|Home | About | Journals | Submit | Contact Us | Français|
Poly(amidoamine) (PAMAM) dendrons were synthesized with c(RGDyK) peptide on the surface to create a scaffold for cellular targeting and multivalent binding. Binary dendron-RGD conjugates were synthesized with a single Alexa Fluor 488, biotin, methotrexate drug molecule, or additional functionalized dendron at the focal point. The targeted-dendron platform was shown to specifically target αVβ3 integrin expressing human umbilical vein endothelial cells (HUVEC) and human glioblastoma cells (U87MG) in vitro via flow cytometry. Specific targeting of the dendron-RGD platform was further confirmed by confocal microscopy. Biological activity of the targeted drug conjugate was confirmed via XTT assay. The orthogonal reaction chemistry used at thedendron focal point gives a precise 1:1 ratio of the attachment of multiple functionalities to a small molecular weight, chemically stable high avidity molecule. These studies serve as a framework to selectively combine biologically relevant functions with enhanced specific binding activity to substitute for antibodies in many diagnostic and therapeutic applications.
The development of specific targeting molecules for diagnostic and therapeutic applications is an important goal being pursued by the biomedical community. Precise targeting of molecules, cellular compartments or tissues has the potential to positively affect a broad range of disciplines including diagnostic and imaging technologies, and could enhance the treatment of a wide range of conditions such as inflammation, infectious diseases, and cancer. The most common molecules currently employed for numerous applications are monoclonal antibodies (1, 2), or immunoconjugates of these antibodies, which take advantage of their high specific binding affinities for antigens. Almost all specific diagnostics are now based on some form of immunoconjugate (3–7), and numerous monoclonal antibody-based therapeutics are FDA-approved to meet clinical needs (8, 9).
Despite these successes, immunoconjugates suffer from significant disadvantages. There are many small molecules that are not immunogenic, and therefore cannot be identified by antibodies. In addition, antibodies are proteins that denature under many chemical and thermal conditions, and therefore necessitate strict refrigeration requirements for diagnostic kits or therapeutics based on immunoconjugates. The large size (approximately 150 kDa) of immunoconjugates limits blood clearance and potentially causes slow tumor penetration (10, 11). Being foreign proteins, antibodies elicit immunogenic responses in patients that limit their therapeutic utility (12, 13). Finally, biological molecules are difficult and costly to produce in bulk (14). Thus, alternative molecules to antibodies are needed.
Poly(amidoamine) (PAMAM) dendritic macromolecules are controllable in size and have defined, monodisperse (PDI = 1.02) structures. The macromolecule is significantly smaller than antibodies, which allows the device to enter the cell to potentially treat multiple proliferation pathways. In addition, the polyamide backbone helps the macromolecule maintain water solubility and minimizes immunogenicity (15, 16). The dendritic macromolecules also have large numbers of surface primary amines created by a branched architecture that allow for the attachment of multiple molecules with various functions. This has been demonstrated with targeting ligands, imaging agents and therapeutics (17, 18). The ability to conjugate multiple targeting moieties provides the opportunity to create a platform with increased binding avidity through polyvalent cell interactions (19). Targeted dendrimer delivery platforms have been successfully tested in in vitro and in vivo model tumor systems (20–29).
To increase the utility of the dendritic scaffold, orthogonal and modular pathways can be used to construct binary devices with a greater degree of specificity and flexibility. However, current research pursuing such pathways to create functional dendritic devices has been sparse. PAMAM dendrimers have been developed with unique surface groups for modular synthesis via click chemistry (30), complimentary oligonucleotide hybridization (31), or orthogonal functionalization (32). Dendrons with a bis-MPA backbone have been utilized by exploiting the unique reactive site at the focal point to create bi-functional materials (33). These approaches offer the potential to construct individual combinations for a specific targeting application from a common library of functionalized devices in a versatile manner and overcome the inefficiencies of step-wise conjugations on a single dendritic macromolecule. However, the central advantage of these techniques is that they provide increased homogeneity products by reducing the complexity of ligand distributions created by multi-functional platforms.
Herein, we present the synthesis and evaluation of in vitro biological activity for a targeted, PAMAM dendron avidity platform (Figure 1). The functionalized PAMAM dendron can maintain the binding specificity and multivalency of previous dendritic models (25, 34) while creating an orthogonally coupled scaffold more suitable for personalized medicine and applications. Binary devices were synthesized through a unique alkyne group at the dendron focal point. The alkyne is reacted with an azide-functionalized dye, biotin, therapeutic molecule or a second dye-conjugated dendron in a copper-catalyzed 1,3-dipolar cycloaddition, commonly referred to as “click chemistry” (35). This strategy is attractive because it provides high versatility and yields, mild reaction conditions and no alteration to neighboring functional groups that eliminates the need for protecting groups. The platform itself gives a precise 1:1 ratio of the attachment of multiple functionalities, preventing aggregate formation. A dendron with multiple Arg-Gly-Asp (RGD) adhesion ligands on the surface and an additional function at the focal point can specifically target αVβ3 integrin expressing human umbilical vein endothelial cells (HUVEC) and human glioblastoma cells (U87MG), displaying the biological utility of the platform. The dendron-RGD platform also successfully inhibits tumor growth when conjugated to a therapeutic drug, such as methotrexate (MTX). Thus, the ability to conjugate specific and varied effector sites to the dendron periphery to generate a high avidity binding agent and use orthogonal reaction chemistry at the focal point to add additional functionalities makes these “dendro–bodies” an attractive alternative to antibodies for targeting applications.
Alexa Fluor 488 azide and Prolong Gold mounting agent were obtained from Invitrogen. c(RGDyK) peptide was obtained from Peptide International. Biotin-dPEG3+4-azide was obtained from Quanta Biodesign. Sephadex G15 and G25 were obtained from GE Life Sciences. Spectra/Por 7 dialysis membrane (MWCO 1000) was obtained from Spectrum Laboratories. NeutrAvidin-Dylight 405 was obtained from Pierce. All other chemicals were obtained from Sigma. All purified reaction products were characterized by 1H NMR, HPLC, and matrix-assisted laser desorption-ionization time-of-flight (MALDI-TOF) mass spectrometry.
Alkyne-G3-PAMAM dendron was prepared as previously reported (36) and analyzed by 1H and 13C NMR and MALDI-TOF. MS (MALDI): 3479 Da; calculated for C153H305N61O30: 3479.45 Da.
Glutaric anhydride (0.0164 g, 0.1437 mmol) dissolved in anhydrous MeOH (0.8 mL) was added dropwise to a solution of 1 (0.0200 g, 0.0057 mmol) and triethylamine (Et3N) (0.0186 g, 0.1838 mol) in anhydrous MeOH (0.2 mL) while stirring, and the reaction mixture was allowed to stir for 24 hours at room temperature. All solvent was removed in vacuo, and the reaction material was dissolved in H2O, purified on a G-15 Sephadex column and lyophilized. MS (MALDI): 5305 Da; calculated for C233H401N61O78: 5305.04 Da.
1-(3-(dimethylamino)propyl)-3- ethylcarbodiimide HCl (EDC) (0.0838 g, 0.4373 mmol) was allowed to react with 2 (0.0145 g, 0.0027 mmol) in 1:1 H2O/DMSO (2 mL) for 2 hours to form an active ester. c(RGDyK) (0.0169 g, 0.0273 mmol) in 4:1 H2O/DMSO (1.25mL) was added dropwise to the EDC and dendron solution and stirred for 2 days at room temperature. The DMSO was removed via dialysis in H2O for 1 hour. The resulting solution was lyophilized, redissolved in H2O, and purified on a G-15 Sephadex column, then again lyophilized. MS (MALDI): 5300–12000 Da (broad).
A mixture of azide-functionalized Alexa Fluor 488 (0.00022 g, 0.0002 mmol) and 3 (0.0015 g, 0.0002 mmol) in H2O (1 mL) in the presence of copper(II) sulfate (2.5 × 10−5 g, 0.0001 mmol) and sodium ascorbate (4.0 × 10−5 g, 0.0002 mmol) was stirred for 36 hours at room temperature. The solution was purified on a G-25 Sephadex column and lyophilized. MS (MALDI): 5300–12600 Da (broad).
Methotrexate (0.0238 g, 0.05237 mmol) was dissolved in DMF (1 mL) and cooled in a water/ice bath. N-Hydroxysuccinimide (0.0067 g, 0.0576 mmol) and EDC (0.0110 g, 0.0576 mmol) were added, and the resulting mixture was stirred in the ice bath for 30 min to give a white precipitate. A solution of 11-azido-3,6,9- trioxaundecan-1-amine (0.0120 g, 0.0550 mmol) in DMF (0.5 mL) was added, and the resulting mixture was allowed to warm to room temperature and stirred for 24 h. The mixture was poured into water (50 mL) and washed with chloroform (50 mL × 3). The aqueous layer was lyophilized and further purified on a G-25 Sephadex column. MS: 654.7 Da; calculated for C28H38N12O7: 654.68 Da.
A mixture of 5 (0.0019 g, 0.0029 mmol) and 3 (0.0025 g, 0.0003 mmol) in H2O (1 mL) in the presence of copper(II) sulfate (2.5 × 10−5 g, 0.0001 mmol) and sodium ascorbate (4.0 × 10−5 g, 0.0002 mmol) was stirred for 36 hours at room temperature. The solution was purified on a G-25 Sephadex column and lyophilized. MS (MALDI): 5300–12600 Da (broad).
A mixture of biotin-dPEG3+4-azide (0.0045 g, 0.0006 mmol) and 3 (0.0080 g, 0.0001 mmol) in 1:1 DMSO/H2O (1 mL) in the presence of copper(II) sulfate (2.5 × 105 g, 0.0001 mmol) and sodium ascorbate (4.0 × 105 g, 0.0002 mmol) was stirred for 36 hours at room temperature. The solution was purified on a G-25 Sephadex column and lyophilized. MS (MALDI): 5300–12700 Da (broad).
Azide-G2-PAMAM dendron was prepared as previously reported (37) and analyzed by 1H and MALDI-TOF. MS (MALDI): 1698.1 Da; calculated for C73H148N32O14: 1698.16 Da.
Alexa Fluor 488 sulfodichlorophenol ester (0.0050 g, 0.0061 mmol) dissolved in 0.5 mL DMF was added dropwise to 8 (0.0028 g, 0.0015 mmol) in 0.5 mL 0.1 M sodium bicarbonate buffer while stirring, and the reaction mixture was allowed to stir for 24 hours at room temperature. Solvent was removed in vacuo, and the resulting product was dissolved in 2 mL H2O. Glycidol (0.0112 g, 0.1506 mmol) was added to the solution and the mixture was allowed to stir for 24 hours at room temperature. The solution was purified on a G-25 Sephadex column and lyophilized. MS (MALDI): 3000–7000 Da (broad).
A mixture of 9 (0.0030 g, 0.0009 mmol) and 3 (0.0044 g, 0.0006 mmol) in H2O (1 mL) in the presence of copper(II) sulfate (2.5 × 105 g, 0.0001 mmol) and sodium ascorbate (4.0 × 105 g, 0.0002 mmol) was stirred for 36 hours at room temperature. The solution was first purified on a G-25 Sephadex column and lyophilized. The product was further purified via HPLC to remove excess azide-G2(AF)(OH). MS (MALDI): 5000–29000 Da (broad), 17500 Da maximum.
HUVEC cells were obtained from Invitrogen (Cascade Biologics). U87MG and KB cells were obtained from American Type Tissue Collection (Manasass, VA). HUVEC cells (cat. # C-015–10C), U87MG and KB cells were grown and maintained according to the supplier’s protocol. Compounds were prepared in non-supplemented cell culture medium (Medium 200) and incubated in cell suspensions for 2 hours at 37 °C in a 5% CO2 incubator. Blocking experiments were run by pre-incubating the cells with a 200 equivalent excess of c(RGDyK) for 10 minutes at room temperature prior to incubation.
Cells were washed with 5mL of PBS, spun at 150 G for 7 minutes, and re-suspended in 1 mL of PBS prior to being examined. Cells were examined quantitatively for dendron uptake via a flow cytometer (Beckman Coulter Epics XL-MCL). The FL1 fluorescence of 10000 cells was measured and the mean fluorescence of cells was quantified.
HUVEC cells were seeded and grown on 35 mm glass-bottomed culture dishes (MatTek Corp., Ashland, MA). After 2 days, cells were treated with 0.5 μM co-dendron 10 or biotin-conjugated dendron 7. Blocked samples were run by pre-incubating the cells with a 200 equivalent excess of c(RGDyK) for 10 minutes at room temperature prior to incubation. After the 2 hour incubation period, biotin-conjugated dendron 7 treated cells were treated with 4 mg/mL BSA for 30 minutes to limit non-specific binding followed by treatment with 0.5 μM NeutrAvidin-Dylight 405 for 30 minutes. Media was aspirated from the cells, washed 3 times with PBS, and fixed with 2% paraformaldehyde. The samples were again aspirated, washed 3 times, and aspirated before mounting with Prolong Gold and a cover slip was added. Samples were imaged using an Olympus FluoView 500 Laser Scanning confocal microscope using 60X, 1.5 NA oil immersion objective. Fluorescence and differential interference contrast (DIC) images were collected simultaneously using either an argon laser with 488 nm excitation or a solid state diode laser with 405 nm excitation and FluoView software.
For cytotoxicity experiments, cells were seeded in 96-well microtiter plates (2500 cells/well) in Medium 200. The cells were treated in triplicate with different concentrations of free MTX or dendron conjugates for 72 hours with a replenishing of 100 μL medium and the conjugates after 48 hours. A colorimetric XTT assay (Invitrogen) was performed following the vendor’s protocol. An aliquot of 30 μL of XTT working solution prepared according to manufacturer instructions was added to each well using a multichannel pipette followed by the addition of 20 μL of PBS. After 1 hour incubation in the dark at 37 °C, the absorbance at 490 nm was recorded using a plate reader. 690 nm was selected as a reference wavelength. Cell viability was calculated based on optical density (OD) from untreated cells. Statistical significance was assessed using Two-Way ANOVA with posthoc Bonferroni tests checking for main effects of conjugate and concentration. P-values less than 0.05 were considered significant.
The surface amines of the G3 PAMAM dendron 1 were allowed to react with excess glutaric anhydride in MeOH, in the presence of Et3N as base, to form carboxylic acids (Figure 2i). The reaction yielded carboxyl terminal groups and eliminates potential nonspecific charge interactions with the cell plasma membrane caused by primary amines. 1H NMR spectroscopy confirmed complete conversion of the primary amines and formation of dendron 2 by the loss of methylene peaks (–CH2CH2NH2) at δ 2.80 and 3.30 ppm and the formation of new methylene peaks (-NHCO(CH2)3COOH) at δ 1.84, 2.21, and 2.27 ppm.
The targeting moiety, RGD peptide, was conjugated to the carboxylated surface of the PAMAM dendron (Figure 2ii). RGD binds to the αVβ3 integrin found on the luminal surface of endothelial cells during angiogenesis (38, 39). This makes the αVβ3 integrin a unique marker that differentiates newly formed capillaries in the tumor vasculature from their mature counterparts (40–42). Antiangiogenic therapies, such as ones that employ RGD targeting, prevent neovascularization by inhibiting differentiation, migration and proliferation of endothelial cells (43).
The surface carboxyl groups were activated with EDC, followed by the addition of 12 equivalents of c(RGDyK) to give dendron 3. The product was purified by size exclusion chromatography and lyophilized. The 1H NMR spectrum of the dendron-RGD body shows some of the aliphatic signals of the peptide overlap with the aliphatic signals of the dendron; however, using the peak integration values in the aromatic region from the phenyl ring of the peptide, an average of 4.1 attached RGD per dendron was calculated.
The dendron AF-G3(COOH)11.9(RGD)4.1 4 (Figure 2iii) was synthesized by the reaction of azide-functionalized Alexa Fluor 488 with 3 in a copper-catalyzed 1,3-dipolar cycloaddition. Alexa Fluor 488 was chosen as the fluorescent label for detection of conjugates by flow cytometry because it is brighter than fluorescein conjugates and more resistant to photobleaching. The products were purified via size exclusion chromatography and lyophilized. The characteristic triazole proton peak of a successful click reaction (δ 8.10 ppm) partially overlaps with the peaks for Alexa Fluor 488.
The specific binding of 4 to αVβ3 integrin expressing HUVEC and U87MG cells was examined via flow cytometry (Figure 3). KB cells, which express lower levels of αVβ3 integrin, were also evaluated. Cells were treated with dendron-RGD conjugates at final concentrations of 0.25, 0.50, 1, 2, and 4 μM, either with or without preincubation with 200 fold molar equivalences of free RGD peptide. Specific binding and dose dependence of dendron bodies to integrin expressing cells were evaluated after the 2 hour incubation at 37 °C. The addition of excess free peptide to inhibit the uptake of the binary dendron by HUVEC and U87MG cells significantly blocked uptake. The inability to completely inhibit binding of the dendron may be due to a higher avidity of the dendron conjugate when compared to free RGD.
A co-dendron structure, (OH)24(AF)2G2-G3(COOH)11.9(RGD)4.1 10 was created by reacting an Alexa Fluor 488 conjugated dendron 9 with an azide focal point with 3. The product was purified via HPLC to remove excess 9. The co-dendron structure offers an increased carrying capacity for conjugates, including the Alexa Fluor 488 used in this study. NMR of 9 showed an average of 2 Alexa Fluor 488 molecules attached per dendron, thus it was expected to emit a stronger signal when conjugated to 3 compared to a single dye at the focal point.
The specific binding of co-dendron 10 to αVβ3 integrin expressing HUVEC cells was determined via flow cytometry (Figure 3D). The co-dendron treated samples see the expected increase in mean fluorescence compared to mono-dye conjugated 4. Specific binding could be significantly blocked (Figure 4), but again the higher avidity dendron conjugate out-competed free RGD. Non-targeted control 9 showed minimal non-specific uptake.
The specific binding evaluated by flow cytometry was confirmed by confocal microscopy. HUVEC cells were treated with 0.5 μM 10 or biotin-conjugated 7 and incubated for 2 hours. The biotin-conjugate was detected by the addition of an avidin-functionalized dye. Confocal images confirm the uptake of the dendrons (Figure 5B, D), as seen from the punctuate pattern within the cytoplasm. Partial inhibition of the dendron conjugates by preincubation with free RGD was also established (Figure 5C, E).
A drug conjugate, 6 was synthesized to show the therapeutic potential of the platform. MTX was reacted with 11-azido-3,6,9-trioxaundecan-1-amine to yield an azide-functional drug. The resulting MTX-azide was then attached to the dendron at the focal point via click chemistry. The inhibition of cell growth induced by the dendron conjugates was tested by XTT colorimetric assay. 6 and free MTX showed comparable cytotoxicity to HUVEC cells after 3 days (Figure 6). Cell numbers were between 55–80% for all 6 and free MTX treated samples at concentrations greater than 0.125 μM. The effects of 6 and MTX were not significantly different at these concentrations. 6 had a significantly greater effect on cell viability versus 3 for concentrations greater than 0.125 μM. 6 has an approximately equivalent concentration of MTX compared to free drug; the drug-conjugated dendron 6 treatments may have slightly lower drug concentration due to any remaining uncoupled dendron. The similar toxicities of 6 and free MTX show that the dendron scaffold is effective at delivering a therapeutic to cells.
We have shown that by exploiting the initiation point as a unique binding site in PAMAM dendrons, selective coupling reactions can yield highly defined binary devices that exhibit the ability to bind to specific cells. Employing a 1,3-dipolar cycloaddition to the focal point of the dendron allows for additional functional moieties to be added in a precise manner without altering other molecules on the dendron. The prototype molecule, synthesized 4, successfully displayed specific targeting to αVβ3 integrin, behaving similarly to RGD conjugates on previously studied multifunctional G5 PAMAM dendrimers (25, 34) that lack the same degree of synthetic control. Detection can be enhanced by attaching an additional dendron with an increased carrying capacity, shown with 10. The dendron can be combined with biotin, allowing it to interact with a variety of avidin-functionalized molecules. 6 was shown to successfully inhibit the growth of HUVEC cells similarly to free drug. The PAMAM dendron platform has potential to be important in targeted delivery chemistry, as our binary device can serve as a framework to selectively attach diagnostic and imaging agents, therapeutics, or other biologically relevant moieties and surfaces. In addition, coupling a second functionalized dendron or dendrimer to the dendron could address other applications requiring the delivery of multiple drugs or imaging agents.
This project was supported with Federal funds from the National Cancer Institute, National Institutes of Health, under award 1 R01 CA119409.