Synergetic Enhancement of Tumor Double-Targeted MRI Nano-Probe

The conventional targeted delivery of chemotherapeutic and diagnostic agents utilizing nanocarriers is a promising approach for cancer theranostics. Unfortunately, this approach often faces hindered tumor access that decreases the therapeutic index and limits the further clinical translation of a developing drug. Here, we demonstrated a strategy of simultaneously double-targeting the drug to two distinct cites of tumor tissue: the tumor endothelium and cell surface receptors. We used fourth-generation polyamideamine dendrimers modified with a chelated Gd and functionalized with selectin ligand and alpha-fetoprotein receptor-binding peptide. According to the proposed strategy, IELLQAR peptide promotes the conjugate recruitment to the tumor inflammatory microenvironment and enhances extravasation through the interaction of nanodevice with P- and E-selectins expressed by endothelial cells. The second target moiety—alpha-fetoprotein receptor-binding peptide—enhances drug internalization into cancer cells and the intratumoral retention of the conjugate. The final conjugate contained 18 chelated Gd ions per dendrimer, characterized with a 32 nm size and a negative surface charge of around 18 mV. In vitro contrasting properties were comparable with commercially available Gd-chelate: r1 relaxivity was 3.39 for Magnevist and 3.11 for conjugate; r2 relaxivity was 5.12 for Magnevist and 4.81 for conjugate. By utilizing this dual targeting strategy, we demonstrated the increment of intratumoral accumulation, and a remarkable enhancement of antitumor effect, resulting in high-level synergy compared to monotargeted conjugates. In summary, the proposed strategy utilizing tumor tissue double-targeting may contribute to an enhancement in drug and diagnostic accumulation in aggressive tumors.


Introduction
Cancer represents a complex disease type that causes substantial mortality in modern society. Despite the plentitude of traditional drugs (e.g. daunorubicin, paclitaxel, etc.) and the development and clinical application of novel monoclonal antibodies (mABs) and small-molecule inhibitors, modern therapy still demands safe and effective anticancer therapeutics and diagnostics.
Nanoparticle (NP)-based drug delivery systems continue to attract enormous attention as important vehicles in cancer therapy [1]. Due to their small size, variability of preparations and modifications, and specific tumor physiology, NPs are widely applied sis. IELLQAR peptide is a synthetic selectin ligand that is utilized for tumor vasculature targeting [42,43].
Thus, regarding the aforementioned data, we chose two peptides as tumor-targeting moieties, KQEFLIN-AFP receptor-binding peptide and IELLQAR-P-selectin-biding peptide, which were immobilized on the surface of polyamideamine dendrimers.
We demonstrated an approach employing distinct binding moieties of tumor tissue for drug delivery. The different location and expression levels of the selectins (tumor endothelial cells) and AFP receptors (tumor cells) at the same tumor nodes allowed us to improve targeted delivery on the nanovehicle compared to monotargeted conjugates. The present and aforementioned studies describing multitargeting strategies resemble the widespread treatment strategies of bacterial infections (combinations of antibiotics), cancer (combinations of anticancer drugs), and viral infections (e.g., highly effective combinations of targeted anti HIV/AIDS drugs). Thus, we expect that the described study will provide not only information about the accumulation efficacy of this system in tumors, but also help researchers choose the appropriate carrier, drug, or vector and evaluate the worthiness of the delivery approach that will be applied.

PAMAM G4 Derivatives Synthesis
PAMAM G4 dendrimers were labeled with p-SCN-Bn-DTPA and FITC (in vitro imaging) or Cy5.5 (in vivo imaging). A total of 5 µM of PAMAM G4 was dissolved in 10 mL of 125 mM NaHCO 3 , mixed with 10 molar equivalents of p-SCN-Bn-DTPA, stirred at room temperature (RT) for 24 h, purified by dialysis (2 kDa MWCO membrane) against aqueous solution of 125 mM NaHCO 3 and then against deionized water, and finally freeze-dried. The DTPA content was analyzed quantitatively using the previously described spectrophotometric method [44]. G4-DTPA metallation was accomplished in 100 mM of ammonium acetate buffer pH 5.5; 1.5 molar equivalents of GdCl 3 was added to DTPA and the reaction mixture was incubated at 37 • C for 2 h. The amount of bound Gd was measured applying ICP-MS analysis and using arsenazo III [45].
The labeling of G4-DTPA with FITC or Cy5.5 was performed in 100 mM of NaHCO 3 over 15 h at RT; the amount of dye bound after purification by dialysis was measured at λ 495 for FITC and λ 680 for Cy5.5.

Characterization
1 H NMR spectra of all intermediate and final polymer products were recorded at 25 • C on a Bruker Avance III HD 500 Mhz using D 2 O as the solvent and a sample concentration of 5 mg/mL [46].
The Gd content was analyzed with ICP-MS (Agilent 7500C, Agilent Technologies, Tokyo, Japan) [47]. The sample introduction system consisted of a robust Babington nebulizer with a Scott spray chamber (Agilent Technologies, Tokyo, Japan) cooled by a Peltier element (2 • C). The data were acquired and processed with the ICP-MS ChemStation (Agilent Technologies, version G1834B; Santa Clara, CA, USA) software package. For sample dilution, the mixture of 0.1% Triton X-100-1% HNO 3 was used. The same solution was used as the mobile phase in the mass spectrometer.
The terminal maleimide as well as IELLQARCRR and KQEFLINGCRR residues were quantified after G4-DTPA-PEG conjugation with GSH excess followed by unreacted glutathione colorimetric analysis with DTDP at λ = 324 nm (ε 324 of DTDP = 19,800 M −1 cm −1 ) [48]. Briefly, 0.1 mL G4-DTPA-PEG, G4-DTPA-PEG-KQ or G4-DTPA-PEG-IEL in a concentration range of 1-20 µM (according to G4 concentration) was mixed with 0.5 mL of 50 µM GSH in 10 mM PBS at pH 7.4, incubated at 37 • C for 40 min, mixed with 0.25 mL of 120 µM DTDP in 10 mM PBS at pH 7.4, and incubated at room temperature for 2 min; unreacted GSH contributed to absorbance at 324 nm after reaction with DTDP and was used for maleimide termini quantification.
The size distribution of the derivatives was determined by dynamic light scattering (DLS) in deionized water, while the zeta-potential was determined by electrophoretic light scattering in 10 mM of NaCl (Nano-ZS ZEN 3600, Malvern-Instruments, Worcestershire, UK) at a sample concentration of 1 mg/mL at 25 • C [49].
The morphology of the dried polymers was analyzed using a transmission electron microscope (TEM) (Talos F200i S/TEM (Thermo Scientific, Waltham, Massachusetts, US). Samples were diluted with water to a concentration of 1 mg/mL, dispersed onto 3 mm carbon-coated Cu grids, and air-dried.

In Vitro Internalization Assay
The intracellular uptake was analyzed by flow cytometry and confocal laser scanning microscopy (CLSM). SKOV-3 cells were maintained in DMEM supplemented with 5% FBS and 50 µg/mL gentamycin. Cells were incubated with corresponding conjugates labeled with FITC for 1 h and washed with PBS. The intracellular FITC fluorescence was analyzed using a Dako CyAn Flow Cytometer equipped with a 488 nm laser (Beckman Coulter, Brea, CA, USA). For CLSM analysis, cells were seeded onto 24-well plates equipped with round glass slides and maintained in corresponding media. Cells were incubated with conjugates labeled with FITC for 2 h, washed with PBS, and fixed in 2% paraformaldehyde. Nuclei were stained with Hoechst33342. Cells were analyzed with a Zeiss Cell Observer Z1 confocal laser scanning microscope and images were acquired with the AxioVision (Germany) and Imaris (Bitplane AG, Zürich, Switzerland) software.

Cytotoxicity Assay
In vitro cytotoxicity of the conjugates was evaluated applying the MTT assay. Human ovarian adenocarcinoma SKOV-3 cells were plated on 96-well plates and incubated for 24 h in a corresponding medium at a density of 5000 cells per well at 37 • C under humidified atmosphere of 5% CO 2 . The next day, the medium was supplemented with G4 conjugates or G4 in a concentration range of 500-3.9 µM dissolved in DMEM. After 72 h of incubation, cell survival was determined using standard MTT assay: 50 µL MTT in DMEM (1 mg/mL) was added into each well. After cell incubation for 4 h at 37 • C, the medium was removed and precipitated formazan crystals were dissolved in 100 µL DMSO. Next, the absorption intensity of formazan was measured at 540 nm on a microplate reader. Cell viability was determined as a percentage of the untreated control. The survival values were normalized to an untreated control [50].

MR Imaging
MRI studies were performed on a 7.05T MRI scanner Bruker BioSpec 70/30 USR with a ParaVision v.5.1 software. A birdcage of 72 mm internal diameter was used as a transceiver.
The phantoms were the test tubes (eppendorfs of 2 mL volume) filled with substances (Magnevist ® or G4-DTPA-PEG-Pep) diluted in distilled water in different Gd concentration: 25 mM, 12.5 mM, 6.25 mM, 3.125 mM, and 1.625 mM. One more phantom filled with distilled water was used as a reference. Thus, there were 11 phantoms in total.
To compare the contrast properties of the substances (G4-DTPA-PEG-Pep and Magnevist ® ), we determined their relaxivities r 1 and r 2 , which reflect a linear correlation function between the relaxation rates of a solution and the concentration of the paramagnetic substance [51]: (1) where T 1 and T 2 are the longitudinal and transverse relaxation times (in seconds), respectively, of the protons in the phantoms; T 1w and T 2w are the relaxation times (in seconds) of the protons in distilled water; and [Gd] is the concentration of the Gadolinium in mM. First, we measured the T 1 and T 2 relaxation times of the protons in each of the phantoms. Then, we calculated the relaxation rate constants R i = 1/T i , where T i -is the longitudinal or transverse relaxation time. The obtained data was used to plot the dependence graphs of the values of the (1/T i − 1/T iw ) on the Gd concentration in the samples.
The T 1 measurements were carried out using a saturation recovery pulse sequence RAREVTR (Rapid Acquisition with Relaxation Enhancement with variable repetition time TR) in coronal projection with the following scan parameters: FOV (field of view): 8 cm × 8 cm, matrix: 32 × 32, in-plane resolution: 2.5 mm × 2.5 mm, number of slices: 1, slice thickness: 10 mm, number of averages: 2. For the phantoms with a Gd concentration of 25 mM, 12.5 mM and 6.25 mM, the TR values were set from 10 ms to 1000 ms and the echo time (TE) was 5 ms with a Rare-factor of 1-total scan time was 3 min and 54 s. For the other phantoms, the TR values were set from 125 ms to 15,000 ms and the effective echo time (TE eff ) was 15 ms with a Rare-factor of 8-total scan time was 5 min and 55 s.
The T 2 measurements were conducted using a spin echo pulse sequence MSME (multi-spin multi-echo) with the same geometric parameters that were applied for the T 1 measurements. The difference was in the number of echoes, 64, and the number of averages, 1. For the phantoms with a Gd concentration of 25 mM, 12.5 mM and 6.25 mM, the TE values were set from 7.8 ms to 500 ms and the TR was 5000 ms-total scan time was 2 min. For the other phantoms, the TE values were set from 150 ms to 9600 ms and the TR was 12,000 ms-the total scan time was 2 min.
Twelve-week-old female BALB/c Nude mice were purchased from Pushino Animal Facility (Russia). The experimental procedures with mice were approved by the N.M. Emanuel Institute of Biochemical Physics Ethics Committee for the use of experimental animals and performed according to the guidelines of European Medicines Agency, Amsterdam, Netherlands. The mice (18-20 g) were inoculated subcutaneously with 10 6 SKOV-3 cells in 100 µL DMEM. After subcutaneous tumor inoculation (day 0) at day 45 tumors reached sizes of~800-1100 mm 3 . Two mice groups (n = 3) were used for the assessment of the MRI contrasting activity of double-targeted G4-DTPA-PEG-Pep immediately after injection; the control group was injected with Magnevist. The amount of drug injected was 0.1 mL at a concentration of 15 mM (according to Gd). A spin echo pulse sequence RARE (Rapid Acquisition with Relaxation Enhancement) was used, and the scan parameters were as follows: FOV: 10 cm × 4 cm; matrix: 400 × 160; in-plane resolution: 0.25 mm × 0.25 mm; number of slices: 18; slice thickness: 1 mm; number of averages: 2; Rare-factor: 4; TR: 1300 ms; TE eff : 9.7 ms; scan time: 2 min and 36 s. To avoid the chest movement artifacts, the scan was produced using the respiratory gating monitoring system. Thus, the total scan time was twice as long and was about~5 min. The graphs were plotted using the Microsoft ® Excel ® 2013 program. The analysis of the in vivo MRI studies was carried out in the ImageJ v.1.51j software [52]. First, the MR images were normalized to the same noise, and then the signal intensity I s was calculated in the location of the tumor before I s1 and after the injection I s2 of the substance (Magnevist ® or G4-DTPA-PEG-Pep). The region of interest (ROI) was set as yellow circle, as shown in Figure 4b. The contrast assessment was performed by calculating the relative change in the signal intensity: I s2 /I s1 .

In Vivo and Ex Vivo Intratumoral Accumulation
Three mice groups (n = 3) were used for the assessment of the intratumoral accumulation of Cy5.5-labeled G4-DTPA-PEG-Pep and two mono-targeted conjugates-G4-DTPA-PEG-KQ and G4-DTPA-PEG-IEL-4 h after injection of 0.2 mL of 100 nM solutions (according to Cy5.5 content); the control group was injected with non-targeted G4-DTPA-PEG. Tumors of the same groups were analyzed for the accumulation of conjugates 24 h after administration ex vivo. The visualization was performed using the IVIS Spectrum CT imaging system (Perkin Elmer, Hopkinton, Massachusetts USA), followed by Living Image software processing. During in vivo experiments, the mice were anesthetized with Isoflurane (Karizoo).

Statistical Analysis
The data visualization and statistical analysis were performed in OriginPro (OriginLab Corporation, Northampton, MA, USA) and Excel (Microsoft Corporation, Redmond, WA, USA). The results are represented as mean ± standard deviation. One-way analysis of variance was applied to determine the significant difference. p < 0.05 was considered as significant. Figure 1 shows a schematic double-targeted G4-DTPA-PEG synthesis. The conjugate was synthesized by a covalent conjugation SCN-DTPA with dendrimer's NH 2 -termini followed by a Gd chelation step. FITC or Cy-5.5 performed a label function for in vitro and vivo analysis. Further, we conjugated G4-DTPA with PEG arms containing methoxy moieties and maleimide followed by modification with cysteine-containing IELLQAR-CRR and KQEFLING-CRR peptides. CRR tags contained cysteine residue essential for maleimide conjugation and double arginine, which enhances peptides' solubility. The final shell consisted of DTPA, peptides, and methoxy-PEG arms. These masked the positive surface charge, increased the hydrodynamic radius of PAMAM dendrimers, reduced the nonspecific cell membrane interaction and opsonization with plasma proteins, and enhanced the blood circulation time and in vivo tumor accumulation [53]. The The final shell consisted of DTPA, peptides, and methoxy-PEG arms. These masked the positive surface charge, increased the hydrodynamic radius of PAMAM dendrimers, reduced the nonspecific cell membrane interaction and opsonization with plasma proteins, and enhanced the blood circulation time and in vivo tumor accumulation [53]. The tumor targeting ligands IELLQARCRR and KQEFLINGCRR were selectively conjugated with PEGs via maleimide residues and SH group of peptides, and provided an active targeting function to the final nanocarrier.

Synthesis and Characterization of Peptides-Conjugated PAMAM-PEG
The G4-DTPA-PEG-Pep conjugate was characterized by a spherical shape (Figure 2a), a size of about 32.1 nm, and a negative zeta potential (−18.4 mV) (Figure 2b). The final shell consisted of DTPA, peptides, and methoxy-PEG arms. These masked the positive surface charge, increased the hydrodynamic radius of PAMAM dendrimers, reduced the nonspecific cell membrane interaction and opsonization with plasma proteins, and enhanced the blood circulation time and in vivo tumor accumulation [53]. The tumor targeting ligands IELLQARCRR and KQEFLINGCRR were selectively conjugated with PEGs via maleimide residues and SH group of peptides, and provided an active targeting function to the final nanocarrier.
The G4-DTPA-PEG-Pep conjugate was characterized by a spherical shape ( Figure  2a), a size of about 32.1 nm, and a negative zeta potential (−18.4 mV) (Figure 2b). We applied ICP-MS analysis to evaluate the amount of Gd bound and H 1 NMR to evaluate the quantity of PEG conjugated to the dendrimer surface (Figure 2c). The major peak at 3.63 ppm corresponded to the methylene protons (CH 2 -O-CH 3 ) of PEG; the peak at 6.75 ppm corresponded to maleimide protons (CH-CH) (R 2 ); the peaks at 2.35, 2.57, 2.76, 3.01, and 3.56 ppm corresponded to the methylene protons of the dendrimer; the peaks at 3.24 and 3.74 ppm corresponded to the protons of DTPA (Figure 2c). We applied GHS/DTDP analysis to quantify the surface maleimide residues and analyze the amount of conjugated peptides.
According to the spectrophotometric and NMR analysis, the composition of conjugate consisted of 29 PEG (approximately 19 oxymethyl-and 10 maleimide-terminated PEG residues) and 18 DTPA residues per dendrimer. PEG quantity was evaluated based on the relative intensity ratio of the peaks at 3.63 and 2.35 ppm (Figure 2c), corresponding to the methylene protons (CH 2 CH 2 O) of PEG and the methylene protons (CH 2 CH 2 CONH) of dendrimers, respectively. The peptide conjugation demonstrated an approximately 60-80% efficiency according to the GHS/DTDP analysis: 1 µM of G4-DTPA-PEG bound approximately 4.1 µM of IELLQARCRR and 3.5 µM KQEFLINGCRR, regardless of the conjugation order.

Effects of Targeting Ligands on the Cellular Uptake
We applied flow cytometry and CLSM to evaluate the cellular uptake of conjugates by AFP receptor-positive SKOV-3 cells. We used mono-peptide-conjugated G4-DTPA-PEG-KQ and G4-DTPA-PEG-IEL as the controls. Figure 3a    Flow cytometry analysis revealed a high level of internalization of G4-DTPA-PEG-Pep after incubation, while counter-targeted G4-DTPA-PEG-IEL and non-targeted G4-DTPA-PEG conjugates demonstrated low internalization, which coincided with the CLSM results (Figure 3b). Cytotoxicity analysis displayed a low level of activity of G4-DTPA-PEG-Pep against SKOV-3 cells in contrast with non-PEGylated G4 (Figure 3c).

In Vivo Biodistribution and Intratumoral Accumulation
We evaluated the efficiency of G4-DTPA-PEG-Pep as a tumor-contrasting agent. The conjugate demonstrated a relaxivity comparable with that of Magnevist in vitro at the same Gd concentration: the r 1 relaxivity was 3.39 for Magnevist and 3.11 for conjugate; the r 2 relaxivity was 5.12 for Magnevist and 4.81 for conjugate (Figure 4a). Slightly lower values for conjugate could be explained by the hindered water accessibility to chelated Gd.   Ex vivo fluorescence analysis 24 h after the conjugate administration confirmed the specific internalization of the conjugate in the tumor tissue ( Figure 5b). Overall, despite some drawbacks, such as initial renal excretion and partial absorption by the liver, these results revealed a certain enhancement of the tumor-specific accumulation of the doubletargeted conjugate over the mono-targeted one and proved the applicability of this strategy for tumor targeting.

Conclusions
Using a PAMAM dendrimer-based system decorated with two vectors for tumor targeting as a delivery system, we assessed whether double-targeting improves the accessibility of tumors for the transported drug compared to non-targeted or mono-targeted vehicles. We chose as the targets P-selectin, a marker of inflamed vasculature that is also characteristic of tumor tissue, and AFP receptor, which is a marker of a wide range of tumors, including SKOV-3 cells. Unfortunately, due to the distinct expression sites, it was impossible to evaluate the targeting properties of double-targeted conjugate in vitro in one experiment. However, double-targeted G4-DTPA-PEG-Pep demonstrated high internalization levels by SKOV-3 cells. In contrast, the same cells incubated with counter-targeted G4-DTPA-PEG-IEL revealed low internalization levels similar to those after incubation with non-targeted G4-DTPA-PEG conjugate. As expected, the PEGylated conjugates displayed low cytotoxicity. An in vitro r1 and r2 relaxivity assessment of G4-DTPA-PEG-Pep revealed similar values to those of commercially available Gd-chelate Magnevist. Immediately after injection, G4-DTPA-PEG-Pep demonstrated an insignificantly lower MRI contrasting ability in vivo compared to Magnevist; the conjugate also contrasted kidneys, which could be explained by a low molecular fraction or non-covalently bounded DTPA-Gd excretion. However, over longer periods, Cy5.5-labeled G4-DTPA-PEG-Pep displayed a clear tumor-associated localization. Thus, these results suggest that the double-targeted strategy, with the use of a proper nanocarrier, surface modification, and convenient vector selection, can be utilized to improve the specific anticancer drug delivery.  Ex vivo fluorescence analysis 24 h after the conjugate administration confirmed the specific internalization of the conjugate in the tumor tissue ( Figure 5b). Overall, despite some drawbacks, such as initial renal excretion and partial absorption by the liver, these results revealed a certain enhancement of the tumor-specific accumulation of the doubletargeted conjugate over the mono-targeted one and proved the applicability of this strategy for tumor targeting.

Conclusions
Using a PAMAM dendrimer-based system decorated with two vectors for tumor targeting as a delivery system, we assessed whether double-targeting improves the accessibility of tumors for the transported drug compared to non-targeted or mono-targeted vehicles. We chose as the targets P-selectin, a marker of inflamed vasculature that is also characteristic of tumor tissue, and AFP receptor, which is a marker of a wide range of tumors, including SKOV-3 cells. Unfortunately, due to the distinct expression sites, it was impossible to evaluate the targeting properties of double-targeted conjugate in vitro in one experiment. However, double-targeted G4-DTPA-PEG-Pep demonstrated high internalization levels by SKOV-3 cells. In contrast, the same cells incubated with counter-targeted G4-DTPA-PEG-IEL revealed low internalization levels similar to those after incubation with non-targeted G4-DTPA-PEG conjugate. As expected, the PEGylated conjugates displayed low cytotoxicity. An in vitro r 1 and r 2 relaxivity assessment of G4-DTPA-PEG-Pep revealed similar values to those of commercially available Gd-chelate Magnevist. Immediately after injection, G4-DTPA-PEG-Pep demonstrated an insignificantly lower MRI contrasting ability in vivo compared to Magnevist; the conjugate also contrasted kidneys, which could be explained by a low molecular fraction or non-covalently bounded DTPA-Gd excretion. However, over longer periods, Cy5.5-labeled G4-DTPA-PEG-Pep displayed a clear tumorassociated localization. Thus, these results suggest that the double-targeted strategy, with the use of a proper nanocarrier, surface modification, and convenient vector selection, can be utilized to improve the specific anticancer drug delivery.