Systemic delivery of cationic liposome-mediated siRNA EGFR enhances therapeutic efficacy in a human colorectal cancer model

1 Systemic delivery of cationic liposome-mediated siRNA EGFR enhances therapeutic 1 efficacy in a human colorectal cancer model 2 3 Damian Kaniowski1,2*, Anna Boguszewska-Czubara3, Katarzyna Ebenryter-Olbińska2, Katarzyna Kulik2, Justyna 4 Suwara2, Artur Wnorowski 1,,4, Jakub Wójcik 1, Barbara Budzyńska 6, Agnieszka Michalak 6, Algirdas Ziogas 1,7, 5 Barbara Nawrot1,2, Olga Swiech1,5* 6 1 Biotechna S.A., Dobrzanskiego 3, 20-262 Lublin, Poland 7 2 Centre of Molecular a nd Macromolecular Studies, Polish Academy of Sciences ( CMMS PAS), Sienkiewicza 8 112, 90-363 Lodz, Poland 9 3 Department of Medical Chemistry, Medical University of Lublin, Chodzki 4A, 20-093 Lublin, Poland 10 4 Department of Biopharmacy, Medical University of Lublin, 20-093 Lublin, Poland 11 5 Faculty of Chemistry, University of Warsaw, Pasteura 1, 02093 Warsaw, Poland 12 6 Independent Laboratory of Behavioral Studies, Medical University of Lublin, 20-093 Lublin, Poland 13 7 Faculty of Fundamental Sciences, Vilnius Tech University, Vilnius, Lithuania 14 *correspondence: [email protected], [email protected] 15 16 Key words: cationic lipids; control release; therapeutic nucleic acids; siRNA, EGFR; targeted delivery; 17 colorectal cancer 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 2 Abstract 34 The clinical translation of RNA interference (RNAi) therapeutics remains limited by i nefficient 35 delivery and cancer -target accumulation . Here, we report the development of a new cationic 36 liposome (CLP) nanocarrier engineered for delivery and controlled-release of small interfering RNA 37 (siRNA) targeting the epidermal growth factor receptor (EGFR) in human colorectal cancer. CLPs 38 were synthesized from ethylphosphocholine-based lipids and PEGylated components, with folic acid 39 (FA) tissue -specific ligand and fluorophore labe lling. These nanocarriers exhibited robust 40 physicochemical stability across a broad pH and temperature range, efficient siRNA complexation, 41 and nuclease-protection of siRNA. Functional studies revealed that CLP -siEGFR achieved effective 42 cytosolic siRNA cargo release and EGFR silencing in vitro, proving to be more effective th an 43 conventional lipid -based transfection systems. In human xenograft models, intravenously 44 administered CLP -siEGFR showed enhanced tumor localization, prolonged siRNA retention, and 45 significant tumor growth suppression, accompanied by marked downregulation of EGFR. 46 Importantly, systemic dosing was well-tolerated, with no evidence of hepatotoxicity, nephrotoxicity, 47 or hematological abnormalities. These results position CLP nanocarriers as an effective platform for 48 targeted RNAi therapeutics, offering translational potential for precision oncology applications. 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 3 1. Introduction 70 Colorectal cancer (CRC) is the third most diagnosed malignancy worldwide, with projections 71 estimating an alarming rise to 3.2 million new cases and 1.6 million fatalities by 2040 .1 Notably, up 72 to 25% of patients with CRC are diagnosed with stage I V disease, while up to 50% of those initially 73 presenting with early-stage CRC progress to metastatic disease over time. The prognosis for stage 74 IV CRC remains dismal, with a 5 -year survival rate of only 12.5% in the United States and 15.4% in 75 Europe.2,3 These statistics highlight a critical unmet clinical need for therapeutic strategies that 76 combine high drug delivery efficacy with a favorable safety profile to effectively address the growing 77 burden of colorectal cancer.4 78 Epidermal growth factor receptor (EGFR) aberrant in colorectal cancer is predominantly driven by 79 protein overexpression, which is observed in approximately 35% to 50% of patients. Numerous 80 studies have established a statistically significant c orrelation between EGFR overexpression and 81 adverse prognostic outcomes .5 The EGFR signalling pathway is integral to cancer progression, 82 leading to the development and clinical introduction of several targeted therapies, including 83 cetuximab and panitumumab.6 These monoclonal antibodies competitively bind to the extracellular 84 domain of EGFR , preventing receptor tyrosine kinase phosphory lation and activation. 7 This 85 mechanism suppresses cell growth, induces apoptosis, and promotes cell cycle arrest . Moreover, 86 EGFR activation can promote tumor growth by promoting VEGF upregulation through a 87 hypoxia‑independent mechanism. 8 Additionally, cetuximab can stimulate antibody -dependent 88 cellular cytotoxicity, contributing to immunogenic cell death and enhancing the immune -mediated 89 effects of therapy. 8,9 These targeted therapies have demonstrated substantial efficacy in s elected 90 subsets of CRC patients, significantly slowing disease progression .10,11 Comprehensive preclinical 91 research on the mechanisms of resistance to anti-EGFR monoclonal antibodies has underscored the 92 critical need for the development of new and EGFR-targeted therapies.12 93 Despite advances in the development and clinical application of 1 st, 2 nd and 3 rd generation EGFR 94 tyrosine kinase inhibitors (TKIs), drug resista nce remains a significant challenge. Clinical trials have 95 highlighted issues such as inadequate tumor accumulation, limited penetration into solid tumor 96 masses, and impediments posed by the tumor microenvironment.13 97 To overcome current limitations in clinical translation, therapeutic nucleic acids represent a 98 powerful strategy for targeted therapy. Small interfering RNAs (siRNAs) are short double -stranded 99 RNA molecules, typically 19 -25 base pairs in length, capable of integrating into the RNA -induced 100 silencing complex (RISC) to induce sequence -specific silencing of target mRNA.14 To date, several 101 siRNA-based therapeutics have received FDA approval for clinical use, targeting gene expression in 102 rare metabolic and cardiovascular disorders. These include Patisiran (Onpattro) for hereditary 103 transthyretin-mediated amyloidosis (hATTR); Givosiran (Givlaari) for acute hepatic porphyria (AHP); 104 Lumasiran (Oxlumo) and Nedosiran (Rivfloza) for primary hyperoxaluria type 1 (PH1); Inclisiran 105 (Leqvio) for lowering LDL cholesterol in patients with atherosclerotic cardiovascular disease 106 (ASCVD); and Vutrisiran (Amvuttra), also for hATTR and Fitusiran (Qfitlia) for hemophilia.15,16 These 107 therapies utilize either lipid nanoparticle or N-acetylgalactosamine ligand ( GalNAc)-conjugate 108 delivery systems to achieve targeted, durable gene silencing via RNAi.17 Despite siRNA therapeutic 109 potential, several significant challenges persist for the in vivo application of siRNA, including off-110 target effects, inefficient delivery with poor cellular uptake, and activation of immune responses, all 111 of which hinder their clinical feasibility.18 These obstacles highlight the need for more efficient and 112 targeted delivery systems to enhance the effectiveness of siRNA -based therapies. Various delivery 113 technologies and strategies, including lipophilic conjugates, cationic polyme r- or micelle-based 114 carriers, protein -antibody conjugates, and modifications to the size and structure of the siRNA 115 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 4 molecule itself, are under development. These approaches have demonstrated robust efficacy in 116 different animal models, with some platforms a dvancing toward clinical translation .14,19,20,21 Lipid 117 nanoparticle delivery was the first successful strategy for gene silencing in human and enabled the 118 approval of the first siRNA-based therapeutic.22,23 119 Despite clinical progress, the use of RNAi nanomedicines in the treatment of solid tumors remains 120 limited by low tumor accumulation and inefficient intracellular transport 24. Quantitative meta -121 analyses indicate that on average only about 0.7% of the injected dose (ID) of nanoparticles 122 accumulates in solid tumors, with a median of 0.6% ID, highlighting limitations related to 123 heterogeneous vascular permeability and clearance by the mononuclear phagocyte system .25,26 A 124 second, well-recognized bottleneck is endosomal entrapment. Image-based single-cell tracking has 125 shown that only about 1 -2% of inter nalized siRNA molecules escape from endosomes into the 126 cytosol, with the vast majority remaining trapped and ultimately degraded in endo -lysosomal 127 compartments.27 128 Polyethylene glycol (PEG) -lipids are widely used to minimize aggregation and opsonization; 129 however, they introduce the well-known ‘PEG dilemma’. While PEG prolongs systemic circulation, 130 excessive surface shielding can impede cellular uptake and endosomal escape. Furthermore, 131 repeated administration may lead to the accelerated blood clearance (ABC) phenomenon .28,29 132 Beyond PEG density, the terminal-group chemistry of PEG is emerging as a tunable parameter that 133 can modulate interfacial charge, protein corona composition, and tumor retenti on. Evidence from 134 reviews and studies on targeted liposomes indicates that functional PEG end-groups (e.g., folate or 135 thiol-reactive moieties) can enhance tumor cell interaction compared to inert methoxy-PEG.30,31 136 Building upon this rat ionale, we employed our proprietary PEG-functionalized cationic liposome 137 (CLP) platform (WIPO WO2023022615 ). Incorporation of PEG -NH₂ lipids combined steric 138 stabilization with enhanced electrostatic association of siRNA, preserving a positive surface charge 139 without inducing aggregation and enabling release modulation through surface chemistry . This 140 system enabled selective accumulatio n of siRNA inhibitor within colorectal tumor tissue, resulting 141 in efficient EGFR gene silencing and significant tumor growth inhibition. The CLP formulation was 142 well tolerated in vivo, with no evidence of systemic toxicity. These findings support the new CLP 143 platform as a promising candidate for further preclinical translation in RNAi -based precision 144 oncology and clinical translation. 145 146 147 148 149 150 151 152 153 154 155 156 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 5 2. Results 157 158 2.1. Design and optimization of cationic liposomes for RNAi-based applications 159 Cationic liposomes (CLP) developed in thi s study were formulated using saturated cationic 160 ethylphosphocholines (EPC) in combination with saturated neutral lipids (DPPC and DMPC) and 161 polyethylene glycol (PEG)-modified lipids bearing a terminal primary amine group. EPCs are cationic 162 derivatives of phosphatidylcholine, where the phosphate oxygen's negative charge is neutralized by 163 the addition of an ethyl group. This modification results in a compound that is chemically stable, 164 biodegradable, and composed entirely of biological metabolites linked by ester bonds. The CLP were 165 synthesized via the conventional thin -film hydration method, followed by extrusion to ensure 166 uniform size distribution (Figure 1A). This approach enabled the preparation of liposomes with a 167 diverse range of lipid compositions, var ying in charge, fatty acid chain length, and surface 168 modifications with either a targeting folic acid-ligand (FA) or a Cyanine 7 dye (Cy7). For this study, 169 cationic liposomes CLP and folic acid -modified cationic liposomes (CLP-FA) were selected as 170 representative formulations, incorporating EPC lipids and PEG -modified lipids with terminal amine 171 functionality (Figure 1B). These formulations were developed and described in our patent 172 applications (PCT.PL2022.000046; WO 2023/022615 A1). The synthesi zed CLPs, CLPs-FA, and their 173 Cy7-labeled counterparts exhibited small hydrodynamic sizes and low polydispersity indices (PDI), 174 indicative of a monodisperse population. Notably, the incorporation of DSPE-PEG(2000)-folate lipid 175 and 1,2-distearoyl-sn-glycero-3-phosphoethanolamine conjugated with Cyanine 7 (18:0 Cy7-PE) 176 lipid did not induce significant alterations in either size or PDI. 177 To assess the stability of the liposomal formulations, temperature profile analysis and 178 physicochemical evaluations across a p H range of 4.0 to 8.0 were conducted (Figure 1C). Both CLP 179 CLP-FA liposomes demonstrated stability across a wide range of pH values and temperatures (Figure 180 1D, 1E). Stability assessments revealed that the hydrodynamic sizes of CLP and C LP-FA remained 181 constant, with PDI values consistently belo w 0.1, in PBS buffers ranging from pH 4.0 to 8.0 (Figure 182 1D, 1E, respectively). Furthermore, these parameters remained unchanged at room temperature 183 (25°C) and after one week of incubation at 37°C, underscoring the structural integrity of the 184 liposomes under physiological conditions relevant for intravenous administration. The thermal 185 stability of the liposomes was further examined across a temperature range of 4°C to 44°C (Figure 186 1F). Throughout this range, PDI values remained below 0.1, and no significant size variations were 187 observed, suggesting the absence of lipid bilayer phase transitions in the selected lipid composition. 188 These findings confirm that CLP and CLP-FA maintain their structural integrity under physiologically 189 relevant conditions. Long-term stability studies have shown that both CLP and CLP-FA remain stable 190 under long-term storage conditions (Figure S1 A-C). 191 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 6 192 193 194 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 7 Figure 1. Engineering the composition of cationic liposomes for the delive ry of siRNA. Schematic 195 representation of the synthetic procedure for CLPs liposomes ( A); a pie chart showing the percentage 196 composition of individual lipids ( B); stability studies (C); changes in hydrodynam ic radius (size, empty bars) 197 and PDI (filled bars) of CLP and CLP-FA in electrolytes with varying pH levels (pH 4 .0 to 8.0) at 25°C or after 198 incubation at 37°C for up to 168h (7 days) (D, E); the temperature profile of liposomes across a range of 4°C 199 to 44°C. The size (upper empty rectangles and circles) and PDI (lower filled rectangles and circles) (F). 200 201 2.2. RNAi-based therapeutic targeting of EGFR in colorectal cancer 202 Kaplan-Meier survival analysis demonstrated that colon cancer patients with high EGFR expression 203 exhibited significantly poorer prognosis compared to those with low EGFR expression, highlighting 204 the clinical relevance of targeting this receptor as a potential therapeutic strategy (Figure 2A). The 205 extracellular domain of EGFR is currently targeted by monoclonal antibodies (e.g., cetuximab) in 206 clinical settings; however, these therapies have shown limited efficacy i n certain patient cohorts.32 207 To address this limitation, we designed a small interfering RNA (siRNA) specifically targeting the 208 mRNA of EGFR within the sequence encoding domain III (L2 ) of EGF R, which is critical for ligand 209 (EGF) binding (Figure 2B). To evaluate efficacy , siRNA EGFR (siEGFR) inhibitors were chemically 210 unmodified or modified with 2′-O-methyl (siEGFR2′-OMe) at the ribose 2′-position, and with or without 211 additional terminal stabil ization using phosphorothioate -modified ( siEGFRPS) internucleotide 212 linkages (Figure 2C). All sequences and chemical modification patterns of the oligonucleotides used 213 are presented in Figure 2D. siRNA compounds were purified from reaction mixtures using standard 214 HPLC procedures and characterized by electrospray ionization quadrupole time -of-flight mass 215 spectrometry (ESI-Q-TOF MS) (Figure S2). 216 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 8 217 218 219 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 9 220 Figure 2. Design and characterization of siRNA targeting EGFR ligand-binding domain. Kaplan-Meier curves 221 for overall survival of human colon cancer (GDC TCGA COAD) according to human EGFR categorized by gene 222 expression RNAseq using UCSC Xena database. Survival time is presented in days; p values are related to Log-223 rank test results ( A), homodimer structure of human EGFR based on UniProt: P00533 with EGF binding 224 domain III. Screened siEGFR targets mRNA within the coding region of this domain (nucleotides mRNA 321-225 334; red) (B); Chemical modification patterns abbreviations: 2'-O-methyl (2' -OMe) - navy blue , ‘_’ - 226 phosphorothioate bonds (PS) - red, phosphodiester bonds (PO) - black, ribonucleic acid (RNA) – blue, 6-227 carboxyfluorescein dye (FAM) and Cyanine5 dye (Cy5) (C) and table of all RNA sense (s) and antisense (as) 228 sequences (D). 229 230 231 2.3. Cationic liposome-mediated siRNA delivery overcomes the limitations of endosomal 232 escape 233 CLP have long been recognized as highly effective gene delivery v ehicles, demonstrating efficient 234 biodegradability, biocompatibility, and high nucle ic acid encapsulation rate .33,24 Zeta po tential 235 analysis provided further mechanistic insights into siRNA-lipid interactions (Figure 3A). The surface 236 charge of CLP nanoparticles was monitored following the addition of either unmodified siEGFR 237 (more hydrophilic) or modified siEGFR 2’-OMe (more hydrophobic, Figure 3B). Pristine CLPs exhibited 238 a zeta potential of +66.7 mV. Upon complexation with siRNA at increasing siRNA:CLP mass ratios 239 (1:100, 1:50, 1:20, 1:10), a progressive reduction in zeta potential was observed for both siRNA 240 variants. Notably, the decrease was consistently less pronounced for 2′-O-methyl modified siRNA, 241 ranging from ~8% at a 1:100 ratio (+58.2 mV vs. +63.9 mV for unmodified and modified siRNA, 242 respectively) to ~25% at a 1:10 ratio (+28.9 mV vs. +39.0 mV, respectively) in Figure 2C. PDI analysis 243 confirmed the colloidal stability of the formulations. Both unmodified and modified siRNA induced 244 only minor increases in PDI (0.07 to 0.19), excluding aggregation and supporting the interpretation 245 that siRNA is incorporated within the CLP interior and outer PEG layer without extensive 246 interparticle bridging (Figure 3C). Comparable zeta potential trends were observed for folate -247 modified CLPs (Figure S1 A -C), indicating that siRNA binding follows a conserved mechanism 248 independent of surface functionalization. 249 Quantitative binding assays using 5′-[32P]-labeled siRNA and CLPs were performed to determine the 250 stoichiometry of siRNA -CLP interactions (Figure 3D) . Subsequently, the binding of negatively 251 charged [32P]-siRNA to positively charge CLP was evaluated using PAGE analysis in both directions -252 from the negative to the positive terminal to visualize the single RNA or siRNA mobility and in the 253 reverse direction to observe the complex (Figure 3D) . The results indicated that the [32P]-siRNA 254 exhibited optimal binding to CLP at a charge mass ratio of 1 :10 (Figure 3 E). The commercial 255 transfection reagent Lipofectamine 2000 was used as a positive control (Figure 3E). 256 Endosomal escape refers to the crucial process by which nucleic acid -loaded nanoparticles are 257 released from endosomes-acidic compartments with a pH ranging from 5.5 to 6.5 -into the cytosol, 258 which maintains a neutral pH of approximately 7.4.34 siRNA release from the CLP-siRNA formulation 259 was assessed using acidified solutions that closely mimic the conditions of endosomal -lysosomal 260 pathway (Figure 3F ). Our results demonst rated that the CLP platform efficiently releases siRNA 261 within the endosomal pH range of 5.5 to 6.5 after 30 min of incubation (Figure 3G). Moreover, the 262 CLP-siRNA complex remains stable at pH 4.0 and 7.4. Next, we evaluated the binding efficiency of 263 chemically modified siRNAs (2′ -OMe, PS, and 2′ -OMe/PS) to the cationic folic acid -liposome (CLP-264 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 10 FA) platform and found comparable binding affinity to tha t of unmodified siRNA EGFR (Figure 3H). 265 Heat map analysis revealed that both modified and unmodified siRNAs are re leased from CLP or 266 CLP-FA within the pH range of 5.5 to 6.0, but n ot at other pH levels (Figure 3I). Interestingly, the 267 release duration of siRNA was extended to 4-hours in the CLP-FA platform compared to unmodified 268 CLP (Figure 3I). To confirm that CLP doe s not degrade unmodified siRNA across sequences (EGFR 269 and scrambled), stability analyses were conducted after 7 days of incubation at 4°C and 37°C using 270 PAGE under release conditions ( pH 5.5). As shown in Figure 3J, the CLP platform supports siRNA 271 stability over one week across temperatures. 272 273 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 11 274 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 12 275 Figure 3. siRNA-loaded cationic liposomes facilitate efficient endosomal e scape-associated intracellular 276 delivery. Scheme of the study on siRNA binding to CLP cationic liposomes and readouts (A); Effect of siRNA 277 loading at different ratios on the zeta potential of CLP (B); hydrodynamic radius (size, empty bars) and PDI 278 (filled bars) after siRNAs addition on the CLP ( C); Schematic showing binding of 5′ -[³²P]-labeled siRNA to 279 cationic CLPs, analyzed by PAGE in both directions (D); 5′-[32P]-labeled siRNA binding to CLP or Lipofectamine 280 2000 (L2000) at various mass ratios (w:w), sEGFR and asEGFR are single -stranded siRNA c omponents (E); 281 Schematic of pH-dependent siRNA release from CLP nanocarriers (lanes 2, 4, 6 and 8) and control siRNA (lanes 282 1, 3, 5, 7) across physiological and endosomal conditions (pH 7.4, blue; 6.0, green; 5.5, orange; 4.0, red) (F, 283 G); Binding efficiency of chemically modified (2′-OMe, PS, and 2′-OMe/PS) and subsequently 5′-[32P]-labeled 284 siRNAs to CLP-FA after 30 min incubation (H); Heat Map of siRNA release from CLP and CLP-FA at pH 4.0-7.4 285 over time were calculated by densitometry (I); CLP-siRNA complex stability was evaluated after 7 days of 286 incubation at 4 °C and 37 °C (lanes 1-4) at pH 7.4 and under controlled-release conditions at pH 5.5 (lanes 2 287 and 4) (J); Analysis were performed using 15% PAGE under non-denaturing conditions. 288 289 2.4. CLP-mediated EGFR siRNA delivery effectively reduces colorectal cancer growth 290 To evaluate uptake of the siRNA-CLP formulation by human colorectal cancer Caco -2 cells , we 291 generated a fluorescently labeled EGFR -targeting siRNA (CLP-FA-siEGFRFAM). Prior to experiments, 292 we assessed binding of siEGFRFAM to CLPs and confirmed stable complex formation for up to 24 h 293 (Figure S3B). Fluorescence microscopy confirmed the efficient and selective uptake of CLP -FA-294 siEGFRFAM by Caco-2 cells in vitro after 30 minutes of incubation, i n contrast to normal mouse 295 fibroblasts (MEF-WT) (Figure 4B). The control experiment include d transfection of siEGFR FAM using 296 Lipofectamine demonstrated comparable uptake of labeled siRNA (Figure S3A). Comprehensive 297 evaluation of Cy5-labeled CLP-siEGFR2′-OMe demonstrated concentration-dependent uptake in Caco-298 2 cells (Fig ure 4C) and confirm ed complete carrier loading (Figure S3B). Both CLP and CLP-FA 299 platforms without siRNA cargo showed no cytotoxicity after 48 h of incubation in Caco -2 cells, 300 normal human co lon epi thelial cells (CCD -841CoN) (Figure 4D) and macrophages (Fig ure S3C), 301 relative to the positive control. Next, to evaluate the EGFR gene-silencing activity of CLP-FA-siEGFR, 302 we used a dual fluorescence assay ( DFA) previously developed in our laborator y.35,36,37 We 303 confirmed that the CLP platform efficiently inhibited the EGFR target by approximately 65% using 304 modified si EGFR ( 2’-OMe and PS) and unmodified siRNA in Caco -2 cells , compared to the CLP -305 scrambled negative control (CLP-SCR), as shown in Figure 4F. In contrast, CLP-FA-siEGFR showed 306 slightly reduced potency but still suppressed EGFR expression by 40% in Caco-2 cells relative to CLP-307 SCR (Figure 4G). In both platforms, GFP-targeting siRNA served as a positive control and effectively 308 reduced EGFR -GFP fluorescence in Caco -2 and A431 cells (Figure S3 D-E), while naked panel of 309 siRNAs were transfected using Lip ofectamine (Figure S3 F -G). Furthermore, flow cytometry 310 demonstrated that CLP -siEGFR2′-OMe markedly reduced EGFR expression in colon cancer cells 311 compared to CLP-SCR (Figure 4H). EGFR inhibitor and CLP-ASO served as positive controls38 (Figure 312 4H and S3H, respectively). To further validate targe t-specific inhibition, we us ed EGFR-313 overexpressing A431 cells, confirming the efficacy of CLP-siEGFR2’-OMe (Figure 4I). Finally, treatment 314 with CLP-siEGFR2′-OMe significantly increased apoptosis in Caco-2 cells, resulting in a reduction in cell 315 survival compared to CLP-SCR (Figure 4J-L). 316 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 13 317 318 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 14 319 Figure 4. CPL-siRNA complexes achieve targeted EGFR silencing in colorectal cancer cells in vitro. Uptake of 320 CLP-FA-siEGFR via folate receptor (FR) -mediated endocytosis in cancer cells, and the mechanism of action 321 through the RNA-induced silencing complex (RISC) (A); Intracellular localization of CLP-FA-siEGFRFAM (100 nM) 322 visualized by fluorescence microscopy (×40 magnification) after 30min incubation with Caco-2 and MEF-WT 323 cells. The nucleus was stained with DAPI (blue), and the endoplasmic reticulum (ER) with ER-Tracker Red (B); 324 Delivery of the CLP-Cy5siEGFR2’-OMe to assess concentration -dependent uptake ( C) Cell viability after 48 h 325 incubation with empty CLP and CLP-FA (10–200 µg/mL), compared to the Staurosporine positive control (STS, 326 1 μM) (D); Schematic of the treatment utilizing a DFA tool (pEGFP-EGFR/DsRED) in Caco-2 cells. Cells were 327 co-transfected with pEGFP -EGFR (green) and pDsRED -N1 (red) plasmids using Lipofectamine 200 0, then 328 treated with CLP-delivered siRNAs. After 48h, EGFP-EGFR/DsRED fluorescence was measured and normalized 329 to untreated plasmid -transfected controls (C -Control, set to 100%) 35,36,37 (E); Silencing activity of tested 330 EGFR-targeted siRNAs (10-200 nM) delivered via CLP (F) and CLP -FA (G), in comparison to scrambled (CLP-331 SCR); Inhibition of EGFR expression by CLP-siEGFR2′-OMe (100 nM) in Caco-2 and A431 cells, assessed by flow 332 cytometry after 48 h incubation ( H-I); Representative Annexin V/Aqua staining (apoptosis/necrosis) flow 333 cytometry panels of Caco-2 cells treated with CLP-siEGFR2′-OMe (100 nM) for 48 h (J), with quantified apoptosis 334 (K) and cell survival (L). EGFR inhibitor (PD153035, 1 µM) served as a positive control (H-L). 335 336 2.5. CLP-siEGFR2’-OMe is safe and well tolerated in mice 337 To assess the translation of our CLP-siEGFR2’-OMe in vitro effects into an in vivo setting and to get 338 more insight into the biodistribution and potential off-target toxicity, we monitored the delivery of 339 CLP-siEGFR2’-OMe into the tumors and also different organs. Organ weights of the heart, lungs, liver, 340 kidneys, spleen, and brain showed no significant differences across treatment groups, indicating no 341 apparent organ toxicity (Figure 5 A). Serum biochemical analysis revealed stable levels of liver 342 enzymes (CK, ALAT, ASPAT) and kidney function markers (urea, creatini ne), suggesting preserved 343 hepatic and renal function (Figure 5B). Additionally, hematological parameters-including red blood 344 cells (RBC), white blood cells (WBC), and platelets (PLT) -remained within normal physiological 345 ranges in PBS group (Figure 5C). These findings confirm that systemic administration of CLP-siEGFR²′-346 OMe formulation does not induce detectable toxicity, supporting its safety for further preclinical 347 mouse model. 348 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 15 349 Figure 5 . Assessment of biochemical , safety and hemocompatibility after CLP-siEGFR²′OMe treatment in 350 mice. Mice received seven intravenous administrations (i.v.) of the tested compounds at a dose of 2 mg/kg. 351 Analyses were performed two days after the final dose . Organ weights (heart, lung, liver, kidney, spleen, 352 brain) showed no s ignificant changes across treatment groups ( A); Serum biochemical markers (CK, ALAT, 353 ASPAT, urea, creatinine) remained within normal ranges, indicating no hepatic or renal toxicity ( B); 354 Hematological parameters (RBC, WBC, PLT) were unaffected, supporting o verall safety ( C). Biochemical 355 parameters (CK: 100-1000 U/L; ALAT: 20-80 U/L; ASPAT: 50-200 U/L; UREA: to 2500-5000 µmol/L; creatinine: 356 18-53 µmol/L) and hematological parameters (RBC: 7-11 ×10⁶/µL; WBC: 1-4 ×10³/µL; PLT: 500-1200 ×10³/µL) 357 remained within the normal reference ranges for NOG mice. Data shown as mean ± SD (n = 6; female mice). 358 In vivo experiments were conducted in accordance with GLP standards. 359 360 361 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 16 362 2.6. CLP-mediated delivery enhances the therapeutic efficacy of siRNA in vivo 363 Clearance kinetics from the bloodstream significantly influence the rate and extent of tissue 364 distribution, with higher tissue exposure often correlating with increased tissue accumulation.39 In 365 this context, we sought to evaluate the impact of the CLP platform on the in vivo clearance kinetics 366 of siRNA, with the objective of enhancing delivery to tumor. A 2 mg/kg dose of either CLPCy7 or CLP-367 FACy7 without siRNA 2’-OMe cargo, naked Cy5siRNA2’-OMe and CLP -Cy5siRNA2’-OMe formulation were 368 intravenously administered to NOG mice bearing Caco -2 xenografts . Whole body imaging was 369 performed at multiple time points post-injection using an IVIS bioluminescence system (Figure 6A). 370 Subsequently, we confirmed that the CLP-Cy5siEGFR2′-OMe had enhanced accumulation in tumor 371 xenografts and demonstrated sustained retention over time, as compared to the naked Cy5siEGFR2′-372 OMe, when administered at the same dose (Figure 6B, E, F) and (Figure S4 A, B). Tissue biodistribution 373 studies confirmed efficient delivery of CLP-Cy5siEGFR2′-OMe to the intestine with subsequent clearance 374 through the liver and kidneys (Figure S4 C-G). 375 To determine whether the enhanced tumor accumulation of the CLP-siEGFR2′-OMe formulation 376 translated into functional modulation of target gene expression, mice bearing human colon cancer 377 xenografts were treated intravenously with 2 mg/kg of the tested compounds for 7 consecutive days 378 (Figure 6 G). Tumors harvested two days after the final treatment showed a significant mass 379 reduction in the CLP-siEGFR2′-OMe group compared to the scrambled group (CLP-SCR2′OMe), as shown 380 in Figure 6H and 6I . Furthermore, western blot analysis confirmed effective target gene silencing, 381 demonstrating a marked decrease in EGFR protein levels in tumors treated with CLP-delivered 382 siEGFR2′-OMe relative to the scrambled control (Figure 6 J, K), but not in the liver (Figure S4 H) . In 383 contrast, naked siEGFR2′-OMe demonstrated no measurable impact on EGFR expression or tumor 384 progression. In vivo safety assessment confirmed that new CLP-siEGFR2′-OMe formulation was well 385 tolerated in mice following repeated intravenous administration (2 mg/kg) over 9 days, with no 386 significant changes in body weight (Figure S4I). These findings confirm the efficacy of the CLP 387 platform in mediating targeted siRNA delivery and EGFR gene silencing in a preclinical colon cancer 388 model. 389 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 17 390 391 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 18 392 Figure 6. Enhanced in vivo EGFR gene inhibition via CLP-enabled siRNA delivery in mice. Graphical study 393 outline illustrating the use of bioluminescence imaging (IVIS) (A); Representative IVIS images of mice treated 394 with 2 mg/kg of CLPCy7, CLP-FACy7, Cy5siEGFR2′-OMe, or CLPCy5siEGFR2′-OMe over time ( B); Quantitative IVIS 395 analysis shows sustained signal intensity over time following administration of the tested compounds in 396 tumor-bearing NOG mice (C-F); Schematic diagram of the treatment regimen: intravenous (i.v.) 397 administration in Caco-2 tumor-bearing NOG mice for 7 consecutive days (G); Tumor volume measurements 398 (H) and tumor weight at endpoint ( I); Western blot analysis of EGFR expression and α -tubulin as a loading 399 control (n=3), was performed independently in the Caco-2_NOG model, (J, K). Error bars represent mean ± 400 SD (n=6 per group; female mice). In vivo experiments were conducted in accordance with GLP standards. 401 402 3. Discussion 403 The development of effective delivery systems remains the principal barrier to the clinical 404 translation of RNA interference therapeutics for solid tumors.40 In this study, we demonstrate that 405 PEGylated cationic liposomes (CLPs), with or without folic acid (FA) functionalization, enable 406 efficient delivery of EGFR-targeted siRNA inhibitor to human colorectal cancer, resulting in robust 407 gene silencing and significant tumor growth inhibition while maintaining a favorable safety profile. 408 Our findings show n that CLP composed of saturated ethylphosphocholine derivatives, neutral 409 phosphatidylcholine derivatives, and PEGylated lipids bearing terminal amine groups can be 410 reproducibly synthesized using thin-film hydration and extrusion (Figure 1 A). Both unmodified and 411 folic acid-modified liposomes (FA-CLPs) exhibited uniform size distribution, low polydispersity, and 412 high stability across a wide range of pH and temperature conditions, during incubation at 37 °C, and 413 under lon g-term storage (Figure 1 B -F). This physicochemical robustness positions EPC -based 414 systems as a valuable alternative to commonly used DOTAP - and ionizable lipid-based carriers for 415 nucleic acid delivery.41 416 Next, the m echanistic in vitro analyses confirmed effici ent siRNA complexation and protection by 417 CLPs, as indicated by gradual ζ-potential reduction, preserved colloidal stability, and validated siRNA 418 binding by PAGE analysis (Figure 3 A-E). Importantly, the complexes remained stable at physiological 419 pH while enabling controlled siRNA release at pH 5.5-6.5 (Figure 3 F, H), consistent with the critical 420 role of endosomal escape in effective RN A delivery .42,43,24 Other studies have shown that 421 encapsulation of therapeutic PD-L1 siRNA and an EGFR-targeting short peptide within cationic PEI-422 LNPs has been reported to enhance stability, enable controlled release, a nd improve transfection 423 efficiency in EGFR-positive lung cancer.44 424 A key improvement of this work is the incorporation of DSPE-PEG(2000)-NH₂ lipids, which improves 425 colloidal stability and prolongs systemic circulation of the CLP platform. However, PEGylation is 426 associated with the well-known ‘PEG dilemma’, as it can reduce cellular uptake and limit endosomal 427 escape28,29. In our system, these effects are mitigated by an optimized lipid composition (Figure 1B), 428 together with ligand -mediated targeting and pH -control release, enabling efficient intracellular 429 delivery despite the presence of a PEG corona. Moreover, t his design provided steric stabilization 430 typical of PEGylation while enabling additional electrostatic binding of anionic siRNA without 431 inducing aggregation and while maintaining a net positive surface charge (Figure 3 D, G, E ). These 432 findings highlight an underexplored role of PEG end -group chemistry in nucleic acid delivery 433 systems. The observed effects are consistent with evidence showing that functionalized PEG 434 influences surface ch arge, protein corona formation, and nanoparticle stability in biological 435 environments.45 Furthermore, the combined presence of PEG -amine groups and folic acid ligands 436 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 19 modulated release kinetics without affecting particle size or disparity, underscoring the importance 437 of surface chemistry in balancing serum stability with pH-dependent release (Figure 3 F, H). 438 Moreover, se lective uptake of CLP -FA-siEGFRFAM formulation by Caco -2 human colorectal cancer 439 cells compared to fibroblasts (Figure 4 A, B) validates our targeting strategy based on differential 440 folate receptor (FR) expression. FR is over expressed in approximately 80% of colorectal 441 adenocarcinomas while showing limited expression in normal tissues, making it an attractive target 442 for selective drug delivery.46 Microscopic analysis confirmed cellular uptake within 30 min, 443 consistent with the rapid receptor -mediated endocyt osis reported for folate -conjugated 444 nanoparticles (Figure 4 B, S3A). 445 The concentration-dependent uptake obser ved with Cy5 -labeled CLP-siEGFR2′-OMe (Figure 4 C) and 446 efficient carrier loading (Figure 3 G, S3 B) indicate that chemical modifications do not interfere with 447 CLP formation or cellular internalization. No significant cytotoxicity was observed in cancer cells, 448 normal colon epithelial cells, or macrophages at therapeutic concentrations of empty CLPs (Figure 449 4 D, S3 C) addresses a longstanding concern regarding cationic lipid formulations i n clinical 450 settings.47 451 Incorporation of 2′-O-methyl ribose modifications and phosphorothioate internucleotide linkages in 452 siRNA was essential for effective gene s ilencing in the CLP system (Figure 2 C, D and 4 E -G). Dual 453 fluorescence assa ys demonstrated that CLP -siEGFR2′-OMe retained approxim ately 65% target 454 inhibition, comparable to unmodified siRNA (Figure 4 F, G ), indicating preserved RISC loading and 455 target recognition.48 An in-depth flow cytometry analysis confirmed significant EGFR inhi bition in 456 Caco-2 cells (Figure 4 H), likely reflecting altered endosomal trafficking following receptor-mediated 457 uptake. These findings are consistent with reports showing that extensively modified siRNAs can 458 maintain full RNAi activit y in vivo .49 EGFR-amplified A431 cells were used as a positive co ntrol, 459 confirming siRNA specificity and broader applicability of the CLP p latform across EGF R-460 overexpressing cancer models (Figure 4 I ). Consistent with previ ous studies using boron cluster -461 modified antisense oligonucleotides (B-ASOs) targeting the same EGFR mRNA fragment38,35,37, CLP-462 delivered siEGFR also reduced target expression and promoted apoptosis in cancer cells (Figure 4 F-463 L). The efficient gene silencing observed with CLP formulations may indicate successful endosomal 464 escape, a key rate -limiting step for lipid -based siRNA delivery (Figure 4 F-I). Although endosomal 465 escape was not directly visualized, functional EGFR inhibition measured by flow cytometry and DFA 466 tool (Figure 4 E-I), confirmed that sufficient siRNA reached the cytoplas m to engage the RNAi 467 machinery. Recently, Mitchell et al. reported the development of an EGFR antibody-conjugated lipid 468 nanoparticle (aEGFR -LNP) platform, enginee red to enhance mRNA delivery to EGFR-expressing 469 placental cells for the treatment of pregnancy -related complications.50 Agnello et al. investigated 470 the therapeutic potential of anti-EGFR and aptamer-functionalized nanostructures in triple-negative 471 breast cancer (TNBC) .51 In addition , LNPs were targeted to EGFR ( αEGFR-LNP) using a well -472 established anti-EGFR antibody (cetuximab) to enable delivery of siRNAs against HPV in head and 473 neck squamous cell carcinoma (HNSCC). The resulting αEGFR-LNP-siHPV formulation improved anti-474 tumor activity, enhanced silencing of HPV targets, and increased apoptosis in EGFR -expressing 475 HNSCC cell lines compared to non-targeted LNPs.52,24 476 Furthermore, biodistribution studies demonstrated favorable pharmacok inetic properties of CLP -477 siEGFR2′-OMe (Figure 6 A -F). Whole-body imaging revealed c learance over 24 h for both CLP Cy7 and 478 CLP-FACy7, indicating prolonged circulation compared with naked siRNA, which is rapidly eliminated 479 by renal filtration (Figure S4 C-G). Heatmap analysis showed more efficient siRNA release from CLP 480 than CLP-FA at endosomal pH 5.5, likely due to steric hindrance imposed by the folate ligand in vitro 481 study (Figure 3 H). Consistently, CLP-Cy5siEGFR2′-OMe exhibited enhanced tumor accumulation in mice 482 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 20 and prolonged rete ntion compared with naked siRNA (Fig ure 6 B-F and S4 A-B), reflecting the 483 combined contributions of passive and active targeting.53,24 484 Significant tumor mass reduction following seven consecutive daily administrations of CLP-siEGFR2′-485 OMe provides proof -of-concept for therapeutic efficacy (Figure 6 H-I). Western blot analysis 486 confirmed downregulation EGFR protein levels in tumor but not in normal tissue (Figure 6 J-K and 487 S4 H), indicating on -target gene silencing rather than nonspecific effects. The administered dose 488 falls within the range used in clinical siRNA studies, supporting translational relevance.54 489 The absence of body weight loss after treatment (Figure S4 I) and normal organ function markers 490 (Figure 5 A-C) indicate the drug was well -tolerated. This is notable given that hepatotoxicity has 491 limited some lipid nanoparticle formulations with high positive charge density.55 492 Overall, CLP -siEGFR2′-OMe achieves tumor -selective delivery, functional gene silencing, and 493 therapeutic efficacy in a human colorectal cancer xenograft model without detectable systemic 494 toxicity. These findings establish a strong preclinical foundation for lipoplex -mediated siRNA 495 delivery to solid tumors and support further development toward clinical translation. 496 497 498 4. Material and Methods 499 Oligonucleotides were synthesized by phosphoramidite solid -phase method using LCA CPG solid 500 support and commercial phosphoramidites (Glen Research, USA, for phosphoramidite 6 -FAM, 501 ChemGenes, USA, for all other phosphoramidites). Hyacinth BMT activator was purchased from 502 empBiotech GmbH (Berlin, Germany), C18 SepPak cartridges from Waters Corp., (Miliford, MA, 503 USA), anhydrous acetonitrile (J.T. Baker brand) and ammonium hydroxide (30%, J.T. Baker brand) 504 from Avantor Performance Materials (Center Valley, PA, USA). 505 Negative ion electrospray mass spectra were obtained on a Synapt G2 Si high -resolution mass 506 spectrometer (Waters) equipped with a quadrupole -time-of-flight mass analyzer (Waters Corp., 507 Miliford,MA, USA). Data collected in a continuous mode was deconvolved using the MaxEnt1 508 algorithm. UV absorption measurements were performed using a Jasco V -770 UV−VIS/NIR 509 spectrophotometer (Jasco Int., Easton, MD, USA). RP-HPLC analyzes were performed on a Shimadzu 510 Prominence HPLC system (Kyoto, Japan) using Kinetex 5 µm C-18 column (100 Å, 250 mm × 4.6 mm 511 column, Phenomenex). Analysis conditions for cyanine 5 labelled oligonucleotides s -Cy5EGFR2’-OMe 512 and s-Cy5scrambled2’-OMe: buffer A - 0.1 M CH 3COONH4, pH 6.7; buffer B: CH 3CN flow: 1 mL / min. 513 Gradient of buffer B: 0 → 2 min 0%, 2 → 25 min 0-48%, 25 → 30 min 48- 60%, 30 → 35 mins 60-0%. 514 UV detection was performed at λmax 260 nm (and at 494 nm for FAM-labelled RNA and at 646 nm 515 for Cy5 -labelled RNA ), and the amount of compound was defined in optical units (OD) at the 516 wavelength of 260 nm. 517 Cationic liposomes were formulated using 1,2-dimyristoyl-sn-glycero-3-ethylphosphocholine (EPC), 518 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), 1,2 -dimyristoyl-sn-glycero-3-phosphocholine 519 (DMPC), and 1,2 -distearoyl-sn-glycero-3-phosphoethanolamine-N-[amino(polyethylene glycol) -520 2000] (DSPE -PEG(2000)-NH₂). For targeted form ulations, 1,2 -distearoyl-sn-glycero-3-521 phosphoethanolamine-N-[folate(polyethylene glycol) -2000] (DSPE -PEG(2000)-Folate) was added. 522 Fluorescently labeled liposomes were prepared using 1,2 -distearoyl-sn-glycero-3-523 phosphoethanolamine conjugated with Cyanine 7 (18:0 Cy7 PE). 524 All lipids were purchased from Avanti Polar Lipids (Alabaster, AL, USA) and used without further 525 purification. Ultrapure Milli-Q water (DNase-free and RNase-free grade) was used for all liposome 526 preparation and nucleic acid handling procedures. 527 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 21 4.1. Synthesis of RNA oligonucleotides 528 All RNA oligonucleotides were synthesized at CMMS PAS on 0.1 µmol scale using GeneWorld 529 DNA/RNA synthesizer H6 (K&A Laborgeraete GbR, Schaafheim, Germany) with standard protocols.56 530 The synthesis of oligonucleotides (s -EGFR, as-EGFR, s-scrambled, as-scrambled, s-GFP, as-GFP, s-531 EGFR2’-OMe, as-EGFR2’-OMe) was carried out in the DMT -on mode, so that it was possible to separate 532 them as pure 5’-protected RNAs according to the described procedure.57 Then, upon removal of the 533 DMT group, oligoribonucleotides were additionally purified by RP -HPLC, desalted on SepPak 534 columns and subjected to the mass spectrometry analysis . For oligonucleotides with 535 phosphorothioate bonds (s-EGFRPS, as-EGFRPS, s-EGFRPS/2’-OMe, as-EGFRPS/2’-OMe), the procedure was 536 modified, as the fists step of deprotection/purification, the β-cyanoethyl protecting groups were 537 removed from phosphorus in internucleotide linkages by washing the bed with piperidine in 538 acetonitrile. Subsequent steps followed the standard protocol for the other oligonucleotides. For 539 fluorescently labelled 5’-FAM-antisense strand (as-FAMEGFR), 5’-FAM-sense strand (s-FAMEGFR) and 540 5’-Cy5-sense strands (Cy5 -EGFR2’-OMe, Cy5-scrambles2’-OMe) oligonucleotides, the cuts off from the 541 bed were carried out by routine procedure and compounds were directly purified by RP -HPLC 542 (conditions described above). All the retention times and MS data obtained for the synthesized 543 oligonucleotides are summarized in Table 1. 544 4.2. Synthesis of CLP, CLP-FA, CLPCy7, CLP-FACy7 liposomes. 545 All liposomes were prepared using the thin -film hydration metho d followed by Bangham et al .,58 546 with our patented components (PCT.PL2022.000046 and WO2023/022615). An appropriate masses 547 of lipids, was dissolved in 2 mL of analytical -grade chloroform. The solution was thoroughly mixed 548 using a vortex mixer. The lipid -chloroform mixture was then transferred to a single round -bottom 549 flask, and the chloroform was evaporated using a rotary evaporator set to 45°C, forming a thin lipid 550 film. Next, 4 mL of sterile saline solution was added to the flask, which was tightly sealed and 551 incubated at 30°C overnight (12 hours) to hydrate the lipid film. After hydration, the lipid layers 552 were vortexed until the entire lipid film detached from the flask walls. The suspension was further 553 incubated for an additional 6 hours at 40°C with continuous shaking. The lipid vesicles were then 554 extruded through a 0.1 µm membrane filter, passing through the membrane 20 times to achieve 555 uniform size. The schematic of the procedure for the preparation of CLPs, CLPs-FA liposomes, and 556 their fluorescently labeled counterparts (CLPsCy7, CLPs-FACy7) is presented in Figure 1A. The liposome 557 fractions were subsequently combined and dia lyzed against ultrapure water for 16 hours using 558 Float-A-Lyzer G2 Dialysis Device membrane MWCO 100kD. The external solution was replaced 559 twice, at 2 and 4 hours. After dialysis, the liposomes were filtered through sterile PTFE syringe filters 560 with a pore size of 0.2 µm. The resulting liposomes were prepared for binding with nucleic acids in 561 subsequent experiments. 562 4.3. Physicochemical characteristic of CLP and CLP-FA liposomes. 563 Dynamic light scattering (DLS) and zeta potential analyses were performed using a Zetasizer Ultra 564 instrument (Malvern Panalytical, UK) equipped with ZS Xplorer software. Measurements were 565 conducted at 25 °C after 30 s equilibration. For size and polydispersity analysis, samples were placed 566 in disposable sizing cuvettes (ZEN0040), with t he material type set to “liposomes” and dispersant 567 set to water. Data were processed using the multiple narrow modes algorithm. For zeta potential 568 measurements, 800 µL of liposome suspension was transferred into a folded capillary zeta cell 569 (DTS1070, Malvern Panalytical). Measurements were carried out under identical conditions (25 °C, 570 30 s equilibration). 571 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 22 4.4. Preparation of CLP-siRNAs for measurements. 572 Cationic liposomes (CLP and CLP -FA) were prepared as described above and dialyzed against 573 ultrapure water. U nmodified siRNA (siEGFR) and ch emically modified siRNA (siEGFR 2’-OMe) were 574 mixed with the liposomes to obtain defined siRNA:CLP mass ratios of 1:100, 1:50, 1:20, and 1:10 575 (w/w). The mixtures were incubated for 30 minutes at room temperature to allow comple x 576 formation. Unless otherwise noted, measurements were performed at 25°C. Two formulation 577 variants were analyzed in parallel (CLP and CLP-FA). 578 4.5. Formation of [32P]-siRNA-lipoplex complex analyzed by PAGE 579 To 0.1 OD of the RNA strand s (sense and antisense) dis solved in 15 µL of Milli -Q water [32P-γ] ATP 580 (37.0 MBq, 1.00 mCi), T4 polynucleotide kinase (1 µL, 10,000 units / mL), and 2 µL of phosphorylation 581 reaction buffer supplied by the kinase manufacturer (final volume 20 µL). The reaction mixture was 582 then incubated at 37 °C for 1 hour. The next step was to inactivate the kinase by incubating the 583 reaction mixture at 85°C for 3 minutes. [32P]-siRNA was assembled by combining solutions of each 584 strand of [32P]-RNA radiolabelled in a 1: 1 molar ratio in sterile milliQ water to a final volume of 20 585 µL. The mixture was heated at 85°C for 5 min, then slowly cooled to room temperature for 1 h. 586 The radiolabeled duplex [32P]-siRNA solution was added to the lipoplex solution (CLP, CLP-FA) in the 587 defined weight ratio (1:1, 1:2.5, 1:10, 1:20), followed by incubation at room temperature for 30 min. 588 After this incubation time, the samples were applied to a 15% polyacrylamide gel (PAGE) under non-589 denaturing conditions (without the addition of urea). The electrophoresis was performe d at room 590 temperature with a voltage of 300 -400 V / cm and a current of 60 mA for 30 minutes. Then, 591 electrophoresis gels were developed autoradiographically on diagnostic membranes. 592 4.6. Preparation of pH buffers for [32P]-siRNA release studies from lipoplex complexes 593 A buffer with pH 4.0 was prepared based on NaOAc sodium acetate and acetic acid (150 mM) in 594 10 ml mQ water. NaOAc was combined with acetic acid to obtain a buffer pH 4.0. In this case, it was 595 not possible to use the buffer prepared on the basis of PBS, due to the different buffer capacity of 596 this solution. The remaining buffers (5.5, 6 .0, 7.4) were made based on Na 2HPO4 and NaH 2PO4 597 (150 mM) in 10 ml mQ water. By combining the two solutions in different proportions, buffers with 598 a higher pH were obta ined (5.5, 6 .0, 7.4). The pH of the solutions was confirmed by the 599 potentiometric method by measuring with a pH-meter. 600 4.7. Release of [32P]-siRNA cargo from the formulation 601 Labeled 32P-siRNA was coated with li poplex (CLP, CLP -FA) in a mass ratio of 1:10, incubation 30 602 minutes, at room temperature. Then added to the [32P]-siRNA (controls) alone as well as the coated 603 systems. In the next step, the samples were incubated with pH buffer (4.0, 5.5, 6 .0, 7.4) for 30 604 minutes, 1 hour, 2 hours and 4 hours at room temperature. Then, mixed with loading buffer (without 605 EDTA) was added to the samples and a double PAGE analysis was performed from the anode (+) to 606 the cathode (-) and inversely as well. 607 4.8. Stability studies of siRNA against lipoplex after 7-day incubation 608 Unmodified siEGFR and scrambled labeled and assembled according to the above procedures were 609 coated with CLP lipoplex in the ratio 1:10 followed by incubation for 30 minutes at room 610 temperature. After this time, the samples were incubated at 37°C or 4°C, and after 7 days their 611 electrophoretic analysis was performed (15% PAGE, non -denaturing conditions) either in the form 612 of a siRNA -lipoplex complex or after release of oligonucleo tides (f rom the siRNA -lipoplex 613 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 23 formulation after the addition of 10 -fold excess, i.e. 15 µL of phosphate buffer at pH 5.5). Prior to 614 loading on the gel, 5 µL of loading buffer without EDTA was added to the samples and 615 electrophoretic analysis was performed. 616 4.9. Cell viability 617 Caco-2 and CCD -841CoN cells were seeded in 96 -well plates and incubated overnight to allow 618 attachment. The culture medium was then replaced with fresh medium containing varying 619 concentrations of empty CLP or CLP -FA lipoplex (ranging from 10 to 200 µg/mL). Staurosporine (1 620 µM) was used as a positive control for cell death. After 48 h of incubation, the MTT assay was 621 performed as described in our previous work .35 Viability of differentiated THP -1 and RAW 264 .7 622 macrophages was assessed using MTS assay. Cells were treated with lipoplexes as described above. 623 Cell viability was determined using the CellTiter 96 AQueous One Solution Cell Proliferation Assay 624 (MTS) from Promega (Madison, WI, USA). Following 24 hours of treatment, 10 µL of the MTS reagent 625 was added directly to each well containing 100 µL of culture medium. The plates were incubated at 626 37°C for 3 h in a humidified, 5% CO2 atmosphere. The absorbance of the soluble formazan product 627 was measured at 490 nm, with a read at λ = 640 for background subtraction, using a BioTek Cytation 628 5 platform ( Agilent, Santa Clara, CA, USA). For both assays, cell viability was calculated as a 629 percentage relative to the untreated control cells (set to 100%). 630 4.10. Assessment of inhibitory activity of siRNAs coated with CLP lipoplexes 631 CaCo-2 and A431 cells were cultured according to the ATCC procedure. On the day before 632 microscopic observations, cells were seeded in black 96 -well plates (clear bottom) at 15 × 10 3 cells 633 in 100 μL of c omplete medium per well and left overnight at 37°C and 5% CO 2. After overnight 634 incubation, full medium was removed from the cells and replaced with OPTI -MEM base medium. 635 Transfection was carried out using commercial Lipofectamine 2000 reagent (Invitrogen) in a ratio of 636 2:1 (2 µL Lipofectamine 2000 per 1 µg nucleic acid) according to manufacturer's protocol. In the dual 637 fluorescence assay (DFA) CaCo -2 cells were transfected with pDsRed -N1 reporter plasmid (30 638 ng/well) and pEGFP -EGFR plasmid (100 ng/well). Th en, appropriate siRNAs (10 -200 nM) were co -639 transfected, delivered to cells with the tested lipoplexes CLP or CLP -FA or as a control with a 640 commercial transfection reagent (Lipofectamine 2000) according to the manufacturer's protocol. 641 After 5 hours of incubation, the transfection mixture was withdrawn from each well and cells were 642 flooded with 200 µL of fresh culture medium containing antibiotics. After 48 h of incubation at 37 643 °C in 5% CO 2, cells were washed twice with PBS buffer (no Ca 2+and Mg2+) and lysed overnight with 644 NP-40 buffer (150 mM NaCl, 1% IGEPAL, 50 mM Tris-HCl pH 7.0 and 1 mM PMSF) at 37°C. Prepared 645 cell lysates were used to determine the level of fluorescence were determined using a FLUOstar 646 Omega reader (BMG LABTECH). Flow cytometry was performed as previously described.38 647 4.11. Western Blotting 648 Total protein was extracted from excised equivolume tissue samples. Tissues were homogenized in 649 ice-cold 1 × Cell Lysis Buffer [Cell Signaling Technology (CST), Danvers, MA, USA] supplemented with 650 a phenylmethanesulfonyl fluoride (1 mM, Merck) and 1 × Protease and Phosphatase Inhibitor 651 Cocktail [Thermo Fisher Scientific (TFS), Waltham, MA, USA] using a using Bio -Gen PRO200 652 Homogenizer (PRO Scientifi c, Oxford, CT, USA). The lysates were centrifuged at 18,188 × g for 10 653 min at 4°C. Protein concentrations were determined in the obtained supernatants usi ng the BCA 654 Protein Assay (TFS). Equal amounts of protein were denatured at 95°C for 5 min in Laemmli sample 655 buffer, separated by 8 -12% PAGE in MES -SDS buffer (90 V for 10 min and 120 V for 60 m in), and 656 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 24 transferred onto PVDF membranes (TFS) using the iBlot2 (TFS) system (20 V for 1 min, 23 V for 4 657 min, and 25 V for 2 min). Membranes were blocked with 3% non-fat dry milk (Merck) in Tris-buffered 658 saline with 0.1% Tween 20 (TBST) for 1 h. After a single rinse with 1 × TBST, the membranes were 659 incubated overnight at 4°C with primary antibodies against EGFR (rabbit, AbCam, Cambridge, UK, 660 #AB52894) or α-tubulin (rat, TFS, #MA180017). Following four washes with TBST, membranes were 661 incubated with anti -rabbit [HRP -conjugated from CST, #7074] or anti -rat (AlexaFluor555 -662 conjugated, #4417) secondary antibodies. After 1 h of incubation at room temperature the 663 membranes were rinsed with 1 × TBST. All antibodies were diluted in 1 × TBST with 3% BSA at 1:2000 664 for primaries and at 1:10000 for secondaries. 665 When necessary, protein bands were visualized using Westar Supernova ECL reagent (Cyanagen, 666 Bologna, Italy). Image acquisition was performed on an Azure c400 (Azure Biosystems Inc., Dublin, 667 CA, USA). Densitometric analysis was performed on sub -saturated images using ImageJ 1.54f 668 software (National Institutes of Health, Bethesda, MD, USA), with EGFR band intensities normalized 669 to the loading control (α-tubulin). 670 4.12. In vivo study 671 Six-week-old female NOG -F mice (strain NOD.Cg -Prkdc^scid Il2rg^tm1Sug/JicTac) were obtained 672 from Taconic Biosciences and used for xenograft experiments. This strain represents a highly 673 immunodeficient model lacking functional T, B, and NK cells and is widely used fo r human tumor 674 xenograft studies. Upon arrival, animals were subjected to a 7-day quarantine, followed by handling 675 and habituation, in full compliance with GLP standards prior to the initiation of experimental 676 procedures. 677 Mice were housed in groups of five per cage under specific pathogen-free conditions with controlled 678 environmental parameters (temperature 20–24 °C, relative humidity 45–65%, 12 h light/dark cycle, 679 and 15 air changes per hour). Animals had free access to a standard rodent diet (Altromin) and water 680 ad libitum. Environmental enrichment, including nesting material and shelters, was provided to 681 promote natural behaviors and reduce stress. All animal procedures were performed in accordance 682 with national regulations for the care and use of laboratory animals and were approved by the Local 683 Ethical Committee for Animal Experiments (approval no. 90/2023). For xenograft implantation, 684 Caco-2 cells were detached at logarithmic growth phase using trypsinization, collected and 685 centrifuged to obtain a pellet. The cell pellet was resuspended in sterile phosphate-buffered saline 686 (PBS). Cell number and viability were determined using a hemocytometer and the suspension was 687 adjusted to the required concentration. Immediately prior to implantation, cells were prepared at a 688 density of 2 × 10⁶ cells per injection in PBS and kept on ice until administration. 689 Female NOG mice bearing Caco-2 xenografts were monitored until tumor volumes reached 50-100 690 mm³. Mice were randomly assigned to five groups (n = 6 per group): PBS, naked siEGFR2’OMe, empty 691 CLP, CLP-SCR2’OMe and CLP-siEGFR2’OMe. Treatments were administered via tail vein injection of siRNA 692 dose of 2 mg/kg body weight and ten -fold more for CLP. Groups receiving treatment were 693 administered formulations three times at defined intervals of 48h. Body weight and tumor volume 694 were recorded at each administration and at subsequent time points. Forty -eight hours after the 695 last injection, animals were euthanized by decapitation. Blood (200 µL) was collected into EDTA -696 containing tubes for immediate hematological analysis. Samples were centrifuged (3000 rpm, 10 697 min, 4 °C) to separate plasma; 50 µL plasma was frozen at -80 °C, while the remainder was used for 698 biochemical analyses. Organs, including brain, heart, lungs, liver, kidneys, and spleen, were excised, 699 weighed, and frozen on dry ice. Tumors were excised, weighed, photographed, and frozen for 700 subsequent analyses. 701 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 25 4.13. Treatment and in vivo imaging 702 Female NOG mice bearing Caco -2 xenografts were randomly assigned to fou r treatment groups (n 703 = 8): CLPCy7, CLP-FACy7, Cy5siRNA2’-OMe and CLPCy5siRNA2’-OMe. The dosing was 2 mg/kg body weight for 704 siRNA and ten -fold higher for CLP formulations. Control animals were left untreated and were 705 included solely as reference to assess background tissue auto-fluorescence. This allowed accurate 706 interpretation of fluorescence si gnals from experimental groups. Formulations were administered 707 via tail vein injection. In vivo fluorescence imaging was p erformed using an IVIS Spectrum 708 (Perkinelmer) imaging system at wavelengths appropriate for Cy5 and Cy7 dye. Tumor and 709 biodistribution imaging were conducted at 0, 6, 12, 18 and 24 hours post -injection. Mice were 710 anesthetized with isoflurane (3 –4% induction, 1.5% maintenance, oxygen flow 1 L/min) during 711 imaging. Images were acquired from both dorsal and ventral sides, and tails were covered to 712 eliminate fluorescence from the injection site. 713 4.14. Blood Hematology and Serum Biochemistry Assessment 714 Blood samples were collected from euthanized mice into EDTA -containing tubes. Hematological 715 analysis was performed using an automated hematology analyzer Abacus Junior Vet 5 following the 716 manufacturer’s instructions. The following parameters were evaluated: White Blood Cells (WBC), 717 Red Blood Cells (RBC) and Platelets (PLC). All measurements were performed in duplicate, and 718 instrument calibration and quality controls were verified prior to analysis to ensure accuracy and 719 reproducibility. 720 Plasma was separated by centrifugation at 3000 rpm for 10 minutes at 4 °C, and serum was collected 721 for biochemical analysis. Serum markers were quantified using commercial kits from Cormay 722 following the manufacturer’s protocols. All measurements were performed on a UV –Vis plate 723 reader (Synergy H1, BioTek) with appropriate blanks and calibration curves. 724 Creatine Kinase MB (CK -MB): CK-MB activity was determined using the Liquick Cor -CK-MB 30 kit 725 (Cormay). The assay is based on a coupled enzymatic reaction in which CK -MB catalyzes the 726 conversion of creatine phosphate and ADP to creatine and ATP. The generated ATP is subsequently 727 utilized in auxiliary reactions leading to NADPH formation, and the change in absorbance at 340 nm 728 is measured spectrophotometrically. 729 Alanine Aminotransferase (ALAT): ALAT activity was measured using the Liquick Cor -ALAT 60 kit 730 (Cormay). The method is based on the conversion of alanine and α -ketoglutarate to pyruvate and 731 glutamate. Pyruvate is subsequently reduced to lactate in the presence of lactate dehydrogenase, 732 accompanied by oxidation of NADH to NAD⁺. The decrease in absorbance at 340 nm is proportional 733 to ALAT activity. 734 Aspartate Aminotransferase (ASAT): ASAT activity was determined using the Liquick Cor-ASAT 60 kit 735 (Cormay). The reaction involves transamination of aspartate and α -ketoglutarate to oxaloacetate 736 and gl utamate. Oxaloacetate is subsequently reduced in a coupled enzymatic reaction involving 737 NADH, and the decrease in absorbance at 340 nm is monitored. 738 Urea: Urea concentration was measured using the Liquick Cor -UREA 60 kit (Cormay). In this 739 enzymatic assay, urea is hydrolyzed by urease to ammonia and carbon dioxide. The produced 740 ammonia participates in a subsequent enzymatic reaction generating a chromogenic product, which 741 is measured spectrophotometrically at 520–550 nm. 742 Creatinine: Creatinine levels were de termined using the Liquick Cor -CREATININE 60 kit (Cormay). 743 The assay is based on the Jaffe reaction, in which creatinine reacts with picrate ions in an alkaline 744 environment to form a colored complex. The intensity of the color, measured 745 spectrophotometrically at 490–510 nm, is proportional to the creatinine concentration. 746 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 26 All assays were calibrated and validated prior to analysis using appropriate calibration and quality 747 control reagents, including CORMAY CK -MB CALIBRATOR, CORMAY MULTICALIBRATOR LEVEL 1, 748 LEVEL 2 and quality control sera (CORMAY SERUM HP, CORMAY SERUM HN). These reagents were 749 used to verify the accuracy and reliability of the analytical methods according to the manufacturer’s 750 recommendations. All measurements were performed in duplicate to ensure reproducibility. 751 4.15. Statistical Analysis 752 An unpaired Student t-test was used to calculate 2-tailed P-values to estimate statistical significance 753 between 2 treatment groups. One -way analysis of variance and Bonferroni post -test were used to 754 assess differences between multiple groups and in tumor growth kinetics, respectively. Statistically 755 significant P-values were indicated in the figures compared to untreated or PBS groups). Data are 756 presented as mean±SD (n=3 -6); *P < 0.05, **P < 0.01, ***P < 0.001, ** **P < 0.0001 by one -way 757 ANOVA with Bonferroni’s correction post hoc test. Data were analyzed using Prism software version 758 10 (GraphPad Software). 759 760 761 5. Conflict of Interest 762 O.S. is an inventor on the international patent application s PCT.PL2022.000046 and 763 WO2023/022615 (RNA binding and stabilizing cationic liposome, its application and method of 764 loading the liposome with emetine) covering the design of lipid formulations. D.K. directs the design 765 and execution of preclinical studies at Biotechna S.A company. A.Z serves on the board , and B.N. 766 provides strategic leadership at Biotechna S.A., actively guiding the organization’s growth and 767 direction. No potential conflicts of interest were declared by other authors. 768 769 770 6. Acknowledgments 771 This research was funded by the Medical Research Agency (Poland) under the project implemented 772 by Biotechna S.A., grant no. 2022/ABM/06/00005. The authors thank Anna Maciaszek (CMMS PAS) 773 for the s ynthesis of oligonucleotides, Ewelina Wielgus (CMMS PAS) for the mass spectrometric 774 analysis. 775 776 777 778 779 780 781 782 783 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 27 784 7. References 785 1. Morgan, E. et al. Global burden of colorectal cancer in 2020 and 2040: incidence and mortality 786 estimates from GLOBOCAN. Gut 72, 338 LP – 344 (2023). 787 2. Siegel, R., DeSantis, C. & Jemal, A. Colorectal cancer statistics, 2014. CA Cancer J Clin 64, 104–788 117 (2014). 789 3. Cardoso, R. et al. Overall and stage -specific survival of patients with screen -detected 790 colorectal cancer in European countries: A population-based study in 9 countries. Lancet Reg 791 Heal - Eur 21, 100458 (2022). 792 4. Siegel, R. L., Miller, K. D., Fuchs, H. E. & Jemal, A. Cancer statistics, 2022. CA Cancer J Clin 72, 793 7–33 (2022). 794 5. Dosso, S. De. EGFR signaling in colorectal cancer : a clinical perspective. 21–38 (2015). 795 6. Liu, J., Zhou, Q., Xu, J., Wang, J. & Zhang, Y. Detection of EGFR express ion in patients with 796 colorectal cancer and the therapeutic effect of cetuximab. 21, 95–100 (2016). 797 7. Misale, S., Di Nicolantonio, F., Sartore -Bianchi, A., Siena, S. & Bardelli, A. Resistance to Anti -798 EGFR Therapy in Colorectal Cancer: From Heterogeneity to Convergent Evolution. Cancer 799 Discov 4, 1269–1280 (2014). 800 8. Khan, K. et al. Targeting EGFR pathway in metastatic colorectal cancer- tumour heterogeniety 801 and convergent evolution. Crit Rev Oncol Hematol 143, 153–163 (2019). 802 9. Khan, K. H. et al. Longitudinal Liquid Biopsy and Mathematical Modeling of Clonal Evolution 803 Forecast Time to Treatment Failure in the PROSPECT -C Phase II Colorectal Cancer Clinical 804 Trial. Cancer Discov 8, 1270–1285 (2018). 805 10. Roh, J. K. et al. Cetuximab and Chemotherapy as Initial Tr eatment for Metastatic Colorectal 806 Cancer. (2009). 807 11. Loree, J. M. et al. Expanded Low Allele Frequency RAS and BRAF V600E Testing in Metastatic 808 Colorectal Cancer as Predictive Biomarkers for Cetuximab in the Randomized CO.17 Trial. Clin 809 Cancer Res 27, 52–59 (2021). 810 12. Pedersen, M. W. et al. Sym004: a novel synergistic anti -epidermal growth factor receptor 811 antibody mixture with superior anticancer efficacy. Cancer Res 70, 588–597 (2010). 812 13. Tariq, A. et al. Exploring 4th generation EGFR inhibitors: A rev iew of clinical outcomes and 813 structural binding insights. Eur J Pharmacol 997, 177608 (2025). 814 14. Tang, Q. & Khvorova, A. RNAi -based drug design: considerations and future directions. Nat 815 Rev Drug Discov 23, 341–364 (2024). 816 15. Traber, G. M. & Yu, A. -M. RN Ai-Based Therapeutics and Novel RNA Bioengineering 817 Technologies. J Pharmacol Exp Ther 384, 133–154 (2023). 818 16. Lu, D., Dou, F. & Gao, J. Fitusiran: The first approved siRNA therapy for hemophilia via 819 reducing plasma antithrombin levels. Drug Discov Ther 19, 131–132 (2025). 820 17. Xiao, B. et al. Development, opportunities, and challenges of siRNA nucleic acid drugs. Mol 821 Ther Nucleic Acids 36, (2025). 822 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 28 18. Lee, S.-Y. et al. A theranostic micelleplex co -delivering SN-38 and VEGF siRNA for colorectal 823 cancer therapy. Biomaterials 86, 92–105 (2016). 824 19. Gary, D. J. et al. Influence of Nano -Carrier Architecture on in Vitro siRNA Delivery 825 Performance and in Vivo Biodistribution: Polyplexes vs Micelleplexes. ACS Nano 5, 3493–3505 826 (2011). 827 20. Chen, Y. & Huang, L. Tumor-targeted delivery of siRNA by non-viral vector: safe and effective 828 cancer therapy. Expert Opin Drug Deliv 5, 1301–1311 (2008). 829 21. Kim, S. H., Jeong, J. H., Lee, S. H., Kim, S. W. & Park, T. G. PEG conjugated VEGF siRNA for anti-830 angiogenic gene therapy. J Control release Off J Control Release Soc 116, 123–129 (2006). 831 22. Adams, D. et al. Patisiran, an RNAi Therapeutic, for Hereditary Transthyretin Amyloidosis. N 832 Engl J Med 379, 11–21 (2018). 833 23. Hoy, S. M. Patisiran: First Global Approval. Drugs 78, 1625–1631 (2018). 834 24. Yang, F. & Scott, T. A. Advances in lipid nanoparticles delivering genetic medicines for solid 835 cancers. Mol Ther Nucleic Acids 37, (2026). 836 25. Cheng, Y. -H., He, C., Riviere, J. E., Monteiro -Riviere, N. A. & Lin, Z. Meta -Analysis of 837 Nanoparticle Delivery to Tumors Using a Physiologically Based Pharmacokinetic Modeling and 838 Simulation Approach. ACS Nano 14, 3075–3095 (2020). 839 26. Wilhelm, S. et al. Analysis of nanoparticle delivery to tumours. Nat Rev Mater 1, 16014 (2016). 840 27. Gilleron, J. et al. Image-based analysis of lipid nanoparticle –mediated siRNA delivery, 841 intracellular trafficking and endosomal escape. Nat Biotechnol 31, 638–646 (2013). 842 28. Hatakeyama, H., Akita, H. & Harashima, H. The polyethylenegly col dilemma: advantage and 843 disadvantage of PEGylation of liposomes for systemic genes and nucleic acids delivery to 844 tumors. Biol Pharm Bull 36, 892–899 (2013). 845 29. Pan, J. et al. Emerging strategies against accelerated blood clearance phenomenon of 846 nanocarrier drug delivery systems. J Nanobiotechnology 23, 138 (2025). 847 30. Suk, J. S., Xu, Q., Kim, N., Hanes, J. & Ensign, L. M. PEGylation as a strategy for improving 848 nanoparticle-based drug and gene delivery. Adv Drug Deliv Rev 99, 28–51 (2016). 849 31. Kesharwani, P. et al. PEGylated PLGA nanoparticles : unlocking advanced strategies for cancer 850 therapy. (2025). 851 32. Ríos-hoyo, A., Monzonís, X., Vidal, J., Linares, J. & Montagut, C. Unveiling acquired resistance 852 to anti -EGFR therapies in colorectal cancer  : a long and winding road. 1 –12 (2024) 853 doi:10.3389/fphar.2024.1398419. 854 33. Hattori, Y. et al. Optimal combination of cationic lipid and phospholipid in cationic liposomes 855 for gene knockdown in breast cancer cells and mouse lung using siRNA lipoplexes. Mol Med 856 Rep 26, 253 (2022). 857 34. Hu, Y. -B., Dammer, E. B., Ren, R. -J. & Wang, G. The endosomal -lysosomal system: from 858 acidification and cargo sorting to neurodegeneration. Transl Neurodegener 4, 18 (2015). 859 35. Ebenryter-Olbińska, K. et al. Versatile Method for the Site-Specific Modification of DNA with 860 Boron Clusters: Anti -Epidermal Growth Factor Receptor (EGFR) Antisense Oligonucleotide 861 Case. Chemistry 23, 16535–16546 (2017). 862 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 29 36. Sierant, M. et al. Specific Silencing of L392V PSEN1 Mutant Allele by RNA Interference. Int J 863 Alzheimer’s Dis 2011, 809218 (2011). 864 37. Gudanis, D. et al. Formation of an RNA Quadruplex -Duplex Hybrid in Living Cells between 865 mRNA of the Epidermal Growth Factor Receptor (EGFR) and a G -Rich Antisense 866 Oligoribonucleotide. Cells vol. 9 (2020). 867 38. Kaniowski, D. et al. EGFR-targeted antisense oligonucleotides modified with boron clusters 868 offer an innovative approach to cancer chemo-radiotherapy. (2025). 869 39. Choung, S., Kim, Y. J., Kim, S., Park, H. -O. & Choi, Y. -C. Chemical modification of siRNAs to 870 improve serum stability without loss of efficacy. Biochem Biophys Res Commun 342, 919–871 927 (2006). 872 40. Hu, B. et al. Therapeutic siRNA: state of the art. Signal Transduct Target Ther 5, 101 (2020). 873 41. Kulkarni, J. A., Cullis, P. R. & van der Meel, R. Lipid Nanoparticles Enabling Gene Therapies: 874 From Concepts to Clinical Utility. Nucleic Acid Ther 28, 146–157 (2018). 875 42. Hamelmann, N. M., Paats, J. -W. D. & Paulusse, J. M. J. Cytosolic Delivery of Single -Chain 876 Polymer Nanoparticles. ACS Macro Lett 10, 1443–1449 (2021). 877 43. Wittrup, A. & Lieberman, J. Knocking down disease: a progress report on siRNA therapeutics. 878 Nat Rev Genet 16, 543–552 (2015). 879 44. Yang, G. et al. Construction of PEI -EGFR-PD-L1-siRNA dua l functional nano -vaccine and 880 therapeutic efficacy evaluation for lung cancer. Thorac Cancer 13, 2941–2950 (2022). 881 45. Ikeda, Y. & Nagasaki, Y. Impacts of PEGylation on the gene and oligonucleotide delivery 882 system. J Appl Polym Sci 131, (2014). 883 46. Scaranti, M., Cojocaru, E., Banerjee, S. & Banerji, U. Exploiting the folate receptor α in 884 oncology. Nat Rev Clin Oncol 17, 349–359 (2020). 885 47. Mizrahy, S., Hazan-halevy, I., Dammes, N., Landesman-milo, D. & Peer, D. Current Progress in 886 Non-viral RNAi-Based Delivery Strategies to Lymphocytes. Mol Ther 25, 1491–1500 (2017). 887 48. Egli, M. & Manoharan, M. Chemistry, structure and function of approved oligonucleotide 888 therapeutics. Nucleic Acids Res 51, 2529–2573 (2023). 889 49. Hassler, M. R. et al. Comparison of partially and fully chemically-modified siRNA in conjugate-890 mediated delivery in vivo. Nucleic Acids Res 46, 2185–2196 (2018). 891 50. Geisler, H. C. et al. EGFR-targeted ionizable lipid nanoparticles enhance in vivo mRNA delivery 892 to the placenta. J Control release Off J Control Release Soc 371, 455–469 (2024). 893 51. Agnello, L. et al. Optimizing cisplatin delivery to triple -negative breast cancer through novel 894 EGFR aptamer-conjugated polymeric nanovectors. J Exp Clin Cancer Res 40, 239 (2021). 895 52. Kampel, L. et al. Therapeutic inhibitory RNA in head and neck cancer via functional targeted 896 lipid nanoparticles. J Control release Off J Control Release Soc 337, 378–389 (2021). 897 53. Sindhwani, S. et al. The entry of nanoparticles into solid tumours. Nat Mater 19, 566–575 898 (2020). 899 54. Zhang, M. M., Bahal, R., Rasmussen, T. P. & Zhong, X. HHS Public Access. 1 –29 (2022) 900 doi:10.1016/j.bcp.2021.114432.The. 901 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 30 55. Akinc, A. et al. The Onpattro story and the clinical translation of nanomedicines containing 902 nucleic acid-based drugs. Nat Nanotechnol 14, 1084–1087 (2019). 903 56. Sanghvi, Y. S. A status update of modified oligonucleotides for chemotherapeutics 904 applications. Curr Protoc nucleic acid Chem Chapter 4, Unit 4.1.1-22 (2011). 905 57. Nawrot, B. & Sochacka, E. Preparation of short interfering RNA containing the modified 906 nucleosides 2-thiouridine, pseudouridine, or dihydrouridine. Curr Protoc nucleic acid Chem 907 Chapter 16, 16.2.1-16.2.16 (2009). 908 58. Bangham, A. D., Standish, M. M. & Watkins, J. C. Diffusion of univalent ions across t he 909 lamellae of swollen phospholipids. J Mol Biol 13, 238–252 (1965). 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 31 935 936 937 938 939 940 941 942 943 944 945 946 947 Supplementary 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 32 966 967 968 Figure S1. Physicochemical characterization of CLP and CLP -FA lipoplexes with and without siRNA. Zeta 969 potential (A); hydrodynamic radius (B); and polydispersity index (PDI) (C) of empty lipoplexes and lipoplexes 970 loaded with siRNA EGFR2′OMe (5 or 10 μg, as shown ), measured immediately after preparation and after 5 971 days of storage. 972 973 974 975 976 977 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 33 978 979 Figure S2 . ESI-Q-TOF mass spectra of unmodified and chemically modified siRNA strands (sense and 980 antisense), including PS, 2′ -O-methyl, and PS/2′ -O-methyl variants, as well as fluorophore -labeled 981 oligonucleotides. Measured masses are consistent with theoretical values. 982 983 984 985 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 34 986 Figure S3. Properties and Activity of EGFR-targeting siRNA lipoplexes. Fluorescence microscopy images (×40 987 magnification) of Caco-2 cells treated with Lipofectamine 2000 -complexed siEGFRFAM (100nM, green dots), 988 demonstrating intracellular localization after 30min incubation; nuclei stained with DAPI (blue) and the 989 endoplasmic reticulum stained with ER -Tracker RED (A); native 15% PAGE analysis confirming complete 990 loading of Cy5-labeled siEGFR into CLP lipoplexes (B); cell viability of THP-1-derived human macrophages and 991 RAW264.7 mouse macrophages following CLP (0.1-200 µg/mL) treatment with IC50 value, assessed after 48 992 h by MTT assay ( C); control GFP silencing in Caco -2 and A431 cells tre ated with CLP- or CLP-FA-formulated 993 siRNA-GFP at the indicated concentrations by DFA ( D-E); EGFR silencing mediated by Lipofec tamine 2000-994 formulated siRNAs in Caco -2 and A431 cells after 48 h by DFA tool (F-G); EGFR expression o n Caco-2 cells 995 following treatment with emp ty CLP, CLP -siRNA EGFR, and CLP -ASO EGFR formulations (100 nM, 48 h; 996 positive controls), compared with the EGFR inhibitor PD153035 (1µM), as determined by flow cytometry. 997 998 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint 35 999 Figure S4. In vivo biodistribution and tumor accumulation of CLP-siEGFR formulations. Quantification of 1000 siRNA Cy5EGFR2’-OMe, Cy5EGFR2’-OMe encapsulated in CLP, and CLP alone in tumor tissue at 6, 12, and 24 hours 1001 post-intravenous injection (2mg/kg) by IVIS (A); Comparative accumulation of CLP-Cy5EGFR2’-OMe and siRNA 1002 Cy5EGFR2’-OMe (2mg/kg) in tumors at the indicated time points (B); Biodistribution of empty CLP Cy7 in major 1003 organs of mice at 6, 12, and 24 hours post -injection (C-D); Organ-specific accumulation of siEGFR Cy5 at the 1004 same time points (E); Biodistribution and accumulation of CLP-Cy5EGFR2’-OMe (2mg/kg) in mouse organs over 1005 6, 12, and 24 hours by IVIS (F-G); EGFR expression in liver and CaCo-2 xenografts, assessed by western blot, 1006 following treatment with indicated compounds (2 mg/kg, intravenous) (H); Body weight of mice monitored 1007 post-treatment, indicating systemic tolerability (I). Data are presented as mean ± SD. 1008 1009 .CC-BY-NC 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted March 31, 2026. ; https://doi.org/10.64898/2026.03.29.715100doi: bioRxiv preprint