Discovery of Neuritogenic Compound Classes Inspired by Natural Products

Neuroactive! An enantioselective, catalytic synthesis strategy provides rapid access to natural-product-inspired classes of neuritogenic compounds (see scheme). The goal is to find interesting chemical probes to shed light on neurodevelopmental processes and foster a better understanding of the complex biology and physiology of neuronal development and related neurodegenerative disorders. The progressive degeneration and impaired restoration of neurons are hallmarks of neurodegenerative disorders for which neuroprotective and neurite-growth-promoting small molecules1 are in high demand.2 The characteristic structural scaffolds of classes of neuroprotective or neurotrophic natural products (NPs)3 may define biologically prevalidated starting points in vast chemical structure space inspiring the design and synthesis of compound collections endowed with the same or similar kind of biological activity.4 Different iridoid glycosides5 and sesquiterpenoids6 possess pronounced neuritogenic properties (for relevant examples see Figure 1 a), and therefore their structural scaffolds may provide valuable targets for the synthesis of natural-product-inspired compound libraries endowed with neurite-growth-promoting activity. a) Natural products with neurobiological activities. b) Plan for the synthesis of two focused libraries based on the pentalenone (9) and the cyclopenta[c]pyranone scaffold (10). Here we report the enantioselective synthesis7 of an iridoid-inspired compound collection, its investigation in phenotypic assays monitoring modulations in neurite outgrowth from primary hippocampal neurons and motor neurons derived from mouse embryonal stem cells, and the discovery of two novel natural-product-inspired neuritogenic compound classes. The screen revealed two novel classes of natural-product-inspired neuritogenic compounds. We envisaged that the [3+2] cycloaddition of allene-derived zwitterions 88 and cyclopentenones 7 catalyzed by a chiral phosphine9 would provide pentalenones 9, and that subsequent regioselective Baeyer–Villiger (BV) oxidation10 would yield the iridoid-inspired scaffold 10 (Figure 1 b). Cyclopentenones have not been employed previously in this zwitterionic annulation.11 While cyclopentenone 7 (R=H) did not undergo the cycloaddition, cyclopentenone 11 a reacted smoothly with the zwitterion derived from allene ester 12 a to yield the desired pentalenone ring system 13 a (Scheme 1).The additional carbonyl group embedded in 11 a enhances the reactivity of the cyclopentenone and provides an additional opportunity to increase the diversity of the final products (see below). Synthesis of target scaffolds. Unexpectedly, BV oxidation of 13 a yielded the undesired regioisomer 14 (Scheme 1), indicating that in the presence of an ester group, formation of the tertiary carbocation is still favored. Introduction of an α,α-dimethyl group, which favors the formation of a stabilized carbocation, and exchange of the ester for a more strongly electron-withdrawing ketone efficiently changed the regioselectivity in the BV oxidation.10 Thus, treatment of 13 b with magnesium bis(monoperoxyphthalate) hexahydrate (MMPP)12 yielded the desired iridoid natural-product-inspired scaffold 15 (Scheme 1). For the synthesis of the compound collection, cyclopentenone building blocks 11 (Table 1) (for their synthesis see the Supporting Information) were subjected to the [3+2] cycloaddition in the presence of either tri-n-butylphosphine or the enantiopure bidentate N-acylaminophosphine catalyst 169d (Table 1). Use of γ-methyl allene ester 12 b (R2=Me, entry 2) resulted in a good yield of a cyclopentenone product with an extra methyl group; product 13 c contains three contiguous stereogenic centers including an all-carbon quaternary center. The relative stereochemistry and the absolute configuration of the cycloadducts were ascertained by means of NOE experiments and modified Mosher ester analysis, respectively (for details, see the Supporting Information). Entry Prod. R1 R2 Yield [%] PBu3/16 ee [%] 1 13 b Ph H 66/72 95 2 13 c Ph Me 54/76 90 3 13 d 1-naphthyl H 61/69 94 4 13 e 4-PhC6H4 H 64/71 93 5 13 f 3,5-tBu2C6H3 H 59/77 92 6 13 g 3,4-BrFC6H3 H 54/70 97 7 13 h 3,4-BrFC6H3 Me 51/73 90 8 13 i 3-(4-ClC6H4)C6H4 H 68/81 95 9 13 j 3-(4-ClC6H4)C6H4 Me 85/90 92 10 13 k 4-(BnO)C6H4 H 58/71 94 11 13 l 4-(BnO)C6H4 Me 65/70 91 12 13 m 3,5-MeOC6H3 H 65/80 93 13 13 n 3,5-MeOC6H3 Me 61/75 91 14 13 o 2-benzofuryl H 60/66 96 15 13 p 2-benzofuryl Me 53/71 88 16 13 q 4-(2-methyloxazolyl) H 58/56 98 17 13 r 4-(2-methyloxazolyl) Me 60/55 81 18 13 s 4-pyridyl H 33/– – 19 13 t 4-pyridyl Me 28/– – 20 13 u isopropyl H 70/90 96 21 13 v isopropyl Me 76/92 91 22 13 w cyclohexyl H 70/88 96 23 13 x cyclohexyl Me 79/94 91 24 13 y t-butyl H 74/89 99 25 13 z t-butyl Me 75/88 92 The asymmetric [3+2] annulation in the presence of organocatalyst 16 proceeded in high yield, and with high regio-, diastereo-, and enantioselectivity at room temperature (Table 1). The yields were significantly better than those obtained from the reactions with tri-n-butylphosphine (Table 1), and generally only one regioisomer was formed with high enantiomeric excess (for details see the Supporting Information). Decreasing the reaction temperature to −20 °C did not improve the enantioselectivity but lowered the yields. The [3+2] annulation tolerates a variety of aryl and alkyl substituents appended to the keto group, including bulky, electron-deficient, and electron-rich aryl groups (Table 1, entries 1–13). The reaction also proceeded well with cyclopentenones bearing heterocyclic ketones (Table 1, entries 14–17). Also in these cases catalyst 16 afforded higher yields than tri-n-butylphosphine and provided high to excellent ee values. Cyclopentenone annulation products containing pyridyl ketones were difficult to handle and decomposed during the reaction and purification (Table 1, entries 18 and 19). Alkyl ketones reacted well in the annulation reaction (despite acidic α-protons) and yielded the bicyclic ring system in high yields (Table 1, entries 20–25). The highest enantiomeric excess was obtained for annulation adduct 13 y bearing a tert-butyl ketone (99 % ee; Table 1, entry 24). The broad scope, high yields, and enantioselectivity of the asymmetric [3+2] annulations of zwitterions 8 with cyclic olefins are unprecedented . For the BV oxidation of ketones 13 to the desired lactones 15 the best results were obtained with four equivalents of MMPP in methanol/water (1:1) at room temperature.12 Under these conditions the desired lactone 15 (R1=Ph) was obtained in 40 % yield after six days at room temperature together with the doubly oxidized product 17 (R1=Ph) (Table 2, entry 1). At higher reaction temperature the doubly oxidized products 17 were formed in higher yields (Table 2, entries 6 and 8). In general, the BV oxidation proceeded without formation of further undesired products and yielded the desired lactones incorporating aromatic, heteroaromatic, and alkyl ketones in moderate to high yields (Table 2). Entry R1 Conv. [%] 15 Yield [%] 17 Yield [%] 1 Ph 60[a] 40 20 2 1-naphthyl 49[b] 49 – 3 3,5-tBu2C6H3 66[b] 52 14 4 3,4-BrFC6H3 85[a] 62 23 5 3-(4-C6H4)C6H4 53[a] 53 – 6 3-(4-C6H4)C6H4 80[b] – 80 7 3,5-MeOC6H3 70[a] 70 – 8 3,5-MeOC6H3 80[b] 50 30 9 4-(2-methyloxazolyl) 78[c] 78 – 10 isopropyl 32[b] 32 – 11 cyclohexyl 37[b] 37 – For the synthesis of a focused collection of analogues the functional groups embedded in or appended to the bicyclic scaffolds were subjected to various transformations. Dihydroxylation of the olefin in 13 and 15 proceeded smoothly and with complete diastereoselectivity providing highly substituted diols 18 and 20, respectively, in excellent yields. Selective acylation of the secondary alcohol in 18 was successfully achieved providing ester 19. Olefin epoxidation in 13, 15/17 smoothly provided 21 and 22, respectively in excellent yields (see the Supporting Information for details). Compounds 13 apparently favored a completely stereoselective epoxidation (21, Scheme 2). Epoxide ring-opening in 21 was performed with methanol as the nucleophile under mildly acidic conditions leading to 23. DibalH reduction of the keto group in 15 provided the iridoid scaffold 24 incorporating a secondary alcohol. Interestingly, upon attempted formation of an acetal, scaffold 15 underwent decarboxylation to yield the product 25 in excellent yield (Scheme 2). Functionalization of the bicyclic scaffolds. Reaction conditons: a) OsO4, NMO, THF/H2O, RT; b) R3COCl or R3CO2H, RT; c) mCPBA, RT; d) pTSA, MeOH, 45 °C, 3 days; e) DibalH, −78 °C; f) pTSA, (CH2OH)2, RT. NMO=N-methylmorpholine N-oxide, mCPBA=meta-chloroperbenzoic acid, pTSA=para-toluenesulfonic acid, DibalH=diisobutylaluminum hydride. The compound collection was screened for neurite-outgrowth-promoting activity in tests employing primary neurons. Although more sensitive and technically more demanding to manipulate and assay, the use of primary neuronal cultures of hippocampi from E18/E19 Sprague Dowley rats was preferred over frequently employed model cell lines, in particular PC12 cells that have a tumorigenic origin.13 During the first days in cell culture, individual neuronal processes emanating from hippocampal neurons can be visualized in their entirety, thus allowing a direct observation of, for example, growing neurites, axons, and the degree of branching. Initially the primary neurons were treated with racemic molecules as DMSO solutions at 10 μM and 1 μM concentration for two days; then the cells were fixed and stained with a membrane dye, and the overall membrane formation was quantified by spectrophotometric readout at 550 nm (see the Supporting Information for details). Enhanced fluorescence relative to that of the DMSO control thus provides a direct measurement of the enhanced neurite outgrowth induced by the compounds. Gratifyingly, the screen identified compounds 18 b and 17 b as interesting hits that substantially enhanced the overall formation of neuronal membranes. Neurons with an enhanced overall membrane signal can have very different morphology which cannot be discriminated without a specific analysis which dissects dendritic branching. In order to visualize the complexity of the neuronal network developed as a consequence of compound treatment, the mean total length of the neurites should be judged against the mean total number of neurites (Figure S1 in the Supporting Information). Thus, the enantiomers 17 b and racemic (±)-18 b were subjected to further indepth analysis. Quantitative analysis of confocal images acquired after exposure for five days revealed that (±)-18 b (1 μM) significantly induced the formation of more complex neuronal networks (Figure 2). Racemic (±)-18 b enhanced the mean total neurite length per cell primarily because it generates a higher number of neurites per cell than BDNF does (Figure 2 a,b). The enantiomers of 17 b displayed remarkable differences in neurotrophic activity. While (R,R)-17 b enhanced the mean single neurite length by twofold relative to the DMSO control (Figure 2 c,d), enantiomeric (S,S)-17 b was cytotoxic as indicated by membrane blebbing (see Figure S2 in the Supporting Information). Neurite-outgrowth-promoting activity on primary cultured rat hippocampal neurons and quantification of morphological changes. a) Morphology of neurons treated with DMSO (negative control; 1:10 000 dilution), BDNF (positive control), and racemic 18 b (1 μM); b) comparative effect of 18 b on the number and length of neurites; c) morphology of neurons treated with DMSO (negative control; 1:1000 dilution), BDNF (positive control), and enantiopure (R,R)-17 b (10 μM); d) comparative effect of (R,R)-17 b on the number and length of neurites.; scale bar: 100 μm. BDNF=brain-derived neurotrophic factor, DMSO=dimethylsulfoxide, CTRL=control, untreated cells. In light of the encouraging neurite-growth-enhancing effects on primary hippocampal neurons we investigated whether the compound collection also contains members that modulate the differentiation and development of motor neurons derived from mouse embryonal stem cells (ESCs). Compounds endowed with such properties are potentially relevant for applications in regenerative medicine and their discovery is challenging. Glial cell line derived neurotrophic factor (GDNF), which is important for neuronal development, neurogenesis, and neuronal survival14 and in neurodegenerative disorders,15 was used as a positive control. The neuronal complexity resulting after the treatment with the compounds (10 μM) was quantitatively analyzed by measuring parameters for neuronal cells that characterize mean and total neurite length, neurite branching, mean and total neurite count, and survival (event type 1 neuron count) of neurons (Figure 3). Neurite-outgrowth-promoting activity and quantification of morphological changes for ESC-derived motor neurons. a) Morphology of neurons treated with 10 μM of compounds (2 days), DMSO, and GDNF (negative and positive controls, respectively); b) chemical structures of active molecules; c) eight different parameters analyzed for differentiation, scale bar: 40 μm. Error bars show standard deviation (n=4). P values were determined following two-tailed Student’s t-test. For all the individual morphological changes analyzed, P<0.05 was observed (except in the case of branch point total count for 18 m and 18 d) using DMSO versus compound and GDNF control (data not shown). A P value of <0.05 was considered significant. Several compounds induced neuronal outgrowth as early as after two days of treatment. Interestingly, (R,R)-17 b, which was active on primary hippocampal neurons, also induced neuronal outgrowth and differentiation of ESC-derived motor neurons (see Figure S3 in the Supporting Information). The most remarkable results were obtained with the racemic pentalenones 18 m and 18 d and the iridoid-inspired epoxide 22 g. The neuronal complexity induced by 18 m emerged mainly because of extensive neurite branching observed for an enhanced number of total neurites. In contrast to 18 m, which induced shorter axons attached to their somas (event type 1 neuron count), 18 d decorated with a naphthyl ketone and an additional methyl substitution on the pentandiol ring scored better than GDNF in increasing the neurite length, branching, and number per wells (Figure 3 a–c). The iridoid-inspired epoxide 22 g induced the highest degree of neuronal complexity in ESC-derived motor neurons. While it increased the total and mean neurite length 1.5-fold as compared to that after treatment with GDNF alone, the number of neurites per well and the type 1 neuron count were increased by more than twofold. In conclusion, we have developed a concise, enantioselective synthesis of natural-product-inspired compound collections endowed with neurotrophic properties which may yield interesting chemical probes to shed new light on neurodevelopmental processes and foster a better understanding of the complex biology and physiology of neuronal development and related neurodegenerative disorders. Dedicated to the Bayer company on the occasion of its 150th anniversary As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.