Discovery of Fluorine-Containing Benzoxazinyl-oxazolidinones for the Treatment of Multidrug Resistant Tuberculosis

Abstract

A novel series of fluorine-containing benzoxazinyl-oxazolidinones were designed and synthesized as antidrug-resistant tuberculosis agents possessing good activity and improved pharmacokinetic profiles. Compound 21 exhibited not only outstanding in vitro activity with a MIC value of 0.25–0.50 μg/mL against drug-susceptible H₃₇Rv strain and two clinically isolated drug-resistant Mycobacterium tuberculosis strains, but also acceptable in vitro ADME/T properties. Moreover, this compound displayed excellent mouse pharmacokinetic profiles with an oral bioavailability of 102% and a longer elimination half-life of 4.22 h, thereby supporting further optimization and development of this promising lead series.

Tuberculosis (TB), one of the top 10 causes of death worldwide in 2015,¹ is caused by the infection of Mycobacterium tuberculosis (MTB). In recent years, the emergence of multidrug-resistant tuberculosis (MDR-TB), extensively drug-resistant tuberculosis (XDR-TB), totally drug-resistant tuberculosis (TDR-TB), and even coinfections with HIV create a serious global challenge for treatment of this fatal disease.^2,3 The prolonged treatment of TB, adverse drug reactions, and drug–drug interactions are commonly encountered, especially to the second-line medicines and in HIV coinfected individuals on antiretroviral treatment. Consequently, there is a pressing need for new antituberculosis agents effective against drug-resistant TB.^4,5 The current situation even necessitates the re-engineering and repositioning of some old drug families to achieve effective control.

The oxazolidinones is a new class of antibacterial protein synthesis inhibitors that block translation through a unique mechanism by binding to 23S RNA in the 50S ribosomal subunit of bacteria.^6,7 Linezolid^8,9 (Figure 1), the first marketed oxazolidinone developed to treat Gram-positive bacterial infection with a dose of 600 mg twice daily in adults, has an off-label use to treat complicated MDR-TB and XDR-TB with improved outcome.^10,11 In addition, sutezolid¹² (Figure 1) is undergoing clinical studies for tuberculosis treatment, making the oxazolidinones a class likely to play a key role in future TB treatment regimens. However, prolonged treatment with linezolid is associated with serious neuropathies and myelosuppression, which is mediated by dose- and time-dependent inhibition of mitochondrial protein synthesis (MPS).^13,14 Hence, it is still needed to develop new oxazolidinones with acceptable safety margins through enhancing efficacy and reducing toxicity.

Figure 1.

Structures of representative oxazolidinones for TB treatment.

Considering the very short half-life (less than 1 h) of linezolid after oral administration in mouse,¹⁵ the drug-like scaffold-tricyclic fused benzoxazinyl-oxazolidinone developed by Yang and co-workers^16,17 has attracted our great interest, due to its benefits regarding pharmacokinetics profile. Additionally, the tetrahydropyridine moiety of AZD5847¹⁸ (Figure 1) provided a possibility to reduce molecular planarity and increase molecular hydrophilicity.¹⁹ To our delight, our first fluoro-benzoxazinyl-oxazolidinone compound 4 containing acetylaminomethyl side chain from linezolid and dihydroxy- propinoyl tetrahydropyridine moiety from AZD5847 demonstrated good anti-TB activity with MIC 0.48 μg/mL against MTB H₃₇Rv and very low Vero cytotoxicity with an IC₅₀ of more than 64 μg/mL. Herein, we disclose the synthesis, biological evaluation, and preliminary structure–activity relationship (SAR) studies from the hit-to-lead optimization of these novel fluorine-containing benzoxazinyl-oxazolidinones bearing a tetrahydropyridine moiety.

As shown in Scheme 1, the synthetic routes for target compounds 4, 17–41 were outlined. The synthesis of the key intermediates Ia–f, IIa and IIIa–c is summarized in Scheme S1 and Scheme S2 (Supporting Information). The desired compounds 24, 26, 28, and 36–41 were obtained through nucleophilic substitution reaction of respective intermediates with 2-bromoethanol. The intermediate Ia reacted with cyclopropanecarbonyl chloride to give compound 18. The remaining acylated target compounds were prepared by condensation reaction of respective intermediates with different carboxylic acids. Among them, compounds 22 and 23 were obtained as a hydrochloride form by the deprotection of Boc group using HCl in EtOAc.

Scheme 1. Synthesis of the Target Compounds 4 and 17–41.

Reagents and conditions: (i) cyclopropanecarbonyl chloride, Et₃N, DCM, rt for 18; carboxylic acids, EDCI, HOBt, Et₃N, DMF, rt for the others; (ii) HCl/H₂O, 0 °C to rt for 4; HCl/EtOAc, 0 °C to rt for 22-23; (iii) 2-bromoethanol, Et₃N, DMF, 100 °C.

All the target compounds were evaluated for their activity against MTB H₃₇Rv using the microplate alamar blue assay (MABA).²⁰ The MIC was defined as the lowest concentration effecting a reduction in fluorescence of ≥90% relative to the mean of replicate bacterium-only controls. The compounds with MIC less than 5 μg/mL were further tested for mammalian cell cytotoxicity using Vero cells measured as a concentration inhibiting 50% growth (IC₅₀) as compared to a no-treatment control. Table 1 summarizes the biological data for 26 new benzoxazinyl-oxazolidinones derivatives. Linezolid and AZD5847 were used as reference compounds.

Table 1. Structures, Anti-Tuberculosis Activity, Cytotoxicity, and Selectivity Index (SI) Values for Target Compounds.

^aSelectivity index (SI) = IC₅₀(Vero)/MIC.

^bUsing in the form of hydrochloride.

Inspired by the observation that the privileged scaffold of tricyclic fused benzoxazinyl-oxazolidinone could increase the antibacterial activity and improve pharmacokinetic properties,¹⁶ compound 4 was prepared to investigate the effect of this scaffold on antituberculosis activity. To our delight, as shown in Table 1, compound 4 displayed the same anti-TB activity as linezolid but superior to that of AZD5847. In addition, this compound exhibited low toxicity. With the encouraging observation, a preliminary SAR investigation was carried out to seek the optimal R₂ and R₁ substituents in this new fluorine-containing benzoxazinyl-oxazolidinone scaffold.

Keeping the tricyclic fused scaffold (section B) and the traditional C-5 acetylaminomethyl side chain (section C), which plays a very important role in linezolid and sutezolid,²¹ our initial exploration of the hit compound 4 was focused on the modification of section A (Table 1) to identify the optimum R₂ substituent. A variety of different substituents including acyclic and cyclic alkylated acyl with or without heteroatoms was evaluated at this position. Unfortunately, the results revealed that R₂ without polar substituent like hydroxyl or amino (17–19) or with a bulkier group (20, 23) caused a more than 10-fold loss of potency. However, replacement of the chiral (S)-2,3-dihydroxypropanoyl group with simple achiral 2-hydroxyacetyl group (21) or even 2-hydroxyethyl group (24) displayed equipotent activity compared to the hit compound 4. Compound 22 bearing glycyl group also displayed good activity. From results described above, it appeared that a small flexible hydrophilic group is beneficial for the antituberculosis activity.

As such, the 2-hydroxyacetyl group or 2-hydroxyethyl group in section A and acetylamino group in section C were selected as optimized fragments to probe the impact of the fluorine on the benzene ring and the double bond in the tetrahydropyridine ring on potency and cytotoxicity. As exhibited in Table 1, compound 26 bearing 2-hydroxyethyl group displayed 5-fold less potency than compound 24. Although compound 25 still exhibited moderate anti-TB activity with MIC 0.725 μg/mL compared to compound 21 (MIC 0.391 μg/mL), these two fluorine-free compounds 25 and 26 showed higher cytotoxicity (25 vs 21, 26 vs 24). Interestingly, almost all compounds listed in Table 1 with fluorine on the benzene ring exhibited very low cytotoxicity, no matter what the MIC values are. The advantage of fluorine on the benzene ring in this tricyclic scaffold was confirmed, consistent with the general SAR studies results of oxazolidinone. Replacement of the 4-tetrahydropyridine with 4-piperidine caused a significant loss of activity (27 vs 21, 28 vs 24). This result demonstrated that the double bond is required for potency.

Although it is well-established that C-5 acetylaminomethyl group in linezolid is essential for good antibacterial activity, further explorations on the C-5 side chain (section C) in this new tricyclic fused oxazolidinone were also undertaken. Different C-5 side chain moieties like trifluoroacetamino (29 and 36), methyl carbamate (32 and 39), aminoisoxazole (33 and 40), hydroxyisoxazole (35), [1,2,3]tirazole (34 and 41), cyclopropanecarboxamide (30 and 37), and cyclobutanecarboxamide (31 and 38) had been introduced to replace the acetylamino group in section C. Most compounds showed good to moderate potency with MIC values 0.473–2.223 μg/mL. However, analogues with [1,2,3]tirazole (34 and 41) and hydroxyisoxazole (35) completely lost the anti-TB activity. The results revealed that N–H is essential for the maintenance of activity.

As a result of their high potency against H₃₇Rv strain and superior selectivity index values, five compounds including 4, 21, 22, 24, and 37 were selected to investigate their anti-DR-TB activity. As summarized in Table 2, all tested compounds displayed high in vitro potency to 16892 strain (MDR-TB). More importantly, compound 21 also exhibited good activity against 16802 strain (XDR-TB), supporting it to move forward for further evaluation.

Table 2. In Vitro Activity of the Selected Compounds against Drug Resistant Tuberculosis.

compd	MIC (μg/mL) (H37Rv)	MIC (μg/mL) (16892)a	MIC (μg/mL) (16802)b
4	0.483	0.25	1.00
21	0.391	0.25	0.50
22	0.732	0.25	2.00
24	0.578	0.50	2.00
37	0.473	0.50	1.00
INH	0.037	>40	2.50
RFP	0.057	>40	20.0

^aResistance to isoniazid (INH) and rifampicin (RFP).

^bResistance to isoniazid, rifampicin, streptomycin, ethambutol, and levofloxacin.

To develop a deeper understanding of compound 21’s druggability, more profiling was performed. As depicted in Table 3, compound 21 exhibited high IC₅₀ value against HepG2 cell, indicating a lack of toxicity to hepatic cells. Notably, this compound showed almost no inhibition of the hERG K⁺ channel using the patch clamp technique. Due to the main adverse effect of myelosuppression during the long-term use of linezolid, likely related to inhibition of mitochondrial protein synthesis (MPS),¹³ the in vitro MPS inhibition activity was conducted to predict myelosuppression risk. Compound 21 showed a comparable inhibition activity (IC₅₀ 9.19 μM) with linezolid (IC₅₀ 8.71 μM). This outcome suggested that 21 and linezolid have a similar safety profile in terms of myelosup- pression risk. Regarding the in vitro physicochemical properties, compound 21 showed good membrane permeability and excellent metabolic stability against mouse and human liver microsomes. These data therefore indicate that compound 21 may exhibit favorable in vivo PK performance. Indeed the PK of 21 in Balb/c mouse was outstanding (Table 4), with high maximal plasma concentration (C_max = 30.0 μg/mL), high plasma exposure (AUC_0–∞ = 122 μg·h/mL), long elimination half-life (t_1/2 = 4.22 h), and excellent oral bioavailability (F = 102%) after oral administration.

Table 3. Representative Properties of Compound 21.

				MLMb		HLMc
HepG2 cytotoxicity IC50 (μg/mL)	hERG K+ inhibition IC50 (μM)	MPSa inhibition IC50 (μM)	Caco-2 Papp (×10–6 cm/s)	substrate remaining (%)d	stabilitye	substrate remaining (%)d	stabilitye
>64	>30	9.19	4.14 ± 0.57	102.0	stable	103.0	stable

^aMitochondrial protein synthesis.

^bMouse liver microsome.

^cHuman liver microsome.

^dSubstrate concentrations were determined in incubations with NADPH after 30 min and normalized to concentrations at time zero.

^eStability was determined without the NADPH cofactor.

Table 4. Mouse Pharmacokinetic Properties of Compound 21.

compd	route	dose (mg/kg)	Cmax (μg/mL)	Tmax (h)	t1/2a (h)	AUC0-∞ (μg·h/mL)	MRT(0-∞)b (h)	clearance (mL/h/kg)	Fc (%)
21	po	100	30.0	2	4.22	122	4.12		102
	iv	10	11.4	0.03	8.23	12	2.54	836

^aPlasma elimination half-life.

^bMean residence time.

^cBioavailability.

In summary, a new series of fluorine-containing benzoxazinyl-oxazolidinones were prepared and characterized. Preliminary exploration of structure–activity relationships on the scaffold revealed that the fluorine atom on the benzene ring has dual function in improving activity and reducing toxicity. Compound 21 displayed promising antituberculosis activity against drug-sensitivity H₃₇Rv and drug-resistant strains (MIC = 0.25–0.50 μg/mL) and an adequate preliminary in vitro safety profile with negligible cytotoxicity and cardiotoxicity. Importantly, compound 21 exhibited notable oral bioavailability (102%) in mouse as well as relatively long half-life, which may help to reduce the dosage and frequency of medication and improve patient compliance. Hence, compound 21 as a promising antituberculosis potential lead compound, warranting further evaluation, will be reported in due course.

Glossary

ABBREVIATIONS

ADME/T

absorption, distribution, metabolism, excretion, and toxicity

Boc

tert-butoxycarbonyl

DCM

dichloromethane

DMF

N,N-dimethylforamide

EDCI

1-(3-(dimethylamino)propyl)-3-ethylcarbodiimide hydrochloride

HOBt

1-hydroxybenzotriazole

PK

pharmacokinetic

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsmedchemlett.7b00068.

• Experimental procedures, compounds characterization, minimum inhibitory concentration assay, in vitro ADME/T, and PK evaluation (PDF)

Author Contributions

^⊥ These authors contributed equally to this work.

The research is supported in part by the National Science & Technology Major Project of China (Grant 2015ZX09102007-013), the National Natural Science Foundation of China (Grant 81502917), and the Fundamental Scientific Research Fund of Institute of Materia Medica (Grant 2014CX15).

The authors declare no competing financial interest.