Bis-γ-carbolines as new potential multitarget agents for Alzheimer’s disease

Synthesis

Bis-γ-carbolines 4a–e were obtained by the reaction of tetrahydro-γ-carboline 1 with alkyldimethanesulfonates 2a–e in the presence of 0.5 equivalents of NaH, as previously described [57], with subsequent conversion of isolated bis-γ-carbolines 3a–e into the corresponding hydrochlorides 4a–e (Scheme 1).

Scheme 1:

Synthesis of bis-γ-carbolines 3 and 4. Reagents and conditions: (a) NaH, THF, r.t., 6 h; (b) HCl, acetone, r.t.

Synthesis of bis-γ-carbolines 7 and 10a–c was performed by reaction of N-propargyltetrahydro-γ-carboline 5 with N-azidoethyltetrahydro-γ-carboline 6 and terminal diazidoalkanes 9a–c, respectively, in the presence of catalytic amounts of Cu(I) (Scheme 2). The reaction mixture was stirred for 3–4 h at 40°C. The target bis-γ-carbolines 7, 10a–c were isolated and purified by column chromatography in 78–81% yields [58], and then were converted into the corresponding hydrochlorides 8, 11a–c.

Scheme 2:

Synthesis of bis-γ-carbolines 7, 8, 10, 11. Reagents and conditions: (a) CuSO₄, Na ascorbate, DCM, 40°C, 4 h; (b) HCl, acetone, r.t.

Study of the inhibitory activities of bis-γ-carbolines toward AChE, BChE and CES

We determined esterase profile of bis-γ-carbolines – their inhibitor activity toward two cholinesterases and CES, a structurally related enzyme. CES is a key enzyme of hydrolytic metabolism of many clinically used medicines containing ester, amide, and carbamate groups [59], [60]. The ability of anticholinesterase compounds used for the therapy of AD to inhibit CES may lead to undesirable drug-drug interactions [61]. A comparative assessment of the inhibitory activity of new compounds against several related serine esterases allows us to elucidate both their potential primary pharmacological effect and possible adverse effects [36], [62], [63], [64], [65], [66].

Results on the esterase profile of the synthesized bis-γ-carbolines are presented in Table 1. As can be seen, the bis-γ-carbolines practically do not inhibit CES, which is a desirable result. The inhibitory activity of the tested compounds against cholinesterases, AChE and BChE, was dependent upon the structure of the spacer.

Table 1:

Esterase profile of bis-γ-carbolines and their ability to displace propidium from the EeAChE PAS.

Compound	Spacer	IC50 (μM)±SEM or inhibition % at 20 μM			Displacement of propidium from the EeAChE PASd by the compound at 20 μM (%)
		AChEa	BChEb	CESc
4a	–(CH2)3–	15.68±1.01	4.96±0.10	n.a.	14.4±1.2
4b	–(CH2)4–	12.31±1.86	7.62±0.26	17.9±1.6%	15.9±1.4
4c	–(CH2)5–	5.37±0.08	6.56±0.58	12.2±0.5%	17.9±1.5
4d	–(CH2)6–	2.74±0.26	3.38±0.16	16.2±0.4%	19.4±1.5
4e		9.03±0.43	0.572±0.053	41.3±1.0%	14.5±1.3
8		10.5±1.0	13.2±0.8	3.0±0.2%	16.3±1.4
11a		9.2±0.2%e	38.3±2.6e	3.2±0.8%e	n.d.e
11b		22.4±2.2e	41.9±1.2e	3.7±1.0%e	13.1±1.1e
11c		0.772±0.005e	4.75±0.06e	7.4±0.4%e	23.7±1.8e
Dimebon			36.3±3.59	5.76±0.51	n.a.	4.0±0.5
Donepezil			0.0400±0.0037	19.2±2.97	>100	10.5±0.8
Galantamine			1.02±0.06	15.8±0.90	>100	n.d.

1. ^aHuman erythrocyte AChE.

2. ^bEquine serum BChE.

3. ^cPorcine liver CES.

4. ^dElectric eel AChE peripheral anionic site.

5. n.a., Not active.

6. n.d., Not determined.

7. ^e[56].

As can be seen from Table 1, in the series of conjugates 4a–d, a smooth increase in anti-AChE activity was obtained with an increase in the length of the alkylene spacer –(CH₂)_n–. An increase in the spacer length from n=3 (4a) to n=6 (4d) led to a 6-fold increase in anti-AChE activity. Bis-γ-carbolines with phenylenedialkylene 4e and monotriazole 8 spacers exhibited almost the same moderate anti-AChE activity with an IC₅₀ of about 10 μM. In contrast, for symmetric bis-γ-carbolines 11a–c with ditriazole spacers there was a sharp increase in inhibitor potency with increasing spacer length: from very weak (9.2±0.2% at 20 μM) for conjugate 11a (n=2) to a rather strong submicromolar value (IC₅₀=0.772±0.005 μM) for 11c (n=6).

Most bis-γ-carbolines, with the exception of compound 11a, inhibited AChE more efficiently than Dimebon; i.e., the duplication of the molecule increased inhibitory potency against AChE. The most active conjugate 11c was 47-fold more potent than the parent compound and more active than the marketed drug galantamine (IC₅₀=1.02±0.06 μM).

As for the inhibition of BChE, doubling the γ-carboline pharmacophore using an alkylene spacer (4a–d) and a long ditriazole spacer (11c), maintained the high micromolar anti-BChE activity of Dimebon, whereas conjugates with short ditriazole spacers (11a,b) and a monotriazole spacer (8) were less active.

A particularly noteworthy finding was the sharp increase in anti-BChE activity (more than 20 times) when replacing the monotriazole spacer (8) with phenylenedialkylene (4e). Conjugate 4e was the most potent BChE inhibitor among the studied bis-γ-carbolines, while compounds 8 and 4e possessed equal anti-AChE activity (Table 1).

In addition, it should be noted that Dimebon is characterized by preferential inhibition of BChE, while bis-γ-carbolines with a long spacer (alkylene 4d or bistriazolalkylene 11c) were more effective toward AChE. A similar effect of inversion of selectivity was seen for other described homobivalent bis-γ-carbolines [43], [51].

Kinetic measurements

All of the studied bis-γ-carbolines showed a mixed-type inhibition of cholinesterases. The mechanism of AChE and BChE inhibition by the conjugates is demonstrated for compound 11c as an example. Graphical analysis using double reciprocal Lineweaver–Burk plots is shown in Fig. 2. The plots demonstrate that the binding of conjugate 11c to either AChE or BChE resulted in changes in V_max and K_m. This is consistent with a mixed-type inhibition. The values of AChE and BChE inhibition constants (K_i=competitive component and αK_i=non-competitive component) are shown in Table 2.

Fig. 2:

Lineweaver-Burk double-reciprocal plots of steady state inhibition of AChE (a) and BChE (b) by compound 11c. Each plot indicates mixed-type inhibition.

Table 2:

The inhibition constants of AChE and BChE by some bis-γ-carbolines.^a

Compound	AChE		BChE
	Ki (μM)	αKi (μM)	Ki (μM)	αKi (μM)
4d	0.93±0.08	1.66±0.05	0.73±0.02	2.33±0.21
8	8.64±0.54	16.0±0.8	5.36±0.35	22.7±1.8
11b	7.18±0.06	11.4±0.4	21.2±2.0	89.0±8.6
11c	0.384±0.037	0.810±0.080	0.905±0.069	5.66±0.02

1. ^aValues for K_i (competitive inhibition constant) and αK_i (non-competitive inhibition constant) were determined from analyses of slopes of 1/V vs. 1/S at various inhibitor concentrations. Values (means±SEM) are from at least three separate experiments.

Displacement of propidium iodide from the peripheral anionic site of EeAChE

A fluorescence assay used to evaluate competitive propidium displacement from the PAS of AChE is commonly used as a primary screening method for AChE pro-aggregation activity inhibitors [15]. Propidium is a selective ligand for the PAS of AChE responsible for Aβ binding [8], [12]. It exhibits a fluorescence increase upon binding to AChE. The fluorescence intensity of propidium iodide bound with AChE increases several times [67]. A decrease of fluorescence intensity of the bound propidium in the presence of the test compounds shows their ability to bind to the PAS of AChE, thereby displacing propidium, which predicts that the compounds would block the AChE-mediated aggregation of Aβ [15]. Donepezil was used as the reference compound, as it is in clinical use, and it is a mixed-type AChE inhibitor for which the ability to block AChE-PAS-induced Aβ aggregation has been demonstrated [10]. The results are presented in Table 1.

The results show that all bis-γ-carbolines displaced propidium from the PAS of AChE more effectively than the reference compound donepezil (i.e. >10.5±0.8%) and substantially exceeded the displacement achieved by Dimebon (i.e. >4.0±0.5%). Elongation of the alkylene spacer from n=3 (4a) to n=6 (4d) in conjugates 4a–d led only to a small increase in the degree of propidium displacement (14.4±1.2% – 19.4±1.5%). A significant effect was not observed when a phenylene 4e (14.5±1.3%) or triazole 8 (16.3±1.4%) fragment was introduced into the spacer. However, for symmetrical bis-γ-carbolines with two triazoles in the spacer, an increase in the length of the intertriazole alkylene fragment from n=4 (11b) to n=6 (11c) led to an almost twofold increase in the degree of propidium displacement (13.1±1.1 vs. 23.7±1.8%, respectively).

Molecular modeling

Molecular docking of the bis-γ-carbolines to AChE and BChE was performed to gain insight into the modes of ligand-protein binding. Results of molecular docking of these homobivalent ligands to AChE PDB ID 4EY7 (Fig. 3a–c) showed that one of γ-carboline fragments binds to the active site of AChE at the bottom of the gorge, mostly due to π-π-stacking interactions with Trp86, regardless of the spacer type. There was some difference in orientation of the γ-carboline ring of conjugates 4a,d (Fig. 3a), 4e, 8 (Fig. 3b) compared with 11a–c (Fig. 3c) allowing the former to have ionic interactions with the Glu202 side chain. Comparison of molecular docking results obtained for different AChE X-ray structures showed sensitivity of the position of the compounds to conformation of the residues inside the gorge as was discussed in detail in an earlier report [68].

Fig. 3:

Results of molecular docking of bis-γ-carbolines into (a–c) human AChE (carbon atoms of side chains are shown in pink) and (d–f) human BChE (carbon atoms of side chains are shown in light blue). Carbon atoms of the catalytic triads of each enzyme are shown in yellow. Orange dashed lines show π-π-stacking and π-cationic interactions, blue dashed lines show ionic interactions, and yellow dashed lines show hydrogen bonds. Carbon atoms of compound 4a are shown in light-green, 4d – light-blue (a, d); 4e – brown, 8 – gold (B,E); 11a – green, 11b – cyan, 11c – magenta (e, f).

With increasing length of the spacer, the position of the second γ-carboline ring in the PAS became more favorable: π-π-stacking and π-cationic interactions with Trp286 were obtained, and this tendency was observed for conjugates with alkylene as well as ditriazole spacers. In the latter case, this effect was more pronounced due to a more rigid structure. One of triazole rings of the spacers in conjugates 11a–c, located below the bottleneck, was able to form weak hydrogen bonding interactions, e.g. with Tyr124 (Fig. 3c). For other compounds, specific interactions between the spacer elements and the gorge residues were not observed. For conjugate 11c with the longest ditriazole spacer, ionic interactions with Glu292 were formed (Fig. 3c). This corresponds well with the experimentally observed sharp increase in anti-AChE activity and propidium displacement. No significant differences in AChE binding were found for compounds with phenylenedialkylene (4e) and monotriazole (8) spacers (Fig. 3b), which agrees with their close inhibitor activity.

In contrast, for inhibition of BChE (Fig. 3d–f), a difference in the positions of compounds 8 and 4e with phenylenedialkylene and monotriazole spacers was observed. Due to a more rigid spacer, compound 4e did not bend inside the wide BChE gorge. Instead, it was found in an elongated conformation and formed specific interactions: an ionic interaction with Glu197 in the active site and a hydrogen bond with Asn83 in the PAS (Fig. 3e). Conjugates with an alkylene spacer 4a–d were also able to form an ion pair between the nitrogen atom of the γ-carboline ring and Glu197, but the second γ-carboline ring formed π-π stacking interactions with the Tyr332 ring (Fig. 3d). In spite of different lengths of the spacer, the positions of 4a and 4d were very similar (Fig. 3d). For compounds 11a–c with a ditriazole spacer, the spacer length also did not affect the position of the compound bound in the BChE gorge. The BChE gorge is wide enough to accommodate compounds with lengthy and flexible spacers in bent conformations [36]. One of the γ-carboline fragments of compounds 11a–c forms π-π interactions and salt bridges with Tyr332 and Asp70, respectively (Fig. 3f). The second γ-carboline fragment is located in a very similar way regardless of different linker lengths, though it is not engaged in specific interactions, mostly due to its geometrical fit to the gorge. Thus, the length of the spacer does not affect significantly the location of γ-carboline rings and binding affinity to BChE.

Thus, the results of molecular docking reveal the reasons for the increased anti-AChE activity of the homodimers of γ-carboline that became double-site inhibitors of AChE, their ability to block AChE-induced aggregation of Aβ, and the observed structure-activity relationships.

Antioxidant activity of bis-γ-carbolines

Primary antioxidant activity of the compounds was determined spectrophotometrically by the ABTS assay [69], in which a decrease in absorbance of the cation-radical ABTS˙⁺ (2,2′-azinobis-(3-ethylbenzothiazoline-6-sulfonic acid)) is measured at λ=734 nm after its interaction with an antioxidant compound as previously described in detail [70]. Trolox was used as a reference antioxidant. The results were expressed as TEAC values (Trolox equivalent antioxidant capacity) calculated by dividing the slope of ABTS˙⁺ concentration decrease vs. the tested antioxidant concentration by the slope for the Trolox plot.

The results of the ABTS test (Table 3) show that conjugates 4a–d linked through an alkylene spacer exhibited rather weak antiradical activity with a tendency to increase slightly with increasing spacer length (TEAC=0.03−0.1). Compound (4e) with a phenylenedialkylene spacer also demonstrated low antiradical activity. On the contrary, symmetric homodimers 11a,b with ditriazole spacers possessed radical-scavenging activity comparable to that of the standard antioxidant Trolox. Moreover, the ABTS-scavenging activity of the compounds depended on the structure and length of the spacer: conjugate 8 with a monotriazole spacer possessed moderate antiradical activity (TEAC=0.25). As for conjugates with ditriazole spacers, compounds with the shorter spacer 11a (n=2) and 11b (n=4) scavenged free radicals with high efficiency – at the level of the standard antioxidant Trolox, while elongation of the intertriazole alkylene chain to n=6 (compound 11c) somewhat reduced antiradical activity.

Table 3:

The radical-scavenging activity of bis-γ-carbolines in the ABTS test and their ability to inhibit lipid peroxidation (LP) in rat brain homogenates.

Compound	Spacer	Antiradical activity, ABTS test		Inhibition of LP in rat brain homogenates
		TEACa	IC50, μM	Fe2+-induced, IC50, μM	Residual tBHP-induced LP, % of control at 30 μM
4a	–(CH2)3–	0.03±0.001	n.d.	25.9±8.4	97±8
4b	–(CH2)4–	0.065±0.003	n.d.	≥30	92±9
4c	–(CH2)5–	0.1±0.006	n.d.	9.9±2.0	89±10
4d	–(CH2)6–	0.086±0.007	n.d.	14.3±4.8	90±14
4e		0.047±0.002	n.d.	23.1±11.1	90±14
8		0.25±0.013	96.3±2.8	13.0±3.1	96±3
11a		0.91±0.05b	20.3±2.2	3.0±0.5b	82±3
11b		0.77±0.04b	26.5±1.3	3.1±0.6b	78±5
11c		0.27±0.015b	104.3±3.2	2.8±0.6b	94±4
Dimebon			0.004±0.001	n.d.	>100	103±7
Trolox			1.0	20.4±1.2	31±6	–

1. n.d., Not determined.

2. ^aTEAC values (Trolox equivalent antioxidant capacity) calculated by dividing the slope of ABTS˙⁺ concentration decrease vs. the tested antioxidant concentration by the slope for the Trolox plot.

3. IC₅₀ values=concentration of compound (μM) required to reduce concentration of ABTS˙⁺ by 50%.

4. Results were obtained from 3 replicates of each sample and 3 independent experiments.

5. ^b[56].

These results suggest a significant role of triazole spacers for the radical-scavenging activity of bis-γ-carbolines in the realization of their antioxidant potency.

To obtain information about the antioxidant activity in a biological system, we studied the ability of the bis-γ-carbolines to suppress lipid peroxidation (LP) in the crude membrane fraction from rat brain homogenates. In the biological in vitro system all the studied bis-γ-carbolines were able to inhibit Fe²⁺-induced LP significantly more effectively than Dimebon (Table 3), but were almost ineffective against tBHP-induced LP. Compounds with ditriazole-containing spacers 11a–c, as in the ABTS test, showed the highest activity as antioxidants. These results allowed us to assume a significant role of the radical-scavenging activity of triazole-bearing bis-γ-carbolines in the realization of their antioxidant activity, except that we could not see a dependence of the degree of LP suppression on the length of the ditriazole spacer. At the same time, conjugates linked through alkylene and phenylenedialkylene spacers did not reveal significant radical scavenging activity, although they possessed antioxidant activity in the LP test.

Effect of bis-γ-carbolines on the characteristics of mitochondria

As a primary screening, mitochondrial potential and swelling assays were performed on rat liver mitochondria (RLM). Mitochondrial membrane potential was measured with the fluorescent probe Safranine O at 580 nm/520 nm upon energization with mitochondrial respiratory chain complex I and II-dependent substrates (glutamate/malate and succinate). Calcium-induced swelling of mitochondria as an index of the MPT was determined. The maximum rate of swelling (V_max) is the quantitative parameter characteristic of this process. Therefore, compounds that effectively decrease V_max are potential inhibitors of the MPT and would be expected to possess neuro(cyto)protective properties.

We have shown that only conjugates 11a–c with a ditriazole-containing spacer cause some increase of depolarization of RLM with time and slightly potentiate calcium-induced mitochondrial swelling. All other studied bis-γ-carbolines do not influence the mitochondrial potential. Moreover, they effectively inhibited in a concentration-dependent manner the calcium-induced mitochondrial swelling of RLM characterizing the process of the MPT (Table 4). Compounds 4b, 4e, 8 are the most active as inhibitors of MPT.

Table 4:

Effects of bis-γ-carbolines on mitochondrial potential and calcium-induced MPT.

Compound	Spacer	Depolarization of mitochondria after 10 min incubation with 30 μM of compounds, %	Vmax of Ca2+-induced swelling of mitochondria with 30 μM of compounds, dA620/dt, %		Inhibition of Ca2+-induced swelling of mitochondria, IC50, μM
			10 μM	30 μM
4a	–(CH2)3–	n.a.	83.1±6.8	33.4±5.9	24.3±7.6
4b	–(CH2)4–	3.6±0.9	33.7±9.9	17.1±3	5.8±1.2
4c	–(CH2)5–	5.3±2	71.2±2.2	48.4±8.9	23.5±8.1
4d	–(CH2)6–	5.0±1.6	68.7±12.5	28.2±5.3	15.9±5.5
4e		7.1±2.4	18.2±4.4	–a	2.4±0.9
8		n.a.	52.4±9.9	24.4±1.9	8.9±2.2
11a		9.1±1.3	n.d.	133.6±9.8	n.d.
11b		8.5±1.3	n.d.	119.6±5.4	n.d.
11c		13.7±1.8	n.d.	107.9±4.8	n.d.
Dimebon			n.a.	70.3±14	49.5±10.3	37±17.5

1. ^aDid not exceed the spontaneous swelling.

2. n.a., Not active.

3. n.d., Not determined.

It should be noted that the mechanism of antioxidant activity of bis-γ-carbolines may be associated not only with radical scavenging but also with inhibition of the MPT and the associated decrease in the production of free radicals. This mechanism of antioxidant activity is characteristic for Dimebon, which does not inhibit Fe²⁺- and tBHP-induced LP (Table 3), but it can weakly (IC₅₀>60 µM) inhibit the spontaneous LP in rat brain homogenate (Fig. 4). For one of the leaders among MPT inhibitory compounds (4e), an effective suppression of spontaneous LP in rat brain homogenates with IC₅₀=10.5±1.3 µM was shown (Fig. 4).

Fig. 4:

Influence of compound 4e on spontaneous lipid peroxidation in rat brain homogenate.

Effect of bis-γ-carbolines on calcium retention capacity of brain mitochondria

Ca²⁺ dysregulation is one of the main features of AD-related cellular pathology. Moreover, it seems obvious that Ca²⁺ retention in mitochondria could play a neuroprotective role by removing excess calcium from the cytoplasm. Therefore, agents that increase mitochondrial calcium retention capacity and, accordingly, increase the threshold of MPT-inducing concentrations of calcium, could improve to some extent the viability of neurons as well as the neuronal activity.

Therefore, we investigated the influence of the leaders in the MPT inhibition test γ-carboline conjugates 4e and 8, on calcium retention capacity (CRC) of brain mitochondria (Fig. 5). The experiments were performed in “bolus” mode [27] with a cell (and mitochondria)-impermeant visible light-excitable Ca²⁺ indicator, Calcium Green-5N.

Fig. 5:

Influence of compounds 4e (a) and 8 (b) on calcium retention capacity of mouse brain mitochondria: a and b – kinetics of CaGreen5N fluorescence during the bolus additions of CaCl₂. c – Area Under the Curve of CaGreen5N fluorescence as an indicator of mitochondrial calcium release, which is inversely proportional to the calcium retention capacity of mitochondria.

The increase of probe fluorescence depicted in Fig. 5a, b reflects an increase of extra-mitochondrial calcium and is connected with calcium “bolus” additions, then with decreased uptake of calcium and finally with calcium release from mitochondria due to induction of the MPT. The value of the area under the fluorescence curve (Fig. 5c) can be used as a relative measure characterizing the CRC and inversely proportional to it. One can see from Fig. 5 that compounds 4e and 8 did not inhibit the initial calcium uptake by brain mitochondria (Fig. 5a, b) and that on the whole, they increased the CRC (Fig. 5a–c).

These results allowed us to assume the neuroprotective potential of conjugates with antioxidant and MPT-inhibition activity. The verification of this assumption was performed with an ionomycin-induced calcium overload model of neuronal death using primary cultures of cerebellar granule cells as described earlier [71]. It was shown (Fig. 6) that compound 8 at 100 nM significantly protected cerebellar granule cells from death in this model of neurodegeneration.

Fig. 6:

Effect of compound 8 on cell viability of cerebellar granule neuron in calcium overload cellular model of neurodegeneration. ***p<0.0001, 2-way ANOVA with multiple comparisons.

Predicted ADMET and physicochemical profiles of bis-γ-carbolines

The results of the computational estimation of a number of ADMET properties for the bis-γ-carbolines are shown in Table 5. As can be seen, all the compounds had high predicted values for intestinal absorption, enabling their oral administration. The conjugates with alkylene linker (4a–d) were predicted to have very good blood-brain barrier permeability (brain concentration is about 7–19 times higher than the plasma concentration), ensuring very efficient action on the CNS, although the permeability of other compounds may be more moderate (brain concentration about 1.3–4.3 times higher than the plasma concentration). Both parameters of the cardiac toxicity risk (pK_i and pIC₅₀) for all analyzed compounds (4.38–6.49 log units) are in the lower or medium part of their possible range (3–9 log units). The predicted lipophilicities and aqueous solubilities of the compounds are also good, although some improvement is desirable for a few compounds. Finally, the integral quantitative estimates of drug-likeness (QED) values for most of the compounds were close to 0.3. Thus, the predicted ADMET properties of the studied bis-γ-carbolines are quite acceptable for potential lead compounds at the early drug development stages.

General

All chemicals were purchased from Sigma-Aldrich and used without further purification, except as otherwise stated. Solutions were prepared using deionized water (18.2 MΩ cm at 25°C).

The ¹H NMR spectra were recorded on a Bruker DXP instrument at 200 or 50 MHz, respectively, in CDCl₃ (compounds 3, 7, and 10) and DMSO-d6 (compounds 4, 8, and 11) using tetramethylsilane as an internal standard. Chemical shifts are reported in ppm units using δ scale. Coupling constants are reported in Hertz (Hz). Melting points were measured in open capillary tubes with a SMP10 melting point apparatus and are uncorrected. The elemental analysis was carried out on an Elementar vario MICRO cube CHN analyzer. Chromatography was carried out on Alfa Aesar Silicagel 60 (0.040–0.063 mm). γ-Carboline 1 was prepared as described in [72]. Dimethanesulfonates 2a–d were obtained by the procedure reported in [73]. Bismesylate 2e was purchased from Chemieliva Pharmaceutical Co., Ltd. and used without preliminary purification. The synthesis of compounds 10a–c was described in [58].

Enzymatic assays