Enhancing Endosomal Escape and Gene Regulation Activity for Spherical Nucleic Acids

Abstract

The therapeutic potential of small interfering RNAs (siRNAs) is limited by their poor stability and low cellular uptake. When formulated as spherical nucleic acids (SNAs), siRNAs are resistant to nuclease degradation and enter cells without transfection agents with enhanced activity compared to their linear counterparts; however, the gene silencing activity of SNAs are limited by endosomal entrapment, a problem that impacts many siRNA-based nanoparticle constructs. To increase cytosolic delivery, SNAs were formulated using calcium chloride (CaCl₂) instead of the conventionally used sodium chloride (NaCl). The divalent Ca²⁺ ions remain associated with the multivalent SNA and have a higher affinity for SNAs compared to their linear counterparts. Importantly, confocal microscopy studies show 22% decrease in accumulation of CaCl₂-salted SNAs within the late endosomes compared to NaCl-salted SNAs, indicating increased cytosolic delivery. Consistent with this finding, CaCl₂-salted SNAs comprised of siRNA and antisense DNA all exhibit enhanced gene silencing activity (up to 20-fold), compared to NaCl-salted SNAs regardless of sequence or cell line (U87-MG and SK-OV-3) studied. Moreover, CaCl₂-salted SNA based forced intercalation probes show improved cytosolic mRNA detection.

Graphical Abstract

Calcium chloride (CaCl₂)-salting instead of conventional sodium chloride (NaCl)-salting during spherical nucleic acid (SNA) synthesis can increase cellular uptake by 36 ×, cytosolic delivery by 22%, and result in up to a 20 × enhancement in gene regulation activity (in the case of siRNA constructs), with no apparent toxicity.

1. Introduction

Small interfering RNA (siRNA) is an important tool that is able to regulate expression of genes in eukaryotic cells via RNA interference.^[1] siRNAs can be designed to specifically degrade complementary target mRNA upon binding to the RNA-induced silencing complex (RISC).^{[2,3] ]} This modularity greatly expands the number of accessible biological targets, including some that are regarded as “undruggable”. While siRNAs exhibit potent gene silencing activity, they do not have the ability to readily enter cells on their own in part due to their negatively charged phosphate backbone, and they are susceptible to rapid degradation by nucleases, making it challenging to broadly use them for biological applications.^[4,5]

Spherical nucleic acids (SNAs) are comprised of a nanoparticle core densely functionalized with a radial array of oligonucleotides.^[6,7] The SNA architecture has unique structure-dependent properties that can be exploited to overcome the limitations associated with the delivery of linear siRNA or antisense DNA. The three-dimensional oligonucleotide shell bestows SNAs with the ability to enter cells via scavenger receptor A-mediated endocytosis,^[8,9] and the dense shell that defines the SNA minimizes nuclease degradation.^[10] When siRNA-based SNAs enter the cytosol, they associate with the RISC complex to mediate gene silencing; thus, SNAs are promising entities for gene regulation therapeutics and have been used in hundreds of cell based knockdown experiments (e.g. EGFP^[11,12], GAPDH^[13], LUC2^[14-17], EGFR^[18-20], Bcl2L12^[21,22], BcL2^[23,24], GM3S^[25], HER2^[26-28], MGMT^[29], Survivin^[20], MFG-E8^[30], PLK1^[12], IL17RA^[31], TGFβ^[32], STAT3^[33], TIMP1^[34], TNF-a^[34], PDL1^[35] and VEGF^[36]).

To realize potent gene regulatory activity, SNAs must be delivered to the cytosol to access the RISC complex and target mRNA, as is the case with all gene silencing therapies.^[37,38] The uptake pathway for SNAs involves trafficking through the endosomal pathway with accumulation in the late endosome, while only a small portion of the SNAs escape to the cytosol where they can engage in gene silencing.^[39] The extent of escape and corresponding activity are highly dependent upon sequence, cell type, and perhaps even cell cycle, making the identification of a lead structure for a given target a difficult-to-predict process. While methods to enhance the cytosolic delivery of SNAs have been explored, these methods have primarily relied on the use of external stimuli. Specifically, treating cells with SNAs comprised of a photosensitizer core followed by irradiation with white^[40] or near infrared light^[41] was found to enhance the cytosolic delivery of SNAs by rupturing lysosomes via singlet oxygen (¹O₂) generation.^[42] However, excessive ¹O₂ exposure is known to result in DNA damage and cell death.^[43,44] To improve the potency and increase the generality of siRNA-SNA based gene regulation constructs, a method that improves the cytosolic delivery of SNAs, without the use of external stimuli, needs to be established and would be highly beneficial for biological and medical applications involving SNAs.

The “proton sponge effect” is a phenomenon that has been exploited in strategies used to promote cytosolic delivery; accumulation of a polyplex inside the endosome induces an influx of protons (H⁺) followed by endosomal membrane rupture due to osmotic swelling and concomitant release of the cargo into the cytosol.^[38,45-47] Therefore, cationic polymers have been utilized to deliver oligonucleotides in vitro, but these polymers are often toxic, preventing their widespread animal and clinical use.^[48-50] A less toxic method that relies on calcium phosphate (Ca₃(PO₄)₂ ) particles has also been used. This method generates a calcium-phosphate-oligonucleotide co-precipitate that facilitates endocytosis of the oligonucleotides into the cells; entry of Ca²⁺ ions into the endosome triggers an influx of protons, resulting in endosomal escape.^[51-56] While Ca₃(PO₄)₂ particles yield high transfection efficiencies (~50%) with long circular oligonucleotides,^[57] particles formed with short linear DNA exhibit low transfection efficiency (1~5% on average), requiring addition of glycerol or DMSO to cells to increase transfection efficiency.^[58-61] Moreover, since it is difficult to control Ca₃(PO₄)₂ particle size and aggregation, transfection efficiency and reproducibility are critically dependent on multiple parameters, including calcium/phosphate molar ratio, temperature, precipitation time, the time to transfect the particles after precipitation, pH and mixing method.^[62-65]

To address the limitations of using Ca₃(PO₄)₂ particles for oligonucleotide delivery, an alternative approach has been developed.^[66,67] This approach relies on Ca²⁺-siRNA complexes formed by mixing aqueous CaCl₂ with linear siRNA duplexes to facilitate uptake of Ca²⁺ into endosomes via complexation to siRNA. Unlike the Ca₃(PO₄)₂ particles, the complexes do not aggregate or precipitate and result in more efficient cell uptake and cytosol delivery. However, the study with Ca²⁺-siRNA was only conducted with mouse embryonic fibroblasts and primary macrophages.

Since SNAs are multivalent and are already capable of readily entering over 60 cell lines, including human derived cell lines,^[7] we hypothesized that they would have a higher affinity for Ca²⁺ and that Ca²⁺-complexed SNAs would have the ability to co-deliver Ca²⁺ and gene-regulatory oligonucleotides into cells through endocytosis. Such properties could potentially lead to improved cytosolic delivery and corresponding gene regulation activity. Herein, we investigated the cellular uptake and gene regulation activities of poly-(lactic-co-glycolic) acid (PLGA) SNAs salted with CaCl₂ to take advantage of PLGA’s biocompatibility and biodegradability.^[68-70] In a U87-MG human glioblastoma cell line, these constructs exhibit enhanced cellular uptake at early time points (36-fold), reduced accumulation (22 %) within the late endosomes compared to NaCl-salted PLGA SNAs, and markedly improved gene silencing capabilities (8 to 18-fold enhancement). Moreover, no apparent toxicity was observed across three different cell lines involving three different targets. Collectively, these data establish the CaCl₂ salting of SNAs as a method for improving cytosolic delivery of SNAs for multiple applications, regardless of nucleic acid sequence. Moreover, we show how this method can be extended to molecular probes such as forced intercalation (FIT) flares^[71] to improve cytosolic mRNA detection.

2. Results and Discussion

Synthesis and characterization of CaCl₂-salted PLGA SNAs

To prepare CaCl₂-salted PLGA SNAs, spherical PLGA particles (36.0 ± 4.3 nm) that incorporate PLGA-b-poly(ethylene glycol)-azides were synthesized using a nanoprecipitation method and subsequently functionalized with dibenzocyclooctyne (DBCO)-modified siRNA duplexes via copper-free click chemistry.^[70] To generate the CaCl₂-salted PLGA SNAs (henceforth termed CaCl₂ PLGA SNAs), instead of the conventionally used 500 mm NaCl,^[8,39,70] CaCl₂ solution was used to raise the total CaCl₂ concentration to 230 mm (Figure 1A).

Figure 1. Synthesis scheme and characterization of CaCl₂-salted PLGA SNAs.

(A) PLGA particles were formed via nanoprecipitation with incorporation of PLGA-PEG5000-N₃. DBCO-modified siRNA duplexes were immobilized on the PLGA particles by copper-free click chemistry and concomitant salting with either NaCl or CaCl₂ to form NaCl-salted PLGA SNAs (PLGA SNAs) or CaCl₂-salted PLGA SNAs (CaCl₂ PLGA SNAs, salted at 230 mm CaCl₂), respectively. (B) DLS of the PLGA particles, PLGA SNAs and CaCl₂ PLGA SNAs. The diameter shown is the number mean. (C) Zeta potential of the PLGA particles, PLGA SNAs, and CaCl₂ PLGA SNAs (D) Fluorescence measurement (λ_exication= 480 nm, λ_emission= 520 nm) from PicoGreen exclusion assay of PLGA SNAs, CaCl₂ PLGA SNAs and CaCl₂-siRNAs treated with PicoGreen solution and with 40 mm EDTA in PicoGreen solution. (E) Percentage of Ca²⁺ ion associated within CaCl₂ PLGA SNAs and CaCl₂-siRNA complexes at multiple time points after dialysis. The percentage of Ca²⁺ remaining was determined by Picogreen exclusion assay. Error bars are the standard deviations (SDs) of three independent measurements. ns: not significant, * p < 0.05, ** p < 0.01, **** p < 0.0001. Only key significances are shown for clarity.

To characterize these SNAs, we first used dynamic light scattering (DLS) to analyze the size of the PLGA nanoparticles before and after functionalization with siRNAs. DLS measurements showed that the hydrodynamic diameter of the standard NaCl-salted PLGA SNAs (henceforth termed PLGA SNAs) (51.2 ± 1.5 nm, PDI: 0.045 ± 0.010) was on average 16 nm larger than that of the bare PLGA particles, which is expected based on the calculated length of the 23 base pair siRNA duplexes and assuming one RNA base pair length is 0.34nm (23 × 0.34 × 2 = 15.64 nm, Figure 1B).^[72] The hydrodynamic diameters and polydispersity indices (PDIs) of the PLGA SNAs and CaCl₂ PLGA SNAs were not significantly different (CaCl₂ PLGA SNAs: 47.0 ± 0.4 nm, PDI: 0.078 ± 0.009; Figure S1, Table S2). Furthermore, the siRNA duplex loading and surface density of the PLGA SNAs and the CaCl₂ PLGA SNAs were not significantly different; with average loading of 639.1 ± 35.3 and 632.7 ± 13.4 duplexes per particle, with a surface density of 13.5 ± 0.7 and 13.6 ± 0.3 pmol/cm² for PLGA SNAs and the CaCl₂ PLGA SNAs, respectively (Figure S2). The zeta potential of the azide-modified PLGA particles was −42.1 ± 0.6 mV due to the negatively charged terminal carboxylic groups of PLGA chains on the nanoparticle surface as reported previously.^[73,74] The zeta potential of the PLGA SNAs was −40.5 ± 2.1 mV; with no significant change in zeta potential compared to PLGA particles (Figure 1C). CaCl₂ PLGA SNAs exhibited a significantly less negative zeta potential (−11.8 ± 1.8 mV), similar to the trend observed for the previously reported CaCl₂-siRNAs (−11.9 ± 0.9 mV);^[66] which is a consequence of Ca²⁺ association with RNA shell (Figure 1C, Figure S3, Table S2).

To further verify the association of Ca²⁺ ions with the SNAs, we conducted a modified ethidium bromide (EtBr) exclusion assay,^[67,75] by employing PicoGreen^™ in place of EtBr due to its lower limit of detection.^[76,77] The fluorescence intensity of the PicoGreen^™ increases when it intercalates into double-stranded nucleic acids.^[78] Since divalent Ca²⁺ ions electrostatically bind to the phosphate backbone of nucleic acids,^[79,80] we hypothesized that the adsorption of Ca²⁺ ions to the siRNA duplexes would compete with and prevent the PicoGreen^™ from intercalating within the oligonucleotide shell. Indeed, upon treatment with PicoGreen^™, the fluorescence intensity of the CaCl₂ PLGA SNAs was only 25.2 ± 3.1 % of that of the PLGA SNAs, whereas the intensity of the CaCl₂-siRNA (prepared with 100 nm siRNA and 230 mm CaCl₂ equivalent to CaCl₂ PLGA SNAs) was 40.4 ± 0.7 % of that of PLGA SNAs (Figure 1D). However, when the CaCl₂ PLGA SNAs and CaCl₂-siRNAs were treated with PicoGreen^™ solution containing 40 mm ethylenediaminetetraacetic acid (EDTA), which chelates Ca²⁺ ions, the fluorescence intensities of CaCl₂ PLGA SNAs and CaCl₂-siRNAs increased to a level similar to that of the PLGA SNAs. These data are consistent with the conclusion that Ca²⁺ ions are indeed associated with the oligonucleotide shell of the SNA and support our hypothesis that the multivalent structure of the SNA, comprised of densely functionalized oligonucleotide shell, has a higher affinity for Ca²⁺ ions compared to its linear subunits.

To test whether Ca²⁺ ions remain associated with the SNAs upon cellular entry, CaCl₂ PLGA SNAs and CaCl₂-siRNAs were dialyzed in solutions that mimic the physiological calcium and sodium chloride concentrations (0.1 m HEPES, 137 mm NaCl and 1.8 mm CaCl₂ at 37°C), and the PicoGreen^™ exclusion assay was subsequently performed at various time points up to 12 h to determine the amount of Ca²⁺ associated with the SNAs or siRNA duplexes at any given time point. Approximately 55.0 ± 5.5 % of the Ca²⁺ ions were still bound to the SNAs after 12 h (Figure 1E) while 84.4 ± 9.2 % is lost from the siRNA at the same time point. This observation indicates that while Ca²⁺ ions dissociate from SNAs under physiological conditions, the rate of dissociation is less than that for siRNA and consistent with the hypothesis that CaCl₂ PLGA SNAs can enter cells with much of the Ca²⁺ still intact. Note that conventional SNAs readily enter cells over the 2-12 h time window.^[7,8,21,70] We attribute this higher affinity for Ca²⁺ ions to the multivalent three-dimensional SNA structure as shown in Figure 1D.

CaCl₂ PLGA SNAs exhibit enhanced cellular uptake compared to PLGA SNAs

We analyzed the cellular uptake of linear siRNAs or SNAs in U87-MG glioblastoma cancer cells using Cyanine 5 (Cy5)-labeled siRNA, and oligonucleotide accumulation was assessed using flow cytometry after 1 h. The cells treated with PLGA SNAs had 3.7-fold higher median fluorescence intensity (MFI) compared to cells that were treated with linear siRNAs (Figure 2A, p < 0.01). This observation is consistent with previous reports that show that the multivalent, three-dimensional spherical architecture of the SNAs enhances the cellular uptake of siRNAs.^[8,27,70] The CaCl₂ PLGA SNAs remarkably had a 36.1-fold higher MFI than PLGA SNAs (p < 0.0001) and a 2.5-fold higher MFI than CaCl₂-siRNAs (p < 0.0001) (Figure 2B). The increased uptake of CaCl₂ PLGA SNAs and CaCl₂-siRNAs compared to SNAs can be attributed to their more positive zeta potential, a consequence of neutralizing the negative charge of the siRNA with the divalent Ca²⁺ ions, which reduces electrostatic repulsion between the SNAs and the negatively charged cell membrane. Moreover, it has been previously reported that Ca²⁺ ions can also facilitate endocytosis in cells due to the Ca²⁺ concentration gradient between the extracellular and intracellular space.^[81,82] Improved uptake of CaCl₂ PLGA SNAs relative to CaCl₂-siRNAs can be attributed to the three-dimensional oligonucleotide shell of the SNAs, which facilitates scavenger A mediated endocytosis through multivalent binding. ^[8,9]

Figure 2. Cellular uptake of PLGA SNAs and CaCl₂ PLGA SNAs in U87-MG cells.

All cells were treated at an Cy5 labeled siRNA concentration of 100 nm for 1 h and the fold change in median fluorescence intensity (MFI) of each treatment was normalized to the MFI of the untreated cells. (A) Comparison of the cellular uptake of linear siRNAs and PLGA SNAs. (B) Comparison of the cellular uptake of CaCl₂-siRNAs and CaCl₂ PLGA SNAs (salted to 230 mm CaCl₂). (C) Effect of inhibitors on the uptake of PLGA SNAs. (D) Effect of inhibitors on the uptake of CaCl₂ PLGA SNAs. Fucoidan concentration: 50 μg/mL, Nifedipine concentration: 20 μM. The error bars are the SD of three independent measurements. ns: not significant, ** p< 0.01, **** p < 0.0001. Note the substantial difference in y-axis scales.

To probe the mechanism of cellular uptake for the CaCl₂ PLGA SNAs, we inhibited various pathways and measured the effect of such inhibition on cellular uptake. When pre-treated with fucoidan (FCD), a class A scavenger receptor blocker, PLGA SNA uptake decreased by 70% (p < 0.0001) (Figure 2C), consistent with previous reports that cellular uptake of conventional SNAs is driven by scavenger receptor A-mediated endocytosis. ^[8,9] Pre-treatment with FCD also significantly decreased the cellular uptake of CaCl₂ PLGA SNAs by 98% (p < 0.0001) (Figure 2D). Treatment with nifedipine, an L-type Ca²⁺ channel blocker,^[83] did not affect the cellular uptake of CaCl₂ PLGA SNAs, indicating that there is no direct cytosolic uptake of the CaCl₂ PLGA SNAs via Ca²⁺ channels. Moreover, transport via Ca²⁺ channels is size dependent, and SNAs are larger than the Ca²⁺ channel pore size of 5.1–6.2 Å.^[84,85]

CaCl₂ PLGA SNAs exhibit decreased accumulation in the late endosome compared to SNAs

SNAs have previously been shown to accumulate in late endosomes, and only a small fraction escape to the cytosol to associate with RNAi machinery and facilitate gene regulation.^[39] Thus, we hypothesized that increased cytosolic delivery would increase the gene regulation activity of SNAs. Hence, to determine whether CaCl₂ PLGA SNAs have improved cytosolic release, we examined the degree of colocalization of PLGA SNAs and CaCl₂ PLGA SNAs with markers of late endosomes via confocal microscopy. We first compared the cellular uptake between PLGA SNAs and CaCl₂ PLGA SNAs after 24 h treatment via flow cytometry and found cells treated with CaCl₂ PLGA SNAs had 4-fold higher MFI compared to the PLGA SNAs, suggesting that the difference in cellular uptake between PLGA SNAs and CaCl₂ PLGA SNAs is not as significant at 24 h as compared to the 1 h time point (Figure 2A, Figure 2B, and Figure 3A, Figure S4). U87-MG cells were pre-treated with CellLight^® Late Endosomes-GFP for 24 h to express the Rab7a-GFP fusion protein and then treated with PLGA SNAs or CaCl₂ PLGA SNAs containing Cy5-labeled siRNA for an additional 24 h (Figure 3B). The cells were then imaged, and quantitative colocalization analysis was conducted by calculating Mander’s overlap coefficients (MOC).^[86] CaCl₂ PLGA SNAs showed a significant decrease in siRNA colocalization with the late endosome marker (Rab7a) compared to PLGA SNAs (Figure 3C, MOC = 0.819 ± 0.025 for PLGA SNAs and MOC = 0.638 ± 0.042 for CaCl₂ PLGA SNAs), suggesting that CaCl₂ PLGA SNAs not only enter cells more readily than PLGA SNAs, but they are also trafficked differently within U87-MG cells, with the escape of SNAs from late endosomes a possible driver of this observation.

Figure 3. Intracellular trafficking analysis of PLGA SNAs and CaCl₂ PLGA SNAs in U87-MG cells via confocal microscopy.

(A) Comparison of the cellular uptake of PLGA SNAs and CaCl₂ PLGA SNAs after 24 h treatment at an siRNA concentration of 100 nM. Fold change in MFI of each treatment was normalized to the MFI of the untreated cells. (B) Representative confocal microscopy image showing Cy5-siRNA colocalization with late endosomes (Rab7a-GFP fusion protein) when treated with PLGA SNAs or CaCl₂ PLGA SNAs (salted to 230 mM CaCl₂). Areas of colocalization appear yellow in the merged image. (Scale bars = 5 μm) (C) Quantitative analysis of colocalization of Cy5-siRNA with Rab7a-GFP fusion protein by Mander’s overlap coefficient (MOC). The error bars are the SD of ten independent cell measurements with ten z-stack images per cell. **** p < 0.0001.

CaCl₂ PLGA SNAs exhibit enhanced silencing of luciferase (Luc2) gene in the U87-MG-Luc2 glioblastoma luciferase reporter cell line

To test our hypothesis that CaCl₂ PLGA SNAs will exhibit improved gene regulation activity compared to conventional PLGA SNAs due to increased cytosolic delivery, we sought to evaluate the gene silencing activity of the CaCl₂ PLGA SNAs using the U87-MG-Luc2 glioblastoma cell line that stably expresses luciferase protein (Luc2) by measuring luciferase luminescence as an indicator of gene expression. We first confirmed that Luc2-targeting siRNAs led to efficient gene silencing by transfecting the siRNA with Lipofectamine RNAiMAX (a cationic lipid transfection reagent) as compared to cells treated with RNAiMAX only or transfected non-targeting siRNAs (Figure S5). Next, cells were treated with PLGA SNAs or CaCl₂ PLGA SNAs containing the same Luc2-targeting siRNAs (Luc2 PLGA SNA and Luc2 CaCl₂ PLGA SNA) and non-targeting control siRNAs (Ctrl PLGA SNA and Ctrl CaCl₂ PLGA SNA) for 48 h. Cells were treated using a siRNA concentration of 100 nm, while the concentration of CaCl₂ with which the PLGA SNAs were salted was varied. CaCl₂ PLGA SNAs showed CaCl₂ concentration-dependent gene regulation activity, as measured by the luminescence assay (Figure 4A). The knockdown efficiency of SNAs improved when salted with increasing concentrations of CaCl₂ with maximum knockdown activity reached at 230 mm CaCl₂ (knockdown efficiency 97.5 ± 0.3%) enhancing SNAs’ gene silencing efficiency up to ~ 20-fold compared to conventional NaCl salted PLGA SNAs (knockdown efficiency 4.9 ± 5.5%). Interestingly, while only 22% decrease in MOC was observed between the siRNA and late endosome marker for PLGA SNA compared to CaCl₂ PLGA SNAs, the significant enhancement in gene silencing activity is consistent with published reports.^[41,87] In other words, systems that exhibit a small decrease in colocalization signal between siRNA and endolysosomal markers can exhibit significantly different downstream cellular processes.

Figure 4. CaCl₂ PLGA SNAs exhibit enhanced gene regulation of the Luc2 gene in U87-MG-Luc2 reporter cell line.

(A) Knockdown potency of CaCl₂ PLGA SNAs after 48 h treatment salted to different CaCl₂ concentrations. CaCl₂ PLGA SNAs were administered at an siRNA concentration of 100 nm. (B) Cell viability of U87-MG-Luc2 cells after 48 h treatment of Luc2 targeting CaCl₂ PLGA SNAs salted to different CaCl₂ concentrations. (C) Comparison of Luc2 knockdown activity of CaCl₂ PLGA SNAs (salted to 230 mm CaCl₂). The final siRNA concentration treated to the cells across treatment groups was 100 nm (D) Effect of endosomal acidification inhibition on CaCl₂ PLGA SNAs. Fold changes in Luc2 protein expression in U87-MG-Luc2 cells with (+) or without (−) bafilomycin A1 co-treatment. The final siRNA concentration treated to the cells across treatment groups was 100 nm. The concentration of bafilomycin A1 pretreated to the U87-MG-Luc2 cells was 200 nm. The error bars are SD of three independent measurements. ns: not significant, * p < 0.05, ** p < 0.01, **** p < 0.0001.

This result also supports the conclusion that SNAs with higher amounts of associated Ca²⁺ led to enhanced cytosolic delivery of SNAs. To rule out the possibility that enhanced gene regulation activity of the SNAs salted at a higher CaCl₂ concentration is a function of the total particle uptake, we have examined the cellular uptake of SNAs salted at 130 mm and 230 mm CaCl₂, respectively, as these constructs exhibited different levels of Luc2 knockdown efficiency after 48 h (20.6 ± 2.2% and 97.5 ± 0.3%, respectively). At 48 h, the cellular uptake of these CaCl₂ salted PLGA SNAs was not statistically different, indicating this was not due to cellular uptake differences (Figure S6). Non-targeting CaCl₂ PLGA SNAs (Ctrl CaCl₂ PLGA SNAs) did not show significant Luc2 knockdown activity, confirming that the knockdown is sequence-specific (Figure 4A) and occurs due to the increased cytosolic delivery of the CaCl₂ PLGA SNAs. Cells treated with CaCl₂ solution at the same concentration used to salt SNAs did not elicit Luc2 gene silencing, indicating that the Luc2 knockdown activity of the CaCl₂ PLGA SNAs is due to RNAi and not CaCl₂ (Figure 4A)_.

Since Ca²⁺ ions are an important secondary messenger involved in apoptosis, excessive intracellular Ca²⁺ concentration may cause cell death.^[88] PrestoBlue^™ cell viability assay was used to determine if the concentration of CaCl₂ added to the cells was cytotoxic. The cells were treated with CaCl₂ PLGA SNAs salted with CaCl₂ concentrations ranging from 0 to 333 mm (Figure 4B). Consistent with previous findings,^[67] no significant (p > 0.05) cytotoxicity was observed at the concentrations studied. This is likely due to the fact that cells can tolerate elevated Ca²⁺ levels and prevent Ca²⁺-induced apoptosis by removing excess Ca²⁺ via Ca²⁺-ATPase pumps on the plasma membrane and mitochondrial calcium uniporter pumps on the mitochondrial membrane.^[38,89]

After determining that SNAs salted to 230 mm CaCl₂ could be used to achieve maximal Luc2 silencing, we next compared the gene silencing activity of CaCl₂ PLGA SNAs to the analogous linear siRNAs transfected with Lipofectamine^™ RNAiMAX or CaCl₂-siRNAs. When treated at an siRNA concentration of 100 nm for 48 h, CaCl₂ PLGA SNAs exhibited significantly enhanced knockdown activity compared to PLGA SNAs (94.38 ± 0.86% vs. 6.16 ± 4.02 % knockdown, respectively) and an equivalent level of knockdown with no statistical significance compared to linear siRNAs transfected with Lipofectamine^™ RNAiMAX (96.08 ± 0.70 % knockdown) (Figure 4C). CaCl₂ salting effectively enhances the cytosolic delivery and results in increased gene silencing activity of SNAs without inducing any significant cytotoxicity after 48 h, whereas RNAiMAX caused 47% decrease in cell viability over this time window (Figure S7). CaCl₂ PLGA SNAs also led to greater knockdown than CaCl₂-siRNAs (76.28 ± 5.39 % knockdown), likely due to enhanced cellular uptake and higher affinity for Ca²⁺ ions to the multivalent SNA architecture. While these experiments measured protein knockdown via luminescence measurements, gene silencing was also confirmed at the mRNA level using reverse transcription quantitative polymerase chain reaction (RT-qPCR) (Figure S8). To test if the increased gene knockdown of CaCl₂ PLGA SNAs is limited to siRNAs or if it can be extended to gene regulation via other pathways, CaCl₂ PLGA SNAs functionalized with Luc2 antisense DNA were used to study luciferase gene knockdown in U87-MG-Luc2 cells. Luciferase knockdown efficiency was measured after treating cells with antisense DNA (1 μm) for 48 h. Our results indicate CaCl₂ PLGA SNAs exhibited 43.93 ± 3.57 % knockdown activity whereas PLGA SNAs only exhibited 10.47 ± 1.86 % knockdown activity (Figure S9).

To better understand the enhanced cytosolic delivery of the CaCl₂ PLGA SNAs, we co-treated U87-MG-Luc2 cells with SNAs and bafilomycin A1, which inhibits the ATPase proton (H⁺) pump and prevents endosomal acidification.^[90] Co-treatment with bafilomycin A1 and CaCl₂ PLGA SNA increased the expression of Luc2 in U87-MG-Luc2 cells by 10.1 ± 0.60-fold compared to cells treated with CaCl₂ PLGA SNAs only, indicating prevention of endosomal acidification (Figure 4D). This data suggests a rapid increase of Ca² in the endosomes due to the entry of CaCl₂ PLGA SNAs results in enhanced influx of protons thereby triggering endosomal membrane disruption and increased cytosolic delivery consistent with previous literatures.^[55,66]

Next, we sought to investigate whether other divalent cations such as Mg²⁺, Zn²⁺, or Mn²⁺ have a similar effect on SNA activity and can be used to replace Ca²⁺ during the salting process of the PLGA SNAs. U87-MG-Luc2 cells were treated with Luc2 targeting PLGA SNAs formulated with MgCl₂, ZnCl₂, and MnCl₂ for 48 hours and then evaluated with the PrestoBlue^™ cell viability assay (Figure S10). The MnCl₂-salted PLGA SNAs and ZnCl₂-salted PLGA SNAs resulted in 84.9 ± 1.19 % and 95.1 ± 1.19 % reduction in cell viability, respectively. MgCl₂-salted PLGA SNAs (MgCl₂ PLGA SNAs) did not result in significant cytotoxicity compared to the untreated cells. Due to their cytotoxicity, the MnCl₂-salted PLGA SNAs and ZnCl₂-salted PLGA SNAs could not be tested for gene silencing activity. We then compared the gene regulation activity of MgCl₂ PLGA SNAs to CaCl₂ PLGA SNA. Surprisingly, the U87-MG-Luc2 cells treated with luciferase targeting PLGA SNAs salted with different concentrations of MgCl₂ did not induce significant down-regulation of Luc2 protein, whereas under similar conditions, Luc2 CaCl₂ PLGA SNAs resulted in 94.8 ± 2.96% knockdown (Figure S11). The inability of the MgCl₂ PLGA SNAs to effect gene regulation may be attributed to the fact that the cytosolic Mg²⁺ concentration (~1 mm),^[91] is ~10× higher than cytosolic Ca²⁺ concentration (~100 nm).^[92] Thus, Mg²⁺ ions delivered into the endosome do not induce an osmotic gradient between the endosome and the cytosol sufficient to result in endosomal membrane rupture and cytosolic delivery. Therefore, CaCl₂ became the salt of choice for facilitating SNA endosomal release.

To further determine whether Ca²⁺ complexation to PLGA SNAs can enhance the function of oligonucleotides in applications where cytosolic delivery is essential, we investigated PLGA SNAs (Ca²⁺ and Na⁺ forms, as well as linear forms) with thiazole orange (TO)-incorporated poly T21 flare DNA probes, which can bind to the poly A tail of mRNAs within the cytosol.^[93] Upon T21 flare DNA binding to the poly A tail, TO will undergo forced intercalation (FIT), resulting in fluorescent turn-on due to the restricted rotation of the methine bridge in TO (Figure S12A).^[71,94-97] First, we confirmed fluorescent enhancement of linear T21 flare DNAs as well as T21 flare PLGA SNAs and T21 flare CaCl₂ PLGA SNAs, where a 15.8 to 16.6-fold fluorescence increase was observed when co-incubated with complementary A21 RNAs (Figure S12B). Fluorescence enhancement was not observed with control A21 flare DNAs, A21 flare PLGA SNAs, or A21 flare CaCl₂ PLGA SNAs.

Next, we treated U87-MG-Luc2 cells with T21 and A21 linear flare DNAs, flare PLGA SNAs and flare CaCl₂ PLGA SNAs for 2 h and fluorescent pixel intensity was analyzed by live confocal fluorescence microscopy (Figure S12C). Cells transfected with linear T21 flare DNAs showed a 2.25 ± 0.53-fold turn-on (increase in mean pixel intensity) compared to cells transfected with control linear A21 flare DNAs (Figure S12D). Importantly, T21 flare CaCl₂ PLGA SNAs led to a 3.77 ± 0.80-fold fluorescent turn-on in cells compared to the control A21 flare CaCl₂ PLGA SNAs. Note that there was no significant difference in mean pixel intensity observed between cells treated with conventional (NaCl salted) T21 flare PLGA SNAs and A21 flare PLGA SNAs. Collectively, T21 flare CaCl₂ PLGA SNAs exhibit significant fluorescence turn-on compared to conventional T21 flare PLGA, consistent with the conclusion that CaCl₂ PLGA SNAs undergo greater endosomal release into the cytosol, compared to conventional PLGA SNAs.

To evaluate the utility of measuring fluorescence enhancement by confocal fluorescence microscopy, we conducted an ex cellulo experiment, incubating T21and A21 linear DNAs, PLGA SNAs and CaCl₂ PLGA SNAs with varying amounts of total RNA extracted from U87-MG-Luc2 cells. When incubated with total RNA amounts ranging from 500-2000 ng (corresponding to 35,000-150,000 cells worth of total RNA), T21 linear flare DNAs, conventional flare PLGA SNAs and CaCl₂ flare PLGA SNAs all exhibited fluorescence enhancement with linear dependence on the amount of RNA in the range of 2 to 5-fold, which is in agreement with the fluorescence intensity results obtained via confocal microscopy (Figure S12E).

CaCl₂ PLGA SNAs exhibit enhanced gene regulation activity of therapeutically relevant oncogenes in U87-MG glioblastoma and SK-OV-3 ovarian cancer cell lines.

Thus far, we have determined that CaCl₂ PLGA SNAs can downregulate Luc2 protein expression with high potency compared to conventional SNAs. We next sought to investigate whether CaCl₂ PLGA SNAs could be used to silence other genes to determine if this strategy is broadly applicable to other target sequences. We used a therapeutically relevant siRNA that targets isocitrate dehydrogenase 1 (IDH1), which is upregulated in primary glioblastoma (GBM) cells, inducing increased macromolecular synthesis, promoting aggressive tumor cell progression, and conferring resistance to radiation therapy.^[98,99]

After treating U87-MG cells with IDH1-targeting CaCl₂ PLGA SNAs for 48 h at 100 nm siRNA concentration, IDH1 protein expression was analyzed by Western blot (Figures 5A-B). CaCl₂ PLGA SNAs reduced IDH1 expression by 91.86 ± 3.40 %, whereas PLGA SNAs did not induce significant knockdown. CaCl₂ PLGA SNAs also exhibited significantly enhanced knockdown activity compared to CaCl₂-siRNAs (65.37± 4.32% knockdown), likely due to their enhanced uptake. Remarkably, CaCl₂ PLGA SNAs achieved a higher level of knockdown compared to that of the siRNAs delivered with Lipofectamine RNAiMAX (74.78 ± 3.65%), the current standard for siRNA transfection.

Figure 5. CaCl₂ PLGA SNAs enhance gene silencing activity of IDH1 in U87-MG cells and HER2 in SK-OV-3 cells.

(A) Representative Western blot showing the knockdown potency of CaCl₂ PLGA SNA targeting IDH1 after 48 h treatment. HSP70 served as a loading control. The band intensity was normalized to HSP70 and then quantified as relative expression compared to the untreated control. The final siRNA concentration treated to the cells across the treatment groups was 100 nm, and the CaCl₂ concentration salted to CaCl₂ PLGA SNA was 230 mm. (B) IDH1 densiometric analysis of Western blots. The error bars are the SDs of three independent experiments. (C) Comparison of HER2 knockdown activity of CaCl₂ PLGA SNAs in SK-OV-3 cell line. HER2 expression was quantified by an in-cell Western assay. The final siRNA concentration treated to the cells across the treatment groups was 100 nm, and the CaCl₂ concentration salted to CaCl₂ PLGA SNAs was 230 mm. (D) Cell viability of SK-OV-3 cells after 48 h treatment of HER2-targeting CaCl₂ PLGA SNAs salted to different CaCl₂ concentrations. Cell viability was measured using a CellTag 700 stain during in-cell Western Blot assay. The error bars are the SDs of three independent measurements. ns: not significant, * p < 0.05, **** p < 0.0001. Only key significances are shown for clarity.

To assess whether CaCl₂ PLGA SNAs can mediate gene regulation in a cell line other than U87-MG, we sought to knock down human epidermal growth factor receptor 2 (HER2) expression in SK-OV-3 human ovarian cancer cells. HER2 is a well-established oncogene that is involved in accelerated tumor growth, progression, and metastasis.^[100] After treating SK-OV-3 cells with CaCl₂ PLGA SNAs or control groups for 48 h at a concentration of 100 nm, HER2 protein expression was measured using an in-cell Western blot assay. Consistent with Luc2 and IDH1 protein knockdown results shown in the U87-MG-Luc2 and U87-MG cell lines, respectively, HER2-targeting CaCl₂ PLGA SNAs exhibited 77.22 ± 0.94 % reduction in HER2 protein levels while the analogous SNAs resulted in 10% reduction in HER2 protein levels (Figure 5C). CaCl₂ PLGA SNAs also appeared to induce greater knockdown than CaCl₂-siRNA complexes (65.99 ± 2.05 % knockdown) and Lipofectamine RNAiMAX-transfected siRNA, although the differences are not statistically significant. No further knockdown was achieved when cells were treated with SNAs salted with CaCl₂ concentrations higher than 230 mm, (Figure S13), indicating that maximal knockdown is achieved at this CaCl₂ concentration. Similar to what was seen with the U87-MG cell line, the use of non-targeting CaCl₂ PLGA SNAs or CaCl₂ alone did not result in significant knockdown, indicating that CaCl₂ PLGA SNAs do indeed achieve sequence-specific gene silencing (Figure S13). In addition, significant cellular toxicity was not observed for cells treated with CaCl₂ PLGA SNAs where SNAs were salted with CaCl₂ concentrations of 230 mm, 290 mm, and 333 mm (Figure 5D). These results indicate that CaCl₂ salting enhances the gene silencing activity of SNAs across a variety of different sequences and cell lines without eliciting significant cytotoxicity. Taken together, these results show that CaCl₂ salting enables SNAs to silence genes more effectively than conventional NaCl-salted SNAs, significantly expanding the applicability of SNAs for a wide range of disease targets for oligonucleotide therapeutics.

Conclusion

Nucleic acid constructs have become powerful biodiagnostic, gene regulation, and gene-editing probes, and are impacting the way we study, track, and treat disease. The use of SNAs to do so potentially expands cellular access, but it has been hampered sometimes by inefficient and uncontrollable endosomal release. This work establishes a new and straightforward methodology to enhance SNA cytosolic delivery, which takes advantage of the unusually high SNA affinity for Ca²⁺ ions. Specifically, this study shows that CaCl₂-salting instead of conventional NaCl-salting during SNA synthesis can increase cellular uptake by 36×, cytosolic delivery by 22%, and result in up to a 20× enhancement in gene regulation activity (in the case of siRNA constructs), with no apparent toxicity. While important for gene regulation, the ability to navigate the endosome and cytosol is also important for cellular detection and potentially gene editing purposes. The multivalent nature of the SNA leads to enhanced Ca²⁺ binding, allowing the Ca²⁺ to remain associated with the constructs, even under physiological conditions. The data are consistent with the notion that once in the endosomes, the Ca²⁺ ions establish a concentration gradient between the endosome and cytosol that leads to disruption of the endosomal membrane and release of SNAs. Consequently, this approach can be used universally with SNAs, regardless of diagnostic, gene regulation, or gene editing intent, as long as the target entities are outside the endosome.

Experimental Procedures

PLGA/PLGA-PEG-N₃ nanoparticle core (NP) synthesis

PLGA/PLGA-PEG-N₃ nanoparticle cores were synthesized using the nanoprecipitation method with slight modifications.^[70] PLGA (Resomer ^® 502H, Sigma Aldrich) / PLGA-PEG-N₃ (AI085, Akina Inc) (15.0 mg; 35%, w/w) was co-dissolved in acetonitrile (ACN) (6 mL) then injected dropwise into a 50 mL glass beaker containing 0.3% (v/v) Poloxamer 188 solution (24 mL) and stirred at 900 rpm. The resulting solution was allowed to evaporate for 2 hours in a fume hood. The NP solution was then concentrated to 1 mL using an Amicon filter (15 mL, size cutoff = 100K) (EMD Millipore).

Quantifying PLGA/PLGA-PEG-N₃ nanoparticle core concentration

Nanoparticle core concentration was quantified using a NanoSight NS300 (Malvern Instruments). A diluted sample solution (1:10,000 dilution, v/v in nanopure water) of the PLGA-PEG-N₃ core was injected using the NanoSight Sample Assistant (Malvern Instruments, United Kingdom). Each nanoparticle tracking analysis was conducted three times in duplicate using a default script provided by the manufacturer. Nanoparticle concentration was calculated based on the average of triplicate measurements.

Synthesis of CaCl₂-salted PLGA SNAs (CaCl₂ PLGA SNAs)

To prepare siRNA duplexes, DBCO-modified sense siRNA strands (20 nmole) and antisense siRNA strands (20 nmole) were hybridized in a duplex buffer (30 mm HEPES and 100 mm potassium acetate, pH 7.5, IDT technologies) by first heating the solution to 95 °C for 2 min, then cooling it to 25 °C in a heat block. The concentration of surface azide on the PLGA core was calculated based on previously reported methods.^[70],[101] PLGA/PLGA-PEG-N₃ (0.0143 nmol, ~500 μL) was added in pH 7.4, 0.1 M HEPES buffered saline, (HBS, 137 mm NaCl) with 0.3% (v/v) Poloxamer 188, and different concentrations of CaCl₂ ranging from 90 mm to 333 mm. Then, 20 nmole of DBCO-modified siRNA duplex was added and the reaction mixture was incubated for 24 hours at room temperature. The unreacted oligonucleotides were removed by 15 min of centrifugation at 10,000 × g using a 100 kDa cutoff Amicon spin filter four times using 1x HBS with 0.3% (v/v) Poloxamer 188 with different concentrations of CaCl₂. After the fourth wash, the particles were resuspended in 1x HBS with 0.3% (v/v) Poloxamer 188 with different concentrations of CaCl₂ and was stored until further use for biological studies.

Synthesis of PLGA SNAs (conventional SNAs that are salted with NaCl)

PLGA/PLGA-PEG-N₃ (0.0143 nmol, ~500 μL) was added mixed with 0.3% (v/v) Poloxamer 188 in pH 7.4, 0.1M HEPES buffered saline (HBS, 137 mm NaCl) and additional 5m NaCl was added to the solution so that the final concentration of NaCl is adjusted to 500 mm. Then, 20 nmole of DBCO-modified siRNA duplex was added to the reaction mixture, and the sample was incubated for 24 hours at room temperature. The un-reacted oligonucleotides were removed by 15 min of centrifugation at 10,000 × g using a 100 kDa cutoff Amicon spin filter four times using 1x HBS with 0.3% (v/v) Poloxamer 188. After the fourth wash, the particles were resuspended in 1x HBS with 0.3% (v/v) Poloxamer 188.

Particle size distribution by dynamic light scattering (DLS) and zeta (ζ) potential measurements

The particle size distribution and the surface charge (zeta potential) of the CaCl₂ PLGA SNAs and PLGA SNAs were measured using a Zetasizer Ultra Red (Malvern Instruments, UK). To measure the size, 1.40 was used as the refractive index.^[70] The hydrodynamic diameter (HD) measurements are derived from the number average value in water at 25 °C. The reported DLS size for each sample was based on at least five measurements per run in triplicates. The surface charge (zeta potential) of the particles was measured in triplicates using the DTS 1070 zeta cell (Malvern Instruments, UK), each run was measured in water at 25 °C, and 10 to 50 measurements were taken using the automated settings in the ZS Xplorer (Malvern Instruments, UK) software.

PicoGreen exclusion assay to evaluate the association of Ca²⁺ ions within the CaCl₂ PLGA SNAs and Ca^2+-siRNA complex

The complexation of Ca²⁺ ions in the SNA architecture was determined using a Quant-iT^™ PicoGreen^™ (Invitrogen) exclusion assay. 50 μL of CaCl₂ PLGA SNAs, PLGA SNAs and Ca^2+-siRNA complexes (CaCl₂-siRNAs, that has equal oligonucleotide concentration of 100 nm compared to CaCl₂ PLGA SNAs and PLGA SNAs and equal CaCl₂ concentration of 230 mm compared to CaCl₂ PLGA SNAs) were first added to the 96-well plate and then 150 μL of PicoGreen^™ solution containing either 1x HBS or 1x HBS + 40 mm EDTA was added to the 96-well plate. The plate was then read at excitation/emission wavelengths of 480 nm/520 nm using a Biotek Cytation 5 plate reader.

PicoGreen exclusion assay to evaluate the dissociation of Ca²⁺ ions within the oligonucleotide shell of the CaCl₂ PLGA SNAs and CaCl₂-siRNA complex within multiple time points.

300 μL of CaCl₂ PLGA SNAs and CaCl₂-siRNAs (salted at 230 mm CaCl₂ with same siRNA concentration) was dialyzed against 50 mL of RNase free 0.1 M HEPES, 137 mm NaCl, 1.8 mm CaCl₂ solution at 37°C using 3.5K MWCO Slide-A-Lyzer^™ MINI Dialysis Device (Thermo Fisher) using a 100 mL beaker with a stir bar at 150 rpm. At multiple time points (0h, 1h, 2h, 4h, 8h and 12h), 30 μL of both CaCl₂ PLGA SNA and CaCl₂-siRNA complex solution was taken out of the dialysis tube. Then 5 μL of the CaCl₂ PLGA SNAs and CaCl₂-siRNAs was first added to the 96 well plate followed by addition of 95 μL of PicoGreen^™ solution containing either 1x HBS or 1x HBS + 40 mm EDTA. The plate was then read at excitation/emission wavelengths of 480 nm/520 nm using a Biotek Cytation 5 plate reader. Then the normalized fluorescence (NF) at each time point was calculated by the following equation, $\frac{\text{RFU (Picogreen})}{\text{RFU (Picogreen}+40\mathrm{m}\mathrm{M}\text{EDTA})}$. Remaining Ca²⁺ ion associated with the siRNA or SNA at each time point (x hour) then was calculated by $100\%\times (1-\frac{Average{NF}_{t=\mathrm{o}h}-{NF}_{t=xh}}{Average{NF}_{t=\mathrm{o}h}})$. The experiment was performed in triplicates.

Cell culture

The U87-MG glioblastoma cell line (ATCC) was cultured in MEM (Gibco) supplemented with 10% fetal bovine serum (FBS) (Gibco) and 1% penicillin-streptomycin (Thermo Fisher). The U87-MG-Luc2 reporter cell line (ATCC) was cultured in MEM supplemented with 10% FBS, 8 μg/mL blasticidin (Thermo Fisher) and 1% penicillin-streptomycin (Thermo Fisher). The SK-OV-3 ovarian cancer cell line (ATCC) was cultured in DMEM (Gibco) supplemented with 10% FBS and 1% penicillin-streptomycin (Thermo Fisher).

Evaluation of CaCl₂ PLGA SNA and PLGA SNA uptake by flow cytometry

100,000 (1×10⁵) U87-MG cells were seeded in a 12-well plate in 2 mL of cell culture media and incubated overnight. The cells were then treated with 100 nm by siRNA duplex concentration with the following treatment groups: Cy5-labeled linear siRNA duplex, Cy5-labeled CaCl₂-siRNA, Cy5-labeled PLGA SNAs, Cy5-labeled CaCl₂ PLGA SNAs (salted to 230 mm CaCl₂) for 1 hour or 24 hours. As a negative control, cells were pre-treated with 50 μg/mL of fucoidan (Sigma Aldrich), a scavenger receptor A blocker, for 30 minutes. Also, cells were pre-treated with nifedipine, a known calcium ion channel blocker at a concentration of 20 μm for 30 minutes. Then, these cell samples were treated under the above-mentioned conditions (but with the presence of the inhibitors). After 1 hour of incubation, the cells were washed with 1X HBS, trypsinized, and washed twice by centrifugation and resuspension in 1X HBS (300 x g for 5 min). For live-dead cell staining, LIVE/DEAD^™ fixable blue dead cell stain (Invitrogen) solution in 1X HBS buffered saline was used; the cells were incubated with it for 15 min at 4 °C. The cells were then washed with 1X HBS, fixed using 4% paraformaldehyde solution for 10 minutes, washed and were re-suspended in 1X HBS. The median fluorescence intensity (MFI) of the Cy5 signal was recorded with a FACSymphony^™ A3 (BD Biosciences). The experiments were performed in triplicate, and the data were analyzed using FlowJo software (BD Biosciences).

Analysis of intracellular trafficking of PLGA SNAs and CaCl₂ PLGA SNAs by confocal microscopy

To visualize SNA uptake and intracellular trafficking, U87-MG cells were plated on an 8-well chambered coverglass slide (Nunc^™ Lab-Tek^® II) with a seeding density of 25,000 cells per well with a total volume of 400 μL of cell culture media. After overnight incubation, the cells were treated with CellLight^™ Late Endosomes-GFP, BacMam 2.0 (Thermo Fisher) at a particle per cell (PPC) of 40 according to the manufacturer’s protocol. After a 24 h incubation, the cells were treated with Cy5-labeled PLGA SNAs or CaCl₂ PLGA SNAs (salted at 230 mm CaCl₂) ([siRNA] =100 nm) for 24 h. The cells were then washed with washing buffer (HBS containing 0.9 mm CaCl₂ and 0.49 mm MgCl₂-6H₂O) three times. Then, the cells were fixed (4% paraformaldehyde) for 15 minutes, washed three times with washing buffer and their nuclei were stained with NucBlue^™ Fixed Cell ReadyProbes^™ Reagent (DAPI) (Thermo Fisher) according to the manufacturer’s protocol. Confocal images of the cells were collected using a Zeiss LSM 800 microscope using equal parameters for image acquisition for each treatment group (e.g., laser power, master gain, offset). Z-stack images (10 slices) of the cells were used to analyze colocalization of Cy5-labeled oligonucleotides within the late endosomes (Rab7a-GFP fusion protein) throughout the entire volume of the cells. Regions of interest (ROI) were assigned by manually tracing the outlines of individual cells. Mander’s overlap coefficients (MOCs) were quantified by reconstruction of the Z-stack images of each cell using Zeiss ZEN Blue software.^[86] Statistical analysis was performed across averages from 10 independent cell images per treatment group.

Cell viability assay

The cell viability of U87-Luc2 cells was determined using PrestoBlue^™ cell viability reagent (Thermo Fisher). The cells were seeded in a black, clear-bottom, 96-well plate at a density of 12,000 cells per well. After overnight incubation, the cells were treated with CaCl₂ PLGA SNAs with different concentrations of CaCl₂. The final concentration of siRNA treated to cells was kept constant at 100 nm. After treatment for 48 hours, cell viability was measured following the manufacturer’s protocol. After incubation, the fluorescence was measured at excitation/emission wavelengths of 560 nm/590 nm using a BioTek Cytation 5 Microplate Reader. The cell viability was normalized to the untreated control and plotted as a percentage of cell viability. The experiment was performed in triplicates, and the error was calculated as the standard deviation of the mean.

Quantification of Luc2 protein down regulation by luciferase assay.

To assess the functionality of CaCl₂ PLGA SNAs, U87-MG-Luc2 cells were seeded in a in a black, clear-bottom, 96-well plate at a density of 12,000 cells per well with a total volume of 200 μL. After overnight incubation, the cells were treated with CaCl₂ PLGA SNA, non-targeting CaCl₂ PLGA SNAs, and CaCl₂ solution equivalent to concentration that the CaCl₂ PLGA SNA was salted in. The final siRNA concentration of the cells samples for all treatment groups was 100 nM. After a 48-hour treatment, the wells were washed with 1x HBS twice, and 90 μL of fresh cell culture media was added to the wells. Subsequently, 10 μL of PrestoBlue^™ Cell viability reagent (Thermo Fisher) was added to measure the relative cell viability compared to the untreated wells. After a 30-minute incubation at 37 °C, fluorescence intensity was measured at excitation/emission wavelengths of 560 nm/590 nm using a BioTek Cytation 5 Reader. After measuring the cell viability, the wells were then washed with 150 μL of HBS three times. The luminescence of U87-Luc2 cells were then measured using the Bright-Glo^™ Luciferase Assay (Promega) using Biotek Cytation 5 plate reader. Luc2 protein expression was analyzed in arbitrary units where the luminescence value was normalized to the fluorescence value from the PrestoBlue assay. . Then, the relative Luc2 expression was normalized to the untreated control group. To compare the Luc2 protein down regulation activity, CaCl₂ PLGA SNAs (salted to 230 mm CaCl₂) were treated to the cells along with CaCl₂-siRNAs and PLGA SNAs. As a positive and negative control, linear Luc2 and Control siRNA duplexes were transfected at a siRNA concentration of 100 nm with Lipofectamine RNAiMAX (Thermo Fisher). To minimize cellular cytotoxity for RNAiMAX treated cells, they were washed with PBS and fresh media was added after a 6-hour treatment step. To analyze the effect of endosomal acidification on CaCl₂ PLGA SNA-mediated gene silencing, the cells were co-treated with bafilomycin A1 at a concentration of 200 nm. Then, the cells were treated with CaCl₂ PLGA SNAs (salted to 230 mm CaCl₂) in the presence of bafilomycin A1. After 48-h incubation, the relative Luc2 protein expression of cells co-treated with bafilomycin A1 and CaCl₂ PLGA SNAs was normalized to cells that were treated with CaCl₂ PLGA SNAs only.

Quantification of IDH1 protein down-regulation by Western blot analysis

100,000 (1×10⁵) U87-MG cells were seeded in a 12-well plate and incubated overnight with a total volume of 2 mL. To compare the IDH1 protein down regulation activity, the cells were treated with CaCl₂ PLGA SNAs (salted to 230 mm CaCl₂), CaCl₂-siRNAs or PLGA SNAs. As a positive and negative control treatment group, linear Luc2 and Control siRNA duplex were transfected at a siRNA concentration of 100 nm with Lipofectamine RNAiMAX (Thermo Fisher). To minimize cellular cytotoxity for the RNAiMAX treated cells, they were washed with PBS and fresh media was added after a 6-hour treatment step. After a 48-h incubation, the wells were washed with 1x HBS twice, and the protein lysates were then extracted using radioimm unoprecipitation (RIPA) buffer with halt protease inhibitor cocktail (Thermo Fisher). After measuring protein lysate concentration using a BCA assay reagent (BioRad), 30 μg of protein lysate per treatment sample was separated using 4-12% NuPAGE^™ Bis-Tris protein gel (Invitrogen) in 100 V for 70 minutes. Then, the protein gel was transferred to nitrocellulose membrane (Life Technologies) using an iBlot^® 2 Gel Transfer Device (Life Technologies). The membranes were blocked with Intercept^® (TBS) Blocking Buffer (LI-COR) in room temperature for 1 hour with shaking and incubated overnight at 4 °C with shaking using the following antibodies: rabbit anti-IDH1 (Cell Signaling Technology, 1:1000 dilution in blocking buffer, 10 mL) and mouse IgG1 anti-HSP70 (BD biosciences, 1:2000 dilution in blocking buffer, 10 mL). After the blots were washed with 1× PBST (0.1% Tween-20) three times for 5 min, the membranes were incubated with IRDye^® 800CW-conjugated goat anti-rabbit secondary antibody (LI-COR, 1:2000 dilution in blocking buffer, 10 mL) and IRDye^® 800CW-conjugated goat anti-mouse IgG1 secondary antibody (LI-COR, 1:2000 dilution in blocking buffer,10 mL) for 1 hour in room temperature with shaking. Then, the nitrocellulose membrane was washed with 1× PBST three times for 5 min. To remove residual Tween-20, the membrane was rinsed in deionized water three times before scanning. Then, the blot image was acquired using an Odyssey^® CLx Imager (Li-COR) at 169 μm resolution in the 800-nm fluorescence channel. Then, the band intensity of the blot was quantified using Image J (NIH, Besthada, MD) ^[102] and normalized to the untreated control group. Western blot was performed in triplicate, and the results were plotted in a bar graph, and the error was calculated as the standard deviation of the mean.

Evaluation of HER-2 protein down-regulation by In-cell Western blot analysis

In a black, clear-bottom, 96-well cell culture plate, 10,000 SK-OV-3 cells were seeded with a total volume of 180 μL per well. After overnight incubation, the cells were treated with CaCl₂ PLGA SNAs, the final siRNA concentration in the cells for all treatment groups was 100 nm. PLGA SNAs were salted at a CaCl₂ concentration of 230 mm, 290 mm, and 333 mm. Moreover, CaCl₂-siRNAs were treated to cells with an equivalent siRNA concentration and CaCl₂ concentrations. 100 nm of linear HER2 siRNA and control siRNA were transfected with lipofectamine RNAiMAX (Thermo Fisher) to the cells as a positive and negative control, respectively. To minimize cellular cytotoxity for RNAiMAX treated cells, they were washed with PBS and fresh media was added after a 6-hour treatment step. After a 48-hour treatment, the wells were washed with 1× HBS with 4 mm EDTA two times, then 1× PBS, then fixed in methanol chilled at −20 °C for 15 min. The wells were then washed with 0.05% Tween-20 in 1× PBS two times, then 1× PBS, then incubated with Intercept^® (TBS) Blocking Buffer (LI-COR) for 90 min with shaking. The cells were then incubated with HER2 antibody (29D8) (Cell Signaling Technology) diluted 1:200 in Intercept^® (TBS) Blocking Buffer for 2 h. The wells were washed with 0.1% Tween-20 in 1× PBS three times and incubated with 2 μg/mL IRDye^® 800CW-conjugated goat anti-rabbit secondary antibody (LI-COR) and 500 nm CellTag 700 (LI-COR) diluted 1:500 in Intercept^® blocking buffer for 1 h protected from light, with shaking. The wells were washed with 0.1% Tween-20 in 1× PBS three times and imaged on an Odyssey CLx system (LI-COR). HER2 protein expression was calculated in arbitrary units by normalizing fluorescence at 800 nm (HER2) to fluorescence at 700 nm (cell viability number). Then, the extent of HER2 protein knockdown was determined by normalizing the HER2 protein expression to that of the untreated control group. Western blot was performed in triplicate, and the results were plotted in a bar graph, and the error was calculated as the standard deviation of the mean.

Statistical analysis

Significant differences between groups were determined using a student’s two tailed t test (Mander’s overlap coefficient analysis) and one-way or two-way ANOVA tests with post-hoc Tukey and Sidak multiple comparison test using Prism Version 9.4.1 (Graphpad). Differences were considered statistically significant at p < 0.05.