The Utility of Tumor-specifically Internalizing Peptides for Targeted siRNA Delivery into Human Solid Tumors

Results and Discussion

Development of the targeted-delivery vector HN-1^TYR. HN-1, a synthetic 12-mer TSLP, was isolated by screening a random peptide-displaying M13 bacteriophage-based library (19). It meets several criteria for targeted drug delivery into solid tumors. Firstly, HN-1 can translocate drugs across the cell membrane into the cytosol. This was originally inferred from the HN-1 internalization study performed using a fluorescent molecule as the cargo (19). More recently, HN-1 was shown to translocate a human protein kinase C-ε (PKC-ε) inhibiting peptide or diphtheria toxin used as therapeutic into human cancer cells (21, 22). Secondly, the uptake of HN-1 occurs in a tumor-specific manner. HN-1 is internalized by human HNSCC cells (i.e. MDA177Tu, MDA138Tu, MDA59Tu, MDA167Tu, MDA686Tu, MDA1986Tu, UMSCC1, UMSCC36) yet poorly by their normal counterparts (i.e. HOK16B, NOE oral epithelial cells) (19, 21). Similarly, HN-1 was internalized by human breast cancer cells (ex. MDA-MB231, SKBR3) but less efficiently by their normal counterpart (ex. MCF10A non-tumorigenic mammary epithelial cells) (21). Thirdly, HN-1 is capable of penetrating solid tumor, which is critical as >90% of human cancers represent solid tumors. This was indicated by the presence of intravenously injected HN-1 at the interior of solid tumors formed from MDA177Tu or MDA167Tu cells in a mouse xenograft model (19). An independent report confirmed that the intravenously injected HN1 localizes at the interior of solid tumors formed from UMSCC1 cells in vivo (21).

HN-1^TYR, which consists of the parental HN-1 backbone with two extra tyrosine residues, was synthesized. Tyrosines were placed N-terminally as the presence of additional residues at this terminus did not interfere with the HN-1 internalization potential (19). HN-1^TYR efficiently solublized in phosphate-buffered saline (PBS) solution. The addition of two tyrosines did not generate novel proteolytic sites. Next, human head and neck cancer KB cells were incubated with FITC-HN-1^TYR at 37°C, for 48 h, as previously described (19). FITC-HN-1^TYR, which represents fluorescein-conjugated HN-1^TYR, was synthesized chemically (see Materials and Methods). It led to fluorescing KB cells. Fluorescence was detected throughout the cell and was not limited to the cell surface (Figure 1; panel 2; also see Figure 2). The enlarged fluorescent micrograph of a single cell (Figure 3A, inset) shows that the internalized peptide primarily resides at the cytoplasm consistent with the properties of the parental peptide HN-1 (19). In the original report describing the HN-1 isolation (19), its internalization was extensively documented using various experimental methods so it is not repeated here. Untreated KB cells exhibited little autofluorescence (Figure 1; panel 1), suggesting that the observed fluorescence (Figure 1; panel 2) was specifically due to the internalized FITC-HN-1^TYR. In contrast, normal human dermal fibroblasts (NHDF) that were similarly incubated with FITC-HN-1^TYR exhibited little fluorescence (Figure 1, panel 4). Thus, HN-1^TYR retains the HNSCC-internalization potential.

HN-1^TYR is internalized by human breast cancer cells. To identify additional cancer targets of HN-1 with the paucity of information regarding the cognate receptor, we relied on the common developmental lineage. Previously, a similar approach (i.e. ‘developmental histogenesis’) was used to identify cancer types with similar chemosensitivity (25). During embryonic development, surface ectodermal epithelium gives rise to three distinct structures—i.e. cranial structures, epidermal structures, ectodermal placodes (26). Cranial structures include multiple cell types within the oral cavity such as the secretory, duct-lining cells of the salivary, palatine of oral glands and the epithelial lining cells of the gums, palate, lips and paranasal sinuses. Epidermal structures include the duct-lining and secretory cells of mammary glands (27). In humans, breast cancer occurs most frequently in the duct (i.e. ductal carcinoma) or lobules (i.e. lobular carcinoma). This fact, along with the recent finding by Bao et al. that HN-1 is internalized by human breast cancer cells (21), prompted us to investigate whether HN-1^TYR could be used to deliver drugs to breast cancer cells.

To examine this point, human breast cancer cells with varying ‘triple status’ were investigated. Triple status refers to the expression level of Her2 receptor, estrogen receptor (ER), and progesterone receptor (PR). Here, we used human breast cancer cells that are distinct from those used in the report by Bao et al. (21). ZR-75-1 cells, which overexpress Her2, are positive for the ER (16 fmol/mg protein) or PR (102 fmol/mg) expression (28, 29). MCF-7 cells, which express little Her2, are positive for the ER (43 fmol/mg protein) or PR (115 fmol/mg protein) expression (30-32). MDA-MB-468 cells, which express little Her2, contain little ER (4 fmol/mg protein) or PR (9 fmol/mg protein (28, 33). Incubation of ZR-75-1 (Figure 2, panel 4), MCF-7 (Figure 2, panel 6) or MDA-MB-468 (Figure 2, panel 2) with FITC-HN-1^TYR at 37°C for 48 h led to fluorescing cells in all three cases. In contrast, untreated ZR-75-1 (Figure 2, panel 3), MCF-7 (Figure 2, panel 5) or MDA-MB-468 (Figure 2, panel 1) cells did not fluoresce, indicative of the lack of autofluorescence. Hence, like HN-1, HN-1^TYR targets human breast cancer cells and the targeting may be independent of their triple status.

Development of HN-1^TYR-anti-hRRM2 siRNA^R for anticancer therapy. To avoid side-effects resulting from the HU chemotherapy, an anti-hRRM2 siRNA that suppresses RR by degrading the mRNA encoding its hRRM2 subunit was previously developed (7). To identify target site, the hRRM2 mRNA sequence was probed for the preferred 27-nucleotide sites using an algorithm designed for predicting end-energy differential and mRNA secondary structure (34). The resulting candidate were distinguished on the basis of their GC content and homology with non-hRRM2 genes using basic local alignment search tool (BLAST) search, which yielded 3 siRNA candidates. Subsequently, numerous duplexes in their immediate vicinities were analyzed, resulting in the identification of a single siRNA with a potent hRRM2 inhibiting property (7). Upon transfection (using lipofectamine), the anti-hRRM2 siRNA suppressed (40-60%) the hRRM2 protein level for as long as 5 days. Transfection of human colorectal adenocarcinoma HT-29 cells with the anti-hRRM2 siRNA led to a near complete cessation of their growth (7). The siRNA does not appear to elicit interferon response in vivo (8, 35).

To translocate across the cell membrane for cellular uptake, the above siRNA was conjugated to HN-1^TYR. To avoid degradation by RNases in vivo, anti-hRRM2 siRNA was synthesized with fluorine, incorporated at its 2’-OH position to yield anti-hRRM2 siRNA^R. FITC-HN-1^TYR was conjugated to 5’-end of anti-hRRM2 siRNA^R (antisense strand only) using hexynyl phosphoramidite linker to generate FITC-HN-1^TYR-anti-hRRM2 siRNA^R. Equimolar concentration of the above sense and antisense strands were hybridized to generate duplex siRNA (molecular mass of the conjugate: ~19.7 kd). The internalization potential of the conjugate was examined. MCF-7 breast cancer cells were incubated with FITC-HN-1^TYR-anti-hRRM2 siRNA^R at 37°C for 48 h, as described above. It led to fluorescing cells (Figure 3A). The inset shown in Figure 3A, representing the amplified view of a single treated cell, shows clearly that the conjugate localizes at the cytoplasm upon entry. Untreated MCF-7 cells did not fluoresce (panel 5 in Figure 2). Thus, HN-1^TYR is capable of translocating anti-hRRM2 siRNA^R into MCF-7 cells in vivo. Next, its effect on the hRRM2 expression was examined. MCF-7 cells were incubated with FITC-HN-1^TYR-anti-hRRM2 siRNA^R for the indicated time period, lysed, and the resulting cell lysate was analyzed by western blot analysis using an anti-human hRRM2 antibody as the primary antibody. It led to a partial reduction (ca. ~50%) in the level of the endogenously expressed hRRM2 (Figure 3B, lanes 2 or 3). The suppression was specific for hRRM2 as the expression level of an irrelevant protein (designated as “NSP” in Figure 3B, lanes 2 or 3) was not affected. Thus, the FITC-HN-1^TYR-anti-hRRM2 siRNA^R conjugate retains the hRRM2-suppressing property.

A key determinant for the successful siRNA therapy is delivery (36). To avoid side-effects, delivery vectors with tumor-homing potential are increasingly sought. For targeted-drug delivery into human solid tumors, multiple requirements must be met (see above). Minute, non-bioactive, tumor-specifically internalizing peptides such as HN-1 may provide a solution. In this report, we described the development of a conjugate composed of the HN-1^TYR peptide and anti-hRRM2 siRNA^R. Our results confirm earlier findings by others that peptides can mediate the transfer of siRNAs into human cells (18). As the uptake seems to occur independently of the ‘triple status’, it may be applicable for a wide spectrum of human breast cancers. The results also suggest that RR remains a tenable target for therapy. Novel chemotherapeutics inhibiting RR are currently being developed. Potentially, siRNAs may represent the next-generation drugs to target RR. Its specificity may be exploited to lessen side-effects while preserving their therapeutic potential. Further works are envisioned. Firstly, despite the recent report (37), the precise identity of the HN-1 receptor remains to be defined. Secondly, a HN-1 derived vector with the endosome-escaping potential may be developed (22). Thirdly, the tumor-suppressing potential of the conjugate will be examined using an animal model. Hopefully, these advances may lead to a safer, yet effective, form of anticancer therapy.