An Evaluation of CXCR4 Targeting with PAMAM Dendrimer Conjugates for Oncologic Applications

The chemokine receptor 4 (CXCR4) is a promising diagnostic and therapeutic target for the management of various cancers. CXCR4 has been utilized in immunotherapy, targeted drug delivery, and endoradiotherapy. Poly(amidoamine) [PAMAM] dendrimers are well-defined polymers with unique properties that have been used in the fabrication of nanomaterials for several biomedical applications. Here, we describe the formulation and pharmacokinetics of generation-5 CXCR4-targeted PAMAM (G5-X4) dendrimers. G5-X4 demonstrated an IC50 of 0.95 nM to CXCR4 against CXCL12-Red in CHO-SNAP-CXCR4 cells. Single-photon computed tomography/computed tomography imaging and biodistribution studies of 111In-labeled G5-X4 showed enhanced uptake in subcutaneous U87 glioblastoma tumors stably expressing CXCR4 with 8.2 ± 2.1, 8.4 ± 0.5, 11.5 ± 0.9, 10.4 ± 2.6, and 8.8 ± 0.5% injected dose per gram of tissue at 1, 3, 24, 48, and 120 h after injection, respectively. Specific accumulation of [111In]G5-X4 in CXCR4-positive tumors was inhibited by the peptidomimetic CXCR4 inhibitor, POL3026. Our results demonstrate that while CXCR4 targeting is beneficial for tumor accumulation at early time points, differences in tumor uptake are diminished over time as passive accumulation takes place. This study further confirms the applicability of PAMAM dendrimers for imaging and therapeutic applications. It also emphasizes careful consideration of image acquisition and/or treatment times when designing dendritic nanoplatforms for tumor targeting.


Modification of POL3026 with the Linker
First, 0.030 g of POL3026 ( Figure 1A, 10.7 µmole) was dissolved in 1 mL DMF, and 0.1 mL DMF containing 1.3 mg succinic anhydride (SA, 10.3 µmole) was added. The pH was adjusted to 7.0 using diisopropylethylamine (DIPEA), and the reaction was carried out for 3 h at room temperature (RT). Obtained SAPOL3026 was purified on reversephase high-performance liquid chromatography (RP-HPLC) as described earlier [28]. SAPOL3026 eluted at~41.5 min was collected, evaporated, dissolved in deionized water, and lyophilized, yielding 0.026 g of the product as a white powder, which was used for further experiments. The resulting SAPOL3026 was analyzed by electrospray ionization mass spectrometry. Theoretical chemical formula: C 99 H 144 N 32

Synthesis of CXCR4-Targeted G5-X4 and Control G5-Ctrl Dendrimers
Preparation of the targeted and control dendrimers involved a multi-step synthesis as presented in Figure 1B: 0.205 g of G5(NH2)128 dendrimer (MW = 28,826 Da, 6.94 µmole) was dissolved in 10 mL DMSO, and 1 mL DMSO containing 4 mole equivalent of DOTA-NHS-ester (0.021 g, 27.7 µmole) was added with vigorously stirring. The resulting reaction mixture was stirred at RT for 4 h, and the solvent was removed on a rotary evaporator. The obtained residue was dissolved in 15 mL of PBS and transferred to an Amicon centrifugal filter with 10 kDa molecular weight cut-off (MWCO), which was centrifuged for 30 min at 4000× g and RT. This process was repeated twice with PBS buffer and 6 times of deionized water. The purified product was dissolved in deionized H2O and lyophilized that provided 0.218 g of G5(NH2)124(DOTA)4 dendrimer (2). In the next step, 0.18 g of 2 (6.51 µmole) was reacted with 10 mole equivalent of Fmoc-NH-PEG2000-NHS-ester (0.13 g) in 10 mL of DMSO for 4 h, followed by evaporation of the solvent on a rotary evaporator

Synthesis of CXCR4-Targeted G5-X4 and Control G5-Ctrl Dendrimers
Preparation of the targeted and control dendrimers involved a multi-step synthesis as presented in Figure 1B: 0.205 g of G5(NH 2 ) 128 dendrimer (MW = 28,826 Da, 6.94 µmole) was dissolved in 10 mL DMSO, and 1 mL DMSO containing 4 mole equivalent of DOTA-NHS-ester (0.021 g, 27.7 µmole) was added with vigorously stirring. The resulting reaction mixture was stirred at RT for 4 h, and the solvent was removed on a rotary evaporator. The obtained residue was dissolved in 15 mL of PBS and transferred to an Amicon centrifugal filter with 10 kDa molecular weight cut-off (MWCO), which was centrifuged for 30 min at 4000× g and RT. This process was repeated twice with PBS buffer and 6 times of deionized water. The purified product was dissolved in deionized H 2 O and lyophilized that provided 0.218 g of G5(NH 2 ) 124 (DOTA) 4 dendrimer (2). In the next step, 0.18 g of 2 (6.51 µmole) was reacted with 10 mole equivalent of Fmoc-NH-PEG 2000 -NHS-ester (0.13 g) in 10 mL of DMSO for 4 h, followed by evaporation of the solvent on a rotary evaporator and purification as described above. Then, 0.281 g of G5(NH 2 ) 116 (DOTA) 4  the pH of the reaction mixture was adjusted to 10 with DIPEA. After stirring for 16 h at RT, DMSO was evaporated, and the G5(NH 2 ) 52 (DOTA) 4 (PEG 2000 -NFmoc) 4 (mPEG 2 ) 64 conjugate was purified using ultrafiltration as described above. Then, the Fmoc protecting group was removed from PEG 2000 using 50% piperidine/DMF solution and purified using Amicon centrifugal filters as described above to obtain non-targeted control G5-Ctrl dendrimer conjugated with an average of 4 DOTA molecules, 62 mPEG 2 capping agents, and 8 PEG 2000 -NH 2 moieties (based on the main generation), as measured by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS). G5-Ctrl was further conjugated SAPOL3026·to generate a CXCR4-targeted G5-X4 dendrimer. Then, 0.011 g of SAPOL3026 (3.96 µmole) was dissolved in 2 mL DMF followed by the addition of 2 mole equivalent of EDC and HOBt to activate the carboxyl group. After 30 min, DMSO containing 0.02 g of G5-Ctrl (0.39 µmole) was added, and pH was adjusted to 7.4 using DIPEA. The resulting reaction mixture was stirred at room temperature for 16 h, followed by evaporation of the solvent on a rotary evaporator and purification of G5-X4 as described above. Starting G5(NH 2 ) 128 dendrimer, intermediated products, and final products were characterized by MALDI-TOF MS, reverse phase high performance liquid chromatography (RP-HPLC), 1 H NMR, dynamic light scattering, and zeta potential analysis (Supporting Information ).

Matrix-Assisted Laser Desorption Ionization-Time-of-Flight
MALDI-TOF spectra were acquired on a Voyager DE-STR spectrophotometer. Matrix (2,5-dihydroxybenzoic acid), starting dendrimers, and synthesized conjugates were dissolved in a 50% MeOH aqueous solution containing 0.1% TFA. Equal volumes of the matrix (10 µL, 20 mg/mL) and dendrimers (4 mg/mL) were mixed, and 1 µL of the resulting mixture was spotted on the target plate, evaporated, and used for data collection.

Dynamic Light Scattering (DLS) and Zeta Potential (ZP)
DLS and zp analyses were carried out using a Malvern Zetasizer Nano ZEN3600. G5(NH 2 ) 128 dendrimer and all synthesized conjugates were prepared in 1xPBS at a concentration of 2 mg/mL Results in Table 1 are a mean of three sequential measurements of number weighted size distribution, as it most accurately represents the size of nanoparticles present in the solution and their content.

Cell Lines
All cell culture reagents were obtained from Invitrogen (Halethorpe, MD, USA) unless otherwise indicated. The CHO (Chinese hamster ovary), U87 (human glioblastoma), H1155 (non-small cell lung cancer (NSCLC), and H69 (human small cell lung cancer SCLC) cell lines were purchased from American Type Culture Collection (Manassas, VA, USA). A human U87 cell line stably transfected with human CXCR4 (U87-stb-CXCR4) and CD4 was purchased from the NIH AIDS Research Reference Reagent Program [29]. All cell lines were cultured per manufacturer protocols. The CHO cell line with stable expression of SNAP-CXCR4 (CHO-SNAP-CXCR4) was generated in our laboratory and maintained as described earlier [30].

Competitive Binding Assays
The ability of G5-X4 and G5-Ctrl dendrimers to inhibit fluorescently labeled CXCL12-Red binding to CXCR4 was evaluated using frequency resonance energy transfer (FRET) assay as described previously [30]. Concentrations of G5-Ctrl and G5-X4 ranged from 0.1 pM to 10 µM. IC 50 values were derived by fitting the data to sigmoidal dose-response curve using Prism 8 Software (GraphPad) and converted to K i using the Cheng-Prusoff equation with K D = 19 for CXCL12-Red and its concentration of 15 nM.

Flow Cytometry
H69 small cell lung cancer (SCLC) cells were detached from the growing flask using a non-enzymatic cocktail (Gibco) when confluency reached 50-70%. Detached cells were washed twice with 5 mL of 1xPBS buffer containing 2 mmol EDTA and 0.5% FBS, which was further used for flow cytometry. CXCR4 expression was evaluated using the allophycoerythrin -conjugated anti-human CXCR4 monoclonal antibody (clone12G5, R&D Systems) per manufacturer instructions. After immunostaining of CXCR4, H69 cells were examined on a FACSCalibur flow cytometer (Becton Dickinson, NJ, USA). Obtained flow cytometry results were analyzed using FlowJo software.

Inhibition of Chemotaxis
The effect of G5-X4, G5-Ctrl, and POL3026 on CXCR4/CXCL12-mediated chemotaxis of H69 small cell lung cancer (SCLC) cells with high CXCR4 expression [31] was evaluated using a CytoSelect cell migration assay (Cell Biolabs, Inc. San Diego, CA, USA) according to the manufacturer's protocol. Briefly, 50,000 cells were suspended in 100 µL RPMI containing G5-X4, G5-Ctrl and POL3026 at concentration of 10 or 100 nM and transferred into insert wells. The insert wells were placed inside the harvesting wells containing RPMI with 100 nM CXCL12 (Peprotech, Rocky Hill, NJ, USA, product # 300-28A). After 17 h incubation at 37 • C, cells that migrated in the harvesting wells were lysed, CyQuant GR dye was added and the fluorescence was measured using Perkin Elmer-2480 Automatic Gamma Counter (PerkinElmer, Waltham, MA, USA).

Radiolabeling
The radiolabeling of G5-Ctrl and G5-X4 with 111 In (t 1/2 = 2.81 days) was carried out in 0.1 M sodium acetate (pH 4.5) for 1 h at 37 • C. To remove loosely bound and unchelated 111 In 3+ , ethylenediaminetetraacetic acid (EDTA) was added to the reaction mixture to achieve a concentration of 5 mM, followed by an additional 5 min incubation at RT and purification on a Zeba™ spin desalting column with 7 kDa MWCO (0.5 mL), then pre-equilibrated with PBS. The radiochemical purity of the resulting [ 111 In]G5-Ctrl and [ 111 In]G5-X4 was 98% as determined using instant thin-layer chromatography (ITLC) with citrate-phosphate-dextrose solution as the mobile phase.

Animal Models
All experimental procedures using animals were conducted according to protocols approved by the Johns Hopkins Animal Care and Use Committee. A total of 35 female nonobese diabetic/severe combined immunodeficiency NOD/SCID or 5 NSG mice, from six to eight weeks old, were purchased from The Johns Hopkins Immune Compromised Animal Core. NOD/SCID mice were implanted subcutaneously (s.c.) with U87 and U87-stb-CXCR4 cells (4 × 10 6 cells/100 µL) in the opposite flanks. Animals were used for ex vivo biodistribution and SPECT/CT imaging studies when the tumor size reached approximately 200-400 mm 3 . An orthotopic mouse model of non-small cell lung cancer was generated by injecting H1155 cells into the left lung of NSG mice using the following surgical procedure. Mice were anesthetized with 2% vaporized isoflurane. An approximately 1 cm incision was made on the skin near the left scapula, and the costal layer was exposed by separating the thoracic muscles. One million H1155 cells (in 30 µL HBSS with 50% matrigel) were injected into the left lung through the intercostal space, using a 29 G needle. Skin incisions were sutured, antibiotics were applied topically, and mice were kept on a heat-pad under observation until fully recovered from anesthesia. Tumor growth was monitored with computed tomography (CT) imaging recorded using an X-SPECT small animal SPECT/CT system (Gamma Medica Ideas, Northridge, CA, USA) and 512 projections.

SPECT/CT Imaging and Analysis
For whole-body SPECT/CT imaging, NOD/SCID mice bearing U87-stb-CXCR4 and U87 tumors were intravenously injected with approximately 12.95 MBq (350 uCi) of [ 111 In]G5-X4, and images were acquired on an X-SPECT small animal SPECT/CT system (Gamma Medica Ideas, Northridge, CA, USA) as described previously [32]. Prior imaging mice were anesthetized with 2% and maintained under 1.5% of isoflurane (v/v) during the acquisition of images. Images were acquired 2 and 8 h after radiotracer injection using 64 projections over 360 • at 45 s per projection and medium energy pinhole collimators. CT was recorded in 512 projections to allow anatomic co-registration. Data were reconstructed using the ordered subsets-expectation maximization algorithm, and 3D-volume-rendered images were generated using Amira 5.3.0 software (Visage Imaging Inc. San Diego, CA, USA). The same procedure was used to monitor the accumulation of [ 111 In]G5-Ctrl in mouse bearing orthotopic H1155 NSCLC tumor.

Data Analysis
Statistical analysis was performed using a Graphpad Prism 8 software and an unpaired two-tailed t-test. When p < 0.05, the difference between the compared groups was considered statistically significant.

Synthesis and Physicochemical Characterization of G5-Ctrl and G5-X4
Syntheses of the CXCR4-targeted (G5-X4) and control (G5-Ctrl) dendrimers are presented in Figure 1. Amine-terminated generation-5 PAMAM dendrimer G5-(NH 2 ) 128 was consecutively conjugated with, on average based on main generation dendrimers, four DOTA chelators (1) to enable radiolabeling, eight PEG 2000 -NH-Fmoc linkers (2) for attachment of targeting moieties and 62 m-PEG 2 short capping agents (3) to reduce net surface positive charge, followed by deprotection of PEG 2000 -NH-Fmoc (4), which generated the G5-Ctrl dendrimers. G5-Ctrl was subsequently conjugated with three SA-POL3026 peptidomimetics (5) to formulate the CXCR4-targeted G5-X4 dendrimers. To enable covalent attachment of POL3026 with G5-Ctrl, the lysine side chain of the peptide was selectively carboxylated with succinic anhydride. Resulting SA-POL3026 was conjugated to a PEG 2000 -NH 2 linker via carbodiimide coupling. Starting dendrimer and synthesized conjugates were characterized by MALDI-TOF, RP-HPLC, 1 HNMR, DLS, and zeta potential (Table 1 and Supporting Information Figures S1-S4). Commercially available PAMAM dendrimers contain three different species trailing generation, main generation, and dimers [33,34]. All three species were detected in the spectrum of the dendrimers that we used as starting material (Supporting Information Figure S1), with the most pronounced peak for main generation-5 dendrimer at 26,111 Da. Trailing generation and dimers were detected at 12,600 and 50,950 Da, respectively. An increase in the molecular weights indicated conjugation of on average two, four, and five DOTA molecules with trailing generation, main generation, and dimers, respectively. Starting from conjugates III trailing generation, main generation-5 and dimers could not be resolved, showing a broad peak indicating a wide distribution of conjugates present in the sample. A shift of the highest signal intensity detected for this peak was used to calculate the average number of covalently attached PEG 2000 , mPEG 2 , and SA-POL3026 moieties. In agreement with MALDI-TOF, RP-HPLC showed the presence of trailing generation, main generation, and dimers with relative percent of 3.5, 90.9, and 5.6%, respectively, in dendrimers used as starting material (Supporting Information Figure S2). RP-HPLC of G5-Ctrl and G5-X4 showed a significant broadening of the peaks and shift to longer retention with no separation of trailing generation, main generation, and dimers, confirming broad distribution of the conjugates. 1 H NMR further confirmed modifications of dendrimers (Supporting Information Figure S3), DLS analysis indicated an increase in dendrimer size upon PEGylation by approximately 3 nm and a significant reduction in the net surface positive charge (Supporting Information Figure S4). G5-Ctrl and G5-X4 dendrimers had size distribution with 7.81 ± 0.85 nm and zeta potential of 12.41 ± 1.43 mV and 9.11 ± 0.92 nm and zeta potential of 14.93 ± 0.51 mV, respectively.

Radiolabeling of G5-Ctrl and G5-X4
To evaluate the pharmacokinetics of the G5-Ctrl and G5-X4 dendrimers, they were radiolabeled with 111 In. Resulting [ 111 In]G5-Ctrl and [ 111 In]G5-X4 radiotracers were prepared with a specific activity of 2 mCi/mg ± 0.1 mCi and radiochemical purity > 95% as confirmed by instant thin-layer chromatography.

Evaluation of In Vitro Specificity
To evaluate the CXCR4 specificity of [ 111 In]G5-X4, we performed in vitro binding assays using U87-stb-CXCR4 and U87 cell lines with high and low expression of CXCR4, respectively ( Figure 2B), as described earlier [35].

Radiolabeling of G5-Ctrl and G5-X4
To evaluate the pharmacokinetics of the G5-Ctrl and G5-X4 dendrimers, they were radiolabeled with 111 In. Resulting [ 111 In]G5-Ctrl and [ 111 In]G5-X4 radiotracers were prepared with a specific activity of 2 mCi/mg ± 0.1 mCi and radiochemical purity > 95% as confirmed by instant thin-layer chromatography.
[ 111 In]G5-Ctrl in the circulation declined by over 120 h. In agreement with biodistribution, SPECT/CT imaging showed higher uptake of [ 111 In]G5-X4 in U87-stb-CXCR4 tumors compared to U87 tumors and high background ( Figure 3D). In mice that received a blocking dose of POL3026 1 h prior to administration of radiolabeled dendrimers, accumulation of [ 111 In]G5-X4 in U87-stb-CXCR4 tumors decreased from 8.2 ±2.1 to 5.05 ± 0.32%ID/g ( Figure 4A,B, p < 0.01), which was similar to radioactivity levels detected in U87 tumors and to [ 111 In]G5-Ctrl accumulation in U87 and U87-stb-CXCR4 tumors. The blocking dose of POL3026 did not affect the blood pool concentration and overall biodistribution of [ 111 In]G5-X4 and [ 111 In]G5-Ctrl. We also evaluated the accumulation of [ 111 In]G5-Ctrl in mice bearing orthotopic H1155 NSCLC tumors ( Figure S6). SPECT/CT images indicated higher [ 111 In]G5-Ctrl uptake in the tumor compared to lungs over 120 h. Ex vivo biodistribution confirmed a relatively high accumulation of [ 111 In]G5-Ctrl in the tumor with 21.49%ID/g.

Discussion
PAMAM dendrimers can be readily tailored to different sizes and compositions depending on the application. Here, we focused on the formulation and evaluation of CXCR4-targeted PAMAM dendrimers. Pharmacokinetics, biodistribution, and tumor targeting of PAMAM dendrimers are strongly influenced by their size and surface properties [36][37][38]. PAMAM dendrimers smaller than 5 nm (up to generation-4) form flexible scaffolds that are rapidly excreted by renal filtration [36][37][38]. Starting from generation 5 (size ≥ 5 nm), dendrimers have more rigid and globular structures and exhibit longer circulation times with hepatobiliary excretion in addition to renal clearance [39]. To synthesize control and CXCR4-targeted dendrimers, we have utilized generation-5 PAMAM dendrimers. To extend circulation for enhanced tumor accumulation and to reduce uptake by the reticuloendothelial system (RES), dendrimers were extensively PEGylated with PEG2000 and mPEG2, which increased their size by approximately 3 nm. As a targeting moiety, we have used the POL3026 peptidomimetic, as it was initially reported by our group to have suitable pharmacokinetics for in vivo targeting and imaging of CXCR4 [28]. G5-X4 dendrimers showed low nanomolar in vitro CXCR4 affinity, with an IC50 value comparable to previously reported for PEGylated POL3026 [28]. G5-X5 also demonstrated superior inhibition of H69 SCLC cell migration toward CXCL12 compared to PLO3026 and G5-Ctrl, confirming the potential therapeutic applications of this construct. In vitro CXCR4 specificity of G5-X4 was further reflected by higher uptake of [ 111 In]G5-X4 in U87-stb-CXCR4 cells compared to U87 cells, which was blocked by unmodified POL3026. In contrast, G5-Ctrl nanoparticles neither inhibited the interaction between CXCR4 and CXCL12 nor showed preferential uptake by U87-stb-CXCR4 cells. Our biodistribution results showed that accumulation of [ 111 In]G5-X4 in U78stb-CXCR4 tumors was nearly 50% higher compared to U87 tumors and then [ 111 In]G5-Ctrl in both tumors at 1 h and 3 h after injection, suggesting faster uptake of targeted

Discussion
PAMAM dendrimers can be readily tailored to different sizes and compositions depending on the application. Here, we focused on the formulation and evaluation of CXCR4targeted PAMAM dendrimers. Pharmacokinetics, biodistribution, and tumor targeting of PAMAM dendrimers are strongly influenced by their size and surface properties [36][37][38]. PAMAM dendrimers smaller than 5 nm (up to generation-4) form flexible scaffolds that are rapidly excreted by renal filtration [36][37][38]. Starting from generation 5 (size ≥ 5 nm), dendrimers have more rigid and globular structures and exhibit longer circulation times with hepatobiliary excretion in addition to renal clearance [39]. To synthesize control and CXCR4-targeted dendrimers, we have utilized generation-5 PAMAM dendrimers. To extend circulation for enhanced tumor accumulation and to reduce uptake by the reticuloendothelial system (RES), dendrimers were extensively PEGylated with PEG 2000 and mPEG 2 , which increased their size by approximately 3 nm. As a targeting moiety, we have used the POL3026 peptidomimetic, as it was initially reported by our group to have suitable pharmacokinetics for in vivo targeting and imaging of CXCR4 [28]. G5-X4 dendrimers showed low nanomolar in vitro CXCR4 affinity, with an IC 50 value comparable to previously reported for PEGylated POL3026 [28]. G5-X5 also demonstrated superior inhibition of H69 SCLC cell migration toward CXCL12 compared to PLO3026 and G5-Ctrl, confirming the potential therapeutic applications of this construct. In vitro CXCR4 specificity of G5-X4 was further reflected by higher uptake of [ 111 In]G5-X4 in U87-stb-CXCR4 cells compared to U87 cells, which was blocked by unmodified POL3026. In contrast, G5-Ctrl nanoparticles neither inhibited the interaction between CXCR4 and CXCL12 nor showed preferential uptake by U87-stb-CXCR4 cells. Our biodistribution results showed that accumulation of [ 111 In]G5-X4 in U78-stb-CXCR4 tumors was nearly 50% higher compared to U87 tumors and then [ 111 In]G5-Ctrl in both tumors at 1 h and 3 h after injection, suggesting faster uptake of targeted dendrimers facilitated by CXCR4. These findings were also supported by SPECT/CT imaging, showing a higher uptake of radioactivity in U87-stb-CXCR4 tumors vs. U87 tumors at early time-points in mice treated with [ 111 In]G5-X4. This initially increased uptake of [ 111 In]G5-X4 in U87-stb-CXCR4 tumors could be blocked by POL3026, further confirming interactions of targeted dendrimers with CXCR4. [ 111 In]G5-X4 showed similar uptake in CXCR4 positive and negative tumors at comparable levels detected for [ 111 In]G5-Ctrl at 24 h that is most likely attributed to prolonged circulation and significant passive tumor targeting due to the enhanced permeability and retention (EPR) effect [40].
[ 111 In]G5-X4 showed higher accumulation in U87-stb-CXCR4 tumors compared to U87 tumors and [ 111 In]G5-Ctrl at 48 and 120 h after injection, suggesting that interactions of targeted dendrimers with CXCR4 could prolong their tumor retention. Similar trends with significant passive and minor active tumor targeting were demonstrated for prostatespecific membrane antigen (PSMA)-targeted polylactic acid (PLA) and non-targeted control nanoparticles with a size of approximately 100 nm [41]. In contrast, generation-4 PAMAM dendrimer-conjugated lysine-glutamate-urea derivative for PSMA targeting with a size of approximately 5 nm showed mostly active targeting, reflected by significantly higher uptake in PSMA + PC3 PIP tumors compared to PSMA -PC3 flu tumors [35]. [ 111 In]G5-X4 and [ 111 In]G5-Ctrl also showed persistent accumulation in the liver and spleen, indicating their uptake by the RES system despite extensive PEGylation. In the lungs, heart, and kidneys, initial high concentration of [ 111 In]G5-X4 and [ 111 In]G5-Ctrl gradually decreased as the blood pool radioactivity declined. We have recently demonstrated that non-targeted generation-4 hydroxy terminated PAMAM dendrimers with size around 5 nm below renal filtration cut-off quickly cleared from bloodstream with major accumulation in kidneys, followed by liver and very low (below 1% ID/g) uptake in PSMA + PC3 PIP and PSMA − PC3 flu prostate tumor models [35]. In contrast, our [ 111 In]G5-Ctrl control dendrimers with size averaging around 8 nm showed significantly longer circulation and substantial passive tumor accumulation, reaching 10.06 ± 1.46%ID/g at 48 h after injection in U87-stb-CXCR4 and U87 tumors. Biodistribution of [ 111 In]G5-Ctrl is similar to the previous report on hydroxy-terminated generation-7 PAMAM dendrimers with size around 7 nm, which showed approximately 7 %ID/g accumulation in orthotopic A2780 ovarian tumors due to passive targeting [39]. Taken together, our results indicate that an increase in dendrimer size by only 3 nm has a profound effect on their pharmacokinetics and passive tumor accumulation. The tumor accumulation of non-targeted dendrimers with a size of approximately 8 nm ranges between 7 and 10 %ID/g, which is comparable to the tumor uptake of PEGylated liposomes or PLA micelles with sizes of 120 nm and 100 nm, respectively [41,42]. Our results also demonstrated relatively high uptake of [ 111 In]G5-Ctrl in the orthotopic H1155 NSCLC mouse model, further supporting its potential in passive tumor targeting for imaging and therapeutic applications.

Conclusions
Our results demonstrate nearly a two-fold enhancement of targeted dendrimers in CXCR4-positive tumors, at earlier time points, compared to control tumors. That uptake differential decreased substantially 24 h after injection. The pharmacokinetics and CXCR4-targeted active tumor uptake of PAMAM dendrimers can be improved through optimization of size and surface properties. In the future, such constructs may be outfitted with imaging agents, drugs, or nucleic acids to detect and treat cancer.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/pharmaceutics14030655/s1, Figure S1. MALDI-TOF spectra recoded for starting dendrimers, inter-mediate products and final conjugates, showing changes in molecular weight upon modification step, which were used to calculated average number of conjugated moieties; Figure S2. RP-HPLC chromatograms of: A-initial G5(NH2)128 dendrimer showing presence of trailing generation (3.5%), main generation (90.9%) and dimeric dendrimers (5.6%); B-G5-Ctrl and C-G5-X4 indicating significant widening of the peak due to presence of broad species distribution in the samples that could not be resolved. Analysis was carried out using previously reported method [35]; Figure S3. Proton-NMR study. A and B-1H NMR spectra of the initial G5(NH2)128 dendrimer. Insert-partial structure of PAMAM dendrimer (n = 4) showing theoretical number of CH2 groups. Assignment of the signals was done based previously reported data [34] and values of integrals are in good agreement with the number of 1H in each group. C-1H NMR spectrum of G5(NH2)124(DOTA)4 (conjugate 2, Figure S1 reaction scheme) showing broadening and shift of the dendrimer signals in particular related to protons V and VI that are close to terminal NH2 groups and 2 additional peaks at 3.05 and 3.5 ppm upon conjugation of on average 4 molecules of DOTA with G5(NH2)128 den-drimer. D-1H NMR spectrum of G5(NH2)116(DOTA)4(PEG2000-Fmoc)8 (conjugate 3, Figure S1 reac-tion scheme) demonstrating dominant peak related to PEG2000-Fmoc at 3.62 ppm and additional peaks of Fmoc in the aromatic region (insert) and significant broadening of dendrimer signals. E-1H NMR spectrum of G5(NH2)54(DOTA)4(PEG2000)8(mPEG2)62 (conjugate 4, Figure S1 reaction scheme) showing additional strong intensity signal at 3.28 ppm related to mPEG2 and further broadening the other peaks. Due to the presence of broad and overlapping signals, their assignment is not feasible. F-1H NMR spectrum of G5(NH2)54(DOTA)4(PEG2000)8(mPEG2)62 (G5-Ctrl, Figure S1 reaction scheme) indicating deprotection of PEG2000 by the absence of peaks in the aromatic region. All spectra were recorded using D2O as a solvent and a Bruker Avance III 500 MHz NMR spec-trometer; Figure S4. Representative number weighted size distribution acquired for starting den-drimers and all synthetized conjugates, demonstrate an increase in dendrimer size after PEGylation; Figure S5