Targeted Delivery of Small Interfering RNA: Approaching Effective Cancer Therapies

Although the use of potent, sequence-specific small interfering RNAs (siRNA) to suppress expression of specific transcripts was originally a useful technique for probing gene function in vitro, their successful application in vivo in animal models against a spectrum of diseases, including cancer (1–5), has spurred interest in developing this approach for siRNA-based therapeutics. However, there are still significant obstacles to be overcome before these molecules can be used in the clinic as anticancer agents, the focus of this minireview. Perhaps foremost among these is the issue of delivery. The in vivo use of siRNAs effectively against cancer hinges on the availability of a delivery vehicle that can be systemically administered to reach both primary and metastatic tumor cells. Moreover, because sufficient intact, functional siRNA must be delivered into the target cell to reach an effective intracellular concentration, and to limit potential side effects due to randomized, general transfection of normal, nontarget tissues, it is also crucial to develop means of directing such a siRNA delivery vehicle specifically to the target cells.

Naked siRNAs, delivered into the bloodstream, even when chemically modified, have extremely short half-lives (seconds to minutes) due to renal clearance (because of their small size; ref. 6). In addition, whereas most RNases are inactive against double-stranded RNA (dsRNA), some serum RNases can degrade siRNA (6). Cellular uptake of naked siRNA is also limited. To overcome some of these challenges, siRNAs have been complexed to a variety of nonviral lipids or protein carriers, including cholesterol, liposomes, antibody protomer fusions, cyclodextrin nanoparticles, fusogenic peptides, aptamers, biodegradable polylactide copolymers, and polymers (4–7). Positively charged cationic liposomes and polymers, such as polyethyleneimine, are currently the two major carriers used to complex with negatively charged siRNA for systemic delivery (3, 8). Although most of the reports in the literature use delivery approaches that are not systemically administered and/or specifically targeted to the tumor, there are a few reports of targeted i.v. delivery of siRNA in animal models of cancer.

The RGD peptide and transferrin (Tf), as well as antibodies and antibody fragments [such as anti-Tf receptor (TfR) and anti–epidermal growth factor receptor], have been used as targeting ligands for i.v. siRNA delivery against tumors (reviewed in ref. 8). Schiffelers et al. (9) linked siRNA against vascular endothelial growth factor (VEGF) receptor 2 to polyethyleneimine that was PEGylated with an RGD peptide ligand at the distal end as a means to target tumor neovasculature. They reported inhibition of both tumor angiogenesis and growth rate in mice bearing murine neuroblastoma N2A tumor xenografts. Similarly, antiangiogenic effects were also observed in ocular neovascularization in the herpes simplex virus disease model (10). Hu-Lieskovan et al. (11), using a Tf-targeted, cyclodextrin-based polycation for delivery of siRNA to tumors in a mouse model of Ewing's sarcoma, saw transient reduction of tumor growth. Recently, Bartlett et al. (12) used a Tf-targeting, ⁶⁴Cu-labeled, cyclodextrin-containing polycation to systemically deliver an anti-luciferase siRNA molecule to Neuro2A-Luc tumor cells. Using simultaneous positron emission tomography/computed tomography to monitor siRNA whole-body biodistribution kinetics and tumor localization to correlate biodistribution data with functional efficacy, they concluded that the primary advantage of the targeting molecule is associated with cellular uptake rather than tumor localization. This same group had previously used Tf-targeted polyplexes carrying an anti-luciferase siRNA to examine the kinetics of siRNA-mediated silencing in mice bearing Neuro2A-Luc tumors (13).

In a proof-of-principal study, Song et al. (14) reported that a protamine-antibody fusion protein using the Fab fragment of HIV-1 envelope antibody (F105-P) as the targeting molecule i.v. delivered FITC-labeled siRNA only to B-16 melanoma tumors modified to express HIV env and not to normal tissues. They also used F105-P to deliver a cocktail of siRNAs against c-myc, MDM-2, and VEGF to mice bearing s.c. B-16 HIV env–expressing xenografts resulting in tumor growth inhibition.

Recently, Pirollo et al. reported the development and use of a tumor-specific, nanosized immunoliposome complex for systemic delivery of siRNA (the scL delivery system; refs. 15–17). This complex [TfR single-chain antibody fragment (TfRscFv)/liposome/siRNA] is composed of an anti-HER-2 siRNA encapsulated by a cationic [1,2-dioleoyl-trimethylammonium-propone (DOTAP) dioleoyl phosphatidylethanolamine (DOPE)] liposome, the surface of which is decorated with a targeting moiety, an anti-TfRscFv (15–17). The TfRscFv contains the complete antibody binding site for the epitope of the TfR recognized by the monoclonal antibody (mAb) 5E9 (18, 19). TfR levels are elevated in various types of cancer cells (20). Elevated TfR levels also correlate with the aggressive or proliferative ability of tumor cells (20). The TfR also recycles during internalization in rapidly dividing cells, such as cancer cells (20), thus contributing to the uptake of TfR-targeted nanocomplexes even in cancer cells where the actual level of the TfR may not be elevated compared with normal cells. Although we have also used Tf as a targeting ligand (21), TfRscFv has certain advantages over the Tf molecule itself or an entire mAb in targeting liposomes to cancer cells with elevated TfR levels. (a) The size of the scFv (∼28 kDa) is much smaller than the Tf molecule (80 kDa) or the parental mAb (155 kDa). The scFv-liposome-DNA complex may thus exhibit better penetration into small capillaries characteristic of solid tumors. (b) The smaller scFv has a practical advantage related to the scaled-up production necessary for the clinical use. (c) Unlike Tf, the scFv is a recombinant molecule and is not isolated from blood.

To increase stability, the siRNA encapsulated in the immunoliposome complex in these studies was a modified hybrid form of siRNA, a type of modification in which the double-stranded molecule is composed of an unmodified antisense RNA strand and a DNA sense strand that may or may not be modified. This approach is different than the use of chemical modifications (reviewed in ref. 8) to improve stability, reduce off-target effects, and maintain efficacy of siRNAs. Three independent publications (15, 22, 23) of RNA interference (RNAi) using these hybrid or modified hybrid duplexes have reported that these antisense RNA/sense DNA hybrid duplexes, which were called “siHybrids” (23), were more potent (15, 22, 23) and led to longer-lasting (23) RNAi, relative to corresponding unmodified siRNA. In contrast to the trend for design of siRNA analogues wherein “more modification is better,” the promising properties of siHybrids indicated that “less may be more” (15). If so, this could translate into more cost-effective RNAi by virtue of using a sense strand in which relatively inexpensive DNA replaces more costly RNA having chemical modifications. In addition, based on molecular “appearance” to Toll-like receptors in the innate immune system (24), a dsRNA/DNA siHybrid might be less immunogenic than a homologous dsRNA/RNA siRNA and may also induce less of an IFN response than dsRNA.

One way to increase the efficacy of the siRNA after the complex has reached the target cell is to enhance siRNA release from the endosome. Thus, more of the siRNA is available in the cytoplasm for knockdown of the target gene rather than being trapped in the endosome and undergoing lysosomal enzymatic degradation. Endosomal compartments are generally acidic in nature. Various methods, including incorporation of pH-sensitive components into liposomes, have been developed to enhance the efficiency of liposomal payload delivery by exploiting this fact (reviewed in ref. 25). The selective destabilization of liposomes following acidification of the surrounding medium with resultant release of the payload has been enhanced by inclusion of specific lipids, many based on phosphatidylethanolamine or modifications thereof (e.g., DOPE). These undergo a lamellar to hexagonal phase transition at low pH, releasing the liposomal contents. Several pH-sensitive synthetic peptides have also been designed in an attempt to produce peptides that can attach to, but not perturb, the surface of a liposome at neutral pH and subsequently fuse adjacent bilayers at acidic pH. Chen et al. (26), as well as others (27), have designed linear and branched histidine-lysine (HK) polymers of varying lengths. The histidine component is believed to buffer and disrupt the endosomes. The inclusion of such HK copolymers significantly increased the transfection efficiency over cationic liposomes alone (24). To further increase the efficacy of the scL complex by facilitating endosomal release while maintaining the small size of the nanoparticle complex, Pirollo et al. (17) also included a small linear pH-sensitive peptide, HoKC {K[K(H)KKK]₅K(H)KKC; adapted from that of Aoki et al. (27)}, in the targeted immunoliposome complex. This HoKC peptide contains a cysteine residue at the end, enabling it to be conjugated to the liposome through a maleimide group.

The results published by Chang and colleagues (15–17) show, using scanning probe microscopy, that this single-chain targeted immunoliposome siRNA complex is a nanoparticle of uniform size, even when the HoKC peptide is included. Their findings also show that modifying the anti-HER-2 siRNA through use of a modified DNA sequence as the sense strand significantly improved the in vitro efficacy of the siRNA compared with standard duplex siRNA and that this approach could significantly sensitize (by over 80-fold) pancreatic cancer cells to the standard chemotherapeutic agent gemcitabine.

The most significant findings of these reports are the tumor-targeting in vivo results. When systemically (i.v. tail vein) administered, both forms of the complex (with and without inclusion of the HoKC peptide) delivered the fluorescently labeled siRNA specifically and efficiently to tumors. This was observed in both large primary prostate tumors (16) and in two metastasis models using human pancreatic cancer and human melanoma MDA435/LCC6 (16, 17). The ability of this approach to efficiently target and transfect metastatic lesions is shown in Fig. 1. The metastasis indicated by the arrow displays a high level of fluorescence with no significant signal in the adjacent normal lung tissue. Moreover, micrometastases near the larger nodule (verified by histology) are also detectable, indicating that even tiny nodules composed of only a few tumor cells can also be reached and transfected via this complex. These results confirm the tumor-targeting ability and the efficient delivery of the siRNA by the nanocomplex containing the pH-sensitive peptide.

More importantly, they were also able to show that this tumor-specific delivery via the HoKC nanocomplex resulted not only in virtually complete knockdown of HER-2 expression in the tumors but also in concomitant changes in expression of pAKT, pMAPK, Bcl-2 (down-modulation), and caspase-3 (up-regulation), all genes involved downstream in the HER-2 signal transduction pathway and apoptotic cell death. Furthermore, the combination of this systemically administered, tumor-specific siRNA nanocomplex and gemcitabine was able to inhibit significantly tumor growth of established PANC-1 xenograft tumors.

Although the field of RNAi has made the transition from basic research to clinical application for localized disease such as macular degeneration in less than 10 years, a time frame perhaps faster than that of any other approach in gene medicine, the lack of efficient targeted delivery, low transfection efficiency, instability to nucleases, poor tissue penetration, and nonspecific immune stimulation have hindered siRNA from reaching its full therapeutic potential. This is particularly true for diseases such as cancer where systemic delivery of targeted therapeutics is essential. In addition to identifying the correct target gene and pathway, effective anticancer siRNA therapies must also be able to deliver a sufficient amount of intact, functional siRNA to the target cell. Progress is being made to address these challenges, with several approaches for cell-specific delivery for cancer, and to increase transfection efficiency being reported. However, the vast majority are still only applicable in vitro or for nonsystemic administration, including intratumoral, i.m., and i.p. injection. The delivery system described by Pirollo et al. (17) seems to be able to overcome the limitations of the current technology. The complex itself is of nanoparticle size and thus able to penetrate through the small capillaries resulting in deeper tumor penetration. The siRNA is a modified hybrid construct that may reduce off-target effects, whereas the encapsulation of this siRNA within the liposome can protect it from degradation and rapid renal clearance while in the bloodstream. This approach has shown exquisite tumor-targeting capabilities to primary and metastatic tumors. The TfR as a target for tumor-specific delivery has been well documented in the literature (20). The inclusion of the anti-TfRscFv also serves to enhance transfection efficiency, as the receptor-bound complex is internalized via receptor-mediated endocytosis. Once internalized, the inclusion of the pH-sensitive, endosomal-disrupting peptide in this nanocomplex enhances release of the payload, increasing the effective cytoplasmic concentration of the siRNA, leading not only to the efficient knockdown of the target as observed in the tumors but, when used in combination with standard chemotherapeutic agents, also to tumor growth inhibition. For maximum efficacy, the use of this targeted siRNA delivery is envisioned not as a single agent but as part of such a combinatorial treatment regimen. Thus, combining all of these factors in one delivery vehicle may be the means to advance the field beyond the current challenges. Successful translation of this approach through clinical trials is the next logical step toward the realization of the potential of siRNA as anticancer therapeutics.

Grant support: SynerGene Therapeutics (K.F. Pirollo), National Foundation for Cancer Research (E.H. Chang), and TriLink Research Award in the form of research grade siRNAs (E.H. Chang).

We apologize to our colleagues whose outstanding publications have not been directly cited due to space constraints.