3-Benzylidene-4-chromanones: a novel cluster of anti-tubercular agents

Abstract

In a quest for developing novel anti-tubercular agents, a series of 3-benzylidene-4-chromanones 1a–l were evaluated for growth inhibition of Mycobacterium tuberculosis H₃₇Rv. Three promising compounds 1d, g, j emerged as the lead compounds with the IC₅₀ and IC₉₀ values of less than 1 µg/mL. Evaluation of the potent compounds 1d, g, j and k against Vero monkey kidney cells revealed that these compounds are far more toxic to M. tuberculosis than to Vero cells. Structure–activity relationships demonstrated that 3-benzylidene-4-chromanones are more potent against M. tuberculosis than the related 2-benzylidene cycloalkanones and the meta substituted chromanone derivatives are more active than their ortho- and para-counterparts. Some guidelines for amplifying the project are presented.

• 4-chromanones
• anti-mycobacterial
• anti-tubercular
• structure–activity relationships
• tuberculosis
• unsaturated ketones

Introduction

A major health problem which extends worldwide is the incidence of tuberculosis particularly in locations with a high prevalence of AIDS1. Currently, one-third of the population of the world is latently infected2 and a little over 9 million new cases are reported each year3. At the moment, there are several major difficulties with the treatment of tuberculosis using drugs. First, strains of Mycobacterium tuberculosis have arisen which are resistant to a number of drugs (MDR-TB) and more recently extremely drug resistant tubercule bacilli (XDR-TB) have been identified. Second, drug therapy is generally for at least 6 months4 and often the side effects are troublesome5 causing distress and on occasions affecting patient compliance. Hence, the creation of novel anti-mycobacterials which are structurally divergent from contemporary medication used against tuberculosis is of paramount importance. In addition, these compounds should be well tolerated in mammals.

The principal interest in our laboratories is the synthesis and bioevaluations of novel conjugated styryl ketones which are designed as thiol alkylators6. Various studies revealed that these compounds react exclusively or preferentially with thiols in contrast to amino or hydroxy groups7^,8 which are found in nucleic acids. Hence, conjugated enones are unlikely to cause genotoxic effects. Thus, the compounds described in the study may demonstrate anti-tubercular properties by reacting with critical thiol groups in the Mycobacterium bacillus. In so doing, they will be exerting their beneficial activity in ways which are different from current drugs such as ethambutol which inhibits cell wall arabinogalactan synthesis, rifampin which interferes with RNA synthesis and streptomycin which inhibits protein synthesis9. Thus conjugated enones acting by a different general process, i.e. thiol alkylation, may be successful in treating tubercule bacilli which have developed resistance to contemporary anti-mycobacterial drugs.

A long-term goal of ours is to find novel compounds which possess both anti-tubercular and anti-neoplastic properties. There are two principal reasons for this focus. First, patients suffering from tuberculosis may develop cancer10 while those suffering from tumors may develop tuberculosis11. Hence, the discovery of single agents to treat both pathological conditions will avoid the potential for drug interactions which can occur with polypharmacy. Second, a question of some significance is whether the relative anti-tubercular and anti-neoplastic potencies are the same. If this proves to be the case, then promising anti-mycobacterials should be evaluated for cytotoxic properties and vice versa. If such relationships can be established in some series of compounds but not others, then the identification of pharmacophoric groups having dual properties may emerge.

This study was undertaken initially with the 3-benzylidene-2,3-dihydro-1-benzopyran-4-ones 1a–l (subsequently referred to as 3-benzylidene-4-chromanones) for the following reasons. First, a number of chalcones, i.e. 1,3-diaryl-2-propen-1-ones, have anti-tubercular properties12 which being flexible acyclic molecules can adopt a range of conformations only some of which may contribute to anti-tubercular properties. The 3-benzylidene-4-chromanones 1 may be regarded as rigid analogs of chalcones. Second, an ether oxygen atom was placed into the heterocyclic ring with a view to serving two purposes, namely, reducing the lipophilicity of the molecules and also having the capacity to assist the alignment of these compounds at a binding site by hydrogen bonding. Third, a previous study revealed that 1a–l displayed excellent cytotoxic properties towards several tumour cell lines and are well tolerated in mice13. As a part of our ongoing effort to develop novel anti-mycobacterial agents, the present investigation is focused on the evaluation of the anti-tubercular potential of 3-benzylidene-4-chromanones 1 and also to develop structure–activity relationships that would afford guidelines for expansion in this area of study.

Materials and methods

Synthesis of compounds

The compounds in series 1 were prepared by a literature procedure13 reported previously from our laboratory and the purity of the compounds was established by ¹H-NMR spectroscopy, melting points and elemental analyses while the synthesis of the analogs in series 2–5 has been described previously14.

Biological assay

The compounds were evaluated against Mycobacterium tuberculosis H₃₇Rv (ATCC 27294) and Vero cells by the Tuberculosis Antimicrobial Acquisition and Coordinating Facility program. For anti-tubercular activity screening, a literature procedure was employed15. In brief, the compounds and the microorganism were incubated in BACTEC 12B medium and the growth inhibition was determined by a broth dilution assay. The IC₅₀ and IC₉₀ figures were generated using the microplate alamar blue assay15.

The effect of 1d, g, j and k on Vero cells was demonstrated by the MTT assay using the protocol of a Promega Celltiter 96 non-radioactive cell proliferation assay kit.

Statistical analysis

Pearson and Spearman correlation coefficients were obtained using statistical program Stata, version 11.2 (64 bit edition), 2012 from StataCorp LP, College Station, TX (http://www.stata.com).

Results and discussion

The compounds in series 1–5 were prepared by condensation between various aryl aldehydes and the appropriate ketones13^,14. These molecules were shown by ¹H-NMR spectroscopy and in several cases by X-ray crystallography to adopt the E stereochemistry13^,14^,16. The choice of aryl substituents was made on the basis of their differing electronic, hydrophobic and steric properties. Thus, the aryl groups display either positive (+) or negative (−) Hammett σ and Hansch π constants, respectively, viz 2-chloro, 4-chloro, 2-bromo, 3-bromo and 4-bromo (+,+), 2-methoxy, 4-methoxy and 2,4-dimethoxy (−,−), 3-methoxy (+,−) and 3-methyl, 4-methyl and 4-dimethylamino (−,+)17^,18. The molar refractivity (MR) values of the aryl substituents in series 1 range from 5.65 (in 1j, k) to 15.74 (in 1i)17. In the case of the analogs in series 2–5, the aryl groups also have a broad range of electronic, hydrophobic and steric properties.

All of the compounds in series 1 were evaluated against Mycobacterium tuberculosis H₃₇Rv and these results are portrayed in Table 1. The biodata generated on series 1 reveals that at a concentration of 10 µg/mL, 83% of the compounds inhibited the growth of M. tuberculosis by varying degrees. The following observations between the magnitude of the anti-tubercular potencies and the substituents in the arylidene aryl rings were noted. First, the average percentage inhibition of 1a–e, g whose aryl substituents have positive Hammett σ values is 69 whereas in the case of 1f, h–l, which bear aryl groups having negative σ values, the figure is 51. This observation suggests that in future investigations, electron-attracting substituents should be placed in the arylidene aryl rings. On the other hand, the average percentage inhibition of the compounds with positive π values in the aryl rings (1a–e, j–l) and negative π constants (1f–i) are 62 and 56, respectively, suggesting that lipophilicity of the molecules has little effect on potency. Second, 1i and 1l which demonstrated no anti-mycobacterial properties at the concentration employed have MR values of 15.55 and 15.74, respectively17, which are substantially greater than the MR values of the aryl substituents of the other compounds in series 1 which are in the range of 5.65–8.8819. These data suggest that the importance of placing small groups into the aryl rings with a view to preparing potent anti-mycobacterials. Third, the position of the substituents in the aryl rings has a significant effect on potency. Thus the average percentage inhibition of the compounds with a single ortho (1a,c,f), meta (1d,g,j) and para (1b,e,h,k,l) substituents are 70, 96 and 44, respectively, revealing a meta > ortho > para relationship. Hence, one may conclude that analog development with a view to finding compounds with increased potencies towards Mycobacterium tuberculosis should have electron-attracting, small substituents in the meta position of the arylidene aryl ring.

Table 1. Evaluation of 1a–l against Mycobacterium tuberculosis H₃₇Rv.

Compound	Substituent in ring A	σ/σ* Value^a	π Values^b	% Inhibition^c
1a	2-Cl	0.37	0.71	71
1b	4-Cl	0.23	0.71	45
1c	2-Br	0.38	0.86	73
1d	3-Br	0.39	0.86	99
1e	4-Br	0.23	0.86	27
1f	2-OCH3	−0.22	−0.02	65
1g	3-OCH3	0.12	−0.02	98
1h	4-OCH3	−0.27	−0.02	59
1i	2,4-(OCH3)2	−0.49	−0.04	0
1j	3-CH3	−0.07	0.56	90
1k	4-CH3	−0.17	0.56	91
1l	4-N(CH3)2	−0.83	0.18	0

^aThe σ values are taken from reference 17 and the σ* constants from the literature18.

^bThe π constants are taken from the literature17.

^cThe figures indicate the percentage inhibition of Mycobacterium tuberculosis H₃₇Rv in the presence of 10 µg/mL of each compound. A reference drug rifampin inhibits the growth of this microorganism by 95% using a concentration of 0.25 µg/mL24.

In the primary bioassay, a lead compound was designated as one which inhibits the growth of M. tuberculosis by 90% or more at a concentration of 10 µg/ml. This criterion was met by 1d,g,j,k and further studies were conducted on these molecules. The IC₅₀ and IC₉₀ figures of these four promising compounds were determined and the data are presented in Table 2. The IC₅₀ values indicate the high potencies of 1d,g,j which contain a single substituent in the meta position of ring A. Movement of the 3-methyl group present in 1j to the para location of ring A yielded 1k whereby potency was reduced threefold which is consistent with the meta > para relationship noted earlier. The IC₉₀ values were obtained which are identical or close to minimal inhibitory concentrations. One may note that 1g rivals rifampin in anti-tubercular potency.

Table 2. Evaluation of 1d, g, j and k against Mycobacterium tuberculosis H₃₇Rv and Vero cells.

		H37Rv		Vero cells
Compound	Substituent in ring A	IC50^a (µg/mL)	IC90^b (µg/mL)	EC50^c (µg/mL)	SI^d
1d	3-Br	0.56	0.77	19.8	25.7
1g	3-OCH3	<0.20	<0.20	9.92	>49.6
1j	3-CH3	0.46	0.72	30.8	42.8
1k	4-CH3	1.43	1.60	>100	>62.5

^aThe IC₅₀ figures are the concentrations of the compounds required to inhibit the growth of Mycobacterium tuberculosis H₃₇Rv by 50%.

^bThe IC₉₀ values are the concentrations of the compounds required to inhibit the growth of Mycobacterium tuberculosis H₃₇Rv by 90%. The minimal inhibitory concentration of rifampin is 0.2 µg/mL15.

^cThe EC₅₀ data are the concentrations of the compounds required to inhibit the growth of Vero cells by 50%. The EC₅₀ figure for rifampin is >100 µg/mL25.

^dThe letters SI refer to the selective index viz the quotients of the EC₅₀ and IC₉₀ values.

To discern whether greater toxicity is demonstrated towards this pathogenic microorganism than a representative mammalian normal cell line, the cytotoxicity of 1d,g,j,k was undertaken using Vero cells (monkey kidney epithelial cells) and the EC₅₀ values in this bioassay are given in Table 2. The data in Table 2 show the huge differential in toxicity towards H₃₇Rv M. tuberculosis compared to Vero cells as indicated by the high selective index (SI) figures. The selective index figures are obtained by dividing the EC₅₀ and IC₉₀ values. Compounds with SI values of greater than 10 are considered important lead molecules. In addition, administration of doses up to and including 300 mg/kg of 1d,g,j,k to mice did not cause deaths13 which further support the development of these molecules as a novel class of anti-tubercular agents.

In summary, therefore, one-third of this series of 3-benzylidene-4-chromanones 1 are lead molecules, and to the best of our knowledge, this is the first report of the anti-tubercular properties of these molecules. A recent study revealed that the percentage of compounds screened by the Tuberculosis Antimicrobial Acquisition and Coordinating Facility Program that meet the criteria (i) in the primary screen of 90% inhibition, (ii) have a MIC value of <1 µg/mL and (iii) possess an SI figure of >10 are 9.9, 9.1 and 0.95, respectively3. Hence, 1d,g,j serve as templates for rapid expansion in the future in the quest for novel anti-tubercular drugs.

A question was addressed whether there is a correlation between the anti-tubercular properties of series 1 and cytotoxic potencies. The chromanones 1a–l have been evaluated against human HSC-2, HSC-3 and HSC-4 squamous cell carcinomas and human HL-60 promyelocytic leukemic cells13. Hence, the percentage inhibition figures in Table 1 were plotted against the CC₅₀ values of these four neoplastic cell lines as well as the average CC₅₀ figure generated from the four cytotoxicity assays. The negative Pearson’s correlation coefficients obtained using HSC-2 (p = 0.01), HSC-3 (p = 0.01), HSC-4 (p = 0.002), HL-60 (p = 0.19) and the average CC₅₀ values (p = 0.006) indicate that, in general, the relative potencies of series 1 towards M. tuberculosis and various neoplasms are the same. The determination of Pearson coefficients takes into consideration both the relative rankings and the magnitude of the potencies. If one considers only the rankings, then the negative Spearman’s correlation coefficients obtained shows an uniform statistically significant correlation of the anti-tubercular data with HSC-2 (p = 0.01), HSC-3 (p = 0.0001), HSC-4 (p = 0.000), HL-60 (p = 0.02) cells as well as the average CC₅₀ figures against all four cell lines (p = 0.006). These observations are an encouragement in the quest to find dual agents capable of simultaneous attack on the tubercule bacillus and various tumors.

In an attempt to find those compounds which have the best combination of anti-tubercular and cytotoxic properties, the quotient of the percentage inhibition of M. tuberculosis and the reciprocal of the average CC₅₀ values were obtained. These figures (given in parentheses) are as follows namely 1a (3.59), 1b (2.73), 1c (4.93), 1d (18.9), 1e (0.61), 1f (5.20), 1g (9.8), 1h (4.37), 1i (0.00), 1j (12.0), 1k (9.1) and 1l (0.00). The four compounds with the highest figures are 1d, g, j, k which afford further reasons for development of these molecules. Thus, in the future, analogs which possess substituents in both meta locations should be prepared and evaluated, e.g. the 3,5-dimethoxy and 3,5-dimethyl compounds as well as the 3-fluoro, 3-chloro and 3-iodo isosteres of 1d.

A further aspect of this investigation was aimed at evaluating the contribution to anti-tubercular potencies of the scaffold to which the arylidene group is attached. Consequently related analogs of 1d,g,j,k in series 1 were examined for anti-tubercular properties. The heterocyclic ether oxygen atom in the chromanones 1 was replaced by an isosteric methylene group (series 2), excised completely (series 3) and substituted by a dimethylene chain (series 4). In addition, series 5 was also prepared with a view to finding if the aryl ring B in series 4 contributed to anti-tubercular properties. The general formulae of the compounds in series 2–5 are presented in Figure 1 while the substituents in aryl ring A are indicated in Table 3. The results portrayed in Table 3 reveal the following structure–activity relationships. (1) Replacement of the ether oxygen atom of the chromanone scaffold by methylene and dimethylene chains led to series 2 and 4 which have significantly reduced anti-tubercular activity or the activity was completely abolished. The ether oxygen atom in series 1 therefore contributes significantly to anti-tubercular properties. It is conceivable that the oxygen atom in series 1 forms hydrogen bonds with the target biomacromolecule(s) which would not be possible with series 2 and 4. (2) The excision of ether oxygen atom in series 1 led to series 3 that show significantly lower anti-tubercular properties and the compounds in series 3 display greater growth inhibition of Mycobacterium as compared to series 2 and 4. A possible reason for this observation is that the five-membered cycloaliphatic ring permits the alignment of series 3 at a receptor while the six- and seven-membered rings in series 2 and 4 are too large to allow facile interactions at the same locations on various cellular constituents. The probability therefore is that the compounds in series 1 and 3 which have good anti-tubercular properties interact at different binding sites. (3) A comparison of the biodata generated in series 4 and 5 suggest that aryl ring B in series 4 is necessary to exhibit anti-tubercular properties.

Figure 1. The general structure of the compounds in series 1–5. The nature of the R¹ and R² groups are presented in Tables 1 and 3.

Table 3. A comparison of the growth inhibition of most potent compounds in series 1 with the related analogs in series 2–5 against Mycobacterium tuberculosis H₃₇Rv.

	% Inhibition in different series^a
Substituents in ring A	1	2	3	4	5
3-Br	99	56	51	63	0
3-OCH3	98	0	79	8	0
3-CH3	90	40	78	19	0
4-CH3	91	0	75	0	3

^aThe data for series 1 are taken from Table 1. In the case of series 2–5, the figures indicate the percentage inhibition of Mycobacterium tuberculosis H₃₇Rv by using 12.5 µg/mL of each compound. A reference drug rifampin inhibits the growth of this microorganism by 95% using a concentration of 0.25 µg/mL24.

Finally, an assessment of the drug-likeness properties of the promising leads 1d,g,j,k was carried out using the web based Molinspiration chemoinformatics tool (www.molinspiration.com) to predict their oral bioavailability. All four compounds satisfy the rule of 520 and also two other important topological parameters, namely, total polar surface area (TPSA)21 and the number of rotatable bonds22. The results suggest that the chromanones 1 possess suitable pharmacokinetic properties and warrant further development as anti-tubercular drug candidates.

From these generalizations, some directions can be made regarding expanding the series of compounds. The general structures of some of the different groups of compounds which should be developed in the future are indicated in Figure 2. First, in general, the compounds in series 1 display the most promising anti-tubercular properties. Thus, in the future, the syntheses of the bioisosteric imino (NH) and sulfur analogs of 1, namely, series 6 and 7 should be undertaken. In addition, series 6 would allow for further analog development, e.g. alkylation and acylation of the basic centre with groups which could assist in making the compound more drug-like such as ensuring the logP values are in the range of 0–3 which is optimal for oral absorption23. The sulfur analogs 7 can be oxidised to the corresponding sulfoxides and sulfones which being electron-attracting groups may enhance the rate of reaction of these compounds with cellular thiols. Second, a hybrid scaffold which incorporates the structural features of series 1 and 3, which have the highest anti-tubercular properties among the analogs 1–5 is benzofuran-3(2H)-one which may be converted to series 8. This heterocycle retains an oxygen atom adjacent to ring B (as in series 1) and a five-membered ring is attached to aryl ring B (which is the case in series 3). Third, analog development should place special emphasis on the placement of a variety of substituents, in particular methyl, methoxy and halogens in one or both of the meta locations in aryl ring A.

Figure 2. The general structures of the compounds proposed for analog development.

Conclusions

A number of cytotoxic 3-benzylidene-4-chromanones 1 display promising anti-tubercular properties. In particular, 1d, g, j, k have low IC₅₀ and IC₉₀ figures towards M. tuberculosis and display significantly greater toxicity to this microorganism than to Vero cells. An assessment of some of the electronic, hydrophobic and steric properties of the aryl substituents in series 1 enabled proposals to be made for the expansion of this series. In addition, an examination of the scaffold in series 1–5 to which the benzylidene group is attached led to various suggestions to be made regarding analog developments.

Declaration of interest

The authors report no conflict of interest and are responsible for the contents and writing of the paper.

Acknowledgements

The authors thank the Canadian Institutes of Health Research for an operating grant to J. R. Dimmock and the Faculty of Medicine Research Fund PTE AOK-KA-2013/20 (University of Pécs) to P. Perjesi. The anti-tubercular results were provided by the Tuberculosis Antimicrobial Acquisition and Coordinating Facility (TAACF) through a research and development contract with the U.S. National Institute of Allergy and Infectious Diseases. Assistance with the literature surveys was given by Vicky Duncan which is recorded with appreciation.