Enhancing Anticancer Efficacy of Chemotherapeutics Using Targeting Ligand-Functionalized Synthetic Antigen Receptor-Mesenchymal Stem Cells

Mesenchymal stem cells (MSCs) have been studied for their potential in facilitating tumor-targeted delivery of chemotherapeutics due to their tumor-homing characteristics. We hypothesized that targeting effectiveness of MSCs can be further enhanced by incorporating tumor-targeting ligands on MSC surfaces that will allow for enhanced arrest and binding within the tumor tissue. We utilized a unique strategy of modifying MSCs with synthetic antigen receptors (SARs), targeting specific antigens overexpressed on cancer cells. MSCs were surface-functionalized by first incorporating recombinant protein G (PG) on the surface, followed by binding of the targeting antibody to the PG handle. We functionalized MSCs with antibodies targeting a tyrosine kinase transmembrane receptor protein, epidermal growth factor receptor (EGFR), overexpressed in non-small-cell lung cancer (NSCLC). The efficacy of MSCs functionalized with anti-EGFR antibodies (cetuximab and D8) was determined in murine models of NSCLC. Cetuximab-functionalized MSCs demonstrated improved binding to EGFR protein and to EGFR overexpressing A549 lung adenocarcinoma cells. Further, cetuximab-functionalized MSCs loaded with paclitaxel nanoparticles were efficient in slowing orthotopic A549 tumor growth and improving the overall survival relative to that of other controls. Biodistribution studies revealed a six-fold higher retention of EGFR-targeted MSCs than non-targeted MSCs. Based on these results, we conclude that targeting ligand functionalization could be used to enhance the concentration of therapeutic MSC constructs at the tumor tissue and to achieve improved antitumor response.


Introduction
Tumors exhibit uneven vascular perfusion; the peripheral or rim regions often have nearly normal blood flow while the cells in the core regions of the tumor are often exposed to hypoxic (low-oxygen) conditions [1][2][3]. Tumors also face challenges related to elevated interstitial fluid pressure and a rigid extracellular matrix, which can hinder the transport of solutes within the tumor microenvironment [4]. These pathophysiologic characteristics limit drug delivery and hinder the effectiveness of treatments [5][6][7]. Cell-based drug delivery systems have shown promise in overcoming the critical tissue barriers presented by elevated interstitial fluid pressure and a rigid extracellular matrix within tumors. These was continued for 24 h at 37 • C. Palmitic-acid-derivatized protein G (PA-PG) was purified on a Sephadex G-25 column to remove the unreacted lipid, as reported previously [27].

Characterization of PA-PG
Purified PA-PG was characterized using ultra-performance liquid chromatography-mass spectrometry (UPLC-MS) to determine the derivatization ratio (number of PA molecules per PG molecule). Liquid chromatography-mass spectrometry (LC-MS) analysis using a Waters Acquity UPLC system coupled to a Waters Synapt G2 HDMS quadrupole orthogonal acceleration time of flight (Q-TOF) mass spectrometer equipped with an LC Column (Waters Acquity UPLC Protein BEH C4, 2.1 mm × 100 mm; particles diameter: 1.7 µm, 35 • C) was performed as described previously [28]. The system was run at linear gradient elution using mobile phase (Phase A: Water containing 0.1% formic acid and Phase B: Acetonitrile containing 0.1% formic acid) at a flow rate of 0.4 mL/min. One mass spectrum (0.2 s scan, m/z 50-1200) was acquired every 10 s. Three lockspray m/z measurements of protonated leucine enkephalin were averaged and mass correction applied to the measured m/z values.

Cell Culture
MSCs, A549-luc cells, and LL/2-luc cells were grown in MSC complete medium, RPMI 1640 medium. and Dulbecco's modified Eagle's medium (DMEM) containing 10% v/v FBS and 1% v/v penicillin-streptomycin antibiotic solution, respectively. All the cell lines were maintained in a controlled humidity chamber (5% carbon dioxide; 37 • C). All our cell culture studies were performed under these conditions, unless specified otherwise.

Optimization of PA-PG Incorporation on MSC Membrane 2.4.1. Effect of PA-PG Concentration on Cell Viability
MSCs (8 × 10 3 cells per well) were incubated with different concentrations of PA-PG-containing media (10,20,50, and 100 µg/mL) for 24 h and the viability of cells was analyzed using MTS assay. The PA-PG-containing media was replaced with MTS/PMS solution in media and incubated for 1 h at 37 • C. The absorbance was recorded at 490 nm using a microplate reader (Spectramax, Molecular Devices, San Jose, CA, USA). Untreated MSCs and MSCs incubated with PG (50 and 100 µg/mL) served as controls.

PA-PG Coating on MSC Surface
Flow cytometry was used to confirm the binding of Ab to PA-PG-coated MSCs. MSCs in culture were harvested via trypsinization and counted. About 2.5 × 10 5 cells were incubated with PA-PG (20-100 µg/mL) in a 1.5 mL Eppendorf tube for 1 h at 37 • C under gentle rotation. The cells were washed twice with DPBS and collected using centrifugation at 1000 rpm for 5 min. The cells were then incubated with Alexa Flour 647-labeled IgG (AF647-IgG) for 1 h at 4 • C under gentle rotation. Following incubation, the cells were washed with DPBS by centrifugation as before and the final pellet was resuspended in flow buffer. Binding of the Ab to MSCs was determined by measuring the mean fluorescence intensity of AF647-IgG-labeled MSCs using a BD LSR II flow cytometer.
The incubation time and temperature of the PA-PG coating reaction were optimized by treating MSCs with PA-PG for different durations (0.5, 1, 2, and 3 h) and at different temperatures (4 • C and 37 • C), followed by multiple washings using DPBS. Cells were then incubated with AF647-IgG for 1 h as described above. Ab binding to MSCs was determined via flow cytometry as above. Cmab and an isotype control IgG were labeled with IR800 dye (LI-COR Biosciences, Lincoln, NE, USA) using established procedures. MSCs were incubated with PA-PG (50 µg/mL) in serum-free media for 1 h at 37 • C. Labeled Ab (either Cmab or isotype Ig) was added to PA-PG-coated MSCs and incubated for 1 h at 4 • C. Post-incubation, the cells were washed with DPBS, resuspended in flow buffer, and analyzed for IR800 fluorescence intensity via flow cytometry.

Quantitative Analysis of Ab Binding to PA-PG-MSCs
The amount of Ab bound to the surface of PA-PG-coated MSCs was determined using a standard curve constructed using known concentrations of IgG-AF647 Ab (0.01, 0.1, 0.25, 0.5, 1, and 10 µg/mL). AF647 IgG-labeled PA-PG-MSCs were prepared as described above. Labeled MSCs (1 × 10 6 ) were then suspended in PBS or a mixture of RIPA buffer and PBS. Fluorescence intensity of the cell suspension was measured using a plate reader (Spectramax), and the concentration of IgG bound to MSCs was calculated from the standard plot.

Qualitative Analysis of Ab Binding to PA-PG-MSCs via Confocal Microscopy
MSCs stably expressing green fluorescent protein were seeded on pre-sterilized 100 mm coverslips in a 6-well plate (50,000 cells/well). One day later, MSCs were incubated with PA-PG (50 µg/mL) in serum-free media for 1 h at 37 • C. The cells were then washed twice with DPBS and then incubated with AF647-labeled Ab (100 µg/mL) for 1 h at 4 • C. The cells were then washed, labeled with Hoechst 33342 dye (to stain the nuclei), and fixed using 4% paraformaldehyde for 15 min at room temperature. The cells were then washed, and the coverslips were mounted on microscopy slides using mounting media (Thermo Fisher Scientific). The slides were then imaged for GFP (MSC), Hoechst (nuclei), and AF647 (Ab) fluorescence using a confocal microscope (Leica, Wetzlar, Germany).
2.6. In Vitro Binding of Cmab-MSCs to Its Target 2.6.1. Binding to Immobilized EGFR Recombinant EGFR protein, dissolved in DPBS (10 µg/mL), was added to 96-well plates, incubated for 2 h at 37 • C, and then washed thrice with DPBS to remove any unbound protein. Non-specific interactions were blocked by incubating the plates with 1% bovine serum albumin in DPBS for 30 min at 37 • C. Unmodified MSCs and Cmab-MSCs were first incubated with calcein AM solution (2 µg/mL) for 1 h and then added to EGFRcoated wells (10,000 cells per well) and incubated for 30 min at 37 • C. The wells were then washed thrice with DPBS to remove unbound and loosely bound cells. The fluorescence intensity of EGFR-bound cells was measured using Spectramax at excitation and emission wavelengths of 485 and 520 nm, respectively.

Binding to A549 Cells
Binding of Cmab-MSCs to A549 cells was evaluated using flow cytometry. First, 3 × 10 5 A549 cells/well were cultured in multi-well plate overnight followed by incubation with Hoechst dye for 10 min in the dark at 25 • C and washed with DPBS. PA-PG-coated MSCs, control IgG-MSCs, and Cmab-MSCs were first stained with the cell membrane labeling dye CellVue 780. Labeled MSCs (1 × 10 5 cells/well) were incubated with A549 cells for 30 min at 37 • C. The plates were washed thrice to remove unbound cells. A549 cells, along with the MSCs attached to them, were harvested by trypsinization, washed, and resuspended in a flow buffer. The cells were then analyzed using flow cytometry (BD LSR-II) to determine doublet cells positive for both CellVue780 (ex:744 nm/em:771 nm) and Hoechst (ex:361 nm/em:497 nm) dyes. Paclitaxel-loaded PLGA nanoparticles were prepared via the emulsion-solvent evaporation method as reported previously [29]. PLGA (32 mg) and paclitaxel (8 mg) were dissolved in 1 mL of chloroform. The oil-in-water emulsion was obtained by mixing the polymer-drug solution with 8 mL of PVA solution (2.5% w/v) using a probe sonicator (Model W-375, Heat Systems Ultrasonics Inc., Farmingdale, NY, USA) for 5 min at 20 W and stirring overnight at room temperature. The resultant dispersion was placed under vacuum for 1 h to remove chloroform. The PLGA nanoparticles were collected via ultracentrifugation (35,000 rpm; 35 min at 4 • C) (Optima XPN-80 Ultracentrifuge, Rotor type: 50.2 Ti, Beckman Coulter, Brea, CA, USA). The nanoparticles were washed two times using distilled water to remove unentrapped drug and PVA. Final nanoparticles dispersion was centrifuged at 1000 rpm for 5 min to remove aggregates. The dried lyophilized (Labconco, FreeZone 4.5, Kansas City, MO, USA) nanoparticles were stored at −20 • C.
The nanoparticles were characterized using a Delsa Nano C particle analyzer (Beckman Coulter, CA, USA). Nanoparticles dispersion (n = 3) in deionized water (0.1 mg/mL) was used for particle size analysis.
To determine paclitaxel loading, the drug was extracted from methanolic dispersion of nanoparticles (1 mg/mL) overnight using a rotary shaker under gentle agitation and analyzed using reverse-phase high-performance liquid chromatography (HPLC) as reported previously [16]. Drug loading and entrapment efficiency were determined using the following equations: Entrapment efficiency (%) = Drug amount in nanoparticles Total amount of drug × 100 Drug loading (%) = Drug amount in nanoparticles Amount of drug loaded nanoparticles × 100

Preparation of Paclitaxel-Loaded Cmab-MSCs
MSCs (2.5 × 10 5 cells/mL) were incubated with paclitaxel nanoparticles (0.1 mg/mL) dispersion for 4 h at 37 • C under periodic shaking followed by incorporation of either Cmab or isotype control IgG on the surface as described previously. MTS assay was used to study the effect of paclitaxel loading on MSC viability. To visualize nanoparticle-loaded, Cmab-functionalized MSCs, MSCs were incubated with 6-coumarin-loaded nanoparticles followed by Cmab functionalization. Cells were imaged using a confocal microscope to visualize green (6-coumarin nanoparticles) and red (AF647-labeled Cmab) fluorescence.

Impact of Paclitaxel Loading on In Vitro Migration
In vitro migration of paclitaxel-loaded Cmab-MSCs was determined using a trans-well assay. Untreated MSCs, Cmab-MSCs, paclitaxel-loaded MSCs, and paclitaxel-loaded IgG-MSCs were used as controls. MSCs were serum-starved for 24 h prior to the assay. Migration towards either serum-free media, 5% serum-containing media, or tumor-conditioned media (5% serum-containing media with A549 cells for 24 h) was assessed at 37 • C for 20 h. To quantitate the cell migration, cells in the bottom wells were stained using Calcein AM and the percentage of cells migrated was determined using a fluorescence plate reader (Spectramax).

Tumor Retention
LL/2-luc lung tumor model was used to determine tumor retention of EGFR-targeted MSCs manufactured using murine anti-EGFR monoclonal antibody (D8). Mice were randomized when tumor bioluminescence reached 5 × 10 6 photons/s into two groups and injected with 1 million unmodified MSCs or D8-MSCs. Mice were euthanized at 1-week and 2-week intervals. Lung tumor tissues were preserved in RNA-later solution at −20 • C. Further, the stored lung tumors were homogenized using a tissue homogenizer and the total RNA extracted was reverse-transcribed using the cDNA Reverse Transcription Kit (Applied Biosystems, Waltham, MA, USA). Primers targeting THY1 (Hs06633377_s1) and RPLP0 (Hs02786624_g1) were used for qPCR analysis. Real-time PCR was run in quadruplicate using a QuantStudio 12K Flex detection system (Applied Biosystems, Waltham, MA, USA) according to the recommended cycle as described previously [30].

Anticancer Efficacy of Paclitaxel-Loaded EGFR-Targeted MSCs
An initial pilot study was performed to determine the optimal dosing regimen for Cmab-MSCs. Mice bearing A549 lung tumors (~5 × 10 6 photons/s bioluminescence) were divided randomly into two groups and injected intravenously with 1 × 10 6 cells/animal with either paclitaxel-loaded IgG-MSCs or paclitaxel-loaded Cmab-MSCs suspended in 200 µL DPBS, once every two weeks.
Following the pilot study, we determined the comparative therapeutic efficacy of paclitaxel-loaded Cmab-MSCs relative to that of paclitaxel-loaded, unfunctionalized MSCs. Saline, Cmab (i.p. administration; 1.1 µg in 200 µL DPBS/mouse), and paclitaxel-loaded IgG-MSCs were used as additional controls. All the treatments were administered biweekly, and bioluminescence measurements were performed to follow tumor progression.

Statistical Analysis
All the data are presented as mean ± standard deviation or mean ± standard error of mean. Population distribution was tested for normality using D'Agostino and Pearson test and Kolmogorov-Smirnov test (GraphPad Prism 8). The significance (p < 0.05) of observed differences between different groups was determined using Student's t-test or ANOVA. Multiple comparisons between individual treatments was performed using Tukey's method. Survival analysis among different treatment groups was performed using Log-rank test.

Results and Discussions
MSCs are a type of adult stem cell that can be isolated from various tissues, including bone marrow, adipose tissue, and umbilical cord tissue [31,32]. MSCs have unlimited potential to differentiate into different cell types depending upon the origin [33]. MSCs play a key role in tissue repair and regeneration [34]. MSCs respond to tissue injury and the consequent release of inflammatory signals by actively migrating to the site of tissue injury [35]. Previous studies show that SDF1-CXCR4/CXCR7 signaling plays a key role in mediating MSC chemotaxis [36]. Tumors co-opt this signaling pathway, resulting in active recruitment of MSCs to the tumor microenvironment, where they can promote immunosuppression [37].
This tumor-homing property of MSCs has been leveraged by several groups to deliver various therapeutic agents to the tumor [35]. MSCs stably transduced with viral and non-viral vectors have been used to deliver cytokines [38], suicidal genes [39][40][41], and antibodies [42] to the tumor. Several investigators, including our group, have engineered MSCs to deliver cytotoxic drugs to the tumor tissue [14][15][16]18,29,[43][44][45]. These studies show that the extent of tumor homing and retention of MSCs is a critical determinant of their therapeutic effectiveness.
In the present study, we hypothesized that the tumor-targeting effectiveness of MSCs can be significantly improved by incorporating tumor-targeting ligands on the MSC surface that will allow for enhanced arrest and binding of MSCs within the tumor tissue. We used EGFR as a model tumor antigen because EGFR overexpression has been reported in  [23,46]. Two different Abs, one specific to human EGFR (Cmab) and another that recognizes mouse EGFR (D8), were used as targeting ligands.
PG was used to incorporate the targeting Ab on the surface of MSCs because PG binds to the Fc region of Ab and does not impact the antigen-binding complementaritydetermining regions (CDRs). To localize PG on the cell membrane, PG was first derivatized with lipophilic PA. Insertion of PA into the cell membrane allows for extracellular localization of PG, which is then available to bind with the Ab [47]. We characterized PA-PG derivatives via LC/QTOF/MS. De-convolution of mass data showed successful conjugation of PA to PG and that up to ten molecules of PA can be added to each molecule of PG ( Figure 1A-E).
MSCs to deliver cytotoxic drugs to the tumor tissue [14][15][16]18,29,[43][44][45]. These studies show that the extent of tumor homing and retention of MSCs is a critical determinant of their therapeutic effectiveness.
In the present study, we hypothesized that the tumor-targeting effectiveness of MSCs can be significantly improved by incorporating tumor-targeting ligands on the MSC surface that will allow for enhanced arrest and binding of MSCs within the tumor tissue. We used EGFR as a model tumor antigen because EGFR overexpression has been reported in NSCLC [23,46]. Two different Abs, one specific to human EGFR (Cmab) and another that recognizes mouse EGFR (D8), were used as targeting ligands.
PG was used to incorporate the targeting Ab on the surface of MSCs because PG binds to the Fc region of Ab and does not impact the antigen-binding complementaritydetermining regions (CDRs). To localize PG on the cell membrane, PG was first derivatized with lipophilic PA. Insertion of PA into the cell membrane allows for extracellular localization of PG, which is then available to bind with the Ab [47]. We characterized PA-PG derivatives via LC/QTOF/MS. De-convolution of mass data showed successful conjugation of PA to PG and that up to ten molecules of PA can be added to each molecule of PG ( Figure 1A-E). The effect of PA-PG incorporation on MSC viability was evaluated. Both PG and PA-PG had a minimal effect on cell viability at up to 50 µg/mL concentration ( Figure 2A). We first optimized the concentration of PA-PG to successfully incorporate both PA-PG and the targeting Ab on MSC surface. PG (100 µg/mL) alone was used as a negative control. As shown in Figure 2A,B, the number of Ab molecules on the MSC surface increased with increasing concentration of PA-PG. Beyond 100 µg/mL concentration, PA-PG caused a significant loss of cell viability. We also evaluated the effect of time and temperature of the incubation reaction on PA-PG incorporation. As shown in Figure 2C, PA-PG The effect of PA-PG incorporation on MSC viability was evaluated. Both PG and PA-PG had a minimal effect on cell viability at up to 50 µg/mL concentration ( Figure 2A). We first optimized the concentration of PA-PG to successfully incorporate both PA-PG and the targeting Ab on MSC surface. PG (100 µg/mL) alone was used as a negative control. As shown in Figure 2A,B, the number of Ab molecules on the MSC surface increased with increasing concentration of PA-PG. Beyond 100 µg/mL concentration, PA-PG caused a significant loss of cell viability. We also evaluated the effect of time and temperature of the incubation reaction on PA-PG incorporation. As shown in Figure 2C, PA-PG incorporation increased with time and plateaued off beyond 2 h. There was not a significant effect of temperature on PA-PG incorporation. The lack of any effect of temperature suggested that PA-PG was bound physically to the cell surface and was not dependent on cellular energy.
The incorporation of the targeting Ab on MSCs was confirmed via flow cytometry. Figure 3A shows flow histograms following the incorporation of fluorescently labeled non-specific IgG, as well as anti-EGFR Ab, on PA-PG-coated MSCs compared to PG-coated MSCs and untreated MSCs. As shown in Figure 3B, an increase in fluorescence was observed for PA-PG-coated MSCs compared to the controls, demonstrating successful incorporation of PA-PG handle and its ability to bind with labeled Ab. There was no significant difference in the amount of Ab incorporated between the isotype IgG control or Cmab groups. We used confocal microscopy to confirm the presence of Ab on the surface of MSCs. As shown in Figure 4, the presence of labeled Ab (red fluorescence) could be clearly visualized on the surface of MSCs. We also determined the concentration of Ab bound to PA-PG-coated MSCs using fluorescence spectroscopy. These studies showed that there was 1.75 µg Ab/10 6 MSCs (~7 × 10 6 Ab molecules/cell). incorporation increased with time and plateaued off beyond 2 h. There was not a significant effect of temperature on PA-PG incorporation. The lack of any effect of temperature suggested that PA-PG was bound physically to the cell surface and was not dependent on cellular energy. The incorporation of the targeting Ab on MSCs was confirmed via flow cytometry. Figure 3A shows flow histograms following the incorporation of fluorescently labeled non-specific IgG, as well as anti-EGFR Ab, on PA-PG-coated MSCs compared to PGcoated MSCs and untreated MSCs. As shown in Figure 3B, an increase in fluorescence was observed for PA-PG-coated MSCs compared to the controls, demonstrating successful incorporation of PA-PG handle and its ability to bind with labeled Ab. There was no significant difference in the amount of Ab incorporated between the isotype IgG control or Cmab groups. We used confocal microscopy to confirm the presence of Ab on the surface of MSCs. As shown in Figure 4, the presence of labeled Ab (red fluorescence) could be clearly visualized on the surface of MSCs. We also determined the concentration of Ab bound to PA-PG-coated MSCs using fluorescence spectroscopy. These studies showed that there was ~1.75 µg Ab/10 6 MSCs (~7 × 10 6 Ab molecules/cell).   The incorporation of the targeting Ab on MSCs was confirmed via flow cytometry. Figure 3A shows flow histograms following the incorporation of fluorescently labeled non-specific IgG, as well as anti-EGFR Ab, on PA-PG-coated MSCs compared to PGcoated MSCs and untreated MSCs. As shown in Figure 3B, an increase in fluorescence was observed for PA-PG-coated MSCs compared to the controls, demonstrating successful incorporation of PA-PG handle and its ability to bind with labeled Ab. There was no significant difference in the amount of Ab incorporated between the isotype IgG control or Cmab groups. We used confocal microscopy to confirm the presence of Ab on the surface of MSCs. As shown in Figure 4, the presence of labeled Ab (red fluorescence) could be clearly visualized on the surface of MSCs. We also determined the concentration of Ab bound to PA-PG-coated MSCs using fluorescence spectroscopy. These studies showed that there was ~1.75 µg Ab/10 6 MSCs (~7 × 10 6 Ab molecules/cell).  We then investigated the binding of Cmab-functionalized MSCs to their target. We initially evaluated the binding of MSCs to immobilized EGFR protein. As can be seen from Figure 5A, a greater number of Cmab-MSCs (as determined by calcein AM fluorescence) adhered to EGFR-protein-coated wells than non-targeted MSCs. We then investigated the binding of MSCs to EGFR-expressing A549 cells. While non-targeted MSCs (MSCs, MSCs coated with PA-PG, and MSCs incorporating non-specific control IgG) demonstrated some binding to A549 cells, Cmab-MSCs resulted in significantly greater binding to A549 cells ( Figure 5B). These results provide evidence for the ability of targeting Ab incorporated on the surface of MSCs to recognize their intended target. PG binds all human subclasses of IgG with high affinity (K D~2 × 10 −10 M) [48]. The K D for Cmab binding to EGFR, on the other hand, is 3.8-11.0 × 10 −10 M [49]. Cmab affinity for the protein expressed on the surface of cells is likely lower than for pure protein.  We then investigated the binding of Cmab-functionalized MSCs to their target. We initially evaluated the binding of MSCs to immobilized EGFR protein. As can be seen from Figure 5A, a greater number of Cmab-MSCs (as determined by calcein AM fluorescence) adhered to EGFR-protein-coated wells than non-targeted MSCs. We then investigated the binding of MSCs to EGFR-expressing A549 cells. While non-targeted MSCs (MSCs, MSCs coated with PA-PG, and MSCs incorporating non-specific control IgG) demonstrated some binding to A549 cells, Cmab-MSCs resulted in significantly greater binding to A549 cells ( Figure 5B). These results provide evidence for the ability of targeting Ab incorporated on the surface of MSCs to recognize their intended target. PG binds all human subclasses of IgG with high affinity (KD ~2 × 10 −10 M) [48]. The KD for Cmab binding to EGFR, on the other hand, is 3.8-11.0 × 10 −10 M [49]. Cmab affinity for the protein expressed on the surface of cells is likely lower than for pure protein. The 2-3-fold (if not higher) difference in binding affinity for PG compared to that for EGFR could explain how MSCs retain the targeting Ab on the surface even in the presence of the target protein. We next evaluated loading of paclitaxel in Cmab-MSCs. In previous studies, we demonstrated different approaches to stably load paclitaxel in MSCs [14,15,18,29]. These included simple endocytic loading by incubating MSCs with paclitaxel-loaded polymeric nanoparticles [29] or conjugating paclitaxel-loaded polymeric nanoparticles to the surface of MSCs [18]. Although the surface conjugation technique results in greater drug loading in MSCs, this approach could interfere with the incorporation of Cmab on the cell membrane.
Hence, in the current study, we loaded the drug via simple incubation of MSCs with paclitaxel-loaded nanoparticles.
Paclitaxel-loaded PLGA nanoparticles were prepared with a mean hydrodynamic diameter of~350 nm with a negative zeta potential and drug loading of 20.3 ± 1.5% (w/w) and an entrapment efficiency of 76.2 ± 5.4%. As we reported previously, loading paclitaxel did not significantly affect the viability ( Figure 6A) or migration ( Figure 6B) of Cmab-MSCs when compared to non-drug-loaded MSCs and non-drug-loaded Cmab-MSCs, or drug-loaded non-targeted MSCs. We recently demonstrated that MSCs upregulate key antioxidant proteins such as Nrf2, superoxide dismutase, and heme oxygenase-1 in response to paclitaxel loading, which may explain how MSCs are resistant to high concentrations of paclitaxel. Other studies have shown that MSCs overexpress drug efflux transporters such as p-glycoprotein [29,50] and upregulate anti-apoptotic proteins [51], which may also contribute to drug resistance. We next evaluated loading of paclitaxel in Cmab-MSCs. In previous studies, w demonstrated different approaches to stably load paclitaxel in MSCs [14,15,18,29]. The included simple endocytic loading by incubating MSCs with paclitaxel-loaded polymer nanoparticles [29] or conjugating paclitaxel-loaded polymeric nanoparticles to the surfa of MSCs [18]. Although the surface conjugation technique results in greater drug loadin in MSCs, this approach could interfere with the incorporation of Cmab on the cell mem brane. Hence, in the current study, we loaded the drug via simple incubation of MSC with paclitaxel-loaded nanoparticles.
Paclitaxel-loaded PLGA nanoparticles were prepared with a mean hydrodynamic d ameter of ~350 nm with a negative zeta potential and drug loading of 20.3 ± 1.5% (w/w and an entrapment efficiency of 76.2 ± 5.4%. As we reported previously, loading paclitax did not significantly affect the viability ( Figure 6A) or migration ( Figure 6B) of Cma MSCs when compared to non-drug-loaded MSCs and non-drug-loaded Cmab-MSCs, drug-loaded non-targeted MSCs. We recently demonstrated that MSCs upregulate k antioxidant proteins such as Nrf2, superoxide dismutase, and heme oxygenase-1 in r sponse to paclitaxel loading, which may explain how MSCs are resistant to hig Figure 5. In vitro binding of EGFR-targeted MSCs to immobilize EGFR protein and EGFR-expressing A549 lung tumor cells. Binding of calcein AM stained unmodified MSCs and Cmab-MSCs to EGFR protein was measured via fluorescence spectroscopy (A). Flow cytometry was used to determine binding of AF647−labeled MSCs to Hoechst-stained A549 cells. Mean fluorescence intensity for the different groups is shown in (B). Individual dot plots are shown in (C). The color scale represents the density of events: blue (low) to red (high). **** p < 0.0001. We next confirmed that loading paclitaxel in MSCs does not affect the surface localization of targeting Ab. Nanoparticles were labeled with green-fluorescent 6-coumarin while the Ab was labeled with red-fluorescent AF647. Confocal microscopy images demonstrated that Ab-associated red fluorescence was present on the MSC surface, while nanoparticleassociated green fluorescence was intracellular ( Figure 6C-E). The distribution pattern of red fluorescence seen in this study was qualitatively similar to that seen in the case of non-drug-loaded MSCs (Figure 4), suggesting that drug loading does not affect Ab localization on MSC surface.  [29,50] and upregulate anti-apoptotic protein which may also contribute to drug resistance. We next confirmed that loading paclitaxel in MSCs does not affect the surface zation of targeting Ab. Nanoparticles were labeled with green-fluorescent 6-cou while the Ab was labeled with red-fluorescent AF647. Confocal microscopy i demonstrated that Ab-associated red fluorescence was present on the MSC surface nanoparticle-associated green fluorescence was intracellular ( Figure 6C-E). The di tion pattern of red fluorescence seen in this study was qualitatively similar to that the case of non-drug-loaded MSCs (Figure 4), suggesting that drug loading does no Ab localization on MSC surface.
We previously reported tumor-targeted delivery of paclitaxel using MSCs to topic lung and ovarian tumors, resulting in improved therapeutic efficacy [15,16,18 the current study, we first investigated the comparative therapeutic efficacy of pac loaded Cmab-MSCs relative to that of paclitaxel-loaded non-targeted IgG-MSCs. be seen from Figure 7B, mice that received Cmab-MSCs demonstrated significantly tumor growth than those that received control IgG-MSCs. These initial results d strated the potential for improved tumor targeting with Cmab-MSCs. Based on promising results, a larger study was conducted to include additional controls Cmab has therapeutic efficacy as a single agent [52,53], we included Cmab alone as trol. As observed in our previous studies, non-targeted MSCs loaded with pac We previously reported tumor-targeted delivery of paclitaxel using MSCs to orthotopic lung and ovarian tumors, resulting in improved therapeutic efficacy [15,16,18,29]. In the current study, we first investigated the comparative therapeutic efficacy of paclitaxel-loaded Cmab-MSCs relative to that of paclitaxel-loaded non-targeted IgG-MSCs. As can be seen from Figure 7B, mice that received Cmab-MSCs demonstrated significantly slower tumor growth than those that received control IgG-MSCs. These initial results demonstrated the potential for improved tumor targeting with Cmab-MSCs. Based on these promising results, a larger study was conducted to include additional controls. Since Cmab has therapeutic efficacy as a single agent [52,53], we included Cmab alone as a control. As observed in our previous studies, non-targeted MSCs loaded with paclitaxel nanoparticles were efficacious in slowing tumor growth and improving the overall survival relative to that of other controls ( Figure 7C). However, the greatest tumor growth inhibition was observed in mice that received paclitaxel-loaded Cmab-MSCs. Compared to a median survival of 21 days for the saline-treated group and 27 days for the paclitaxel-loaded nonfunctionalized MSC-treated group, paclitaxel-loaded Cmab-MSC treatment resulted in a median survival of 35 days (p < 0.0001) ( Figure 7E). nanoparticles were efficacious in slowing tumor growth and improving the overall survival relative to that of other controls ( Figure 7C). However, the greatest tumor growth inhibition was observed in mice that received paclitaxel-loaded Cmab-MSCs. Compared to a median survival of 21 days for the saline-treated group and 27 days for the paclitaxelloaded non-functionalized MSC-treated group, paclitaxel-loaded Cmab-MSC treatment resulted in a median survival of 35 days (p < 0.0001) ( Figure 7E).  We next evaluated the tumor-targeting effectiveness of EGFR-targeted MSCs. To fully understand the impact of a functional immune system on in vivo disposition of EGFRtargeted MSCs, the syngeneic, LL/2 lung tumor model was used. Tumors were induced by i.v. injection of LL/2-luc cells in immunocompetent C57/Bl6J mice. Further, since the goal was to evaluate the tumor accumulation of MSCs, we used RT-PCR to quantify human CD90 mRNA, which is specific to human MSCs and can be used to specifically and accurately quantify the MSC levels in the mouse tissue [54,55]. It has been previously shown that owing to the differential vascularity of organs, intravenously administered MSCs in healthy mice first travel to the lung, followed by the liver and spleen. However, the migration pattern and final residence could be different in the presence of different tumors and sites of inflammation [33,56]. There was no difference in MSC levels between nontargeted and targeted MSCs one day post-injection. However, there was a six-fold higher level of EGFR-targeted MSCs compared to non-targeted MSCs at 7 days post-injection ( Figure 8). These results support the hypothesis that ligand functionalization will enhance MSC accumulation and retention in the tumor tissue.
We next evaluated the tumor-targeting effectiveness of EGFR-targeted MS fully understand the impact of a functional immune system on in vivo dispos EGFR-targeted MSCs, the syngeneic, LL/2 lung tumor model was used. Tumors w duced by i.v. injection of LL/2-luc cells in immunocompetent C57/Bl6J mice. Furthe the goal was to evaluate the tumor accumulation of MSCs, we used RT-PCR to q human CD90 mRNA, which is specific to human MSCs and can be used to spe and accurately quantify the MSC levels in the mouse tissue [54,55]. It has been pre shown that owing to the differential vascularity of organs, intravenously admin MSCs in healthy mice first travel to the lung, followed by the liver and spleen. Ho the migration pattern and final residence could be different in the presence of d tumors and sites of inflammation [33,56]. There was no difference in MSC levels b non-targeted and targeted MSCs one day post-injection. However, there was a higher level of EGFR-targeted MSCs compared to non-targeted MSCs at 7 days jection ( Figure 8). These results support the hypothesis that ligand functionalizat enhance MSC accumulation and retention in the tumor tissue. Tumor homing of MSCs involves tethering of MSCs (mediated by CD44) to s expressed on endothelial cells followed by rolling on the vascular wall [57]. G-p coupled chemokine receptors (especially CXCR 4 and 7) activate MSCs in respons expression of SDF-1 [58]. The activation of integrin α4β1 on MSCs in response to kines such as SDF-1 and subsequent binding to VCAM-1 on endothelial cells is a mechanism that facilitates the arrest and extravasation of MSCs in the tumor mic ronment, enabling their tumor-targeted delivery and therapeutic potential [59, searchers have explored various strategies to enhance the homing of MSCs to the tissue, including overexpression of CXCR4 [43], magnetic targeting [35,57], and m tion of MSC surface to secrete proinflammatory cytokines [61]. The baseline expre various cell surface proteins required for effective homing and arrest has been sh be low (~10-13%) [62,63], which could explain the low efficiency of tumor accum with native MSCs.
We advance here a novel approach of incorporating tumor-targeting ligands MSC surface that allows for enhanced arrest and binding within the tumor tissu al. previously demonstrated a surface coating method to target MSCs to inflam bowel disease (IBD) [47,64]. MSCs were coated with Abs against addressins to i their colon delivery for treating Crohn's disease and ulcerative colitis. Ab-coated showed improved treatment efficacy as evidenced by body weight and colon len covery along with increased survival and immune cell suppression [64]. There rently no reports on using ligand-functionalized MSCs for improved tumor ta Tumor homing of MSCs involves tethering of MSCs (mediated by CD44) to selectins expressed on endothelial cells followed by rolling on the vascular wall [57]. G-proteincoupled chemokine receptors (especially CXCR 4 and 7) activate MSCs in response to the expression of SDF-1 [58]. The activation of integrin α4β1 on MSCs in response to chemokines such as SDF-1 and subsequent binding to VCAM-1 on endothelial cells is a critical mechanism that facilitates the arrest and extravasation of MSCs in the tumor microenvironment, enabling their tumor-targeted delivery and therapeutic potential [59,60]. Researchers have explored various strategies to enhance the homing of MSCs to the tumor tissue, including overexpression of CXCR4 [43], magnetic targeting [35,57], and modulation of MSC surface to secrete proinflammatory cytokines [61]. The baseline expression of various cell surface proteins required for effective homing and arrest has been shown to be low (~10-13%) [62,63], which could explain the low efficiency of tumor accumulation with native MSCs.
We advance here a novel approach of incorporating tumor-targeting ligands on the MSC surface that allows for enhanced arrest and binding within the tumor tissue. Ko et al. previously demonstrated a surface coating method to target MSCs to inflammatory bowel disease (IBD) [47,64]. MSCs were coated with Abs against addressins to improve their colon delivery for treating Crohn's disease and ulcerative colitis. Ab-coated MSCs showed improved treatment efficacy as evidenced by body weight and colon length recovery along with increased survival and immune cell suppression [64]. There are currently no reports on using ligand-functionalized MSCs for improved tumor targeting. Notably, the surface functionalization approach is non-genetic and highly modular. Thus, this technology can be used to incorporate any IgG, and can therefore be used to treat other tumor types by utilizing specific Abs against other tumor antigens such as HER2 or EphA2.

Conclusions
Our results demonstrate that incorporating the EGFR-targeting Ab Cmab improved MSC binding to immobilized EGFR protein and to tumor cells overexpressing EGFR. Further, Cmab-MSCs carrying paclitaxel resulted in a significantly improved antitumor effect in lung tumor-bearing mice. Cmab functionalization improved MSC accumulation and retention in a syngeneic mouse lung tumor model. Based on these results, we conclude that ligand functionalization could be used to enhance the concentration of therapeutic MSC constructs at the tumor tissue and to achieve improved antitumor response.

Data Availability Statement:
The data that support the findings of this study will be made available by the corresponding author, SP, upon reasonable request.