Properties of Poly(B-amino Ester)s

The poly(b-amino ester)s, a class of biodegradable cationic polymers, were firstly prepared by Chiellini in 198340. These polymers were based on poly(amidoamine)s developed in 1970 by Ferruti41, that contain tertiary amines in their backbones and can be synthesized by a simple Michael addition reaction of bifunctional amines and bisacrylamides. However, the interest over the use of poly(b-amino ester)s rised significantly after its use as transfection reagent at Langer Lab in 200042. The development of poly(b-amino ester)s emerged by the need to develop a cationic polymer for gene delivery with high transfection efficiency and long-term biocompatibility including hydrolyzable moieties easily degradable into non-toxic small molecule byproducts. The synthesis of this polymer can easily be accomplished: without necessity of independent preparation of specialized monomers; the use of stoichiometric amounts of expensive coupling reagents, or amine protection strategies prior to polymerization42. The main general objective of the work of mentioned research group was to develop a polymer-based non-viral vector more efficient and less cytotoxic than other cationic polymers used at that time for this purpose (such as, polyethylenimine (PEI) or poly(L-lysine) (PLL)).

In fact, poly(b-amino ester) approach exhibited a particularly attractive basis for the development of new polymer-based transfection vectors for several reasons: the polymers contain the required amines (positive charges to complex genetic material); readily degradable linkages (by hydrolysis of ester bonds in the polymer backbones may increase the biodegradability and biocompatibility); and multiple analogues could be synthesized directly from compounds commercially available (easy and inexpensive synthesis) allowing to tune polymer properties (like buffering capacity)42.

Besides being used as transfection vector, PbAEs has been also applied in others biomedical areas, such as delivery systems for drugs43;44 or proteins45;46, magnetic resonance imaging agents47;48, or as scaffold for tissue engineering49;50.

Synthesis and main physicochemical properties of poly(b-amino ester)s

The poly(b-amino ester)s are easily synthesized by the conjugate addition of a primary amine or bis(secondary amine) and a diacrylate, in a one-step reaction without any side product that need be removed through further purification steps. It can be prepared without solvents, catalysts, or complex protecting group strategies42;51.

Depending on the ratio of monomers during the synthesis, poly(b-amino ester)s can be tailored to have either amine- or diacrylate-terminated chains. An excess of either diacrylate or amine monomer results in a prevalence of acrylate- or amineterminated poly(b-amino ester)s, respectively52;53.

The synthesis is performed either neat (solvent free) or in anhydrous organic solvents to mitigate hydrolytic degradation during synthesis42;54. Normally, experiments using solvents occur at lower temperature and over long periods of time compared to solvent-free formulations. Table 1.3 summarizes the main reactions for the synthesis of PbAE and the obtained properties such as molecular weight, polydispersity index (Ð), solvent solubility or yield.

The most common solvents used are dimethylsulfoxide (DMSO), chloroform (CHCl3), or dichloromethane (CH2Cl2)57. However, others solvents have also been used, such as methanol, N,N-dimethylformamide (DMF) or N,N-dimethylacetamide (DMA)59;61–63. The solvent used has influence on the final molecular weight of the PbAE. For example, the use of CH2Cl2 typically yields higher molecular weight polymer compared to THF42.

On the other hand, solvent-free polymerizations maximize monomer concentrations, thus favoring the intermolecular addition over intramolecular cyclization reaction64.

The absence of solvent also allows rising temperature resulting in a higher reaction rate and a lower viscosity of the reacting mixture, assisting to compensate the higher viscosity found on the solvent-free systems. The combination between increased monomer concentration and reaction temperature resulting in a reduction in the reaction time64. The solvent-free reactions also allows the generation of higher molecular weight polymers, besides increasing the reaction rate and obviating the solvent removal step53;64.

After polymerization, PbAE can be precipitated, normally in cold diethyl ether, hexane42, ether65 or ethyl ether58 and/or then dried under vacuum57;65. Frequently, PbAEs are immediately used or stored in the cold conditions (4 _C52;66;67, 0 _C62, or -20 _C68–70). Some PbAEs should be also kept airproof due to its strong moisture absorption ability and easy degradation71.

Concerning to the biodegradation and biocompatibility, PbAEs have been shown generally to possess low cytotoxicity and good biocompatibility42;52;61;55;72. Different studies have suggested that PbAEs are significantly less toxic than currently available cationic polymers, such as, PEI and PLL51;64. Nevertheless, the increase of the number of carbons in the backbone or side chain is associated to the increase of the cytotoxicity73. PbAE degrade under physiological conditions via hydrolysis of their backbone ester bonds to yield small molecular weight b-amino acids biologically inert derivatives42;51;55;74. Some results revealed that the degradation rate of poly(b– amino ester)s was highly dependent on the hydrophilicity of the polymer, i.e., the more hydrophilic the polymer is, the faster the degradation occurs75;76.

In Table 1.4 are summarized the main characteristic of PbAEs which make them a promising polymeric non-viral vector for gene delivery.

Combinatorial libraries – a fast and efficient way to evaluate different poly(b–amino ester)s

A fast and efficient way to study the relationships between structure and function in particular material that could be prepared with different reagents is using combinatorial libraries. Due to promising preliminar results of PbAEs as non-viral vectors, Langer research group reported a parallel approach for the synthesis of hundreds of PbAEs with different structures and the application of these libraries to a rapid and high throughput identification of new transfection reagents and structure-function trends. For this purpose, major contributions have been reported52;53;57;66;67;72;75;77;78 not only exploring the possible structure/function relationships, but also imposing an assortment of monomers (amines were denoted by numbers and acrylates by latin alphabet letters) used in order to facilitate cataloging of different PbAEs (Table 1.5 and Tables A.1 and A.2 (Appendix A)).

The first initial library screening was synthesized in 2001 by Lynn51. 140 Different PbAEs from 7 diacrylates and 20 amines were prepared with molecular weights between 2,000 and 50,000 g.mol-1. From this, polymers C93 (Mw = 3180 g.mol-1) and G28 (Mw = 9170 g.mol-1) revealing transfection levels 4-8 times higher than control experiments employing PEI. At same time, it was observed that for transfection efficiency, high molecular weight was not an important parameter. This work was then completed in 2003 by Akinc57, where biophysical properties and the ability of each polymer/DNA complex to overcome important cellular barriers to gene deliver were investigated. As previous experiments, complexes formed from polymers C93 and G28, revealed higher levels of internalization compared to ”naked” DNA, displaying 18- and 32-fold more internalization, respectively. In contrast, the majority of the polyplexes were found to be uptake-limited. Regarding diameter and zeta potential, out of 10 polymer/DNA complexes with the highest internalization rates, all had diameters lower than 250 nm and 9 had positive zeta potentials. By measuring the pH environment of delivered DNA through fluorescence-based flow cytometry protocol using plasmid DNA covalently labeled with fluorescein (pH sensitive) and Cy5 (pH insensitive) it was possible to investigate the lysosomal trafficking of the polyplexes. The results demonstrated that complexes based on polymers C93 and G28 were found to have near neutral pH measurements, indicating that they were able to avoid acidic lysosomal trafficking. In the same year, Akinc64 studied the effect of polymer molecular weight, polymer chain end-group, and polymer/DNA ratios on in vitro gene delivery. For this purpose, 12 different structures were synthesized based only in two different PbAE (C28 prepared from 1,4-butanediol diacrylate and 1-aminobutanol and E28 prepared from 1,6-hexanediol diacrylate and 1-aminobutanol) (Figure 1.6.)

These structures were synthesized by varying amine/diacrylate stoichiometric ratios, resulting in PbAEs with either acrylate or amine end-groups and with molecular weights ranging from 3,350 to 18,000 g.mol-1. Polymers were then tested, using high throughput methods, at nine different polymer/DNA ratios between 10/1 (w/w) and 150/1 (w/w). Concerning terminal groups, it was found that amino-terminated polymers transfected cells more effectively than acrylate-terminated polymers. In contrast, none of the acrylate terminated PbAEs mediated appreciable levels of transfection activity under any of the assessed conditions. These findings suggest that end-chains of PbAE have crucial importance in transfection activity. Concerning molecular weight effect, highest levels of transfection occurred using the higher molecular weight samples of both amine-terminated C28 (Mw _13100 g.mol-1 and E28 (Mw _13400 g.mol-1). Regarding the optimal polymer/DNA ratios for these polymers, it was observed a markedly difference, 150/1 (w/w) for C28 and 30/1 for E28. These results highlighted the importance of polymer molecular weight, polymer/DNA ratio, and the chain end-groups in gene transfection activity. Moreover, it has found the fact that two similar polymer structures, differing only by two carbons in the repeating unit, have different optimal transfection parameters emphasizing the usefulness of library screening to perform these optimizations for each unique polymer structure. Meanwhile, in 2003, Anderson52 described, for the first time, a high-throughput and semi-automated methodology using fluid-handling systems for the synthesis and screening of a library of PbAEs to be used as gene carrier.

A crucial feature of these methods was that all process of synthesis, storage, and cell-based assays were performed without removing solvent (DMSO). By using these methods, it was possible to synthesize a library of 2350 structurally unique, degradable and cationic polymers in a single day and then test those as transfection reagent at a rate of _1000 per day. Among PbAEs tested, it was identified 46 polymers that transfect in COS-7 as good as or better than PEI. The common characteristic among them was the use of a hydrophobic diacrylate monomer. Moreover, in the hit structures mono- or dialcohol side groups and linear, bis(secondary amines) are over represented. From data obtained from this library, Anderson67, in 2004, continued his study developing a new polymer library of >500 PbAE using monomers that led higher transfection efficiency in the previous studies and optimizing their polymerization conditions. The top performing polyplexes were assessed by using an in vitro high-throughput transfection efficiency and cytotoxicity assays at different N/P ratios. As previously observed, the most promising polymers are based on hydrophobic acrylates and amines with alcohol groups. Among those, C32 stood out due to higher transfection activity with no associated cytotoxicity. The efficiency to deliver DNA was evaluated in mice after intra-tumoral (i.t.) and intra-muscular (i.m.) injection. The results revealed important differences. While by i.t injection C32 delivered DNA 4-fold better than jetPEI R , a commercial polymeric non-viral vector, by i.m. administration transfection was rarely observed. C32 was then assessed for DNA construct encoding the DT-A (DT-A DNA) deliver to cells in culture and to xenografts derived from androgen-sensitive human prostate adenocarcinoma cells (LNCaP). Results showed that DT-A DNA was successfully delivered and the protein expressed in tumor cells in culture. In human xenografts, the growth was suppressed in 40% of treated tumors. The fact of C32 is non-toxic and it is able to transfect efficiently tumors locally and transfects healthy muscle poorly turned it as a promising carrier for the local treatment of cancer.

From here, a panoply of results based in PbAE combinatorial library appeared. In 2005, Anderson53, prepared a new library of 486 second-generation PbAE based on polymers with 70 different primary structures and with different molecular weights.

These 70 polymers were synthesized using monomers previously identified as common to effective gene delivery polymers. This library was then characterized by molecular weight of polymers, particle size, surface charge, optimal polymer/DNA ratio and transfection efficiency in COS-7 of polymer/DNA complexes. Results showed that from 70 polymers with primary structures, 20 possess transfection activities as good as or better than Lipofectamine R 2000, one of the most effective commercially available lipid reagents. Results also revealed that, in general, the most effective polymers/DNA complexes had <150 nm of particle size and a positive surface charge. Among them, the 2 most effective PbAEs complexed with DNA in a smallest particle sizes, 71 nm (C32) and 79 nm (JJ28), and have positively surface charge (over 10 mV) nanoparticles. Interestingly, the 9 most effective polymer structures were all constituted with amino alcohols, and the chemical structure of the 3 best performing PbAEs (C28, C32 and JJ28) differs by only one carbon. These results show a convergence in structure of the top performing polymers and suggest a common mode of action providing a framework for the design of efficient gene delivery systems.

In 2006, Green79, synthesized, on a larger scale and at a range of molecular weights, the top 486 of 2350 PbAEs previously assessed52 and studied their ability to deliver DNA. These PbAEs were tested, firstly, on the basis of transfection efficacy in COS-7 cells in serum-free conditions, and then, the 11 of the best-performing PbAEs structures were further analyzed. The transfection conditions were optimized in human umbilical vein endothelial cells (HUVECs) in the presence of serum. In this study, the influence of the factors like polymer structure and molecular weight, and biophysical properties of the polyplexes (such as, particle size, zeta potential, and particle stability throughout time) were studied. The results showed that many of the polyplexes formed have identical biophysical properties in the presence of buffer, but, when in the presence of serum proteins their biophysical properties changed differentially, influencing the transfection activity. Concerning to the size, the results showed that in spite of all vectors condensed DNA into small particles below 150 nm in buffer, only a few, such as C32, JJ32 and E28, formed small (_200 nm) and stable particles in serum. C32, JJ32 and E28 revealed also high transfection activity both in the absence of serum in COS-7 cell line as in the presence of serum in HUVEC cell line. Moreover, C32 transfected HUVECs in the presence of serum significantly higher than jetPEI R and Lipofectamine R 2000, the two top commercially available transfection reagents. The 3 mentioned PbAEs share a nearly identical structure.

The acrylate monomers of these polymers, C, JJ, and E, differ by only their carbon chain lengths (4, 5, and 6 carbons, respectively). Similarly, amines 20, 28, and 32 differ also by only the length of their carbon chain (3, 4, and 5 carbons, respectively).

For example, polymers prepared with the same acrylate monomer (C) in which itwas increased the length of the carbons chain of the amine monomer resulted in an increased transfection efficacy (C32 (5 carbons) > C28 (4 carbons) > C20 (3 carbons)) of these polymers-based polyplexes. Interestingly, this study reinforced C32 as the lead PbAE vector and revealed other potential two, JJ28 and E28, which previously showedto be poor vectors. On the other hand, C28 and U28, previously recognized as an efficient transfection reagent, were found to transfect inefficiently HUVEC in serum. By constructing a new library of end-modified PbAE, the research was continued78 in order to understand the structure-function relationship of terminal modification of PbAE in transfection activity. For this purpose, it was used twelve different amine capping reagents to end-modify C32, D60 and C20. The choice of these 3 PbAEs was based in their transfection activity: C32, the most effective; D60, an effective transfection reagent with a significantly different structure from that of C32; and, C20, a poor transfection reagent but with similar structure to C32 differing only in the length of the amine monomer. The results showed that some PbAEs-based vectors (C32-103 and C32-117) were able to deliver DNA by approximately two orders of magnitude higher than unmodified C32, PEI (25,000 g.mol-1) or Lipofectamine R2000, and, at levels comparable to adenovirus at a reasonably high level of infectivity (multiplicity of infection = 100). Once again, it was demonstrated that small structural changes influence greatly gene delivery, from biophysical properties (such as, DNA binding affinity, particle size, intracellular DNA uptake) until final protein expression. From these 3 polymers assessed, C20 was the one who transfected cells much less effectively, although it has seen a remarkably improvement with end-modifications. As expected, C32-based polyplexes, based on C32-103 and C32-117, revealed the higher transfection efficiency enhancing cellular DNA uptake up to five-fold compared to unmodified C32. Interestingly, and in a general way, terminal modifications of C32 with primary alkyl diamines were more effective than those with PEG spacers, revealing that a degree of hydrophobicity at the chain ends is an added value for these polymers. Another interesting fact in terminal modification of C32 was that at least a three carbon spacer between terminal amines is necessary to obtain an efficient gene delivery. For example, results showed that C32-103 transfection efficiency is 130- and 300-fold higher than C32-102 on the COS-7 and HepG2 cell lines, respectively. As the molecular weight was the same, this result demonstrated the critical role of the chain ends in transfection activity.

In order to better understand the role of the chain ends in transfection efficiency a new library of end-modified C32 was synthesized by Zugates80 in 2007 using 37 different amine molecules to end-modify the PbAE. In a general way, it was observed that polymers end-capped with hydrophilic amine end groups containing hydroxyls or additional amines led to higher transfection efficiency. On the other hand, terminal-modifications with hydrophobic amines containing alkyl chains or aromatic rings proved to be much less effective. Concerning to cytotoxicity, terminal modification with primary monoamine reagents (independently of functional group extending from the amine, such as aromatic, alkyl, hydroxyl, secondary and tertiary