Chemicals were provided by either Merck or Sigma-Aldrich. Restriction enzymes and T4 ligase were purchased from Fermentas (Vilnius, Lithuania). The helper plasmid PMJS205 was kindly provided by Prof. Lloyd Ruddock. Escherichia coli BL21 (DE3) was from Invitrogen (Thermo Fisher Scientific) and the pMAL-c2X plasmid was provided from Addgene (#75286). The optimized DNA sequence encoding the D16F in pUC57 was synthesized by Genscript company. Dulbecco′s modified Eagle′s medium (DMEM), trypsin-EDTA, Penicillin/Streptomycin and fetal heat-inactivated bovine serum (FBS) were purchased from Gibco^® (Gaithersburg, USA). Cancerous cell lines including MCF-7 (human breast cancer), LnCap (human prostate cancer), U87MG (human glioblastoma) and human skin fibroblast (HSF) were obtained from Department of Cell Bank, Pasteur Institute of Iran.

2.5 g succinic anhydride was added to the solution of 500 mg Ca-Gly NPs in 120 ml tetrahydrofuran. The dispersion solution was refluxed for 12 h. The obtained Ca-Gly-COOH NPs were washed several times with deionized water and ethanol and dried in an oven (Bi et al., 2018).

0.004 mol of maltose and 0.004 mol of NH₄HCO₃ were added to 20 ml of an aqueous NH3 solution. This solution was heated at 42°C for 36 h, concentrated to half the volume, and then lyophilized (Zhou L. et al., 2012).

500 mg of maltose–NH2 was added to 50 ml of deionized water containing 100 mg Ca-Gly-COOH NPs. 0.5% (w/v) cyanamide and 0.5% (w/v) 1,6-diaminohexan were added to the solution and incubated overnight at RT. After the reaction, the Ca-Gly-Maltose NP was washed several times with deionized water and ethanol. The product was dried in an oven.

The crystal structure of the synthesized NPs was analyzed by Powder X-ray diffraction (PXRD) analysis (Bruker, D8 advance, Germany). Morphology and size of nanoparticles were studied by SEM (TESCAN Vega three microscope) at an accelerating voltage of 20 kV and samples were sputter-coated with gold. The carboxylation and maltose surface modification of Ca-Gly-Maltose NPs (BioMOF) were analyzed by Fourier transform infrared (FT-IR) spectroscopy (Perkin Elmer RXI, United States). The thermal behavior was investigated by a Thermogravimetric analyzer (TGA). Surface area measurements were analyzed using the BET instrument (BELSORP Mini II). The hydrodynamic diameter (Dh) and ζ-potential measurement were performed with a DLS instrument (HORIBA SZ-100).

Prodigiosin (PG) and Simvastatin (SIM) loading was performed by adsorption method for both, alone and in combination. Briefly, 1 mg of Ca-Gly-Maltose NPs were dispersed in 10 ml of PBS buffer solution with a pH∼7.4 for 10 min. Drug loading was performed by iteratively adding drug to the nanoparticle solution at 10 min time intervals while stirring on a rotator (30 rpm) at 25°C.

To estimate the encapsulation efficiency (EE), the protocol previously reported was followed (Rastegari et al., 2017). Briefly, a certain amount of drug-loaded NPs was collected using a centrifuge and the aqueous phase was discarded. After adding 10 ml of methanol, the mixture was shaken for 10 min to completely re-dissolve the entrapped drug. The supernatant was assessed by a spectrophotometer at 535 nm for PG and at 239 nm for SIM. (Shah and Pathak, 2010; Rastegari and Karbalaei-Heidari, 2016). The drug encapsulation efficiency (EE%) and loading capacity (LC) were calculated by following equations: E E ( % ) = drug initial weight (mg) - unloaded drug in aqueous phase(mg) drug initial weight(mg) × 100 L C = Drug initial weight (mg) - unloaded drug in aqueous phase(mg) Nanoparticle weight  ( mg )

In vitro drug release from the Ca-Gly-Maltose nanoparticles were monitored by UV-visible spectrophotometer at 239 nm and 535 nm for PG and SIM, respectively. In order to simulate the biological condition of lysosome, tumor microenvironment and physiological pH of blood stream, the drug release profile of PG and SIM from the newly designed BioMOF was investigated in PBS solution at 37°C during 2 h at pH 5.0, 6.5 and 7.4.

1 mg of drug loaded nanoparticles were dissolved in 10 ml phosphate buffered saline and dispersed with sonication for 5 min. At specified time points, the nanoparticles were collected by centrifuge and 500 µl of the aqueous solution was mixed with acidified ethanol in 1:1 ratio. Concentration of PG and SIM in supernatant was measured by spectrophotometer in a method as described above.

A gene encoding for D16F7 scFv was chemically synthesized after codon optimization to be expressed in bacterial host and ordered as an insert in the pUC57 plasmid. The gene consisted of the sequences encoding the anti-VEGFR-1 antibody variable heavy (VH) and variable light (VL) chains (Sanjuan et al., 2016), a peptide linker (Gly4Ser)3 between them, and a hexa-histidine tag at the C terminus (VL-(Gly4Ser)3-VH-H6) (Supplementary Figure S1). The synthetic gene was then cloned into the expression plasmid, pMAL-c2X which contains a maltose-binding protein (MBP) partner upstream of multiple cloning sites (Alexandrov et al., 2001). The correctness of the new construct, pMAL-c2X-D16F7, was confirmed by DNA sequencing.

The helper plasmid pMJS205, and the recombinant construct (pMAL-c2X-D16F7) were co-transformed into E. coli BL21 (DE3) and grown on LB-agar plate containing 100 μg/ml ampicillin and 35 μg/ml chloramphenicol. A single colony was transferred to LB medium containing both antibiotics overnight at 37°C. The preculture was used to inoculate ZYM-5052 autoinduction medium (Studier, 2005) with the appropriate antibiotics and incubated at 25°C and 250 rpm overnight. Cultures were harvested by centrifugation and pellets were resuspended in a lysis buffer containing 50 mM sodium phosphate buffer, pH = 7.4, 150 mM NaCl and 1% triton x-100. To purify the overexpressed MBP-scFv fusion protein, the soluble fraction of cell extract was loaded onto a Ni-NTA agarose column (1 ml). The partially purified MBP-scFv protein was dialyzed against 20 mM Tris -HCl (pH 7.5), and the homogeneity of the sample was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) in 12.5% acrylamide gels.

Immobilization of MBP-D16F7 scFv on maltose-coated BioMOF was performed using 1:6, 1:8, 1:10, and 1:20 ratios of MBP-scFv and Ca-Gly-Maltose NPs. The MBP-scFv fusion protein was incubated with 2 mg Ca-Gly-Maltose NPs overnight at 4°C. Unbound MBP-scFv was monitored by centrifugation of the nanoparticle solution and analysis of the supernatant by Vis-UV spectrophotometer at 280 nm (ε = 113,,220 M⁻¹ cm⁻¹) (Liu et al., 2009).

All cell lines were cultured in modified Dulbecco’s Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 1% penicillin/streptomycin and maintained at 37°C and 5% CO₂ in a humidified atmosphere in cell incubator. The MTT assay was used to assess the cytotoxicity of free drugs and various nanoformulations against MCF-7, LnCap, U87MG and human skin fibroblast (HSF) cell lines. Cells (1 × 10⁴ cells/well for MCF-7 and U87MG, 1.5 × 10⁴ cells/well for LnCap and 1.2 × 10⁴ cells/well for fibroblast) were seeded in 96-well plates and incubated overnight. After 24- and 48-h incubation, the cells were washed twice with PBS and MTT was added at final concentration of 0.5 mg/ml. After 4 h of incubation at 37°C, the culture media were removed, and the formazan crystals were solubilized in a solution containing 40.0% (v/v) DMF, 16.0% (w/v) SDS and 2% (vol/vol) glacial acetic acid pH∼4.7. Then, the absorbance was measured using a SPECTROstar Nano (BMG Labtech, Germany) microplate reader at 570 nm after background correction at 630 nm. The percentage of cell viability was calculated as follow: v i a b i l i t y ( % ) = A T   ( s a m p l e ) A T   ( c o n t r o l ) × 100 A T = A 570 − A 630

Cellular uptake of the formulations in cells was determined by fluorescence microscopy. Briefly, 2.5 × 10⁴ cells of MCF-7, LnCap, U87MG and HSF were seeded in 48 well plates and incubated at 37°C for 24 h. The culture media were then replaced with fresh medium containing IC₂₀ concentrations of free PG-SIM, PG + SIM-NP and PG + SIM-NP-D16F7 for periods of 30, 90 min, and 24 h. After incubation, cells were washed three times with PBS and fixed with 4% formaldehyde for 20 min. Then, the cells were stained with 300 nM DAPI for 5 min and washed three times. Finally, cells were washed three times with PBS and observed with a Florescence microscope (Olympus; IX51).

MCF-7 cells were seeded at a concentration of 4 × 10⁴ cells/ml in 24 well plates and allowed to grow to a confluence level of 70–80%. Then, wounds were created in each well using a 10 µl pipette tip. The debris was removed by washing with PBS. IC₂₀ concentrations of the above formulations were added and incubated for 48 h. Images were taken immediately after the incision as well as 24 and 48 h after incubation. Migration distance was measured using ImageJ software and wound healing rate was calculated with the following equation: Scratch   healing   rate = ( W 0 − W ) / W 0 × 100 where W₀ is wound width of sample at 0 h and W are the wound width after 24or 48 h (Li et al., 2019).

Data in this study were analyzed using the software GraphPad Prism, version 8.3.0 (GraphPad, San Diego, CA, United States). Comparison between groups was made using the one way with Dunnett’s or two-way ANOVA with Tukey test. Differences with p-values less than 0.05 indicated significance. Combination index (CI) values were determined by CompuSyn software version 1.0 (freeware, The CompuSyn, Inc, Paramus, NJ, United States).

In this study, we synthesized a non-toxic Ca-based BioMOF, which is named Ca-Gly. As shown in Figure 1, Ca²⁺coordinates with the carboxyl (-COOH) and amino (-NH₂) groups of glycine (Yin et al., 2017). Water molecules are also present in the crystal lattice and the Ca²⁺ ions are associated with the oxygen of water. The BioMOF surface was modified by treatment with succinic anhydride containing carboxylic groups. Then, reacted with the amine group of the maltose–NH₂ to obtain Ca-Gly-Maltose nanoparticles (Figure 1).

To investigate the crystalline structure and ensure the synthesis of Ca-Gly, Ca-Gly-COOH, Ca-Gly-Mal, XRD analysis was performed (Figure 2A). The strong peaks at small angles (2ɵ) prove that ample pores are present in the synthesized BioMOF structure. As observed, the main peaks are at 2ɵ values equal to 5.63, 6.49, 8.1, 14.4, 15.23, 16.7, 18.05, 20.79, 22.08, 25.65, 30.47, and 31.08° indicating the crystalline nature of the nanoparticles. The XRD patterns confirmed that the crystalline structure of BioMOF NPs was not changed after surface modification. In Ca-Gly-Maltose, the 29° peak associated with maltose was enhanced.

The average crystallite sizes were determined to be 287, 325, and 319°Afor Ca-Gly, Ca-Gly-COOH, Ca-Gly-Mal using the Debbie Scherrer equation with X’pert software.

The FT-IR spectra of the synthesized Ca-Gly, Ca-Gly-COOH and Ca-Gly-Mal are shown in Figure 2B. The observed peak at 716 cm⁻¹ is associated with Ca-O vibration, which is also seen with a shift in Ca-Gly-COOH and Ca-Gly-Maltose spectra. The peaks at 1,312, 1,335 and 1,366 cm⁻¹ belong to the C-N bond in the amino acid which has been disappeared in the Ca-Gly-Maltose nanocarrier. Compared with the spectrum of Ca-Gly, the new broad peak in the wavenumber region of 2,964–3,448 cm⁻¹ (the carboxyl group) and an enhanced peak at 1,557 and 1,591 cm⁻¹ (the amide group) and a peak at 1,410 cm⁻¹ (for OH of the carboxylic acid group) in the spectrum of Ca-Gly-COOH revealed the successful modification of carboxyl groups on the surface of Ca-Gly (Bi et al., 2018).

The thermal stability of the prepared formulations was evaluated by thermo gravimetric analysis (TGA). As shown in Figure 2C, weight loss occurred in three stages. In the first stage, the lattice water molecules trapped in the pores were lost, with 7% of the weight Ca-Gly evaporating in the temperature ranges of 129–163°C. In the temperature ranges of 163–245 °C, coordinated water molecules had disappeared. The evaporation of the water molecules in this stage resulted in a weight reduction of 20% of Ca-Gly. Finally, in the third stage, from 245°C glycine degradation occurred (Kathalikkattil et al., 2015) and up to 600°C, Ca-Gly lost about 53.44% of its weight. The stability of Ca-Gly-COOH, and Ca-Gly-Maltose increased, so the Ca-Gly-COOH lost 30% of the weight at 150–427°C and 39.64% of the total weight up to 600°C. For Ca-Gly-Maltose, the weight loss up to 600°C was about 21%.

The BET surface areas of Ca-Gly, Ca-Gly-COOH, and Ca-Gly-Maltose BioMOFs decreased from 19.733 to 10.9 m²/g and 2.8153 m²/g after carboxylation and coating with Maltose (Table 1). To address this issue, the N2 adsorption isotherms and pore size distribution diagrams of the nanoparticles have been shown in Figure 2D. The BJH method was used to determine the pore size distribution, which confirmed the mesoporous character of the nanoparticles.

The hydrodynamic size of the Ca-Gly, Ca-Gly-COOH and Ca-Gly-Maltose NPs were 111 nm, 122 and 181 nm, respectively (Table 1). The increase in the hydrodynamic diameter of the NPs after coating is consistent with the existence of carboxylic acid and maltose around the NPs. The ζ-potentials of Ca-Gly, Ca-Gly-COOH and Ca-Gly-Maltose NPs were 6.2 ± 4.66, −4.63 ± 3.18 and −2.8 ± 0.6 mV in PBS solution, respectively (Table 1). Therefore, the presence of carboxylic acid negatively decreased the surface charge of Ca-Gly-COOH, while maltose coating positively increased the surface charge.

Scanning electron microscopy analysis of Ca-Gly, Ca-Gly-COOH, and Ca-Gly-Maltose NPs was also performed. The particle sizes were 60–79 nm, 62–105 nm and 61–126 nm for Ca-Gly, Ca-Gly-COOH and Ca-Gly-Maltose NPs, respectively. SEM analysis shows that the Ca-Gly was spherical, and no morphological changes were observed after coating (Figure 2F).

At first, the PG and SIM loading on synthesized BioMOFs was performed individually in the concentration range of 50–400 nmol (∼16.17–129.37 μg/mg for PG and 20.93–167.5 μg/mg for SIM). PG was attempted to be loaded up to 64.90 μg/mg with a loading efficiency of 94.72%, and SIM was loaded up to 110 μg/mg with an efficiency of 66.25%. The combined effect of PG and SIM (PG + SIM) on cancer cells was investigated by applying a 1:1 ratio of the drugs (PG IC50/SIM IC50) based on their IC₅₀ values. Due to the high sensitivity of the cells to the combination of PG and SIM, the drugs were loaded at a lower ratio (PG:SIM = 1:3.5) for co-loading. Therefore, 3% (w/w) of PG (30 μg/mg) with an efficiency of 98.3 ± 1.24 and 11% (w/w) of SIM (110 μg/mg) with an efficiency of 66.25 ± 2.67% were loaded onto the Ca-Gly-Maltose BioMOF for further investigations. After 4 days of incubation at 4°C, 94.57% of PG and 64.54% of SIM were still present.

In order to simulate the biological condition of lysosome, tumor microenvironment, and blood stream, the drug release profile of PG and SIM loaded BioMOF was investigated In vitro and monitored by UV-visible spectrophotometer at 239 nm and 535 nm for PG and SIM, respectively.

As shown in Figure 3A, the PG release at physiological pH (∼7.4) was slower than that at pH 6.5 and 5.0, with an initial ∼4% release during 30 min and ∼15.72% release after 150 min. The PG release accelerates up on increasing of acidic conditions, showing ∼5.5 and 7% drug release after 30 min, respectively. However, the release was increased in acidic conditions after 60 min and reached to ∼31% and ∼35% for PG within 90 min at pH 6.5 and 5.0, respectively.

Also, the SIM shown the same pattern of release at neutral pH, with an initial ∼10.38% release during 30 min and 20.5% release in 150 min (Figure 3B). The drug release rate is much faster in acidic conditions, showing ∼45% and ∼54% unloading after 60 min incubation at pH 6.5 and 5.0, respectively. The results showed that the pH- dependent release of PG was much lower than Simvastatin. The drug release profiles for PG and SIM were pH-dependent and time-dependent. In acidic conditions, release of drugs is faster than physiological pH and 6.5. PH-dependent release behavior can be useful for the development of the drug delivery systems in cancer cells due to the acidic pH of the tumor environment. According to reports, drug release from MOFs is pH-dependent (Liu et al., 2019; Xiong et al., 2020; Ma et al., 2021). Rapid release in acidic conditions may be due to protonation of NPs and their structure degradation. With the decrease of the pH values of PBS media, the NPs exhibit an accelerated degradation. Our data also showed that the release rate of Simvastatin was higher than Prodigiosin which is may because of more hydrophobic nature of PG in compare to SIM.

As shown in Figure 4A, double digestion of the new construct yielded 6,619 and 990 bp DNA fragments, confirming proper integration of the scFv insert in downstream of a maltose binding protein gene sequence followed by a Ser-Asn linker into the pMAL-c2X plasmid (Supplementary Figure S1 for more detail).

Co-expression of the enzymes of the pMJS205 plasmid and the MBP partner protein in pMAL-c2X-D16F7 resulted in a sufficient amount of soluble MBP-scFv in the cytoplasm of E. coli, when cultured in ZYM-5052 autoinduction medium for 24 h. As shown in Figure 4B, partial purification of the fusion protein was achieved using Ni-NTA agarose column. A single band with suitable homogeneity and a molecular mass of approximately 71.5 kDa was seen in SDS-PAGE.

The MTT assay was performed to determine the viability of cells after treatment with the different formulations as a function of time and concentration. First, the cytotoxicity of blank nanoparticles (Ca-Gly-Maltose) was investigated in cancerous and normal cell lines. The unloaded synthesized nanoparticles showed very low toxicity in the concentration ranges of 6.25–300 μg/ml, and no significant differences were observed between the negative control and cells exposed to increasing concentrations of the nanoparticles (Supplementary Figure S2A). The purified free MBP-D16F7 protein had very low cytotoxicity in the concentration range of 3.125–200 μg/ml, although the viability of cancer cells decreased significantly compared with normal cell at concentrations above 100 μg/ml (Supplementary Figure S2B). The empty BioMOF-D16F7 showed a slightly stronger toxic effect than the blank nanoparticles and the free MBP-scFv (Supplementary Figure S2C).

The IC₅₀ values for free PG and SIM were measured after 24 and 48 h in three different cancer cell lines (MCF-7, LnCap and U87MG) and one non-cancerous cell (Human Skin Fibroblast, HSF). Given the known selectivity of PG for cancer cells (Pérez-Tomás and Vinas, 2010; Rastegari and Karbalaei-Heidari, 2016), the free PG at a range of concentrations (from 0.161 μg/ml to 16.17 μg/ml) and exposure times showed a higher cytotoxicity for the cancer cell lines than for the HSF cells (Supplementary Figure S3). In accordance with these differences, the selective index (SI) values of the free PG on MCF-7, LnCap, and U87MG cell lines were determined as 2.53, 1.94, and 1.90 for 24 h and 2.01, 1.42 and 1.78 for 48 h, respectively. Similarly, SIM also showed an increase in cytotoxicity with increasing of the concentrations (from 0.209 μg/ml to 41.87 μg/ml) and exposure time, although its IC50 values were higher than those of PG (Supplementary Figure S3).

As shown in Figure 5, the IC₅₀ of the PG-NPs was higher than that of free PG at 24 and 48 h, indicating lower toxicity of the compounds after loading, probably due to the slow release of drug from the nanoparticles. However, SIM-NPs showed higher toxicity on cells than the free SIM (lower IC₅₀). For SIM-loaded BioMOFs, higher cytotoxicity of drug-loaded NPs and D16F7-functionalized NPs was observed in all tested cell lines.

To have better comparison, we summarized the IC₅₀ values of all formulations after 24 and 48 h incubation in different cell lines (Supplementary Table S1). PG-NP-D16F7 and SIM-NP-D16F7 showed significantly improved cytotoxic activity in different cell lines after 24 and 48 h treatment compared with non-targeted nanoparticles. However, the highest cytotoxicity for both PG-NP-D16F7 and SIM-NP-D16F7 was in U87MG cells, with IC₅₀ values of 1.17 and 3.86 μg/ml after 24 h and 0.91 and 2.31 μg/ml after 48 h, respectively.

These results are consistent with the aim of the study to improve drug cytotoxicity on cancerous cells after loading and functionalization of Ca-Gly BioMOFs. Statistical analysis revealed that although there were no significant differences between the PG-loaded NPs and PG-D16F7-BioMOFs on most cell lines, but there were statistically significant effects for SIM-loaded BioMOFs (Figure 5). Overall, incubation of cells with the free drugs and drug-loaded nanoparticles showed that cytotoxicity exhibited a concentration and time-dependent behavior showing an enhancement of toxicity with increasing drug concentration and exposure time, and in some cases when the BioMOFs were functionalized with an anti-VEGFR1 antibody.

The combined cytotoxic effect of PG and SIM was evaluated by the MTT cytotoxicity assay at an equipotent molar ratio ([PG] at IC₅₀ [SIM] at IC₅₀). The combination index (CI) has been used to evaluate the synergistic, antagonistic or additive effects of drugs combination. CI > 1 indicates antagonism, CI < 1 indicates synergy and CI = 1 indicates an additive effect (Chou and Martin, 2007; Chou, 2010). The CI of each drug combination was plotted as a function of fractional inhibition (Fa) by computer simulation from Fa = 0.10 to 0.95 (Supplementary Figure S4). As summarized in Table 2, the combination of PG and SIM (PG + SIM) had a greater anticancer effect in their equipotency ratio and showed a lower value of the IC₅₀ than the free drugs alone. For example, after 24 h incubation on MCF-7 cells, the PG + SIM combination showed an IC₅₀ = 5.99 μg/ml (0.33 μg/ml PG+ 5.66 μg/ml SIM), while the IC₅₀ values for the free PG and free SIM were 0.73l and 9.22 μg/ml, respectively (Figure 5; Supplementary Table S1). Table 2 shows the CI values with their interpretations in different cells for PG + SIM and PG + SIM-NP-D16F7 treatment at 24 and 48 h. This result indicates that the PG + SIM inhibits cancer cells growth more effectively than free PG and SIM. Moreover, the CI values obtained for PG + SIM-NP-D16F7 after 24 and 48 h treatment showed synergistic cytotoxicity (Table 2).

The CI value of PG + SIM was less than PG + SIM-NP-D16F7 after 48 h in U86MG cells. Cells are more sensitive to free drugs than drug-loaded nanoparticles where have slower release. The PG + SIM-NPs-D16F7 showed lower IC₅₀ values than the free PG + SIM and PG + SIM-NPs which could be attributed to the functionalization of the nanoplatform with the D16F7 scFv, confirming the effect of targeting on the VEGFR1 receptor on cancer cells, although no significant differences were seen between them except for MCF-7 (Figure 5). Easy accessibility of cells for the hydrophobic drugs is the main reason of these observations which obscures the targeting behavior of antibodies in Ca-Gly-Maltose-D16F7.

The efficiency of cellular uptake was assessed using an inverted fluorescence microscope. The results showed that these formulations were able to deliver the drugs into cells, and the fluorescence increased dramatically after 90 min (Supplementary Figure S5). Since the PG has autofluorescence, its penetration into cells can be assessed with a fluorescence microscope. The hydrophobicity of the compounds contributes to their ease of penetration into the cell in free form. However, the higher fluorescence emission of cells treated with PG + SIM-NP-D16F7 compared with PG-SIM-NP in a short time (30 min) after addition showed that the targeting design worked well and uptake of the smart nanocarrier occurred via antibody-receptor interactions (Supplementary Figure S5).

Cellular migration or tumor invasion is a crucial phenomenon in carcinogenesis, as metastasis of tumor cells occurs in this way. According to the MTT results, the developed nanoplatforms significantly decreased the cell viability of MCF-7 cell line. In our study, a scratch assay was performed at IC₂₀ concentrations to examine the potential of the free drugs and the developed drug-loaded nanocarriers to inhibit cell migration. Since the cells will more rapidly die at IC₅₀ of the drugs, for wound healing assay, a sub-toxic (IC₂₀) concentration is generally used to perform the experiment. As shown in Figure 6, the cell migration inhibition assay revealed a reduction in cellular proliferation and migration when treated with the various formulations compared to the control group (without treatment).

For the free PG and SIM at IC₂₀ concentration, there was a significant difference in cell migration (18.32% at 24 h and 16,79% at 48 h for the free PG and 20.51 and 31.28% at 24 and 48 h, for the free SIM); while the wound healing rate was 38.04 and 6.82% at 24 and 48 h for the combination of free PG and SIM. The PG-NP, the SIM-NP, and the PG + SIM-NP were also able to inhibit the cell proliferation weaklier than the free PG, SIM and the PG + SIM, probably due to the slow penetration of the nanoparticles. Inhibition of cell migration by PG-NP-D16F7 and PG + SIM-NP-D16F7 was significantly higher than that of PG-NP and the PG + SIM-NP, respectively; while there was no significant difference between SIM-NP-D16F7 and SIM-NP. The D16F7 scFv and functionalized nanoparticles (D16F7-NP) showed a reduction in cell migration compared to the blank nanoparticles and untreated control, highlighting the superiority of D16F7 for targeting cancer cells.