UCSF

The large molecular weight, poly-anionic, and labile nature of small interfering RNA (siRNA) necessitate that these macromolecules are associated with carriers for cellular delivery. Lipid-based particles (LNPs) are very efficient at both cellular internalization and contents delivery; however, the cell has evolved mechanisms to efficiently degrade exogenous macromolecules and prevent the alteration of cellular function by foreign nucleic acids. An important aspect to understand is how cells differentially traffic LNPs so that more efficacious formulations can be designed for therapeutic purposes.

The majority of this dissertations presents a series of investigations into how cells endocytosis and sort LNPs encapsulating siRNA. First, I developed a set of fluorescently tagged endosomal marker proteins for clathrin-, caveolin-, and flotillin-dependent internalization pathways. In addition, I prepared markers for early and late endosomes (Rab5a and Rab7a), recycling endosomes (Rab11a), multivesicular bodies (CD82), and lysosomes (LAMP1). Second, I utilized the technique of ratiometric fluorescence microscopy to measure endosomal pH. Third, I developed an image-processing program built into ImageJ/Fiji to accurately colocalize the lipid, siRNA, and endosomal marker or pH probe signals. I then used these techniques to determine the pathway cells employ to internalize three different LNP formulations comprised of cholesteryl hemisuccinate (CHEMS), 1,2-dioleoyl-3-trimethylammonium-propane (DOTAP), and 1,2-dilinoleyloxy-3-dimethylaminopropane (DLinDMA).

I present evidence that DOTAP LNPs are more potent at delivering siRNA compared to CHEMS or DLinDMA LNPs. I also demonstrate that cells endocytose the LNPs in measurably different ways, specifically, DOTAP exhibits at 0.62 pH unit increase in endosomal pH, pH dependent delivery, 30% colocalization with Rab11a and CD82 during the first 30 minutes of transfection, and a slow accumulation into LAMP1 positive lysosomes over a 4 hour transfection ending at 25% colocalization. In contrast, CHEMS is the least efficient LNP at delivering siRNA and colocalizes with LAMP1 at 25% during a 4 hour transfection. DLinDMA shows a 0.4 pH unit increase in endosomal pH and low colocalization (10 - 15%) with the Rab11a, CD82, and LAMP1 markers tested, despite being present at similar levels as the other formulations, and is presumably trafficked into compartments I did not label. Correlations between cellular trafficking and siRNA delivery efficiency are discussed that may aid in the design of potent LNP formulations.

Finally, I also investigated the human dsRNA transporter SIDT-1, and present evidence that high expression levels are required for cellular siRNA uptake, and the that extracellular domain does not bind directly to dsRNA. In the appendix, I present efforts to use site directed mutagenesis to convert human cytidine deaminase into an enzyme with cytosine deaminase activity.