RNAi Screening of the Glucose-Regulated Chaperones in Cancer with Self-Assembled siRNA Nanostructures

Abstract

The emerging field of RNA nanotechnology has been used to design well-programmed, self-assembled nanostructures for applications in chemistry, biology, and medicine. At the forefront of its utility in cancer is the unrestricted ability to self-assemble multiple siRNAs within a single nanostructure formulation for the RNAi screening of a wide range of oncogenes while potentiating the gene therapy of malignant tumors. In our RNAi nanotechnology approach, V- and Y-shape RNA templates were designed and constructed for the self-assembly of discrete, higher-ordered siRNA nanostructures targeting the oncogenic glucose regulated chaperones. The GRP78-targeting siRNAs self-assembled into genetically encoded spheres, triangles, squares, pentagons and hexagons of discrete sizes and shapes according to TEM imaging. Furthermore, gel electrophoresis, thermal denaturation, and CD spectroscopy validated the prerequisite siRNA hybrids for their RNAi application. In a 24 sample siRNA screen conducted within the AN3CA endometrial cancer cells known to overexpress oncogenic GRP78 activity, the self-assembled siRNAs targeting multiple sites of GRP78 expression demonstrated more potent and long-lasting anticancer activity relative to their linear controls. Extending the scope of our RNAi screening approach, the self-assembled siRNA hybrids (5 nM) targeting of GRP-75, 78, and 94 resulted in significant (50-95%) knockdown of the glucose regulated chaperones, which led to synergistic effects in tumor cell cycle arrest (50-80%) and death (50-60%) within endometrial (AN3CA), cervical (HeLa), and breast (MDA-MB-231) cancer cell lines. Interestingly, a nontumorigenic lung (MRC5) cell line displaying normal glucose regulated chaperone levels was found to tolerate siRNA treatment and demonstrated less toxicity (5-20%) relative to the cancer cells that were found to be addicted to glucose regulated chaperones. These remarkable self-assembled siRNA nanostructures may thus encompass a new class of potent siRNAs that may be useful in screening important oncogene targets while improving siRNA therapeutic efficacy and specificity in cancer.

Keywords: GRP; RNAi nanotechnology; cancer gene therapy; cervical and breast cancer; chaperones; endometrial; glucose regulated proteins; siRNA nanostructures.

Figure 1

Design and self-assembly of siRNA nanostructures. The RNA templates (namely, linear and V-, and Y-shaped RNA) were designed and synthesized according to our previously described methodology. The V- and Y-shaped templates incorporate a branchpoint ribouridine (rU), which facilitates the hybridization of complementary sense (S) and antisense (A) RNA. These templates preorganizes the self-assembly of siRNA hybrid nanostructures having discrete sizes and shapes, including those belonging to circles, triangles, squares, rectangles, pentagons, hexagons, and poroustype structures. These siRNA nanostructures are genetically encoded to target a single (1), double (1, 2) and triple (1, 2, 3) sites of oncogenic GRP-75, 78, and 94 mRNA.

Figure 2

siRNA self-assembly. Native, nondenaturing 16% PAGE of self-assembled Y- (lanes 2 and 3), V-shape (lanes 4–7) siRNA hybrids, Y-shape RNA templates hybridized to linear complementary RNA sequences (lanes 8 and 9), V-shape RNA templates hybridized to linear complementary RNA sequences (lanes 10 and 11), and linear siRNA (lane 12) along with Y-shape (lane 13), V-shape (lane 14), and linear (lane 15) RNA templates. The RNA ladder (23–500 bp) was used to track the relative sizes of the siRNA hybrids on the gel (lane 1).

Figure 3

Sizes and shapes of siRNA nanostructures. TEM images and particle size distribution plots of (A) V-shape siRNA hybrids A11–S11, (B) V-shape siRNA hybrids A12–S12, (C) V-shape siRNA hybrids A11–A2S1–S12, (D) V-shape siRNA hybrids A11–A2S1–S12–A12, (E) Y-shaped siRNA hybrids A111–S111, and (F) Y-shaped siRNA hybrids A123–S123.

Figure 4

Biological evaluation of the siRNA leads. (A) The AN3CA cell growth curves (0–66 h) obtained from the Incucyte following RNAiMAX (7 µL) transfections of the siRNAs (5 nM). (B) LDH release assay following siRNA transfections. The percent LDH released was measured for the siRNAs and normalized according to the NS RNA. (C) Western blot of the total GRP78 and Cl-PARP levels following siRNA transfections. The loading control, GAPDH, was used to normalize the detected bands for quantitative densitometry using NIH imager (ImageJ). (D) The percent GRP78 knockdown and the percent Cl-PARP levels were normalized according to the NS RNA control and quantitated following densitometry of the Western blot. The linear (r² = 0.9243) correlation diagram in between the percent GRP78 knockdown and the percent Cl-PARP levels is provided as an inset. All experiments were replicated in triplicate, with average values presented with their standard deviations about the mean. Statistical analyses produced error bars with acceptable variance ± SEM; N = 3; p < 0.05.

Figure 5

RNAi screening. (A) Grp mRNA levels detected by RT-PCR. HeLa cells were transfected with NS RNA, linear GRP (78 + 94 + 75), and Y-shape GRP(789475) siRNA (5 nM) using RNAiMAX (7 µL). Grp78, Grp94, Grp75, and Gapdh mRNA levels were normalized according to Gapdh and quantitated with respect to NS RNA. (B) LDH release assay. The percent LDH released was measured following transfections of all treated cell lines (MRC5, AN3CA, MDA-MB-231, and HeLa). The LDH levels were quantitated and normalized according to the NS RNA. (C) Western blots measuring GRP78, GRP94, and GRP75 (percent protein) levels following siRNA (5 nM) transfections in normal lung, MRC5, endometrial, AN3CA, breast, MDA-MB-231, and cervical HeLa cancer cells. The GRP78, GRP94, and GRP75 levels were normalized according to GAPDH and quantified with respect to the NS RNA. Data represents knockdown efficiency of V-shape siRNAs (GRP7894, GRP7875, and GRP9475), Y-shape siRNA (GRP789475), and the linear siRNAs (GRP78 + GRP94 + GRP75) added in combination. All experiments were replicated in triplicate, with average values presented with their standard deviations about the mean. Statistical analyses produced error bars with acceptable variance ± SEM; N= 3; p < 0.05.