A practical synthesis of amino limonin/deoxylimonin derivatives as effective mitigators against inflammation and nociception

A practical synthetic route, consisting of 5 steps, has been developed and applied successfully for converting limonin/deoxylimonin into the corresponding amino derivatives I-5a–I-5e and II-5a–II-5e.

Abstract

A practical synthetic route, consisting of 5 steps, has been developed and applied successfully for converting limonin/deoxylimonin into the corresponding amino derivatives I-5a–I-5e and II-5a–II-5e. Deoxylimonin analogs II-5a and II-5b bearing an open A ring structure underwent ring closure reaction by employing the Mitsunobu procedure to afford II-6a and II-6b. All of the 12 newly synthesized compounds were subjected to preliminary screening for evaluating their inflammation and nociception inhibition efficacy. The most promising compounds, I-5b and II-5d, were selected for further investigation by a carrageenan-induced mouse paw edema model, both of which displayed a dose–response dependent manner and demonstrated superior anti inflammation capacity to that of indomethacin in the first 2 hours.

Introduction

Limonoids, mainly existing in the Rutaceae and Meliaceae plant families, are a group of highly oxygenated triterpenoids.1 As shown in Fig. 1, these typical limonoids, limonin, nomilin, obacunone and ichangensin, all contain the 4,4,8-trimethyl-17-furyl steroidal signature-skeleton, and a vast number of limonoid family products either possess this unique structure or have oxidized or rearranged skeletons derived from such precursors.2 The pharmacological studies on these abundant naturally occurring products suggested that many limonoids display a broad scope of biological activities, such as anti-inflammatory,3,4 analgesic,3,4 antitumor,5–8 anti-microbial9 and anti-feedant10 activities.

Fig. 1. Structures of limonoids.

For the past fifty years, tremendous efforts have been made by medicinal chemists towards converting these limonoids into therapeutics. However, the poor aqueous solubility of limonoids (limonin < 0.005 mg mL^–1),11 caused by the cyclic triterpenoid structure, results in low bioavailability, which always impedes scientific research and medical application. On the other hand, the cyclic triterpenoid structure of limonin, possessing C-7 ketone, C17 furan and lactone, makes it amenable to chemical modification. Our lab has attempted to chemically transform limonin into corresponding alkaloids by either introducing an amide group on the A ring to give compound 4,12 adding tertiary amines onto the C(7)-position on the B ring to afford compound 2 (ref. 11, 13 and 14) or incorporating alkylamines on the C17 furan ring to produce compound 3 (ref. 15) (Fig. 2). Most of the resulting limonin derivatives showed optimized antinociceptive and anti-inflammatory activity with favorable physicochemical properties. Interestingly, pioneering research showed that the enhanced biological activity doesn't solely depend on the optimized biophysical properties.11 The diversified structures16,17 also plays a critical role that involves adopting alternative binding pattern with protein, changed in vivo stability, all of which caused by the different structures could render distinct anti-inflammation effects. Thus, diversified limonoid portfolio, either by enzyme metabolism18,19 or chemical synthesis,11–13,15 could enable medicinal chemists to advance the discovery of novel anti-inflammatory agents. In this study, we developed a detour but still a practical synthetic route so as to insert an amino group into the A ring, which led to the discovery of open A ring amino limonin analogues. Moreover, by applying the Mitsunobu reaction, deoxylimonin derivatives were further converted into their conformationally restrained cyclic amino counterparts. The newly synthesized limonin/deoxylimonin analogues, which could be easily prepared into hydrochloric acid salt forms harboring substantially increased water solubility and exhibiting improved anti-inflammatory and analgesic capacities, could serve as promising drug candidates.

Fig. 2. Chemical modification of limonin leads to structurally diverse limonin derivatives.

Results and discussion

Synthesis

The synthetic routes for the proposed limonin derivatives are illustrated in Schemes 1 and 2. As shown in Scheme 1, the preparation of target molecules I-5a–I-5e commenced with the amide exchange on the A lactone ring of limonin (I) with corresponding substituted phenethylamine or benzylamine, which afforded intermediates I-1a–I-1e. Although amide reduction has proven to be a well established method to prepare alkylated amines, the process often requires high hydrogenation pressures and reaction temperatures to be effective (i.e. often requiring pressures above 197 atm and temperatures exceeding 200 °C).20 Alternative reducing agents, such as lithium aluminium hydride or lithium borohydride, are able to affect this reaction.21 However, other susceptible functional groups, lactone, hexanone and furan, can't survive under such harsh reaction conditions or with overwhelmingly reductive agents. Therefore, we designed a detour strategy by first selective thiolation of the amide followed by reduction of thioamide under mild conditions to yield the target molecules. Although I-1a–I-1e could be converted to the corresponding thioamides upon treatment with Lawesson's reagent (1 equiv., toluene), the free hydroxy group on C-19 reacted competitively with Lawesson's reagent, which inevitably yielded unwanted thiolated side products. To prevent the competitive thiolation, tert-butyldimethylsilyl chloride (TBSCl) and trimethylsilyl chloride (TMSCl) were applied to convert the hydroxy group into silyl ether. Unfortunately, oversized silyl protecting groups failed to fit in the congested open A ring, subsequently unable to reach the hydroxy group on C-19 of I-1a–I-1e. Thereafter, di-tert-butyl dicarbonate (Boc₂O), providing a longer linker between the bulky tert-butyl group and C-19 hydroxy group, was successfully applied to transform the hydroxy group into carbonate (I-2a–I-2e), and subsequent thiolation with Lawesson's reagent (1 equiv., toluene, 55–60 °C, 2–3 h) provided I-3a–I-3e in superb conversions (85–92%) with complete selectivity for the amide. Treatment of I-3a–I-3e with RANEY®-Ni in the presence of a H₂ flow afforded the reduced intermediates I-4a–I-4e, which then underwent the decarbonation with HCl saturated methanol to yield exclusively the target amino limonin derivatives I-5a–I-5e. Deoxylimonin analogs II-5a–II-5e were prepared starting from deoxylimonin (II) in the same manner as I-5a–I-5e, with the alkene on the deoxylimonin remaining untouched in the synthetic strategy.

Scheme 1. Preparation of compounds I-5a–I-5e and II-5a–II-5e. Reagents and conditions: (a) corresponding substituted phenethylamine or benzylamine, THF, reflux. (b) Boc₂O, DMAP, CH₂Cl₂, 0–25 °C. (c) Lawesson's reagent, PhMe, reflux. (d) RANEY®-Ni, H₂, MeOH, 25 °C. (e) HCl saturated MeOH, 0 °C.

Scheme 2. Preparation of compounds II-6a and II-6b. Reagents and conditions: (a) DEAD, PPh₃, anhydrous PhMe, 0–25 °C.

The ring closure reaction was efficiently performed by employing the Mitsunobu procedure22 as depicted in Scheme 2. Treatment of II-5a and II-5b with diethyl azodicarboxylate (DEAD) and triphenylphosphine in dry toluene at 0 °C, followed by warming up the reaction mixture to 25 °C over 2 hours gave the deoxylimonin analogs II-6a and II-6b quantitatively. Interestingly, only the substituted benzylamino derivatives (II-5a and II-5b) were successfully transformed into the closed ring products (II-6a and II-6b). The substituted phenethylamino compounds (II-5c and II-5d), sharing the same functional moiety on the aromatic ring as II-5a and II-5b, have higher electron density on the nitrogen, which are theoretically prone to nucleophilic substitution resulting in the removal of triphenyl phosphate to give the closed ring counterparts. In practice, we were unable to obtain these closed ring molecules. Hypothetically, the oversized phenethylamino group impeded the ring closure reaction. As a consequence, the Mitsunobu cyclization of II-5c and II-5d possessing the bulkier substituted phenethylamino moiety failed to proceed, while it worked for II-5a and II-5b bearing a smaller substituted benzylamino group.

Biological evaluation

All target molecules (I-5a–I-5e, II-5a–II-5e, II-6a and II-6b) were evaluated for in vivo anti-inflammatory effects by utilizing the xylene-induced ear swelling test.23 Limonin (I, 100 mg kg^–1), deoxylimonin (II, 100 mg kg^–1) and naproxen (150 mg kg^–1) were used as the references (Fig. 3A). All the synthetics, limonin derivatives I-5a–I-5e and deoxylimonin analogs II-6a–II-6e, showed an increased inhibition rate than limonin (1). Among all the tested molecules, the anti-inflammatory efficacy of compound II-5d (51.53%) was five times that of the precursor, deoxylimonin (11.23%), and exceeded that of naproxen (25.03%), implying that the alkenyl and 3-methoxyphenethylamino moieties contributed substantially to the improved potency.

Fig. 3. A: Xylene-induced ear swelling test in mice: limonin/deoxylimonin (100 mg kg^–1), naproxen (150 mg kg^–1) and limonin/deoxylimonin derivatives (100 mg kg^–1). B: Acetic acid writhing response in mice: limonin/deoxylimonin (70 mg kg^–1), aspirin (200 mg kg^–1), and limonin/deoxylimonin derivatives (70 mg kg^–1). The inhibition rate was expressed as the percentage change from that of vehicle controls. The experiment was performed in 8 replicates.

In terms of antinociception assay, acetic acid writhing response in mice24 was applied as the model to efficiently screen the analgesic effect of all target compounds (Fig. 3B). The results revealed that compounds I-5a (63.77%), I-5b (81.51%), I-5c (55.80%) and I-5d (58.70%) demonstrated optimized analgesic effects compared to limonin (I) (44.20%) and slightly enhanced efficacy compared to aspirin (51.54%). Unlike the anti-inflammation outcome, the oxygen bridge in limonin (I) was an essential pharmacophore for nociception relief properties.

To confirm the results of the single dose preliminary screening, the most promising compounds, I-5b and II-5d, were subjected to further investigation with the carrageenan-induced mouse paw edema model25 treated with different doses (Fig. 4A and B). Both compounds I-5b and II-5d exhibited dose-dependent inhibition of paw edema during the 4 hour test. It is noteworthy that compounds I-5b and II-5d, even at the lowest dose level (25 mg kg^–1), produced a comparable anti-inflammatory response with that of aspirin at 1 hour. While the efficacy of analogs I-5b and II-5d (25 mg kg^–1 and 50 kg kg^–1) decreased accordingly as time went on, the groups treated with a high dose of I-5b and II-5d (100 mg kg^–1) reached the peak of inhibition of swelling development at 2 hour interval, which displayed longer-lasting effects than the reference drug indomethacin.

Fig. 4. A: Inhibitory effect on carrageenan-induced paw edema of compound I-5b^ (25 mg kg^–1), I-5b^^ (50 mg kg^–1), I-5b^^^ (100 mg kg^–1), indomethacin (22 mg kg^–1) and limonin (50 mg kg^–1). The experiment was performed in 8 replicates. B: Inhibitory effect on carrageenan-induced paw edema of compound II-5d^ (25 mg kg^–1), II-5d^^ (50 mg kg^–1), II-5d^^^ (100 mg kg^–1), indomethacin (22 mg kg^–1) and limonin (50 kg kg^–1). The experiment was performed in 8 replicates.

Conclusions

In summary, a series of modified A-ring limonin and deoxylimonin amino derivatives were prepared by employing a mild and efficient synthetic strategy. The preliminary pharmacological studies demonstrated that most of the synthesized target molecules exhibited improved anti-inflammatory and analgesic activity compared to the lead natural product, limonin. It is noteworthy that compound II-5d exhibited more potent anti-inflammatory effects than naproxen, and compound I-5b displayed stronger analgesic activity than aspirin. Moreover, in-depth evaluation of I-5b and II-5d with the mouse paw edema model revealed the dose-dependent inhibition, and the anti-inflammatory activity of both compounds (100 mg kg^–1) peaked after 2 hours and exceeded that of the reference drug indomethacin. In the short term, the novel limonin/deoxylimonin derivatives (I-5b and II-5d) with favorable anti-inflammatory/analgesic activity could be utilized as chemical probes to elucidate the unspecified underlying biological mechanism of limonoid agents.

Compliance with ethical standards

All animal procedures were performed in accordance with the National Institutes of Health Guidelines “GUIDE FOR THE CARE AND USE OF LABORATORY ANIMALS” and approved by the Animal Ethics Committee of the “China Pharmaceutical University”.

Conflicts of interest

The authors declare no conflict of interest, financial or otherwise.