Amino acid derivatives and resins were obtained from Iris Biotech (Marktredwitz, Germany). Reagents, such as N,N′-diisopropylcarbodiimide (DIC), triisopropylsilane (TIS), 1-hydroxybenzotriazole (HOBt), 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU), Isoniazid (INH), glyoxylic acid and 5(6)-carboxyfluorescein (Cf) were purchased from Sigma (Budapest, Hungary). Trifluoroacetic acid (TFA) and acetonitrile (AcN) were from VWR (Budapest, Hungary). N,N-dimethylformamide (DMF), dichlormethane (DCM), diethyl ether, and ethanol were purchased from Reanal (Budapest, Hungary)

For the in vitro assays, RPMI-1640 medium, fetal calf serum (FCS), and trypan blue were obtained from Sigma (Budapest, Hungary), while DMEM medium and 2 mM l-glutamine were from Lonza (Basel, Switzerland). Trypsin, nonessential amino acids, and penicillin-streptomycin were from Gibco (Thermo Fisher Scientific, Waltham, MA, USA). HPMI buffer (9 mM glucose, 10 mM NaHCO₃, 119 mM NaCl, 9 mM HEPES, 5 mM KCl, 0.85 mM MgCl₂, 0.053 mM CaCl₂, 5 mM Na₂HPO₄ × 2H₂O, pH = 7.4) was prepared in-house using components obtained from Sigma (Budapest, Hungary). BBL Löwenstein-Jensen medium was from Becton Dickinson (Környe, Hungary).

Peptides were produced on solid phase (Fmoc-Rink Amide MBHA, capacity = 0.67 mmol/g) resin in an automated peptide synthesizer (Syro-I, Biotage, Uppsala, Sweden) using standard Fmoc/tBu strategy with DIC/HOBt coupling reagents. Cf was coupled to the N-terminus of the peptides by using DIC/HOBt coupling method. Peptides were cleaved from the resin with TFA/H₂O/TIS (9.5:2.5:2.5, v/v) mixture (2 h, RT). After filtration, compounds were precipitated in cold diethyl ether, centrifuged (4,000 rpm, 5 min) and freeze-dried from water.

RP-HPLC purification was performed on an UltiMate 3000 Semiprep HPLC (Thermo Fisher Scientific) with a Phenomenex Jupiter Proteo C-12 column (250 × 10 mm) using gradient elution, consisting of 0.1% TFA in water (eluent A) and 0.1% TFA in acetonitrile/water = 80/20 (v/v) (eluent B).

Purified peptides were analyzed by LC-MS using a Thermo Scientific Q Exactive Focus Hybrid Quadrupole-Orbitrap Mass Spectrometer. For the separation, a Waters Acquity UPLC BEH C18 (1.7 µm, 150 × 2.1 mm) column was used with a flow rate of 0.3 ml/min.

Residual TFA and TFA counter-ion was removed by using an acetate-exchange resin. For the detailed description, see Supplementary Information 1.1 ).

Prior to conjugation, INH was derivatized as described previously (44). Briefly, 20.0 g, 0.146 mol INH (M = 137.1) was dissolved in 200 ml acetonitrile/water (1:1, v/v)) and reacted with 13.6 g, 0.148 mol glyoxylic acid monohydrate (M = 92.1, 1.01 equiv, dissolved in 50 ml water). After 1 h of stirring, the precipitate was filtered, washed with water and acetonitrile, and dried over P₂O₅ under vacuum (27.1 g, 96% yield, ESI MS calcd. 193.1 (Mmo) found 193.1).

The resulted isonicotinoylhydrazonoacetic acid (3.415 g, 17.7 mmol, M = 193.2) was reduced with 1.110 g, 17.7 mmol NaBH₃CN (M = 62.8) in 30 ml abs. ethanol (12 h stirring, RT). The obtained yellowish solution was filtered, evaporated to dryness, and recrystallized from methanol (2.777 g, 81.2% yield, ESI MS calcd. 195.1 (Mmo) found 195.1)

The product, isonicotinoylhydrazinoacetic acid, (7 equiv) was added to the peptidyl resins described above, in the presence of 7 equiv. of DIC and HOBt (2 h, RT). INH-peptide conjugates were then cleaved from the resin with TFA in the presence of scavengers (2.5% H₂O, 2.5% TIS; 2 h, RT), precipitated with cold diethyl ether, dissolved in water, freeze-dried, and purified as mentioned above.

MonoMac-6 human monocytic cell line (45) (DSMZ No. ACC 124, Braunschweig, Germany) was maintained as an adherent culture in RPMI-1640 supplemented with 10% heat-inactivated fetal calf serum (FCS) l-glutamine (2 mM) and gentamicin (35 μM) at 37°C in a humidified atmosphere containing 5% CO₂.

EBC-1 lung squamous carcinoma (RRID: CVCL_2891) (46–48) and CALU-1 human lung epidermoid carcinoma (Sigma 93120818) (49, 50) were maintained in DMEM medium containing 10% FBS and supplemented with 2 mM l-glutamine, 1% nonessential amino acids, 1 mM sodium pyruvate, and 1% penicillin-streptomycin (from 10,000 units penicillin and 10 mg streptomycin/ml). Cells were cultured at 37°C, 5% CO₂-humidified atmosphere.

Mycobacterium tuberculosis H37Rv (ATCC 27294) was grown in Sauton’s liquid medium to exponential growth phase (approx. 3–4 weeks). The Sauton’s medium was prepared in-house as described in (51) with the addition of 0.05% (w/v) Tween-80 to prevent bacterial aggregation (since Mtb H37Rv tends to form clumps). The bacterial suspension was homogenized by ball-milling using sterilized stainless steel grinding balls, and after dilution, it was used for the inoculation of the test tubes.

All experimental procedures with infectious Mtb were performed in a biosafety level 3 (BSL-3) laboratory at the National Public Health Center (Hungary), respecting the institutional containment level 3 laboratory management and biosecurity standards based on applicable national and EU Directives.

Antimycobacterial activity of INH and INH-conjugates was determined on Mtb H37Rv using broth dilution method in Sula semisynthetic medium (pH = 6.5) (52–54). Dilution series of the compounds were prepared in DMSO and added to 5 ml Sula medium-containing test tubes. The Mtb bacterial suspension (0.5 McFarland equal to 1.5 × 10⁸ CFU/ml) were diluted 10⁴ times and added to the test tubes. The minimal inhibitory concentration (MIC, reported in micromolars) was determined after incubation at 37°C for 28 days. MIC was the lowest concentration of a compound at which no visible growth of the bacteria occurred. The antitubercular effect of the tested compounds was confirmed using a colony-forming unit (CFU) determination by subculturing 100 µl of the supernatant onto drug-free Löwenstein-Jensen solid medium (37°C, 28 days), which is a selective medium specifically used for the culture and isolation of Mtb. Experiments were repeated at least two times.

The internalization of the compounds was measured in MonoMac-6 cells. For the assay, cells were treated with Cf-labeled peptides at 5, 10, and 20 µM final concentration and were incubated for 2 h. After centrifugation (1,000 rpm, 5 min) and washing with serum-free RPMI medium, the supernatant was removed, and 100 μl 0.25% trypsin was added to the cells. After 5 min incubation, 0.8 ml 10% FCS/HPMI medium was added, then cells were washed and resuspended in 0.3 ml HPMI medium. The intracellular fluorescence intensity of the cells was measured on a BD LSR II flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA) on channel FITC (emission at λ = 505 nm), and data were analyzed with FACSDiva 5.0 software (BD Biosciences). All measurements were performed in triplicates, and the mean fluorescent intensity together with the standard error of the mean (SEM) was graphically presented. For analysis of statistical significance, unpaired t-test was used.

Cellular uptake and localization were visualized by confocal laser scanning microscopy. MonoMac-6 cells were seeded 1 day prior to treatment into coverslips (thickness 1, Assistent, Karl Hecht GmbH, Sondheim vor der Rhön, Germany) containing a 24-well plate (Sarstedt, Nümbrecht, Germany) for microscopy studies at a density of 7.5 × 10⁴ cells/well, in 1 ml complete DMEM medium. The following day, cells were treated with Cf-peptide at a concentration of 10 µM (diluted in serum-free DMEM medium) for 2 h. Lysosomes were stained by LysoTracker Deep Red (Invitrogen, Waltham, MA, USA) for 30 min, and nuclei were stained by Hoechst 33342 solution (Thermo Scientific, USA). After each step, cells were washed three times with serum-free medium, then cells were fixed by 4% paraformaldehyde for 15 min and mounted to microscopy slides with Mowiol 4-88 (Sigma, Budapest, Hungary). Imaging was performed by a Zeiss LSM-710 system (Carl Zeiss microscopy GmbH, Oberkochen, Germany) with a ×40/1.4 Plan-Apochromat oil immersion objective using lasers with the parameters: Cf peptides λex = 488 nm, λem = 541 nm, nuclei λex = 405 nm, λem = 467 nm (Hoechst 33342), lysosomes λex = 633 nm, λem = 720 nm (LysoTrackerTM Deep Red) or λex = 577 nm, λem = 590 nm (LysoTrackerTM Red DND-99) were used. Zeiss ZEN lite software (Carl Zeiss Microscopy GmbH, Jena, Germany) was used for image processing.

EBC-1 lung squamous carcinoma and CALU-1 human lung epidermoid carcinoma were seeded on 24-well plates with Transwell inserts (TW) [Nunc, Sigma polycarbonate microporous membrane, 0.4 µm pore size (growth area 0.412 cm²)]. Before use, TW inserts were incubated with ICM. On day 1, 300 μl of CALU-1 suspension in complete DMEM medium (8.0 × 10⁵ cells) was pipetted onto the surface of the TW apical chamber and 500 μl complete DMEM medium was added to the basolateral side. In order to avoid nonattached cells, after incubation at 37°C for 5–6 h, the apical medium was removed and replaced with fresh complete DMEM medium. On day 3, the medium was changed, and CALU-1 cells were grown up to confluence (which was checked prior and after the experiments) with CellTracker Green (CMFDA (5-chloromethylfluorescein diacetate, Invitrogen, C2925). CMFDA was dissolved in serum-free DMEM medium to reach 10 µM concentration. After 35 min of incubation, the cells were washed with serum-free DMEM and cell nuclei of CALU-1 monolayers were stained with Hoechst 33342 solution, according to the manufacturer’s suggestions. ZEISS Axio Observer Inverse Imaging Platform (Zeiss) (equipped with Led3 470 blue, Led4 555 green; filter set: 25HE) was used to capture the TW insert membrane with CALU-1 monolayer. On day 5, the medium was changed, then Cf-peptide was added to the CALU-1-containing apical side at 10 and 20 µM concentrations, and the system was incubated for 45 min or 3 h (37°C, 5% CO₂). The TW chamber was removed after incubation. EBC-1 cells were studied using flow cytometry (BD LSR II).

EBC-1 cells were treated without the presence of the TW inserts as controls (on day 4 prior to treatment, EBC-1 cells (10⁵ cells/well in complete DMEM).

In this study, we have employed a modified method based on (55). Briefly, micromolds for casting 3D Petri dishes (MicroTissues, Sigma, 5 × 7 array) were filled with molten agarose (2% (w/v) in PBS). The gelled agarose dishes were equilibrated with serum-free DMEM (2 ml/2 h, 37°C). Cells were seeded after incubation (1.8 × 10⁴ cell/µL in DMEM CM). Prior to seeding process, nuclei were stained with Hoechst 33342 solution (0.2 µM) for 30 min ( Figure 2 ). Based on our previous work (55), the nuclear stain could not successfully penetrate into deeper layers of the spheroids (mostly localized at the spheroid surface), and towards the center of the spheroids the Hoechst 33342 signal was gradually decreasing. Stained cells were incubated in 2 ml DMEM CM for 48 h while cell-to-cell adhesion drives the aggregation and formation of spheroids. To monitor the condition of the spheroids, bright-field images were captured (Olympus CX41). After 48 h of incubation, spheroids were washed two times with fresh ICM DMEM and were treated with the Cf-Dhvar4 peptide (10 µM/2 ml ICM DMEM for 3 h). After the treatment, spheroids were washed two times with ICM DMEM and three times with PBS. Spheroids were fixed with 4% paraformaldehyde for 15 min (37°C) and were washed three times with PBS, and then harvested and transferred from the agarose micro-wells into µ-Slide 8-well uncoated Ibidi chambers for imaging (Zeiss LSM 710; ×10 dry objective (×10/0.45 M27)). The same excitation and emission wavelengths were used as for the 2D confocal microscopy imaging. Z-stack images were obtained by scanning the spheroids from the bottom of the spheroid with 11 µm distance between each scanning plane ( Figure 2 ). Images were processed with ZEN 3.0 blue lite software. Line scan analysis was performed by NIH ImageJ software, two spheroids were analyzed (eight scans per spheroid, n = 16 in total). Line scans were carried out using greyscale images; grey value corresponds to the intensity of a given pixel on a scale of 0 to 255. As spheroids are of various shapes and sizes, the diameter of the scanned area (horizontal axis distance) was normalized to 1 for better comparison. To average the intensity values from line scans of spheroid sections of slightly different sizes and shape, all line scan lengths were normalized to 1, yielding normalized diameter. Calculation of average and SEM of multiple curves was carried out using the software OriginPro 2018.

Prior to experiment, MonoMac-6 cells were cultured in a 24-well plate for 24 h (2 × 10⁵ cells/1 ml medium well). Adherent cells were infected with Mtb H37Rv at a multiplicity of infection (MOI) of 10 for 4 h. Nonphagocytized extracellular bacteria were removed, and the culture was washed three times with serum-free RPMI. The infected monolayer was incubated for 1 day before antitubercular treatment. Infected cells were then treated with the compounds at 5, 10, and 50 μM concentrations. After 3 days, the treatment was repeated with fresh solution of the compounds for an additional 3 days. After washing steps to remove the compounds, infected cells were lysed with 2.5% sodium dodecyl sulfate solution. Ten times serial dilutions were then prepared from the lysates and plated onto BBL Löwenstein-Jensen tubes. After 4 weeks of incubation, the CFU of Mtb was enumerated, and the bacterial number was estimated and graphically presented. Experiments were repeated at least twice.

Animal experiments were carried out in accordance with the guidelines of EU Directive 2010/63/EU and Hungarian laws and were approved by the Hungarian Scientific Ethical Committee on Animal Experimentation under the following protocol number: PE/EA/2569-4/2016.

For infection experiments, 6–8-week-old BALB/c female mice were kept in ventilated cages and allowed ad libitum access to water and to a standard pellet diet. Mice were infected with a mid-logarithmic culture of Mycobacterium tuberculosis H37Rv (10⁶ CFU/ml in PBS, 200 μl, i.p.). Two weeks after, therapy of 20 mg/kg bw of INH in 100 µl PBS (per os) and 30 mg/kg bw of INH-Dhvar4 in 100 µl PBS (s.c.) has started. Mice were treated twice a week for 5 weeks. Mice were monitored twice per day for the parameter attitude, locomotion, breathing, curiosity, nasal secretion, grooming, and dehydration. Mice that lost >20% body weight and had evidence of severe symptoms, such as scruffy coat, inactivity, loss of appetite, poor locomotion, or painful posture, were sacrificed before the termination of the experiments.

One week after the 10th treatment, mice were euthanized, and their organs were removed. To determine the viable bacteria, a portion of lung and spleen were homogenized in a tissue homogenizer using ceramic beads (MagNA Lyser Green Beads, MagnaLyser, Green Beads, Roche, Switzerland) in Bouillon broth (two times 60 s, 7,000 Hz), then 100 μl of the supernatant was plated onto BBL Löwenstein-Jensen Medium. CFU were counted after 4 weeks of incubation. For histopathological analysis, each animal’s lung, spleen, liver, and kidney were removed and fixed in 8% neutral buffered formalin for 24 h at room temperature before removing from the high-containment facility. Tissue specimens were dehydrated in a series of ethanol and xylene baths and embedded in paraffin wax. Sections (3–4 μm) were stained with hematoxylin and eosin (HE). For in situ visualization of the acid-fast bacilli, the Ziehl-Neelsen (ZN) staining method was applied on similar pretreated sections. Slides were analyzed in an Olympus BX53 microscope (Japan), and photomicrographs were obtained with an Olympus SC100 high-resolution digital color camera using the Olympus cellSens imaging software platform.

For INH conjugation, representative cell-penetrating (CPP) and antimicrobial peptides (AMP) were chosen. These cationic, amphiphilic peptides are favored in the transportation of various cargos due to their high internalization rate. The used peptide series contains well-known CPPs such as penetratin, transportan, and Tat, frequently studied AMPs such as magainin and buforin, and designed analogs such as Dhvar4, Crot(1-9,38-42), and CM15. Also, a receptor-binding tuftsin analog (OT20), which was reported as a macrophage-targeting peptide, was also applied. Table 1 shows the theoretical and experimental monoisotopic molecular masses of the peptide, measured on a high-resolution mass spectrometer. Mass spectra and RP-HPLC chromatograms are available at https://figshare.com/s/48a023a25068f7658adf, DOI: 10.6084/m9.figshare.16644901. Retention times observed on the analytical RP-HPLC chromatograms were also presented, which had a strong correlation to the overall hydrophobicity of the compounds. To study the internalization rate and intracellular localization, fluorescently labeled analogs were also synthesized. Note that the retention time of Cf-peptide are somewhat higher than unlabeled analogs, and in some cases, two peaks were detected due to the isomeric distribution of 5(6)-Cf molecule.

INH was coupled to the N-terminus of the peptides, using a previously described method (44). Resulted INH-peptide conjugates were then cleaved from the resin with TFA in the presence of scavengers ( Figure 3 ).

Generally, after the cleavage and purification of synthetic peptides, TFA can remain in the lyophylisates. Residual TFA content can be deteriorative during the in vivo experiments; therefore, the removal of TFA and the replacement of TFA counter-ion to a less harmful acetate ion is strongly recommended. For that, an anion-exchange method was applied, which is described in the Supplementary Information 1.1 ).

Next, we examined the enzymatic stability of two peptide conjugates in rat liver lysosomal homogenate (see Supplementary Information 2 for the method description). Peptides were dissolved at 0.025 µg/µl concentration then allowed to react with the lysosomal homogenate. Degradation of the conjugates was fast. After 30 min of incubation, the peaks of the intact conjugates almost disappeared from the chromatograms ( Supplementary Figure S1 ). In the case of INH-Dhvar4, the most stable fragment was conjugated, which gave the main peak after 24 h of incubation, containing a tripeptide (KRL) covalently attached to INH ( Figure 4A ). In the case of INH-penetratin, the most stable fragment was a tetrapeptide (RQIK) with INH ( Figure 4B ).

First, cellular uptake of MonoMac-6 human monocytes was studied. This cell line was utilized as a host for Mtb in the next set of experiments. Cells were treated with Cf-labeled peptides at 5, 10, and 20 µM concentrations for 2 h. In order to remove the surface-bound peptides, trypsin was added to the cells; then, the intracellular fluorescent intensity was measured by flow cytometry. At the studied concentration, none of the peptides showed more than 20% cytotoxicity (the relative viability was >80%), except melittin, where the cytotoxicity was almost 50% at the highest treatment concentration ( Figure 5A ). Highly efficient internalization was measured for the conjugates. The two most effective peptides were the Dhvar4 and penetratin ( Figures 5B, C ).

Parallel with flow cytometry measurements, confocal laser scanning microscopy images were also captured in order to assess qualitative information on intracellular localization. LysoTrackerTM Deep Red was used for lysosome and Hoechst 33342 for nuclear staining to distinguish subcellular localization of the Cf-peptide. The experiment was carried out after 2 h of incubation time, and representative images are presented in Figure 6A . As was expected, based on the results of flow cytometry measurements, all of the peptides were internalized efficiently. In the case of peptide transportan and CM15, fluorescent signals were observed mainly in lysosomal compartments ( Figure 6A ). Peptide Dhvar4 and penetratin could be imaged in the cytosol with almost no or low level of colocalization with lysosomal staining, which suggests that these peptides internalize and display a ubiquitous distribution in the cytosol. Considering all these data, it was presumed that the Dhvar4 and penetratin peptides enter the cells in a concentration-dependent manner by direct penetration or by endocytosis, followed by an endosomal release ( Figure 6B ). This was further proved by the effectivity of INH-Dhvar4 and INH-penetratin on intracellular bacteria. MonoMac-6 cells are excellent as host cell models; however, preparing them for confocal microscopy is quite challenging. We have optimized the cell preparation process (including the Cf-peptide treatments, fixation, staining protocol, incubation times, concentrations, washing steps, etc.). Our main goal was to achieve no cell morphology deformation, no staining differences in order to increase the confidence that Cf-peptide and lysosomal stain occupy the same or different structure. Colocalization of Cf-peptide and LysoTracker may be subjectively identified by the cooccurrence, the simple spatial overlap which combined the contribution of both signals (green and red) when the images of each signal are superimposed (merged, see Supplementary Figure S3 ). So, for example, colocalization of Cf-transportan and LysoTracker was apparent in vesicles that appear orange because of the combined contributions of green signal (Cf-transportan) and red fluorescence (LysoTracker). We have chosen a representative Cf-peptide set to quantify roughly the differences among peptide localization using greyscale analysis. The detailed process and graphs were presented in the Supplementary Material . We should point out, however, that this approach has its limitation, because our main goal was to visualize the differences and our experiments were mainly designed for qualitative rather than quantitative comparisons.

In addition to the internalization experiments, the surface morphology of INH-peptide-treated and fixed EBC-1 cells was also studied with atomic force microscopy (AFM) ( Supplementary Information 1.3 ). High-resolution imaging of the cell surfaces was performed with a Flex-Axiom AFM system (Nanosurf, Liestal, Switzerland).The representative images and surface roughness of the native and treated cells are shown in Supplementary Figure S2 . Mean Rq values are somewhat higher in the case of peptide and conjugate-treated cells, compared with untreated cells, but the differences do not suggest membrane disruption. As a result, we can conclude that membrane integrity maintained after the Dhvar4 and INH-Dhvar4 conjugate treatment.

First, we studied the penetration ability of the Cf-peptide on Transwell using submerged, cocultured noncontact monolayers of CALU-1 and EBC-1 cells as a simple in vitro bronchial interface model ( Figure 7 ). Cf peptides were added to the CALU-1 containing apical chamber at 10 and 20 µM concentration and after 45 min or 3 h of incubation. EBC-1 cells from the basolateral chamber were stained and fixed prior to confocal microscopic analysis. As control experiments, treatments without TW were also carried out ( Figure 7G ). All four peptides were able to cross the CALU-1 monolayer, and they were internalized into EBC-1 cells. According to the green signal intensity, the internalization rates followed the same pattern as obtained in the cellular uptake studies ( Figure 5C ), namely, penetratin, and Dhvar4 have higher internalization rates (after crossing membrane and CALU-1 barrier) than the other peptides. Similar tendencies were observed after 3 h of incubation (data not shown).

Next, we assessed the penetration ability of Cf-Dhvar4 peptide on spheroids formed from EBC-1 cells. Cells were treated with the peptide for 3 h at 10 µM concentration. Nuclear staining with Hoechst 33342 was also applied. Fixed spheroids were visualized by confocal laser scanning microscopy. The EBC-1 spheroids had an average diameter of 350 and 400 nm, respectively ( Figure 8 ). To visualize the penetration profile of the Cf-Dhvar4, we have carried out line scan analysis extracted from the images. Spheroids were scanned in the z-direction with a step size of 11 µm (z1–z5). The deepest z-section from the surface of a spheroid presented here is approximately 55 µm (z5). To have comparable fluorescent intensity values among the zones, normalized grey value data were standardized as the proportion of the mean normalized intensity in the peripheries (0%–100%). Towards the center of the spheroids, the Cf signal was gradually decreasing ( Figure 8 ). Based on the presented results, we can conclude that Cf-Dhvar4 has fair penetration ability on EBC-1 spheroids.

Antibacterial effect against extracellular Mtb was determined using a conventional broth dilution method. The MIC value for INH was between 0.1 and 0.5 µM (0.01–0.07 mg/L), which is in accordance with literature data (56, 57). INH-peptide conjugates exhibited similar MIC values, namely 0.5–1.25 µM, which proved that INH maintained its activity in the conjugated form. This observation is in line with our previous results published recently (39, 44).

Intracellular killing activity of the INH-peptide conjugates was tested on Mtb H37Rv-infected MonoMac-6 human monocytes. Infected cells were treated two times for 3 + 3 days at 5, 10, and 50 µM concentrations. After washing and SDS lysis, Löwenstein–Jensen tubes were inoculated with the lysates, and bacterial colonies (CFU) were enumerated after 4 weeks of incubation. The most efficient compound was the INH-Dhvar4; therefore, this conjugate was studied in more detail. Cytotoxicity to the host cells was also measured ( Figure 9A ). The viability of MonoMac-6 cells, compared with medium-treated control cells, was >80% at the highest treatment concentration (50 µM) ( Figure 9B ).

Intracellular killing efficacy was further studied with different controls. INH itself was not effective against the intracellular form of Mtb at the studied concentration range, which is in line with previous observations (58, 59). Dhvar4 peptide itself was also not effective on the intracellular population of Mtb, but INH and Dhvar4 mixture showed a moderate antibacterial effect. When INH was covalently conjugated to the Dhvar4 peptide, significantly reduced bacterial growth was measured within the monocytes ( Figure 9C ). As low as 10 µM concentration was enough from the conjugate to be able to eliminate Mtb from the cells.

Taking together all in vitro data obtained for the conjugates, INH-Dhvar4 proved to be the optimal candidate for the in vivo evaluation.

In this study, a BALB/c murine model was applied with i.p. injection of 10⁶ CFU/ml Mtb bacteria ( Figure 10A ). This type of infection induces a course of slowly progressive and systemic disease and provides an opportunity for a prolonged experimental setup. The well-being of the animals, monitored by the body weight, showed an unbiased gain till the last week of the experiment ( Figure 10D ). After euthanasia, organ homogenates were cultured on Löwenstein-Jensen media and CFU were counted. Beside the lung, mice contained a high number of Mtb within the spleen, leading to the progressive systemic spread of the bacteria. In histological sections of the spleen of untreated control animals, rod-shaped acid-fast bacteria were observed, indicating the successful experimental design to model TB ( Figures 10B, C ). INH was administered per os (20 mg/kg bw), while INH-Dhvar4 conjugate was administered s.c. (30 mg/kg bw). Note that the INH content of the conjugate was only 2 mg. Each treatment group contained five animals, but after the seventh week of the experiment, one mouse from the untreated (but infected) control group was euthanized. The measured weight loss indicated this action (4 data points on the CFU/organ panels for the untreated control group). Clearance of Mtb from the lung was experienced after INH treatment ( Figure 10E ), while only a moderate decrease was observed in the spleen ( Figure 10F ). The same decrease in the number of Mtb counted from the spleen homogenates was observed for the INH-Dhvar4 conjugate, which contains only 1/10 of INH compared with the orally administered drug. The progress of the infection is extremely a complex process in vivo, and it is affected by the relative density of bacterial subpopulations, tissue microenvironments and consequently the respective susceptibility of the individual animals to therapy. The aforementioned factors can result in the heterogeneity of the efficacy data. Our murine model was intended to be used as a slow progress one. The drawback of this model is that some animals had higher spleen and/or lung bacterial loads to resist the activity of the conjugate and to achieve the sterile cure.