**2. Results**

### *2.1. Designing of Alendronate BTL-NPs and Their Characterization*

Alendronate tagging of PEG-PE was carried out for formulating Alen-PEG-PE as illustrated (Supplementary Figure S1). Resulting Alen-PEG-PE was used for the synthesis of BTL liposomes using additional PE-PEG and PC (Figure 1A,B). INH, RIF and Coumarin6 (fluorescent marker C6) were separately encapsulated in BTLs to formulate three separate BTLs: INH BTL, RIF BTL and C6 BTL followed by characterization (Figure 2). C6 BTLs were used for bone a ffinity experiments (Figure 3), cellular uptake (Figure 4) and tissue uptake studies (Figure 5C). Transmission electron microscopy results showed that all liposomal formulations consist of unilamellar vesicles in the size range of 80 to 110 nm having a spherical morphology. Dynamic light scattering data showed that the mean diameter of the prepared formulations, C6 BTLs, INH BTLS and RIF BTLs, were 109 ± 12.6 nm, 100 ± 16.3 nm and 84 ± 18.4 nm in diameter, respectively. Zeta potential values were negative which facilitates the long circulation of liposomes (Figure 2A–C). Average drug encapsulation e fficiency determined for C6 BTLs, INH-BTLs and RIF BTLs were 70% ± 3.7%, 69.5% ± 4.2% and 70.6% ± 4.7%, respectively (Table 1). Non-targeted liposomes exhibited similar size range and drug encapsulation e fficiency.

**Figure 1.** Designing of alendronate bone-homing PEGylated liposome (BTL) nanoparticles. (**A**) Alendronate and PE-PEG undergo chemical reactions for the binding of alendronate to form alendronate PE-PEG. Alendronate-PE-PEG is attached to the surface of liposomes. (**B**) Schematic of isoniazid (INH) and rifampicin (RIF) encapsulated alendronate PEG BTL-NPs. PC = Phosphatidyl-choline, CH2CH2N+(CH3)3; PE = Phosphatidylethanolamine, CH2CH2NH3<sup>+</sup>.

**Figure 2.** Characterization of liposomal formulations (TEM and size distribution of BTLs). Transmission electron microscopy (JEOL 2100F) images of all the three BTLs encapsulating C6, INH and RIF were performed. Dynamic light scattering and zeta potential were measured using Malvern Zetasizer. The above figure is a representative image for each liposomal formulation.


**Table 1.** Size and drug encapsulation efficiency of BTL-NPs.

### *2.2. Binding Assay of Alendronate BTL-NPs*

Alendronate has a natural affinity toward the calcium component of Hydroxyapatite (HA) crystal. To investigate the affinity of engineered BTLs for BM, bone-chip binding and HA binding experiments were performed. In order to visualize qualitative binding, C6 (a fluorescent marker)-loaded BTLs were used. Alen-tagged BTLs showed a high fluorescence signal (green), whereas the non-targeted liposomes did not show any binding (Figure 3A), proving the high affinity of BTLs for mice bone fragments (femur). In the second binding assay, HA microparticles were allowed to bind with alen BTLs and non-targeted liposomes. TEM imaging results showed that HA binds to alen BTLs with greater affinity while non-targeted liposomes did not show any affinity to HA microparticles (Figure 3B). Overall, these results confirmed that BTLs exhibited a strong affinity to bone HA and therefore may possess bone-homing capabilities.

**Figure 3.** (**A**) Bone-chip binding assay of BTL-NPs. Bone fragments from femur were washed and incubated with alendronate BTLs encapsulating Coumarin6 (alen BTL C6) and non-targeted liposomes C6 (NT liposomes C6) for 30 min followed by washing and visualization through a confocal microscope (scale bar = 0.1 mm). (**B**) BTL's affinity to Hydroxyapatite (HA)**.** For HA binding assay of BTL-NPs, HA microparticles were incubated with BTLs or NT liposomes. A 200 μL volume of 0.05 mg/mL HA was incubated with 40 μL of BTL or NT followed by rotation for 30 min at room temperature. The samples were washed 3 times gently and loaded on a TEM grid followed by staining and visualization with TEM (JEOL 2100F). Black rods are cylindrical HA microparticles and round nanoparticles around the HA are liposomes; scale bar = 100 nm HA + BTL (left panel), HA + NT (right panel).

### *2.3. Cellular Uptake of Alendronate BTL-NPs by CD271*+*BM-MSCs*

To assess the cellular uptake efficiency of BTLs, two different cells were investigated: macrophages (RAW 264.7 cell line murine macrophages) and freshly isolated murine CD271+BM-MSCs. Using live cell imaging, RAW 264.7 cell uptake was observed from 5 min to 30 min after the addition of BTL. Figure 4A indicates the successful uptake of BTLs by macrophages. Next, mice CD271+BM-MSCs were incubated with alen BTLs which exhibited similar uptake patterns showing localization inside the cells at 30 min (Figure 4B).

**Figure 4.** Confocal microscopy of C6BTLs uptake by CD271+BM-MSCs. Cells were incubated with C6BTL followed by live cell imaging to evaluate the time-bound uptake of BTLs by cells. (**A**) RAW 246.7 cells capture every 5 min up till 30 min after the addition of C6BTL. (**B**) CD271+BM-MSCs incubated with C6BTLs, fixed and imaged after 30 min.

### *2.4. In Vitro Drug Release Profile of INH and RIF BTL-NPs and Tissue Uptake*

In vitro drug release profile was enumerated by an in vitro dialysis method. The INH-encapsulated alen BTLs and RIF-encapsulated alen BTLs showed sustained slower release till 72 h in comparison to their counterpart non-targeted NPs as indicated by % cumulative drug release pattern (CDR%). Notedly, INH-BTLs showed faster release compared to RIF BTLs (Figure 5A,B). Nevertheless, both targeted NP's cumulative drug release was slow and sustained. Further, we investigated the bone uptake of BTLs, where mice were injected with C6 loaded alen BTLs, C6 non-targeted (NT) liposomes or C6 PBS. Pairs of femur and tibia were excised followed by C6 extraction with chloroform: methanol. Fluorimetric analysis was carried out for determining the percentage of C6 dose injected, recovered from both bone pairs. We found that there was an approximate eight times increase in the homing of C6BTLs in comparison to C6NT liposomes (Figure 5C) demonstrating the bone-seeking nature of BTL-NPs.

**Figure 5.** (**A**) Schematic of in vitro drug release. (**B**) Cumulative drug release pattern (CDR%) of targeted (BTLs) versus non-targeted (NT) for INH and RIF. A 500 μL volume of INH- and RIF-encapsulated BTLs and NTs were suspended in dialysis bags and stirred for 72 h in 25 mL of 1× PBS (pH 7.4). At each time point, 500 μL was aliquoted from a 25 mL reservoir for drug release assay. Data represents mean CDR% ± SEM (*n* = 3). (**C**) Tissue uptake of targeted (C6 BTL) versus non-targeted (C6 NT). Data are presented as % of C6 injected recovered per g of organ (organ = pair of tibia and femur). SEM (*n* = 3), \*\* *p* < 0.001 vs. C6NT.

### *2.5. <sup>E</sup>*ffi*cacy of INH and RIF BTL-NPs in H37Rv Infected Mice*

For evaluating the efficacy of INH and RIF alendronate BTLs, standard Cornell model of persistent Mtb was followed (Figure 6A): Mice infected with pathogenic strains of Mtb H37RV were administered standard INH and RIF dosage for 90 days as a result of which all organs except for bone marrow were free of bacilli load [11,13]. After one week of the above regimen, animals were treated with the drug-loaded BTLs/NT NPs for 12 weeks, twice per week, with an effective concentration of INH = 4 mg/kg and RIF = 3 mg/kg. Appropriate controls were used. Animal health and weight were monitored for four months followed by CFU enumeration of CD271+BM-MSCs. A significant reduction of Mtb CFU in CD271+BM-MSCs was observed (Figure 6B). Notedly, Mtb culture from lungs was found negative in all experimental groups.

Relapse studies were carried out on Mtb-infected mice as illustrated in Figure 7A. Mice were monitored up to four months for relapse post-dexamethasone administration (10 mg/kg/ four weeks). Gross pathology inspection of the lung lobes was carried out for granuloma formation and the remaining portion of the lung was homogenized and cultured to score positive or negative Mtb load (Figure 7). Appropriate controls were included for relapse study viz. only vehicle and empty BTLs. All untreated mice succumbed to death within two months of infection. The relapse rate was 11% for the BTL group while the other experimental groups showed a higher relapse rate (Figure 7).

**Figure 6.** *Mycobacterium tuberculosis* (Mtb) load from CD271+BM-MSCs post-INH and RIF BTLs treatment. (**A**) Schematic of in vivo treatment. (**B**) C57bl/6 mice were infected with H37Rv, administered with standard anti-Tb drugs and next administered with drug-loaded BTL or NTs for 12 weeks/twice a week/intravenously. The effective dosage of INH and RIF was maintained at 4 mg/kg and 3 mg/kg, respectively, for all drug groups. Mtb CFU were enumerated by plating CD271+ BM-MSCs population obtained from immuno-magnetic sorting. *n* = 3, data expressed as mean CFU = +/− SEM per 10<sup>7</sup> cells. \*\* *p* < 0.001.

**Figure 7.** Relapse rate for INH and RIF BTLs versus NT liposomes. (**A**) Schematic of in vivo treatment for relapse. (**B**) Relapse % for targeted liposomes. Mice from each group were monitored for 4 months for signs of relapse. Animals were sacrificed for relapse signs: positive culture from lungs and gross lung pathology/granuloma formation. Data expressed as mean = +/− SEM. \*\* *p* < 0.001.
