**3. Discussion**

We and others have shown that CD271+BM-Mesenchymal stem cells are natural reservoirs for dormant Mtb [11,13,14]. In this study, we engineered bone-homing liposomes decorated with Ald and PEG to actively target Mtb in its novel niche. Our BTL-NPs showed significant bone binding capabilities and resulted in the decreased Mtb load of CD271+BM-MSCs. Although bone-homing liposomes have been reported earlier [42,43], this is the first study to formulate and test alen-conjugated PEG liposomes to target BM-resident Mtb. Multiple functions were kept in mind and were incorporated into the engineering of BTL-NP: (1) bone-homing capability, (2) stealth properties by PEG and (3) size. For bone-homing, alendronate, a biphosphate was attached to the liposome surface resulting in high affinity and homing of BTLs to the bone microenvironment. Alendronate belongs to the bisphosphonate family of drugs and is generally used for cancer therapy. Previous studies show that alendronate has no cytotoxic effects and it does not alter the differentiation or self-renewal properties of stem cells residing in the bone marrow [35,43–45]. It is known that alendronate treatment can be helpful in the treatment of bone disease, especially bone cancer [46]. Alendronate, 5 or 10 mg/d, may increase the bone mineral density in women (postmenopausal) and men with primary osteoporosis [47]. Alendronic acid, administered at a 70 mg once weekly or 35 mg twice weekly dose, has shown to increase BMD (bone mineral density) as 10 mg/d and is already used clinically in the treatment or prevention of osteoporosis [48]. While alendronate on systemic administration homes to the bone microenvironment, PEG improves the systemic circulation of liposomes [34,49]. High circulation time owing to PEGylation, high bone-homing capability of BTL-NPs due to alendronate and optimal size in our study may have rendered them effective for clearance of CD271+BM-MSCs resident Mtb.

Our in vivo data indicates that BTL-NPs may have increased the available intracellular RIF and INH leading to Mtb clearance inside BM-MSCs. We also speculated that the PEG-modification of BTLs passively increased the bone marrow selectivity by inhibiting the hepatic uptake, resulting in BTL-NPs to target BM macrophages [35,50]. Moreover, it is possible that after getting trapped in macrophages, BTL-NPs gradually ge<sup>t</sup> digested by lysosomal enzymes leading to controlled drug release from macrophages to the surrounding BM tissue, further increasing intracellular drug. While MSCs are known to have drug efflux pumps that gradually efflux out drugs/antibiotics to protect the cells from cytotoxicity, our data sugges<sup>t</sup> that the BTL-NPs may be delivering the drugs at a faster pace in comparison to their efflux rate. In aggregate, it is possible that BTL-NPs may increase the available intracellular drugs, rendering them more effective in Mtb killing compared [51] to free drugs or NT NPs. However, we acknowledge that this speculation requires further investigation.

Lung-homing liposomes targeting Mtb lowers treatment duration and hepatotoxicity in pulmonary tuberculosis [40,52,53]. Swami et al. recently showed that PLGA nanocarriers could be used for the targeted delivery of drugs to bone cancer [35]. These two studies taken together led to the hypothesis that BM-homing liposomes could aid in clearing tubercle bacilli present in the BM cells with reduced off-target toxicity. We acknowledge that, although the entire Mtb population from BM cells was not cleared leading to relapse in very few mice, a decrease in relapse rate is of considerable clinical significance. Few Mtb cells remaining in CD271+BM-MSC (Figure 6B) despite BTL treatment, indicate that there might be some subpopulation of Mtb which (a) may be pushed into deeper stage dormancy, (b) may have developed into drug resistance variants or (c) were present in an unreachable compartment of CD271+BM-MSCs. Any of these three possibilities require further investigation as deeper understanding may lead to complete clearance in BM cells. Nevertheless, currently, there is no treatment/drug which targets BM-MSCs resident Mtb and since alendronate is already in clinical use for bone strengthening, alendronate-based bone targeting may be of clinical significance. Once clinically investigated, BTLs may be used to reduce the relapse rate in post-chemotherapy patients. Overall, our study demonstrates that BTL-NPs can be used for the effective delivery of existing anti-TB drugs to bone microenvironment through specially formulated liposome carriers.

### **4. Materials and Methods**

### *4.1. Alendronate Tagging of DSPE-PEG*

First, 23 mg of DSPE-PEG (2000)-carboxylic acid (1,2-distearoyl-*sn*-glycero-3-phosphoethanolamine -N-[carboxy(polyethylene glycol)-2000]ammonium salt(Avanti lipids polar, Inc., USA, Cat Number 880125P) was dissolved in 5 mL of acetone. Next, 4.2 mg N,N-Dicyclohexylcarbodiimide (DCC) and 2.4 mg N-hydroxy succinimide (NHS) were added for activation overnight at RT. Syringe assisted removal of insoluble by-product (Dicyclohexylurea ) was carried out. Following this drying of lipids was carried out for 2 h through nitrogen. For alendronate tagging, activated lipid and 2 mg alendronate sodium trihydrate (Ald, A-4978, Sigma-Aldrich, St. Louis, MO, USA) were dissolved in a mixture of DMSO and water for 24 h. This was followed by 24 h dialysis against water and subsequent drying under nitrogen.

### *4.2. Preparation of PEG Liposome and BTL PEGylated Liposome NPs with Ald Tagged PEG-PE and PC (with INH and RIF Encapsulation)*

Liposome NPs were prepared as described previously [21,54,55]. In brief, Egg-PC and Egg-PE in a molar ratio of 8:2 were mixed in a 100 mL round bottom flask in chloroform and rotated under vacuum at 37 ◦C until a thin lipid layer is formed followed by desiccation for 2 h. For alendronate PEGylated liposomes a.k.a. BTLs, PC (2% of total lipids) + Ald-PE-PEG2000 (2.5% of total lipid content) + PE-mPEG2000 (2.5% of total lipid content) ratio was used. The resultant dry lipid was vortexed with PBS until complete lipid dispersion. For RIF-encapsulated BTL-NPs, Rifampicin (M.P Biomedicals, CAS #13294-4-1, Irvin, CA, USA) was dissolved in chloroform, and for INH-encapsulated BTL-NPs, isoniazid (Sigma-Aldrich, CAS #54-85-3, St. Louis, MO USA) was dissolved in saline at appropriate concentrations. Next, liposomal dispersion was centrifuged at 50,000 rpm (60 min) at 4 ◦C. This step was done to ge<sup>t</sup> rid of un-encapsulated drugs. The pellet obtained was resuspended and centrifuged twice to remove the traces of un-encapsulated drugs. For equal size and decreased variation, the final suspension obtained after three washing was extruded sequentially through 3 membrane filters. The liposomal suspension was assayed spectrophotometrically for the presence of RIF and INH, according to the modified method described [56]. Millipore membrane filters used were of Type RA from Millipore Corp., Bedford, MA, USA.

### *4.3. TEM and Size Distribution*

The liposomal suspension was diluted in distilled water and placed upon 300-mesh carbon-coated copper grids and air dried for analysis [21,54]. The liposomes were visualized under a transmission electron microscope (JEOL 2100F). Zetasizer Nano ZS (ZEN 3600; Malvern Instruments, Worcestershire, UK) was used to measure NP size and zeta potential. Detailed NP characterization was performed as previously described by Rajendran et al. [55]. Briefly, NPs were diluted 1000-fold in water in order to make homogenous suspension (scatter angle = 90 ◦C, temperature = 25 ◦C) and analyzed by placing in a zeta cell (DTS-1060C).

### *4.4. Drug Concentration Estimation and Encapsulation E*ffi*ciency*

Drug encapsulated liposomes were lysed by vortexing in 70% ethanol and incubated at 60 ◦C. Drugs (INH and RIF) were quantified by spectrophotometric absorbance method against a free drug standard curve at 263 nm and 479 nm, respectively [56].

Encapsulation e fficiency = (Drug concentration in eluent/drug concentration loaded) × 100

### *4.5. Cellular Uptake by Mouse CD271*+ *BM-MSCs*

Briefly, 2 × 10<sup>3</sup> CD271+BM-MSCs and RAW 246.7 cells were seeded in 8-well glass-bottom chambered slide (Ibidi) in D-MEM/F-12 medium with GlutaMAX ™-I and DMEM (Thermo Fisher Scientific, Cat. Number 10565018, Cat. Number 12100046, Waltham, MA, USA) serum supplemented media, respectively. After 4–5 h of seeding, prior to imaging, Coumarin6 NP (e ffective C6 concentration = 2 μg/mL) was added. Live cell imaging was carried out to monitor time lapsed uptake of engineered BTL with Coumarin6 using Nikon Camera Nikon Real-Time Laser Scanning Confocal Microscope, Model A1R. Notedly, C6 is a common fluorescent model drug used for tracking NP uptake [57].

### *4.6. NP Binding Assay with Bone Chips*

Mice were sacrificed to the excise femur bone. Next, bone chips from femur were washed with 1× PBS, incubated with alendronate BTLs encapsulating Coumarin6 (alen BTL C6) or non-targeted liposomes C6 (NT liposomes C6) for 30 min followed by washing. The whole procedure was carried out in 8-well glass-bottom confocal slides. Lastly, each well was added with PBS for visualization through a confocal microscope (Nikon Camera Nikon Real-Time Laser Scanning Confocal Microscope, Model A1R) for the binding of C6 NPs with mouse femur bone chips obtained [35].

### *4.7. In Vitro Release from BTL-NPs and Tissue Uptake*

For in vitro release assay 500 μL of NPs suspension in phosphate-buffered saline (PBS) were added to a dialysis bag (MWCO 12,000–14,000; Sigma-Aldrich, St. Louis, MO, USA). The dialysis bag was placed into 25 mL of PBS (pH 7.4) taken in a dissolution vessel at 37 ◦C and stirred at 50 rev min−1. At periodic intervals, 500 μL samples were aliquoted from the dialysate and then an equal volume of PBS (pH 7.4) was added. INH and RIF were quantified by measuring absorbance against the free drug standard curve at 263 and 479 nm, respectively. The modified procedure described previously [58] was performed for tissue uptake assay. Briefly, after 8 h of administration of C6 BTL, C6 NT or C6 PBS intravenously, femur and tibia pair were excised followed by chloroform: methanol extraction. The same effective concentration of C6 (150 mg/kg) was administered to all three mice groups. Fluorescence of resulting suspension was determined using Cary Eclipse Fluorescence Spectrophotometer (Agilent) in a quartz cuvette at the emission/excitation of 460/525 nm. The concentration of C6 in the suspension was determined by using the standard curve of C6. % injected a dose of C6 = Conc. of C6/Total injected C6) × 100 per g of organ. Data were represented as % dose injected per organ (per organ in biodistribution study = a pair of tibia and femur from single mice).

### *4.8. In Vivo Administration and Assessment of BTL-NPs <sup>E</sup>*ffi*cacy*

Pathogen-free, 6 weeks female C57bl/6J mice were obtained from the National Institute of Nutrition (Hyderabad, India). Animals were housed and fed according to the standard norms of Animal biosafety level 3 (ABSL3) at Jawaharlal Nehru University, New Delhi. All studies were carried out post-approval from Institution of Animal Ethics Committee (IAEC) and the Institution of Biosafety Committee (IBSC) at Jawaharlal Nehru University, New Delhi.

Mice infections were performed as previously described [13]. Briefly, C57bl/6J female age 6 weeks were infected with H37Rv 2 × 10<sup>6</sup> CFU through tail vein and followed by anti-TB drug therapy (RIF = 0.1 g/L, INH = 0.25 g/L) for 90 days to achieve a status where only Mtb survives in CD271+BM-MSCs while other organs are rendered sterile [11,13]. This model known as the Cornell model has been known for decades where animals treated with anti-Tb drugs wipe out Mtb from lung, liver and spine rendering them sterile, however for no clear reason the disease is relapsed [7,59,60]. It has been shown in mice that this relapse could be due to Mtb present in a novel host cell namely CD271+BM-MSCs [13]. It is postulated that Mtb mesenchymal stem cells may travel from BM to the lung via blood circulation and may cause a relapse of pulmonary TB [11,13]. We have used

this Cornell model for targeting Mtb in BM-MSCs. Animals were treated with various treatments including free drugs or BTL-NPs intraperitoneally. The concentration of INH and RIF remained under the safe limits which are 10 mg/kg (standard) for both the drugs [61]. Briefly, INH and RIF average concentrations fell in the range of 2–2.3 mg/mL and 2.66–3.1 mg/mL for their respective NP suspensions (BTL or NT). All liposomes and free drug solutions were diluted to the e ffective dosage of 3 mg/kg (RIF) or 4 mg/kg (INH) per mice in 150 μL. To be noted, NPs were freshly prepared for each administration and concentrations of drug in each formulation was determined using a standard curve for each drug as mentioned in method Section 4.4. At the end of the regimen, mice were sacrificed and Mtb CFU from CD271+ BM-MSCs sorted cells were enumerated by plating on 7H11 agar plates supplemented with 10% Middlebrook OADC.

### *4.9. Enumeration of Mtb CFU from CD271*+*BM-MSCs*

The assay was performed as previously described [11–13]. BM cells were aseptically obtained from each group followed by magnetic sorting CD271+ BM-MSCs as described previously [13]. Briefly, an average 4 × 10<sup>7</sup> bone marrow cells were obtained from a femur pair per mice after RBC lysis. BM cells were then used for performing CD271+ BM-MSCs selection by magnetic sorting (Cat. Number S10467, Life Technologies, Carlsbad, CA, USA; Cat. Number 18554, Stem Cell Technologies, Vancouver, BC, Canada). Sorted cells were lysed followed by plating on 7H11 plates supplemented with 10% Middlebrook OADC.

### *4.10. TB Relapse Assay*

A similar experiment as described in the assessment of BTL-NPs E fficacy (Section 4.8) was carried out. NP treatment regimen was followed by steroid administration (glucocorticoid dexamethasone, intraperitoneal injections of 200 μL of dexamethasone at 10 mg/kg of body weight every 2 days for 4 weeks) and were sacrificed after 4 weeks to look for signs of reactivation. Signs for relapse were (a) granuloma formation in lungs by gross pathology and (b) culture-negative or culture-positive for lungs. Relapse percentage was calculated as follows:

> Relapse % = (No. of mice with culture-positive relapse/total No. of mice) × 100
