Nonviral Vectors for the Delivery of Small Interfering RNAs to the CNS

Inmaculada Posadas; Francisco Javier Guerra; Valentín Ceña


Nanomedicine. 2010;5(8):1219-1236. 

In This Article

Cationic Lipids & Cationic Liposomes

In 1987 the term 'lipofection' was used for the first time when Felgner and coworkers described it as a highly efficient lipid-mediated DNA-transfection procedure using the synthetic cationic lipid (N-2,3-dioleoyloxypropyl)-N,N,N-trimethylammonium (DOTMA).[51] Depending upon the cell line, lipofection was found to be five- to ten-fold more effective than calcium phosphate or dextran, which were the two transfection techniques available at the time.[51] The combination of DOTMA, a cationic amphiphile lipid, with dioleoylphosphatidylethanolamine (DOPE), a neutral phospholipid, was commercialized as the well-known reagent Lipofectin®. Subsequently, the combination of a cationic cholesterol derivative (3β-[N-(N'N'-dimethylaminoethane)-carbarmoyl cholesterol][DC-Chol]) combined with DOPE was reported to improve Lipofectin efficiency transfection with lower toxicity in A431, A549 and L929 cell lines.[52]

Simultaneously, it was described that the lipospermine dioctadecylamidoglycylspermine produced an efficient transfection in a broad range of cell lines of diverse origin (LMTK, Ras4, CHO, F9, BU4, HeLa and AtT-20 cells) and, interestingly, in primary cultures, including anterior pituitary cells, chromaffin cells and neurons, it was an approach that was not available at that time.[53]

Cationic lipids were used in combination with retroviral vectors to improve the rate of transduction in target cells. DC-Chol/DOPE was shown to enhance the transduction efficiency of a retroviral vector without altering the basic properties of retroviral transduction in TE671 and NIH 3T3 cell lines.[54] Moreover, in a nude-mouse xenograft model of mesothelioma, DC-Chol/DOPE complexed with murine leukemia virus-A administered intraperitoneally induced a 3.9-fold increase in transduction efficiency, compared with injection of murine leukemia virus-A alone.[54]

Cationic lipids were extensively studied as a new tool to efficiently deliver DNA,[51,53,55] mRNA,[56] antisense oligomers[57] and proteins[58] into living cells. Several reagents based in cationic lipids became commercially available, including LipofectACE®,[59] Lipofectamine® or Lipofectamine ®.[60]

Transfection of Neurons in vitro

Despite the good prospects and great advantage in transduction reported for these cationic lipids in vitro in a great variety of cell lines, low transfection efficiencies have been consistently reported for primary cultures, especially for CNS cells, including neurons and glial cells. It has been reported that 1,2-dioleyloxy-3-(trimethylammonium)propane (DOTAP) has an efficiency of 0.5–3% in primary cultures of rat hippocampal neurons;[61] Lipofectin is a very poor transfectant (0.02–0.5%) in primary septohippocampal neuronal cultures;[62] and Transfectam® yields 1–5% efficiency in rat striatal neurons,[63] whereas an efficiency of 2.4% can be obtained using Fugene 6 in both primary glial and primary neuronal cultures.[64] Lipofectamine 2000 increases transfection efficiency to approximately 25% in primary E18 rat cortical neurons and E18 rat hippocampal neurons.[16,17,65]

Thus, except for Lipofectamine 2000, the maximum transfection efficiency reported for delivery of plasmids using cationic lipids is relatively low. Although this transfection efficiency might be enough to perform single cell studies, it falls short for the required efficiency to carry on lack-of-function studies.

The small transfection efficiency was attributed to insufficient endosomal escape of cationic lipids which meant that the endocytosed DNA remained trapped in the endosome pathway, where it is degraded, lowering the efficiency of the gene transfer process.[66,67]

Another problem with transfection derives from the fact that nonviral gene carriers must traverse multiple extracellular and intracellular barriers, including delivery of DNA through the nuclear membrane. It is difficult for large DNA molecules to cross the nuclear membrane and cytoplasmic nucleases contribute to DNA degradation, thus reducing transfection efficiency.[68] While less than 1% of plasmid DNA (pDNA) delivered to the cytoplasm eventually reaches the nucleus, pDNA injected directly into the nucleus shows high gene expression.[69] In dividing cells, the nuclear membrane breaks down at the end of each mitosis and hence allows for passive inclusion of transfected DNA. However, in nondividing cells, including the postmitotic neuronal cells in which mitotic activity is absent,[70] low pDNA nuclear translocation occurs, which is probably the result of passive movement through the nuclear pore complex[71] or by fusion of lipoplexes with the nuclear membrane.[72] Aronsohn and Hughes demonstrated that the amount of DNA able to cross the nuclear envelope in nonproliferating cells by simple diffusion is negligible.[73]

Thus, different strategies were developed in order to improve gene delivery in primary neuronal cultures. One interesting approach was the use of the cationic lipid 1,2-dioleoyl-sn-glycero-3-succinyl-2-hydroxyethyl disulfide ornithine (DOGSDSO), which contains a disulfide bond between the positively charged head group and the lipophilic backbone, on the basis that disulfide linkage of lipids may be cleaved by reductive substances in the cytosol leading to higher concentrations of DNA.[74] Quantification of luciferase activity and protein content demonstrated that introduction of disulfide bonds into the lipid structure resulted in more efficient transfection compared with nondisulfide formulations. Gene expression using DOGSDSO was much higher in glia than in neurons.[75] However, the transfection efficiency of these formulations was markedly lower than that obtained in the neuroblastoma cell line SK-N-SH.[75]

Another strategy consisted of the addition of molecules that undergo conformational changes to disrupt endosomal membranes when exposed to the low pH of the endosome. One of these molecules is the pH-sensitive surfactant dodecyl 2-[1'-imidazoyl propionate] (DIP) that becomes ionized at low pH, leading to surfactant properties inside the endosome.[76] The addition of DIP to a mixture of DOTAP/DOPE increased the transfection efficiency approximately threefold compared with DOTAP/DOPE alone in primary glial and neuronal cultures. However, the percentage of neurons or glial cells transfected was still low.[77]

A different approach was the combination of cationic lipids with peptides or proteins in order to increase transfection. The combination of Lipofectamine 2000™ with an amino acid synthetic peptide (polylysine molossin) yielded a transfection efficiency higher than 30% in primary neuron cultures with no detectable toxicity.[78] Addition of human transferrin (Tf; an iron-transporting protein that interacts with receptors ubiquitously expressed in tissues) to liposome/DNA complexes resulted in high levels of transfection in a large variety of dividing cells. However, association of transferrin to DOTAP:Chol, DOTAP:DOPE or DOTAP:DOPE:Chol liposomes produced low levels (~4%) of transfection in both primary hippocampal and cortical neurons, whilst a detected toxicity of 10–15% indicated that caution should be taken when such vectors are used to mediate gene delivery in postmitotic neurons.[79]

Conversely, Rakotoarivelo and coworkers described an original method to transfect motoneurons. 'Surfection', as they have named their method, consists of transfecting these cells by seeding them onto a cationic lipid:DNA-precoated coverslip.[80] The use of a DOTAP:PC-based lipid vector gave a transfection yield of 15%, while Lipofectamine and other well-known auxiliary lipids provided lower surfection rates.[80]

One of the major goals of current neuroscience is to understand the mechanism involved in neuronal neurodegeneration and to develop effective therapies to treat it. In this context, siRNA has emerged as a powerful tool in functional genomic studies and the interest to find a good tool to efficiently deliver siRNA inside different CNS cells has increased. Cationic lipids, such as Lipofectamine 2000, have been described to efficiently introduce siRNA in rat primary hippocampal neurons, resulting in reductions of target mRNA and protein by more than 70% in transfected cells.[81] In the same way, in primary cortical cultures, both Tf-DOTAP:Chol and Lipofectamine 2000 efficiently delivered siRNA inside the cells, leading to a reduction of approximately 50% in luciferase activity, without affecting cellular viability when the final siRNA concentration was 50 nM.[82] More recently, a significant reduction in target mRNA of 70% (which corresponds with knockdown protein levels of 50% in mouse cortical cultures) by using Tf-DOTAP:Chol has been described.[83]

Transfection of Neurons in vivo

Gene therapy for the CNS has enormous therapeutic potential but suffers from limitations, such as blood–brain barrier (BBB) crossing, the difficulties to introduce genes into neuronal cells and the low level of gene expression.[84]

The in vivo performance of a lipidic vector depends on the administration route. Several studies have shown that the intravenous administration of cationic liposome/DNA or cationic liposome/cationic lipid/DNA gives systemic gene expression, particularly in lungs,[85,86] whereas no expression in the brain has been detected.[87] The choice of helper lipids plays an important role in determining the in vivo activity of lipidic vectors; when intravenous administration is used, the in vivo activity of lipidic vectors decreases by 100–1000-fold when DOPE is used as a helper lipid, whilst cholesterol enhances it.[88] Conversely, intratracheal or intratissue administration yields better lipofection when DOPE is used as the helper lipid.[87]

One of the main problems in intravenous administration of cationic lipids is vector aggregation in the presence of serum leading to vector disintegration, DNA release and degradation.[87] Thus, optimization of different routes of administration is needed to reach the expected vector levels at the target tissue. In this regard, DC-Chol liposome/DNA complex injected locally into the intact rat spinal cord gray matter produced an increased expression of target mRNA in the injection areas, whereas no tissue damage was detected.[89]

Similar results have been recently obtained when Tf-lipoplexes (DOTAP:Chol) were used to deliver luciferase siRNA to the brain by stereotactic injection in the right hemisphere striatum. The siRNA-treated group showed a significant decrease in luciferase levels in the ipsilateral striatal tissue (approximately 40%), whereas no luciferase knockdown was detected in the ipsilateral cortex or corpus callosum of these animals.[82] In any case, no relevant signs of toxicity were found in the ipsilateral injected hemisphere when compared with the contralateral hemisphere.

Nevertheless, the low efficiency of expression mediated by cationic lipids compared with viral vectors, together with the need of local injections to obtain significant transfection in the CNS, indicates that much additional work is needed in order to improve their design and to develop them from an experimental concept to a useful tool for the therapy of neurological disorders. One of the possible alternatives are the 'Trojan horse' liposomes that are composed of a lipid core containing polyethylene glycol (PEG), conjugated to specific monoclonal antibodies that allow BBB crossing.[90]


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.