Novel siRNA delivery system to target podocytes in vivo

Abstract

Podocytes are injured in several glomerular diseases. To alter gene expression specifically in podocytes in vivo, we took advantage of their active endocytotic machinery and developed a method for the targeted delivery of small interfering ribonucleic acids (siRNA). We generated an anti-mouse podocyte antibody that binds to rat and mouse podocytes in vivo. The polyclonal IgG antibody was cleaved into monovalent fragments, while preserving the antigen recognition sites. One Neutravidin molecule was linked to each monovalent IgG via the available sulfohydryl group. Protamine, a polycationic nuclear protein and universal adaptor for anionic siRNA, was linked to the neutravidin via biotin. The delivery system was named shamporter (sheep anti mouse podocyte transporter). Injection of shamporter coupled with either nephrin siRNA or TRPC6 siRNA via tail vein into normal rats substantially reduced the protein levels of nephrin or TRPC6 respectively, measured by western blot analysis and immunostaining. The effect was target specific because other podocyte-specific genes remained unchanged. Shamporter + nephrin siRNA induced transient proteinuria in rats. Control rats injected with shamporter coupled to control-siRNA showed no changes. These results show for the first time that siRNA can be delivered efficiently and specifically to podocytes in vivo using an antibody-delivery system.

Figure 1. Design of Shamporter (sheep anti mouse podocyte & transporter).

Shamporter is a modified anti podocyte antibody that piggybacks siRNA to a target cell. (A) Specific cleaving of the anti mouse podocyte divalent IgG at the inter-heavy chain disulfide bond, using 2-Mercaptoethylamine, results in monovalent IgG. (B) A Neutravidin binding site is conjugated to the available sulfohydryl group. (C) Protamine is biotinylated which then binds to the monovalent IgG. (D and E) Negatively charged siRNA molecules bind to the positively charged protamine domain of the shamporter construct. (F) The modified antibody (shamporter) carrying siRNA binds to the podocytes. The bound antibody is internalized and transported to the cytoplasm. Uncoating of the endosomatic vesicle releases the transported siRNA into the cytoplasm. The released siRNA activates RISC (RNA induced silencing complex), which is followed by degradation of the target mRNA.

Figure 2. Affinity of the anti-podocyte antibody.

The antibody binds to the apical, basal and the slit diaphragm regions of the podocyte (A). A small number of immunogold particles were present in the glomerular basement membrane (GBM) and endothelial cells (GEN). The arrows indicate examples of immunogold particles. Sheep anti mouse podocyte antibody staining was detected only in the glomerulus (B + D arrows) and was absent in the tubular compartment (D asterix). Control animals injected with normal sheep IgG showed no staining (C + E).

Figure 3. Podocyte antibody specificity and uptakes by podocytes.

Western blot analysis with the podocyte antibody on membrane fractions from cultured immortalized mouse podocytes (P), mouse mesangial cells (MC), mouse fibroblasts (FB), rat proximal tubular epithelial cells (TC), and rat glomerular endothelial cells (GE), showed a band only in podocytes (A lanes 1 & 5). After 30 minutes at 37°C the cytoplasmic fraction of podocytes incubated with shamporter, showed a marked increase in detectable sheep IgG (55 kDa). Incubation of cells on 4°C instead significantly slowed shamporter uptake (B).

Figure 4. Western blot analyses for p57, CDK5 and TRPC6.

In cultured immortalized mouse podocytes transfected with shamporter + siRNA directed against p57 (A) or CDK5 (B), there was a progressive decrease in protein levels at 48 h and 72 hours. In cultured podocytes transfected with shamporter + siRNA directed against TRPC6, there was a dose dependent decrease in protein levels for TRPC6 (C). Cells transfected with nonfunctional control-siRNA served as control. The panels on the right show densitometry performed against loading controls GAPDH or β-Actin.

Figure 5. Western blot analysis for Nephrin and TRPC6 in rats injected with shamporter and siRNA.

Rats were injected via tail vein with shamporter + control siRNA or shamporter + nephrin siRNA or shamporter + TRPC6 siRNA. Compared with control siRNA injected rats, there was a significant decrease in nephrin protein levels in the rats injected with nephrin siRNA (A). Likewise, there was a significant decrease in TRPC6 protein levels in the rats injected with TRPC6 siRNA (B). Staining for GAPDH or beta actin was used as a protein loading control, and the podocyte specific protein podocin or nephrin were used to ensure that the same amount of podocyte protein was loaded. Densitometry and of the results are shown in the graph. Ratios were compared using t-test (p<0.05); ratios are expressed as median ± min and max values.

Figure 6. Immunohistochemistry for nephrin, podocin and TRPC6.

Animals injected with shamporter + control siRNA show a strong linear staining signal for nephrin and podocin, as well as glomerular staining for TRPC6. In animals injected with shamporter + nephrin or TRPC6 specific siRNA, the staining signal was significantly reduced, while the signal for podocin remained unchanged. When nephrin, podocin, or TRPC6 antibodies were omitted, no staining was detected.

Figure 7. Proteinuria and Podocyte abnormalities induced in nephrin siRNA injected rats.

Urinary protein (UP) to urinary creatinine (UC) ratio (mg/mg) was used to assess renal function. UP∶UC ratio was increased significantly in rats injected with shamporter + nephrin siRNA after 24 and 72 h compared to baseline (ANOVA; p<0.001). 72 h after injection, shamporter + nephrin siRNA injected rats exhibited a significantly higher UP∶UC ratio then animals from the control group (t-test; p<0.005). No significant differences in UP∶UC ratio was found between control animals at 72 h and baseline (A) No difference was found in BUN in either group (data not shown). There were no abnormalities in podocytes in control rats injected with shamporter + control siRNA (C). Arrows indicate normal slit diaphragms. In rats injected with shamporter + nephrin siRNA, there was loss of normal slit diaphragm architecture which included crowding of the slit diaphragms and reduced space between adjacent foot processes (B arrows). Quantification of foot process width showed a significant increase in rats injected with shamporter + nephrin siRNA compared with rats injected with shamporter + control siRNA (D).