Design and synthesis of multifunctional dendritic polymers aiming at the development of effective Drug Delivery Systems is described. Commercially available dendritic polymers were functionalized for obtaining Drug Delivery Systems (DDS) of low toxicity coupled with high loading capacity, targeting ability to specific cells and transport through their membranes. This was achieved by functionalizing the surface of dendritic polymers with targeting ligands which induce specificity to certain cells, with polyethylene glycol chains (PEG chains) which enhance their water solubility and stability in biological milieu and therefore result in prolonged circulation. Furthermore, molecular transporters facilitate transport through cell membranes while fluorescence probes are introduced to determine their intracellular localization. Crucial feature of these DDS is multi-valency which considerably enhances their binding strength to complementary cell receptors. Controlled and/or triggered release coupled with high loading of drugs are also advantageous properties of DDS. These properties when simultaneously fulfilled upgrade the potential and effectiveness of drugs.
Keywords: Dendrimers, Hyperbranched Polymers, Dendritic Polymers, Drug Delivery Systems, Folate Targeting Ligand
Dendritic polymers exhibit tree-like structures and include the subgroups of hyperbranched polymers, dendrigrafts, dendrons and dendrimers each group reflecting specific structural features [1-5]. The most extensively investigated subgroup is that of dendrimers which are highly branched, symmetrical and nano-sized macromolecules, having well-defined molecular weight and consist of a central core, repeating units and terminal functional groups. For their synthesis tedious experimental conditions are employed which increase their cost. However this may not be a serious problem if effective drugs would finally be prepared. Hyperbranched polymers, the other major subgroup, are on the contrary conveniently prepared being therefore of low cost but they are non-symmetrical and poly-dispersed.
Due to their structural features both classes of polymers exhibit nanocavities in the interior of which various molecules including drugs can be encapsulated. Drugs have also been conjugated at the dendritic scaffold. The significant number of functional groups on their surface enhances their binding to cells, due to the so-called multivalent effect, because more than one receptors can be simultaneously accessed by one dendritic polymer [6-7].
The property of dendritic polymers i.e. to encapsulate or conjugate various drugs together with multifuctionalization of their surface has been taken advantage in the development of effective drug delivery systems (DDS) [8-18]. Each type of external group plays a specific function. Thus, specificity to certain cells has been achieved by surface attachment of targeting ligands. Folate targeting ligand [19-20], due to its simple structure and convenience of surface attachment is extensively used while others such as cRGD peptides and herceptin (humanized monoclonal antibody), have been used to a lesser extent. Cell penetrating peptides [21-25] and guaninidium bearing oligomers are primarily being used for preparing molecular transporting dendritic nanoparticles. Guanidinium moiety on the surface of the dendritic polymers interacts with the phosphate and/or carboxylate group of the cells’ surface, inducing membrane transport [26-28]. Water solubility, decreased toxicity, biocompatibility, stability and protection in the biological milieu has been achieved by introducing on the surface of dendritic polymers poly(ethylene glycol) chains (PEG) in analogy with the most established liposomal carriers [29-30]. Modification of the internal groups of dendritic polymers can affect their solubilizing character, rendering possible the encapsulation of a diversity of drugs. Non-covalently encapsulated drugs inside nanocavities and covalently conjugated can be released from the dendritic scaffold by changes in the biological environment at the site of action [31,32]. Schematic representations of a multifunctional dendrimer and a hyperbranched polymer, loaded covalently or non-covalently with drugs, are shown in Figure 1.
Figure 1. Schematic representation of a multifunctional dendrimer (left) and hyperbranched polymer (right). End groups A, B, C depict the presence of targeting ligands, protective coating, transport agents, fluorescent probes, or in some cases a drug conjugated on the surface of the dendritic carrier.
In this short review, highlighting on the investigations conducted by our laboratory at “NCSR Demokritos” we will focus on multifunctional dendrimers and hyperbranched polymers discussing the strategy of their preparation through molecular engineering of the surface of commercialy available dendritic polymers. In this review, the two most extensively investigated dendrimers, i.e. diaminobutane poly(propylene imine), PPI, (Figure 2) and poly(amidoamine), PAMAM, (Figure 2) have been used as scaffolds for multifunctionalization. Also, the hyperbranched polyglycerol, PG, (Figure 2) and the recently introduced and very promising biodegradable hyperbranched aliphatic polyester, BoltornTM, (Figure 2)  was employed as basic dendritic polymer.
Figure 2. Commercially available dendrimers (left column) and hyperbranched polymers (right column) which can be subjected to multifunctionalization.
2. Multifunctional Drug Delivery Systems Based on Dendrimers
Employing diaminobutane poly(propylene imine) PPI, Figure 2, with 64 amino end groups, a model multifunctional dendrimeric DDS was prepared . Such a system can, in principle, simultaneously address stability and prolonged circulation in biological milieu, enhanced water solubility, strong binding to cell surface, transport through its membrane and also pH - triggered release. Thus, the surface of PPI, was multi-functionalized, Figure 3 by the introduction of protective poly(ethylene glycol) chains, guanidinium moieties for interacting with phosphate groups or other anionic groups on the cells surface the binding of which is amplified due to multivalency effects. Simultaneously, the accumulation of guanidinium groups on the surface of the dendrimer can also facilitate its transport through cells membrane in an analogous manner to cell-penetrating oligoarginine peptides [21-25]. Furthermore, due to the polyamine nature of the nanocavities the release of the encapsulated drug can be triggered by pH changes . Finally toxicity will be reduced by the decrease of toxic amino groups on dindrimeric surface [34,35].
Figure 3. A multifunctional dendrimeric derivative based on diaminobutane poly(propylene imine) dendrimer with 64 end groups.
Encapsulation and release of this multifunctional dendrimer were investigated with betamethasone valerate (BV). The multifunctional dendrimer encapsulated significantly higher concentrations of BV compared to the parent dendrimer. Thus seven molecules of the hydrophobic BV were solubilized per dendrimer. Specifically, BV loading capacity is 11 wt% inside the multifunctional dendrimer, i.e. almost double compared to the loading capacity of the simply PEGylated dendrimer (6 wt%)  and more than five times higher compared to the loading capacity of the parent dendrimeric solution (1.7 wt%) .
Although the above multifunctional dendrimer is pH-responsive, due to the presence of tertiary amino groups in the nanocavities interior, a two-step triggered process was required for completing the release of BV. Release of the drug with hydrochloric acid has not been achieved because BV remained solubilized within the dendrimeric environment and preferably within the poly(ethylene glycol) chains. However, BV encapsulated in this dendrimeric derivative was completely released upon addition of sodium chloride. Within the concentration range of sodium cation in extracellular fluids, i.e. 0.142 M, BV was released at a relatively low percentage. Sodium chloride of extracellular fluids can interact with PEG chains affecting the overall release profile of the drug. Thus, the possibility of triggering premature drug release in the extracellular fluid, i.e. before endocytosis to the target-cells, should be taken into account when designing a targeted PEGylated drug delivery system.
Based on a fourth generation diaminobutane poly(propylene imine) dendrimer, a novel targeted DDS nanocarrier was prepared , bearing protective PEG chains and a folate targeting ligand at the end of the chain, Figure 4. As a control, a PEGylated derivative without folate was also synthesized. For assessing the loading capacity and release properties of the above PEGylated dendrimeric derivatives, anticancer drug etoposide was employed. The concentration of encapsulated etoposide in dendrimers is higher than in phosphate buffer saline (PBS), i.e. 0.2 mM. Specifically, solubility of etoposide increased by a factor of ~14 in PPI aqueous solution, or otherwise ~21 molecules of drug were solubilized per dendrimeric molecule.
Figure 4 A PEGylated and folate functionalized dendrimeric derivative based on diaminobutane poly(propylene imine) dendrimer with 64 end groups.
Etoposide, in a simulated release experiment, was completely diffused (~100%) through the dialysis membrane following a linear release rate within 5 hours, while ~60% etoposide was released from PPI dendrimer in the first 4 hours, followed by the relatively slow release of about 15% in the next 3 hours. Even slower release rate is observed when PEGylated derivatives were employed. Specifically, ~40% of drug was released in a steady rate from PPI-6PEG and PPI-6PEG-Folate within the first 4 hours, while a relatively slow release of about 25% of etoposide is observed during the next 7 hours.
The cytotoxicity of the parent PPI and dendrimeric functional derivatives was assessed by flow cytometry. Human breast adenocarcinoma cells (MDA-MB-468) were seeded in 6-well plates and incubated for 48h at various concentrations of the functional dendrimers ranging from 50 to 250 μg mL-1. The viability (%) of cells gradually decreased as the concentration of dendrimeric derivatives increased while it was found that the introduction of PEG chains at their external surface leads to a concomitant decrease of the dendrimer's cytotoxicity.
Cellular uptake of PPI-6PEG and PPI-6PEG-Folate was assessed by confocal microscopy investigating the ability of folate moieties to enhance internalization of PPI-6PEG into MDA.MB.468 cells, which overexpress the folate receptors. Images A and C in Figure 5 depict cells incubated with PPI-6PEG-FITC, while images B and D depict cells incubated with PPI-6PEG-Folate-FITC. PPI-PEG and PPI-PEG-Folate images were acquired using the same laser power for excitation. From a detailed analysis of the results it is deduced that internalization was more efficient for PPI-6PEG-Folate-FITC i.e. folate moieties enhance internalization of PPI-6PEG in MDA.MB.468 cells.
Figure 5. Confocal microscopy images of MDA-MB-468 cells incubated with PPI-6PEG-FITC (A and C) and PPI-6PEG-Folate-FITC (B and D) for 24h.
The work with multifunctional dendrimeric derivatives is intensively continuing employing primarily the biodegradable PAMAM basic derivatives, Figure 2. Thus, step-wise modification of PAMAM dendrimers afforded multifunctional polymers  in which the drug is covalently attached on the dendrimeric scaffold. Partial acetylation of the primaryamino groups of fifth generation (G5) PAMAM dendrimer prevented non-specific interactions both in vitro  and in vivo . The remaining primary amino groups were used for attaching functional moieties e.g. fluorescein isothiocyanate (FITC), which is an imaging agent, folic acid (FA), which targets folate receptors on specific cancer cells, and methotrexate (MTX) anticancer drug, Figure 6, A. Folate moiety is overexpressed in a broad variety of human cancers and on activated macrophages  and, therefore, folate-mediated targeting has been widely applied to a diversity of DDS such as liposomes, dendrimers, hyperbranched polymers and other nanoparticles.
Figure 6 Multifunctional PAMAM dendrimeric derivative with covalently bound MTX (A) or paclitaxel (B), a PAMAM nanodevice with cyclic RGD and biotin moieties (C), and a fourth generation PAMAM dendrimer conjugated, through a spacer, to a PEG-folate moiety (D).
By an analogous strategy another multifunctional system  was prepared, based on fifth generation PAMAM scaffold which had covalently attached the anticancer drug paclitaxel (Taxol®), Figure 6, B. Experiments in vitro have shown efficient cell uptake of this multifunctional dendrimeric system and which is completely non-toxic to KB cells at the concentration of 200 nM.
Following preliminary reports in which a Arg-Gly-Asp (RGD) peptide was conjugated to a fifth generation PAMAM for in vitro targeting to ανβ3 integrin receptor expressing cells, the same dendrimer was subjected to multifunctionalization affording the nanocarrier shown in Figure 6, C . This carrier bears on its surface a cyclic Arg-Gly-Asp peptide (cRGD) and biotin groups that amplify detection of the carrier by anti-biotin antibody or avidin linked to horseradish peroxidase. This drug delivery system exhibits selective targeting of ανβ3 integrins, when compared either to the same, free cRGD peptide or to the biotinylated nanocarrier without any covalently attached peptide.
For the synthesis of this DDS surface amino groups of PAMAM were first partially acetylated and then biotinylated, while the remaining primaryamino groups were converted to succinamic moieties, some of which were conjugated with cRGD peptide residues through the amino group of the lysine side chain. Cytotoxicity was investigated employing B16F10 melanoma cell cultures using the XTT colorimetric assay. Binding of the nanocarriers to the target was determined using plates coated with human ανβ3 integrin and ανβ3 receptor expressing human dermal microvascular endothelial cells. The carrier is non-toxic within physiologic concentration ranges and is binding to the ανβ3 integrins much stronger than the cyclic cRGD peptide itself.
In another investigation  folate moiety was conjugated to a fourth generation PAMAM dendrimer either directly or indirectly through a PEG4000 chain as shown in Figure 6, D. The anticancer drug 5-fluorouracil, was encapsulated in both dendrimers and investigated in in vitro and in vivo experiments. Folate-PEG-dendrimeric derivative was significantly safer and more effective in tumor targeting compared to the non-PEGylated carrier. Functionalization with PEG-folate moiety reduced haemolytic toxicity, which secured sustained drug release as well as higher accumulation to the tumor site.
Furthermore, in another report;  herceptin targeting ligand was employed i.e. a humanized monoclonal antibody binding to human growth factor receptor-2 (HER2). It was covalently attached to a fifth-generation PAMAM in which methotrexate anticancer drug was also conjugated. The specificity of this FITC labelled nanocarrier and its internalization were demonstrated in cell lines overexpressing HER2 by flow cytometry as well as confocal microscopy. In addition, binding and uptake of these antibody conjugated dendrimers was completely blocked by excess non-conjugated herceptin. Reduced cytotoxicity of the conjugate in comparison to free methotrexate was attributed to the slow release of methotrexate from the carrier and its long retention in the lysosomes. Although the conjugate was less toxic to cells than methotrexate alone in vitro, it is however possible that this carrier might be superior for in vivo experiments due to its reduced toxicity for HER2-overexpressing breast cancers as a result of its specificity to target tumor cells.
3. Multifunctional Drug Delivery Systems Based on Hyperbranched Polymers
Hyperbranched polyglycerols, PG  have been widely employed exhibiting low toxicity and biocompatibility. They have been functionalized by an analogous strategy to dendrimers leading to multifunctional DDS. PEGylated (PG-PEG) and PEGylated-Folate (PG-PEG-Folate) functional derivatives of polyglycerol were prepared, Figure 7, and investigated as prospective drug carrier systems . Encapsulation properties of these dendritic derivatives were assessed by employing pyrene fluorescence probe and tamoxifen (TAM), a hydrophobic anticancer antiestrogen drug widely used for the treatment and prevention of breast cancer . Its encapsulation and simulated release properties were comparatively investigated for the parent polyglycerol, the PEGylated derivative, PG-PEG, and multifunctional PG-PEG-Folate. The solubility of TAM in water is 1.9 x 10-6 M which increases by a factor of 5 when solubilized in 1 mM PG solution. The solubility of TAM is considerably further enhanced by a factor of 65 in the presence of PG-PEG. This significant increase indicates that TAM is not only encapsulated inside the hyperbranched interior but also within poly(ethylene glycol) coating. Therefore, the introduction of poly(ethylene glycol) chains enhances, in general, the solubilization efficiency of dendritic polymers. It should be noted that for PG-PEG-Folate a ~1300-fold increase of TAM solubility was observed.
Figure 7 PEGylated (PG-PEG) and PEGylated-Folate (PG-PEG-Folate) functional derivatives of hyperbranched polyglycerol.
The release of TAM encapsulated in PG and its derivatives, was triggered by the addition of NaCl in the solution. Since molecules encapsulated in PEG coating can be replaced by metal ions it is interesting to investigate whether sodium cation complexation may cause premature release of the drug in the extracellular fluids, i.e. before the drug-loaded nanocarrier reaches the target cell. By the addition of 0.142 M NaCl solution, 39 % and 24 % of the solubilized TAM in PG and PG-PEG were released in aqueous media, respectively. Under the same conditions, in the presence of PG-PEG-Folate, only 6 % of the solubilized TAM was released. Thus by employing the multifunctional derivative as nanocarrier most of TAM remained encapsulated and it is not released in the extracellular fluid at a concentration of 0.142 M NaCl solution. It can therefore be assumed that this nanocarrier when employed in in vivo experiments will reach, in principle, target cells appreciably loaded with TAM.
Biodegradable dendritic hyperbranched polyesters have recently been prepared  based on 2,2-bis(hydroxymethyl)propanoic acid as an AB2 monomer and 2,2-bis(hydroxymethyl)-1,3-propanediol as central core. These polymers are supplied in various molecular weights, bearing a varying number of primary hydroxyl groups, i.e. 16, 32, 64 for Boltorn H20, H30 and H40, respectively. Two glycodendritic polymers  have been initially prepared with 16 and 32 mannose moieties based on Boltorn H20 and H30 hyperbranched polymers which have been used as controls. These glycodendritic polymers are water soluble, exhibit low toxicity and have the capability to interact with lectin receptors. They can, therefore, be considered as promising candidates for drug delivery.
Boltorn H40 was functionalized with poly(ethylene glycol) chains affording a water soluble PEGylated dendritic derivative, BH40-PEG, Figure 8, A . PEGylation of the dendritic scaffold is indispensable not only for protecting the drug carrier in biological milieu but also for enhancing its water solubility. Paclitaxel solubility increased by a factor of 65, 110, 210 and 350, in 1%, 3%, 6% and 9% w/v BH40-PEG solutions respectively, compared to aqueous solubility. Paclitaxel simulated release was determined by the dialysis method and showed that about 60% of the encapsulated drug was released in the aqueous phase during the first 6 hours and almost completed in 12 hours. The cytotoxicity was assessed in vitro with A549 human lung carcinoma cells and found to be non-toxic for 3h incubation at concentrations equal or lower to 50 µM, while LD50 was 100 µM.
A folate-conjugated amphiphilic hyperbranched block copolymer (H40–PLA-b-MPEG/PEG–FA) based on Boltorn H40 was also synthesized , Figure 8, B. Thus a hydrophobic poly(l-lactide) (PLA) inner shell, a hydrophilic methoxy poly(ethylene glycol) (MPEG) with a folate moiety attached on poly(ethylene glycol) chain terminal group (PEG–FA), were introduced on Bolton H40 scaffold. This block copolymer forms unimolecular micelles in aqueous solutions as established by dynamic light scattering, fluorescence spectroscopy, and transmission electron microscopy. Encapsulation of the anticancer drug doxorubicin (DOX) in its base form afforded a DDS H40–PLA-b-MPEG/PEG–FA. Drug release of these DOX-loaded micelles showed an initial “burst release” (up to 4 h) followed by a sustained release of DOX over a period of about 40 h. Cellular uptake of the DOX-loaded H40–PLA-b-MPEG/PEG–FA micelles was higher than that of the DOX-loaded H40–PLA-b-MPEG micelles because of the folate-receptor mediated endocytosis. Micelles therefore exhibit higher cytotoxicity against the 4T1 mouse mammary carcinoma cell line. Degradation in vitro revealed that H40–PLA-b-MPEG/PEG–FA block copolymer is hydrolytically degraded within six weeks. These results have indicated that these carriers have great potential as tumor-targeted drug delivery nanocarriers.
Figure 8 Chemical structures of a Boltorn H40 PEGylated dendritic derivative (A), a Boltorn H40 poly(l-lactide) - PEGylated-Folate derivative (B) and of a poly(epsilon-caprolactone) -PEGylated-Folate dendritic derivative (C).
In an analogous manner, amphiphilic core-shell multifunctional dendritic hyperbranched polymers bearing folate group as targeting ligand were synthesized, Figure 8, C . The core consisted of polyester Boltorn H40, the inner part of the amphiphilic polymers were hydrophobic poly(epsilon-caprolactone) chains while the outer shell were hydrophilic poly(ethylene glycol) chains. Tumor cell targeting was achieved by folate moiety attached at the terminal hydroxyl group of the PEG chain. Paclitaxel and 5-fluorouracil were encapsulated into these dendritic nano-carriers. Experiments in vitro showed that drug-loaded nanoparticles exhibited enhanced cell growth inhibition, attributed to folate targeting increased cytotoxicity of drug-loaded nanoparticles.
Surface multifunctionalization of dendrimers and hyperbranched polymers results in the development of DDS of low toxicity, significant loading capacity, specificity to certain biological cells and transport ability through their membranes. Drug delivery effectiveness, attributed primarily to targeting due to molecular engineering of dendritic surface, is also affected by covalent or non-covalent binding of selected drugs, the release of which can be tuned by changes in the biological environment such as pH or ionic strength. Multivalent interactions of dendritic DDS functional groups with cell complementary receptors is finally of crucial significance.
1. Vögtle F et al, Prog Polym Sci 2000;25:987-1041
2. Svenson S & Tomalia DA, Adv Drug Delivery Rev 2005;57:2106-29
3. Lee CC et al, Nature Biotechnol 2005;23:1517-26
4. Newkome GR & Shreiner CD, Polymer 2008;49:1-173
5. Khadare J et al, Chem Soc Rev 2012;41:2824-2848
6. Mammen M et al, Angew Chem, Int Ed 1998;37:2755-94
7. Badjic JD et al, Acc Chem Res 2005;38:723-32
8. Rolland O et al, New J Chem 2009; 33:1809-24
9. D' Emanuele A & Attwood D, Adv Drug Delivery Rev 2005; 57:2147-62
10. Gajbhiye V et al, Curr Pharm Des 2007;13:415-29
11. Paleos CM et al, Mol Pharm 2007;4:169-88
12. Paleos CM et al, Curr Top Med Chem 2008;8:1204-24
13. Paleos CM et al, Exp. Opinion on Drug Delivery 2010: 7: 1387-1398
14. Stiriba S-E et al, Angew Chem, Int Ed 2002;41:1329-34
15. Gillies ER & Fréchet JMJ, Drug Discovery Today 2005;10:35-43
16. Baker JR, Jr, in Hematology, American Society of Hematology 2009; 708-19
17. Nanjwadea BK et al, Eur J Pharm Sci 2009; 38:185–96
18. Svenson S, Eur J Pharm Biopharm 2009;71:445-62
19. Low PS et al, Acc Chem Res 2008;41;120-9
20. Vlahov IR et al, Bioconjugate Chem 2012;23:1357-69
22. Futaki S, Adv Drug Delivery Rev 2005;57;547-58
23. Rothbard JB et al, Adv Drug Delivery Rev 2005;57:495-504
24. Foged C & Nielsen HM, Expert Opin Drug Deliv 2008;5:105-17
25. Bonduelle CV & Gillies ER, Pharmaceuticals 2010;3:636-66
26. Tsogas I et al, Langmuir 2006;22:11322-8
27. Tsogas I et al, ChemBioChem 2007;8:1865-76
28. Tsogas I et al, Biomacromolecules 2007;8:3263-70
29. Lasic DD & Needham D, Chem Rev 1995;95:2601-28
30. Kaasgaard T et al, Int J Pharm 2001;214:63-5
31. Paleos CM et al, Biomacromolecules 2004;5:524-9
32. Calderón M et al, Biochimie 2010;92:1242-51
33. Zăgar E et al, Polymer 2006;47:166-75
34. Malik N et al, J Control Release 2000;65:133-48
35. Jain K et al, Int J Pharm 2010;394:122-42
36. Sideratou Z et al, J Colloid Interface Sci 2001;242:272-6
37. Sideratou Z et al, Biorganic & Medicinal Chemistry Letters 2010; 20:6513-6517
38. Majoros IJ et al, J Med Chem 2005;48:5892-9
39. Thomas TP et al, J Med Chem 2005;48:3729-35
40. Kukowska-Latallo JF et al, Cancer Res 2005;65:5317-24
41. Low PS et al, Acc Chem Res 2008;41:120-9
42. Majoros IJ et al, Biomacromolecules 2006;7:572-9
43. Lesniak WG et al, Bioconjugate Chem 2007;18:1148-54
44. Singh P et al, Bioconjugate Chem 2008;19:2239-52
45. Shukla R et al , Nanotechnol 2008;19/295102:1-7.
46. Wilms D et al, Acc Chem Res 2010;43:129-41
47. Tziveleka L-A et al, Macromol Biosci 2006;6:161-9
48. Mocanu EV& Harrison RF., Rev Gynaecol Practice 2004;4:37-45
49. Arce E et al, Bioconjugate Chem 2003;14:817-23
50. Kontoyianni C et al, Macromol Biosci 2008;8:871-81
51. Prabaharan M al, Biomaterials 2009;30:3009-19
52. Chen S et al, Biomacromolecules 2008;9:2578-85
Constantinos M. Paleos
NCSR "Demokritos", 15310 Aghia Paraskevi,Attiki,Greeceand
Regulon AE, 7 G. Afxentiou st.,
Tel: +30 210 6503666; E-mail:firstname.lastname@example.org