Carbosilane dendrimers as carriers of siRNA

J. DRUG DEL. SCI. TECH., 22 (1) 75-82 2012

Carbosilane dendrimers as carriers of siRNA J.L. Jiménez1, 4, R. Gómez2, 4, V. Briz1, R. Madrid1, M. Bryszewsk1, 3, F.J. de la Mata2, 4, M.Á. Muñoz-Fernández1, 4* Plataforma de Laboratorio, Laboratorio de Inmunobiología Molecular, Hospital General Universitario Gregorio Marañón, C/Doctor Esquerdo 46, 28007 Madrid, Spain 2 Departamento de Química Inorgánica, Universidad de Alcalá, Campus Universitario, Alcalá de Henares, Spain 3 Departament of General Biophysics, University of Lodz, Lodz, Poland 4 CIBER de Bioingeniería, Biomateriales y Nanomedicina (CIBER-BBN), Instituto de Salud Carlos III, Spain *Correspondence: [email protected]

1

Despite the enormous possibilities of RNAi, there still exist many problems that need to be addressed. Obstacles in delivery, target cell transfection, stability/degradation, transient activity, secondary effects, toxicity caused by the delivery vector, and resistance all hinder the path of carrying out in vivo experiments with RNAi and further developing RNAi as a new therapy for clinical use. Notwithstanding, the majority of research that uses RNAi depends on a delivery vector of some kind. This review offers a brief overview of the current status of carbosilane dendrimers as siRNA delivery vectors. Key words: RNAi – Carbosilane dendrimers – Dendrimers – Non-viral vectors – HIV infection.

Despite the enormous possibilities of RNAi, there still exist many problems that need to be addressed. Obstacles in delivery, target cell transfection, stability/degradation, transient activity, secondary effects, toxicity caused by the delivery vector, and resistance all hinder the path of carrying out in vivo experiments with RNAi and further developing RNAi as a new therapy for clinical use. However, a great deal of these issues can be addressed by using the correct delivery vector. Thus, a great deal of research is being put into finding a suitable partner and chauffeur for these valuable small nucleic acid molecules. The ideal delivery agent should protect the siRNA from degradation, transport it to the target cells or tissues, facilitate its transfection into the cytoplasm of the cells and all the while should remain fairly innocuous, causing little or no adverse effects.

I. Drug Delivery

The main limiting problems with developing gene therapy and RNAi into pharmaceuticals are their ineffectiveness in vivo and the difficulties in converting them into clinical therapies. Although some success has been achieved by localized delivery to such areas as the eye, skin, or local tumors [1-5], systemic RNAi therapy still faces many issues [6]. These problems arise from the difficulty of directing the plasmids or oligonucleotides to target cells or tissues and then being able to successfully transfect them into the cell [7-9] where they can exert an effect. In vivo, free oligonucleotides are rapidly filtered from the blood by the kidneys or are processed by the reticulo-endothelial system and delivered to the liver and spleen [10]. This causes problems if the target cells are located in other regions of the organism. Furthermore, oligonucleotides, siRNA and plasmids are very unstable once in circulation because of the presence of nucleases and serum proteins in the bloodstream which kidnap and degrade these potential therapeutic agents. The area of study dedicated to addressing these problems is termed “drug delivery”. Despite the existence of several methods that allow successful transfection of cells with siRNA oligonucleotides in vitro, transferring these methods to in vivo studies is often unfeasible.

Figure 1 - Physiological barriers to the systemic delivery of siRNA vectors. A systemic delivery system confronts numerous obstacles in reaching its objective: 1) filtration, degradation and phagocytosis in the blood stream; 2) crossing the endothelial barrier; 3) transversing the extracellular matrix to the target cell; 4) cell transfection; 5) endosomal escape: 6) delivery of siRNA to cell cytoplasm.

barriers that must be overcome if the delivery system is to achieve stability in circulation. Crossing the vascular endothelial barrier presents a problem for most delivery systems (Figure 1, 2). This is especially true in the central nervous system (CNS) where difficulties in crossing the blood brain barrier (BBB) with small molecule pharmaceuticals are widely recognized. It is believed that particles with diameters greater than 5 nm do not readily cross the vascular endothelial barrier. Some organs, including the liver, spleen, lungs, and kidneys, can be reached with larger molecules due to accumulation in the reticuloendothelial system, which is a method of eliminating foreign particles from circulation. Likewise, some tumors can be accessed with larger molecules because of increased vasculature permeability in tumors referred to as the enhanced permeability and retention effect [6]. However, for diseases involving leukocytes or other parts of the im-

II. Physiological obstacles to systemic RNAi therapy

There exist numerous problems with applying systemic RNAi therapy in vivo (Figure 1). Aside from the stability, degradation and filtration issues in the bloodstream (Figure 1, 1), there are several other 75


Carbosilane dendrimers as carriers of siRNA J.L. Jiménez, R. Gómez, V. Briz, R. Madrid, M. Bryszewsk, F.J. de la Mata, M.Á. Muñoz-Fernández

mune system, obtaining stable circulation in the bloodstream may be sufficient to reach all the cells affected by the disease. HIV infection and other immunological and neurological disorders fall under this category. However, the extent with which the virus may infect other non-immune system cells such as neuronal cells [11] and establishes regions of latently infected cells that do not circulate throughout the bloodstream is still under research. Once a drug crosses the endothelial barrier, the extracellular matrix possesses a myriad of obstacles, like polysaccharides and fibrous proteins, that hinder the progression of the treatment on their path to the target cells (Figure 1, 3). Whether the target cells are reached by crossing the endothelial membrane and penetrating the thick forest of the extracellular matrix or are reached from within the bloodstream, the barrier of crossing the cell membrane and gaining access to the interior of the cell is the next obstacle that must be overcome (Figure 1, 4). Indeed, cell transfection/uptake is one of the principle topics to which research in drug delivery is devoted [12]. The methods by which particles are taken up are receptor mediated endocytosis and membrane fusion [13]. Viral vectors use the viruses’ evolutionarily developed methods for evading obstacles and gaining entrance to the cell. Many of these vectors use cell membrane receptors to facilitate fusion with the membrane, which allows for the deposit of their genetic payload to the interior of the cell. Non-viral vectors face greater challenges in achieving entry to the cell. Non-targeted vectors are believed to interact with the cell membrane via electrostatic interactions that cause endocytosis. Targeted vectors use ligands that are recognized by cell membrane receptors and their association initiates endocytosis [14]. Endocytosis of delivery vehicles results in their localization in endosomal vesicles from which the delivery system must achieve escape (Figure 1, 5). If all these obstacles can be overcome, the final step to be completed is the release of the intact and functional DNA or siRNA so that it can exert its effect (Figure 1, 6). In the case of plasmid DNA delivery for gene therapy, the additional problem of achieving access to the cell nucleus must be addressed. Notwithstanding, the majority of research that uses RNAi depends on a delivery vector of some kind. Drug delivery vectors can be divided into two different categories. Viral vectors are chiefly used to transport and deliver genes, while non-viral vectors are the method of choice for smaller oligonucleotides such as siRNA. In many cases, viral vectors are not a feasible option, and fortunately, there exist many kinds of non-viral vectors that can function just as well as viral vectors. Dendrimers are one of the most used types of nanoparticles for gene therapy applications.

The dendrimer structure can be divided into three different parts: i) a central core from which the arms extend outward that contains a microenvironment protected by the external structural components, ii) multiple branching arms that radiate out from the central core whose structure, length and number can all be altered depending on the desired size and functionality, and external functional groups located at the periphery of the branches and, therefore, on the molecular surface. These groups generally provide the primary functionality of the dendrimer. Similarly, the dendrimer structure can be viewed as having layers, or “generations”. The number of generations a particular dendrimer has depends on the number of branching or focal points that exist in the branching arms as they extend outward from the core [20]. Since at each focal point, the number of arms extending outwards increases by at least a multiple of two, the higher the generation, the greater the number of functional groups at the extremities. For a typical dendrimer that doubles the number of exterior branches at each focal point, a fourth generation dendrimer would possess thirty-two functional groups at the periphery. Monodispersity refers to the homogeneity in which the dendrimers are synthesized (i.e., all dendrimer molecules are made with the same structure, generation, and amount of branching). The branching structure of dendrimers allows them to possess multiple functional groups on the branch extremities all radiating out from a central core [21]. A primary advantage of dendrimers is the flexibility in which they can be generated. This allows for a wide variety of functional groups to be located along the branches as well as at the far extremities. The structure of the molecule depends entirely on the function it is designed to carry out. Internal pockets near the core of the dendrimer can provide microenvironments to contain and protect small molecules [19]. In other cases, external charged groups function to tightly bind oppositely charged molecules such as DNA or oligonucleotides via electrostatic interactions. The size of dendrimers depends on the structure and, in particular, the number of generations that the dendrimers possess. Dendrimers range in size from around 1.5 nm to 15 nm in diameter. This size range categorizes dendrimers as nanoparticles and allows them to imitate proteins in terms of their interactions with small molecules or other macromolecules [22]. This characteristic is essential for these synthetic molecules to achieve functions based on biological interactions. Such functions include imaging probes, vaccine adjuvants, and the encapsulation, delivery and/or controlled release of drugs [23-25]. In addition to their size, dendrimers benefit from their unique structural characteristics in order to function as biologically useful molecules. Dendrimers make use of this structural characteristic to optimize their biological interactions at a molecular level [26]. Furthermore, the highly branched nature of the structure results in a high density of functional groups at the molecular surface. This attribute, in addition to maximizing the efficiency of the dendrimer in forming molecular interactions, could also theoretically allow for the use of several different types of peripheral groups on the same molecule in order to achieve multiple functionalities. A possible example would be dendrimers with cell targeting moieties as well as cationic functional groups to bind oligonucleotides or siRNAs to achieve cell specific delivery.

III. Dendrimers

Dendrimers are defined as nanoscopic, monodisperse, polybranched, synthetic polymers [15]. The term “dendrimer” comes from the Greek “dendros” for tree because of the tree-like or branching nature of the dendrimer structure. Due to a diameter size of from 1.5 to 15 nm [16], dendrimers are considered to be nanoscopic. Dendrimers are synthesized under highly controlled conditions so that the product is as close to completely monodisperse as possible. This characteristic comes from a system of synthesis in which many steps are followed to gradually create a uniform structure and all of the dendrimer molecules have the exact same structure. The reason for creating nanoparticles with such a specific structure is for the functionality that arises because of that structure. As in the case with proteins, the macromolecules that dendrimers were largely designed to imitate, structure determines function. The highly systematic synthesis allows for a great deal of control in size, shape and terminal group functionality [17, 18]. Dendrimers are synthesized by two different methods [19]. One method of synthesis called “divergent synthesis” generates the molecule from the core outwards, while the other, “convergent synthesis”, initially starts with the external functional groups and generates the dendrimer by joining the external branches together and working from the outside in.

IV. Carbosilane dendrimers

Carbosilane dendrimers with a silicon branch point in an exclusively carbon-silicon skeleton are non-polar, inert, neutral and thermally and hydrolytically stable compounds [27]. The absence of polar bonds facilitates the use of many derivatization reactions and creates the possibility of a strong physic-chemical contrast between the core and the outer corona. The synthesis of dendrimers is almost always by a divergent process from the core to the interior generations and to periphery, with the number of reactions per dendrimer increasing geometrically with each generation. The divergent synthesis 76



of carbosilane dendrimers consists of generational repetition of a sequence of two clean, high-yield reactions: a) hydrosilylation and b) nucleophilic substitution by Grignard or organolithium reagents. The hydrosilylation reaction (A) introduces the branch juncture and creates the next generation while the substitution reaction (B) introduces the branch. Both reactions are performed with excess of highly reactive hydrosilane and Grignard or organolithium compounds in order to promote quantitative conversions. Due to the high sensitivity of these reagents to oxygen, moisture, carbon dioxide, etc., these reactions have to be protected by high vacuum and/or Schlenk techniques.

physiological conditions and are responsible for ionic condensation with negatively charged siRNA molecules. The siRNA/dendrimers, dendriplexes, are able to bind to the cell surface, enter the cells through endocytosis before being trapped within endosomes where they release the siRNA into the cytosol [51]. Dendrimers harboring tertiary amines in their interior can preferentially promote siRNA release via the “proton sponge” effect [52]. The released siRNA molecules eventually join the RNAi machinery where gene silencing can occur. Although carbosilane dendrimers have different uses in the diagnostic and clinical therapy, this review offers a brief overview of the current status of carbosilane dendrimers as siRNA delivery vectors. Novel amine or ammonium-terminated carbosilane dendrimers of type nG-[Si{OCH 2 -(C 6 H 3 )-3,5-(OCH 2 CH 2 NMe 2 ) 2 }] x , nG-[Si{O(CH2)2N(Me)(CH2)2NMe2}]x and nG-[Si{(CH2)3NH2}]x or nG[Si{OCH2-(C6H3)-3,5-(OCH2CH2NMe3+I–)2}]x, nG-[Si{O(CH2)2N(Me) (CH2)2NMe3+I–}]x and nG-[Si{(CH2)3NH3+Cl–}]x (Figure 3) were synthesized up to the third generation and fully characterized based on two strategies: (i) the alcoholysis of Si-Cl bonds by the use of amine-alcohols and the subsequent quaternization with MeI, and (ii) the hydrosilylation of allyl amine with Si-H bonds of the dendritic systems and the latter quaternization with HCl [30-32, 53, 54]. Quaternized carbosilane dendrimers are soluble in water, although degradation is apparently via hydrolysis of Si-O bonds. However, hydrolysis can be prevented or reduced, under very dilute conditions, by means of the introduction of a rigid phenyl group as a linker between the amine or ammonium groups and the Si-O bonds. In the same way, dendrimers containing Si-C bounds were shown to be water-stable [55].

A) Hydrosilylation reaction n

+

Pt

R(3-m)ClmSi-H

n

SiR(3-m)Clm

B) Substitution reaction n

SiR(3-m)Clm + H2 C=C-(CH2 )n -MgX H

n

SiR(3-m)[(CH2 )n –CH=CH2 ]m

In convergent synthesis, a dendron containing the peripheral groups is repeatedly added via its focal group to a branched molecule and in the final step to a multifunctional core molecule. A possible scheme for a convergent synthesis of a carbosilane dendron consists of the repletion of the following two reactions (C and D) where the definition of subscripts m and n are the same for the divergent synthesis. The convergent approach has seldom been applied to carbosilane dendrimers.

C)

R1 R3-m R2 − Si−H + CH2 =CH-(CH2 )n Si −Cl m 3 R

D)

R1 R3-m R2 − Si−CH2 -CH2 -(CH2 )n Si −Cl 3 R m

Pt

LiAlH4

R1 R3-m R2 − Si−CH2 -CH2 -(CH2 )n Si −Cl 3 m R R1 R3-m R2 − Si−CH2 -CH2 -(CH2 )n Si −H 3 m R

Embryonic carbosilane dendrimers with 12 and 18 Si-Cl reactive peripheral (i.e., end) groups were described as early as 1978 and 1980 [28, 29]. The first synthesis of carbosilane dendrimers was reported almost simultaneously by three groups in the early 1990s [30-32], and several reviews of the field were published recently [33-38] (Figure 2).

Figure 2 - Two-dimensional representations of selected carbosilane dendrimers. A) 1G (4,2) dendrimer with ethanediyl (C2) branches and vinyl end groups; B) 2G(4,2,2) dendrimer with C2 branches and C1 end-groups; C) 2G (4,3,3) dendrimer with propanedyl (C3) branches and allyl end-groups. These figures do not include expected backfolding of a fraction of dendrimer branches.

V. Creation of an effective combination: carbosilane dendrimer-siRNA

Up to now, a multitude of dendrimers have been explored for siRNA delivery, including poly(amidoamine) (PAMAM) dendrimers, poly(propylene imine) (PPI) dendrimers, poly(L-lysine) dendrimers, triazine dendrimers polyglycerol dendrimers, polyphenylenevinylene PAMAN dendrimer, polycationic “viologen”-based dendrimers and carbosilane dendrimers [39-50]. These dendrimer vectors bear positively charged amine functionalities at the dendrimer surface under

Figure 3 - Molecular representations of the second generation ammonium-terminated carbosilane dendrimers. A) IM8: 2G-(NMe3I)8; mw= 3011 g/mol; 8 positive chargers; B) IM16: 2G-(NMe3I)16; mw = 4731,59 g/mol; 16 positive chargers; C) ClNH4: 2G-(NH3Cl)8; Mw= 1927,6 g/ mol; 8 positive chargers, D) Phe: 2G-(Ph(NMe3I)2)8; mw= 5692,80 g/ mol; 16 positive chargers; E) NN: 2G-(NNMe3I)8; mw= 3468,08 g/mol; 8 positive chargers. mw: molecular weight 77



macrophages and PBMC, that ranged from immortalized adherent cells to suspension cells and primary cells showed that the dendriplex causes less toxicity than the dendrimer alone [46-48, 55-57]. Noncomplexed dendrimer without the neutralizing presence of siRNA has all of its positively charged groups exposed. This indicates that the toxicity is most likely a result of the high positive charge density, an attribute that can be improved upon or shielded through the use of stabilizing moieties like polyethylenglycol-graft-trimethyl [62]. On top of the good cytotoxicity profile, the dendriplex distinguishes itself from other transfection agents like Lipofectin in that it functions in medium containing serum and antibiotics while with Lipofectin these additives must be removed. This is a fundamental advantage that allows dendrimers to make the transition to in vivo scenarios. A general problem for gene therapy in chronic viral infections or other immunological or neurological diseases is that these cells are very difficult to transfect. For example, the study of the role of different proteins in neuronal physiology or pathology requires an approach that should include the selective knockdown of such proteins to study a lack-of–function effect. siRNA is becoming a very popular method to acutely and selectively knockdown specific proteins. This method allows the role of a specified protein in different physiological and pathological functions to be studied [63]. However, the general use of siRNA technology in postmitotic neuronal cells has mainly relied on the use of viral vectors with neuronal tropism. The transfection efficiency of these vectors is estimated to range from 70 to 90 % for either adenoviral [64], adeno-associated [65] or lentiviral vectors [66]. However, viral vectors have several drawbacks such as the complexity of preparation and possible immune and inflammatory responses described for adenoviruses when used in animals, including humans [64, 67]. These drawbacks have led to the search for new methods, based in non-viral vectors, for the safe and efficient delivery of genetic material to postmitotic neurons. However, the classical non-viral vectors have shown low transfection efficiency ranging from 0.01 % for calcium phosphate co-precipitation [68] to about 25 % using Lipofectamine 2000 [69]. The 2G-NN16 carbosilane dendrimer coupled to siRNA has a transfection efficiency of about 85 % in rat cortical neurons [48]. This is similar to that achieved using viral vectors [66] and much higher than that obtained with other methods previously described [68, 69]. None of the methods previously described results in a reduction in protein levels to an extent that would allow the study of a lack-of-function effect. Moreover, human astrocytoma cells, PBMC, T lymphocytes and dendritic cells were successfully transfected by fluorochrome-labeled siRNA complexed with the 2G-NN16 carbosilane dendrimer. These data indicate that this is a new, simple, fast and non-toxic method to efficiently deliver siRNA to cells from the immune and nervous systems. In addition, the introduction of the dansyl chromophoric group (5-(dimethylamino)-1-naphthalene-sulfonamido) at the periphery of carbosilane dendrimers as a potent fluorescence label demonstrated that systems based on carbosilane dendrimers can be internalized into the cells [70]. Therefore, confocal microscopy images showed the uptake of siRNA by T lymphocytes and human astrocytoma cells using 2G-NN16 carbosilane dendrimer [46-48, 71] and the similarity of the compartmentalized fluorescent siRNA with perinuclear endosomes in terms of morphology and spatial localization indicate the possibility of the dendriplexes entering via endocytosis. In several polarized T cells, fluorescence appears clearly accumulated in the uropod. This peculiar localization only occurs in polarized lymphocytes [72], offering further evidence of endocytosis. On the other hand, the use of nanoparticles for siRNA delivery to the brain remains at the experimental stage. Nanoparticles of size