Bioorganic & Medicinal Chemistry 23 (2015) 3426–3435
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry journal homepage: www.elsevier.com/locate/bmc
Synthesis, biological activities and pharmacokinetic properties of new fluorinated derivatives of selective PDE4D inhibitors Chiara Brullo a, Matteo Massa a, Carla Villa a, Roberta Ricciarelli b, Daniela Rivera b, Maria Adelaide Pronzato b, Ernesto Fedele c, Elisabetta Barocelli d, Simona Bertoni d, Lisa Flammini d, Olga Bruno a,⇑ a
Department of Pharmacy, Section of Medicinal Chemistry, School of Medical and Pharmaceutical Sciences, University of Genoa, Viale Benedetto XV, 3, 16132 Genoa, Italy Department of Experimental Medicine, Section of General Pathology, School of Medical and Pharmaceutical Sciences, University of Genoa, Via LB Alberti, 2, 16132 Genoa, Italy Department of Pharmacy, Section of Pharmacology and Toxicology, School of Medical and Pharmaceutical Sciences, University of Genoa, Viale Cembrano, 4, 16147 Genoa, Italy d Department of Pharmacy, University of Parma, Parco Area delle Scienze 27/A, 43124 Parma, Italy b c
a r t i c l e
i n f o
Article history: Received 20 February 2015 Revised 9 April 2015 Accepted 11 April 2015 Available online 16 April 2015 Keywords: Phosphodiesterases type 4D inhibitors PDE4D cAMP enhancers Pharmacokinetic properties Microwave-assisted reaction
a b s t r a c t A new series of selective PDE4D inhibitors has been designed and synthesized by replacing 3-methoxy group with 3-difluoromethoxy isoster moiety in our previously reported cathecolic structures. All compounds showed a good PDE4D3 inhibitory activity, most of them being inactive toward other PDE4 isoforms (PDE4A4, PDE4B2 and PDE4C2). Compound 3b, chosen among the synthesized compounds as the most promising in terms of inhibitory activity, selectivity and safety, showed an improved pharmacokinetic profile compared to its non fluorinated analogue. Spontaneous locomotor activity, assessed in an open field apparatus, showed that, differently from rolipram and diazepam, selective PDE4D inhibitors, such as compounds 3b, 5b and 7b, did not affect locomotion, whereas compound 1b showed a tendency to reduce the distance traveled and to prolong the immobility period, possibly due to a poor selectivity. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Phosphodiesterase enzymes (PDEs) regulate the intracellular levels of the second messengers cAMP and cGMP, thus exerting control over a number of cellular functions.1,2 The large PDEs family consists of eleven enzymes ubiquitously expressed in the body. Four different PDE type 4 isoforms (PDE4A, PDE4B, PDE4C and PDE4D) and twenty five splice variants (e.g., PDE4A2, PDE4B2. . .) constitute a subfamily that specifically hydrolizes cAMP in several districts.3–5 Some PDE4 inhibitors (PDE4Is), enhancing cAMP dependent signal transduction in lung and endothelial cells, exert beneficial effects in several inflammatory pathologies (chronic obstructive pulmonary disease, psoriatic arthritis) and have been recently approved for therapeutic use (roflumilast, ibudilast, apremilast). On the other hand, PDE4Is have been proposed as
Abbreviations: AD, Alzheimer disease; cAMP, cyclic adenosine monophosphate; CREB, cAMP response element-binding protein; cGMP, cyclic guanosine monophosphate; OLT, object location test; ORT, object recognition test; PDE, phosphodiesterase; PDE4Is, PDE4 inhibitors; PTC, phase transfer catalyzes; SAR, structure– activity-relationship. ⇑ Corresponding author. Tel.: +39 010 353 8367; fax: +39 010 353 8358. E-mail address: [email protected] (O. Bruno). http://dx.doi.org/10.1016/j.bmc.2015.04.027 0968-0896/Ó 2015 Elsevier Ltd. All rights reserved.
successful agents for the treatment of neurological disorders characterized by cognitive deficits. Particularly, in the hippocampus, cAMP increases cAMP response element-binding protein (CREB) phosphorylation and activates transcription of proteins related to synaptic plasticity and memory formation.6–11 Different lines of evidence have recently indicated PDE4D as the isoform mainly involved in cognitive mechanisms.12–14 However, most of the first generation of pan class PDE4-Is, such as rolipram, showed some undesirable side effects, in particular sedation15 and emesis.16 Our research on PDE4Is led to several series of small molecules structurally related to rolipram and characterized by good and selective PDE4D inhibition.17–19 Among them, compound GEBR7b (7b, Fig. 1) increased memory performances in the object location test (OLT) and the object recognition test (ORT) in rodents,13 and improved spatial memory in the APPswe/PS1dE9 mouse model of Alzheimer’s disease (AD).20 Interestingly, 7b did not induce any emetic-like effect at doses 33–100 times higher than those effective on memory, as assessed with the taste reactivity test on rats and the xylazine/ketamine test on mice.13 In 2010, Burgin et al. reported a PDE4D allosteric modulator able to revert the scopolamine-induced cognitive impairment in
3427
C. Brullo et al. / Bioorg. Med. Chem. 23 (2015) 3426–3435
O O R 1a-c 2b 3a-c 4b 5a-c 6b 7a-c 8b
OH
N O H
R = CHF2 R = CHF2 R = CHF2 R = CHF2 R = CH3 R = CH3 R = CH3 R = CH3
NR2
A
R1 = A R1 = B R1 = C R1 = D R1 = A R1 = B R1 = C R1 = D
NR2 = a-c NR2 = b NR2 = a-c NR2 = b NR2 = a-c NR2 = b NR2 = a-c NR2 = b
R1
B
C D
H N NR2 OH NR2 O NR2
NR2
a
N
O
b
N
O
c
N
OH
O
Figure 1. Molecular structure of new compounds 1a–c, 2b, 3a–c, and 4b and of former 5a–c, 6b, 7a–c, and 8b.
the Y maze test and to improve memory performance in the ORT. That compound was 100–3000 less emetic than rolipram in three different species that are able to vomit (Suncus murinus, beagle dog and Cynomolgus monkey).12 In line with this study, our results, obtained with 7b, supported the idea that PDE4D selective modulators or inhibitors could exert pro-cognitive effects with reduced emetic potential.21 In addition, two different studies reported no changes in locomotor activity between PDE4D knock out and wild type mice, indicating that PDE4D inhibition does not necessarily induce sedation.22,14 The halogen or halogenated group introduction in different active molecules enhances their lipophilicity and improves the blood–brain barrier (BBB) crossing, with a little change in their steric hindrance. Moreover, the fluorine inductive effects could increase the binding force to specific molecular sites and therefore could increase drug potency.23–26 On the basis of these well-established lead optimization paradigms, we planned a further chemical development of our PDE4D inhibitors to improve their druggability. To investigate the role of fluorine substitution on enzyme inhibition activity and to improve the pharmacokinetic properties, a new series of derivatives (compounds 1a–c, 2b, 3a–c, and 4b, Fig. 1) were designed by replacing the 3-methoxy group with a 3-difluoromethoxy isoster moiety in our previously synthesized compounds. In particular, we selected, among the numerous derivatives, some that were active and selective (5a–c and 7a–c, Fig. 1) and some that were inactive (6b and 8b, Fig. 1). Preliminary tests on PDE4D3 activity were performed using the recombinant human enzymes expressed in a baculoviral system. All compounds were tested in duplicate at the concentration of 10 lM, as previously reported.18 We subsequently assessed their inhibitory potency on PDE4D3 (IC50 values) and their selectivity compared with PDE4A4, PDE4B2, and PDE4C2 isoforms. All compounds were analyzed to preliminarily evaluate their cytotoxicity on human neuroblastoma cells (HTLA); subsequently, genotoxicity was assessed only on selected, not cytotoxic, compounds. On the basis of its PDE4D3 inhibitory activity and selectivity, and given the absence of cytotoxic and genotoxic effects, 3b has been selected among the synthesized compounds and tested in vitro to verify its capability of enhancing cAMP accumulation, using 7b as reference compound. Moreover, on the basis of their PDE4D3 inhibitory activity and selectivity, we selected some compounds to assess their capability of affecting spontaneous locomotor activity on mouse exploratory behavior using the open-field test. In detail, we tested the previously reported compound 5b (PDE4D3 active and selective), its fluorinated analogue 1b (PDE4D3 very active, but not selective), our lead 7b, and its fluorinated analogue 3b (PDE4D3 selective, but less active than 7b), in comparison with rolipram and diazepam used as reference compounds.
Finally, preliminary in vivo pharmacokinetic studies were performed on our lead 7b and its fluorinated analogue 3b, to verify the effects of isosteric substitution on BBB crossing ability.
2. Results and discussion 2.1. Chemistry The limitative step to obtain all title compounds is the already described starting product 4-(difluoromethoxy)-3-hydroxybenzaldehyde 9. Most methods reported in the literature use 3,4-dihydroxybenzaldehyde as a substrate and methyl 2-chlorodifluoroacetate as a reagent in the presence of different basic salts (as potassium or caesium carbonate) in a solvent medium.27a–j The yields are ranging from 12% to 44%, but the crude requires an expensive and time-consuming phase of work up and chromatographic purification due to the large amount of byproducts. A recent alternative strategy starting from vanillin gives, in two steps, the desired compound 9 in high yields.28 However, this method involves the use of chlorodifluoromethane, a gas reagent difficult to manipulate and responsible for ozone depletion. Thus, we developed an alternative green method to optimize the reaction conditions and to improve the product yields. A microwave irradiation system has been used and an experimental design was applied by changing the type and amount of basic salts, the heating method and the reaction time (Supporting information). The new microwave assisted procedure achieved an improved yield (57%) and cleaner crude, which was purified by rapid work up and simple Silicagel filtration. The second intermediate 3-(cyclopentyloxy)-4-(difluoromethoxy)benzaldehyde (10), was obtained by alkylation with cyclopentylbromide, as reported in the literature.29 Compound 10 reacted then with hydroxylamine in ethanol 95% and water to give the corresponding oxime (11).27g The latter, by reacting with epichlorohydrin in anhydrous dimethylformamide at 40–50 °C for 24 h, in the presence of sodium ethoxide, gave the 3(cyclopentyloxy)-4-(difluoromethoxy)benzaldehyde O-(oxiran-2ylmethyl)oxime (12), which finally reacted with the suitable amines or with 2,6-dimethylmorpholin-4-amine to give the desired products 1a–c and 2b, respectively, with good yields (Scheme 1). A convergent reaction between the oxime intermediate (11) and the suitable chloroacetylamines or 4-(3-chloropropanoyl)2,6-dimethylmorpholine gave compounds 3a–c and 4b, as reported in the Scheme 1. The syntheses of the building blocks 2,6-dimethylmorpholine4-amine, 4-(chloroacetyl)morpholine, 4-(chloroacetyl)-2,6dimethylmorpholine and 4-(3-chloropropanoyl)morpholine have been already reported in the literature,30–33 but we prepared them
3428
C. Brullo et al. / Bioorg. Med. Chem. 23 (2015) 3426–3435
O
O
O
N O
O F2HC
O
NR2
O
N O
O F2HC
3a-c
N
4b
vii viii F2HC
F2HC
OH OH
O
F2HC
O OH
i or ii
H
O
O O
iii
H
O
9
iv N OH
H
10
O O
v
vi
11
NR2
O F2HC
O
O
N O
O F2HC
1a-c
OH
O N O
12 O
N NH
O
N O
O F2HC
2b
OH
Scheme 1. Synthetic route for the preparation of 1a–c, 2b, 3a–c, and 4b. Reagents and conditions: (i) Methyl 2-chlorodifluoroacetate, Cs2CO3, an. DMF, under N2 atm, 60– 70 °C, 6 h, 39%. (ii) Methyl 2-chlorodifluoroacetate, Cs2CO3, an. DMF, 300 W, under pressure, 90 °C, compressed air cooling, 5 cycles of 3 min, 57%. (iii) Cyclopentylbromide, K2CO3/KI, an. DMF, 65 °C, 22 h, 87%. (iv) NH2OHHCl, NaHCO3, EtOH/H2O, rt, 4 h, 95%. (v) (1) Na, abs EtOH, rt; (2) epichlorohydrin, an. DMF, 45 °C, 24 h, 67%. (vi) For 1a and 1b: morpholine or dimethylmorpholine, 40–50 °C, overnight, 36% and 60%, respectively. For 1c and 2b: 4-hydroxypiperidine or 2,6-dimethylmorpholine-4-amine, an. DMF, 40– 50 °C, overnight, 45% and 21%, respectively. (vii) Chloroacetylamines, an. K2CO3, an. DMF, 50–60 °C, 24 h, 47–85%. (viii) 4-(3-Chloropropanoyl)-2,6-dimethylmorpholine, an. K2CO3, an. DMF, 50–60 °C, 24 h, 18%.
using the suitable amine and chloroacetyl- or chloropropanoyl chloride in toluene in the presence of potassium carbonate, a new profitable method that we have recently described.19 We used the same method also for the synthesis of the 4-(3-chloropropanoyl)-2,6-dimethyl-morpholine, while, 1-(chloroacetyl)piperidin4-ol was prepared from 4-hydroxypiperidine and chloroacetyl chloride in the presence of a sodium carbonate saturated solution/ethyl acetate mixture (1:2), as previously reported in the literature.34 2.2. Enzymatic assays Preliminary tests on PDE4D3 activity were performed using the recombinant human enzymes expressed in a baculoviral system. All compounds were tested in duplicate at the concentration of 10 lM, as previously reported and briefly described in the Section 4.19 Subsequently, all compounds showing percent of inhibition higher than 50% were tested on PDE4D3 at five concentrations (5 10 8–10 4 M) and IC50 values were determined by non linear regression analysis of their inhibition curves. All synthesized compounds showed a good inhibitory activity on PDE4D3 enzyme, ranging from 60% to 87% and IC50 values in the micromolar or sub-micromolar range, being 1b, 2b and 4b the most active compounds with IC50 values of 0.19, 0.19 and 0.47 lM, respectively (Table 1). Compounds 1b, 3a, 3c and 4b showed a good increase in potency with respect to the isosteric
analogues (5b, 7a, 7c and 8b, respectively). In particular, the result of compounds 3c and 4b (IC50 values of 2.63 and 0.47 lM, respectively) has to be underlined, since the corresponding non fluorinated analogues (7c and 8b) were almost inactive. Moreover, compounds 1a and 2b showed the same potency against PDE4D3 as the analogues 5a and 6b, while compounds 1c and 3b were less potent than 5c and 7b, respectively.
Table 1 Compounds 1–8 inhibitory activity and IC50 values toward the PDE4D3 isoforma Compd
PDE4D3 % inhib (10 lM)
IC50 (lM)
Compd
PDE4D3 % inhib (10 lM)
IC50 (lM)
1a 1b 1c 2b 3a 3b 3c 4b
83 84 51 87 63 74 60 63
1.49 0.19 2.54 0.19 1.12 4.84 2.63 0.47
5ab 5bc 5cd 6bd 7ab 7bc 7cd 8b
81 68 65 80 73 67 30 24
1.45 3.46 1.79 0.21 3.48 1.91 nde nd
a Results are expressed as % inhibition in respect to the control. Compounds have been tested in duplicate at 10 lM concentration. b Results already reported for generic PDE4.18 c Results already reported.18 d Results already reported.19 e Not determined.
3429
C. Brullo et al. / Bioorg. Med. Chem. 23 (2015) 3426–3435
We, therefore, suggest that introducing fluorine in the cathecolic system generally improves the interaction with the catalytic domain and causes a better PDE4D3 inhibition. Subsequent screening on different PDE4 isoforms (particularly PDE4A4, PDE4B2 and PDE4C2) were performed on all compounds in order to determine their selectivity (Table 2). The majority of the tested compounds, except 1b and 2b, showed a good PDE4D3 selectivity, having inhibitory activity, at 10 lM concentration, lower than 50% toward PDE4A4, PDE4B2 and PDE4C2 (Table 2). The MOLCAD surface studies performed by Srivani et al.35 showed a higher lipophilicity in the PDE4B M-loop compared to that of PDE4D. Such observations could explain the decreased selectivity of compounds 1b and 2b in comparison with their isoster 5b and 6b.19 On the other hand, the in silico calculations on PDE4B and PDE4D interaction with our previous compounds were not particularly explicatory and only gave an approximate picture of the ligand–enzyme interaction phenomenon, which is actually very complex and influenced by many different factors. 2.3. Cytotoxicity and genotoxicity assays To evaluate the cytotoxic potential of the new molecules, we analyzed the lactate-dehydrogenase release in human HTLA cells that were exposed for 24 h to a very high concentration (100 lM) of the test compounds and 7b (as reference compound). Compounds exhibiting a cytotoxic potential higher than 3% (i.e., 1b, 1c and 3c) have also been tested at a lower concentration (10 lM), finally showing negligible cytotoxicity (Table 3). To evaluate the genotoxicity, we analyzed the phosphorylation of the chromatin-bound histone H2AX (c-H2AX), which is a quantitative marker for the DNA damage response at the site of doublestrand breaks.36 To this purpose, we performed immunoblot analysis on protein extracts from HTLA cells that had been exposed for 24 h to the compounds 1a, 3b and 4b, which showed a cytotoxic potential below 1% at the concentration of 100 lM. As a positive control, we used etoposide, a topoisomerase II inhibitor that induces DNA double-stranded breaks.37 While the etoposide treatment of HTLA cells led to a robust DNA damage, compounds 1a, 3b and 4b, as well as 7b, did not produce the c-H2AX immunoreactive band, typical marker of genotoxicity (Fig. 2). 2.4. In vitro intracellular cAMP quantification To verify the capability of enhancing cAMP intracellular accumulation, we used a cAMP-specific enzymatic immunoassay (EIA) in HTLA cells that had been exposed for 30 min to compound 3b,
Table 3 Relative cytotoxic potential of compounds 1a–c, 2b, 3a–c, and 4b and 7b as reference compounda
a
Compd
Cytotoxicity%
Positive control DMSO 1a (100 lM) 1b (100 lM) 1b (10 lM) 1c (100 lM) 1c (10 lM) 2b (100 lM) 3a (100 lM) 3b (100 lM) 3c (100 lM) 3c (10 lM) 4b (100 lM) 7b (100 lM)
100.00 ± 1.98 0.67 ± 0.24 0.21 ± 1.40 8.83 ± 6.36 0.01 ± 0.02 31.37 ± 4.13 3.42 ± 2.49 1.54 ± 0.91 2.06 ± 1.61 0.01 ± 0.42 4.62 ± 4.62 0.01 ± 0.17 0.01 ± 0.40 0.14 ± 0.18
Data represent the mean ± SEM for three independent experiments.
which had been selected among all the new and previous synthesized compounds on the basis of its PDE4D3 inhibitory activity and selectivity, and the absence of cytotoxic and genotoxic effects. In the same assay, the non fluorinated analogue 7b was used as reference compound. In order to slightly stimulate the activity of adenylyl cyclase, the cells received forskolin (FSK) during the last 20 min of incubation. Both 7b and 3b have been able to significantly increase the FSK-induced cAMP accumulation, without affecting the basal levels of the cyclic nucleotide (Fig. 3). Moreover, compound 3b showed higher efficacy in increasing cAMP intracellular levels than 7b, despite having a lower PDE4D3 inhibitory activity. As the introduction of fluorine is known to increase the lipophilicity of drugs, we feel that the greater effect of compound 3b on cAMP accumulation is likely to depend on its increased cell penetration and intracellular availability (see also below). However, given that, in addition to PDE4A4, B2, C2 and D3 (see Table 2), there are other different PDE4D variants, the possibility that 3b inhibits more enzymes than 7b, leading higher levels of intracellular cAMP, cannot be completely ruled out. 2.5. Spontaneous motor activity in an open field task In order to verify the potential sedative effects, compounds 1b, 3b, 5b, and 7b (10 mg/kg) were tested on mouse exploratory behavior using the open-field test in comparison with rolipram (10 mg/kg) and diazepam (10 mg/kg), used as reference drugs. Compounds 3b, 5b and 7b did not affect mice locomotor activity at the same dose of rolipram and diazepam, which, on the
Table 2 Inhibition activity % toward different PDE4 isoforms of compounds 1a–c, 2, 3a–c, and 4 in comparison with previously reported compounds 5–8, and rolipram as reference compounda
a b c d
Compd
PDE4A4 %inhib
PDE4B2 %inhib
PDE4C2 %inhib
Compd
PDE4A4 %inhib
PDE4B2 %inhib
PDE4C2 %inhib
1a 1b 1c 2b 3a 3b 3c 4b Rolipram
41 68 15 58 36 32 28 34 79
43 68 15 51 31 29 26 37 65
ntb 62 nt nt nt 9 nt nt 57
5a 5bc 5cd 6bd 7ac 7bc 7c 8b —
nt 34 10 41 nt 34 nt nt —
nt 39 5 51 nt 23 nt nt —
nt 19 9 25 nt 25 nt nt —
Results expressed as % inhibition in respect to the control at 10 lM concentration, in duplicate. Not tested. Results already reported.18 Results already reported.19
3430
C. Brullo et al. / Bioorg. Med. Chem. 23 (2015) 3426–3435
Figure 2. Genotoxic potential of compounds 1a, 3b, 4b, and 7b. Western blot analysis of c-H2AX in HTLA cells treated for 24 h with 100 lM etoposide, 7b and the test compounds, or an equal volume of solvent (DMSO). The H2AX signals represent the internal loading control. The figure is representative of three independent experiments all showing essentially similar results.
contrary, significantly, reduced the distance traveled by mice (P
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry journal homepage: www.elsevier.com/locate/bmc
Synthesis, biological activities and pharmacokinetic properties of new fluorinated derivatives of selective PDE4D inhibitors Chiara Brullo a, Matteo Massa a, Carla Villa a, Roberta Ricciarelli b, Daniela Rivera b, Maria Adelaide Pronzato b, Ernesto Fedele c, Elisabetta Barocelli d, Simona Bertoni d, Lisa Flammini d, Olga Bruno a,⇑ a
Department of Pharmacy, Section of Medicinal Chemistry, School of Medical and Pharmaceutical Sciences, University of Genoa, Viale Benedetto XV, 3, 16132 Genoa, Italy Department of Experimental Medicine, Section of General Pathology, School of Medical and Pharmaceutical Sciences, University of Genoa, Via LB Alberti, 2, 16132 Genoa, Italy Department of Pharmacy, Section of Pharmacology and Toxicology, School of Medical and Pharmaceutical Sciences, University of Genoa, Viale Cembrano, 4, 16147 Genoa, Italy d Department of Pharmacy, University of Parma, Parco Area delle Scienze 27/A, 43124 Parma, Italy b c
a r t i c l e
i n f o
Article history: Received 20 February 2015 Revised 9 April 2015 Accepted 11 April 2015 Available online 16 April 2015 Keywords: Phosphodiesterases type 4D inhibitors PDE4D cAMP enhancers Pharmacokinetic properties Microwave-assisted reaction
a b s t r a c t A new series of selective PDE4D inhibitors has been designed and synthesized by replacing 3-methoxy group with 3-difluoromethoxy isoster moiety in our previously reported cathecolic structures. All compounds showed a good PDE4D3 inhibitory activity, most of them being inactive toward other PDE4 isoforms (PDE4A4, PDE4B2 and PDE4C2). Compound 3b, chosen among the synthesized compounds as the most promising in terms of inhibitory activity, selectivity and safety, showed an improved pharmacokinetic profile compared to its non fluorinated analogue. Spontaneous locomotor activity, assessed in an open field apparatus, showed that, differently from rolipram and diazepam, selective PDE4D inhibitors, such as compounds 3b, 5b and 7b, did not affect locomotion, whereas compound 1b showed a tendency to reduce the distance traveled and to prolong the immobility period, possibly due to a poor selectivity. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Phosphodiesterase enzymes (PDEs) regulate the intracellular levels of the second messengers cAMP and cGMP, thus exerting control over a number of cellular functions.1,2 The large PDEs family consists of eleven enzymes ubiquitously expressed in the body. Four different PDE type 4 isoforms (PDE4A, PDE4B, PDE4C and PDE4D) and twenty five splice variants (e.g., PDE4A2, PDE4B2. . .) constitute a subfamily that specifically hydrolizes cAMP in several districts.3–5 Some PDE4 inhibitors (PDE4Is), enhancing cAMP dependent signal transduction in lung and endothelial cells, exert beneficial effects in several inflammatory pathologies (chronic obstructive pulmonary disease, psoriatic arthritis) and have been recently approved for therapeutic use (roflumilast, ibudilast, apremilast). On the other hand, PDE4Is have been proposed as
Abbreviations: AD, Alzheimer disease; cAMP, cyclic adenosine monophosphate; CREB, cAMP response element-binding protein; cGMP, cyclic guanosine monophosphate; OLT, object location test; ORT, object recognition test; PDE, phosphodiesterase; PDE4Is, PDE4 inhibitors; PTC, phase transfer catalyzes; SAR, structure– activity-relationship. ⇑ Corresponding author. Tel.: +39 010 353 8367; fax: +39 010 353 8358. E-mail address: [email protected] (O. Bruno). http://dx.doi.org/10.1016/j.bmc.2015.04.027 0968-0896/Ó 2015 Elsevier Ltd. All rights reserved.
successful agents for the treatment of neurological disorders characterized by cognitive deficits. Particularly, in the hippocampus, cAMP increases cAMP response element-binding protein (CREB) phosphorylation and activates transcription of proteins related to synaptic plasticity and memory formation.6–11 Different lines of evidence have recently indicated PDE4D as the isoform mainly involved in cognitive mechanisms.12–14 However, most of the first generation of pan class PDE4-Is, such as rolipram, showed some undesirable side effects, in particular sedation15 and emesis.16 Our research on PDE4Is led to several series of small molecules structurally related to rolipram and characterized by good and selective PDE4D inhibition.17–19 Among them, compound GEBR7b (7b, Fig. 1) increased memory performances in the object location test (OLT) and the object recognition test (ORT) in rodents,13 and improved spatial memory in the APPswe/PS1dE9 mouse model of Alzheimer’s disease (AD).20 Interestingly, 7b did not induce any emetic-like effect at doses 33–100 times higher than those effective on memory, as assessed with the taste reactivity test on rats and the xylazine/ketamine test on mice.13 In 2010, Burgin et al. reported a PDE4D allosteric modulator able to revert the scopolamine-induced cognitive impairment in
3427
C. Brullo et al. / Bioorg. Med. Chem. 23 (2015) 3426–3435
O O R 1a-c 2b 3a-c 4b 5a-c 6b 7a-c 8b
OH
N O H
R = CHF2 R = CHF2 R = CHF2 R = CHF2 R = CH3 R = CH3 R = CH3 R = CH3
NR2
A
R1 = A R1 = B R1 = C R1 = D R1 = A R1 = B R1 = C R1 = D
NR2 = a-c NR2 = b NR2 = a-c NR2 = b NR2 = a-c NR2 = b NR2 = a-c NR2 = b
R1
B
C D
H N NR2 OH NR2 O NR2
NR2
a
N
O
b
N
O
c
N
OH
O
Figure 1. Molecular structure of new compounds 1a–c, 2b, 3a–c, and 4b and of former 5a–c, 6b, 7a–c, and 8b.
the Y maze test and to improve memory performance in the ORT. That compound was 100–3000 less emetic than rolipram in three different species that are able to vomit (Suncus murinus, beagle dog and Cynomolgus monkey).12 In line with this study, our results, obtained with 7b, supported the idea that PDE4D selective modulators or inhibitors could exert pro-cognitive effects with reduced emetic potential.21 In addition, two different studies reported no changes in locomotor activity between PDE4D knock out and wild type mice, indicating that PDE4D inhibition does not necessarily induce sedation.22,14 The halogen or halogenated group introduction in different active molecules enhances their lipophilicity and improves the blood–brain barrier (BBB) crossing, with a little change in their steric hindrance. Moreover, the fluorine inductive effects could increase the binding force to specific molecular sites and therefore could increase drug potency.23–26 On the basis of these well-established lead optimization paradigms, we planned a further chemical development of our PDE4D inhibitors to improve their druggability. To investigate the role of fluorine substitution on enzyme inhibition activity and to improve the pharmacokinetic properties, a new series of derivatives (compounds 1a–c, 2b, 3a–c, and 4b, Fig. 1) were designed by replacing the 3-methoxy group with a 3-difluoromethoxy isoster moiety in our previously synthesized compounds. In particular, we selected, among the numerous derivatives, some that were active and selective (5a–c and 7a–c, Fig. 1) and some that were inactive (6b and 8b, Fig. 1). Preliminary tests on PDE4D3 activity were performed using the recombinant human enzymes expressed in a baculoviral system. All compounds were tested in duplicate at the concentration of 10 lM, as previously reported.18 We subsequently assessed their inhibitory potency on PDE4D3 (IC50 values) and their selectivity compared with PDE4A4, PDE4B2, and PDE4C2 isoforms. All compounds were analyzed to preliminarily evaluate their cytotoxicity on human neuroblastoma cells (HTLA); subsequently, genotoxicity was assessed only on selected, not cytotoxic, compounds. On the basis of its PDE4D3 inhibitory activity and selectivity, and given the absence of cytotoxic and genotoxic effects, 3b has been selected among the synthesized compounds and tested in vitro to verify its capability of enhancing cAMP accumulation, using 7b as reference compound. Moreover, on the basis of their PDE4D3 inhibitory activity and selectivity, we selected some compounds to assess their capability of affecting spontaneous locomotor activity on mouse exploratory behavior using the open-field test. In detail, we tested the previously reported compound 5b (PDE4D3 active and selective), its fluorinated analogue 1b (PDE4D3 very active, but not selective), our lead 7b, and its fluorinated analogue 3b (PDE4D3 selective, but less active than 7b), in comparison with rolipram and diazepam used as reference compounds.
Finally, preliminary in vivo pharmacokinetic studies were performed on our lead 7b and its fluorinated analogue 3b, to verify the effects of isosteric substitution on BBB crossing ability.
2. Results and discussion 2.1. Chemistry The limitative step to obtain all title compounds is the already described starting product 4-(difluoromethoxy)-3-hydroxybenzaldehyde 9. Most methods reported in the literature use 3,4-dihydroxybenzaldehyde as a substrate and methyl 2-chlorodifluoroacetate as a reagent in the presence of different basic salts (as potassium or caesium carbonate) in a solvent medium.27a–j The yields are ranging from 12% to 44%, but the crude requires an expensive and time-consuming phase of work up and chromatographic purification due to the large amount of byproducts. A recent alternative strategy starting from vanillin gives, in two steps, the desired compound 9 in high yields.28 However, this method involves the use of chlorodifluoromethane, a gas reagent difficult to manipulate and responsible for ozone depletion. Thus, we developed an alternative green method to optimize the reaction conditions and to improve the product yields. A microwave irradiation system has been used and an experimental design was applied by changing the type and amount of basic salts, the heating method and the reaction time (Supporting information). The new microwave assisted procedure achieved an improved yield (57%) and cleaner crude, which was purified by rapid work up and simple Silicagel filtration. The second intermediate 3-(cyclopentyloxy)-4-(difluoromethoxy)benzaldehyde (10), was obtained by alkylation with cyclopentylbromide, as reported in the literature.29 Compound 10 reacted then with hydroxylamine in ethanol 95% and water to give the corresponding oxime (11).27g The latter, by reacting with epichlorohydrin in anhydrous dimethylformamide at 40–50 °C for 24 h, in the presence of sodium ethoxide, gave the 3(cyclopentyloxy)-4-(difluoromethoxy)benzaldehyde O-(oxiran-2ylmethyl)oxime (12), which finally reacted with the suitable amines or with 2,6-dimethylmorpholin-4-amine to give the desired products 1a–c and 2b, respectively, with good yields (Scheme 1). A convergent reaction between the oxime intermediate (11) and the suitable chloroacetylamines or 4-(3-chloropropanoyl)2,6-dimethylmorpholine gave compounds 3a–c and 4b, as reported in the Scheme 1. The syntheses of the building blocks 2,6-dimethylmorpholine4-amine, 4-(chloroacetyl)morpholine, 4-(chloroacetyl)-2,6dimethylmorpholine and 4-(3-chloropropanoyl)morpholine have been already reported in the literature,30–33 but we prepared them
3428
C. Brullo et al. / Bioorg. Med. Chem. 23 (2015) 3426–3435
O
O
O
N O
O F2HC
O
NR2
O
N O
O F2HC
3a-c
N
4b
vii viii F2HC
F2HC
OH OH
O
F2HC
O OH
i or ii
H
O
O O
iii
H
O
9
iv N OH
H
10
O O
v
vi
11
NR2
O F2HC
O
O
N O
O F2HC
1a-c
OH
O N O
12 O
N NH
O
N O
O F2HC
2b
OH
Scheme 1. Synthetic route for the preparation of 1a–c, 2b, 3a–c, and 4b. Reagents and conditions: (i) Methyl 2-chlorodifluoroacetate, Cs2CO3, an. DMF, under N2 atm, 60– 70 °C, 6 h, 39%. (ii) Methyl 2-chlorodifluoroacetate, Cs2CO3, an. DMF, 300 W, under pressure, 90 °C, compressed air cooling, 5 cycles of 3 min, 57%. (iii) Cyclopentylbromide, K2CO3/KI, an. DMF, 65 °C, 22 h, 87%. (iv) NH2OHHCl, NaHCO3, EtOH/H2O, rt, 4 h, 95%. (v) (1) Na, abs EtOH, rt; (2) epichlorohydrin, an. DMF, 45 °C, 24 h, 67%. (vi) For 1a and 1b: morpholine or dimethylmorpholine, 40–50 °C, overnight, 36% and 60%, respectively. For 1c and 2b: 4-hydroxypiperidine or 2,6-dimethylmorpholine-4-amine, an. DMF, 40– 50 °C, overnight, 45% and 21%, respectively. (vii) Chloroacetylamines, an. K2CO3, an. DMF, 50–60 °C, 24 h, 47–85%. (viii) 4-(3-Chloropropanoyl)-2,6-dimethylmorpholine, an. K2CO3, an. DMF, 50–60 °C, 24 h, 18%.
using the suitable amine and chloroacetyl- or chloropropanoyl chloride in toluene in the presence of potassium carbonate, a new profitable method that we have recently described.19 We used the same method also for the synthesis of the 4-(3-chloropropanoyl)-2,6-dimethyl-morpholine, while, 1-(chloroacetyl)piperidin4-ol was prepared from 4-hydroxypiperidine and chloroacetyl chloride in the presence of a sodium carbonate saturated solution/ethyl acetate mixture (1:2), as previously reported in the literature.34 2.2. Enzymatic assays Preliminary tests on PDE4D3 activity were performed using the recombinant human enzymes expressed in a baculoviral system. All compounds were tested in duplicate at the concentration of 10 lM, as previously reported and briefly described in the Section 4.19 Subsequently, all compounds showing percent of inhibition higher than 50% were tested on PDE4D3 at five concentrations (5 10 8–10 4 M) and IC50 values were determined by non linear regression analysis of their inhibition curves. All synthesized compounds showed a good inhibitory activity on PDE4D3 enzyme, ranging from 60% to 87% and IC50 values in the micromolar or sub-micromolar range, being 1b, 2b and 4b the most active compounds with IC50 values of 0.19, 0.19 and 0.47 lM, respectively (Table 1). Compounds 1b, 3a, 3c and 4b showed a good increase in potency with respect to the isosteric
analogues (5b, 7a, 7c and 8b, respectively). In particular, the result of compounds 3c and 4b (IC50 values of 2.63 and 0.47 lM, respectively) has to be underlined, since the corresponding non fluorinated analogues (7c and 8b) were almost inactive. Moreover, compounds 1a and 2b showed the same potency against PDE4D3 as the analogues 5a and 6b, while compounds 1c and 3b were less potent than 5c and 7b, respectively.
Table 1 Compounds 1–8 inhibitory activity and IC50 values toward the PDE4D3 isoforma Compd
PDE4D3 % inhib (10 lM)
IC50 (lM)
Compd
PDE4D3 % inhib (10 lM)
IC50 (lM)
1a 1b 1c 2b 3a 3b 3c 4b
83 84 51 87 63 74 60 63
1.49 0.19 2.54 0.19 1.12 4.84 2.63 0.47
5ab 5bc 5cd 6bd 7ab 7bc 7cd 8b
81 68 65 80 73 67 30 24
1.45 3.46 1.79 0.21 3.48 1.91 nde nd
a Results are expressed as % inhibition in respect to the control. Compounds have been tested in duplicate at 10 lM concentration. b Results already reported for generic PDE4.18 c Results already reported.18 d Results already reported.19 e Not determined.
3429
C. Brullo et al. / Bioorg. Med. Chem. 23 (2015) 3426–3435
We, therefore, suggest that introducing fluorine in the cathecolic system generally improves the interaction with the catalytic domain and causes a better PDE4D3 inhibition. Subsequent screening on different PDE4 isoforms (particularly PDE4A4, PDE4B2 and PDE4C2) were performed on all compounds in order to determine their selectivity (Table 2). The majority of the tested compounds, except 1b and 2b, showed a good PDE4D3 selectivity, having inhibitory activity, at 10 lM concentration, lower than 50% toward PDE4A4, PDE4B2 and PDE4C2 (Table 2). The MOLCAD surface studies performed by Srivani et al.35 showed a higher lipophilicity in the PDE4B M-loop compared to that of PDE4D. Such observations could explain the decreased selectivity of compounds 1b and 2b in comparison with their isoster 5b and 6b.19 On the other hand, the in silico calculations on PDE4B and PDE4D interaction with our previous compounds were not particularly explicatory and only gave an approximate picture of the ligand–enzyme interaction phenomenon, which is actually very complex and influenced by many different factors. 2.3. Cytotoxicity and genotoxicity assays To evaluate the cytotoxic potential of the new molecules, we analyzed the lactate-dehydrogenase release in human HTLA cells that were exposed for 24 h to a very high concentration (100 lM) of the test compounds and 7b (as reference compound). Compounds exhibiting a cytotoxic potential higher than 3% (i.e., 1b, 1c and 3c) have also been tested at a lower concentration (10 lM), finally showing negligible cytotoxicity (Table 3). To evaluate the genotoxicity, we analyzed the phosphorylation of the chromatin-bound histone H2AX (c-H2AX), which is a quantitative marker for the DNA damage response at the site of doublestrand breaks.36 To this purpose, we performed immunoblot analysis on protein extracts from HTLA cells that had been exposed for 24 h to the compounds 1a, 3b and 4b, which showed a cytotoxic potential below 1% at the concentration of 100 lM. As a positive control, we used etoposide, a topoisomerase II inhibitor that induces DNA double-stranded breaks.37 While the etoposide treatment of HTLA cells led to a robust DNA damage, compounds 1a, 3b and 4b, as well as 7b, did not produce the c-H2AX immunoreactive band, typical marker of genotoxicity (Fig. 2). 2.4. In vitro intracellular cAMP quantification To verify the capability of enhancing cAMP intracellular accumulation, we used a cAMP-specific enzymatic immunoassay (EIA) in HTLA cells that had been exposed for 30 min to compound 3b,
Table 3 Relative cytotoxic potential of compounds 1a–c, 2b, 3a–c, and 4b and 7b as reference compounda
a
Compd
Cytotoxicity%
Positive control DMSO 1a (100 lM) 1b (100 lM) 1b (10 lM) 1c (100 lM) 1c (10 lM) 2b (100 lM) 3a (100 lM) 3b (100 lM) 3c (100 lM) 3c (10 lM) 4b (100 lM) 7b (100 lM)
100.00 ± 1.98 0.67 ± 0.24 0.21 ± 1.40 8.83 ± 6.36 0.01 ± 0.02 31.37 ± 4.13 3.42 ± 2.49 1.54 ± 0.91 2.06 ± 1.61 0.01 ± 0.42 4.62 ± 4.62 0.01 ± 0.17 0.01 ± 0.40 0.14 ± 0.18
Data represent the mean ± SEM for three independent experiments.
which had been selected among all the new and previous synthesized compounds on the basis of its PDE4D3 inhibitory activity and selectivity, and the absence of cytotoxic and genotoxic effects. In the same assay, the non fluorinated analogue 7b was used as reference compound. In order to slightly stimulate the activity of adenylyl cyclase, the cells received forskolin (FSK) during the last 20 min of incubation. Both 7b and 3b have been able to significantly increase the FSK-induced cAMP accumulation, without affecting the basal levels of the cyclic nucleotide (Fig. 3). Moreover, compound 3b showed higher efficacy in increasing cAMP intracellular levels than 7b, despite having a lower PDE4D3 inhibitory activity. As the introduction of fluorine is known to increase the lipophilicity of drugs, we feel that the greater effect of compound 3b on cAMP accumulation is likely to depend on its increased cell penetration and intracellular availability (see also below). However, given that, in addition to PDE4A4, B2, C2 and D3 (see Table 2), there are other different PDE4D variants, the possibility that 3b inhibits more enzymes than 7b, leading higher levels of intracellular cAMP, cannot be completely ruled out. 2.5. Spontaneous motor activity in an open field task In order to verify the potential sedative effects, compounds 1b, 3b, 5b, and 7b (10 mg/kg) were tested on mouse exploratory behavior using the open-field test in comparison with rolipram (10 mg/kg) and diazepam (10 mg/kg), used as reference drugs. Compounds 3b, 5b and 7b did not affect mice locomotor activity at the same dose of rolipram and diazepam, which, on the
Table 2 Inhibition activity % toward different PDE4 isoforms of compounds 1a–c, 2, 3a–c, and 4 in comparison with previously reported compounds 5–8, and rolipram as reference compounda
a b c d
Compd
PDE4A4 %inhib
PDE4B2 %inhib
PDE4C2 %inhib
Compd
PDE4A4 %inhib
PDE4B2 %inhib
PDE4C2 %inhib
1a 1b 1c 2b 3a 3b 3c 4b Rolipram
41 68 15 58 36 32 28 34 79
43 68 15 51 31 29 26 37 65
ntb 62 nt nt nt 9 nt nt 57
5a 5bc 5cd 6bd 7ac 7bc 7c 8b —
nt 34 10 41 nt 34 nt nt —
nt 39 5 51 nt 23 nt nt —
nt 19 9 25 nt 25 nt nt —
Results expressed as % inhibition in respect to the control at 10 lM concentration, in duplicate. Not tested. Results already reported.18 Results already reported.19
3430
C. Brullo et al. / Bioorg. Med. Chem. 23 (2015) 3426–3435
Figure 2. Genotoxic potential of compounds 1a, 3b, 4b, and 7b. Western blot analysis of c-H2AX in HTLA cells treated for 24 h with 100 lM etoposide, 7b and the test compounds, or an equal volume of solvent (DMSO). The H2AX signals represent the internal loading control. The figure is representative of three independent experiments all showing essentially similar results.
contrary, significantly, reduced the distance traveled by mice (P
