Whole-Cell Screening-Based Identification of Inhibitors against the Intraphagosomal Survival of Mycobacterium tuberculosis

Abstract

We describe here an efficient strategy that employs whole-cell-based screening and further short listing of the compounds by cytotoxicity- and fluorescence-based intracellular assays, resulting in potential bactericidal hits against the growth of Mycobacterium tuberculosis in broth culture as well as in phagosomes. These compounds also inhibited multidrug-resistant strains of M. tuberculosis but showed no or poor inhibition of nonpathogenic mycobacteria or other bacterial species such as Escherichia coli.

TEXT

The existing tuberculosis (TB) drugs isoniazid, rifampin, ethambutol, and pyrazinamide can cure TB, but the chemotherapy requires 6 to 9 months of continuous medication, which oftentimes becomes a hurdle to success (1, 2). The failure to complete the chemotherapy schedule has historically been the main reason for the development of resistance against TB drugs, leading to a continuous rise in the number of multidrug-resistant (MDR) TB patients (1–4). TB drug discovery programs during the past half century have met with very limited success, and bedaquiline, the new TB drug that was approved by FDA recently, represents the sole new TB drug developed after the introduction of rifampin in 1967 (5–7). In this study, by using a whole-cell-based approach employing (i) a resazurin reduction method for the assessment of cellular viability and (ii) a fluorescence-based method to assess the intraphagosomal survival of the pathogen, we report an efficient strategy to identify the compounds that potently inhibit the growth of M. tuberculosis in broth culture as well as intraphagosomal survival of the pathogen (Fig. 1).

Fig 1.

Screening strategy for the identification of hits against M. tuberculosis. Shown is a schematic representation of various screens employed in this study, with the number of compounds tested at every step.

The antimycobacterial activity was determined by using a resazurin reduction microtiter assay (REMA) (8, 9). We screened 2,538 compounds belonging to three compound libraries, namely, Diversity Set III, Mechanistic Set, and Natural Products Set II, which were procured from the National Cancer Institute Developmental Therapeutic Program (NCI-DTP) (http://dtp.nci.nih.gov/repositories.html). A preliminary M. tuberculosis single-dose susceptibility assay was carried out at 25 μM in 96-well plate format; the reaction volume was 200 μl. An exponential-phase culture of M. tuberculosis diluted to a final A₆₀₀ of 0.02 in MB7H9 supplemented with 0.2% Tween 80 and 10% albumin-dextrose-catalase (ADC) was mixed with each compound to a final concentration of 25 μM. The plates were then incubated at 37°C for 7 days, followed by the addition of 30 μl of 0.01% resazurin solution. The plates were further incubated at 37°C for 24 h, and the percent bacterial viability was calculated by measuring the fluorescence (excitation wavelength, 530 nm; emission wavelength, 590 nm) of the reduced resazurin, represented by a change in its color from blue to pink. Based on 90% inhibition of M. tuberculosis growth, 171 hits were identified (see Table S1 in the supplemental material). The screening of the libraries was also carried out similarly in the fast-growing, nonpathogenic Mycobacterium smegmatis, except that the incubation time for the growth of cells with the compounds was 3 days. It was observed that only 27 compounds showed inhibition of M. smegmatis growth as against 171 hits for M. tuberculosis. A comparative analysis showed <10% overlap in the compounds (13 out of 171) inhibiting both the slow-growing, pathogenic M. tuberculosis and the fast-growing, saprophytic M. smegmatis. Hence, these two organisms may have differences in the presence of target proteins or their sequences. Thus, our findings, as observed in a few studies earlier (10, 11), suggest that caution should be exercised in making the choice of the organism for screening the compounds to identify M. tuberculosis inhibitors, although it has not escaped our attention that the only new drug approved against TB, i.e., TMC207, has been identified by a strategy that employed M. smegmatis for screening (12).

The preliminary 171 hits were cherry-picked and subjected to a dose-response assay (0.78 μM to 25 μM) for MIC₉₀ determination by using REMA. For ease of comparison with the known TB drugs, the MIC₉₀ values for 171 hits were converted to μg/ml units. Forty compounds exhibiting the highest efficacy toward M. tuberculosis (with MIC₉₀ values of 10 μg/ml or less) were selected for further work, and based on availability, 26 of these were reordered in larger quantities (see Table S2 in the supplemental material).

The selected 26 compounds were numbered according to their ascending National Service Center (NSC) numbers (N1 to N26) and were retested in a dose-response assay (0.0156 μg/ml to 8 μg/ml) for the determination of MIC₉₉ against M. tuberculosis H37Rv and M. tuberculosis Erdman. All the 26 compounds exhibited significant inhibition, with 15 compounds exhibiting an MIC₉₉ of 1 μg/ml or less for both strains (Table 1). In fact, 9 of these compounds exhibited an MIC₉₉ of 0.5 μg/ml or less for both the strains, indicating their high inhibitory efficacy for these two pathogens (Table 1). These MIC₉₉ values were further confirmed by spotting 5 μl of the culture from each well on MB7H11 agar plates, followed by incubation at 37°C for 3 to 4 weeks. A perfect correlation was observed in the results obtained from REMA and the growth on MB7H11 plates; no growth was observed at the MIC₉₉s (Fig. 2).

Table 1.

MIC₉₉ values of the 26 short-listed compounds for M. tuberculosis H37Rv and M. tuberculosis Erdman

Compound	MIC99 (μg/ml) for M. tuberculosis
	H37Rv	Erdman
N1	0.5	0.25
N2	0.25	0.5
N3	4	4
N4	2	2
N5	4	2
N6	4	2
N7	2	2
N8	4	4
N9	4	8
N10	0.0625	0.0625
N11	1	1
N12	0.0156	0.0156
N13	2	2
N14	2	2
N15	0.5	1
N16	1	0.25
N17	0.5	0.25
N18	0.5	0.5
N19	0.5	0.5
N20	0.0312	0.0312
N21	2	0.5
N22	1	1
N23	1	0.5
N24	0.25	0.125
N25	2	0.5
N26	1	0.25

Fig 2.

Determination of MIC₉₉ values of the selected 26 compounds obtained from preliminary screening. MIC₉₉ values were determined by employing REMA (8, 9). M. tuberculosis H37Rv and M. tuberculosis Erdman cells were incubated for 7 days in the presence of various concentrations (0.0156 μg/ml to 8 μg/ml) of each compound. The MIC₉₉ was considered the lowest concentration that inhibited the growth of the cells by ≥99%. Shown is the growth of M. tuberculosis H37Rv from the 5-μl sample from each well on MB7H11 agar plates. The lowest concentration of a compound at which there is no growth coincides with the MIC₉₉ of the compound. Rif, rifampin; Iso, isoniazid.

Next, we analyzed the cytotoxicity of the selected 26 compounds against mammalian cell lines THP-1, RAW 264.7, HepG2, HeLa, and CHO to show that the compounds were not generally cytotoxic and determined each compound's selectivity index (SI), which defines the therapeutic value of a candidate compound. Cells were seeded in 96-well plates (1 × 10⁴ cells/well) in the presence of various concentrations of the compounds (1 μg/ml to 100 μg/ml) in an assay volume of 200 μl. After 48 h, resazurin dye was added, followed by the measurement of fluorescence of reduced resazurin 24 h later to determine the 50% inhibitory concentrations (IC₅₀s) and SI values (IC₅₀ for cell line/MIC₉₉ against M. tuberculosis growth) for each of the compounds (see Tables S3 and S4 in the supplemental material). Out of the 26 selected compounds, 15 compounds that displayed SI values of >10 in at least three of the cell lines tested were selected as hits for further work (N1, N2, N5, N6, N7, N10, N12, N13, N14, N16, N17, N18, N20, N23, and N24).

For evaluating the influence of the selected hits on the intracellular growth of M. tuberculosis in THP-1 macrophages, we employed the strategy developed by Christophe et al. (13) with certain modifications. For this, a highly fluorescent strain of M. tuberculosis, H37Rv, was employed; we generated this strain by using a pSD5 derivatized plasmid containing a modified high-strength synthetic promoter to drive constitutive episomal expression of enhanced green fluorescent protein (EGFP) (red-shifted variant of wild-type green fluorescent protein [GFP]). For this, the −35 region sequence (TTGTAC) of the A37 promoter was changed to the mycobacterial consensus −35 sequence (TTGCGA) and a single base pair was inserted between the −10 and −35 regions in order to provide an optimal 18-bp distance between −10 and −35 regions of the promoter. The resulting promoter, A37⁺, exhibited ∼50% enhanced activity in comparison to its parental counterpart (14). THP-1 cells were activated in RPMI media containing 10% fetal bovine serum (FBS) and an antibiotic-antimycotic mix by the addition of 30 nM phorbol myristic acetate (PMA) and incubated at 37°C for 16 h in the presence of 5% CO₂. Media containing PMA were discarded, and the cells were washed once with RPMI medium to remove PMA and antibiotics. The adhered cells were then scraped off and harvested at 1,000 rpm for 10 min, and the resulting cell pellet was resuspended in a small volume of RPMI medium containing 10% fetal bovine serum. Cells were counted by trypan blue exclusion staining. Exponentially growing M. tuberculosis H37Rv-GFP cells were harvested at 4,500 rpm for 10 min and washed twice with MB7H9 medium. The resulting bacterial cells were resuspended in 10 ml of RPMI medium containing 10% FBS to which ∼6 g of 0.5-mm glass beads was added, followed by vortexing for 15 min. The suspension was judiciously centrifuged in order to remove any remaining bacterial clumps. The CFU of the resulting single-cell homogeneous suspension of bacteria were estimated by measuring the absorbance of this suspension at 600 nm (A₆₀₀ of 0.5 corresponds to 3 × 10⁷ M. tuberculosis CFU). Activated THP-1 cells were infected with M. tuberculosis H37Rv-GFP in suspension at a multiplicity of infection (MOI) of 5:1 (bacteria/macrophage) in RPMI medium containing 10% FBS at 37°C for 2 h with a constant shaking at 100 rpm. A control set of uninfected THP-1 cells was set up in a similar manner. The cells were then harvested by centrifugation to ensure the removal of any nonphagocytosed extracellular bacteria and washed once with RPMI medium. The remaining extracellular bacilli, if any, were killed by the addition of 200 μg/ml amikacin at 37°C for 1 h with constant shaking at 100 rpm. The final two rounds of washing with RPMI medium were carried out to get rid of the remaining extracellular bacteria, if any. One hundred thousand infected cells were then seeded in each of the 96 wells in the presence of various concentrations of the compounds (0.0625 μg/ml to 100 μg/ml) in a final volume of 250 μl. Along with this, the uninfected THP-1 cells were also seeded at 10⁵ cells/well in the presence of the same concentrations of the compounds to serve as a control for the toxicity of the compounds to the THP-1 cells. The plates were then kept at 37°C in the presence of 5% CO₂. After 5 days of infection, the plates containing the infected cells were photographed by using a fluorescence microscope. The inhibition of the growth of intracellular bacteria was indicated by the absence of fluorescence. For further confirmation of growth inhibition, the medium was removed from the wells, followed by lysis of the THP-1 cells by the addition of 50 μl sterile water. After ∼2 h of lysis, the cells were vigorously pipetted and the entire volume was spotted on MB7H11 oleic acid-albumin-dextrose-catalase (OADC) agar plates followed by incubation at 37°C for 2 to 3 weeks to analyze the inhibition of bacterial growth. The results of the fluorescence assay were confirmed by our observation that the samples from the 96-well plates that showed no fluorescence exhibited no growth on the MB7H11 OADC agar plates. Additionally, resazurin solution (30 μl of a 0.01% solution in each well) was added to all the wells of the control plates containing the uninfected cells; these cells were then incubated at 37°C for 24 h, following which their fluorescence resulting from a change in the color from blue to pink was monitored. The MIC₉₉ was calculated as the minimum concentration of the compound at which no fluorescence as well as growth on MB7H11 was observed, along with no toxicity to the uninfected THP-1 cells. The growth of the bacteria, as measured by increase in the fluorescence, followed a typical growth pattern for 7 to 8 days, after which the lysis of macrophages began due to cytopathogenic effects of M. tuberculosis (Fig. 3; see Fig. S1 in the supplemental material). Before screening the hits for the inhibition of intracellular mycobacterial growth, we employed standard TB drugs isoniazid and rifampin to validate the assay. As expected, we observed a dose-dependent decrease in the fluorescent bacteria. Following this, all the hits were evaluated for their ability to inhibit the intracellular survival of the pathogen and were employed in a dose-dependent assay (0.0625 μg/ml to 100 μg/ml) for MIC₉₉ determination. Table 2 displays the list of the intracellular MIC₉₉ values for the selected hits wherein five compounds (N1, N12, N13, N17, and N18) were identified as inhibitors of the growth of intracellular M. tuberculosis (Table 2; see Fig. S2 and S3 in the supplemental material). Compounds N1 (5-nitro-1,10-phenanthroline) and N17 (quinolinium dichloride), which exhibited potent inhibition of M. tuberculosis growth in broth culture, with an MIC₉₉ of 0.5 μg/ml, also inhibited the growth of intracellular M. tuberculosis, with MIC₉₉ values of 0.5 μg/ml and 20 μg/ml, respectively. Compound N12, which exhibited the most potent activity against the in vitro growth of M. tuberculosis, with an MIC₉₉ of 0.0156 μg/ml, displayed an MIC₉₉ of 5 μg/ml against the intracellular growth of the pathogen. Another compound, N18, with an in vitro MIC₉₉ of 0.5 μg/ml, exhibited an MIC₉₉ of 10 μg/ml against the intracellular growth of M. tuberculosis. Compounds N12 and N18 represent the same chemical scaffold as rifamycin; however, compounds N1 and N17 represent classes of chemical series hitherto not used for the development of antitubercular drugs. Compound N13 exhibited an MIC₉₉ of 2 μg/ml against M. tuberculosis growth in broth culture; however, it displayed a very poor intracellular MIC₉₉ of 80 μg/ml.

Fig 3.

Monitoring of the growth of fluorescent M. tuberculosis H37Rv bacilli inside the macrophages. Shown are fluorescence microscopy pictures of the same field of THP-1 macrophages infected with a GFP-expressing strain of M. tuberculosis at an MOI of 5:1, in which there is a clear increase in the intraphagosomal survival of the bacilli. The middle panels show bright-field and fluorescence overlaid images of the areas circled in the top and bottom panel fluorescence images for visualization of the presence of fluorescent M. tuberculosis inside macrophages.

Table 2.

Compounds which displayed both in vitro and intracellular inhibition of M. tuberculosis

The four most effective compounds, namely, N1, N12, N17, and N18, were also subjected to determination of MIC₉₉ against cross-species mycobacteria (M. smegmatis mc²155, M. vaccae, M. phlei, and M. fortuitum) and other bacterial species such as Escherichia coli strains BL21(λDE3), XL1-Blue, and DH5α by employing a resazurin reduction assay (Table 3). It was observed that all four compounds exhibited either no inhibition or much higher MIC₉₉ values against the cross-species mycobacterial strains and E. coli strains in comparison to the inhibition exerted on M. tuberculosis, as shown in Table 3. Moreover, in view of the divergence exhibited by these four compounds in their MIC₉₉s against M. tuberculosis and other cross-species mycobacteria and E. coli, we assessed all 26 selected compounds for their efficacy against M. smegmatis and E. coli strains (see Table S5 in the supplemental material). The compounds either showed very poor or no inhibition at up to 8 μg/ml against M. smegmatis and E. coli strains in comparison to the MIC₉₉ values for M. tuberculosis (see Table S5), further confirming their divergence in MIC₉₉s against various bacterial species. The poor activity (no inhibition till 8 μg/ml) of N12 and N18 (which belong to the family of rifamycins) against E. coli strains, is consistent with the earlier findings that rifampin exhibits poor MIC₉₉s of ∼8 μg/ml against E. coli strains (15, 16). Next, we assessed the efficacy of these four compounds against the growth of drug-resistant strains of M. tuberculosis. It was observed that all the compounds showed potent inhibition of the drug-resistant strains of M. tuberculosis and displayed, in general, efficacy against the resistant strains that was equal to or better than that against M. tuberculosis H37Rv and M. tuberculosis Erdman, as shown in Table 3. The susceptibility of the rifampin-resistant strain M. tuberculosis JAL2261 to compounds N12 and N18 (rifamycin analogs) might appear unanticipated; however, the mechanism of rifampin resistance provides a possible explanation for it. The development of resistance to rifampin is mostly due to mutations in a well-defined, 81-base-pair central region of the gene encoding the β-subunit of RNA polymerase (17, 18). The most common mutations alter either codon 526 or codon 531 and result in a high level of resistance to rifampin (17, 18). The precise mutation conferring resistance to rifampin in the M. tuberculosis JAL2261 strain is not known; however, it is known that not all mutations within this 81-bp region produce the same level of resistance (17, 18). For example, alterations in codons 511, 516, 518, and 522 result in a low level of resistance to rifampin and rifapentine but the organisms remain susceptible to two other rifamycins (rifabutin and rifalazil) (17, 18).

Table 3.

MIC₉₉ values of the 4 selected compounds for M. smegmatis, M. vaccae, M. fortuitum, M. phlei, various E. coli strains, and drug-resistant strains of M. tuberculosis

Compound	NSC no.	MIC99(μg/ml) for:
		M. smegmatis	M. vaccae	M. fortuitum	M. phlei	E. coli strainsa	M. tuberculosis JAL2261b	M. tuberculosis 1934c	M. tuberculosis BND320d
N1	4263	8	NIe	NI	NI	NI	0.125	0.125	0.125
N12	133100	4	1	2	4	NI	0.0156	0.0156	0.0156
N17	218439	NI	NI	NI	8	NI	0.125	0.25	0.25
N18	239375	NI	8	NI	NI	NI	0.0625	0.125	0.0625

^aE. coli strains BL21(λDE3), XL1-Blue, and DH5α were employed, and none showed inhibition in the presence of any of the four compounds even at the highest concentration evaluated (8 μg/ml).

^bResistant to isoniazid, rifampin, and ethambutol.

^cResistant to streptomycin, isoniazid, and ethambutol.

^dResistant to isoniazid.

^eNI, no inhibition of the growth was observed even at the highest concentration of the compound tested (8 μg/ml).

In addition, to determine the nature of inhibition by these compounds, we carried out MIC₉₉ determination in two identical sets, and the second set was employed to determine if the inhibition of M. tuberculosis was bactericidal. After 7 days of incubation of M. tuberculosis cells with the compounds, the entire culture (200 μl) was appropriately diluted, plated on MB7H11 OADC agar plates, and incubated at 37°C for 3 to 4 weeks. The bactericidal nature was confirmed by the complete absence of viable cells from the wells, which showed no conversion of the blue color of resazurin. We observed that the inhibition exerted by the compounds N1, N12, N17, and N18 at their MIC₉₉ values was bactericidal.

In conclusion, we employed a high-throughput whole-cell screening assay based on the reduction of resazurin dye, further strengthened by the use of several other screens, including intracellular survival assays. Thus, by using a highly fluorescent strain of M. tuberculosis and a high-throughput 96-well-format assay, we have described an important strategy to identify potent inhibitors against the intraphagosomal growth of M. tuberculosis.