2-Hydroxy-substituted cinnamic acids and acetanilides are selective growth inhibitors of Mycobacterium tuberculosis

Abstract

Selective chemical hits are required for feeding the initial discovery phase of the anti-tuberculosis therapeutics pipeline. These chemical entities should ideally target novel mechanisms of action in order to tackle drug resistance in Mycobacterium tuberculosis. In this work, hydroxycinnamic acid and acetamidophenol skeleta were employed for assessing the effects of constitutional isomerism on in vitro anti-TB activity. The whole cell evaluation of minimum inhibitory concentration values of different substituted cinnamic acids and acetamidophenols showed that the free ortho hydroxyl group conferred both potency and selectivity. Both 2-coumaric acid and 2-acetamidophenol showed minimum inhibitory concentration below 150 μM against M. tuberculosis H₃₇Rv and selectivity index higher than 30.

---

Introduction

Tuberculosis (TB) continues to be a great scourge for humanity. According to the last report from the WHO, in 2009 almost 10 million children were orphaned as a consequence of parental death caused by TB.¹ Drug-resistant strains of the causal pathogen Mycobacterium tuberculosis, are rising everywhere and there were recently reports of isolates resistant to all the tested drugs in Iran, Italy and India, which once again triggered alarm.² These drug-resistant strains of M. tuberculosis are extremely difficult to treat. In HIV-prevalent regions, infection by drug-resistant TB almost always has fatal consequences.³ Certainly there is an urgent need for new anti-TB drugs that are effective in reducing the length of treatment and for combatting drug-resistant strains. Target-based exploration efforts have been unfruitful in antimicrobial drug discovery.^4,5 The reasons for this failure have been attributed to the difficulty in crossing the bacterial cell wall to reach their intracellular targets, or to the physicochemical properties of the members of the screening libraries which are not ideal for targeting microbial pathogens.⁶ In fact, historically all of the antibiotics and current drugs in the pipeline against TB have been developed based on their whole–cell activity.

Cinnamic acid, which is typically found in cinnamon (Cinnamomum verum) and storax (Liquidambar orientalis), was used in the 19th century for treating TB.⁷ Substituted cinnamic acids have been recently reported as novel scaffolds for the development of selective anti-TB agents.^8–10 In particular, O-prenylation and O-geranylation of 4-hydroxycinnamic acids have been found to accord increased anti-TB activity in a series of thioester, amide and azide derivatives.⁹ In this work we evaluated different positional isomers of hydroxycinnamic (coumaric) acid with the aim of extending the study of O-prenylation and O-geranylation of coumarates on anti-TB activity. In addition the positional isomers of 4-acetamidophenol, a widely used and relatively safe analgesic known as paracetamol,¹¹ were also included in this study to examine a possible trend of anti-TB activity among positional isomers of aromatic compounds. This class has never been reported to display anti-mycobacterial properties.

Results and discussion

Chemistry

The synthesis of 4-, 3- and 2-O-prenyl-substituted coumaric acids 4a–c was carried out by a slight modification of the synthetic route reported by De et al. in 2011.⁹ Firstly, coumaric acids 1a–c (Fig. 1) were esterified under Fischer conditions, to yield the corresponding methyl coumarates 2a–c (Scheme 1). The O-prenylated methyl coumarates 3a–c were obtained by refluxing the esters in acetone with 3,3-dimethylallyl bromide and potassium carbonate. Several attempts to deprotect the methyl esters 3a–c to yield the carboxylic acids using potassium carbonate under different conditions were unsuccessful. Saponification was successfully performed by treating 3a–c with an excess of lithium hydroxide in a methanol/water (3 : 1) mixture at room temperature for 16 h which yielded the O-prenylated coumaric acids 4a–c in satisfactory yields (Scheme 1). The O-geranyl derivatives of methyl 4-coumarate and 4-coumaric acid, 5a and 5b (Fig. 2), were prepared following the same procedure described in Scheme 1, using geranyl bromide instead of prenyl bromide. For a description of experimental procedures and characterisation data refer to Supplementary Information.

Fig. 1 Structure of the commercially available cinnamic acids and acetamidophenols used in this study. Chemical structures of 4-coumaric acid (1a), 3-coumaric acid (1b), 2-coumaric acid (1c), cinnamic acid (1d), 3,4-methylenedioxycinnamic acid (1e), 4-acetamidophenol (6a), 3-acetamidophenol (6b) and 2-acetamidophenol (6c).

Scheme 1 Synthetic route for prenylation of the structural isomers of coumaric acid. Reagents and conditions: (i) H₂SO₄, MeOH, reflux, 16 h, >97%; (ii) 3,3-dimethylallyl bromide, K₂CO₃, acetone, reflux, 24 h, 37–91%; (iii) LiOH, MeOH/H₂O, rt, 16 h, 42–96%.

Fig. 2 Structures of the geranylated 4-coumaric derivatives. Chemical structures of methyl O-geranyl-4-coumarate (5a) and O-geranyl-4-coumaric acid (5b).

Biological activity

Minimum inhibitory concentration (MIC) values were determined against both M. tuberculosis H₃₇Rv and M. bovis BCG using the spot culture growth inhibition assay (SPOTi) in 24 or 96 well-plates.^12,13 Half growth inhibitory concentrations (GIC₅₀) against RAW264.7 murine macrophages enabled the evaluation of cytotoxicity and anti-TB selectivity. Among the coumaric acids 1a–c, clearly the most active was the ortho substituted 1c with an MIC value of 122 μM (Table 1), followed by the para substituted 1a (MIC = 244 μM) and then the meta substituted 1b (MIC = 366 μM) against M. tuberculosis H₃₇Rv. The same trend of activity was observed against M. bovis BCG. The three positional isomers 1a–c showed comparable cytotoxicity with GIC₅₀ values between 4560 and 5294 μM. The most selective coumaric acid was the 2-hydroxycinnamic acid (1c), achieving an SI value higher than 37.

Table 1 Anti-tubercular activity and cytotoxicity of substituted cinnamic acids (1a–e, 2a–c, 3a–c, 4a–c, 5a–b) and acetamidophenols (6a–c) against M. tuberculosis H₃₇Rv, M. bovis BCG and murine macrophage RAW264.7 cell line. Isoniazid (INH) and rifampin (RIF) were included as positive controls

Entry	MIC (μM) H37Rv	MIC (μM) BCG	GIC50 (μM) RAW264.7	Selectivity index (SI)
1a	244	366	4560 ± 160	18.7
1b	366	609	5294 ± 226	14.5
1c	122	244	4563 ± 310	37.4
1d	270	405	6547 ± 205	24.2
1e	312	416	2774 ± 178	8.80
2a	224	337	609 ± 194	2.72
2b	224	561	850 ± 148	3.80
2c	112	112	1125 ± 247	10.0
3a	81.2	60.9	487 ± 183	6.00
3b	244	244	2972 ± 110	12.2
3c	>406	>406	2598 ± 151	<6.40
4a	86.1	86.1	1339 ± 172	15.6
4b	172	258	1296 ± 302	7.53
4c	258	258	1683 ± 124	6.52
5a	127	254	159 ± 105	1.25
5b	66.8	134	210 ± 82	3.14
6a	>661	>661	1002 ± 68	<1.52
6b	>661	>661	5338 ± 602	<8.08
6c	132	132	4204 ± 265	31.8
INH	0.73	0.73	21 900	30 000
RIF	60.7 nM	60.7 nM	850	14 000

---

Cinnamic acid (1d) completely inhibited M. tuberculosis H₃₇Rv growth at 270 μM (Table 1), showing a GIC₅₀ of 6547 μM against the macrophage cell line. Its selectivity index (SI) was 24. 3,4-Methylenedioxycinnamic acid (1e) showed an MIC value of 312 μM, however it was the most cytotoxic of the commercially available cinnamic acids with a GIC₅₀ value of 2774 μM, attaining an SI less than 10.

Methyl coumarates 2a–c displayed MIC values (Table 1) comparable to the corresponding carboxylic acids 1a–c, with values ranging from 112 to 224 μM. However the esters were significantly more cytotoxic than the free carboxylic acids having GIC₅₀ values between 609 and 1125 μM. Among the methyl coumarates, the ortho-substituted 2c was the most potent and selective, with an MIC value of 112 μM and an SI of 10.

O-Prenylated methyl 4-coumarate 3a inhibited the growth of M. tuberculosis H₃₇Rv at a concentration of 81.2 μM, whereas the meta and ortho derivatives, 3b and 3c, were less active with MIC values of 244 and >406 μM respectively. However 3a displayed high cytotoxicity with a GIC₅₀ value of only 487 μM, achieving a low SI of 6. O-Prenylation increased the toxicity of the para isomer (from 609 to 487 μM) whereas the meta and ortho isomers reduced their cytotoxicity (from 850 to 2972 μM, and from 1125 to 2598 μM respectively) compared to the non-prenylated compounds. O-Geranylation of 2a yielded 5a which was found to be the most cytotoxic of its class, with a GIC₅₀ of 159 μM, without significant gain in anti-mycobacterial activity. These results led us to hypothesise that the para-O-prenyl/geranyl derivatives may act by a different biological mechanism compared to the meta and ortho-O-prenyl derivatives, because O-prenylation of the para derivative increased cytotoxicity whereas the opposite effect was observed for the meta and ortho methyl coumarates.

The O-prenyl 4-coumaric acid 4a was equally as active as the ester 3a displaying an MIC value of 86.1 μM against both M. tuberculosis H₃₇Rv and M. bovis BCG (Table 1). This free carboxylic acid derivative showed higher selectivity against mycobacterial cells than mammalian macrophages with an SI of 15.6. The other prenylated coumaric acids 4b and 4c were less active, with MIC values between 172 and 258 μM, achieving SI values lower than 10. The O-geranyl 4-coumaric acid 5b was more active than the methyl ester 5a, with an MIC value of 66.8 μM, but was toxic, achieving a low SI value of 3.

The presence of unsaturation on the propenyl chain of the cinnamic acids was found to be essential for anti-TB activity. All of the five dihydro derivatives of 1a–e, obtained by catalytic hydrogenation using Pd/C in methanol (data not shown) were found to be inactive against M. bovis BCG with MIC values higher than 600 μM. Moreover the free phenolic OH was also found to be important for anti-TB activity and selectivity. Zosteric acid (Fig. 3), a naturally occurring para-sulfooxy (–OSO₃H) derivative of coumaric acid is known for its potent antifouling and antimicrobial effects.¹⁴ This natural compound and its ortho and meta constitutional isomers were prepared by treating 1a–c with sulphur trioxide/pyridine complex¹⁵ (data not shown). All of the sulfooxy derivatives were devoid of growth inhibitory activity against both M. tuberculosis H₃₇Rv and M. bovis BCG.

Fig. 3 Chemical structures of 1c, 6c, PAS, and zosteric acid. Chemical structures of 2-coumaric acid (1c), 2-acetamidophenol (6c) and para-aminosalycilic acid (PAS) and zosteric acid.

The constitutional isomers of the acetamidophenols were also screened for their anti-TB activity and cytotoxicity. The para and meta isomers 6a–b were inactive, with MIC values higher than 661 μM against M. tuberculosis H₃₇Rv and M. bovis BCG. However the ortho isomer 6c was a notable inhibitor of mycobacterial growth, with an MIC value of 132 μM against both species. This compound displayed a RAW264.7 GIC₅₀ close to 4200 μM and its SI was calculated to be 32. Surprisingly the para isomer (which is widely used as analgesic) was far more cytotoxic than the other structural isomers.

Compounds 1a–1e, 2a–c, 3a–c, 4a–c, 5a–b and 6a–c were tested against the rapid-growing species Mycobacterium smegmatis mc²-155, each displaying MIC values higher than 600 μM. The narrow spectrum of activity of 1c and 6c suggested the possibility that these compounds operate by interfering with a unique biological mechanism present specifically in members of the M. tuberculosis complex.

The most potent molecules evaluated in this study, 2-coumaric acid (1c) and 2-acetamidophenol (6c) share structural similarity with 4-aminosalicylic acid (PAS), a well known second-line anti-TB drug (Fig. 3). It can be hypothesised that 2-coumaric acid and 2-acetamidophenol may act via a similar biological mechanism to PAS, either by interference with the folate pathway¹⁶ or with iron uptake.¹⁷ In addition, the structural similarity serves to support predictions of potentially more active anti-TB agents. Introducing amino groups to 1c and 6c in the meta position to the phenolic hydroxyl (and para with respect to the other substituent), leading to 4-amino-2-coumaric acid and 2-acetamido-5-aminophenol respectively, could be the first step for obtaining a new generation of anti-tuberculars. It seems that the minimum structural requirement for anti-TB activity of this group is the presence of an aromatic nucleus with a conjugated group in position ortho to a free phenolic hydroxyl. Further structural requirements for increasing potency on this basic skeleton will be elucidated in future studies.

Conclusions

Among a variety of cinnamic acids, 2-coumaric acid (1c) was identified as the most potent and selective anti-TB agent with an MIC value of 122 μM against M. tuberculosis H₃₇Rv and an SI value of 37. Methyl esterification and O-prenylation of coumaric acids, yielded derivatives usually more toxic and/or less active than the parent compounds. Structural isomers of acetamidophenol were also screened for anti-TB activity, and interestingly 2-acetamidophenol was found to be the most active, with an MIC value of 132 μM against M. tuberculosis H₃₇Rv and a SI value of 32. The presence of an aromatic nucleus substituted with a conjugated group in the position ortho to a free phenolic hydroxyl were found to be the essential structural characteristics in this preliminary structure-activity study.

Acknowledgements

The authors are grateful to Professor Simon Croft at the London School of Hygiene and Tropical Medicine for his support to Containment Level III Lab access. We also acknowledge Mr Emanuel Samuel at UCL School of Pharmacy for his technical support.

Notes and references

1. Global Tuberculosis Control: WHO report 2011, World Health Organization, 2011.
2. S. Loewenberg, Lancet, 2012, 379, 205.
3. N. R. Gandhi, N. S. Shah, J. R. Andrews, V. Vella, A. P. Moll, M. Scott, D. Weissman, C. Marra, U. G. Lalloo and G. H. Friedland, Am. J. Respir. Crit. Care Med., 2010, 181, 80–86.
4. M. Scott D, Biochem. Pharmacol., 2006, 71, 1096–1102.
5. D. J. Payne, M. N. Gwynn, D. J. Holmes and D. L. Pompliano, Nat. Rev. Drug Discovery, 2007, 6, 29–40.
6. S. T. Cole and G. Riccardi, Curr. Opin. Microbiol., 2011, 14, 570–576.
7. J. P. Warbasse, Ann. Surg., 1894, 19, 102–111.
8. G. K. Yoya, F. Bedos-Belval, P. Constant, H. Duran, M. Daffé and M. Baltas, Bioorg. Med. Chem. Lett., 2009, 19, 341–343.
9. P. De, G. Koumba Yoya, P. Constant, F. Bedos-Belval, H. Duran, N. Saffon, M. Daffé and M. Baltas, J. Med. Chem., 2011, 54, 1449–1461.
10. M. D. Kakwani, P. Suryavanshi, M. Ray, M. G. R. Rajan, S. Majee, A. Samad, P. Devarajan and M. S. Degani, Bioorg. Med. Chem. Lett., 2011, 21, 1997–1999.
11. K. Pavelka, Int. J. Clin. Pract., 2004, 58, 5–12.
12. A. Gupta and S. Bhakta, J. Antimicrob. Chemother., 2012, 67, 1380–1391.
13. J. D. Guzman, D. Evangelopoulos, A. Gupta, K. Birchall, S. Mwaigwisya, B. Saxty, T. D. McHugh, S. Gibbons, J. Malkinson and S. Bhakta, BMJ Open, 2013, 3, e002672–e002685.
14. F. Villa, B. Pitts, P. Stewart, B. Giussani, S. Roncoroni, D. Albanese, C. Giordano, M. Tunesi and F. Cappitelli, Microb. Ecol., 2011, 62, 584–598.
15. F. Villa, D. Albanese, B. Giussani, P. S. Stewart, D. Daffonchio and F. Cappitelli, Biofouling, 2010, 26, 739–752.
16. V. Mathys, R. Wintjens, P. Lefevre, J. Bertout, A. Singhal, M. Kiass, N. Kurepina, X.-M. Wang, B. Mathema, A. Baulard, B. N. Kreiswirth and P. Bifani, Antimicrob. Agents Chemother., 2009, 53, 2100–2109.
17. K. A. Brown and C. Ratledge, Biochim. Biophys. Acta, Gen. Subj., 1975, 385, 207–220.

---

Footnote

† Electronic supplementary information (ESI) available: Materials, synthetic procedures, experimental data of compounds 2a–c, 3a–c, 4a–c, 5a–b and biological evaluation methods can be found as ESI. See DOI: 10.1039/c3md00251a

---