficits compared to WT (Figure 7). We found that memantine treatment resulted in the recovery of cognitive function by increasing the discrimination index (DI) in both short-term (Figure 7A) and long-term (Figure 7B) memory tests. Importantly, IIc also enhanced working memory function in 5XFAD mice, with treated mice showing a significant DI increase at both tests (Figure 7).

Figure 7.

A. Discrimination index (DI) calculated by using exploration time for novel and familiar object in the short-term (2 h) memory test. B. DI in the long-term memory test session (24 h). Values are the mean ± Standard error of the mean (SEM); (n = 14 for WT, n = 8 for 5XFAD control groups, n = 11 for memantine (Mem, 5 mg / kg) and n = 10 for IIc (5 mg/kg) groups). ***p < 0.001; ****p < 0.0001 for WT vs 5XFAD control. ###p < 0.001; ####p < 0.0001 for memantine (Mem, 5 mg / kg) or IIc (5 mg / kg) vs 5XFAD control.

2.7.4. NMDAR molecular alterations in 5XFAD mouse model after IIc treatment

NMDARs are heterotetramers composed of two GluN1 subunits and typically two GluN2 subunits, of which there are four subtypes (GluN2A-D). In the adult forebrain, GluN2A and GluN2B are the two predominant subtypes. GluN2A subunits, incorporated into synaptic NMDARs, mediate long-term synaptic plasticity and their activation is associated with neuroprotective pathways. Therefore, selective activation of NMDARs containing GluN2A subunits would have a beneficial effect on AD pathology.^94–95

Western-blot analysis showed that 5XFAD mice exhibit less GluN2A protein levels than control mice, and interestingly, IIc treatment was able to rescue GluN2A to control levels (Figures 8A–B).

Figure 8.

A. Representative Western Blots and quantification. B. GluN2A. C. Ratio p-Fyn / Fyn. D. Ratio p-GluN2B (Tyr1472) / GluN2B. Values plotted in bar graphs are adjusted to 100% for protein levels of wild type (WT). Values are the mean ± Standard error of the mean (SEM); (n = 3 for WT and 5XFAD Control groups and n = 4 for memantine (Mem, 5 mg / kg) and IIc (5 mg / kg) groups. **p < 0.01 for WT vs 5XFAD Control. #p < 0.05; ##p < 0.01; ###p < 0.001 for memantine (Mem, 5 mg / kg) or IIc (5 mg / kg) vs 5XFAD Control.

Extrasynaptic NMDARs are enriched in GluN2B subunits and are believed to promote neuronal death, as occurring in neurodegenerative diseases. The activation of NDMARs containing the GluN2B subunit may be increased by Aβ, contributing to neurodegenerative and apoptotic pathways.⁹⁶ Memantine preferentially blocks extrasynaptic NMDARs over those located at the synaptic cleft, contributing to the clinical tolerability of this drug.^21,97–98

Additionally, GluN2B-subunit-containing NMDARs may induce synaptic NMDAR internalization after dephosphorylation, disrupting normal synaptic function.⁹⁹ Furthermore, phosphorylated Fyn kinase (p-Fyn) is responsible for phosphorylation of GluN2B at tyrosine 1472, which prevents GluN2B endocytosis and maintains NMDARs at the synaptic surface, preventing neuronal death.¹⁰⁰ To evaluate GluN2B subunit and Fyn kinase phosphorylation in control mice and the effects of memantine and IIc on 5XFAD mice, the p-GluN2B(Tyr1472) / GluN2B and p-Fyn / Fyn ratios (Figures 8A, 8C–D) were determined. Of note, we found that treatment of 5XFAD mice with memantine or IIc increased Fyn phosphorylated levels and correspondingly elevated GluN2B phosphorylation at Tyr1472. These findings support the literature relating GluN2B -subunit phosphorylation at Tyr1472 and the prevention of NMDAR internalization, which may prevent synaptic dysfunction.^101–102

2.7.5. IIc treatment modifies CREB pathways leading to synaptic improvements

Fyn activity is related to proteins implicated in cell-survival, such as cAMP response element binding protein (CREB).¹⁰³ Therefore, we evaluated p-CREB protein levels in cytosol and nuclei. p-CREB levels were found to be dysregulated in 5XFAD compared to WT. p-CREB protein in the nucleus was reduced in 5XFAD mice compared to the WT group, whereas treatment with memantine or IIc significantly increased p-CREB protein levels in the nucleus compared to untreated 5XFAD mice. The enhanced p-CREB translocation to the nucleus of the treated groups (Figure 9A, 9B) indicated a beneficial effect of NMDAR antagonists by modulating signaling pathways involved in cell survival.

Figure 9.

A. Representative Western Blot and quantifications for B. Ratio of p-CREB in the nucleus vs in the cytosol C. Calbindin D. PSD95. Values in bar graphs are adjusted to 100% for protein levels of the wild type (WT). Values are the mean ± Standard error of the mean (SEM); (n = 3 for WT and 5XFAD Control groups and n = 4 for memantine (Mem, 5 mg / kg) and IIc (5 mg / kg) groups. *p < 0.05; ****p < 0.0001 for WT vs 5XFAD Control. #p < 0.05; ##p < 0.01 for memantine (Mem, 5 mg / kg) or IIc (5 mg / kg) vs 5XFAD Control.

To further evaluate the effect of memantine and IIc on AD-related proteins, the calciumdependent protein calbindin was evaluated. Calbindin D-28K is a Ca²⁺ binding protein and an important regulator of Ca²⁺ homeostasis found to be dysregulated in AD animal models.¹⁰⁴ Remarkably, we found a significant, unprecedented dysregulation of calbindin protein levels in 5XFAD mice (Figure 9A, 9C), suggesting the important role of calbindin in the Aβ pathology presented by this transgenic mouse model. We also demonstrated that treatment with memantine or IIc reverts calbindin D-28K protein levels to WT levels (Figure 9A, 9C). It is noteworthy that p-CREB is a transcription factor that increases the expression of several proteins implicated in neuroprotection and neuronal signaling.¹⁰⁵ Interestingly, we found that treatment with memantine or IIc rescues expression of postsynaptic density protein (PSD) 95 in 5XFAD mice (Figure 9D). Together, these data may explain the improvement in working memory observed in 5XFAD after treatment with NMDAR channel blockers.

Overall, we provide promising evidence that IIc treatment rescues Fyn-GluN2B-CREB signaling and PSD95 expression in 5XFAD mice. Our results suggest a mechanism by which IIc may improve cell survival and synaptic function in AD through increasing the activity of cell-survival signaling pathways and preventing internalization of synaptic NMDARs. Thus, IIc could be a feasible candidate for further testing as a new clinically therapeutic NMDAR channel blocker.

3. Conclusions

Despite scientific efforts to develop new clinically useful NMDAR blockers, memantine remains the only noncholinergic drug approved for symptomatic AD treatment. Therefore, we further explored the benzohomoadamantane scaffold with the main aim to develop new memantine analogs. Although most of the changes in the chemical structure of benzohomoadamantane negatively impact potency at NMDARs, 2, IIa and IIc showed low micromolar potencies. Furthermore, these results were rationalized and strongly supported by MD simulations. Subsequently, of the most active derivatives, IIc was found to possess electrophysiological characteristics similar to memantine, including similar potencies, voltage dependencies, and blocking sites. Additionally, our DMPK studies (microsomal stability, cytochrome and hERG inhibition, Caco-2 and PAMPA assays) revealed that IIc was a good candidate for in vivo studies. The administration of IIc to both C. elegans and the 5XFAD mouse model of AD reverted the condition of the animals to that of controls, both at the behavioral and the biochemical levels. The effect observed was comparable with that produced by memantine, demonstrating that IIc could be a feasible and promising candidate for a new clinically therapeutic NMDAR blocker.

4. Experimental part

4.1. Chemistry. General methods

Commercially available reagents and solvents were used without further purification unless stated otherwise. Preparative normal phase chromatography was performed on a CombiFlash Rf 150 (Teledyne Isco) with pre-packed RediSep Rf silica gel cartridges. Thin-layer chromatography was performed with aluminum-backed sheets with silica gel 60 F254 (Merck, ref 1.05554), and spots were visualized with UV light and 1% aqueous solution of KMnO₄. Melting points were determined in open capillary tubes with a MFB 595010M Gallenkamp. 400 MHz ¹H and 100.6 MHz ¹³C NMR spectra were recorded on a Varian Mercury 400 or on a Bruker 400 Avance III spectrometers. The chemical shifts are reported in ppm (δ scale) relative to internal tetramethylsilane, and coupling constants are reported in Hertz (Hz). Assignments given for the NMR spectra of selected new compounds have been carried out on the basis of COSY ¹H / ¹³C (gHSQC) experiments. IR spectra were run on Perkin-Elmer Spectrum RX I, Perkin-Elmer Spectrum TWO or Nicolet Avatar 320 FT-IR spectrophotometers. Absorption values are expressed as wavenumbers (cm⁻¹); only significant absorption bands are given. High-resolution mass spectrometry (HRMS) analyses were performed with an LC / MSD TOF Agilent Technologies spectrometer. The elemental analyses were carried out in a Flash 1112 series Thermofinnigan elemental microanalyzator (A5) to determine C, H, N. HPLC / MS were determined with a HPLC Agilent 1260 Infinity II LC/MSD coupled to a photodiode array and mass spectrometer. 5 μL of sample 0.5 mg/mL in methanol: acetonitrile were injected, using an Agilent Poroshell 120 EC-C18, 2.7 μm, 50 mm × 4.6 mm column at 40 °C. The mobile phase was a mixture of A = water with 0.05% formic acid and B = acetonitrile with 0.05% formic acid, with the method described as follows: flow 0.6 mL/min, 5% B-95% A 3 min, 100% B 4 min, 95% B-5% A 1 min. Purity is given as % of absorbance at 210 or 275 nm. The structure of all new compounds was confirmed by elemental analysis and/or accurate mass measurement, IR, ¹H NMR and ¹³C NMR. The analytical samples of all the new compounds, which were subjected to pharmacological evaluation, possessed purity >95% as evidenced by their elemental analyses or HPLC-UV.

4.1.1. 5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-9-d-7-amine (2)

1. 9-bromo-6,7,8,9,10,11-hexahydro-9-methyl-5,7:9,11-dimethano-5H-benzocyclononen-7-amine hydrochloride. Thionyl bromide (6.5 mL, δ = 2.68, 83.8 mmol) was added to a solution of amine 1 (500 mg, 2.81 mmol) in toluene (17 mL) containing a few drops of DMF. The resulting red solution was stirred at rt for 1.5 hours. The reaction mixture was concentrated to dryness in vacuo. Toluene (50 mL) was added, and the resulting solution concentrated in vacuo. The procedure was repeated five times more until an orange solid was obtained. The crude was partitioned between DCM (15 mL) and saturated aq NaHCO₃ solution (15 mL) and the phases were separated. The aq layer was extracted with further dichloromethane (100 mL) and the combined organic extracts were dried over anh Na₂SO₄, filtered and concentrated in vacuo to give a colourless oil (530 mg, 83% yield) that was used in the next step without further purification.

2. 5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-9-d-7-amine. To a suspension of 9-bromo-6,7,8,9,10,11-hexahydro-9-methyl-5,7:9,11-dimethano-5H-benzocyclononen-7-amine hydrochloride (prepared from the addition of an excess of HCl / Et₂O to the compound) (250 mg, 0.76 mmol) in anh and deoxygenated toluene (3 mL) under Ar atmosphere were added tributylstannane-d (372 μL, 1.37 mmol) and AIBN (16 mg, 0.10 mmol). The resulting solution was heated to 95 °C for 4 hours adding more AIBN (16 mg, 0.10 mmol) each 60 min. Saturated aq Na₂CO₃ solution was added until pH = 12 and then the aq layer was extracted with 10% MeOH / EtOAc (3 × 15 mL). All organic layers were joined, dried over anh Na₂SO₄, filtered and concentrated in vacuo. The resulting crude was purified by Biotage^® purification system in silica gel using a MeOH / NH₃ 0.7 M in DCM mixture (0.4/9.6) to obtain 2 as a pale-yellow solid (100 mg, 61% yield). Its hydrochloride was obtained by adding an excess of HCl / Et₂O in DCM, followed by filtration of the resulting white precipitate. mp 72–73 °C. IR (ATR) ν: 2906, 2838, 1653, 1560, 1490, 1439, 1352, 1115, 1047, 990, 933, 914, 874, 826, 755, 702, 610 cm⁻¹. ¹H-NMR (400 MHz, CD₃OD) δ: 1.66–1.72 [complex signal, 6H, 6(12)-H_a, 10(13)-H_a and 8-H₂], 1.85 [m, 2H, 6(12)-H_b], 1.92 [dd, J = 12.8 Hz, J’ = 6.0 Hz, 2H, 10(13)-H_b], 3.03 [tt, J = 6.0, J’= 1.5 Hz, 2H, 5(11)-H], 7.02 (br s, 4H, Ar-H). ¹³C-NMR (100.6 MHz, CD₃OD) δ: 32.7 (t, J_C-D = 20.0 Hz, C9), 35.4 [CH₂, C10(13)], 42.9 [CH, C5(11)], 44.1 [CH₂, C6(12)], 45.0 (CH₂, C8), 48.8 (C, C7), 127.3 [CH, C2(C3)], 129.0 [CH, C1(C4)], 148.0 [C, C4a(C11a)]. HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₅H₁₉DN]⁺: 215.1658, found: 215.1653.

4.1.2. 7-iodo-5,6,7,8,9,10-hexahydro-5,8:7,10-dimethanobenzo[8]annulene (4)

1. 5,8,9,10-tetrahydro-5,8:7,10-dimethanobenzo[8]annulen-7(6H)-yl methanesulfonate. To a solution of 5,8,9,10-tetrahydro-5,8:7,10-dimethanobenzo[8]annulen-7(6H)-ol (1.19 g, 5.40 mmol) in pyridine (9 mL), methanesulfonyl chloride (2.32 mL, δ = 1.48, 29.9 mmol) was added slowly with stirring at rt. The mixture was then heated at 120 °C for 5 hours. After cooling, crushed ice (100 g) was added and the mixture was extracted with DCM (5 × 40 mL). The combined organic phase was washed with 2 N HCl solution (2 × 40 mL), H₂O (2 × 40 mL), saturated aq NaHCO₃ solution (2 × 40 mL) and dried over anh Na₂SO₄. The crude mesylate was obtained after filtration and removal of the solvent under reduced pressure as a dark oil that was used in the next step without further purification (1.32 g, 80% yield). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₅H₁₈O₃S+NH₄]⁺: 296.1315, found: 296.1318.

2. 7-iodo-5,6,7,8,9,10-hexahydro-5,8:7,10-dimethanobenzo[8]annulene. In an RBF equipped with a reflux condenser, H₃PO₃ (99%, 135 g), mesylate (1.32 g, 4.75 mmol) and NaI (63 g, 42.0 mmol) were added. The mixture was stirred at 150 °C for 6 hours. After cooling, H₂O (150 mL) was added slowly to the mixture. The resulting red solution was extracted with DCM (4 × 80 mL) and the combined organic phase was washed with saturated aq sodium thiosulfate solution (1 × 100 mL), dried over anh Na₂SO₄, then filtered and concentrated under vacuum to give 4 as a white solid (1.39 g, 95% yield). mp 132–133 °C (dec.). IR (NaCl disk) ν: 3052, 3013, 2950, 2892, 2852, 1490, 1447, 1304, 1278, 1232, 1215, 1095, 1046, 1032, 967, 830, 778, 755 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 1.59 [dd, J = 12.4 Hz, J’ = 2.4 Hz, 2H, 9(12)-H_a], 2.01 [dd, J = 11.6 Hz, J’ = 2.4 Hz, 2H, 6(11)-H_a], 2.24 [m, 2H, 6(11)-H_b], 2.39 [m, 2H, 9(12)-H_b], 2.55 (t, J = 8.8 Hz, 1H, 8-H), 3.23 [t, J = 6.0 Hz, 2H, 5(10)-H], 7.08 [m, 2H, 1(4)-H or 2(3)-H], 7.13 [m, 2H, 2(3)-H or 1(4)-H]. ¹³C-NMR (100.6 MHz, CDCl₃) δ: 42.8 [CH₂, C9(12)], 48.3 [CH, C5(10)], 49.6 (CH, C8), 51.0 [CH₂, C6(11)], 90.5 (C, C7), 126.2 [CH, C1(4) or C2(3)], 129.5 [CH, C2(3) or C1(4)], 145.3 [C, C4a(10a)]. GC-MS (El), t_r = 17.8 min: 310 [(M)^·+, 2], 183 [(M-I)⁺, 100], 141 (73), 129 (23), 128 (15).

4.1.3. 5,8,9,10-tetrahydro-5,8:7,10-dimethanobenzo[8]annulene-7(6H)-carboxylic acid (5)

To a solution of 7-iodo-5,6,7,8,9,10-hexahydro-5,8:7,10-dimethanobenzo[8]annulene (2.03 g, 6.54 mmol) in dry and degassed toluene (20 mL), methyl oxalyl chloride (1.8 mL, δ = 1.33, 19.5 mmol) and bis(tributyltin) (4.55 g, 7.85 mmol) were added. The mixture was irradiated in a quartz reactor under argon atmosphere with a 125 W Hg lamp for 20 hours. Then, DCM (15 mL), methanol (0.6 mL) and triethylamine (1.2 mL) were successively added to the reaction mixture at 0 °C and it was concentrated under vacuum to give a dark oil (3.99 g). A solution of this oil in a 40% methanol solution of KOH (50 mL) was heated to reflux for 2 hours. H₂O (50 mL) was added, and the reaction was refluxed for 3 hours. The reaction mixture was allowed to cool to rt and the methanol was removed under vacuum. H₂O (40 mL) was added to the residue and the aq layer was washed with DCM (4 × 50 mL). After that, the aq phase was acidified with conc. HCl until pH = 1 and extracted with DCM (4 × 50 mL). The organic extracts were dried over anh Na₂SO₄, filtered and concentrated under reduced pressure to give the acid as a brown solid (555 mg, 37% overall yield). An analytical sample of the acid was obtained by crystallization from DCM/Pentane, mp 188–189 °C. IR (NaCl disk) ν: 3300–2800 (3065, 3011, 2946, 2858), 1690, 1488, 1450, 1410, 1318, 1290, 1231, 1218, 1092, 1052, 1038, 941 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 1.71 [dd, J = 12.8Hz, J’ = 2.4 Hz, 2H, 9(12)-H_a], 1.89 [dd, J = 12.4 Hz, J’ = 2.4 Hz, 2H, 6(11)-H_a], 2.23 [m, 2H, 9(12)-H_b], 2.57 [dm, J = 12.4 Hz, 2H, 6(11)-H_b], 3.20 (t, J = 8.8 Hz, 1H, 8-H), 3.27 [t, J = 6 Hz, 2H, 5(10)-H], 7.08 [m, 2H, 1(4)-H or 2(3)-H], 7.15 [m, 2H, 2(3)-H or 1(4)-H]. ¹³C-NMR (100.6 MHz, CDCl₃) δ: 43.3 [CH₂, C9(12)], 46.4 [CH₂, C6(11)], 49.4 [CH, C5(10)], 49.9 (CH, C8), 59.9 (C, C7), 126.2 [CH, C1(4) or C2(3)], 129.5 [CH, C2(3) or C1(4)], 145.4 [C, C4a(10a)], 184.1 (C, C7-COOH). HRMS-ESI+ m/z [M-H]⁻ calcd for [C₁₅H₁₅O₂]⁻: 227.1078, found: 227.1078.

4.1.4. 5,8,9,10-tetrahydro-5,8:7,10-dimethanobenzo[8]annulen-7(6H)-amine hydrochloride (6)

To a solution of 5,8,9,10-tetrahydro-5,8:7,10-dimethano-benzo[8]annulene-7(6H)-carboxylic acid (90 mg, 0.39 mmol) in toluene (1.2 mL), Et₃N (73 μL, 0.53 mmol) and diphenylphosphoryl azide (159 mg, 0.58 mmol) were added and the mixture was heated at reflux for 3 hours. The mixture was cooled and washed with 1 N HCl solution (2 × 10 mL). Thereafter, to the organic layer was added 6 N HCl solution (1.6 mL) and the suspension was heated at reflux for 24 hours. The reaction mixture was then cooled to rt and the two phases were separated. The aq phase was extracted with ethyl acetate (3 × 3 mL). The combined organic phases were washed with 5 N NaOH solution (3 × 10 mL), dried over anh Na₂SO₄, filtered and concentrated under vacuum to give 6. Its hydrochloride was obtained by adding an excess of HCl / methanol to a solution of the amine in methanol. The methanol was removed under reduced pressure yielding 6 as a brown solid (35 mg, 38% yield). An analytical sample was obtained by crystallization from MeOH / Et₂O, mp > 250 °C. IR (KBr disk) ν: 3100–2500 (2943, 2881, 2665), 2043, 1622, 1598, 1501, 1448, 1336, 1246, 1175, 1089, 1057, 1028, 952, 769, 749, 614 cm⁻¹. ¹H-NMR (400 MHz, MeOD) δ: 1.67 [dd, J = 12.4 Hz, J’ = 2.4 Hz, 2H, 9(12)-H_a], 2.02 [dd, J = 12.0 Hz, J’ = 2.4 Hz, 2H, 6(11)-H_a], 2.30 [m, 2H, 6(11)-H_b], 2.42 [m, 2H, 9(12)-H_b], 2.80 (t, J = 8.8 Hz, 1H, 8-H), 3.33 [t, J = 6.0 Hz, 2H, 5(10)-H], 7.08 [m, 2H, 1(4)-H or 2(3)-H], 7.15 [m, 2H, 2(3)-H or 1(4)-H]. ¹³C-NMR (100.6 MHz, MeOD) δ: 43.4 [CH₂, C9(12)], 48.2 [CH₂, C6(11)], 49.1 (CH, C8), 49.4 [CH, C5(10)], 70.6 (C, C7), 127.6 [CH, C1(4) or C2(3)], 130.5 [CH, C2(3) or C1(4)], 145.3 [C, C4a(10a)]. HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₄H₁₈N]⁺: 200.1434, found: 200.1432.

4.1.5. 2-chloro-N-(9-methyl-2-nitro-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-yl)acetamide (8).

To a cold (0°C) solution of chloroacetamide 7 (3.0 g, 9.87 mmol) in acetic anhydride (10 mL), glacial acetic acid (1.6 mL) and fuming nitric acid (1.9 mL) were carefully added. The mixture was allowed to reach rt and left stirring overnight. The obtained yellow solution was then poured into ice–water (20 mL) and extracted with DCM (3 × 40 mL). The combined organic extracts were washed with 2 N NaOH solution (1 × 40 mL), H₂O (1 × 40 mL) and brine (1 × 40 mL). The organic layer was dried over anh Na₂SO₄, filtered and concentrated under vacuum to give a yellow residue. Purification by Combiflash^® in silica gel using as eluent a gradient from hexane to ethyl acetate / hexane mixture (1 / 9) gave 8 as a white solid (2.92 g, 85% yield), mp 174–176 °C. IR (ATR) ν: 3269, 3233, 3080, 2951, 2922, 2903, 2846, 1670, 1563, 1518, 1460, 1405, 1362, 1346, 1336, 1303, 1279, 1226, 1166, 1138, 1083, 884, 839, 794, 739, 708, 667 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 0.96 (s, 3H, 9-CH₃), 1.50–1.58 [complex signal, 2H, 10(13)-H_a], 1.67–1.76 [complex signal, 2 H, 10(13)-H_b], 1.81 (dm, J = 12.0 Hz, 1H, 8-H_a), 1.87 (dm, J = 12.0 Hz, 1H, 8-H_b), 2.10–2.23 [complex signal, 4H, 6(12)-H₂], 3.19–3.25 [complex signal, 2H, 5(11)-H], 3.93 (s, 2H, CH₂Cl), 6.34 (broad s., 1H, NH), 7.21 (d, J = 8.8 Hz, 1 H, 4-H), 7.94 (dd, J = 8.8 Hz, J’ = 2.4 Hz, 1 H, 3-H), 7.95 (d, J = 2.4 Hz, 1H, 1-H). ¹³C-NMR (100.6 MHz, CDCl₃) δ: 32.1 (CH₃, C9-CH₃), 33.8 (C, C9), 38.0 (CH₂) and 38.3 (CH₂) (C6 and C12), 40.5 (CH₂) and 40.7 (CH₂) (C10 and C13), 40.88 (CH) and 40.90 (CH) (C5 and C11), 43.0 (CH₂, CH₂Cl), 46.9 (CH₂, C8), 54.6 (C, C7), 121.9 (CH) and 123.1 (CH) (C1 and C3), 129.2 (CH, C4), 146.6 (C, C11a), 147.6 (C, C2), 153.8 (C, C4a), 164.9 (C, CO). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₈H₂₂ClN₂O₃]⁺: 349.1313, found: 349.1313.

4.1.6. 9-methyl-2-nitro-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-amine hydrochloride (9)

Thiourea (177 mg, 2.33 mmol) and glacial acetic acid (1 mL) were added to a solution of chloroacetamide 8 (677 mg, 1.94 mmol) in abs. ethanol (33 mL) and the mixture was heated at reflux overnight. The resulting suspension was then tempered to rt, H₂O (20 mL) was added and the pH adjusted to 12 with 5 N NaOH solution. DCM (40 mL) was added, the phases were separated, and the aqueous phase was extracted with further DCM (2 × 40 mL). The combined organic layers were dried over anh Na₂SO₄, then filtered and concentrated in vacuo to give 9 as a light brown solid. Its hydrochloride was obtained by adding an excess of HCl / Et₂O to a solution of the amine in ethyl acetate, followed by filtration of the resulting white precipitate (443 mg, 74% yield), mp > 225 °C (dec). IR (ATR) ν: 3381, 3100–2800 (2997, 2927, 2856), 2065, 1613, 1520, 1486, 1350, 1329, 1305, 1286, 1138, 1033, 896, 837, 801, 762, 736 cm⁻¹. ¹H-NMR (400 MHz, MeOD) δ: 1.02 (s, 3H, CH₃), 1.55–1.63 [complex signal, 2H, 10(13)-H_a], 1.71–1.75 (complex signal, 1H, 8-H₂), 1.73–1.81 [complex signal, 2 H, 10(13)-H_b], 1.85–1.94 [complex signal, 2H, 6(12)-H_a), 2.07–2.14 [complex signal, 2H, 6(12)-H_b], 3.36–3.42 [complex signal, 2H, 5(11)-H], 7.37 (d, J = 8.4 Hz, 1H, 4-H), 7.99 (dd, J = 8.2 Hz, J’ = 2.4 Hz, 1H, 3-H), 8.02 (d, J = 2.4 Hz, 1H , 1-H). ¹³C-NMR (100.6 MHz, MeOD) δ: 32.0 (CH₃, C9-CH₃), 34.7 (C, C9), 38.5 (CH₂) and 38.7 (CH₂) (C6 and C12), 40.8 (CH₂) and 41.0 (CH₂) (C10 and C13), 41.3 (CH) and 41.4 (CH) (C5 and C11), 46.8 (CH₂, C8), 55.2 (C, C7), 123.0 (CH, C3), 124.0 (CH, C1), 130.6 (CH, C4), 148.1 (C, C11a), 148.3 (C, C2), 154.2 (C, C4a). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₆H₂₁N₂O₂]⁺: 273.1598, found: 273.1604.

4.1.7. 9-methyl-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulene-2,7-diamine dihydrochloride (10)

To a solution of amine 9·HCl (200 mg, 0.65 mmol) in methanol (25 mL), Pd on charcoal (20 mg, ca. 10% Pd) was added and the resulting suspension was hydrogenated at 1 atm of H₂ at rt for 48 hours. The black suspension was filtered, and the solvent removed by concentration in vacuo to give 10 as a brown solid. Its hydrochloride was obtained by addition of an excess of HCl / MeOH to a solution of the amine in methanol followed by evaporation to obtain a brown solid (204 mg, quant. yield), mp 294–295 °C. IR (KBr disk) ν: 3200–2500 (3024, 2912, 2847, 2588), 1994, 1598, 1502, 1454, 1381, 1365, 1303, 1261, 1173, 1131, 1021, 957, 877, 827, 576, 473 cm⁻¹. ¹H-NMR (400 MHz, MeOD) δ: 1.01 (s, 3H, CH₃), 1.50–1.60 [complex signal, 2H, 10(13)-H_a], 1.68–1.80 [complex signal, 2H, 10(13)-H_b], 1.72 (broad s., 2H, 8-H₂), 1.84–1.94 (complex signal, 2H, 6-H_a, 12-H_a), 2.03–2.14 (complex signal, 2H, 6-H_b, 12-H_b), 3.26–3.29 [complex signal, 5(11)-H], 7.16 (m, 1H, 3-H), 7.18 (d, J = 2.0 Hz, 1H, 1-H), 7.29 (m, 1H, 4-H). ¹³C-NMR (100.6 MHz, MeOD) δ: 32.1 (CH₃, C9-CH₃), 34.7 (C, C9), 38.7 (CH₂) and 38.8 (CH₂) (C6 and C12), 41.1 (CH₂) and 41.15 (CH₂) (C10 and C13), 41.21 (CH₂) and 41.4 (CH₂) (C5 and C11), 46.8 (CH₂, C8), 55.3 (C, C7), 122.2 (CH) and 123.7 (CH) (C1 and C3), 130.4 (C, C4a), 131.1 (CH, C4), 147.1 (C) and 148.9 (C) (C2 and C11a). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₆H₂₃N₂]⁺: 243.1856, found: 243.1856.

4.1.8. N-(2-acetyl-9-methyl-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-yl)-2-chloroacetamide (11).

A solution of 7 (557 mg, 1.83 mmol) and acetyl chloride (1.31 mL, 18.4 mmol) were dissolved in DCM (14 mL) and treated with aluminium trichloride (1.22 g, 9.15 mmol). The resulting black mixture was stirred for 60 minutes then poured over a mixture of ice and saturated aq Na₂CO₃ solution (10 mL). After stirring 20 minutes the mixture was extracted with DCM (3 × 30 ml). The organic layer was dried over anh Na₂SO₄, filtered and concentrated in vacuo to give a yellow solid. Purification by Combiflash^® in silica gel using as eluent a gradient from hexane to ethyl acetate / hexane mixture (2 / 8) gave 11 as a white solid (340 mg, 54% yield), mp 186–187 °C. IR (ATR) ν: 3276, 3083, 2959, 2916, 2845, 1670, 1603, 1562, 1456, 1419, 1405, 1364, 1336, 1281, 1271, 1249, 1238, 1228, 1166, 1108, 1051, 1013, 966, 948, 842, 795, 704, 689, 643 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 0.95 (s, 3H, CH₃), 1.50–1.58 [complex signal, 2H, 10(13)-H_a], 1.66–1.74 [complex signal, 2H, 10(13)-H_b], 1.82 (dm, J = 12.0 Hz, 1H, 8-H_a), 1.87 (dm, J = 12.0 Hz, 1H, 8-H_b), 2.08–2.22 [complex signal, 4H, 6(12)-H₂], 2.57 (s, 3H, COCH₃), 3.13–3.23 [complex signal, 2H, 5-(11)-H], 3.92 (s, 2H, CH₂Cl), 6.32 (broad s., 1H, NH), 7.15 (d, J = 7.8 Hz, 1H, 4-H), 7.67 (m, 1H, 3-H), 7.69 (d, J = 1.6 Hz, 1H, 1-H). ¹³C-NMR (100.6 MHz, CDCl₃) δ : 26.7 (CH_3, CH₃CO), 32.2 (CH₃, C9-CH₃), 33.8 (C, C9), 38.3 (CH₂) and 38.6 (CH₂) (C6 and C12), 40.8 (CH₂) and 40.97 (CH₂) (C10 and C13), 41.0 (CH) (C5 and C11), 43.1 (CH₂, CH₂Cl), 46.9 (CH₂, C8), 54.8 (C, C7), 127.1 (CH, C1), 128.0 (CH, C3), 128.6 (CH, C4), 135.7 (C) and 146.5 (C) (C2 and C11a), 151.9 (C, C4a), 164.8 (C, NHCO), 198.1 (C, COCH₃). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₂₀H₂₅ClNO₂]⁺: 346.1568, found: 346.1570.

4.1.9. 1-(7-amino-9-methyl-6,7,8,9,10,11-hexahydro-5H-5,9:7,11-dimethanobenzo[9]annulen-2-yl)ethan-1-one hydrochloride (12)

Thiourea (233 mg, 3.06 mmol) and glacial acetic acid (1.6 mL) were added to a solution of chloroacetamide 11 (845 mg, 2.44 mmol) in abs. ethanol (42 mL) and the mixture was heated at reflux overnight. The resulting suspension was then tempered to rt, then H₂O (20 mL) was added and the pH was adjusted to 12 with 5 N NaOH solution. DCM (40 mL) was added, the phases were separated, and the aq phase was extracted with further dichloromethane (2 × 40 mL). The combined organic layers were dried over anh Na₂SO₄, filtered and concentrated in vacuo to give a light brown solid. Purification by Combiflash^® in silica gel using as eluent a gradient from hexane to ethyl acetate / hexane mixture (3 / 7) gave 12 as a white solid. Its hydrochloride was obtained by adding an excess of HCl / Et₂O to a solution of the amine in ethyl acetate, followed by filtration of the resulting white precipitate (385 mg, 52% yield), mp >200 °C (dec.). IR (ATR) ν: 3000–2400 (2988, 2904, 2858), 1678, 1603, 1508, 1451, 1428, 1362, 1272, 1256, 1207, 1135, 1115, 1031, 1002, 949, 912, 835, 827, 649, 640 cm⁻¹. ¹H-NMR (400 MHz, MeOD) δ: 1.01 (s, 3H, CH₃), 1.53–1.61 [complex signal, 2H, 10(13)-H_a], 1.70 (broad s., 2H, 8-H₂), 1.70–1.80 [complex signal, 2 H, 10(13)-H_b], 1.82–1.92 [complex signal, 2H, 6(12)-H_a], 2.03–2.12 [complex signal, 2H, 6(12)-H_b], 2.57 (s, 3H, COCH₃), 3.33–3.35 [complex signal, 2H, 5(11)-H], 7.26 [d, J = 7.6 Hz, 1H, 4-H), 7.75–7.78 (complex signal, 2H, 1-H, 3-H). ¹³C-NMR (100.6 MHz, MeOD) δ: 26.7 (CH₃, CH₃CO), 32.1 (CH₃, C9-CH₃), 34.7 (C, C9), 38.8 (CH₂) and 39.1 (CH₂) (C6 and C12), 41.1 (CH₂) and 41.4 (CH₂) (C10 and C13), 41.5 (CH) and 41.6 (CH) (C5 and C11), 47.0 (CH₂, C8), 55.3 (C, C7), 128.5 (CH) and 129.2 (CH) (C1 and C3), 129.8 (CH, C4), 137.2 (C) and 147.0 (C) (C2 and C11a), 152.3 (C, C4a), 200.2 (C, CO). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₈H₂₄NO]⁺: 270.1852, found: 270.1858.

4.1.10. 1-(7-(dimethylamino)-9-methyl-6,7,8,9,10,11-hexahydro-5H-5,9:7,11-dimethanobenzo[9]annulen-2-yl)ethan-1-one hydrochloride (14)

AlCl₃ (419 mg, 3.14 mmol) was added to a solution of 13 (161 mg, 0.63 mmol) and AcCl (0.45 mL, 6.31 mmol) in DCM (5 mL). The resulting mixture was stirred for 60 minutes and then poured over a mixture of ice and saturated aq Na₂CO₃ solution (10 mL). After stirring 20 minutes the mixture was extracted with DCM (3× 10 mL). The organic layer was dried over anh Na₂SO₄, filtered and concentrated in vacuo to give 14 as a yellow solid. Its hydrochloride was obtained by adding an excess of HCl / Et₂O to a solution of the amine in ethyl acetate, followed by filtration of the resulting light brown precipitate (83 mg, 44% yield), mp 180–181 °C. IR (ATR) ν: 3300–2400 (3221, 2947, 2928, 2903, 2850), 1666, 1657, 1651, 1646, 1518, 1505, 1460, 1399, 1363, 1355, 1285, 1277, 1258, 1245, 1211, 1190, 1158, 1133, 1124, 1091, 1082, 1025, 1006, 960, 900, 864, 813, 803, 740, 708, 689 cm⁻¹. ¹H-NMR (400 MHz, MeOD) δ: 1.04 (s, 3H, CH₃), 1.52–1.61 [complex signal, 2H, 10(13)-H_a], 1.71–1.81 [complex signal, 2 H, 10(13)-H_b], 1.81 (broad s., 2H, 8-H₂), 1.97–2.07 [complex signal, 2H, 6(12)-H_a], 2.10–2.21 [complex signal, 2H, 6(12)-H_b], 2.57 (s, 3H, COCH₃), 2.83 [s, 6H, N(CH₃)₂], 3.41 [complex signal, 2H, 5(11)-H], 7.27 (d, J = 8.0 Hz, 1H, 4-H), 7.75–7.80 (complex signal, 1-H, 3-H). ¹³C-NMR (100.6 MHz, MeOD) δ : 26.7 (CH₃, CH₃CO), 32.3 (CH₃, C9-CH₃), 34.0 (CH₂) and 34.2 (CH₂) (C6 and C12), 35.5 (C, C9), 37.3 [CH₃, N(CH₃)₂], 40.9 (CH₂) and 41.2 (CH₂) (C10 and C13), 41.5 (CH) and 41.6 (CH) (C5 and C11), 44.2 (CH₂, C8), 67.0 (C, C7), 128.6 (CH) and 129.2 (CH) (C1 and C3), 129.8 (CH, C4), 137.2 (C, C2), 146.9 (C, C11a), 152.2 (C, C4a), 200.1 (C,CO). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₂₀H₂₈NO]⁺: 298.2165, found: 298.2176.

4.1.11. N,N,9-trimethyl-2-nitro-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-amine hydrochloride (15)

To a solution of 9·HCl (222 mg, 0.72 mmol) in MeOH (7 mL), NaBH₃CN (129 mg, 2.05 mmol), AcOH (0.3 mL) and formaldehyde (0.22 mL, 37% in H₂O solution, 2.16 mmol) were added and the mixture was stirred at rt for 8 hours. An additional portion of NaBH₃CN (129 mg, 2.05 mmol) and formaldehyde (0.22 mL, 37% in H₂O solution, 2.16 mmol) were added, then the mixture was stirred at rt for 18 hours and subsequently was concentrated in vacuo to dryness. H₂O (20 mL) was added to the residue, then the suspension was basified with 1 N NaOH solution (10 mL) and was extracted with ethyl acetate (4 × 15 mL). The combined organic extracts were washed with brine (2 × 10 mL), dried with anh Na₂SO₄, then filtered and concentrated in vacuo to give 15 as a yellow oil. Its hydrochloride was obtained by adding an excess of HCl / Et₂O to a solution of the amine in ethyl acetate, followed by filtration of the white solid precipitate (180 mg, 74% yield), mp 168–169 °C. IR (ATR) ν: 3452, 3411, 2947, 2906, 2860, 2655, 2631, 2601, 1514, 1484, 1460, 1445, 1348, 1309, 1281, 1174, 1123, 1108, 1088, 1000, 982, 905, 896, 833, 801, 766, 736 cm⁻¹. ¹H-NMR (400 MHz, MeOD) δ: 1.05 (s, 3H, CH₃), 1.55–1.62 [complex signal, 2H, 10(13)-H_a], 1.75–1.82 [complex signal, 2 H, 10(13)-H_b], 1.82–1.85 (complex signal, 2H, 8-H₂), 2.01–2.11 [complex signal, 2H, 6(12)-H_a], 2.14–2.22 [complex signal, 2H, 6(12)-H_b], 2.84 [s, 6H, N(CH₃)₂], 3.44–3.50 [complex signal, 2H, 5(11)-H], 7.39 (d, J = 8.0 Hz, 1H, 4-H), 8.00 (dd, J = 8.0 Hz, J’ = 2.4 Hz, 1H, 3-H), 8.04 (d, J = 2.4 Hz, 1H, 1-H). ¹³C-NMR (100.6 MHz, MeOD) δ : 32.2 (CH₃, C9-CH₃), 33.6 (CH₂) and 33.9 (CH₂) (C6 and C12), 35.5 (C, C9), 37.3 [CH₃, N(CH₃)₂], 40.6 (CH₂) and 40.8 (CH₂), (C10 and C13), 41.3 (CH) and 41.4 (CH) (C5 and C11), 44.0 (CH₂, C8), 66.8 (C, C7), 123.0 (CH, C3), 124.0 (CH, C1), 130.6 (CH, C4), 148.1 (C) and 148.2 (C) (C2 and C11a), 154.1 (C, C4a). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₈H₂₅N₂O₂]⁺: 301.1911, found: 301.1916.

4.1.12. N-(9-methyl-2-nitro-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-yl)acetamide (17)

To a cold (0°C) solution of 16 (2.68 g, 9.95 mmol) in acetic anhydride (10.6 mL), glacial acetic acid (1.6 mL) and fuming nitric acid (1.9 mL) were carefully added. The mixture was allowed to reach rt and left stirring overnight. The obtained yellow solution was then poured into ice–water (20 mL) and extracted with DCM (3 × 40 mL). The combined organic extracts were washed with 2 N NaOH solution (1 × 40 mL), H₂O (1 × 40 mL) and brine (1 × 40 mL). The organic layer was dried over anh Na₂SO₄, filtered and concentrated in vacuo to give a yellow residue. Purification by Combiflash^® in silica gel using as eluent a gradient from hexane to ethyl acetate / hexane mixture (3 / 7) gave 17 as a white solid (1.87 g, 60% yield), mp 174–176 °C. IR (NaCl disk) ν: 3398, 3307, 3201, 3063, 2943, 2917, 2863, 1653, 1588, 1523, 1455, 1346, 1322, 1304, 1268, 1245, 1214, 1166, 1141, 1124, 1081, 1037, 1010, 945, 893, 865, 838, 798, 763, 740, 701, 645 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 0.93 (s, 3H, CH₃), 1.52 [complex signal, 2H, 10(13)-H_a], 1.67–1.73 [complex signal, 2H, 10(13)-H_b], 1.76 (dt, J = 12.0 Hz , J’ = 2.0 Hz, 2H, 8-H_a), 1.84 (dt, J = 12.0 Hz , J’ = 2.0 Hz, 2H, 8-H_b), 1.90 (s, CH₃, CH₃CO), 2.06–2.19 [complex signal, 4H, 6(12)-H₂], 3.17–3.21 [complex signal, 2H, 5(11)-H], 5.27 (broad s, 1H, NH), 7.20 (d, J = 8.8 Hz, 1H, 4-H), 7.91–7.94 (complex signal, 2H, 1-H, 3-H). ¹³C-NMR (100.6 MHz, CDCl₃) δ : 24.8 (CH₃, CH₃CO), 32.2 (CH₃, C9-CH₃), 33.7 (C, C9), 38.2 (CH₂, C6 or C12), 38.7 (CH₂, C12 or C6), 40.6 (CH₂, C10 or C13), 40.8 (CH₂, C13 or C10), 40.98 (CH, C5 or C11) and 41.00 (CH, C11 or C5), 47.2 (CH₂, C8), 54.1 (C, C7), 121.8 (CH, C1 or C3), 123.1 (CH, C3 or C1), 129.2 (CH, C4), 146.5 (C2 or C11a), 147.8 (C, C11a or C2), 154.1 (C, C4a), 169.5 (C, NHCO). HRMS-ESI+ m/z [M-H]⁺ calcd for [C₁₈H₂₃N₂O₃]⁺: 315.1703, found: 315.1714.

4.1.13. N-(2-amino-9-methyl-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-yl)acetamide (18)

To a solution of acetamide (2.65 g, 8.43 mmol) in abs. ethanol (140 mL), PtO₂ (190 mg, 0.84 mmol) was added, and the resulting suspension was hydrogenated at 1 atm of H₂ at rt for 4 hours. The black suspension was filtered, and the solvent removed by concentration in vacuo to give a dark solid. Purification by Combiflash^® in silica gel using as eluent a gradient from hexane to ethyl acetate / hexane mixture (5 / 5) gave 18 as a white solid (1.63 g, 68% yield), mp 113 °C. IR (NaCl disk) ν: 3432, 3324, 3224, 3056, 3004, 2938, 2903, 2856, 2835, 1651, 1618, 1546, 1507, 1447,1362, 1344, 1300, 1262, 1194, 1164, 1136, 1065, 862, 818, 735, 701 cm⁻¹. ¹H-RMN (400 MHz, CDCl₃) δ: 0.89 (s, 3H, CH₃), 1.45–1.54 [complex signal, 2H, 10(13)-H_a], 1.59–1.65 [complex signal, 2H, 10(13)-H_b], 1.76 (dm, J = 12.0 Hz, 1H, 8-H_a), 1.85 (dm, J = 12.0 Hz, 1H, 8-H_b), 1.89 (s, 3H, CH₃CO), 1.90 (m, 1 H, 6-H_a or 12-H_a), 1.96 (broad d, J = 12.0 Hz, 1 H, 12-H_a or 6-H_a), 2.09 (ddd, J = 12.8 Hz, J’ = 6.4 Hz, J” = 2.4 Hz, 2H, 6-H_b or 12-H_b), 2.17 (ddd, J = 12.8 Hz, J’ = 6.4 Hz, J” = 2.4 Hz, 2H, 12-H_b or 6-H_b), 2.90 (t, J = 6.0 Hz, 1H, 5-H or 11-H), 2.96 (t, J = 6.0 Hz, 1H, 11-H or 5-H), 5.21 (broad s., 1H, NH), 6.40 (dd, J = 8.0 Hz, J’ = 2.2 Hz, 1H, 3-H), 6.42 (d, J = 2.2 Hz, 1H, 1-H), 6.83 (d, J = 8.0 Hz, 1H, 4-H). ¹³C-NMR (100.6 MHz, CDCl₃) δ : 25.0 (CH₃, CH₃CO), 32.4 (CH₃, C9-CH₃), 33.6 (C, C9), 39.1 (CH₂, C12 or C6), 39.9 (CH₂, C6 or C12), 40.2 (CH, C11 or C5), 41.2 (CH, C5 or C11), 41.3 (CH₂, C10 or C13), 41.9 (CH₂, C13 or C10), 47.3 (CH₂, C8), 54.6 (C, C7), 112.6 (CH, C3), 115.4 (CH, C1), 129.1 (CH, C4), 136.9 (C, C4a), 144.6 (C, C11a), 147.5 (C, C2), 169.3 (C, NHCO). HRMS-ESI+ m/z [M-H]⁺ calcd for [C₁₈H₂₅N₂O]⁺:285.1961, found: 285.1972.

4.1.14. N-(2-chloro-9-methyl-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-yl)acetamide (19)

18·HCl (1.04 g, 3.24 mmol) was dissolved in H₂O (6 mL) and conc. HCl solution (6 mL) then cooled to 0 °C and treated with a solution of sodium nitrite (448 mg, 6.49 mmol) in H2O (2 mL) dropwise. To the resulting solution was added a CuCl (691 mg, 6.98 mmol) in conc. HCl solution (3 mL) and over 10 min gas was observed. The resulting solution was warmed to 60 °C for 90 minutes, then was cooled to rt, diluted in H₂O (60 mL) and extracted with DCM (4 × 90 mL). The combined organic extracts were washed with saturated aq NaHCO₃ solution, brine and were dried over anh Na₂SO₄, filtered and concentrated in vacuo to give a dark green solid. Purification by Combiflash^® in silica gel using as eluent a gradient from hexane to ethyl acetate / hexane mixture (1 / 9) gave 19 as a white solid (210 mg, 21% yield), mp 190–191 °C. IR (NaCl disk) ν: 3301, 3196, 3071, 2921, 2855, 1651, 1594, 1549, 1487, 1454, 1414, 1364, 1343, 1308, 1281, 1263, 1211, 1139, 1109, 1012, 950, 875, 820 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 0.91 (s, 3H, CH₃), 1.46–1.54 [complex signal, 2H, 10(13)-H_a] 1.60–1.68 [complex signal, 2H, 10(13)-H_b], 1.78–1.81 (complex signal, 2H, 8-H₂), 1.90 (s, 3H, CH₃CO), 1.98–2.06 [complex signal, 2H, 6(12)-H_a], 2.09–2.16 [complex signal, 2H, 6(12)-H_b] 3.00 (tm, J = 6.0 Hz, 1H, 5-H or 11-H], 3.04 (tm, J = 6.0 Hz, 1H, 11-H or 5-H], 5.19 (broad s., 1H, NH), 6.97 (dd, J = 7.2 Hz, J’ = 1.6 Hz, 1H, 3-H), 7.02–7.04 [complex signal, 2H, 1(4)-H]. ¹³C-NMR (100.6 MHz, CDCl₃) δ: 24.9 (CH₃, CH₃CO), 32.3 (CH₃, C9-CH₃), 33.7 (C, C9), 38.9 (CH₂, C12 or C6), 39.0 (CH₂, C6 or C12), 40.5 (CH, C11 or C5), 40.9 (CH, C5 or C11), 41.1 (CH₂, C10 or C13), 41.2 (CH₂, C13 or C10), 47.3 (CH₂, C8), 54.4 (C, C7), 126.2 (CH, C1 or C4), 128.1 (CH, C4 or C1), 129.6 (CH, C3), 131.6 (C, C2), 144.6 (C, C4a), 148.2 (C, C11a), 169.3 (C, NHCO). HRMS-ESI+ m/z [M-H]⁺ calcd for [C₁₈H₂₃ClNO]⁺: 304.1463, found: 304.1460.

4.1.15. 2-chloro-9-methyl-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-amine hydrochloride (20)

To a solution of acetamide 19 (190 mg, 0.63 mmol) in 2-propanol (6 mL), H₂O (8 mL) and conc. HCl (4 mL) were added, and the mixture was stirred under reflux for 6 days. The solution was cooled and washed with ethyl acetate (3 × 4 mL). The aq phase was basified with 1 N NaOH solution (10 mL) and extracted with ethyl acetate (4 × 15 mL). The combined organic layers were dried over anh Na₂SO₄, filtered and concentrated in vacuo to give 20 as a light brown solid. Its hydrochloride was obtained by adding an excess of HCl / Et₂O to a solution of the amine in ethyl acetate, followed by filtration of the resulting white precipitate (36 mg, 19% yield). The analytical sample was obtained by crystallization from MeOH / Et₂O, mp >250 °C. IR (KBr disk) ν: 3200–2500 (2990, 2950, 2916, 2861), 2058, 1597, 1570, 1509, 1488, 1454, 1416, 1380, 1365, 1302, 1256, 1217, 1155, 1133, 1110, 1093, 1032, 1000, 948, 875, 820, 771, 673 cm⁻¹. ¹H-NMR (400 MHz, MeOD) δ: 1.00 (s, 3H, CH₃), 1.52–1.60 [complex signal, 2H, 10(13)-H_a], 1.65–1.68 (complex signal, 2H, 8-H₂), 1.68–1.74 [complex signal, 2H, 10(13)-H_b], 1.82–1.90 [complex signal, 2H, 6(12)-H_a], 1.99–2.07 [complex signal, 2H, 6(12)-H_b], 3.18 (tm, J = 6.0 Hz, 1H, 5-H or 11-H], 3.22 (tm, J = 6.0 Hz, 1H, 11-H or 5H], 7.09–7.102 [complex signal, 2 H, 3(4)-H], 7.14 (broad signal, 1H, 1-H). ¹³C-NMR (100.6 MHz, MeOD) δ : 32.2 (CH₃, C9-CH₃), 34.7 (C, C9), 39.0 (CH₂, C12 or C6), 39.1 (CH₂, C6 or C12), 41.0 (CH, C11 or C5), 41.3 (CH, C5 or C11 and CH₂, C10 or C13), 41.4 (CH₂, C13 or C10), 47.0 (CH₂, C8), 55.3 (C, C7), 127.7 (CH, C3 or C4), 129.1 (CH, C1), 130.9 (CH, C4 or C3), 133.2 (C, C2), 145.2 (C, C4a) , 148.5 (C, C11a). HRMS-ESI+ m/z [M+]⁺ calcd for [C₁₆H₂₁ClN]⁺: 262.1357, found: 262.1359.

4.1.16. 2-fluoro-5,6,8,9-tetrahydro-7H-5,9-propanobenzo[7]annulene-7,11-dione (25).

1. 4-fluorophthalaldehyde (1 g, 6.57 mmol) was dissolved in methanol (18 mL) then dimethyl-1,3-acetonedicarboxylate (2.33 g, 13.4 mmol) and diethylamine (2 drops) were added. The yellow solution was brought to reflux for 1.5 hours. The solution was allowed to cool to rt and more diethylamine (2 drops) was added. The solution was left at 4 °C overnight and then the solid was filtered off, obtaining the tetraester as a white solid (675 mg, 22% yield). Crude was used in the next step without further purification.

2. The aforementioned tetraester (675 mg, 1.45 mmol) was mixed with conc. HCl (1 mL) and glacial acetic acid (4 mL). The yellow solution was heated to reflux for 12 hours. The solvent was removed under vacuum and the solid residue was diluted with hot diethyl ether (10 mL) over 15 minutes and the solution was stored at 4°C overnight. The residue was filtered off obtaining a white solid (250 mg) that was constituted by a mixture of diketone 25 and its hydrate in an approximate ratio of 1:2. The pure diketone was obtained as a white solid using 7 mL of toluene as a solvent in a Dean-Stark apparatus for 24 hours (200 mg, 59% yield), mp 105–107 °C. IR (NaCl disk) ν: 2923, 2848, 1710, 1607, 1593, 1490, 1428, 1380, 1346, 1253, 1208, 1153, 1119, 1074, 985, 944, 865, 806 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 2.68–2.78 [complex signal, 4H, 6(12)-H_a, 8(10)-H_a], 2.84–2.93 [complex signal, 4H, 6(12)-H_b, 8(10)-H_b], 3.31 (tt, J = J’ = 4.4 Hz, 1H, 5-H or 9-H), 3.37 (tt, J = J’ = 4.4 Hz, 1H, 9-H or 5-H), 7.01 (td, J = 8.0 Hz, J’ = 2.8 Hz, 1H, 3-H), 7.06 (dd, J = 9.2 Hz, J’ = 2.8 Hz, 1H, 1-H), 7.29 (dd, J = 8.0 Hz, J’ = 5.6 Hz, 1H, 4-H). ¹³C-NMR (100.6 MHz, CDCl₃) δ: 37.3 (CH) and 38.1 (CH), (C5 and C9), 49.0 (CH₂) and 49.3 (CH₂) [C6(12) and C8(10)], 115.0 (CH, d, J_C-F = 21 Hz, C3), 115.8 (CH, d, J_C-F = 22 Hz, C1), 130.7 (CH, d, J_C-F = 8.0 Hz, C4), 138.9 (C, C4a), 145.2 (C, d, J_C-F = 7.1 Hz, C9a) , 162.4 (C, d, J_C-F = 247.3 Hz, C2), 208.4 [C, C7(11)-CO]. HRMS-ESI+ m/z [M+]⁺ calcd for [C₁₄H₁₄FO₂]⁺: 233.0972, found: 233.0967.

4.1.17. 2,3-dimethoxy-5,6,8,9-tetrahydro-7H-5,9-propanobenzo[7]annulene-7,11-dione (26)

1. 4,5-dimethoxyphthalaldehyde (6.54 g, 33.7 mmol) was dissolved in methanol (130 mL) then dimethyl-1,3-acetonedicarboxylate (11.7 g, 67.4 mmol) and diethylamine (19 drops) were added. The brown solution was brought to reflux for 1.5 hours. The brown suspension was allowed to cool to rt and more diethylamine (19 drops) was added. The suspension was left at 4 °C over the weekend and was then filtered under vacuum. The white residue was dried under vacuum (6.84 g, 40% yield) and was used in the next step without further purification.

2. The aforementioned tetraester (6.84 g, 13.5 mmol) was mixed with conc. HCl (10 mL) and glacial acetic acid (35 mL). The yellow paste was heated to reflux for 24 hours. The resulting orange suspension was allowed to cool to rt and was then concentrated under vacuum. The brown solid was dissolved in toluene and the brown solution was heated to reflux in a Dean-Stark set up for 24 hours. The yellow solution was allowed to cool to rt and was then concentrated under vacuum. The resulting brown solid was washed twice with hot methanol and was dried again to obtain diketone 26 as a brown solid (2.80 g, 76% yield), mp 236–237 °C. IR (NaCl disk) ν: 2952, 2840, 1698, 1605, 1516, 1467, 1451, 1416, 1355, 1336, 1254, 1221, 1192, 1162, 1025, 1002, 880, 811 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 2.73 [dd, J = 15.2 Hz, J’ = 4.0 Hz, 4H, 6(8,10,12)-H_a], 2.85 [dd, J = 15.2 Hz, J’ = 4.4 Hz, 4H, 6(8,10,12)-H_b], 3.27 [tt, J = 4.8 Hz, J’ = 3.6 Hz, 2H, 5(9)-H], 3.90 (s, 6H, OCH₃), 6.82 [s, 2H, 1(4)-H]. ¹³C-NMR (100.6 MHz, CDCl₃) δ: 37.7 [CH, C5(9)], 49.4 [CH₂, C6(8,10,12)], 56.3 [CH₃,OCH₃], 112.4 [CH, C1(4)], 135.2 [C, C4a(9a)], 148.5 [C, C2(3)], 209.1 [C, C7(11)]. HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₆H₁₉O₄]⁺: 275.1278, found: 275.1279.

4.1.18. 1-fluoro-5,6,8,9-tetrahydro-7H-5,9-propanobenzo[7]annulene-7,11-dione (27)

1. 3-fluorophthalaldehyde (11.2 g, 73.6 mmol) was dissolved in methanol (220 mL) then dimethyl-1,3-acetonedicarboxylate (25.6 g, 147 mmol) and diethylamine (33 drops) were added. The yellow solution was brought to reflux for 1.5 hours. The solution was allowed to cool to rt and more diethylamine (33 drops) was added. The solution was left at 4 °C overnight and was then filtered under vacuum. The white crystals were dried under vacuum and the tetraester was used in the next step without further purification (18.5 g, 54% yield).

2. The aforementioned tetraester (18.5, 39.9 mmol) was mixed with conc. HCl (31 mL) and glacial acetic acid (103 mL). The yellow solution was heated to reflux for 24 hours. The pale-yellow solution was allowed to cool to rt and was then concentrated under vacuum. The off-white solid was dissolved in toluene and the yellow solution was heated to reflux in a Dean-Stark set up for 24 hours. The yellow solution was allowed to cool to rt and was then concentrated under vacuum to give 27 as a an off-white solid (8.73 g, 94% yield), mp > 150 °C. IR (NaCl disk) ν: 2940, 2908, 1701, 1619, 1585, 1468, 1421, 1370, 1304, 1245, 1222, 1203, 1072, 1052, 988, 931, 897, 789, 746 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 2.69 [dd, J = 15.6 Hz, J’ = 3.6 Hz, 2H, 8(10)-H_a], 2.75 [dd, J = 15.6 Hz, J’ = 4.0 Hz, 2H, 6(12)-H_a], 2.90 [dd, J = 15.6 Hz, J’ = 4.8 Hz , 4H, 6(12)-H_b, 8(10)-H_b], 3.42 (m, 1H, 5-H), 4.00 (tt, J = 4.4 Hz, J’ = 4.0 Hz, 1H, 9-H), 7.08–7.15 [complex signal, 2H, 2(4)-H], 7.29 (m, 1H, 3-H). ¹⁹F-NMR (376 MHz, CDCl₃) δ: −117.3 (t, J = 7.52 Hz). ¹³C-NMR (100.6 MHz, CDCl₃) δ: 26.4 (CH, d, J_C-F = 6.4 Hz, C9), 37.9 (CH, d, J_C-F = 2.5 Hz, C5), 48.6 [CH₂, C8(10)], 49.0 [CH₂, C6(12)], 115.1 (CH, d, J_C-F = 24.1 Hz, C2), 124.3 (CH, d, J_C-F = 3.2 Hz, C4), 129.3 (CH, d, J_C-F = 9.2 Hz, C3), 130.0 (C, d, J_C-F = 13.8 Hz, C9a), 146.0 (C, d, J_C-F = 1.9 Hz, C4a), 159.5 (C, d, J_C-F = 245.8 Hz, C1), 208.5 (C, CO). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₄H₁₄FO₂]⁺: 233.0972, found: 233.0976.

4.1.19. 2-methoxy-5,6,8,9-tetrahydro-7H-5,9-propanobenzo[7]annulene-7,11-dione (28)

1. 4-methoxyphthalaldehyde (10.2 g, 62.1 mmol) was dissolved in methanol (380 mL) then dimethyl-1,3-acetonedicarboxylate (21.6 g, 124 mmol) and diethylamine (28 drops) were added. The dark red solution was brought to reflux for 1.5 hours. The solution was allowed to cool to rt and more diethylamine (22 drops) was added. The solution was left at 4 °C over 5 days and was then filtered under vacuum. The red residue was dried under vacuum (19.4 g, 66% yield) and was used in the next step without further purification.

2. The aforementioned tetraester (250 mg, 0.53 mmol) was mixed with conc. HCl (0.4 mL) and glacial acetic acid (1.4 mL). The white suspension was heated to reflux for 24 hours. The dark red solution was allowed to cool to rt and was then concentrated under vacuum. The red oil was dissolved in toluene and the red solution was heated to reflux in a Dean-Stark set up for 24 hours. The red solution was allowed to cool to rt and was then concentrated under vacuum to give 28 as a red solid (125 mg, 98% yield), mp 157–158 °C. IR (NaCl disk) ν: 2941, 2910, 2837, 1701, 1610, 1585, 1504, 1431, 1414, 1370, 1321, 1300, 1266, 1166, 1094, 1033, 989 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 2.68–2.78 [complex signal, 4H, 6(12)-H_a, 8(10)-H_a], 2.82–2.91 [complex signal, 4H, 6(12)-H_b, 8(10)-H_b], 3.27 (tt, J = 4.4 Hz, J’ = 4.0 Hz, 1H, 9-H), 3.32 (tt, J = 4.4 Hz, J’ = 4.0 Hz, 1H, 5-H), 3.84 (s, 3H, OCH₃), 6.83 (dd, J = 8.4 Hz, J’ = 2.8 Hz, 1H, 3-H), 6.87 (d, J = 2.8 Hz, 1H, 1-H), 7.23 (d, J = 8.4 Hz, 1H, 4-H). ¹³C-NMR (100.6 MHz, CDCl₃) δ: 37.2 (CH, C5), 38.4 (CH, C9), 49.2 (CH₂) and 49.6 (CH₂), C6(12) and C8(10), 55.5 (CH₃, OCH₃), 112.9 (CH, C3), 114.7 (CH, C1), 130.2 (CH, C4), 135.2 (C, C4a), 144.2 (C, C9a), 160.0 (C, C2), 209.1 [C, C7(11)-CO]. HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₅H₁₇O₃]⁺: 245.1172, found: 245.1180.

4.1.20. 2-fluoro-7,11-dimethylene-6,7,8,9-tetrahydro-5H-5,9propanobenzo[7]annulene (29)

A suspension of sodium hydride, 60% dispersion in mineral oil (85 mg, 2.13 mmol), in anh DMSO (2 mL) was heated to 90 °C for 45 minutes under N₂ atmosphere. After the reaction mixture was tempered, a solution of triphenylmethylphophonium iodide (1.48 g, 3.66 mmol) in anh DMSO (3.5 mL) was added, and the resulting yellow solution was stirred at rt for 20 minutes. Then a suspension of diketone (200 mg, 0.86 mmol) in anh DMSO (2 mL) was added and the obtained solution was heated to 90 °C overnight. The resulting black solution was allowed to cool to rt and then poured into H₂O (20 mL). Hexane (20 mL) was added and the phases were separated. The aq phase was extracted with further hexane (4 × 20 mL) and the combined organic phases were washed with brine (20 mL), dried over anh Na₂SO₄, filtered and concentrated in vacuo. The crude was purified by Combiflash^® in silica gel using 100% hexane to give diene 29 as a white solid (118 mg, 60% yield), mp 108–109 °C. IR (NaCl disk) ν: 3072, 2985, 2921, 2844, 1639, 1612, 1592, 1494, 1451, 1444, 1363, 1246, 1162, 1135, 1095, 1048, 974, 951, 930, 887, 820, 716, 658, 638, 598, 528 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 2.41–2.51 [complex signal, 4H, 6(12)-H_a, 8(10)-H_a], 2.55–2.62 [complex signal, 4H, 6(12)-H_b, 8(10)-H_b], 3.04 (tt, J = J’ = 4.4 Hz, 1H, 5-H or 9-H), 3.09 (tt, J = J’ = 4.4 Hz, 1H, 9-H or 5-H), 4.68 [s, 4H, 7(11)-CH₂], 6.75–6.82 [complex signal, 2H, 1(3)-H], 7.02 (m, 1H, 4-H). ¹³C-NMR (100.6 MHz, CDCl₃) δ: 40.7 (CH₂) and 40.9 (CH₂) [C6(12) and C8(10)], 40.9 (CH) and 41.6 (CH) (C5 and C9), 112.8 (CH, d, J_C-F = 20 Hz, C1), 115.4 [complex signal, C7(11)=CH₂, C3), 130.0 (CH, d, J_C-F = 8.0 Hz, C4), 137.1 (C, C4a), 146.3 [C, C7(11)], 162.8 (C, C2). The signal from C9a was not observed.

4.1.21. 2,3-dimethoxy-7,11-dimethylene-6,7,8,9-tetrahydro-5H-5,9-propanobenzo[7]annulene (30)

In a flame dried 3-necked round bottom flask, sodium hydride 60% in mineral oil (1.67 g, 41.8 mmol) was added under N₂ and was dissolved in anh DMSO (22 mL). The grey suspension was heated to 75 °C for 45 min and was then cooled to rt.

Methyltriphenylphosphonium iodide (16.9 g, 41.8 mmol) dissolved in anh DMSO (36 mL) and diketone 29 (2.80 g, 10.2 mmol) dissolved in anh DMSO (23 mL) were sequentially added under N₂. The brown solution was stirred and heated to 90 °C overnight. The black solution was allowed to cool to rt and it was then poured into H₂O (200 mL). The H₂O was extracted with hexane (4 × 200 mL) and the organic layer was dried over anh Na₂SO₄, filtered and concentrated under vacuum. The crude was purified by Combiflash^® in silica gel using a solvent system of 100% hexane. Diene 30 was isolated as a yellow solid (633 mg, 23% yield), mp 74–75 °C. IR (NaCl disk) ν: 3068, 2977, 2913, 2832, 1639, 1606, 1515, 1464, 1450, 1429, 1414, 1358, 1342, 1293, 1261, 1240, 1225, 1191, 1173, 1103, 1023, 956, 931, 889, 804, 656, 634 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 2.48 [dd, J = 14.0 Hz, J’ = 5.2 Hz, 4H, 6(8,10,12)-H_a], 2.59 [dd, J = 14.0 Hz, J’ = 4.4 Hz, 4H, 6(8,10,12)-H_b], 3.00 [tt, J = 4.8 Hz, J’ = 4.4 Hz, 2H, 5(9)-H], 3.86 (s, 6H, OCH₃), 4.67 (s, 4H, 7(11)-C=CH₂), 6.63 [s, 2H, 1(4)-H]. ¹³C-NMR (100.6 MHz, CDCl₃) δ: 41.1 [CH₂, C6(8,10,12)], 41.3 [CH, C5(9)], 56.1 [CH₃, OCH₃], 112.7 [CH, C1(4)], 115.1 [CH₂, C=CH₂], 136.8 [C, C4a(9a)], 146.8 [C, C7(11)], 148.5 [C, C2(3)]. HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₈H₂₃O₂]⁺: 271.1693, found: 271.1688.

4.1.22. 1-fluoro-7,11-dimethylene-6,7,8,9-tetrahydro-5H-5,9-propanobenzo[7]annulene (31)

In a flame dried 3-necked round bottom flask, sodium hydride 60% dispersion in mineral oil (2.82 g, 70.5 mmol) was added under N₂ and was dissolved in anh DMSO (34.8 mL). The grey suspension was heated to 75 °C for 45 min, yielding a green suspension, and was then cooled to rt. To the green suspension, methyltriphenylphosphonium iodide (28.5 g, 70.5 mmol) diluted in anh DMSO (75 mL) and the diketone 27 (4.00 g, 17.2 mmol) diluted in anh DMSO (38 mL) were added sequentially. The resulting mixture was heated at 90 °C overnight. The resulting black solution was allowed to cool to rt and it was then poured into H₂O (200 mL). The H₂O was extracted with hexane (4 × 200 mL) and the organic layer was dried over anh Na₂SO₄, filtered and concentrated under vacuum. The crude was purified by Combiflash^® using as eluent a gradient from pure hexane to ethyl acetate / hexane mixture (1 / 9). Diene 31 was isolated as a colourless oil (2.69 g, 69% yield). IR (NaCl disk) ν: 3071, 3033, 2981, 2921, 2838, 1639, 1614, 1583, 1464, 1446, 1429, 1365, 1248, 1155, 1046, 991, 935, 919, 895 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 2.43–2.56 [complex signal, 4H, 6(12)-H_a, 8(10)-H_a], 2.64 [dd, J = 14.4 Hz, J’ = 4.2 Hz, 4H, 6(12)-H_b, 8(10)-H_b], 3.19 (m, 1H, 5-H), 3.79 (tt, J = J’ = 4.8 Hz, 1H, 9-H), 4.72 [d, J = 0.8 Hz, 4H, 7(11)-CH₂], 6.88–6.96 (complex signal, 2H, 2-H, 4-H), 7.08 (m, 1H, 3-H). ¹⁹F-NMR (376 MHz, CDCl₃) δ: −119.9 (t, J = 4.89 Hz). ¹³C-NMR (100.6 MHz, CDCl₃) δ: 29.5 (CH, d, J_C-F = 5 Hz, C9), 40.2 (CH₂) and 40.6 (CH₂), C6(12) and C8(10), 41.7 (CH, d, J_C-F = 2.0 Hz, C5), 113.7 (CH, d, J_C-F = 24 Hz, C2), 115.5 [CH₂, C7(11)-CH₂], 124.1 (CH, d, J_C-F = 3.0 Hz, C4), 127.4 (CH, d, J_C-F = 9.1 Hz, C3), 131.0 (C, d, J_C-F = 13.1 Hz, C9a), 146.2 [C, C7(11)], 147.6 (C, d, J_C-F = 2.5 Hz, C4a), 159.7 (C, d, J_C-F = 242.4 Hz, C1). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₆H₁₈F]⁺: 229.1387, found: 229.1392.

4.1.23. 2-methoxy-7,11-dimethylene-6,7,8,9-tetrahydro-5H-5,9-propanobenzo[7]annulene (32)

In a flame dried 3-necked round bottom flask, sodium hydride 60% in mineral oil (2.68 g, 67.0 mmol) was added under N₂ and was dissolved in anh DMSO (35 mL). The grey suspension was heated to 75 °C for 45 min and was then cooled to rt.

Methyltriphenylphosphonium iodide (27.1 g, 67.0 mmol) dissolved in anh DMSO (57 mL) and 28 (4.00 g, 16.4 mmol) dissolved in anh DMSO (38 mL) were sequentially added under N₂. The brown solution was stirred and heated to 90 °C overnight. The black solution was allowed to cool to rt and it was then poured into H₂O (200 mL). The H₂O was extracted with hexane (4 × 200 mL) and the organic layer was dried over anh Na₂SO₄, filtered and concentrated under vacuum. The crude was purified by Combiflash^® in silica gel using a solvent system of 100% hexane. Diene 32 was isolated as a yellow waxy oil (1.50 g, 38% yield), mp 68–69 °C. IR (NaCl disk) ν: 3068, 2979, 2911, 2833, 1639, 1609, 1580, 1501, 1464, 1449, 1431, 1363, 1313, 1260, 1203, 1172, 1152, 1109, 1034, 955, 929, 889, 809, 661, 613 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 2.43–2.53 [complex signal, 4H, 6(12)-H_a, 8(10)-H_a], 2.55–2.64 [complex signal, 4H, 6(12)-H_b, 8(10)-H_b], 3.03 (tt, J = 4.8 Hz, J’ = 4.4 Hz, 1H, 5-H), 3.07 (tt, J = 4.8 Hz, J’ = 4.4 Hz, 9-H), 3.79 (s, 3H, CH₃), 4.68 [s, 4H, 7(11)-CH₂], 6.65 (dd, J = 8.0 Hz, J’ = 2.8 Hz, 1H, 3-H), 6.66 (d, J = 2.8 Hz, 1H, 1-H), 7.00 (d, J = 8.0 Hz, 1H, 4-H). ¹³C-NMR (100.6 MHz, CDCl₃) δ: 40.7 (CH, C9), 40.9 (CH₂) and 41.2 (CH₂) C6(8,10,12), 41.9 (CH, C5), 55.2 (CH₃, OCH₃), 110.8 (CH, C3), 114.9 (CH, C1), 115.1 [CH₂, C7(11)-CH₂], 129.6 (CH, C4), 136.9 (C, C4a), 146.0 (C, C9a), 146.7 [C, C7(11)], 158.3 (C, C2). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₇H₂₁O]⁺: 241.1587, found: 241.1588.

4.1.24. 2-chloro-N-(2-fluoro-9-methyl-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-yl)acetamide (33)

Chloroacetonitrile (84 μL, δ = 1.19, 1.32 mmol) was added to a solution of 2-fluoro-7,11-dimethylene-6,7,8,9-tetrahydro-5H-5,9-propanobenzo[7]annulene (75 mg, 0.32 mmol) in acetic acid (0.25 mL) and the mixture was cooled to 0−5 °C with an ice bath. Conc. H₂SO₄ (0.11 mL, δ = 1.84, 1.97 mmol) was added dropwise (<10 °C). After the addition, the mixture was allowed to reach rt and was stirred overnight. The solution was added to ice (2 g) and the mixture was stirred at rt for a few minutes. DCM (5 mL) was added, the phases were separated, and the aq phase was extracted with further DCM (2 × 5 mL). The combined organic layers were dried over anh Na₂SO₄, then filtered and evaporated in vacuo to give 33 as a yellow solid (50 mg, 48% yield), mp 141–144 °C. IR (NaCl disk) ν: 3399, 3313, 3067, 2944, 2920, 2851, 1657, 1607, 1591, 1518, 1498, 1451, 1361, 1345, 1252, 1179, 1145, 1086, 1049, 1009, 966, 963, 863, 820 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 0.94 (s, 3H, 9-CH₃), 1.49–1.59 [complex signal, 2H, 10(13)-H_a], 1.62–1.70 [complex signal, 2 H, 10(13)-H_b], 1.81 (m, 1H, 8-H_a), 1.84 (m, 1H, 8-H_b), 2.05–2.19 [complex signal, 4H, 6(12)-H₂], 3.03 (m, 1H, 5-H or 11-H), 3.09 (m, 1H, 11-H or 5-H), 3.93 (s, 2H, CH₂Cl), 6.30 (broad s, 1H, NH), 6.73 (dd, J = 8.4 Hz, J’ = 2.8 Hz, 1H, 1-H), 6.76 (m, 1H, 3-H), 7.00 (dd, J = 8.0 Hz, J’ = 6.0 Hz, 1 H, 4-H). ¹³C-NMR (100.6 MHz, CDCl₃) δ: 32.2 (CH₃, C9-CH₃), 33.7 (C, C9), 38.5 (CH₂) and 38.9 (CH₂) (C6 and C12), 40.3 (CH, C11), 40.93 (CH, C5), 40.97 (CH₂) and 41.26 (CH₂) (C10 and C13), 43.1 (CH₂, CH₂Cl), 47.0 (CH₂, C8), 55.0 (C, C7), 112.5 (CH, d, J_C-F = 20 Hz, C1), 115.0 (CH, d, J_C-F = 21Hz, C3), 129.6 (CH, d, J_C-F = 8 Hz, C4), 141.9 (C, d, J_C-F = 3 Hz, C4a), 148.2 (C, d, J_C-F = 6.9 Hz, C11a), 161.4 (C, d, J_C-F = 244 Hz, C2), 164.7 (C, NHCO). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₈H₂₂ClFNO]⁺: 322.1368, found: 322.137.

4.1.25. 2-chloro-N-(2,3-dimethoxy-9-methyl-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-yl)acetamide (34)

Diene 30 (498 mg, 1.84 mmol) was dissolved in glacial acetic acid (1.6 mL) and chloroacetonitrile (467 μL, δ = 1.19, 7.38 mmol) was added. The solution was cooled to 0 °C and conc. H₂SO₄ (601 μL, δ = 1.84, 11.0 mmol) was added dropwise with care not to raise above 10 °C. The solution was allowed to warm to rt and was left stirring overnight, then it was poured over ice (2 g) and extracted with DCM (3 × 10 mL). The organic layer was dried over anh Na₂SO₄, then filtered and concentrated under vacuum. The crude was crystallized from DCM to obtain 34 as a white solid (501 mg, 75% yield), mp 204–205 °C. IR (NaCl disk) ν: 3306, 2941, 2907, 2861, 2838, 1666, 1605, 1516, 1467, 1452, 1415, 1381, 1361, 1345, 1293, 1252, 1231, 1191, 1168, 1092, 1021, 948, 861, 802 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 0.93 (s, 3H, 9-CH₃), 1.55 [d, J = 13.2 Hz, 2H, 10(13)-H_a], 1.65 [dd, J = 13.2 Hz, J’ = 6 Hz, 2H, 10(13)-H_b], 1.83 (s, 2H, 8-H₂), 2.07 [d, J = 12.8 Hz, 2H, 6(12)-H_a], 2.15 [ddd, J = 12.8 Hz, J’ = 6.4 Hz, J” = 2.0 Hz, 2H, 6(12)-H_b], 3.00 [t, J = 6.4 Hz, 2H, 5(11)-H], 3.84 (s, 6H, OCH₃), 3.93 (s, 2H, CH₂-Cl), 6.29 (broad s., 1H, NH), 6.60 [s, 2H, 1(4)-H]. ¹³C-NMR (100.6 MHz, CDCl₃) δ: 32.3 (CH₃, C9-CH₃), 33.7 (C, C9), 39.0 [CH₂, C6(12)], 40.7 [CH, C5(11)], 41.4 [CH₂, C10(13)], 43.1 (CH₂, CH₂-Cl), 47.1 (CH₂, C8), 55.1 (C, C7), 56.2 (CH_3, OCH₃), 112.4 [CH, C1(4)], 138.4 [C, C4a(11a)], 146.8 [C, 2(3)], 164.7 (C, NHCO). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₂₀H₂₇ClNO₃]⁺: 364.1674, found: 364.1674.

4.1.26. 2-chloro-N-(1-fluoro-9-methyl-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-yl)acetamide (35)

Chloroacetonitrile (2.7 mL, δ = 1.19, 42.7 mmol) was added to a solution of diene 31 (2.37 g, 10.4 mmol) in acetic acid (7 mL) and the mixture was cooled to 0–5 °C with an ice bath. Conc. H₂SO₄ (3.4 mL, δ = 1.84, 63.8 mmol) was added dropwise (<10 °C). After the addition, the mixture was allowed to reach rt and stirred overnight. The solution was added to ice (10 g) and the mixture was stirred at rt for few minutes. DCM (100 mL) was added, the phases were separated, and the aqueous phase was extracted with further DCM (2 × 100 mL). The combined organic layers were dried over anh Na₂SO₄, filtered and evaporated in vacuo to give 35 as a white solid (2.28 g, 68% yield), mp 154–155 °C. IR (NaCl disk) ν: 3402, 3308, 3073, 2947, 2911, 2863, 2840, 1660, 1613, 1583, 1529, 1463, 1363, 1348, 1312, 1242, 1186, 1155, 1069, 1053, 979, 875, 798, 748 cm⁻¹. ¹H-RMN (400 MHz, CDCl₃) δ: 0.93 (s, 3H, 9-CH₃), 1.45–1.61 (complex signal, 2H, 10-H_a, 13-H_a), 1.63–1.71 (complex signal, 2H, 10-H_b, 13-H_b), 1.80 (dt, J = 12.0 Hz, J’ = 2.0 Hz, 1H, 8-H_a), 1.87 (dt, J = 12.0 Hz, J’ = 2.0 Hz, 1H, 8-H_b), 2.08–2.11 (complex signal, 2H, 6-H₂ or 12-H₂), 2.14–2.17 (complex signal, 2H, 12-H₂, 6-H₂), 3.13 (m, 1H, 5-H), 3.73 (tm, J = 6.0 Hz, 1H, 11-H), 3.93 (s, 2H, CH₂Cl), 6.31 (broad s., 1H, NH), 6.84 (d, J = 8.0 Hz, 1H, 4-H), 6.86 (dd, J = 8.4 Hz, J’ = 1.2 Hz, 1H, 2-H) 7.01 (m, 1 H, 3-H). ¹³C-NMR (100.6 MHz, CDCl₃) δ : 28.6 (CH, d, J_C-F = 5.7 Hz, C11), 32.2 (CH₃, C9-CH₃), 33.81 (C, C9), 38.3 (CH₂) and 38.4 (CH₂) (C6 and C12), 40.6 (CH₂) and 40.96 (CH₂) (C10 and C13), 41.03 (CH, d, J = 2.3 Hz, C5), 43.1 (CH₂, CH₂Cl), 47.0 (CH₂, C8), 55.0 (C, C7), 113.5 (CH, d, J_C-F = 24.8 Hz, C2), 123.6 (CH, d, J_C-F = 3.1 Hz, C4), 127.2 (CH, d, J_C-F = 9.3 Hz, C3), 132.6 (CH, d, J_C-F = 13.2 Hz, C11a), 149.3 (C, d, J_C-F = 2.2 Hz, C4a), 159.3 (C, d, J_C-F = 243.5 Hz, C1), 164.7 (C, NHCO). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₈H₂₂ClFNO]⁺: 322.1368, found: 322.1374.

4.1.27. 2-chloro-N-(2-methoxy-9-methyl-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-yl)acetamide (36)

Diene 32 (1.50 g, 6.24 mmol) was dissolved in glacial acetic acid (4.8 mL) and chloroacetonitrile (1.58 mL, δ = 1.19, 24.9 mmol) was added. The solution was cooled to 0 °C and conc. H₂SO₄ (2.00 mL, δ = 1.84, 37.5 mmol) was added dropwise with care not to raise above 10 °C. The dark red solution was allowed to warm to rt and was left stirring overnight. The dark red solution was poured over ice (5 g) and it was then extracted with DCM (3 × 20 mL). The organic layer was washed with 10 N NaOH solution (1 × 20 mL), dried over anh Na₂SO₄, then filtered and concentrated under vacuum to give a brown solid (1.18 g, 57% yield), mp 144–145 °C. IR (NaCl disk) ν: 3403, 3304, 3062, 2997, 2945, 2905, 2860, 2838, 1662, 1609, 1582, 1528, 1499, 1454, 1382, 1361, 1311, 1267, 1242, 1198, 1180, 1154, 1043, 1013, 955, 873 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 0.93 (s, 3H, 9-CH₃), 1.481.59 [complex signal, 2H, 10(13)-H_a], 1.61–1.70 [complex signal, 2H, 10(13)-H_b], 1.82 (dm, J = 12.0 Hz, 1H, 8-H_a), 1.86 (dm, J = 12.0 Hz, 1H, 8-H_b), 2.00–2.10 [complex signal, 2H, 6(12)-H_a], 2.10–2.21 [complex signal, 2H, 6(12)-H_b), 2.99–3.09 [complex signal, 2H, 5(11)-H], 3.77 (s, 3H, OCH₃), 3.92 (s, 2H, CH₂-Cl), 6.60 (dd, J = 8.0 Hz, J’ = 2.8 Hz, 1H, 3-H), 6.63 (d, J = 2.8 Hz, 1-H), 6.97 (d, J = 8.0 Hz, 1H, 4-H). ¹³C-NMR (100.6 MHz, CDCl₃) δ: 32.3 (CH₃, C9-CH₃), 33.7 (C, C9), 38.8 (CH₂) and 39.2 (CH₂) (C6 and C12), 40.1 (CH) and 41.2 (CH) (C5 and C11), 41.2 (CH₂) and 41.6 (CH₂) (C10 and C13), 43.0 (CH₂, CH₂-Cl), 47.0 (CH₂, C8), 54.8 (C, C7), 55.3 (CH_3, OCH₃), 110.6 (CH, C3), 114.3 (CH, C1), 129.2 (CH, C4), 138.5 (C, C4a), 147.5 (C, C11a), 158.1 (C, C2), 164.6 (C, NHCO). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₉H₂₅ClNO₂]⁺: 334.1568, found: 334.1569.

4.1.28. 2-fluoro-9-methyl-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-amine hydrochloride (37).

Thiourea (15 mg, 0.20 mmol) and glacial acetic acid (0.1 mL) were added to a solution of 33 (50 mg, 0.16 mmol) in absolute ethanol (3 mL) and the mixture was heated at reflux overnight. The resulting suspension was then tempered to rt and concentrated under vacuum. Crude was partitioned in H₂O (2 mL) and DCM (3 mL) and the phases were separated. Then the pH of the aq phase was adjusted to 12 with 5 N NaOH solution and extracted with further DCM (3 × 3 mL). The combined organic layers were dried over anh Na₂SO₄, filtered and concentrated in vacuo to obtain amine 37. Its hydrochloride was obtained by adding an excess of HCl / Et₂O to a solution of the amine in ethyl acetate, followed by filtration of the resulting white precipitate (16 mg, 36% yield), mp > 300 °C (dec.) IR (KBr disk) ν: 3200–2500 (2983, 2945, 2917, 2867), 2059, 1612, 1595, 1501, 1456, 1444, 1431, 1379, 1364, 1302, 1283, 1256, 1246, 1186, 1157, 1143, 1132, 1030, 1004, 962, 863, 814 cm⁻¹. ¹H-NMR (400 MHz, MeOD) δ: 1.0 (s, 3H, 9-CH₃), 1.51–1.61 [complex signal, 2H, 10(13)H_a], 1.68 (s, 2H, 8-H₂), 1.68–1.75 [complex signal, 2H, 10(13)-H_b], 1.82–1.91 [complex signal, 2 H, 6(12)-H_a], 2.00–2.08 [complex signal, 4H, 6(12)-H_b], 3.18 (tm, J = 6.2 Hz, 1H, 5-H or 11-H), 3.23 (tm, J = 6.2 Hz, 1H, 11-H or 5-H), 6.81 (td, J = 8.4 Hz, J’ = 2.8 Hz, 3-H), 6.87 (dd, J = 9.4 Hz, J’ = 2.8 Hz, 1-H), 7.11 (dd, J = 8.4 Hz, J’ = 5.6 Hz, 1 H, 4-H). ¹³C-NMR (100.6 MHz, MeOD) δ: 32.2 (CH₃, C9-CH₃), 34.7 (C, C9), 39.0 (CH₂) and 39.3 (CH₂) (C6 and C12), 40.8 (CH) and 41.4 (CH) (C5 and C11), 41.3 (CH₂) and 41.6 (CH₂) (C10 and C13), 47.1 (CH₂, C8), 55.3 (C, C7), 113.8 (CH, d, J_C-F = 20 Hz, C3), 115.9 (CH, d, J_C-F = 22 Hz, C1), 131.0 (CH, d, J_C-F = 8.1 Hz, C4), 142.5 (C, d, J_C-F = 3 Hz, C4a), 148.8 (C, d, J_C-F = 7.1 Hz, C11a), 162.9 (C, d, J_C-F = 243 Hz, C2). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₆H₂₁FN]⁺: 246.1653, found: 246.1649.

4.1.29. 2,3-dimethoxy-9-methyl-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-amine hydrochloride (38)

Chloroacetamide 34 (436 mg, 1.20 mmol) was dissolved in abs. ethanol (23 mL) and glacial acetic acid (0.75 mL). Thiourea (109 mg, 1.44 mmol) was added, and the solution was left stirring overnight at reflux. The yellow solution was concentrated under vacuum and the solid was dissolved in MeOH (10 mL). HCl / Et₂O was added, and the solution was concentrated to dryness. The off-white solid was washed with DCM (2 × 3 mL). The combined supernatants were concentrated under vacuum to obtain the hydrochloride of 38 as a white solid (315 mg, 81% yield), mp > 200 °C (dec.). IR (KBr disk) ν: 3200–2500 (2993, 2918, 2831), 2047, 1701, 1606, 1517, 1451, 1416, 1386, 1365, 1327, 1310, 1291, 1252, 1237, 1192, 1174, 1131, 1098, 1031, 973, 950, 863, 798, 586, 543 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 0.95 (s, 3H, 9-CH₃), 1.55 [d, J = 13.6 Hz, 2H, 10(13)-H_a], 1.63 [dd, J = 13.6 Hz, J’ = 6.0 Hz, 2H, 10(13)-H_b], 1.80 (s, 2H, 8-CH₂), 2.04 [d, J = 12.6 Hz, 2H, 6(12)-H_a], 2.15 [dd, J = 12.6 Hz, J’ = 6.0 Hz, 2H, 6(12)-H_b], 3.01 [t, J = 5.2 Hz, 2H, 5(11)-H], 3.83 (s, 6H, OCH₃), 6.58 [s, 2H, 1(4)-H], 8.38 (broad s., 3H, 7C-NH₃⁺). ¹³C-NMR (100.6 MHz, CDCl₃) δ: 31.9 (CH₃, C9-CH₃), 33.9 (C, C9), 38.6 [CH₂, C6(12)], 40.1 [CH, C5(11)], 40.7 [CH₂, C10(13)], 46.5 (CH₂, C8), 55.7 (C, C7), 56.2 (CH_3, OCH₃), 112.5 [CH, C1(4)], 137.4 [C, C4a(11a)], 147.1 [C, 2(3)]. HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₈H₂₆NO₂]⁺: 288.1958, found: 288.1954.

4.1.30. 1-fluoro-9-methyl-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-amine hydrochloride (39)

Thiourea (284 mg, 3.73 mmol) and glacial acetic acid (2 mL) were added to a solution of 35 (1.00 g, 3.11 mmol) in abs. ethanol (54 mL) and the mixture was heated at reflux overnight. The resulting suspension was then tempered to rt and concentrated under vacuum. Crude was partitioned in H₂O (40 mL) and DCM (30 mL) and the phases were separated. Then the pH of the aq phase was adjusted to 12 with 5 N NaOH solution and extracted with further DCM (3 × 30 mL). The combined organic layers were dried over anh Na₂SO₄, then filtered and concentrated in vacuo. The hydrochloride of 39 was obtained by adding an excess of HCl / Et₂O to a solution of the amine in ethyl acetate followed by filtration of the resulting white precipitate (732 mg, 84% yield). The analytical sample was obtained by crystallization from methanol, mp >200 °C (dec.). IR (KBr disk) ν: 3200–2500 (2945, 2717, 2586), 2060, 1677, 1608, 1584, 1511, 1464, 1380, 1366, 1317, 1303, 1248, 1214, 1199, 1165, 1132, 1071, 1052, 1032, 1000, 977, 946, 885, 877, 854, 798, 747, 623 cm⁻¹. ¹H-NMR (400 MHz, MeOD) δ: 1.0 (s, 3H, CH₃), 1.50 [d, J = 14.0 Hz, 1H, 10-H_a or13-H_a], 1.57 (d, J = 14.0 Hz, 1H, 13-H_a or 10-H_a), 1.72 (s, 2H, 8-H₂), 1.68–1.75 [complex signal, 2H, 10(13)-H_b], 1.79 (d, J = 12.8 Hz, 1H, 6-H_a or 12-H_a) 1.86 (d, J = 12.8 Hz, 1H, 12-H_a or 6-H_a), 2.00–2.08 (complex signal, 2H, 6-H_b, 12-H_b), 3.27 (m, 1H, 5-H), 3.80 (t, J = 6.2 Hz, 11-H), 6.90 (ddd, J = 8.4 Hz, J’ = 8.0 Hz, J” = 1.4 Hz, 1H, 2-H), 6.92 (m, 1H, 4-H), 7.09 (ddd, J = 8.4 Hz, J’ = 7.6 Hz, J” = 5.6 Hz, 1H, 3-H). ¹³C-NMR (100.6 MHz, MeOD) δ : 29.5 (CH, d, J_C-F = 6 Hz, C11), 32.7 (CH₃, C9-CH₃), 34.7 (C, C9), 38.6 (CH₂) and 39.0 (CH₂) (C6 and C12), 40.9 (CH₂) and 41.3 (CH₂) (C10 and C13), 41.6 (CH, d, J_C-F = 3 Hz, C5), 47.0 (CH₂, C8), 55.3 (C, C7), 114.7 (CH, d, J_C-F = 25 Hz, C2), 125.2 (CH, d, J_C-F = 4 Hz, C4), 129.1 (CH, d, J_C-F = 10 Hz, C3), 132.8 (CH, d, J_C-F = 14.1 Hz, C11a), 149.7 (C, d, J_C-F = 2 Hz, C4a), 160.4 (C, d, J_C-F = 242.4 Hz, C1). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₆H₂₁FN]⁺: 246.1653, found: 246.1649.

4.1.31. 2-methoxy-9-methyl-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-amine hydrochloride (40)

Chloroacetamide 36 (1.10 g, 3.29 mmol) was dissolved in abs. ethanol (60 mL) and glacial acetic acid (2.2 mL). Thiourea (300 mg, 3.94 mmol) was added and the orange solution was left stirring overnight. The orange solution was concentrated under vacuum and the solid was dissolved in MeOH (10 mL). HCl / Et₂O was added, the solution was concentrated, and the off-white solid was extracted with DCM (2 × 3 mL). The combined supernatants were concentrated under vacuum to obtain 40 as a brown solid that was crystallized from DCM / pentane (779 mg, 81% yield), mp > 200 °C (dec.). IR (KBr disk) ν: 3200–2500 (2985, 2942, 2908), 2056, 1735, 1609, 1582, 1499, 1449, 1379, 1364, 1334, 1305, 1268, 1252, 1205, 1170, 1132, 1103, 1040, 1001, 954, 869, 849, 815, 756, 692 cm⁻¹. ¹H-NMR (400 MHz, CD₃OD) δ: 0.99 (s, 3H, 9-CH₃), 1.50–1.59 [complex signal, 2H, 10(13)-H_a], 1.65 (s, 2H, 8-CH₂), 1.65-1.73 [complex signal, 2H, 10(13)-H_b], 1.79–1.89 [complex signal, 2H, 6(12)-H_a], 1.97–2.06 [complex signal, 2H, 6(12)-H_b], 3.10–3.18 [complex signal, 2H, 5(11)-H], 3.74 (s, 3H, OCH₃), 6.64 (dd, J = 8.0 Hz, J’ = 2.8 Hz, 1H, 3-H), 6.67 (d, J = 2.8 Hz, 1H, 1-H), 7.00 (d, J = 8.0 Hz, 4-H). ¹³C-NMR (100.6 MHz, CD₃OD) δ: 32.3 (CH₂, 9-CH₃), 34.6 (C, C9), 39.3 (CH₂) and 39.6 (CH₂) (C6 and C12), 40.8 (CH) and 41.6 (CH) (C5 and C11), 41.7 (CH₂) and 42.0 (CH₂) (C10 and C13), 47.2 (CH₂, C8), 55.5 (C, C7), 55.7 (CH₃, OCH₃) 112.1 (CH, C3), 115.2 (CH, C1), 130.3 (CH, C4), 138.6 (C, C4a), 147.7 (C, C11a), 159.9 (C, C2). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₇H₂₄NO]⁺: 258.1852, found: 258.1862.

4.1.32. 1-fluoro-11-methylene-6,7,8,9-tetrahydro-5H-5,9-propanobenzo[7]annulen-7-one (41)

In a flame dried 3-necked round bottom flask, sodium hydride 60% dispersion in mineral oil (1.01 g, 25.2 mmol) was added under N₂ and was dissolved in anh DMSO (50 mL). The grey suspension was heated to 90 °C for 45 min and was then cooled to rt. After the reaction mixture was tempered, a solution of triphenylmethylphophonium iodide (10.6 g, 26.2 mmol) in anh DMSO (58 mL) was added and the resulting yellow solution was stirred at rt for 20 minutes. Then a suspension of diketone 27 (4.72 g, 20.3 mmol) in anh DMSO (50 mL) was added and the obtained solution was heated to 90 °C overnight. The resulting black solution was allowed to cool to rt and then poured into H₂O (200 mL). The H₂O was extracted with hexane (4 × 200 mL) and the organic layer was dried over anh Na₂SO₄, filtered and concentrated under vacuum. The crude was purified by Combiflash^® using as eluent a gradient from pure hexane to an ethyl acetate / hexane mixture (2.5 / 7.5). The product 41 was isolated as a yellow solid (2.16 g, 46% yield), mp 96 °C. IR (ATR) ν: 2928, 2913, 2895, 2849, 1688, 1612, 1583, 1462, 1432, 1406, 1366, 1247, 1196, 1104, 1049, 1034, 1002, 970, 921, 911, 883, 819, 789, 749, 657 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 2.39 (d, J = 13.2 Hz, 1H, 10-H_a or 12-H_a), 2.46 (d, J = 12.4 Hz, 1H, 10-H_a or 12-H_a), 2.48–2.62 [complex signal, 2H, 6(8)-H_a], 2.64–2.72 [complex signal, 2H, 10(12)-H_b], 2.83–2.91 [complex signal, 2H, 6(8)-H_b], 3.22 (m, 1H, 5-H), 3.80 (m, 1H, 9-H), 4.96 (complex signal, C=CH₂), 6.97 (ddd, J = 8.4 Hz, J’ = 1.6 Hz, 1H, 2-H), 6.99 (d, J = 8.0 Hz, 4-H), 7.15 (ddd, J = 8.4 Hz, J’ = 7.6 Hz, J” = 5.6 Hz, 1H, 3-H). ¹³C-NMR (100.6 MHz, CDCl₃) δ: 28.3 (CH, d, J_C-F = 5 Hz, C9), 40.4 (CH, d, J_C-F = 3 Hz, C5), 41.4 (CH₂) and 41.7 (CH₂) (C10 and C12), 48.1 (CH₂) and 48.5 (CH₂) (C6 and C8), 114.3 (CH, d, J_C-F = 24 Hz, C2), 120.3 (CH₂, 11-CH₂), 124.1 (CH, d, J_C-F = 3 Hz, C4), 128.3 (CH, d, J_C-F = 9 Hz, C3), 131.2 (C, d, J_C-F = 14.1 Hz, C9a), 143.4 (C, C11), 147.5 (C, d, J_C-F = 3 Hz, C4a), 159.5 (C, d, J_C-F = 244.5 Hz, C1), 211.2 [C, C7-CO]. HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₅H₁₆FO]⁺: 231.1180, found: 231.1180.

4.1.33. 2-chloro-N-(1-fluoro-9-hydroxy-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-yl)acetamide (42)

Chloroacetonitrile (0.62 mL, δ = 1.19, 9.77 mmol) was added to a solution of 41 (2.06 g, 8.94 mmol) in DCM (21 mL) and the mixture was cooled to 0–5 °C with an ice bath. Conc. H₂SO₄ (0.75 mL, δ = 1.84, 14.1 mmol) was added dropwise (<10 °C). After the addition, the mixture was allowed to reach rt and stirred overnight. The solution was added to ice (10 g) and the mixture was stirred at rt for a few minutes. DCM (15 mL) was added, the phases were separated, and the aq phase was extracted with further DCM (2 × 15 mL). The combined organic layers were dried over anh Na₂SO₄, then filtered and evaporated in vacuo to give 42 as a white solid (921 mg, 32% yield), mp 150 °C. IR (ATR) ν: 3406, 3272, 3217, 3075, 2926, 2905, 2850, 1661, 1585, 1561, 1466, 1443, 1428, 1409, 1362, 1341, 1311, 1298, 1243, 1218, 1158, 1105, 1037, 991, 974, 891, 884, 791, 744, 734, 679, 625 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 1.67–1.80 [complex signal, 2H, 10(13)-H_a], 1.93–2.00 [complex signal, 2H, 10(13)H_b], 2.02–2.22 [complex signal, 4H, 6(12)-H₂], 3.24 (m, 1H, 5-H), 3.84 (t, J = 6.8 Hz, 1H, 11-H), 3.93 (s, 2H, CH₂Cl), 6.38 (broad s., 1H, NH), 6.86–6.91 [complex signal, 2 H, 2(4)-H], 7.05 (ddd, J = 8.2 Hz, J’ = 7.6 Hz, J” = 5.6 Hz, 1H, 3-H). ¹³C-NMR (100.6 MHz, CDCl₃) δ : 27.8 (CH, d, J_C-F = 6 Hz, C11), 37.9 (CH₂) and 38.0 (CH₂) (C6 and C12), 39.9 (CH, d, J = 3 Hz, C5), 41.9 (CH₂) and 42.2 (CH₂) (C10 and C13), 43.0 (CH₂, CH₂Cl), 48.0 (CH₂, C8), 57.4 (C, C7), 70.8 (C, C9), 113.8 (CH, d, J_C-F = 24 Hz, C2), 123.7 (CH, d, J_C-F = 3 Hz, C4), 127.6 (CH, d, J_C-F = 9 Hz, C3), 131.6 (CH, d, J_C-F = 13.1 Hz, C11a), 148.2 (C, d, J_C-F = 2 Hz, C4a), 159.2 (C, d, J_C-F = 243.5 Hz, C1), 164.9 (C, NHCO). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₇H₂₀ClFNO₂]⁺: 324.1161, found: 324.1162.

4.1.34. 2-chloro-N-(1,9-difluoro-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-yl)acetamide (43)

A solution of chloroacetamide 42 (611 mg, 1.89 mmol) in DCM (10 mL) was cooled to −30 °C. Then, DAST (2.8 mL, 1M solution in DCM, 2.8 mmol) was added and the reaction mixture was stirred at rt overnight. To the resulting solution was added H₂O (10 mL) and the pH adjusted to 12 with 5N NaOH solution. The phases were separated, and the aq phase was extracted with further DCM (3 × 10 mL), and the combined organic layers were dried over anh Na₂SO₄, then filtered and concentrated in vacuo. The crude was purified by Combiflash^® in silica gel using as eluent a gradient from pure hexane to ethyl acetate / hexane mixture (6 / 4) to give 43 as a white solid (420 mg, 69% yield), mp 180 °C. IR (ATR) ν: 3276, 3075, 2964, 2940, 2901, 2858, 1671, 1650, 1584, 1552, 1463, 1442, 1360, 1337, 1331, 1317, 1282, 1242, 1219, 1175, 1143, 1104, 1018, 1066, 1018, 1004, 979, 901, 887, 865, 799, 746, 737, 696, 662 cm⁻¹. ¹H-NMR (400 MHz, CDCl₃) δ: 1.85–1.92 [complex signal, 2H, 6(12)-H_a], 2.01–2.12 [complex signal, 2H, 10(13)-H_a], 2.13–2.24 [complex signal, 4H, 6(10,12,13)-H_b], 2.24–2.35 (complex signal, 2H, 8-H₂), 3.31 (m, 1H, 5-H), 3.90 (m, 1H, 11-H), 3.94 (s, 2H, CH₂Cl), 6.40 (broad s., 1H, NH), 6.88 (m, 1H, 4-H), 6.93 (ddd, J = 8.8 Hz, J’ = 8.4 Hz, J” = 1.6 Hz, 1H, 2-H), 7.07 (ddd, J = 8.4 Hz, J’ = 7.2 Hz, J” = 5.2 Hz, 1H, 3-H). ¹³C-NMR (100.6 MHz, CDCl₃) δ: 27.4 (CH, dd, J_C-F = 13.6 Hz, J’_C-F = 6.4 Hz, C11), 37.8 (CH₂, d, J_C-F = 1.1 Hz, C6 or C12), 37.9 (CH₂, d, J_C-F = 1.0 Hz, C12 or C6), 39.4 (CH, dd, J_C-F = 13.3 Hz, J’_C-F = 2.4 Hz, C5), 39.7 (CH₂, d, J_C-F = 19.5 Hz, C10 or C13), 39.8 (CH₂, d, J_C-F = 20.4 Hz, C13 or C10), 43.0 (CH₂, CH₂Cl), 45.7 (CH₂, d, J_C-F = 19 Hz, C8), 58.1 (C, d, J_C-F = 11 Hz, C7), 93.0 (C, d, J_C-F = 178.1 Hz, C9), 114.1 (CH, d, J_C-F = 24.8 Hz, C2), 123.8 (CH, d, J_C-F = 3.2 Hz, C4), 127.9 (CH, d, J_C-F = 9.3 Hz, C3), 131.2 (CH, d, J_C-F = 13.4 Hz, C11a), 147.7 (C, d, J_C-F = 2.1 Hz, C4a), 159.2 (C, d, J_C-F = 244.1 Hz, C1), 164.9 (C, NHCO). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₇H₁₉ClF₂NO]⁺: 326.1118, found: 326.1116.

4.1.35. 1,9-difluoro-5,6,8,9,10,11-hexahydro-7H-5,9:7,11-dimethanobenzo[9]annulen-7-amine hydrochloride (44)

Thiourea (56 mg, 0.74 mmol) and glacial acetic acid (0.46 mL) were added to a solution of 43 (240 mg, 0.74 mmol) in absolute ethanol (14 mL) and the mixture was heated at reflux overnight. The resulting suspension was then tempered to rt and concentrated under vacuum. Crude was partitioned in H₂O (30 mL) and DCM (30 mL) and the phases were separated. Then the pH of the aq phase was adjusted to 12 with 5 N NaOH solution and extracted with further DCM (3 × 40 mL). The combined organic layers were dried over anh Na₂SO₄, then filtered and concentrated in vacuo. The hydrochloride of 44 was obtained by adding an excess of HCl / Et₂O to a solution of the amine in ethyl acetate, followed by filtration of the resulting white precipitate (167 mg, 79% yield). The analytical sample was obtained by crystallization from methanol, mp >200 °C (dec.) IR (ATR) ν: 3000–2700 (2981, 2950, 2911, 2867, 2831), 2063, 1611, 1588, 1509, 1465, 1445, 1363, 1321, 1246, 1194, 1105, 1095, 1008, 1002, 988, 967, 903, 888, 860, 801, 743, 673 cm⁻¹. ¹H-RMN (400 MHz, MeOD) δ: 1.74–1.96 [complex signal, 4 H, 6(10,12,13)-H_a], 2.07–2.15 [complex signal, 4H, 6(12)-H_b and 8-H₂], 2.15–2.26 [complex signal, 2H, 10(13)-H_b], 3.48 (broad s, 1H, 5-H), 3.99 (t, J = 5.6 Hz, 1H, 11-H), 6.98 (ddd, J = 8.8 Hz, J’ = 8.0 Hz, J” = 1.2 Hz, 1H, 2-H), 7.00 (m, 1H, 4-H), 7.17 (ddd, J = 8.4 Hz, J’ = 8.0 Hz, J” = 5.6 Hz, 1H, 3-H). ¹³C-NMR (100.6 MHz, MeOD) δ : 28.4 (CH, dd, J_C-F = 13.4 Hz, J’_C-F = 6.9 Hz, C11), 38.0 (CH₂, C6 or C12), 38.4 (CH₂, C12 or C6), 39.8 (CH₂, d, J_C-F = 21.0 Hz, C10 or C13), 40.0 (CH, dd, J_C-F = 13.1 Hz, J’_C-F = 2.1 Hz, C5), 40.2 (CH₂, d, J_C-F = 20.8 Hz, C13 or C10), 45.8 (CH₂, d, J_C-F = 20.6 Hz, C8), 58.0 (C, d, J_C-F = 11.1 Hz, C7), 93.9 (C, d, J_C-F = 179.9 Hz, C9), 115.2 (CH, d, J_C-F = 25 Hz, C2), 125.2 (CH, d, J_C-F = 4 Hz, C4), 129.6 (CH, d, J_C-F = 9 Hz, C3), 131.7 (CH, d, J_C-F = 13 Hz, C11a), 148.3 (C, C4a), 160.3 (C, d, J_C-F = 243.4 Hz, C1). HRMS-ESI+ m/z [M+H]⁺ calcd for [C₁₅H₁₈F₂N]⁺: 250.1402, found: 250.1401.

4.2. Intracellular calcium-based determination of IC₅₀ of NMDAR channel blockers

A functional assay of antagonist activity at NMDA receptors was performed using primary cultures of rat cerebellar granule neurons that were prepared according to established protocols.⁴⁹ Cells were grown on 10 mm poly-L-lysine coated round glass cover slips and used for the experiments after 6–9 days in vitro. Cells were loaded with 6 μM Fura-2 AM (ThermoFisher) for 30 min. After loading, a coverslip was mounted on a quartz cuvette containing a Mg²⁺-free Locke-HEPES buffer using a specialized holder. Measurements were performed using a PerkinElmer LS-55 fluorescence spectrometer equipped with a fast filter accessory, under mild agitation and at 37 °C. Analysis from each sample was recorded real-time over 1600 s. After stimulation with NMDA (100 μM, in the presence of 10 μM glycine), increasing concentrations of the compound to be tested were added after signal stabilization. The percentage of inhibition at every tested concentration was analyzed using non-linear regression curve fitting (variable slope) using the software Prism 5.04 (GraphPad Software Inc.).

4.3. Molecular docking and MD simulations

To understand structural aspects of ligand binding using molecular docking and MD simulations, we used an MD-refined computational structural model of the NMDAR (GluN1/2A) transmembrane domain with the channel gate closed. Simulations of receptor-bound IIa, IIb, IIc and 6 were initiated with the compounds in their protonated forms docked into the channel oriented in a “flipped down” conformation such that their amines were interacting with the asparagine (ASN) cluster on the tip of M2 helices. Docking was performed in AutoDock Vina,¹⁰⁶ which predicts interactions between small molecules and proteins. Protein and ligand coordinate files were converted to pdbqt format using AutoDock Tools, where non-polar hydrogen atoms were merged to the bonded heavy atoms. The active site of the protein was designated as a box with size of 26 × 26 × 40 at 1Å grid spacing. Docked receptor-ligand complexes of NMDAR-IIa, NMDAR-IIb, NMDAR-IIc and NMDAR-6 were built and assembled with POPC membrane using CHARMM-GUI Membrane Builder.^107–108 All systems were solvated with TIP3P water and neutralized by adding Na⁺ and Cl⁻ ions to the bulk solution until the salt concentration was 0.15M, leading to a total system size of approximately 169,070 atoms and box size of 127 × 127 × 101 Å. All initial membrane-protein-ligand complex model systems for MD equilibration simulations were built with tleap program in AMBERTOOLS18.¹⁰⁹ FF14SB force field parameters were used for the protein;¹¹⁰ Amber Lipid14 FF for POPC lipid¹¹¹ and general AMBER force field¹¹² (GAFF) was used for all four ligands. To confirm convergence in binding stability of ligands, we used two replicates for each of NMDAR-IIa, NMDAR-IIb, NMDAR-IIc and NMDAR-6 systems.

All 8 initial membrane-protein-ligand complex model systems were energy minimized for 500 cycles of each steepest-descent and conjugate gradient algorithm while keeping restraints on the protein Cα atoms. Water and ions were equilibrated at constant volume as the temperature was gradually increased from 0 to 300K with restraints of 40 kcal/mol/Å on all protein and lipid heavy atoms. This was followed by equilibration for 20 ns at 1 atm and 300 K with a time step of 2 fs using the pmemd.cuda program of the AMBER18 molecular dynamics package.¹⁰⁹ The restraints on the protein residues were gradually reduced from 20 to 0.5 kcal/mol/Å. For each system, MD production runs without any restraints using AMBER18 were performed for 200 ns. The NPT equilibration and production MD simulations were performed with an integration time step of 2 fs while pressure and temperature of the simulated system were maintained at 1 bar and 300 K respectively. We used Langevin thermostat with damping coefficient of 1 ps⁻¹ and a semi-isotropic pressure scaling algorithm as implemented in AMBER18, with pressure relaxation time of 5 ps. During whole simulations, hydrogen bonds were constrained via the SHAKE algorithm.¹¹³ Periodic boundary conditions and all atom wrapping were employed in all simulations as implemented in AMBER18. Long range electrostatics were calculated using the Particle Mesh Ewald method¹¹⁴ where Non-bonded Lennard-Jones and Coulombic interactions were truncated at 8Å. Results were viewed and analyzed with VMD¹¹⁵ and CPPTRAJ tools.¹¹⁶

4.4. Cell culture, transfection, recording, and analysis for electrophysiology experiments

Electrophysiological characterization of memantine-like compounds for Figure 4 was carried out using tsA201 cells transiently transfected with plasmids codifying for NMDAR subunits. tsA201 (Sigma catalog #85120602) is a cell line derived from HEK293 cells stably expressing the temperature sensitive gene for SV40 T-antigen to allow plasmid replication using the SV40 origin and thus enabling high levels of recombinant proteins. Cells were maintained in DMEM supplemented with 10% fetal bovine serum and 1% streptomycin/penicillin as described.⁶⁰ Prior to transfection, cells were plated at 1 × 10⁶ cells/coverslip in glass coverslips previously coated with poly D-lysine (0.1 mg/ml). 24 hours after plating, cells were transiently transfected with 1 μg total cDNAs using PEI transfection reagent (1 mg/ml) in a 3:1 ratio (PEI:DNA). Cells were transfected with plasmids encoding the rat GluN1 subunit, rat GluN2A subunit and enhanced green fluorescent protein (eGFP) for identification of transfected cells. cDNA ratios of 0.5 eGFP: 1 GluN1: 1 GluN2A were used. Culture medium was supplemented with the competitive NMDAR antagonist D,L-2-amino-5-phosphonopentanoate (dl-APV, Sigma, 200 μM) at the time of transfection to prevent NMDAR-mediated cell death.

Whole-cell voltage-clamp recordings were obtained on tsA201 cells 18–30 hours after transfection. Pipettes were pulled from borosilicate capillary tubing (OD = 1.5 mm, ID = 0.86 mm; Harvard Apparatus) using a P-97 electrode puller (Sutter Instruments) and subsequently fire-polished to a resistance of 2 – 5 MΩ using a MF-830 forge (Narishige). Intracellular pipette solution contained contained (in mM): 140 CsCl, 10 HEPES, 5 EGTA, 4 Na₂ATP and 0.1 Na₃GTP with pH adjusted to 7.25 with CsOH. Extracellular recording solution contained (in mM): 140 NaCl, 5 KCl, 1 CaCl₂, 10 HEPES and 10 glucose, adjusted to pH 7.42 with NaOH.

Whole-cell currents were recorded using an Axopatch 200B patch-clamp amplifier (Molecular Devices). Current signal was low-pass filtered at 1 kHz, and sampled at 2 kHz in pClamp10 (Molecular Devices). Solutions containing agonists (100 μM NMDA and 10 μM glycine) or blockers (10 μM) were applied by piezoelectric translation (P-601.30; Physik Instrumente) of a theta-barrel application tool made from borosilicate glass (1.5 mm o.d.; Sutter Instruments). All recordings were performed at rt.

The percentage of channel block, unblock, and recovery were measured with the following protocol: At a holding of −60 mV, NMDA (100 μM) and glycine (10 μM) were applied until current reached a clear steady-state. Then, IIb, IIc or 2 were rapidly applied by piezo control (1 ms solution exchange) for 30 seconds as described.²⁸ During the application of the blocker, a 5 s jump to +60mV was performed to study the voltage dependence of channel block.

Blockers were then removed in the presence of the agonists to allow recovery of the current. During this period (around 1 min) a second jump (5 s duration) to +60mV was performed. Finally, agonists were removed. Percentage of block was calculated by dividing steady state current in the presence of the blocker by steady state current in the absence of the blocker. Percentage of unblock was calculated by dividing the steady state current after blocker removal by the steady state current before blocker application. Finally, the voltage dependence (% of block at +60mV) was calculated by using Equation 1:

$$\%of\mathit{\text{block}}=100\left(1-\frac{I_{+60}\left(\mathit{\text{Drug}}\right)}{I_{+60}\left(\mathit{\text{Agonist}}\right)}\right)$$

where I_+60(Drug) is the current after 5 second at the holding voltage of +60mV in the presence of the blocker and I_+60(Agonist) is the current after 5 seconds at +60mV in the absence of the blocker. The kinetics of blocking and unblocking of NMDAR responses were fitted according to a double-exponential function to calculate the weighted time constant (τ_weighted):

$${\tau }_{\mathit{\text{weighted}}}={\tau }_f\left(\frac{A_f}{A_f+A_S}\right)+{\tau }_S\left(\frac{A_S}{A_f+A_S}\right)$$

where A_f and τ_f are the amplitude and time constant of the fast component of recovery and A_s and τ_s are the amplitude and time constant of the slow component.

For Figure 5, cells were maintained as previously described.¹¹⁷ Briefly, tsA201 cells (European Collection of Authenticated Cell Cultures) were cultured and plated in DMEM supplemented with 10% fetal bovine serum and 1% GlutaMAX (Thermo Fisher Scientific). 18–24 hours after plating, the cells were transfected using FuGENE 6 (Promega) with cDNA coding for enhanced green fluorescent protein (eGFP), WT rat GluN1-1a (GluN1; GenBank X63255), and GluN2A (GenBank M91561 in pcDNA1). GluN1-1a and eGFP were expressed using a plasmid containing cDNA encoding eGFP inserted between the promoter and the GluN1 open reading frame.¹¹⁸ dl-APV was added to the medium at the time of transfection to prevent NMDAR-mediated cell death.

For Figure 5, all whole-cell voltage-clamp recordings were performed at rt 18–30 hours after transfection. Pipettes were fabricated from borosilicate capillary tubing (OD = 1.5 mm, ID = 0.86 mm) using a Flaming Brown P-97 electrode puller (Sutter Instruments) and fire-polished to a resistance of 3.0 – 4.0 MΩ. The intracellular pipette solution contained (in mM): 130 CsCl, 10 HEPES, 10 BAPTA, and 4 MgATP; adjusted to pH 7.2 ± 0.05 with CsOH. Pipette solution osmolality was 275 ± 10 mOsm. Extracellular recording solutions contained (in mM): 140 NaCl, 2.8 KCl, 1 CaCl₂, 10 HEPES, 0.01 EDTA, and 0.1 glycine; adjusted to pH 7.2 ± 0.05 with NaOH and to osmolality 290 ± 10 mOsm with sucrose. Concentrated stocks of glutamate, MgCl₂, and/or IIc were diluted in extracellular solution on the day of experiments (glutamate stock = 1 M; MgCl₂ stock = 1 M; IIc stock = 10 mM in 140 mM NaCl and 2% DMSO, maximum [DMSO] = 0.075%). Extracellular solutions were applied to the patched cell via an in-house built fast perfusion system.¹¹⁷ Whole-cell currents were recorded and digitized using an Axopatch 200A patch-clamp amplifier and Digidata 1440A digitizer (Molecular Devices). The current signal was low-pass filtered at 5 kHz and sampled at 20 kHz using pClamp 10.7 (Molecular Devices). Series resistance was compensated 85 – 90% in all experiments. An empirically determined −6 mV liquid junction potential between the internal and external solutions was corrected in all experiments.

Concentration-inhibition relations were measured using the protocol shown in Figure 5A. Glutamate (1 mM) was applied until current reached steady-state, then sequentially increasing concentrations of IIc were applied for 10 s in the presence of glutamate. Glutamate alone was reapplied for 80 s following IIc application to allow recovery from inhibition. Cells in which current did not recover to 90% of steady-state current during the initial glutamate application were excluded from analysis to prevent underestimation of potency. We estimated IC₅₀ by fitting Equation 2 to concentration-inhibition data:

$$\frac{{\text{I}}_{IIc}}{{\text{I}}_{\text{Glu}}}=\frac{1}{1+{\left(\frac{[IIc]}{{\text{IC}}_{50}}\right)}^{{\text{n}}_{\mathrm{H}}}}$$

where I_IIc is mean current steady state current during application of IIc, I_Glu is the average of the mean steady state currents before and after application of IIc and n_H is the Hill coefficient. IC₅₀ and n_H were free parameters during fitting. I_IIc and I_Glu were measured as the mean current over the final 1 s of each application. Baseline current was subtracted from all current measurements. For Mg²⁺ competition experiments, IIc IC₅₀s were measured in the constant presence of extracellular 0.2 mM Mg²⁺.

We measured the voltage dependence of block by IIc using the protocol shown in Figure 5D. Inhibition by IIc was measured in each cell at nine voltages ranging from −105 to +55 mV. During each voltage step, the cell was exposed to: 5 s in extracellular solution following the voltage step; a 10 s application of 1 mM glutamate; a 15 s application of IIc with 1 mM glutamate; a second application of 1 mM glutamate for 15 s to allow drug unbinding; application of extracellular solution for 2 s to allow a return to baseline current. Voltage was then returned to −65 mV for 3 s before the next voltage jump was made. Voltage dependence of block was quantified using Equation 3:

$$\frac{{\text{I}}_{IIc}}{{\text{I}}_{\text{Glu}}}=\frac{1}{1+\frac{[IIc]}{\left({\text{IC}}_{50}^{-65\text{mV}}\right){\text{e}}^{\frac{{\mathrm{V}}_{\text{m}}+65}{ {\text{V}}_0}}}}$$

where ${\text{IC}}_{50}^{-65\text{mV}}$ is the IC₅₀ at −65 mV (calculated from concentration-inhibition experiments), V_m represents the holding potential, and V₀ represents the change in voltage (in mV) that results in an e-fold change in the IC₅₀ of the drug. I_IIc and I_Glu were measured as in analysis of concentration-inhibition data. V₀ was the only free parameter during fitting. The fraction of the total membrane voltage field felt by the blocker at its binding site (δ)¹¹⁹ was estimated using Equation 4:

$$\delta =\frac{\text{RT}}{{\text{V}}_0\text{ZF}}$$

where R is the ideal gas constant, T is the absolute temperature, z represents the valence of the blocker, and F is the Faraday constant. Note that while δ is useful for comparing voltage dependence of blockers, permeant ions influence the voltage dependence of NMDAR channel block.¹²⁰ Therefore, δ only provides a rough estimate of the location of a channel blocker’s binding site in the voltage field.

4.5. In vitro DMPK

4.5.1. Microsomal stability

The human, rat, and mouse pooled microsomes employed were purchased from Tebu-Xenotech. The compound was incubated at 37 °C with the microsomes in a 50 mM phosphate buffer (pH = 7.4) containing 3 mM MgCl₂, 1 mM NADP, 10 mM glucose-6-phosphate and 1 U/mL glucose-6-phosphate-dehydrogenase. Samples (75 μL) were taken from each well at 0, 10, 20, 40 and 60 min and transferred to 4 °C in a plate containing 75 μL acetonitrile and 30 μL of 0.5% formic acid in water, which were added to improve the chromatographic conditions. The plate was centrifuged (46000 g, 30 min) and supernatants were taken and analyzed in a UPLC-MS/MS (Xevo-TQD, Waters) using a BEH C18 column and an isocratic gradient of 0.1% formic acid in water: 0.1% formic acid acetonitrile (60:40). The metabolic stability of the compounds was calculated from the logarithm of the remaining compounds at each time point studied.

4.5.2. Permeability

The Caco-2 cells were cultured to confluency, trypsinized and seeded onto a 96-filter transwell insert (Corning) at a density of ~10,000 cells/well in DMEM cell culture medium supplemented with 10% fetal bovine serum, 2 mM L-Glutamine and 1% penicillin/streptomycin. Confluent Caco-2 cells were subcultured at passages 58–62 and grown in a humidified atmosphere of 5% CO₂ at 37 °C. Following an overnight attachment period (24 hours after seeding), the cell medium was replaced with fresh medium in both the apical and basolateral compartments every other day. The cell monolayers were used for transport studies 21 days post seeding. The monolayer integrity was checked by measuring the transepithelial electrical resistance (TEER), obtaining values ≥ 500 Ω/cm². On the day of the study, after the TEER measurement, the medium was removed and the cells were washed twice with pre-warmed (37 °C) Hank’s Balanced Salt Solution (HBSS) buffer to remove traces of medium. Stock solutions were made in DMSO, and further diluted in HBSS (final DMSO concentration 1%). Each compound and reference compounds (Colchicine, E3S) were tested at a final concentration of 10 μM. For A → B directional transport, the donor working solution was added to the apical (A) compartment and the transport media as receiver working solution was added to the basolateral (B) compartment. For B → A directional transport, the donor working was added to the basolateral (B) compartment and transport media as receiver working solution was added to the apical (A) compartment. The cells were incubated at 37 °C for 2 hours with gentle stirring.

At the end of the incubation, samples were taken from both donor and receiver compartments and transferred into 384-well plates and analyzed by UPLC-MS/MS. Detection was performed using an ACQUITY UPLC /Xevo TQD system. After the assay, Lucifer yellow (LY) was used to further validate the cell monolayer integrity, cells were incubated with LY 10 μM in HBSS for 1 hour at 37 °C, obtaining permeability (Papp) values for LY of ≤ 10 nm/s confirming the well-established Caco-2 monolayer.

4.5.3. Cytochrome P450 inhibition assay

The objective of this study was to screen the inhibition potential of the compound using recombinant human cytochrome P450 enzymes (CYP1A2, CYP2C9, CYP2C19, CYP2D6, CYP3A4 (BFC) and CYP3A4 (DBF)) and probe substrates with fluorescent detection. Incubations were conducted in a 200 μL volume in 96-well microtiter plates (COSTAR 3915). The addition of the mixture buffer-cofactor (KH₂PO₄ buffer, 1.3 mM NADP, 3.3 mM MgCl₂, 3.3 mM glucose-6-phosphate and 0.4 U/mL glucose-6-phosphate dehydrogenase), control supersomes, standard inhibitors (furafyline, tranylzypromine, ketoconazole, sulfaphenazole and quinidine; Sigma Aldrich), and previously diluted compound to plates was carried out by a liquid handling station (Zephyr Caliper). The plate was then preincubated at 37 °C for 5 min, and the reaction was initiated by the addition of prewarmed enzyme/substrate (E/S) mix. The E/S mix contained buffer (KH₂PO₄), c-DNA-expressed cytochrome P450 enzymes in insect cell microsomes, substrate (3-cyano-7-ethoxycoumarin, for CYP1A2 and CYP2C19; 7-methoxy-4-(trifluoromethyl)coumarin for CYP2C9; 3-[2-(N,N-diethyl-N-methylammonium)ethyl]-7-methoxy-4-methylcoumarin for CYP2D6; and 7-benzyloxytrifluoromethyl coumarin (7-BFC) and dibenzylfluorescein (DBF) for CYP3A4) in a reaction volume of 200 μL. Reactions were terminated after various times (a specific time for each cytochrome) by addition of STOP solution (ACN/TrisHCl 0.5 M 80:20 or 2 N NaOH). Fluorescence per well was measured using a fluorescence plate reader (Tecan Infinity M1000 pro) and percentage of inhibition was calculated.

4.5.4. hERG inhibition assay

The assay was carried out on a CHO cell line transfected with the hERG potassium channel. 72 hours before the assay, 2500 cells were seeded on a 384 well black plate (Greiner 781091). Cells were maintained at 37 °C in a 5% CO₂ atmosphere for 24 hours and at 30 °C in a 5% CO₂ atmosphere for 48h plus. hERG activity was measured by using the FluxorTM Potassium Ion Chanel Assay Kit (Thermo Fisher F10016). Medium was replaced for 20 μl Loading Buffer and the cells were incubated for 60 minutes at RT, protected from direct light. After incubation, Loading Buffer was replaced with Assay buffer and the compounds were incubated for 30 minutes at RT. 5 μl of Stimulus Buffer was added to each well and the fluorescence was read (λex = 490 nm, λem = 525nm) using imaging plate reader system (FDSS7000EX, Hamamatsu^®) every second after the establishment of a baseline line. Astemizole (IC₅₀ = 152 nM) was used to validate the assay.

4.5.5. Parallel Artificial Membrane Permeation Assays-Blood-Brain Barrier (PAMPA-BBB)

To evaluate the brain penetration of the compounds, a parallel artificial membrane permeation assay for blood-brain barrier was used, following the method described by Di et al.¹²¹ The in vitro permeability (P_e) of fourteen commercial drugs through lipid extract of porcine brain membrane together with the test compounds were determined. Commercial drugs and assayed compounds were tested using a mixture of PBS:EtOH (70:30). The assay was validated by comparing experimental and reported P_e values of a set of fourteen commercial drugs (Table S2), and the following correlation was obtained: P_e (exp) = 1.6662 P_e (lit) – 1.4174 (R² = 0.9334). On the basis of this equation and the limits established for BBB permeation,¹¹⁹ the threshold for high BBB permeation (CNS+) was set at P_e (10⁻⁶ cm s⁻¹) > 5.247; low BBB permeation (CNS−) was set at P_e (10⁻⁶ cm s⁻¹) < 1.915; and the range for uncertain BBB permeation (CNS±) at 5.247 > Pe (10⁻⁶ cm s⁻¹) > 1.915. Three different experiments, each performed in triplicate, were carried out for each compound.

4.6. In vivo evaluation of IIc and memantine

4.6.1. Worm strains, maintenance, and general methods

Strains used are listed in Table S3. Standard methods were used for culturing and examination C. elegans.⁹² Wild-type strains were cultured at 20 °C, while transgenic strains were cultured at 16 °C in a temperature-controlled incubator on solid nematode growth medium (NGM) seeded with Escherichia coli (E. coli) OP50 strain as a food source. To collect age synchronized populations of eggs, adults were treated with alkaline hypochlorite solution (0.5 M NaOH, ~2.6% NaCl) for 5–7 min. Fertilized eggs were re-suspended in S-medium for 12 hours, and L1 larvae were incubated to hatch overnight in the absence of food.

4.6.2. Compound Preparation and Treatment

The compounds were dissolved in 100% DMSO. Each concentration was then dissolved in MilliQ purified water to achieve a final concentration between 100 mg/L and 16 mg/L in 1% DMSO in well. For the locomotion assay, treatments were carried out in liquid culture for 4 days at 20 °C. Each well contained a final volume of 60 μL, consisting of 25–30 animals in the L1 stage, compounds under study at the appropriate doses, and OP50 inactivated by freeze-thaw cycles and incubated in S-medium complete to a final optical density of 595 (OD595) of 0.9–0.8 measured in the microplate reader. For the chemotaxis assay synchronized CL2355 and CL2122 strains were treated with DMSO or compounds on NGM plates seeded with inactivated E. coli (OP50). They were cultured at 16 °C for 36 hours, and then at 23 °C for 36 hours.

4.6.3. Locomotion assay

A locomotion assay of NMDA receptor channel blockers was run to obtain a dose-response profile and evaluate their effects on motor function of the CL2006 transgenic strain. The animals were synchronized by alkaline hypochlorite treatment, and the OP50 bacteria OD595 was adjusted to 0.9 to perform the assay. WT strain (N2), and the transgenic strain (CL2006) were treated in liquid media for 4 days at 20 °C with continuous shaking at 180 rpm, starting at the L1 stage. On day 5 of age, the nematodes were moved from the 96-well plates to an unseeded NGM plate for 45 minutes before starting the trial, allowing the plate to dry. Locomotion assays were conducted in 30 mm NGM plates, where the entire plate was covered by OP50. 5 to 10 adult nematodes were positioned in the center of a circle (1 cm diameter) in seeded 30 mm NGM plates. After 1 minute, the number of animals remaining inside the circle were considered locomotor defective. A locomotor defective score (LD) was calculated by dividing the number of worms inside the circle by the total number of worms on the plate. Motor behavior assays were run with a total of at least 200 animals tested per compound concentration. The motor index score ranges from 0% (animals that exhibit the same motor defect as untreated CL2006 animals) to 100% (animals that exhibit motor behavior comparable to WT animals). All results included in the figures are calculated using the following formula:

$$\mathit{\text{Motor}}\mathit{\text{index}}(\%)=\frac{LDCL{2006}_{\mathit{\text{vehicle}}}-LDCL{2006}_{\mathit{\text{drug}}}}{LDCL{2006}_{\mathit{\text{vehicle}}}-LD{W}_{\mathit{\text{vehicle}}}}\times 100$$

4.6.4. Chemotaxis assay

Nematodes were sampled after the treatments with memantine and IIc and washed with M9 buffer (3 g/L KH₂PO₄, 6 g/L Na₂HPO₄, 5 g/L NaCl, 1 mL/L 1 mM MgSO₄, 800 mL/L ddH₂O). In brief, the assay was performed in 100 mm NGM plates. 10 μL of odorant (0.025% benzaldehyde in 96% ethanol) was added to the “attractant” spot, and on the oppose side of the NGM, 10 μL of control odorant (96% ethanol) was added. Then, 70–80 worms were positioned in the center of the plate. Assay plates were incubated at 23 °C for 1 hour and were scored according to the chemotaxis index (CI) as follows: CI = (number of worms at attractant-number of worms at control)/(number of worms at attractant spot + number of worms at control spot). Two trials were performed in triplicate. In each experiment, at least 250 worms from each group were analyzed.

4.6.5. 5XFAD mice

The 5XFAD strain is a well-established and suitable transgenic AD mouse model expressing five familial mutations of human AD. 5XFAD mice also exhibit early-onset cognitive impairment and Aβ pathology.⁸⁸ Six months old female Wild Type (WT) and 5XFAD mice (n = 43) were randomly divided into four groups: WT (n = 14), 5XFAD Control (Control) (n = 8), 5XFAD treated with memantine (Mem) (n = 11) and 5XFAD treated with IIc (IIc) (n = 10). Animals had free access to food and water under standard temperature conditions (22 ± 2 °C) and were maintained on 12 hour : 12 hour light-dark cycles (300 lux/0 lux). Water consumption was controlled each week; drug concentrations were adjusted accordingly to reach the appropriate dose for each cage. Memantine and IIc were dissolved in 1.8% 2-hydroxypropyl-β-cyclodextrin and administered at 5 mg/kg/day through drinking water for four weeks and up to euthanasia. Studies were approved by the Animal Experimentation Ethics Committee (CEEA) at the University of Barcelona. All efforts were made to reduce the number of animals used and their suffering.

4.6.6. Novel Object Recognition Test (NORT)

NORT is a cognitive test used to assess short- and long-term recognition memories. The apparatus used consisted of a 90 °, two-arm, 25-cm-long, 20-cm-high, and a 5-cm-wide black maze of black polyvinyl chloride. The objects to be discriminated were made of plastic and chosen not to frighten mice and without any part likely to be bitten. The test was performed over 5 days. On the first three days, animals were individually habituated to the apparatus for 10 min each day. On the fourth day, the animals were individually exposed during 10 min to the apparatus and were allowed to freely explore the zone inside the apparatus (First trial / Familiarization) where we had placed two identical novel objects (A+A’ or B+B’) at the end of each arm. After 10 min, the animals were removed and returned to their home cage, and 2 hours later the retention trial (second trial) was performed. In this second trial, objects A’ and B’ were swapped (A+B’ or A’+B) and the mice were allowed to explore the maze for 10 min. Twenty-four hours after the first trial, animals were exposed again to the apparatus, and in this case, objects A’ and B’ were substituted by two new objects with different shapes and colours (A+C or B+C’), and animals were allowed to explore them for 10 min. The time exploring the novel object (TN) and the old object (TO) were measured and recorded using a camera. Exploration of an object was defined as pointing the nose towards the object at a distance ≤ 2 cm and/or touching it with the nose. Turning or sitting around the object was not considered exploration. To avoid object preference biases objects A and B were counterbalanced so that one-half of the animals in each experimental group were first exposed to object A and then to object B, whereas the other half first saw object B and then object A. Finally, to quantify cognitive performance, the Discrimination Index (DI) was calculated, which is defined as (TN−TO) / (TN+TO).

4.6.7. Brain Tissue Preparation

Mice groups were euthanized by cervical dislocation and brains were immediately removed from the skull. For molecular experiments, the hippocampus was isolated and frozen in powdered dry ice. Hippocampi were maintained at −80 °C for further use for protein extraction, RNA and DNA isolation. For thioflavin-S staining, mice were anesthetized (ketamine 100 mg / kg and xylazine 10 mg / kg, intraperitoneally) and then perfused with 4% paraformaldehyde (PFA) diluted in 0.1 M phosphate buffer solution intracardially. Brains were removed and postfixed in 4% PFA overnight at 4 °C. Afterwards, brains were changed to PFA + 15% sucrose. Finally, the brains were frozen on powdered dry ice and stored at −80°C until sectioning. Brain coronal sections of 30 μm were obtained (Leica Microsystems CM 3050S cryostat, Wetzlar, Germany) and kept in a cryoprotectant solution at −20 °C until use.

4.6.8. Protein Level Determination by Western Blotting

For Western Blotting (WB), aliquots of 15 μg of hippocampal protein extraction per sample were used. In brief, tissues were homogenized in lysis buffer containing phosphatase and protease inhibitors (Cocktail II, Sigma-Aldrich). Total protein levels were obtained, and protein concentration was determined by the Bradford method. Protein samples were separated by Sodium dodecyl sulphate-Polyacrylamide gel electrophoresis (SDS-PAGE) (8–20%) and transferred onto Polyvinylidene difluoride (PVDF) membranes (Millipore). Afterwards, membranes were blocked in 5% non-fat milk in Tris-buffered saline (TBS) solution containing 0.1% Tween 20 TBS (TBS-T) for 1 hour at rt, followed by overnight incubation at a 4 °C with the primary antibodies listed in Table S4. Then, membranes were washed and incubated with secondary antibodies for 1 hour at RT. Immunoreactive proteins were viewed with the chemiluminescence-based detection kit, following the manufacturer’s protocol (ECL Kit, Millipore), and digital images were acquired using ChemiDoc XRS+ System (BioRad). Semi-quantitative analyses were done using ImageLab software (BioRad), and results were expressed in Arbitrary Units (AU), considering control protein levels as 100%. Protein loading was routinely monitored by immunodetection of Glyceraldehyde-3-phosphate dehydrogenase (GAPDH).

4.6.9. Data Acquisition and Statistical Analysis

Data analysis was conducted using GraphPad Prism ver. 6 statistical software. Data are expressed as the mean ± standard error of the mean (SEM) of at least 3 samples per group. Statistical analysis was performed by way of two-tail Student’s t-test or one-way analysis of variance (ANOVA) followed by Tukey post-hoc analysis. Statistical significance was defined as p-value < 0.05. Statistical outliers were determined with Grubbs’ test and, when necessary, were removed from the analysis. The cognitive analysis was performed blindly: an experimenter unaware of the treatment groups performed the tests and recorded the number of animals. Another experimenter analyzed the videos and the behavioral scoring.

4.7. Cytotoxicity

The cytotoxicity of IIc was measured in Neuro2A cells. Cells were plated in 96-well plates at a density of 40.000 cells/well and grown under conventional conditions (0,2 mL per well in RPMI medium 10%FBS and supplemented with pyruvate and antibiotics). The day after, cells were exposed during 24h to 1, 10 or 100 μM EV-19. Afterwards, cell viability was assessed using the MTT method. Briefly, 20 μL of a solution of 5mg/ml of 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide in PBS was added in each well and after 1h in the cell incubator the supernatant was discarded and cells lysed with 200 μL DMSO. After shaking, absorbance was measured at 595 nm in a Biotek PowerWave HT plate reader. Assays were performed in triplicate and no difference was found between controls and drug-treated cells at any concentration.

Several benzohomoadamantane-based amines were synthesized.

The compounds were evaluated as blockers of the NMDA receptor.

A previously reported compound was as potent as the clinically approved memantine.

Electrophysiological characterization of a selected amine was carried out.

The selected amine showed positive effects in in vivo models of Alzheimer’s disease.