An Orthogonal Protection Strategy for Synthesizing Scaffold-Modifiable Dendrons and Their Application in Drug Delivery.

Dendrons have well-defined dendritic structures. However, it is a great challenge to preserve their high structural definition after multiple functionalization because the site-selective conjugation of different functional molecules is quite difficult. Scaffold-modifiable dendrons that have orthogonal reactive groups at the scaffold and periphery are ideal for achieving the site-specific bifunctionalization. In this paper, we present a new strategy for synthesizing scaffold-modifiable dendrons via orthogonal amino protection and a solid-phase synthesis method. This strategy renders the reactive sites at the scaffold and periphery of the dendrons a super selectivity, high reactivity, and wide applicability to various reaction types. The fourth-generation dendrons can be facilely synthesized within 2 days without structural defects as demonstrated by mass spectrometry. We conjugated doxorubicin (DOX) and phenylboronic acid (PBA) groups to the scaffold and periphery, respectively. Thanks to the PBA-enhanced lysosome escape, tumor targeting ability, and tumor permeability as well as the high drug loading content larger than 30%, the dendron-based prodrug exhibited extraordinary antitumor efficacy and could eradicate the tumors established in mice by multiple intravenous administration. This work provides a practical strategy for synthesizing scaffold-modifiable dendrons that can be a promising nanoplatform to achieve function integration in a precisely controlled manner.

Introduction

Dendrimers and dendrons are two types of monodisperse dendritic polymers. Dendrons can be looked as the wedge-shaped units of dendrimers.[1−3] Both dendrimers and dendrons have intriguing structural characteristics, such as a well-defined and widely tunable chemical structure, highly branched topology, nanoscale size, globular shape, abundant peripheral groups and available internal cavities, etc., which make them very promising in biomedical applications, for example, as nanocarriers for drugs or imaging contrast agents.[4−11] Their precisely controllable and widely tunable chemical structure not only ensures perfectly reproducible pharmacokinetic and pharmacological profiles but also renders versatile and tailorable biological functionalities.[12] Their abundant groups and large internal cavities provide plenty of sites for drug loading. Their low-nanometer size is greatly favorable for deep tissue penetration.[13−15]

It is notable that, for traditional dendrimers or dendrons, their scaffold is not chemically modifiable due to the absence of reactive groups, which greatly limits their expansibility in structure and functionality. When multifunctionalization is needed, their peripheral groups have to be shared by different functional moieties such as drugs, targeting groups, contrast agents, etc., which would place restrictions on the content of each functional moiety and also cause structural heterogeneity, leading to indefinite activity. Furthermore, when used in living systems, the functional molecules anchored onto the periphery of dendrimers or dendrons may suffer from enzymatic decomposition in the circulatory system and lose their functions before reaching target tissues. Encapsulation in their internal cavity can protect the cargoes; however, high drug loading and controllable drug release are still challenging.[16] Taken together, introducing reactive groups to the internal repeating units of dendrimers or dendrons is an effective way to achieve site-specific multifunctionalization and attain the optimized placement of the enzyme-labile function molecules. So far, only a very limited number of studies have involved the dendrimers or dendrons bearing reactive groups at the scaffold. The reactive groups include hydroxyl groups introduced by using an epoxy–amine reaction,[17−19] and alkynyl groups[20−22] and carboxylic ester groups[23] introduced, separately, by multifunctional building blocks. These synthesis methods are generally complicated, time-consuming, and tedious. Therefore, an efficient synthesis strategy for internally functionalized dendrimers or dendrons as well as their application potentials in the biomedical field are worth further exploration and development.

In this paper, we present the synthesis of a novel class of scaffold-modifiable dendrons up to the fourth generation (G4) via an orthogonal amino protection strategy and a solid-phase synthesis method. The dendron G4 can be synthesized within 2 days. This synthesis strategy for scaffold-modifiable dendrons makes the reaction sites at the scaffold and periphery highly selective, reactive, and widely applicable to various reaction types. Moreover, by using the dendrons as drug carriers, we achieved extraordinary tumor treatment efficacy. In the synthesis of the dendrons, we introduced a benzoyloxycarbonyl (Cbz)-protected acylhydrazine group in each repeating unit and a tert-butyloxycarbonyl (Boc)-protected amino group at each end (Scheme ). The Cbz and Boc protecting groups could be selectively cleaved, and hence, controlled bifunctionalization of the dendrons could be achieved specifically at the scaffold and periphery. After the selective deprotection, we conjugated the antitumor agent doxorubicin (DOX) to the scaffold and the poly(ethylene glycol) (PEG)-bound phenylboronic acid (PBA) group to the periphery, respectively, since we previously demonstrated that PBA modification could effectively facilitate the lysosome escape of nanomaterials.[24] Thanks to PBA-enhanced lysosome escape, a tumor targeting ability, and tumor permeability as well as the high drug loading content larger than 30%, high antitumor efficacy was achieved. After multiple intravenous administration, the tumors established in mice were eradicated.

Scheme 1

Synthetic Routes of Dendron G4

Results and Discussion

Synthesis and Characterizations of Dendrons

To achieve the controlled bifunctionalization of dendrons, we designed a building block named M1 with an AB2C2-type structure containing one carboxylic acid group, two Boc-protected amino groups, and two Cbz-protected acylhydrazine groups. As shown in Scheme , M1 was synthesized by the amidation between 2-azido-6-tert-butoxycarbonylamino-hexanoic acid [azido-Lys(Boc)–OH, B, derived from lysine] and carbobenzoxyhydrazide (C) followed by coupling the obtained azido-bearing intermediate with 2,2-bis (propargyl) propionic acid (A) via Cu(I)-catalyzed 1,3 dipolar azide–alkyne cycloaddition (CuAAC). The detailed synthesis procedures and characterization data for M1 and its precursors can be found in the Supporting Information text and Figures S1–S4.

With M1 in hand, we synthesized dendrons of G1–G4 by a continuous solid-phase synthesis method using oxime resin as a solid support. For the first incorporation on oxime resin via the esterification reaction, M1 was controlled at a relatively low loading of ∼0.15 mmol/g to reduce the steric hindrance in the subsequent branching processes.[25,26] After the first incorporation, the terminal Boc groups were selectively removed by trifluoroacetic acid (TFA) to liberate the amines, making them available for M1 coupling via the amidation mediated by benzotriazol-1-yloxy(tripyrrolidin-1-yl) phosphonium hexafluorophosphate (PyBOP). The completion of each coupling reaction was driven by feeding largely excessive M1 and was confirmed by a ninhydrin assay to ensure the defect-free chemical structures of the dendrons.[27] By repeating the Boc cleavage and M1 coupling, stepwise growth of dendrons was achieved. All of the target dendrons of successive generations were finally cleaved from the resin by propargylamine to introduce an alkynyl group at the focal. The dendron G4 can be synthesized within 2 days. A desired yield for each generation was attained as shown in Table .

Table 1

Molecular Weights, Synthesis Periods, and Yields of the Dendrons G1–G4

	numbers of Cbz/Boc groups	calculated molecular weight (g/mol)	m/z determined by MALDI-TOF MS	synthesis period	yield
G1	2/2	1088.234	1110.773 [M + Na]+	15 h	99%
G2	6/4	2954.308	2977.006 [M + Na]+	22 h	95%
G3	14/8	6686.456	6709.719 [M + Na]+	27 h	86%
G4	30/16	14 150.752; 12 548.880 (after Boc cleavage)	6285.251 [M + Na + H]2+ (after Boc cleavage)	32 h	74%

The chemical structures and purities of the dendrons were characterized by proton and carbon nuclear magnetic resonance (1H and 13C NMR), matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS), and gel permeation chromatography (GPC) (Figure and Figures S5–S12). The 1H NMR spectra of dendrons G1–G4 are presented in Figure A,B. As shown in Figure A, the proton signals of G1 are clearly identified, and the presence of several signals arising from Boc (∼1.36 ppm), Cbz (∼7.36 ppm, aromatic protons), and triazole (∼8.16 ppm) groups reflects its structural characteristics. The spectra of G2–G4 are similar to that of G1 (Figure B). The integral intensity ratios of the signals in each spectrum are well consistent with the corresponding calculated values. Typically, it can be observed that the integral intensity ratio of the signal from the Cbz group (∼5.09 ppm) to the signal from the methylene in the focal propargyl group (∼3.82 ppm) increases with the generation number in an exponential manner, conforming with the structural evolution of the dendrons.

Figure 1

(A) 1H NMR spectrum of dendron G1 in DMSO-d6 with signal identification. (B) 1H NMR spectra of dendrons G1–G4 in DMSO-d6. The asterisked signals originate from water and diethyl ether. MALDI-TOF MS spectra (C) and GPC curves (D) of dendrons G1–G4.

The accurate formations of the dendrons can be definitely determined by MALDI-TOF MS. As shown in Figure C, the MALDI-TOF MS spectrum of each dendron from G1 to G3 shows a strong peak from the cation adduct (Na+ adduct) of the target dendron, and no signals from the incompletely reacted species are observed, indicating their defect-free structures. In the case of dendron G4, the ionization is very difficult due to its large molecular weight (14 150 Da). We tried different matrix and ionizing agents and failed to obtain a desirable spectrum. We turned to test the MALDI-TOF MS of the amine-terminated G4 after Boc cleavage. A signal from [M + Na + H]2+ was detected, demonstrating its precise structure. Further, we used GPC to examine the size evolution of the dendrons with generation increasing (Figure D). As can be seen, the retention time decreases with the increase of generation, and each generation exhibits a monomodal molecular weight distribution with low polydispersity indexes (Đ) ranging from 1.02 to 1.06 (Table S1), confirming their high purity and monodispersity.

Synthesis and Characterizations of Dendron-Based Prodrugs

To improve the hydrophilicity and in vivo properties of the dendrons, we conjugated PEG chains to their periphery before drug loading. The PEGylated dendron G3 named PEG–G3 was also synthesized by the solid-phase synthesis method (Figure A). A heterobifunctional PEG (Boc-PEG1k-NHS) bearing a Boc-protected amino group at one end and a N-hydroxysuccinimide (NHS) ester at the other end was used and conjugated to the periphery of the resin-bound dendrons via amidation. To achieve the tumor targeted therapy, we further introduced PBA groups to the periphery of the resin-bound dendrons via the amidation with 4-carboxyphenylboronic acid after the cleavage of the Boc protecting groups, since PBA has been well-demonstrated to be effective in improving the lysosome escape, tumor targeting ability, and tumor permeability of biomaterials.[24,28,29] The PBA-modified PEGylated dendron G3 was named PBA–PEG–G3 (Figure A). Propargylamine was also used to cleave the modified dendrons from the resin. In the following studies, we chose the dendron G3 to load drugs because it is more easily obtainable and its smaller size as compared to the dendron G4 is more favorable to tumor penetration. Before the acylhydrazine groups at the scaffold can be used to link DOX, the Cbz protecting groups have to be cleaved by catalytic hydrogenolysis. The complete cleavage of the Cbz in PEG–G3 is demonstrated by the 1H NMR spectra before and after the cleavage. As shown in Figure S13, after the cleavage of the Cbz, the proton signals from the benzyl group (∼5.09 ppm, ∼7.36 ppm) completely disappear, and the hydrazide proton signals shift from ∼9.40 to ∼8.45 ppm, indicating the complete removal of the protecting groups. This is also confirmed in the case of the PBA–PEG–G3 (Figure S14). We further verified the Cbz cleavage efficiency by measuring the MALDI-TOF MS of the deprotected dendron G3 without PEG segments since the wide molecular weight distribution of the PEG segments would interfere with the evaluation of the cleavage efficiency. As shown in Figure S15, compared to the spectrum of the dendron G3 before deprotection, the spectrum of the deprotected dendron shows a strong signal from the target molecule with all of the 14 Cbz groups cleaved, and no trace of the incompletely deprotected species is observed, confirming the high Cbz cleavage efficiency. After removing the Cbz protecting groups in PEG–G3, we conjugated DOX to the dendron scaffold through the reaction of the acylhydrazine groups with the carbonyl group in DOX, forming a pH-sensitive acylhydrazone linkage between DOX and the dendrons. The PBA-modified DOX-loaded and DOX-free G3 dendrons were named PDG3 and PG3, respectively. For a performance evaluation, the PBA-free DOX-loaded dendron G3 with Boc-protected amino groups at the periphery were used as a reference sample and named DG3, and the corresponding dendron G3 without PBA and DOX was named NG3 (Figure A).

Figure 2

(A) Synthetic routes of PDG3 and DG3. TEM and DLS data of PDG3 (B) and DG3 (C).

The drug loading contents of PDG3 and DG3 were determined to be about 31.8 ± 2.8 and 33.2 ± 2.4 wt % by UV–vis spectroscopy using a pre-established calibration curve, respectively. The high drug loading contents are attributable to the abundant acylhydrazine groups at the scaffold of the dendrons (14 acylhydrazine groups per dendron G3). It can be calculated that, on average, about 11.0 and 11.4 acylhydrazine groups in PDG3 and DG3 are occupied by DOX, respectively. Both PDG3 and DG3 have desirable solubility in an aqueous medium such as normal saline and fetal bovine serum (FBS) as evidenced by the photograph of their solutions in Figure S16A. From the UV–vis absorption spectra of PDG3 and DG3 in phosphate-buffered saline (PBS, 10 mM, pH = 7.4), it can be observed that the absorption band arising from the DOX moieties exhibits a bathochromic shift of ∼10 nm compared to the free DOX in the same medium, which should be associated with the aggregation state of the DOX moieties in the dendrons (Figure S16B). We further examined the morphological structures and sizes of PDG3 and DG3 by transmission electron microscopy (TEM) and dynamic light scattering (DLS) (Figure B,C). Based on the TEM observations, both PDG3 and DG3 exhibit a spherical morphology with a narrowly distributed diameter of 3.8 ± 0.4 nm, and their hydrodynamic diameters in water are determined to be 7.8 ± 1.0 nm by DLS. These results indicate that both PDG3 and DG3 exist in the unimolecular state in water. We also checked the size distributions and purities of the DOX-conjugated dendrons by GPC (Figure S17). The Đ values of PDG3 and DG3 are about 1.10 and 1.12, respectively, slightly larger than those of PG3 (1.07) and NG3 (1.06), which should result from the nonidentical number of DOX moieties in each dendron molecule. No peaks from impurities are observed in the GPC curves of the dendrons, indicating their high purities.

In Vitro Drug Release

The in vitro drug release behaviors of PDG3 and DG3 were studied by dialysis against PBS at different pH values of 7.4, 6.0, and 5.0 at 37 °C, separately. As shown in Figure S18, at pH 7.4, DOX release from either PDG3 or DG3 is very slow, and the cumulative released portions are about 26.3% and 25.3% within 48 h, respectively. By comparison, weak acidic conditions of pH 6.0 and 5.0 can trigger faster DOX release due to the acid-labile acylhydrazone linkage. Within 48 h, about 69.4% and 65.7% of DOX releases from PDG3 and DG3 at pH 6.0, respectively, while the values increase to about 81.1% and 81.3% at pH 5.0. The pH-sensitive drug release behaviors of PDG3 and DG3 are greatly favorable to reduce unwanted side effects and enhance antitumor efficacy since the drug release in blood circulation is remarkably suppressed, and tumor-specific drug release is basically realized.

In Vitro Cellular Uptake and Cytotoxicity

As is well-documented, PBA can rapidly and reversibly react with sialic acid (SA) residues that are overexpressed on various tumor cells and thus can augment the tumor targeting ability and cellular uptake.[28] We compared the cellular uptake behaviors of PDG3 and DG3 in the SA-positive murine hepatic H22 cells and human hepatoma HepG2 cells by using confocal laser scanning microscopy (CLSM) and flow cytometry. Figure A,B shows the CLSM images of the cells incubated with PDG3 and DG3 at 37 °C for 2 h, respectively. As can be seen, for both H22 and HepG2 cells, the red signals from DOX are mainly distributed in the cell cytoplasm, with a small fraction inside nuclei, indicating that, during the 2 h incubation, only a small part of DOX released and entered nuclei. The cells treated with PDG3 display stronger DOX fluorescence signals than the corresponding cells treated with DG3, which is confirmed by the quantitative analyses of flow cytometry. The mean fluorescence intensity (MFI) from the DOX moieties in the PDG3-treated H22 cells is ∼1.48-fold that in the DG3-treated H22 cells, and for HepG2 cells, the value is ∼1.96, signifying that PBA modification can indeed facilitate the cellular uptake. It is noticed that the enhancement effect of the PBA groups on the cellular uptake in HepG2 cells is higher than that in H22 cells, which is attributable to the higher SA expression level of HepG2 cells as compared to that of H22 cells.[30]

Figure 3

Typical CLSM images and MFIs measured by flow cytometry of the H22 (A) and HepG2 (B) cells after 2 h of incubation with PDG3 and DG3 at 37 °C. The cell nuclei were stained by Hoechst. Scale bars = 20 μm. Data represent mean values ± SD (n = 3). *** P < 0.005 (PDG3 versus DG3). In vitro cytotoxicities of PDG3, DG3, and DOX against H22 (C) and HepG2 (D) cells after 48 h of incubation. Data are presented as the mean ± SD (n = 5). ** P < 0.05, *** P < 0.005 (PDG3 versus DG3). (E) IC50 values calculated from the MTT assay data. (F) Typical CLSM images of the equatorial planes of the HepG2MCs coincubated with the FITC-labeled PG3 and RBITC-labeled NG3 together for 6 and 24 h. Scale bars = 50 μm. 2.5 D models of the fluorescence distributions and MFIs of PG3 and NG3 in the equatorial planes of the MCs after 24 h of incubation. *** P < 0.005 (PG3 versus NG3).

To evaluate the potential antitumor activity of the dendron-based prodrugs and the biosafety of the drug-free dendrons, we measured the in vitro cytotoxicities of PDG3, DG3, PEG–G3, and PBA–PEG–G3 against H22 and HepG2 cells by MTT [3-(4′,5′-dimethylthiazol-2′-yl)-2,5-diphenyltetrazolium bromide] assay with DOX as the positive control. As shown in Figure S19, after 24 h of incubation, the drug-free dendrons do not exhibit significant cytotoxicities at all of the test concentrations, indicating their good cytocompatibility. In either the H22 or HepG2 cell line, each of the DOX formulations displays a dose-dependent cytotoxicity (Figure C,D). Their IC50 values are calculated and compared in Figure E. The IC50 value increases in the order of DOX < PDG3 < DG3 in either the H22 or HepG2 cell line, which is ascribed to the sustained drug release of PDG3 and DG3 and the targeting effect of the PBA groups in PDG3.

Lysosome Escape

In-time lysosome escape is crucial for a nanomedicine to exert its function. To verify the PBA effect on the lysosome escape of the dendrons, we labeled PG3 and NG3 with fluorescein isothiocyanate (FITC) via the reaction between the acylhydrazine groups in the dendrons and the isothiocyanate group in the dye, incubated them with HepG2 cells, and determined their lysosome escape behaviors by fluorescence colocalization of the labeled dendrons and LysoTracker red via CLSM imaging and Pearson colocalization analyses (Figures S20 and S21). It is notable that, for the case of PG3, the Pearson’s correlation coefficient decreases rapidly from 0.62 at 2 h post-treatment to 0.11 at 24 h post-treatment; in comparison, for NG3, the coefficient decreases relatively slowly from 0.76 to 0.53 (Figure S21), indicating that the PBA groups in PG3 can indeed significantly promote its lysosome escape, which is very favorable to its therapy efficacy.

Penetration in Three-Dimensional (3D) Multicellular Spheroids (MCs)

Higher tumor permeability enables nanomedicines to penetrate deeper in avascular regions and hence imparts a higher therapy efficacy. Dendron-/dendrimer-based nanomedicines generally have excellent tumor permeability due to their low-nanometer sizes.[15] We studied the permeabilities of the dendrons PG3 and NG3 in HepG2MCs to simulate and evaluate their tumor penetration and the effect of PBA groups on it. To compare their penetration under exactly the same conditions, we labeled PG3 and NG3 with FITC and rhodamine B isothiocyanate (RBITC), respectively, and treated the MCs with the labeled PG3 and NG3 together. The DOX-free dendrons were used for eliminating the interference of the DOX fluorescence. After incubation for 6 and 24 h, separately, the MCs were observed by CLSM. As shown in the typical CLSM images (Figure F), time-dependent penetration behaviors of PG3 and NG3 are observed. After 24 h of incubation, both PG3 and NG3 can penetrate to the centers of the MCs, whereas, in the interior of the MCs, the fluorescence signals from PG3 are significantly stronger than those from NG3, which can be evidenced clearly by the 2.5 D model images and the MFIs from PG3 and NG3 in the equatorial planes of the MCs (Figure F), indicating that PBA groups can indeed facilitate the penetration of the dendrons in MCs, which can be explained by the PBA-enhanced transcellular transfer mechanism.[24]

Biodistribution in Tumor-Bearing Mice

The biodistribution of a nanomedicine is of paramount importance for its therapeutic effectiveness and biosafety. We first studied the biodistributions of PG3 and NG3 by using real-time near-infrared (NIR) fluorescence imaging after labeling them with an NIR dye NIR-797-isothiocyanate and intravenously injecting the labeled dendrons into hepatic H22 tumor-bearing mice. The NIR fluorescence images of the administrated mice at different time points postinjection (p.i.) are shown in Figure A. As can be seen, during the initial 2 h, the fluorescence signals from the labeled dendrons appear in the regions of the tumor and abdomen, suggesting that both PG3 and NG3 can rapidly accumulate in tumors via the enhanced permeability and retention (EPR) effect, and some of the dendrons also suffer body elimination. From 4 to 96 h p.i., the fluorescence signals mainly appear at the tumor sites, and for each of the two samples, the tumor signal intensity reaches its maximum at 24 h p.i. It is noticed that over the whole monitoring duration, the tumor signal intensity of the PG3-treated group is always higher than that of the NG3-treated group, indicating that PBA modification can really enhance the tumor targeting ability of the dendrons. To further understand the biodistributions of PG3 and NG3, at 48 h p.i., we excised and imaged the main organs, tumors, and other tissues (Figure B). Their MFIs are acquired and compared in Figure S22. It can be seen that, for both PG3 and NG3, strong fluorescence signals are only observed in the liver and tumor. For PG3, the MFI of the tumor is 1.60-fold that of the liver, and for NG3, the value is 1.52. Additionally, compared to the NG3-treated group, the tumor MFI of the PG3-treated group is 1.68-fold stronger thanks to the targeting effect of the PBA groups.

Figure 4

(A) NIR fluorescence images of the H22 tumor-bearing mice at different time points after tail-vein injection of the NIR-797-labeled PG3 and NG3. (B) NIR fluorescence images of the organs, tumors, and other tissues excised at 48 h after injecting the labeled PG3 and NG3. (C) DOX concentrations in different tissues at different time points after tail-vein injection of PDG3 and DG3. (D) DOX concentrations in plasma versus time after tail-vein injection of PDG3 and DG3. (E) AUC of DOX accumulation in tumors in PDG3 and DG3 groups. Data are presented as the mean ± SD (n = 3).

We further evaluated the biodistributions and drug delivery properties of PDG3 and DG3 by determining the DOX concentrations in different organs and tissues after intravenous injection. To this end, PDG3 and DG3 were injected into H22 tumor-bearing mice via the tail vein. At predetermined time points, different organs and tissues including the heart, liver, spleen, lung, kidney, tumor, and blood were collected. DOX concentrations in these samples were measured by using a fluorescence quantification technique after extracting DOX from the tissue homogenates under acidic conditions and are expressed as the percentage of injected dose per gram (% ID/g) of wet tissues (the data for blood are expressed as % ID/mL, Figure C,D). In blood, the concentration–time profiles for both PDG3 and DG3 fit well into two-compartment models, by which their elimination half-lives are calculated to be about 9.2 and 9.1 h, respectively (Figure D). Significant tumor accumulation of PDG3 and DG3 can be observed from Figure C. For the case of PDG3, the maximal DOX concentration in the tumors appears at 4 h p.i. and is about 7.6% ID/g, while for DG3, the maximum tumor uptake appears at 2 h p.i. and is about 4.3% ID/g. We further compared their tumor enrichment abilities by the tumor area-under-the-curve (AUC) based on DOX concentration over time. The tumor AUC of the PDG3-treated mice is ∼280.57% ID h/g, which is about 1.73-fold that of the DG3-treated mice (∼162.15% ID h/g, Figure E), demonstrating the role of PBA modification in augmenting drug delivery to tumors. In the heart, liver, spleen, lung, kidney, and tumor, the maximal DOX concentrations and corresponding appearance time for PDG3 and DG3 are compared in Table . It can be observed that, compared to DG3, PDG3 exhibits higher liver and spleen uptakes, comparable heart and lung uptakes, and a lower kidney uptake. It is also notable that both PDG3 and DG3 exhibit a significant ability to transport drugs to lungs, which might make them promising in the treatments of pulmonary diseases.

Table 2

Overview of the Maximal DOX Concentrations and Corresponding Appearance Time in the Main Organs and Tumor (PDG3 versus DG3)

PDG3 versus DG3
	heart	liver	spleen	lung	kidney	tumor
maximal DOX concentration (% ID/g)	3.7/3.7	10.0/7.3	5.7/1.2	4.5/4.8	17.4/23.0	7.6/4.3
appearance time (h)	2/4	2/2	1/0.5	0.5/2	1/1	4/2

Intratumoral Distribution

The intratumoral distribution of therapeutic agents closely relates to their tumor permeability, which is another important factor influencing the therapy efficacy apart from the tumor targeting ability. We first examined the intratumoral distributions of the RBITC-labeled PG3 and NG3 by tumor slice analyses after tail-vein injection into subcutaneous H22 tumor-bearing mice. At 36 h p.i., the tumors were excised and sectioned with cryostat. The blood vessels were stained with anti-CD31 (green), and the nuclei were stained with DAPI (blue). Figure A presents the slice images covering the whole equatorial plane of the tumors as well as their local enlarged views, since these slices can essentially reflect the overall intratumoral distributions of the samples. It can be seen that, when compared with the RBITC-labeled NG3, the red signals from the RBITC-labeled PG3 are significantly stronger and exhibit a more uniform distribution in the whole slice. Based on the fluorescence colocalization of the labeled dendrons and blood vessels, it can be observed that both PG3 and NG3 can extravasate from the blood vessels and penetrate into the tumor tissues, whereas PG3 penetrates deeper with a smaller fraction left in the vessels, which can be observed more clearly from the local enlarged pictures, revealing the higher tumor permeability of PG3 as compared to NG3.

Figure 5

(A) Typical CLSM images of the frozen tumor sections covering the whole equatorial plane of the tumors as well as their local enlarged views (scale bar = 50 μm) at 36 h after tail-vein injection of the RBITC-labeled PG3 and NG3 (red). Blood vessels were stained with anti-CD31 (green), and nuclei were stained with DAPI (blue). Scale bars = 2 mm. In vivo PA images of the tumors (B) and intratumoral mean intensities and distribution percentages of the PA signals from the labeled dendrons (C) at different time points after tail-vein injection of the NIR-797-labeled PG3 and NG3. (D) In vivo 3D PA images of the tumors at 36 h after tail-vein injection of the NIR-797-labeled PG3 and NG3. Scale bars = 2 mm. Data are presented as the mean ± SD (n = 3). ** P < 0.05 (PG3 versus NG3). (E) In vivo tumor growth curves of the H22 tumor-bearing mice after different treatments, expressed as the mean values of the relative tumor volume. (F) Mean tumor weights and photographs of the tumors excised from the mice on the 15th day after the treatments. (G) Evolution of the body weights during the treatments, expressed as the fold of initial body weight. “multi. i.v.” means multiple intravenous injection. Data are presented as the mean ± SD (n = 10). ** P < 0.05, *** P < 0.005.

We further studied the intratumoral distributions of PG3 and NG3 by using photoacoustic (PA) imaging after labeling them with NIR-797-isothiocyanate, since PA imaging can trace the contrast agents in 3D mode with high imaging depth and high resolution and can provide visual information on the intratumoral distribution in the whole tumor.[31] The PA tomographic images of the whole equatorial plane of the tumors at 4, 12, and 36 h after tail-vein injection are shown in Figure B. It can be clearly seen that, in the tumors, both the intensity and distribution area of the PA signals from the dendrons increase with time over the 36 h monitoring duration (Figure C). By contrast, as discussed above, the maximal DOX concentrations in the PDG3-treated and DG3-treated tumors appear at 4 and 2 h p.i., respectively, which are different from the continuous increase with time of the tumor concentrations of PG3 and NG3 determined by PA imaging. This may be caused by the faster metabolism and clearance of the released DOX as compared to the dendrons in tumors. At 36 h p.i., the PA signal intensity from the dendrons per unit tumor area of the PG3-treated tumor is about 1.2-fold that of the NG3-treated tumor (0.071 vs 0.059), and the percentage of the PG3 distribution area in the whole tumor section is also about 1.2-fold that of NG3 (82.7% versus 68.4%). All of these data indicate the significantly higher tumor targeting ability and higher tumor permeability of PG3 as compared to those of NG3. The tumor accumulation and intratumoral distribution of PG3 versus NG3 can also be observed clearly in the 3D PA images shown in Figure D.

In Vivo Antitumor Performance

Based on the in vitro and in vivo results discussed above, the controlled bifunctionalization with DOX and PBA groups endows PDG3 with improved cellular uptake, cytotoxicity to tumor cells, and tumor accumulation and penetration. We expect that these desirable properties would bring about a high antitumor efficacy. We studied the antitumor performance of PDG3 and DG3 with DOX and neat saline as the positive and negative control, respectively, by using the subcutaneous H22 tumor-bearing mice as the model animals. The mice were randomly divided into five groups for the treatments of a single dose of PDG3, DG3, free DOX, and neat saline, and multiple doses of PDG3 at a dose of 4.5 mg/kg DOX equivalent, respectively, with 10 mice per group. Each DOX formulation was injected as a solution in 0.2 mL of saline. The tumor volumes were measured every other day, and the tumor growth rates were evaluated by the evolution with time of the relative tumor volume (the ratio of measured volume to initial volume, Figure E). It can be seen that PDG3 and DG3 inhibit tumor growth much more efficiently than free DOX with single-dose administration. On the 15th day, the relative tumor volumes of the groups treated with PDG3 and DG3 are 4.8 and 8.1, respectively, while those of the groups treated with free DOX and saline are 14.0 and 20.9, respectively. The final average tumor weights of the groups treated with PDG3 and DG3 are 0.47 and 0.75 g, versus 1.58 g of the DOX-treated group and 1.93 g of the saline-treated group. For the group of the multiple doses of PDG3, the test mice were administrated every other day from day 1 to day 9. As shown in Figure E, the tumors started to shrink after two doses. Half of the tumors were eradicated within 7 days, and the rest were eradicated in the following 4 days. No recidivation was observed. The different inhibition effects of the treatments can be reflected by the photographs of the sarcomas excised from the mice on the 15th day (Figure F). The high antitumor efficacy of PDG3 should be contributed by its high tumor accumulation and deep tumor penetration. The body weights (Figure G) and clinical situations of all of the test groups were monitored, and no significant adverse effects were found in the groups treated by the four DOX formulations when compared with the saline-treated group, indicating the well-tolerated dose level of DOX even for the multiple doses.

Conclusions

We synthesized a new class of scaffold-modifiable bifunctional dendrons up to the fourth generation via an orthogonal amino protection strategy and a solid-phase synthesis method. The dendron G4 can be synthesized facilely within 2 days. The defect-free structures of the dendrons were definitely demonstrated by using mass spectrometry. We conjugated DOX and PBA groups to the scaffold and periphery of the dendrons, respectively. The PBA-modified dendron-based prodrug has a hydrodynamic diameter of ∼7 nm, higher drug loading content larger than 30%, and desirable water solubility. Thanks to the PBA-enhanced lysosome escape, tumor targeting ability, and tumor permeability as well as the high drug loading, the dendron-based prodrug exhibited an extraordinary antitumor efficacy and could eradicate the tumors established in mice by multiple intravenous administration.