Sodium alginate (ALG, from brown algae, viscosity = 10 cps in 1% H₂O at 25°C, M/G ratio = 3.42), rifampicin (RIF) and isoniazid (INH) were purchased from Sigma (St. Louis, MO, United States). Silver nitrate, D (+)-glucose, and sodium hydroxide were purchased from Merck (Darmstadt, Germany).

Green synthesis of ALG-capped silver nanoparticles (ALG-AgNPs).

ALG-AgNPs were synthesized according to our previous reports (Chang et al., 2017; Young et al., 2018) by the reduction of Ag⁺ ions in an alkaline aqueous medium using glucose as a reducing agent and ALG as a stabilizing and/or reducing agent. Briefly, an aqueous solution of silver nitrate was added dropwise to an aqueous solution of ALG, glucose, and NaOH at room temperature while stirring. After the reaction, the mixture was diluted 50-fold to measure the surface plasmon resonance (SPR), size, and zeta potential. The excess base and reactants remaining in the suspension were dialyzed against water using a Dialysis Cassette (Slide-A-Lyzer®, G2, 2000 MWCO) until the pH of the aqueous solution became neutral. Table 1 summarizes the conditions used to synthesize the ALG-AgNPs, and the particle size and zeta potential of the as-prepared and dialyzed colloidal solution. The purified ALG-AgNPs were then subjected to different spectral measurements, TEM imaging, and in vitro or in vivo tests. The dialyzed aqueous solutions of ALG-AgNPs were stable for more than 6 months at 4°C.

SPR values were measured using UV–vis absorption spectra on a Perkin Elmer Lambda 35 spectrophotometer (PerkinElmer, Waltham, MA, United States), scanning from 300 to 700 nm. Samples were run in disposable 1.5 ml plastic cuvettes. Milli-Q water was used in the reference cell. Concentrated samples (0.02 ml) were diluted 50-fold with 0.98 ml of Milli-Q water before detection.

The particle size was determined by photon correlation spectroscopy (Zetasizer Nano-ZS; Malvern Instruments, United Kingdom). The instrument contains a 4 mW He-Ne laser that operates at a wavelength of 633 nm, and an avalanche photodiode (APD) detector. The scattered light was detected at an angle of 173°. Raw data were subsequently correlated to the mean hydrodynamic size by cumulant analysis (Z-average mean). The zeta potentials (ζ) of all ALG-AgNPs were analyzed using laser Doppler anemometry (Zetasizer Nano-ZS; Malvern Instruments, United Kingdom).

The morphology of ALG-AgNPs following purification as described previously, was examined using a JEOL JEM-1200 EXII transmission electron microscope (JEOL, Tokyo, Japan). A typical method of preparing TEM samples was as follows: one drop of the ALG-AgNP suspension was deposited on a 200-mesh Formvar/carbon-coated copper grid, and excess solution was removed by wicking with filter paper to avoid particle aggregation. The images were examined by TEM at 80 kV.

The crystalline structure of the ALG-AgNPs was examined by X-ray diffraction (XRD), which was carried out on a Bruker D2 Phaser powder diffractometer (Karlsruhe, Germany) operated with a Cu Kα radiation source filtered with a graphite monochromator (0.154 nm). X-rays were generated at 30 kV and 10 mA. The XRD scans were recorded at 2θ from 10 to 90° at a scan rate of 0.05°/0.5 s. X-ray photoelectron spectra (XPS) were collected using a VG-Scientific ESCALAB 250 instrument. Elemental analysis was conducted using a pass energy of 20 and 160 eV for the survey. Al Kα radiation at 1486.6 eV was used and the peak positions were calibrated internally to the C1s peaks at 284.6 eV. A colloidal sample was deposited on a silicon substrate and dried in a vacuum chamber at 110°C for XPS signal measurement. The operating power was 300 W, and the scan rate was 0.1 eV/150 ms. The XRD and XPS analysis were measured by the Precious Instrument Center of Ministry of Science and Technology of Taiwan. Colloid silver samples were lyophilzed to powder before measurement.

The Mtb strains used in this study were H37Ra, H37Rv, W6 (Beijing strain); two multidrug-resistant (MDR) strains, KVGH376 and KVGH264, from Kaohsiung Veterans General Hospital (Kaohsiung, Taiwan); two extensively drug-resistant (XDR) strains, TCHL002, TCHL017, from Taipei City Hospital-Linsen branch (Taipei, Taiwan); and two strains from Changhua Christian Hospital (Changhua, Taiwan), CHCH005 (Beijing strain) and CHCH029 (East African-Indian strain). All strains were routinely grown at 37°C with 5% CO₂ in 7H10 agar (Becton, Dickinson and company (BD), MD, United States) supplied with 10% oleic acid-albumin-dextrose-catalase (OADC) (Creative life sciences, Taipei, Taiwan) enrichment and 0.5% glycerol (Union Chemical works LTD., Hsinchu, Taiwan) or 7H9 broth (BD, MD, United States) supplied with 10% OADC enrichment and 0.5 mg/ml Tween 80 (Sigma, St. Louis, MO, United States).

Minimum inhibitory concentration (MIC) of the ALG-AgNPs was determined using a microplate Alamar blue assay (Franzblau et al., 1998). The Alamar blue assay was performed in sterile 96-well plates. The different Mtb strains in log phase were cultured in 7H9 medium supplemented with 10% OADC and cultured at 37°C with shaking at 220 rpm for 3–4 weeks before use. 1.56 × 10⁵ colony forming units (CFUs) per 100 μL of different Mtb strains were inoculated into 96-well plates containing 100 μL 7H9 broth with different amounts of ALG-AgNPs. Additional controls included wells with Mtb cells without ALG-AgNPs and wells with only ALG-AgNPs and media. After 5 days of incubation at 37°C, 50 μL of 10× Alamar blue (Invitrogen, CA, United States) with 10% Tween 80 at a 1:1 ratio was added to control wells of Mtb cells alone for 24 h to confirm the growth of Mtb cells, then 50 μL of Alamar blue mix solution was added to each well, and the plates were re-incubated for 24 h. Each well either remained blue or turned pink, representing no growth or growth, respectively. Some wells showed a violet color and were recorded as growth. The MIC value was determined as the lowest concentration of ALG-AgNPs which prevented the color change in the well. For drug susceptibility testing, cultured mycobacteria (100 μL) from the Alamar blue assay representing four Mtb strains [H37Rv, W6 (Beijing), KVGH264 (MDR), and TCHL017 (XDR)] were treated with PBS or the indicated concentrations of ALG-AgNPs for 0 and 7 days, then titrated for three dilutions (10×, 100× and 1000×) and performed in three independent experiments. CFU values were determined by using 100 μL bacterial mixture diluted on 7H10 agar plates. Plates were kept at 37°C for 3–4 weeks and the number of colonies on each plate was counted.

The human monocytic cell line THP-1 was purchased from BCRC (Bioresource Collection and Research Center, Hsinchu, Taiwan). THP-1 cells were cultured in RPMI 1640 supplemented with 10% FBS, 2 mM L-glutamine, 100 U/ml penicillin, 100 µg/ml streptomycin, and 0.05 mM 2-mercaptoethanol and incubated at 37°C under 5% CO₂. 10⁵ THP-1 cells/well were seeded into 96-well culture plates in RMPI (10% FBS, L -glutamine, no antibiotic) with phorbol 12-myristate 13-acetate (PMA, Sigma-Aldrich) at a final concentration of 100 nM. Following 3 days of PMA stimulation, THP-1 cells differentiated into plastic-adherent, well-spread cells with macrophage-like morphology. Four strains of Mtb [H37Rv, W6 (Beijing), KVGH264 (MDR), and TCHL017 (XDR)] were used to infect PMA-differentiated THP-1 macrophages at an MOI of 1 for 24 h. Cells were then wash thoroughly with PBS to remove excessive extracellular bacilli. After that, PBS, 1 μg/ml rifampicin or 25, 50, 100 μg/ml ALG-AgNPs were added to the wells and then incubated for five more days. Finally, the cells were lysed in distilled water containing 0.1% SDS for 15 min at room temperature. The cell lysates from each sample were used to determine the number of CFUs on Middlebrook 7H10 agar after incubation for 21 days.

Cultures were taken from a frozen stock of Mtb strain H37Rv. The cells were thawed, grown to log phase (OD₆₀₀ = 0.6–0.8) and then diluted 1:100 into a 15-ml screw-capped conical tube filled with 10 ml of Middlebrook 7H9 medium supplemented with 0.2% glycerol, 10% OADC and 0.025% Tween 80 at 37°C. The head space to volume ratio (HSR) was maintained at 0.5. The tubes were tightly sealed with Parafilm^® (Bemis NA). A visual indication of oxygen depletion was observed by the addition of sterile methylene blue solution (500 mg/ml) to a final concentration of 1.5 mg/ml to 10-ml standing cultures maintained under hypoxic conditions in Middlebrook 7H9 supplemented as above. Control tubes containing Middlebrook 7H9 medium and methylene blue, but no bacteria, were also set up. The decline in absorbance of methylene blue is routinely used for measuring hypoxia to confirm O₂ depletion. In the cultivation of Mtb, methylene blue decolorization starts when the dissolved oxygen concentration falls below 3%. Hence, complete decolorization of methylene blue, indicative of hypoxia, was reached at 12 days. After the cultures had reached hypoxia and grown to OD₆₀₀ = 0.5, approximately 5 × 10⁷ CFU/ml were treated with PBS, rifampicin (1 μg/ml), isoniazid (0.2 μg/ml), and different amounts of ALG-AgNPs (25, 50, and 100 μg/ml) for 5 days. Aliquots from each sample were used to determine the number of CFUs on Middlebrook 7H10 agar after 21 days.

Oxygen-replete and non-replicating persistent (NRP) stage Mtb cultures were harvested. Mtb cultures exposed to hypoxia were removed from the hypoxia chamber and kept on ice with minimal disturbance for 30 min to arrest the cells at their actual metabolic state. This cell suspension was centrifuged in a pre-cooled rotor at 10,000 × g for 10 min at 4°C. Finally, the supernatant was removed, and the pellet was snap-frozen in liquid nitrogen and stored at -80°C. In the case of log phase cells from oxygen replete cultures, bacteria were harvested as per the hypoxic cultures. Total RNA was extracted by homogenizing the cells with glass beads in 1 ml Trizol reagent (Thermo Fisher Scientific, United States) on ice for 10 min, then the bacterial clumps were homogenized using a Minilys^® homogenizer (Bertin technologies, France) with bead-beating for 15 s 6 times at medium speed (4,000 rpm). The homogenates were left on ice for 20 min, then chloroform was added for 2–3 min. The sample was centrifuged for 15 min at 12,000 × g at 4°C. The mixture separates into a lower red phenol-chloroform organic phase, an interphase, and a colorless upper aqueous phase. The aqueous phase containing the RNA was transferred to a new tube containing 0.5 ml isopropanol for 10 min. The tube was centrifuged for 10 min at 12,000 × g at 4°C. Total RNA precipitate forms a white gel-like pellet at the bottom of the tube. The pellet was washed in 1 ml of 75% ethanol and resuspended in 20–50 μL of RNase-free water. RNA concentration and quality were measured using a Nanodrop^® ND-1000 spectrophotometer (Thermo Fisher Scientific, United States). The cDNA was reverse-transcribed using Moloney murine leukemia virus reverse transcriptase and random hexamer oligonucleotides for priming (Life Technologies, CA, United States). Expression of the hspX and esat-6 genes was determined by real-time quantitative PCR analysis using an ABI VII7 system (Applied Biosystems, Foster City, CA, United States). Transcript levels between various RNA samples were normalized using 16S rRNA. Gene expression was quantified using the ΔΔCt method. Primers used for RT-PCR were as follows: hspX Fwd: 5′-GGA​AGA​CGA​GAT​GAA​AGA​GG-3′ and Rev: 5′-AAC​CGC​CAC​CGA​CAC​AGT​AAG-3′; esat-6 Fwd: 5′-CCA​TTC​ATT​CCC​TCC​TTG​ACG-3′ and Rev: 5′-TGC​GAA​CAT​CCC​AGT​GAC​G-3′; and 16S rRNA Fwd: 5′-TTG​ACG​GTA​GGT​GGA​GAA​GAA​GC-3′ and Rev: 5′-CCT​TTG​AGT​TTT​AGC​CTT​GCG​G-3′.

Log phase (OD₆₀₀ = 0.8) cells of Mtb strain H37Ra (8 × 10⁷ CFU/ml) were treated with 50 μg/ml ALG-AgNPs in a flask for 24 h, while untreated H37Ra cells were used as a control. The bacteria were harvested by centrifugation at 3,000 × g for 10 min and washed with 0.9% NaCl solution. Then, the bacteria were stained with 8 μg/ml propidium iodide dye (Sigma) in 0.9% NaCl solution for 15 min and kept in the dark. After staining, the bacteria were centrifuged and washed twice with 0.9% NaCl solution. The pellets were resuspended in 0.9% NaCl solution. Bacteria with damaged membranes appeared red when visualized by fluorescence microscopy, whereas intact bacterial cells were not stained.

The mycobacterial shuttle vector pMV261-DsRed was constructed and electroporated into Mycobacterium marinum (M. marinum) cells. Transformants were selected on 7H10 plates supplemented with 50 μg/ml kanamycin. The AB/TL wild-type zebrafish embryos and larvae used for these experiments were obtained from a breeding stock of adult animals kept in our zebrafish core facility (NHRI, Taiwan). Embryos and larvae were maintained in fish water (FW) at 28.5°C from collection and throughout the experiments. The larvae were transferred to FW containing 0.2 mM 1-phenyl-2-thiourea (PTU) at 24 h post fertilization (hpf). The larvae were maintained in the FW with PTU for the duration of the experiment to prevent pigment formation and maintain optical transparency. The larvae were dechorionated within 26–28 hpf. Larvae were anesthetized by tricaine in FW for 5–10 min to allow infection by microinjection of M. marinum-DsRed via the caudal vein. Prior to microinjection, M. marinum-DsRed cells were grown to log phase, OD₆₀₀ = 0.7–0.8 ( ≑ 7–8 × 10⁷ CFU/ml), then 4.6 nL (322–368 CFU/larvae) were microinjected into the caudal vein per zebrafish larvae. The infected larvae were treated with PBS, 100 μg/ml rifampicin or 200 μg/ml ALG-AgNPs for 5 days (the medium was changed every 2 day). For observation of bacterial infection, five fish per group were inoculated for each experiment and two experiments were performed. At 5 days post infection (dpi), bacterial infection was observed with a fluorescence microscope. Bright field and fluorescence images were recorded and analyzed by ImageJ software. After observation of bacterial infection, five zebrafish larvae were pooled together and lysed in 1 ml PBS containing 0.1% SDS. The homogenates were diluted at 10 × and 100 × in triplicate and plated on 7H10 agar plates. The bacterial burden was analyzed by CFU assay, and two experiments were performed. All procedures involving zebrafish embryos were performed in compliance with NIH guidelines for the use and care of laboratory animals and approved by the Institutional Animal Care and Use Committee of the National Health Research Institutes, Taiwan (#NHRI-IACUC-106123, 10/15/2018).

Female BALB/c mice 6–8 weeks old were purchased from the National Laboratory Animal Center (Taipei, Taiwan). All mice were kept in individually ventilated cages at the Animal Center of the National Health Research Institutes (Miaoli, Taiwan). BALB/c mice were initially administered with 10, 50, 100 mg/kg of ALG-AgNPs by oral gavage or intravenous injection for the sighting study. Before the start of the toxicity tests, mice were fasted for 3–4 h. Following the period of fasting, the mice were weighed, and the fixed doses of ALG-AgNPs administered. During this period, mice were observed daily for at least 14 days without clear signs of toxicity. As there was no evident toxicity or mortality, the mice were further dosed at 500 mg/kg ALG-AgNPs by oral gavage or 250 mg/kg by intravenous injection once daily for 2 weeks. Mice were observed post-administration at 4 and 6 h, and then monitored daily for body weight and survival for 2 weeks. For determining the efficacy of ALG-AgNPs in the TB mouse model, BALB/c mice were infected with 10⁶ CFU of Mtb strain H37Ra via intravenous injection. The treatments started at 2 weeks post-infection (day 14) and the mice were given intravenous injections of PBS (control) or different amounts of ALG-AgNPs (10, 50 mg/kg) and rifampicin (10 mg/kg) 5 days per week for 2 weeks (from day 14 to day 28). After euthanasia, the lungs were collected, homogenized and plated onto 7H10 agar at appropriate dilutions. The efficacy of the ALG-AgNP treatments was determined by the CFU of H37Ra in the lungs. All procedures were performed and approved by the Institutional Animal Care and Use Committee of the National Health Research Institutes, Taiwan (#NHRI-IACUC-108077, 04/01/2019).

Cytotoxicity of ALG-AgNPs in THP-1 cells was performed by using an LDH Assay kit (Dojindo). LDH catalyzes dehydrogenation of lactate to pyruvate, thereby reducing NAD to NADH. NADH reduces a water-soluble tetrazolium salt (WST) in the presence of an electron mediator to produce an orange formazan dye. The amount of formazan dye thus formed is proportional to the amount of LDH released into the medium from damaged cells, which is an indication of cytotoxicity. Briefly, 10³ THP-1 cells per well were seeded into 96-well plates and incubated at 37°C with 5% CO₂. After 24 h, the indicated concentrations of ALG-AgNPs were added and the cells incubated for 48 h. After 48 h, working solution was added to each well and incubated for 30 min, and the reactions were terminated by adding stop solution. The absorbance was then measured at 490 nm by a microplate reader.

Statistical analyses were performed using GraphPad Prism version 7 software (GraphPad Software Inc., San Diego, CA, United States). All values were given as mean ± SEM. The t-test (two-tailed) was used to determine the statistical significance of the difference between two groups. For analyzing multiple groups, one-way ANOVA with a multiple comparison test (Tukey, Bonferroni, Newman-Keuls) or two-way ANOVA with Tukey’s multiple comparison test was used and p values < 0.05 were considered statistically significant.

Our previous report showed that polysaccharides can be used as both a reducing agent and a stabilizing agent to prepare AgNPs (Cheng et al., 2014); however, the reduction reaction progressed very slowly at room temperature. In the present study, we used glucose as a reducing agent, and found that the reaction time could be shortened to 1 h under alkaline conditions. The characteristic SPR peak at around 410 nm confirmed the formation of AgNPs and was also used to estimate the reaction rate of AgNP formation (Chang et al., 2017; Young et al., 2018). The reaction rate, particle size, and zeta potential of the AgNPs were all affected by the concentrations of base, reducing agent, and stabilizer (Supplementary data; Supplementary Figure S1; Supplementary Table S1). Table 1 summarizes the conditions used to synthesize the ALG-AgNPs, as well as the particle size and zeta potential of the as-prepared and dialyzed colloidal solutions. The average particle size and zeta potential used in the following in vitro and in vivo studies were 70 nm and -47 mV, respectively. The concentration of Ag⁰ was 396 μg/ml, as measured by ICP-Mass, which is smaller than the calculated concentration (539.3 ppm) because the conversion yield is not 100% conversion yield. TEM images showed the purified ALG-AgNPs to be well dispersed, well-defined, and round in shape (Figure 1A). As per our previous observation, the particle size measured using TEM (mostly <50 nm) was smaller than the Z-average diameter (∼70 nm) measured by DLS (39, 40, 42). The X-ray diffraction (XRD) spectrum (Figure 1B) ) of lyophilized colloid sample showed four peaks at 2θ values of 38.1, 44.2, 64.8, and 77.8 deg, corresponding to (111) (200), (220), and (311) reflections, which are consistent with the standard JCPDS database for AgNPs (No. 04-0783). Elementary XPS analysis indicated two peaks located at binding energies of 372.8 and 378.8 eV with a spin-orbit separation of 6.0 eV, which correspond to the emission of the 3d photoelectrons of Ag 3d_5/2 and Ag 3d_3/2 (Figure 1C).

ALG-AgNPs have potential anti-mycobacterial activity against various pathogenic strains of Mycobacterium tuberculosis in vitro.

According to epidemiological surveillance data, the Beijing strain of Mtb is endemic throughout Asia as well as other parts of the world, and is often associated with highly virulent, multiple-drug resistance (Liao et al., 2015). The East African–Indian (EAI) lineage is prevalent in many tropical Asian countries (Liao et al., 2015; Wada et al., 2017). To determine the anti-mycobacterial activity of ALG-AgNPs against various pathogenic Mtb strains, we tested 7 clinical isolates of Mtb—two Beijing strains (W6, CHCH005), one EAI strain (CHCH029), two MDR strains (KVGH264 and KVGH376), and two XDR strains (TCHL002 and TCHL017)—and one standard reference strain (H37Rv). H37Rv, a pan-susceptible strain, is the quality-control strain used for anti-mycobacterial drug susceptibility testing by most laboratories (Ghiraldi et al., 2011; Zhao et al., 2018). In vitro anti-mycobacterial activity of ALG-AgNPs, reported as the minimum inhibitory concentration (MIC), was evaluated by the microplate Alamar blue assay (MABA). Alamar blue (AB) is an oxidation–reduction indicator dye that has been widely employed to evaluate the sensitivity of mycobacteria to anti-mycobacterial drugs (Franzblau et al., 1998). Reduction of AB dye occurs during mycobacterial growth, which converts the dye from blue to pink. Drug-mediated growth inhibition interferes with AB reduction and therefore the development of pink color. Treatment of pathogenic Mtb with increasing concentrations of ALG-AgNPs showed dose-dependent inhibition of Mtb growth (i.e., blue wells at higher ALG-AgNP concentrations) (Figure 2A). All clinical Mtb strains tested, except CHCH005 (Beijing strain), were sensitive to ALG-AgNPs at an effective concentration of <20 μg/ml; CHCH005 was more resistant (∼40 μg/ml) (Figure 2B). Specifically, the MICs of four drug-sensitive strains were 4.17 ± 1.04 μg/ml for H37Rv, 7.29 ± 2.76 μg/ml for W6 (Beijing strain), 41.67 ± 8.33 μg/ml for CHCH005 (Beijing strain), 2.60 ± 0.52 μg/ml for CHCH029 (EAI strain). For the two MDR-TB strains, the MICs were 1.04 ± 0.26 μg/ml for KVGH376 and 16.67 ± 4.17 μg/ml for KVGH264. Finally, for the two XDR-TB strains, the MICs were 7.29 ± 2.76 μg/ml for TCHL002 and 8.33 ± 2.08 μg/ml for TCHL017 (Figure 2B). This result demonstrates the potential anti-mycobacterial effect of ALG-AgNPs against various pathogenic Mtb strains.

To confirm the bactericidal activity of ALG-AgNPs against Mtb, we evaluated H37Rv, W6 (Beijing strain), KVGH064 (MDR), and TCHL017 (XDR) for enumeration of colony-forming units (CFUs) on 7H10 agar plates after 7 days of treatment with ALG-AgNPs or PBS (control). Colony formation after PBS treatment was significantly increased for all 4 Mtb strains compared with day 0 (Figures 3A–D). When treated with 1× MIC of ALG-AgNPs, colony formation of the 4 different Mtb strains showed almost no increase at day 7 compared with day 0, indicating that bacterial growth was inhibited by ALG-AgNPs. Of note, the 2-fold and 4-fold MIC treatments caused significant reductions in colony formation of the 4 different Mtb strains (Figures 3A–D). These results indicate that ALG-AgNPs have potency against both virulent and drug-resistant Mtb strains. Because the MICs for all but one of the strains are less than 20 μg/ml (the exception is < 50 μg/ml) (Figure 2B), we chose ALG-AgNP concentrations of 25, 50, and 100 μg/ml for use in further experiments.

ALG-AgNPs significantly inhibit the growth of cytosolic Mycobacterium tuberculosis in a macrophage infection model.

Mtb is recognized predominantly as an intracellular pathogen. When tubercle bacilli are inhaled into the lungs, they are engulfed by phagocytic cells such as pulmonary macrophages (Guirado et al., 2013; Srivastava et al., 2014). To mimic the behavior of human disease, we used PMA to differentiate THP-1 cells into macrophages in vitro, then infected the cells with Mtb [H37Rv, W6 (Beijing strain), KVGH264 (MDR), and TCHL017 (XDR)] to mirror Mtb internalization. The efficacy of ALG-AgNPs against intracellular Mtb was determined by treating Mtb-infected macrophages for 5 days with PBS (negative control), RIF (positive control), and increasing concentrations of ALG-AgNPs (25, 50, and 100 μg/ml), then measuring the mycobacterial burden in all infected samples by CFU analysis (Figure 4A). Significantly higher CFU counts were measured in the PBS treatment groups compared with the RIF and ALG-AgNP treatment groups for all four Mtb stains (Figures 4B–E). More interestingly, ALG-AgNPs showed good potency in the Beijing- and drug-resistant strains. These results confirm the ability of ALG-AgNPs to inhibit intracellular Mtb.

Latent TB, also known as non-replicating persistent (NRP) TB, develops when mycobacteria become adapted to the host’s immune system and survive within phagocytic immune cells despite hypoxia, nutrient deficiency, and low pH. These mechanisms subvert the immune system and allow infection without any symptoms for decades (de Martino et al., 2019). Current anti-TB drugs have poor activity against NRP Mtb. Thus, we investigated whether ALG-AgNPs have potential sterilizing activity against latent TB, by culturing the Mtb strain H37Rv either under oxygen-replete or hypoxic conditions to induce an NRP state. We then treated the cells with PBS, RIF, INH, and different amounts of ALG-AgNPs (25, 50, and 100 μg/ml) and assessed Mtb survival by CFU analysis. The NRP state was confirmed by measuring the expression levels of two genes, hspX, which is known to be transcribed in stationary-phase Mtb (Jee et al., 2018), and esat-6, which is known to be downregulated when Mtb cells enter dormancy (Arroyo et al., 2018). The results of quantitative polymerase chain reaction (QPCR) analysis showed that hypoxic stress significantly induced hspX mRNA expression (Figure 5A), whereas the expression of esat-6 was downregulated (Figure 5B). These data indicate that Mtb H37Rv had entered the NRP state in response to hypoxia. Moreover, CFU analysis showed significantly fewer colonies in the NRP state compared with the oxygen-replete condition after PBS treatment, which further confirmed the entry of H37Rv cells into dormancy (Figure 5C). Treatment with RIF and INH, which is used for preventive therapy in LTBI (Villa et al., 2019), produced significant reductions of the bacterial burden in both the oxygen-replete and hypoxia-induced NRP states. More importantly, treatment with ALG-AgNPs produced dose-dependent reductions in CFU counts of NRP H37Rv (Figure 5C). These results indicate that ALG-AgNPs have potential as a preventive therapy against Mtb reactivation by suppression of dormant bacilli during latent infection.

Previous studies have reported that the bactericidal activity of AgNPs is mediated by damage to bacterial cell walls via the release of silver ions that produce reactive oxygen species (ROS) to form a redox reaction (Dakal et al., 2016; Yin et al., 2020). Thus, we wondered whether our anti-mycobacterial ALG-AgNPs adopt this mechanism against Mtb. To test this possibility, we employed a bacterial cell-wall permeability assay using the Mtb strain H37Ra (we used this avirulent strain rather than H37Rv because our P2⁺ facility is not set up for fluorescence microscopy). We treated the cells with or without ALG-AgNPs at 50 μg/ml, which is the concentration that successfully suppressed the growth of the replicating-active drug-sensitive strains H37Rv and W6, the MDR strain KVGH264, and the XDR strain TCHL017 (Figures 4B–E) as well as the dormant Mtb strain H37Rv (Figure 5C) without inducing cytotoxicity in THP-1 cells [as measured by lactate dehydrogenase (LDH) released into culture supernatant from damaged or apoptotic cells (Supplementary Figure S3A)]. The bacterial cell-wall permeability assay showed that H37Ra bacilli treated with ALG-AgNPs had greatly increased permeability to the cell-wall-impermeable nucleic acid-binding dye propidium iodide, compared to cells treated with PBS (Figure 6). These results indicate that the anti-mycobacterial mechanism of ALG-AgNPs is related to disrupting the cell wall directly.

To confirm the anti-mycobacterial activity of ALG-AgNPs in vivo, we used zebrafish larvae as a TB animal model, which we infected with Mycobacterium marinum, a strain closely related to human Mtb but which in addition expresses DsRed fluorescent protein. We then treated the larvae with PBS, RIF or ALG-AgNPs, and assessed the anti-mycobacterial efficacy by measuring the red fluorescence signals of M. marinum-DsRed as well as the bacterial burden from the infected larvae. The images of M. marinum-DsRed-infected larvae show a decrease in DsRed fluorescent signal after treatment with either RIF or ALG-AgNPs compared with the PBS control (Figures 7A,B). The CFU analysis similarly demonstrated significant reductions in the mycobacterial burden in RIF- or ALG-AgNP-treated larvae compared with those treated with PBS (Figure 7C).

To examine Mtb infection in a mammalian model, BALB/c mice were infected with the Mtb strain H37Ra then treated with PBS, RIF, and different amounts of ALG-AgNPs (10, 50 mg/kg). Before the infection experiment, we determined the maximum tolerated dose (MTD) of ALG-AgNPs in these mice by two routes of administration. Mice were given the ALG-AgNPs by oral gavage or intravenous injection once daily for 2 weeks. During this period, mice were observed daily and no significant changes in animal behavior and clear signs of toxicity were found. The mice tolerated a dose up to 500 mg/kg by oral gavage and 250 mg/kg by intravenous injection. Body weight showed an increase of about 8.1% for oral gavage and 6.2% for intravenous administration on day 14 compared with day 1 (Supplementary Figure S3B) and there were no animal deaths (Supplementary Figure S3C). In the mouse TB model, mice treated with PBS showed a significant increase in H37Ra-mycobacterial burden in the lungs on day 28 compared to day 14 (Figure 7D). Importantly, ALG-AgNP treatment dose-dependently decreased the mycobacterial burden in the lungs on par with RIF (Figure 7D). These data demonstrate that ALG-AgNPs are non-toxic in vivo, and that their anti-mycobacterial potential is strong in both zebrafish and mouse TB animal models. These results suggest ALG-AgNPs could provide a new therapeutic option to treat Mtb.