Biotech

Comparative optimization of polysaccharide-based nanoformulations for cardiac RNAi therapy – Nature Communications

Time Stamp: June 25, 2024 8:00 PM
Source Node: 623759

A robotic assisted microfluidics platform for nanoparticles formation screening

An ideal RNAi delivery system was customized to screen and determine the feasibility and physical properties over a range of dynamic parameters. Followed with our design criteria, a rapid, one-step synthetic method was developed through the high-throughput microfluidics workstation coupled with streamlined dynamic light scattering (DLS) analysis system (Fig. 2a, Supplementary Fig. 2; Supplementary Video S1). Such integrated NPs’ production-characterization system automates the screen process while maintain the reproducibility, which is ideal for in vitro screening of NP-formulations. Briefly, EEPG was immobilized as ionic part, which was further co-assembled with oppositely charged molecules, including polymers, lipids, small molecular drugs and cell-penetrating peptides (Fig. 2b). The fabrication of EEPG-based NPs was achieved by microfluidics assisted nanoprecipitation method, where EEPG was fixed as outer phase and different cationic molecules were adopted as inner phase, and the NPs formation was triggered by passive microfluidic mixing in a co-flow microfluidic devide^13,14. By changing flow conditions of injection system, different weight ratios or nitrogen: phosphate (N/P) ratios between cationic materials and EEPG were evaluated and simultaneously analyzed by the online size characterization system, which integrates an automatic dynamic light scattering (DLS) instrument. The detailed information in collection system (solvents, flow conditions, inner phase versus outer phase) are listed in Supplementary Table 1. As such, multiple formulations were tested, and the average derived counting rate over 6000 kcps was indicative to the formation of NPs. The overall screening results are summarized in Fig. 2c. The formation of stable NPs is shown as blue dot, whereas grey dot indicates the weak interaction of the two compartments without pervasive formation of NPs.

**Fig. 2: “Robot assisted” microfluidic preparation and screening of EEPG-paired nanoparticles.**

In the screening pool, we set 7 different weight ratios (w/w) between each category of cationic biomaterials and EEPG for our initial screens. From the standpoint of N/P ratios, all the tested w/w ranges covered the N/P ratio at 1/1, where the intermolecular electrostatic interaction between protonated amino groups and corresponded phosphorylated groups was expected to be maximally evident¹⁵. For the category 1 (C1), cationic polymer-based biomaterials, due to their significant variation in molecular weight and amino substitution rate among different polymers¹⁶, we chose to adopt the corresponding N/P ratios to better interpret the NPs formation in an ionotropic gelation scenario. The detailed information regarding weight ratios and corresponded N/P ratios in each subpanel is shown in Supplementary Table 2–5. The hydrodynamic diameter and zeta potential of representative nanoformations is shown in Fig. 2d, e, and promising pairings from each category are imaged by transmission electron microscopy (TEM) (Supplementary Fig. 3).

For C1, we noticed that under a same N/P ratio, amine-modified dextran failed to form NPs with EEPG at all tested sets, whereas other cationic polymers, including chitosan and PEI, can feasibly form NPs under the N/P ratio near 1/1, which is consistent with previous studies¹⁷. In C2, which represents the small molecular drugs, despite their highly enriched amine groups in molecular structure, none of them can form into stable NPs around N/P 1/1, where spermine can only form into NPs at high weight ratio (8/1, corresponding to N/P = 30/1). For cationic cell-penetrating peptides (C3), we observed pervasive NPs formation in all tested molecules at weight ratio 1/1 and 1/2 (which covered the N/P range 1/1), where K9 and KALA can also form into NPs at weight ratio 8/1. In C4, which represents cationic lipids, despite Lipid 8 and 10 cannot form into NPs with EEPG at the N/P around 1/1, they can form into stable NPs with EEPG with extra high amount of EEPG, where the N/P is lower than 0.04/1, whereas Lipid 14 cannot form NPs with EEPG in all tested formulations (Supplementary Table 6).

Next, a subset of 16 nanosystems with the most optimal physical characters (smallest size with the PDI < 0.2) were adopted for further screen and evaluations (Fig. 2d, e).

siRNA encapsulation efficiency of nanoformulations

Efficient siRNA encapsulation and protection is another critical aspect for formulation optimization, therefore we further assessed the siRNA encapsulation efficiency using afore-defined nanosystems with different cationic species, and Ribogreen assay was initially adopted for quantifying the RNA encapsulation¹⁸. Previous studies emphasize the efficient encapsulation of siRNA majorly relies on the electrostatic interaction between cationic biomaterials and negatively charged phosphate backbone of nucleic acid, as indicated by the (N/P) ratio¹¹. Therefore, we evaluated the siRNA encapsulation efficiency of afore-screened formulations at fixed N/P ratios of 15/1 or 20/1 respectively (Figs. 3a, b). As expected, under the same N/P ratio, the selection of cationic compounds significantly affected the encapsulation efficiency of siRNA. Meanwhile, for most of the nanosystems, the encapsulation efficiency (EE%) was positively correlated with N/P ratio, with the exception of Transportan-EEPG NPs, where we noticed a significant reduction of EE at N/P 20/1 comparing to that at 15/1. For cationic polymers, the EE% for PEI is pervasively higher than that from chitosan. A comparable EE% was observed between spermine-EEPG NPs and PEI-EEPG NPs. For cell-penetrating peptides, besides of Transportan, K9/KALA/Penetratin-paired NPs showed enhanced EE (> 80%) after increasing N/P ratio to 20/1. In category of cationic lipids, the positive correlation between N/P ratio and EE% was observed across all tested molecules, with Lipid 2/Lipid 10-paired NPs excelling by achieving an EE > 80%, outperforming other lipids.

**Fig. 3: Step-wise screening of EEPG-paired nanoparticles for potential siRNA delivery.**

From a pharmaceutical standpoint, to maximize the siRNA encapsulation with low consumption of biomaterials, we chose the nanosystems with EE% over 80%, that included Spermine/PEI800/PEI25K/Lipid 2/Lipid 10/K9/KALA-EEPG NPs, for following screening.

Biocompatibility of the nanoformulations

With a satisfied siRNA encapsulation efficiency, we then sought to evaluate the in vitro biocompatibility of afore-filtered nanosystems. Considering the macrophages tropism capability from the NPs, which is inherently generated by the addition of EEPG⁶, the biocompatibility of promising formulations was tested on human monocytic cells THP-1 and RAW 264.7 murine macrophages cells. THP-1 monocytes were pre-stimulated by 100 ng mL⁻¹ phorbol myristate acetate (PMA) for 24 h to obtain macrophages. Afterwards, THP-1 derived macrophages and RAW 264.7 cells were separately incubated with 10–500 µg mL⁻¹ of various EEPG-paired NPs for 48 h. We have previously confirmed the satisfied biocompatibility from EEPG towards different macrophages cell lines⁶. Upon the ionotropic gelation with different cationic compounds to form NPs, spermine-EEPG NPs showed the most significant loss of cell viability even at lowest doses (10 µg mL⁻¹). Low-molecular-weight PEI (PEI, Mw = 800) based NPs (PEI800-EEPG) showed an overall better cellular tolerability compared to their high-molecular-weight counterparts (PEI25K-EEPG, IC50 = 298 µg mL⁻¹ vs. 227 µg mL⁻¹), which is consistent with previous reports, where the molecular weight of PEI is positively associated with cytotoxicity¹⁹. For cationic lipids, Lipid 2 showed substantial loss of cell viability (< 50%) at dosages higher than 75 µg mL⁻¹, where Lipid 10 only showed significantly cytotoxic at dosages over than 250 µg mL⁻¹. Compared to other categories, K9-EEPG and KALA-EEPG NPs showed an altered, yet relatively satisfied biocompatibility towards macrophages, where a significant cellular toxicity was only observed at highest dosages conditions (500 µg mL⁻¹, Fig. 3c). In addition, the biocompatibility of different EEPG-paired NPs in RAW 264.7 cells were pervasively higher than that in THP-1 cells (Supplementary Fig. 4). Considering the inherent targeting capability from EEPG towards a C-type lectin-1 receptor, Dectin-1²⁰, the difference may be partly attributed to the distinct Dectin-1 expression profile from different cells, where RAW 264.7 is known as a Dectin-1^– cell line and THP-1 cells are Dectin-1⁺^21,22. The biased biocompatibility of EEPG-based NPs also partly confirmed the preserved Dectin-1 targeting capability from EEPG after fabricating into NPs. Nevertheless, with these results, we excluded spermine-EEPG NPs for further screening to assess their efficacy on gene silencing.

Gene silencing efficacy of nanoformulations

With the identification of in vitro biocompatibility, finally, we evaluated the gene silencing efficacy of afore-filtered nanosystems. In total 6 EEPG-based NPs were selected, including PEI800-EEPG, PEI25K-EEPG, Lipid 2-EEPG, Lipid 10-EEPG, K9-EEPG and KALA-EEPG. GAPDH-counteracted siRNA (siGAPDH) was adopted as a model siRNA to be encapsulated. Due to the Dectin-1 targeting efficiency from EEPG²³, Dectin-1⁺ human monocytic cells, THP-1, were utilized to test the knockdown efficiency of different formulations⁶. The working concentration of siGAPDH was kept at 50 nM in all tested NPs, after incubating with cells for 48 h, the silencing efficiency towards GAPDH was quantitatively evaluated by western blotting analysis, with α-Tubulin serving as internal control (Fig. 3d). The western blotting results indicated a superior gene silencing efficacy from siGAPDH-loaded PEI25K-EEPG NPs comparing to PEI800-EEPG. For cationic lipid-paired NPs, despite Lipid 2 and Lipid 10 exhibited almost identical chemical composition and molecular structure, we observed the gene silencing efficiency from Lipid 2-EEPG NPs was slightly higher than that from Lipid 10-EEPG NPs, despite no statistical significancy was detected. Similar to cationic lipids, cationic peptides-paired NPs also exhibited potent gene silencing efficacy. Similarly, in a same siGAPDH concentration (50 nM), K9-EEPG and KALA-EEPG NPs showed slightly difference in regarding to gene knockdown efficiency, and among which, KALA-EEPG NPs exhibited the highest knockdown levels (73%) at the siRNA concentration of 50 nM (Fig. 3e).

In summary, we conducted a step-wise screening of formulation optimization, including NP-formation, siRNA encapsulation, biocompatibility and in vitro knockdown efficiency. The generated pairing NPs composed by EEPG and cell-penetrating peptide KALA exhibited preferred properties in all screening steps and therefore will be further tested for biological applications in the following experiments.

Computational-assisted nanoformulation screening

Besides of the identified promising NPs, we sought to understand the pattern beneath the selection process. However, considering the limited size of the dataset, to design a robust algorithm for cationic compounds prediction by conventional cheminformatics method is challenging. Our goal was rather to extract potential scientific understanding from the screening process and to provide perspectives for interpretation methods, as this interpretability holds particular value for medical applications^12,24. We first produced a classification model in the form of a decision tree (DT) to summarize the screening process (Fig. 4a). Dark blue squares and grey squares separately indicated the positive and negative compound sets for each screening, and the light blue diamond described the screening criteria. For NPs formation, in a typical scenario of polyelectrolytes-based gelation, the binding and aggregation is usually prone to be observed in an electro-neutral manner²⁵. However, in the current study, the phenomenon can be divided into 3 categories: (1) cationic reagents that did not form stable NPs with EEPG at any concentration (samples 15, 16, 17, and 20); (2) cationic compounds formed colloidal stable NPs only at N/P ratios significantly deviating from unity (< 0.1 or > 10, samples 14, 18, and 19); and (3) cationic reagents that can steadily form into NPs (samples 1–13, Fig. 4b, Supplementary Fig. 5a). We first thought whether a molecular descriptor (MolDes) based quantitative structure-activity relationship (QSAR) model can provide potential explanation. MolDes of each cationic compound was generated by alvaDesc (Methods), which were sequentially adopted for logistic regression based feature selection with binary formation of NPs as label²⁶. However, even for MolDes with the lowest p-value (indicating the MolDes holds the most significant association with the NPs formation process)²⁷, its corresponding receiver operating characteristic (ROC) curve only showed an area under the curve (AUC, indicating the performance of the MolDes in distinguishing between the positive and negative classes) of 0.85 (Supplementary Fig. 5b, c), which limited the extraction of scientific understanding from the current model. To gain better insights, we turned to short (1 ns) molecular dynamics (MD) simulations to quantify non-covalent interaction potentials among EEPG and different cationic compounds on the basis of electrostatic interaction (Coulombs potential) and electronically neutral van der Waals interaction (Lennard-Jones, LJ potential)²⁸. For negative subsets (sample 14–20, Supplementary Fig. 5a), the inter-Coulomb potential between EEPG and counter-reagents consistently approached or even exceeded 0, whereas the corresponding values for positive group were typically below 0 (Fig. 4c), suggesting minimal contribution from electrostatic interaction in cationic reagents that did not form stable NPs with EEPG. We further evaluated the difference between the cationic compounds that did not form stable NPs with EEPG at any concentration (samples 15, 16, 17, and 20), and cationic compounds formed colloidal stable NPs only at N/P ratios significantly deviating from unity (< 0.1 or > 10, samples 14, 18, and 19, Supplementary Fig. 5a). Upon simulation, the inner LJ potential of EEPG in non-electronic binding groups were significantly lower, resembling a self-aggregation tendency from EEPG (Fig. 4d), partly suggesting the NPs formation may be attributed to the self-aggregation from EEPG through non-polar interaction. Overall, MD simulations may be adopted as a valuable supplementary tool for initial evaluations and insights into the inotropic gelation based NPs formation process, which is mainly reflected by observing the Coulomb’s potential during the simulation.

**Fig. 4: Computational assisted step-wise interpretation of formulation screening process.**

Next, we try to elucidate the critical factors in governing the siRNA encapsulation efficiency (EE%). Considering that the data generated from MD simulations may offer supplementary insights beyond the conventional MolDes computed by alvaDesc, we combined the data generated from MD with the conventional MolDes calculated by alvaDesc, which were further subjected to feature selection with EE% as labels. Feature selection analysis indicated that the average LJ potential between EEPG and counter-reagent had the highest r-square value (r² = 0.7, Fig. 4e), which is substantially higher than those conventionally calculated MolDes²⁹. Considering the LJ potential is correlated with the distance between two molecules, we hypothesize this correlation might be attributable to steric repulsion exerted by the cationic reagents during siRNA encapsulation³⁰, whereas the stronger interaction between EEPG and the cationic compound might induce spatial hindrance for siRNA residence, negatively affecting siRNA encapsulation. To confirm our hypothesis, we separately visualized binding of EEPG with KALA and Transportan. These two peptides were selected for comparison based on their similar molecular weight, hydrophobicity (LogP) and surface charges, yet markedly different EE% (KALA, EE% = 95%; Transportan, EE% = 25%). Different from EEPG-KALA complex, where KALA was mobilized in one tail of EEPG in a flexible manner, Transportan was enveloped by EEPG in a more compact structure (Fig. 4f). This was also partly reflected as EEPG-Transportan complex formed over ten times more close contacts (< 6 Å) than the EEPG-KALA complex (Supplementary Fig. 5d).

In our sequential investigation into the factors contributing to the observed biocompatibility of various formulations, we employed a similar independent feature selection methodology, utilizing the IC50 values as labels and the MolDes of each cationic compound as variables. Multiple studies have adopted MolDes to identify association between physicochemical properties and toxicity. Notably, molecular weight (Mw), lipophilicity (log P), and basicity (pKa of strongest base) have been identified in multiple pharmaceutical studies^31,32. However, the current results potentially suggested the significance of number and topological location of negatively charged atoms in mediating the promiscuity of a compound. This was evidenced as a MolDes, qnmax, which represented the maximum negative charge of the compound, potentially showed highest coefficient of determination (r² = 0.93, Fig. 4g, h, Supplementary Table 7). Similarly, the third leading MolDes, SPP, was also correlated with the absolute difference between the maximum negative charges and maximum positive charges (Supplementary Table 7)³³. Other leading MolDes included B02[O-O] and B10[N-N], which separately represented the presence/absence of O-O (N-N) at topological distance 2/10 (Supplementary Table 7)³⁴. Despite the hydrophobicity and molecular charges, mostly positive, have been widely acknowledged for assessing biocompatibility of a compound, from the above, we suggested that the distribution and ratio between negative and positive parts of a compound may also occupy a significant place in mediating the biocompatibility of the molecules.

In our last pursuit to elucidate the relationship between the chemical properties of cationic counterparts and the gene knockdown (KD) efficacy of the resultant NPs, feature selection was conducted with KD efficiency of each formulation as the label and MolDes of each cationic compounds as variables. Coefficients of determination of each MolDes is adopted for evaluating the impact of each MolDes, as it represents the proportion of the variance for KD efficiency to be affected by a specific MolDes. We noted that among the top 5 MolDes with the highest coefficient of determination, 3 of them (TDB10p, E3m and H2p) belonged to 3D MolDes^27,35,36, with TDB10p exhibited the highest r² value of 0.93 (Fig. 4i–j). These MolDes typically encoded both the 3D-geometrical information and chemical information with different weighting scheme such as polarizability (TDB10p, H2p) and mass (E3m) from the atoms. The detailed description of each MolDes is shown in Supplementary Table 7. Previous studies demonstrated the transfection efficiency from different cationic polyelectrolytes, even with the identical chemical compositions, were largely affected by their 3D orientation^37,38. Thus, conventional MolDes with clear chemical or physical attributes, despite their better interpretability, may not be enough for understanding the divergent phenomenon observed during the experiments³⁹, as partly confirmed by the unsatisfied r² value from MolDes describing hydrophobicity, polar surface area and positive charges (Supplementary Fig. 5e, Supplementary Table 7). Therefore, the current results highlighted the potential importance of 3D geometry of a cationic compound in affecting the transfection efficiency of the fabricated nanosystem.

Identification and characterization of GluCARDIA NPs

With a systematic evaluation of key parameters for NPs formation and siRNA delivery, we identified pairing NPs between EEPG and KALA to form siRNA loaded NPs. Considering the precise Dectin-1 targeting efficiency from EEPG⁶, along with the abnormally activated expression of Dectin-1 in injured myocardium upon myocardial ischemic/reperfusion (IR) injury⁴⁰, as a proof of concept, we determined to test the therapeutic potency of identified KALA-EEPG NPs for ameliorating cardiac reperfusion injuries, and the NPs were further termed as β-Glucan based Combinatory Assessed delivery system for Reperfusion Damage Inhibition and Aid (GluCARDIA) NPs. Prior to this, the morphology of bare GluCARDIA, as well as siRNA-loaded GluCARDIA NPs was analyzed by transmission electron microscopy (TEM, Fig. 5a, b). TEM images show that both of them showed a homogenous size and morphology, which is in good accordance with the physicochemical parameters evaluated by DLS (Fig. 5c). The mean size of GluCARDIA-siRNA NPs was slightly increased from 132 ± 0.6 nm to 137 ± 1.6 nm due to the incorporation of siRNA, whereas siRNA incorporated GluCARDIA NPs still exhibited a narrow polydispersity index (PDI) of 0.12.

**Fig. 5: Characterization and in vitro evaluation of GluCARDIA NPs.**

Cytosolic delivery of siRNA

Next, we investigated the GluCARDIA NPs cytosolic delivery of siRNA, which is critical for RNAi machinery function^6,41. To this end, the fluorescence dye labelled siRNA (Cy5-siRNA) encapsulated GluCARDIA NPs (GluCARDIA-Cy5-siRNA NPs) were incubated with THP-1 cells for 1 h, 2 h and 6 h (Fig. 5d), and lysotracker was co-stained to unravel the acidic endosomes/lysosomes intracellular compartments during endocytosis. After 1 h incubation, GluCARDIA-Cy5-siRNA NPs was observed, whereas a major colocalization between NPs and lysotracker was measured, indicating their major entrapment in endosomes/lysosomes organelles. Followed by increasing the incubation time, a gradual disassociation between Cy5-siRNA and lysotracker was observed in a time-dependent manner, suggesting the efficient endosome escape of siRNA facilitated by GluCARDIA NPs. This process was further quantified by evaluating the Manders’ colocalization coefficient (MCC) between lysotracker and GluCARDIA-Cy5-siRNA NPs⁴², where the MCC was gradually decreased from 0.6 to 0.2, suggesting the time-dependent endosome escape process mediated by GluCARDIA NPs (Fig. 5e). Our previous studies confirmed that EEPG with a pH-dependent membrane-lytic capability, upon acidification, EEPG can destabilize the membrane and facilitate siRNA delivery into cytosol during the maturation process of endosomes⁶. Overall, these results indicate the efficient cellular uptake and endosomal escape properties of GluCARIDIA-based siRNA delivery system.

Targeting Dectin-1⁺ macrophages with gene silencing

One major advantage of polysaccharide-based nanosystems is their innate targeting capability towards specific lectins, enabling a targeted delivery without subsequent modifications⁷. β-glucans, the building block for EEPG, are inherent and precise ligands for Dectin-1^40,43. Thus, we sequentially tested the Dectin-1 targeting capability from GluCARDIA NPs. We compared the cellular interaction of GluCARDIA NPs on different macrophage cell lines, including Dectin-1⁺ THP-1 and Dectin-1^– RAW 264.7 macrophages^22,44. First, the cellular uptake of GluCARDIA-Cy5-siRNA NPs in different cell-lines was quantitatively evaluated by flow cytometry analysis. As expected, free Cy5-siRNA barely showed any cellular uptake in both cell lines, and the obvious cellular uptake were only observed when Cy5-siRNA were incorporated in GluCARDIA NPs (Supplementary Fig. 6–7). Moreover, compared to RAW 264.7 cells, THP-1 showed a statistically significant higher cellular uptake towards GluCARDIA as shown by the positive cell percentages differences (Fig. 5f).

We further evaluated whether this targeting efficiency is associated with gene knockdown efficiency in macrophages with different Dectin-1 expression pattern. To this end, siGAPDH with varied dosage (ranging from 10 nM to 100 nM) were loaded into GluCARDIA NPs (GluCARDIA-siGAPDH NPs), followed by incubating with THP-1 or RAW 264.7 cells for 48 h. Cells from different groups were then subjected to western blotting analysis to evaluate the gene knockdown efficiency (Fig. 5g, h). The results showed a superior gene silencing efficacy from GluCARDIA-siGAPDH NPs in THP-1 cells compared to that from RAW264.7 cells, indicating the enhanced genetic cargo delivery into Dectin-1⁺ cells.

To exclude the possibility that the promoted cellular interaction and the concomitantly achieved higher gene silencing efficacy in Dectin-1⁺ macrophages from GluCARDIA NPs were endowed by the addition of KALA, but rather generated by the incorporation of EEPG. As a side-by-side comparison, we compared the gene silencing ability of different siGAPDH encapsulated EEPG-based NPs (namely PEI800-EEPG, PEI25K-EEPG, Lipid 2-EEPG, Lipid 10-EEPG, K9-EEPG, KALA-EEPG) in both THP-1 cells (Fig. 3d) and RAW 264.7 cells (Supplementary Fig. 8). The results suggested under the same siRNA dosage (50 nM), the gene knockdown efficiency from the same NPs was pervasively higher in THP-1 cells comparing to RAW 264.7 cells, confirming the promoted gene silencing was more likely to be correlated to the EEPG but not from the cationic counterparts.

Overall, we confirmed that GluCARDIA NPs have a higher cellular uptake as well as gene knockdown efficiency in Dectin-1⁺ macrophages comparing to their Dectin-1^– counterparts, partially indicating the Dectin-1 correlated gene silencing from GluCARDIA in vitro.

In vivo targeting to cardiac IR

After the in vitro results, we sought to investigate whether this targeting efficiency is also valid in vivo and its potential biological benefits in promoting the cardiac accumulation of GluCARDIA NPs in mice with myocardial reperfusion injury. Prior to this, the biosafety property was evaluated with intravenous (i.v.) administration of GluCARDIA NPs at the dosage of 4.8 mg kg⁻¹ (pre-determined by the potential siRNA dosage of 2.64 μg per mouse). After 48 h, the serum levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), and total bilirubin (TBIL) were quantified to assess the liver function (Fig. 6a–c). Negligible differences were shown within two groups, suggesting that GluCARDIA NPs were well-tolerated without undesirable liver toxicity in short term.

**Fig. 6: Biodistribution and gene silencing efficacy of GluCARDIA NPs in murine cardiac IR model.**

Next, we evaluated the biodistribution of GluCARDIA NPs in a well-established preclinical IR murine model⁸. To this end, indocyanine green labelled GluCARDIA NPs (ICG-GluCARDIA NPs) was synthesized through a previous established protocol for fluorescent imaging purposes⁸. As a comparison, we applied dextran, a type of α-glucan, served as a negative polysaccharide control without Dectin-1 binding affinity⁴⁵. ICG was encapsulated in acetalated dextran NPs according to previous protocol to fabricate ICG-Dextran NPs⁴⁶. Mice were subjected to 50 min of ischemia followed by 24 h of reperfusion, afterwards, ICG-GluCARDIA or ICG-Dextran NPs was i.v. injected for further analysis. Major organs including heart, lung, liver, spleen and kidney were harvested at 8 h post-injection for fluorescence imaging. The cardiac accumulation of different NPs was calculated based on a previously established term, heart-targeting index (HTI, HTI = heart fluorescence emission/liver fluorescence emission)⁴⁷. The results show a clear statistically significant difference in cardiac accumulation between ICG-GluCARDIA NPs and ICG-Dextran NPs (Fig. 6d), with a significant higher HTI observed from mice treated with ICG-GluCARDIA (Fig. 6e), suggesting the feasibility of adopting GluCARDIA NPs for promoted cardiac accumulation post-IR.

Despite the substantial lower cardiac accumulation comparing to hepatic accumulation, we further determined to evaluate its potential impact on gene silencing, and the knockdown efficiency was evaluated by western blotting analysis (Supplementary Fig. 9). Results showed that while the overall accumulation of GluCARDIA NPs in cardiac tissue is significantly lower than that in liver, the gene knockdown efficiency was still comparable to that observed in hepatic tissue. This might be explained by the “siRNA saturation” effect as described by Lieberman et al., as only marginal amount of the siRNA (~ 10^-9 pmol) is needed to achieve a maximal gene knockdown efficiency, and further increasing of the siRNA showed no significant difference in terms of gene knockdown yield⁴⁸. Therefore, despite the relative cardiac accumulation of GluCARDIA is still substantially lower than that in liver, the currently developed nanosystem can significantly promote the cardiac accumulation and sequentially achieve the envisioned design. Meanwhile, RNA-seq suggested the administration of GluCARDIA-siIRF3 exhibited minimal impact on major liver function in the context of myocardial IR, indicative of a satisfied biosafety property for the current therapeutic strategy (Supplementary Fig. 10, Supplementary Data 3).

To further identify the potential association between GluCARDIA NPs with other types of immune cells in injured myocardium, we conducted double-immunofluorescence staining between FITC labelled GluCARDIA NPs (FITC-GluCARDIA) and Dectin-1 (Figs. 6f) or Ly-6G (neutrophil marker, Fig. 6g) in the mice underwent IR injury^6,8. The results show a major co-localization between Dectin-1 and FITC-GluCARDIA NPs, with negligible correlation observed from Ly-6G. Further quantitative assessment using MCC calculation revealed a value of 0.96 between GluCARDIA and Dectin-1, which was significantly higher compared to that with Ly6G (0.38, Fig. 6h).

A more comprehensive investigation on the potential interaction between the administrated NPs with different immune cells were evaluated by flow cytometry analysis. Three groups were subjected to FACS analysis: fluorescein isothiocyanate (FITC)-labelled GluCARDIA NPs; FITC-labelled Dextran NPs (utilized as a negative control due to its absence of Dectin-1 targeting affinity); and FITC-labelled SM-102 LNPs (utilized as a comparative standard for commercially available NPs in gene delivery). Mice were subjected to 50 min of ischemia followed by 24 h of reperfusion, afterwards, FITC-labelled GluCARDIA NPs/Dextran NPs/SM-102 LNPs was i.v. injected for further analysis. 8 h post-injection, organs including heart and liver were harvested for flow cytometry analysis (Supplementary Fig. 11a). In the lesioned myocardium, GluCARDIA NPs mainly co-localized with monocytic cells, which was substantially higher comparing to Dextran NPs and SM-102 LNPs (Supplementary Fig. 11b). Notably, compared with Dextran NPs and SM-102 LNPs, GluCARDIA NPs also exhibited an increased recognition by dendritic cells, which are also known as major Dectin-1⁺ cellular subsets upon sterile inflammation⁴⁹, while exhibiting a comparable cellular uptake with Dectin-1^– cells such as neutrophils⁴⁰.

Regarding to the potential cellular interaction with hepatic immune cells, despite the major colocalization was still observed in monocytic cells, in contrary to the enhanced monocytic targeting efficiency in lesioned myocardium, GluCARDIA NPs exhibited a similar monocytic internalization efficiency comparing to SM-102 LNPs (Supplementary Fig. 11c). This may be attributed to the restrained Dectin-1 expression profile in hepatic macrophages upon sterile injury⁶. In conclusion, here we have demonstrated that in the setting of myocardial IR, our designed GluCARDIA NPs are primarily recognized by monocytic cells, significantly exceeding its uptake in other immune cells. Moreover, GluCARDIA NPs exhibited substantially higher targeting efficiency towards cardiac monocytes comparing to α-polysaccharides (dextran) based NPs and commercialized benchmark NPs (SM-102 LNPs), offering a promise delivery system for efficient cardiac gene delivery post IR.

Gene silencing after cardiac IR

We further aimed to construct GluCARDIA NPs loaded with functional siRNA for potential cardiac therapeutic interventions. Previous studies indicated cardiac injury provoked interferon regulatory factor 3 (IRF3) activation in recruited macrophages in infarct myocardium, potentially leading to adverse ventricular remodeling and poor prognosis⁵⁰, therefore, we hypothesized that siIRF3-loaded GluCARDIA NPs (GluCARDIA-siIRF3) yield a beneficial effect for cardiac protection after IR. Mice were subjected to 50 min of ischemia followed by different period of reperfusion, and GluCARDIA loaded with non-targeting siRNA controls (GluCARDIA-siNC), free siIRF3 and GluCARDIA-siIRF3 were i.v. injected at 30 min, 24 h and 72 h post-reperfusion at the siRNA dosage of 0.1 mg kg⁻¹. We evaluated the protein expression level of IRF3 and phosphorylated IRF-3 (p-IRF3) after therapeutic intervention. To this end, injured left ventricular (LV) tissues were harvested at day 4 post-reperfusion, and then subjected to western blotting analysis. We detected marked decrease of IRF3 and p-IRF3 (Fig. 6i–k) level in LV tissue solely from GluCARDIA-siIRF3, and this phenomenon was not observed from GluCARDIA-siNC or free siIRF3 administration, indicative to a potent in vivo gene silencing efficacy from GluCARDIA NPs.

Protection of heart after myocardial IR

With identifying the satisfied genetic silencing efficacy in vivo, we finally sought to determine the functional significance of GluCARDIA-siIRF3. To this end, mice underwent same IR scenario were i.v. injected with GluCARDIA-siNC and GluCARDIA-siIRF3 at 30 min, 24 h and 72 h post-reperfusion at the siRNA dosage of 0.1 mg kg⁻¹ (Fig. 7a). The 24 h post the reperfusion, a reduced terminal-deoxynucleotidyl transferase mediated nick end labeling (TUNEL) cells were observed from GluCARDIA-siIRF3 group, indicative of a potential cardioprotective effects of the current treatment (Supplementary Fig. 13). Furthermore, we assessed the transcription levels of key regulatory genes in each group, including TGF-β, interferon-β, and caspase-3 to reflect the fibrosis, immune responses, and cellular apoptosis status following myocardial IR. Comparing to GluCARDIA-siNC or IR group, marked decreases in transcriptional levels of these genes were noted in the GluCARDIA-siIRF3 group, underscoring the potential beneficial effects of GluCARDIA-siIRF3 treatment following myocardial IR (Supplementary Fig. 14). At day 7 post-reperfusion, RNA-seq analysis was conducted on IR, GluCARDIA-siNC and GluCARDIA-siIRF3 groups to collect a genome-wide expression profiling (Supplementary Data 1). Comparing to IR group, Gene Set Enrichment Analysis (GSEA) identified significant negative association between the GluCARDIA-siIRF3 treatment and signaling pathways related to transforming growth factor β (TGF-β) signaling pathway (Fig. 7b) and dilated cardiomyopathy (Supplementary Fig. 12a)⁵¹, indicating a potentially better prognosis regarding to LV remodeling. Similarly, the GluCARDIA-siIRF3 treatment, comparing to GluCARDIA-siNC group, also inhibited fibrosis progress, promoted cardiac contraction and facilitated the inflammation resolution (Supplementary Fig. 12b–e)⁵², suggesting the potentially beneficial effects from IRF3 silencing at the initial stages of myocardial IR.

**Fig. 7: GluCARDIA-siIRF3 NPs ameliorated cardiac IR prognosis.**

Echocardiograph (ECG) was also conducted to evaluate the cardiac function from different groups. Follow-up with 7 days of reperfusion, the left ventricular ejection fraction (EF%) improved to 36% with GluCARDIA-siIRF3 treatment, comparing to an EF of 29% from IR group, while GluCARDIA-siNC (29%) showed no beneficial impact on cardiac functions. However, at day 7 post-reperfusion, no statistical difference was observed in regarding to LV end-systolic internal dimension (LVIDs) and LV end-diastolic internal dimension (LVIDd, Fig. 7c, d). Sequentially, the preserved cardiac function by GluCARDIA-siIRF3 was still valid at day 28 post-IR, as reflected by the substantially higher EF comparing to IR group. Meanwhile, GluCARDIA-siIRF3 also ameliorated the LV remodeling as reflected by the significantly reduced LVIDs (Fig. 7e, f). Masson’s trichrome staining of heart section from each group was also conducted at day 28 post-IR (Fig.7g), demonstrating a preserved wall thickness (Fig. 7h) and lower percentage of fibrosis in the left ventricle (Fig. 7i) in mice treated with GluCARDIA-siIRF3. Adverse remodeling post myocardial IR injury can lead to cardiomyocytes hypertrophy⁵³, we observed that mice treated with GluCARDIA-siIRF3 showed considerably decreased myocyte cross-sectional area in comparison to GluCARDIA-siNC or IR (Fig. 7j).

Considering the major accumulation of NPs in liver, we then checked the potential liver damage from different groups. RNA-seq suggested the administration of GluCARDIA-siIRF3 exhibited minimal impact on major liver function in the context of myocardial IR, indicative of a satisfied biosafety property for the current therapeutic strategy (Supplementary Fig. 10). Furthermore, with four weeks follow-up, liver fibrosis status was evaluated by sirius red staining of liver samples from different groups at end points, and we did not observe significant fibrosis in harvested liver tissues from afore-mentioned three groups (Supplementary Fig. 15), suggesting the GluCARDIA was not overtly toxic when administered to mice over long term.

SEO Powered Content & PR Distribution. Get Amplified Today.
PlatoData.Network Vertical Generative Ai. Empower Yourself. Access Here.
PlatoAiStream. Web3 Intelligence. Knowledge Amplified. Access Here.
PlatoESG. Carbon, CleanTech, Energy, Environment, Solar, Waste Management. Access Here.
PlatoHealth. Biotech and Clinical Trials Intelligence. Access Here.
Source: https://www.nature.com/articles/s41467-024-49804-x

Alkaloid extract of seed Citrullus colocynthis as novel green inhibitor for mild steel corrosion in one molar HCl acid solution: DFT and MC/MD approaches – Scientific Reports

July 22, 2024 8:00 pm

Publisher Correction: Oral mitochondrial transplantation using nanomotors to treat ischaemic heart disease – Nature Nanotechnology

July 22, 2024 8:00 pm

Nima Farzan named CEO of Latigo

July 22, 2024 7:56 pm

Comparative optimization of polysaccharide-based nanoformulations for cardiac RNAi therapy – Nature Communications

A robotic assisted microfluidics platform for nanoparticles formation screening