pH-Responsive Nanoparticles That Reprogram Immunity in MSS Colorectal Cancer by Disrupting Protein Balance, Targeting PD-L1, and Activating Immune Receptors

Poly(L-glutamic acid) (PGA) was obtained from Vedan Biotechnology Corporation (Taichung, Taiwan). DOTAP and DSPE-PEG derivatives were sourced from Avanti Polar Lipids (Alabaster, AL, USA). The TLR7 agonist R837 (R) and CB were purchased from MedChemExpress (Monmouth Junction, NJ, USA). FAM-labeled and unlabeled miR-142 oligonucleotides were custom synthesized by GenePharma (Shanghai, China). The targeting peptides were prepared by Kelowna International Scientific (Taipei, Taiwan). Cholesterol, Tween^® 80, and paraformaldehyde were acquired from Acros Organics (Geel, Antwerp, Belgium). Cell culture media and supplements were provided by Gibco (Carlsbad, CA, USA). Additional analytical-grade reagents were procured from Cayman Chemical (Ann Arbor, MI, USA), Merck KGaA (Darmstadt, Germany), and MilliporeSigma (Burlington, MA, USA).

The pH-sensitive polymer PGA-PEG was synthesized by coupling PGA (Vedan Biotechnology Corporation, 100 mg, molecular weight 15–50 kDa) with methoxy PEG amine (mPEG-NH₂; Nanocs, New York, NY, USA, molecular weight 2000 Da, 50 mg) using N-hydroxysuccinimide (NHS; Alfa Aesar, Ward Hill, MA, USA, 11.5 mg, 0.1 mmol) and 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC; Acros Organics, 19.2 mg, 0.1 mmol) at a molar ratio of 1:1:1:1 in 10 mL of aqueous buffer (pH 7.4, 50 mM phosphate buffer). The mixture was stirred continuously using a magnetic stirrer (300 rpm) at 25 °C overnight under nitrogen atmosphere to prevent oxidation. Following dialysis (MWCO 3.5 kDa, Spectrum Laboratories, Rancho Dominguez, CA, USA) against phosphate-buffered saline (PBS, pH 7.4, 4 × 1 L, changed every 6 h) for 48 h at 4 °C, the product was freeze-dried using a lyophilizer at −50 °C and <0.1 mbar for 48 h. Successful conjugation of PGA-PEG was confirmed by proton ¹H nuclear magnetic resonance (¹H NMR; 400 MHz, Bruker Avance III, Billerica, MA, USA) in deuterium oxide (D₂O). This acid-labile polymer was designed to remain shielded in circulation but de-shield in the acidic MSS CRC tumor microenvironment (pH~6.0–6.5), enabling ligand exposure and payload release through protonation of carboxyl groups (-COOH) on PGA at acidic pH, leading to reduced electrostatic repulsion and polymer detachment.

Lipid-peptide conjugates (DSPE-PEG-C, -S, or -W) were synthesized as described previously [7] with minor modifications. Briefly, DSPE-PEG-maleimide was reacted with each peptide (C-, S-, or W-peptide) in a 1:1 molar ratio overnight (16–18 h) in aqueous buffer at 25 °C under gentle magnetic stirring (200 rpm) protected from light. The maleimide group of DSPE-PEG reacted with the thiol group of cysteine residues in the peptides via Michael addition, forming stable thioether linkages. The conjugation product was dialyzed against PBS (MWCO 3.5 kDa, changed every 8 h) for 24 h at 4 °C and freeze-dried as described above. Successful conjugation was confirmed by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry (Bruker, Billerica, MA, USA) in positive ion mode using α-cyano-4-hydroxycinnamic acid (MedChemExpress, Monmouth Junction, NJ, USA) as the matrix.

The SLNs were prepared using a modified emulsification-solvent evaporation method combined with electrostatic adsorption. Briefly, monostearin (glyceryl monostearate), cholesterol, and DOTAP were dissolved in methanol to form a lipid phase at 65 °C under magnetic stirring (400 rpm). Subsequently, CB, R, and DSPE-PEG-peptide conjugates were added sequentially to the lipid phase and stirred for 30 min at 65 °C to ensure complete dissolution and molecular mixing, followed by the addition of Tween^® 80 (0.1% w/v) and mixing for 1 h. The anionic miR was incorporated by dropwise addition via a micropipette into the pre-formed SLN dispersion under gentle stirring to allow electrostatic complexation. For the preparation of pH-responsive CB + miR + R/PGA-SLN-CSW, the pH-sensitive polymer PGA-PEG (0.1% w/v in PBS pH 7.4) was added dropwise to the pre-formed SLN suspension (containing CB, miR, R, and peptide conjugates) and mixed at 25 °C for 2 h under gentle magnetic stirring to allow polymer adsorption onto the nanoparticle surface.

The particle size, polydispersity index (PDI), and zeta potential of the nanoparticles were measured by dynamic light scattering (DLS) and laser Doppler electrophoresis using a Zetasizer Nano ZS (Malvern Instruments, Worcestershire, UK) equipped with a 633 nm He-Ne laser at 25 °C and a detection angle of 173° (backscattering mode). Each sample was diluted 1:20 in PBS (pH 7.4, filtered through 0.22 μm membrane) and equilibrated for 2 min before measurement, then analyzed in triplicate (n = 3) with 12 sub-runs per measurement. The results were expressed as mean ± standard deviation. The refractive index was set to 1.47 for lipids and 1.33 for PBS. Size distribution was analyzed using the intensity-weighted distribution (Z-average) and reported as mean diameter ± SD. To determine the EE% and DL% of CB- or miR-loaded nanoparticles, 1 mL of nanoparticle suspension was ultracentrifuged at 15,000 rpm for 60 min at 4 °C using a Beckman Coulter ultracentrifuge. The supernatant (containing free unencapsulated drug) was carefully removed without disturbing the pellet, and the resulting pellets were washed once with cold PBS (500 μL), re-centrifuged under the same conditions, and then lysed in 0.5% Triton X-100 (500 μL, incubated at 37 °C for 30 min with vortexing every 10 min to ensure complete lysis of nanoparticles). Following centrifugation at 15,000 rpm for 30 min at 4 °C to pellet debris, the concentrations of CB and miR in both the supernatant (free drug) and the lysed pellet (encapsulated drug) were quantified using a UV/Vis spectrophotometer (Ultrospec 8000, Biochrom, Holliston, MA, USA) at wavelengths of 280 nm for CB and a NanoDrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) at 260 nm for miR, respectively. The EE% and DL% were calculated according to standard equations.EE% = [(W_e − W_f)/W_e] × 100%(1)DL% = [(W_e − W_f)/W_t] × 100%(2) where W_e represents the initial weight of CB or miR, W_f denotes the weight of CB or miR in the filtrate, and W_t corresponds to the total nanoparticle weight.

The morphological features of the nanoparticles were examined by transmission electron microscopy (TEM; JEM-1400 PLUS, JEOL, Tokyo, Japan). For TEM imaging, nanoparticle suspensions (10 μL, diluted 1:100 in deionized water) were deposited on carbon-coated copper grids (300 mesh, Electron Microscopy Sciences, Hatfield, PA, USA), fixed by adding 2% glutaraldehyde (5 μL) for 5 min, and then negatively stained with 2% uranyl acetate (10 μL for 1 min). Multiple fields (at least 10 random fields per sample) were examined, and representative images were captured. To evaluate long-term stability, particle size, PDI, and zeta potential were monitored at predetermined time points over a 24-week period under controlled storage conditions (4 °C in amber glass vials, protected from light, n = 3 independent batches). At each time point, samples were brought to room temperature, gently mixed by inversion (10 times), and analyzed immediately. Physical stability was defined as a <10% change in particle size and <0.05 increase in PDI.

To investigate the acid-sensitive behavior and confirm the de-shielding mechanism under mildly acidic conditions, particle size and zeta potential of CB- and/or miR-loaded formulations were measured at pH 7.4 and 6.0 using a Zetasizer Nano-ZS (Malvern Instruments).

Murine colorectal carcinoma CT-26 cells (MSS phenotype, derived from BALB/c mice, ATCC catalog CRL-2638), RAW 264.7 murine macrophage (derived from BALB/c mice, ATCC catalog TIB-71), and IEC-6 rat intestinal epithelial cells (non-transformed normal intestinal epithelial cells, ATCC catalog CRL-1592) were obtained from the Bioresource Collection and Research Center (Hsinchu, Taiwan). CT-26 and IEC-6 cells were cultured in RPMI-1640 medium, while RAW 264.7 cells were cultured in DMEM, all supplemented with 10% fetal bovine serum (FBS; Gibco) and 1% penicillin–streptomycin–glutamine (PSG; Corning, NY, USA). The CT-26 line was explicitly chosen to model the poor T cell infiltration and TAM-rich environment characteristic of MSS CRC.

Confocal laser scanning microscopy (CLSM) imaging was used to detect PD-L1/EGFR dual-targeting and ER localization in the MSS CRC context. CB was fluorescently labeled with 1,1′-dioctadecyl-3,3,3′,3′-tetramethyl indocarbocyanine perchlorate (DiI) for visualization. For trafficking, CT-26 cells were treated for 1, 3, and 24 h, stained for nuclei and ER (CytoPainter ER Staining Kit; Abcam, Cambridge, UK), and immunolabeled for EGFR and PD-L1. Briefly, cells were fixed with 4% paraformaldehyde for 15 min, permeabilized with 0.1% Triton X-100 in PBS for 10 min, blocked with 5% bovine serum albumin (BSA) in PBS for 1 h at room temperature, and incubated overnight at 4 °C with primary antibodies: rabbit anti-EGFR (1:200) and mouse anti-PD-L1 (1:200). After washing three times with PBS-Tween 20 (0.05%), cells were incubated with Alexa Fluor 488-conjugated goat anti-rabbit IgG (1:500) and Alexa Fluor 647-conjugated goat anti-mouse IgG (1:500) for 1 h at room temperature in the dark. Imaging was performed with an Olympus CLSM (FV10i, Tokyo, Japan). For the internalization of DiI-CB and FAM-miR, CT-26 and IEC-6 cells were incubated with DiI-CB (200 nM), FAM-miR (100 nM), and/or R (1.5 µM) formulations for 24 h. Uptake was quantified by flow cytometry (FACSCalibur, BD Biosciences, San Jose, CA, USA) using 488 nm (FAM detection, FL1 channel) and 561 nm (DiI detection, FL2 channel) lasers. A minimum of 10,000 gated events (based on forward scatter and side scatter to exclude debris) were acquired per sample. Data were analyzed using FlowJo v10 software (FlowJo LLC, Ashland, OR, USA).

CT-26 and IEC-6 cells were treated with different miR-, CB-, and/or R-loaded formulations. Cells were then fixed with 1% trichloroacetic acid (TCA) at 4 °C for 1 h. Fixed cells were stained with 50 μL per well of 0.04% SRB (sulforhodamine B, Sigma-Aldrich, Saint Louis, MO, USA) dissolved in 1% acetic acid for 30 min at room temperature with gentle shaking, and washed with 1% acetic acid. After air drying, 10 mM Tris base was added to each well, and absorbance at 540 nm was measured using an ELISA reader (TECAN, Männedorf, Switzerland). To assess synergistic effects, CT-26 cells were treated with varying concentrations of miR and/or CB. The combination index (CI) was analyzed using CompuSyn software, v. 1.0 (Paramus, NJ, USA) from SRB-derived dose–response data following the Chou-Talalay method. CI values were interpreted as follows: CI < 0.7 (strong synergism), 0.7–0.85 (moderate synergism), 0.85–0.9 (slight synergism), 0.9–1.1 (additive effect), and >1.1 (antagonism). All experiments were performed in triplicate.

Following a 24 h treatment with various formulations (CB: 200 nM; miR: 100 nM; R: 1.5 μM) or vehicle control, CT-26 cells seeded at 5 × 10⁴ cells per well in 8-well chambered coverslips were washed twice with PBS and stained with 4′,6-diamidino-2-phenylindole (DAPI; Millipore Sigma, Burlington, MA, USA) for nuclear counterstaining and 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA; Millipore Sigma) for ROS detection (20 min incubation for both stains). Cells were fixed with 4% paraformaldehyde for 10 min, and images were subsequently acquired using CLSM (Olympus FV10i). For quantitative analysis, intracellular DCF fluorescence was then measured via flow cytometry (FACSCalibur).

To quantify apoptotic and necrotic cell populations, an Annexin V-fluorescein isothiocyanate (FITC)/propidium iodide (PI) apoptosis detection kit (Strong Biotech Corporation, Taipei, Taiwan) was utilized after 24 h of treatment. Cell suspension (~1 × 10⁵ cells) was transferred to flow cytometry tubes and incubated with Annexin V-FITC and PI (50 μg/mL) for 15 min at room temperature in the dark. Samples were analyzed within 1 h. Fluorescence was measured using FL1 (FITC, 530 nm) and FL3 (PI, 670 nm) channels, acquiring 10,000 events per sample. Cell populations were classified as follows: viable (Annexin V⁻/PI⁻), early apoptotic (Annexin V⁺/PI⁻), late apoptotic (Annexin V⁺/PI⁺), and necrotic (Annexin V⁻/PI⁺). Data were analyzed using FlowJo v10 software with appropriate compensation settings. The results were expressed as percentage of cells in each quadrant. For cell cycle analysis, CT-26 cells were seeded at 3 × 10⁵ cells per well in 6-well plates, cultured overnight, and treated for 24 h as described above. Cells were then fixed with 70% ice-cold ethanol at −20 °C overnight and subsequently stained with PI (50 μg/mL), 0.1% Triton X-100, and RNase A (Millipore Sigma) for 30 min at 37 °C in the dark. Cell cycle distribution across G0/G1, S, and G2/M phases was then assessed via flow cytometry (FACSCalibur) using the FL3 channel (670 nm).

Wound closure in CT-26 monolayers was quantified after 15 h treatment using Ibidi culture inserts (Culture-Insert 2 Well in μ-Dish 35 mm, Ibidi GmbH, Gräfelfing, Germany). Briefly, CT-26 cells were seeded at 3 × 10⁴ cells per well of the insert in 70 μL of complete medium and cultured overnight until confluent monolayers formed on both sides of the insert. The insert was then gently removed using sterile forceps, creating a defined 500 μm gap. Serum-free medium containing various formulations (CB: 200 nM; miR: 100 nM; R: 1.5 μM) or vehicle control was added. Cell migration was visualized before and after treatment using light microscopy (Olympus IX70, Tokyo, Japan), and the migration area was quantified using ImageJ software, v. 1.53t.Relative migration area (% of area at 0 h) = 100% − [blank area_(15h)/blank area_(0h) × 100%](3)

After 24 h treatment with various formulations (CB: 200 nM; miR: 100 nM; R: 1.5 μM) or vehicle control, total cellular proteins were extracted using radioimmunoprecipitation assay (RIPA) buffer. Protein concentrations were quantified using a BCA assay (Thermo Fisher, Waltham, MA, USA). Equal amounts of protein (30–50 μg per lane, normalized across samples) were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred to polyvinylidene fluoride (PVDF) membranes (Bio-Rad, Hercules, CA, USA), and probed with specific primary antibodies followed by horseradish peroxidase (HRP)-conjugated IgG secondary antibodies (JacksonImmunoResearch, West Grove, PA, USA). Protein bands were visualized using Immobilon enhanced chemiluminescence (Millipore, Billerica, MA, USA) and imaged with a Luminescence Imaging System (Amersham Imager 680; GE Healthcare, Chicago, IL, USA).

To evaluate the effect of formulations on TAM polarization, RAW 264.7 macrophages were co-cultured with CT-26 cells in cell culture inserts (Greiner Bio-One, Kremsmünster, Austria). Briefly, RAW 264.7 cells (1 × 10⁵ cells) were seeded in the upper insert chamber in 1.5 mL of DMEM complete medium, and CT-26 cells (2 × 10⁵ cells) were seeded in the lower well in 2.5 mL of RPMI-1640 complete medium. After 24 h of co-culture to establish cell–cell communication, CT-26 cells were treated with various formulations (CB: 200 nM; miR: 100 nM; R: 1.5 μM) for an additional 48 h at 37 °C. RAW 264.7 cells, CT-26 cells, and conditioned medium were collected for Western blot analysis.

Cytokine and HMGB1 concentrations were determined using ELISA kits (BioLegend, San Diego, CA, USA). Each well was coated with 200 μL of capture antibody and incubated overnight, followed by blocking with assay diluent buffer for 1 h. Subsequently, standards and conditioned media were introduced and incubated for 2 h, followed by sequential incubation with detection antibodies (HMGB1, IL-10, TNF-α, TGF-β, IFN-γ), avidin-HRP, and substrate solutions (tetramethylbenzidine (TMB), provided in kit) for 15–30 min at room temperature in the dark until color development. Absorbance at 450 nm was measured using a TECAN ELISA reader to quantify protein levels.

Following 24 h of treatment, cells were stained with ER-Tracker Green to visualize the endoplasmic reticulum, then fixed and blocked prior to antibody labeling. Primary antibodies against CRT or HMGB1 were applied and incubated overnight at 4 °C. The next day, Alexa Fluor 680- and Cy5.5-conjugated secondary antibodies were introduced to detect the respective targets. The cellular distribution of CRT and HMGB1 was assessed using CLSM on an Olympus FV10i system.

ATP levels were assessed using an ATP detection assay kit (Cayman, Ann Arbor, MI, USA) based on the luciferin-luciferase bioluminescence reaction. Intracellular ATP was quantified from lysed cells, whereas extracellular ATP was measured from conditioned media. Luminescence signals were recorded using a TECAN ELISA reader to evaluate ATP concentrations. The results are presented as percentage changes relative to untreated controls.

To assess hemocompatibility, rat red blood cells (RBCs) were collected into heparinized tubes (10 U/mL), washed three times with sterile PBS (pH 7.4) after centrifugation at 1000× g for 5 min at 4 °C, and resuspended in PBS to a final concentration of 2% (v/v). RBC suspensions (200 μL) were incubated with 200 μL of the formulations or 0.5% Triton X-100 (positive control, 100% hemolysis) or the PBS control (negative control, 0% hemolysis) for 24 h. Hemolysis was assessed by incubating supernatants from lysed RBCs with Drabkin's reagent (Millipore Sigma) at 25 °C for 15 min in equal volume proportions (1:1, v/v). The intensity of the chromogenic signal was determined by measuring the optical absorbance at 540 nm with a TECAN microplate spectrophotometer.

BALB/c mice weighing approximately 20 g and aged 4–6 weeks were sourced from the National Laboratory Animal Center (Taipei, Taiwan). All in vivo procedures strictly adhered to ethical guidelines approved by the Institutional Animal Care and Use Committee (IACUC) of National Yang Ming Chiao Tung University (approval number: 1100411, 19 April 2021). Mice were subcutaneously implanted in the right flank with CT-26 cells (3 × 10⁶ cells suspended in 100 μL of PBS). When tumors reached ~60 mm³, mice (n = 3 per group) received the designated treatments twice weekly for 14 days via tail vein injection. The treatment groups were: (1) saline control (CTR), (2) miR/SLN-CSW, (3) CB, (4) CB/SLN-CSW, (5) CB + miR/SLN-CSW, (6) CB + miR + R/SLN-CSW, and (7) CB + miR + R/PGA-SLN-CSW, administered at 10 mg/kg for CB, 1.25 mg/kg for miR-142, and 6 mg/kg for R. Tumor dimensions were measured using digital calipers every 3–4 days (length and width in mm), and tumor volume was calculated using the modified ellipsoid formula:V = (L × W²)/2(4) where L is the longest diameter and W is the shortest perpendicular diameter.

Tumor localization was evaluated using PET/MRI. Mice were fasted for at least 6 h before intravenous administration of [¹⁸F]-2-deoxy-2-fluoro-D-glucose (¹⁸F-FDG; 0.282 mCi per mouse, equivalent to approximately 10.43 MBq, diluted in 100 μL of sterile saline) on day 15. PET scans were acquired at 45 min post-injection using a Bruker 7T PET/MRI system (Bruker, Billerica, MA, USA), followed by MRI scans for anatomical reference and attenuation correction. Analysis of PET and MRI datasets was performed using AMIDE, an open-source imaging software, v. 1.0.6 (SourceForge, Ames, IA, USA).

Tumors and spleens from treated CT-26 tumor-bearing mice were harvested on day 15 (24 h after the final treatment), immediately placed in cold RPMI-1640 medium on ice, and processed into single-cell suspensions for characterization of tumor-infiltrating and immune cell populations. Single-cell suspensions (1 × 10⁶ cells in 100 μL of flow cytometry buffer) were stained with fluorescein-conjugated antibodies for 20 min in the dark. Flow cytometric analysis of dual-labeled cells stained with FITC- and PE-conjugated immune antibodies was conducted using FlowJo v10 software (FlowJo LLC, Ashland, OR, USA).

Circulating serum cytokine and chemokine levels were measured using a Bio-Plex multiplex assay (Bio-Rad, Hercules, CA, USA). Serum samples (50 μL, diluted 1:4 in assay diluent) and standards (provided in kit, creating 7-point standard curves) were incubated with antibody-coupled magnetic beads in a 96-well plate for 30 min at room temperature with shaking (300 rpm), washed three times using a magnetic washer, incubated with detection antibodies for 30 min, washed again, and incubated with streptavidin-phycoerythrin (SA-PE) for 10 min. After final washing, the beads were resuspended in assay buffer and analyzed using the Bio-Plex 200 system, which identifies individual beads by their spectral signature and quantifies bound PE fluorescence as a proportion of the analyte concentration. Analyzed targets included IL-1β, IL-2, IL-4, IL-5, IL-8 (CXCL-1), IL-10, IL-12, IL-17, interferon-gamma (IFN-γ), tumor necrosis factor-alpha (TNF-α), transforming growth factor-beta (TGF-β), and G-CSF.

Major organs including tumors, liver, kidneys, heart, spleen, and intestines (jejunum) were fixed in 4% paraformaldehyde, paraffin-embedded, sectioned into 4–5 μm thick sections, and stained with hematoxylin and eosin (HE). Stained samples were imaged using an Olympus VS120 slide scanner. Tumor apoptosis was assessed using a TUNEL assay (In Situ Cell Death Detection Kit, Roche, Basel, Switzerland), with imaging conducted using an Olympus FV10i confocal microscope with 488 nm excitation for TUNEL (green) and 405 nm for DAPI (blue), using a 20× objective. Tumor and spleen sections were subjected to IHC staining using primary antibodies targeting CD4, CD8, FoxP3, and PD-L1, followed by incubation with HRP-conjugated secondary antibodies. Signal development was performed using 3,3′-diaminobenzidine (DAB) with hematoxylin counterstaining. Stained sections were scanned using an Olympus VS120 slide scanner at 20× magnification, and digital images were analyzed. Quantitative assessment of DAB staining intensity was performed using Image Pro Plus 4.0 software.

Following the 14-day treatment regimen, primary tumors were surgically excised. Seven days post-surgery, CT-26 cells (1 × 10⁶ cells in 100 μL of PBS) were re-implanted into the contralateral flank of the same mice under brief isoflurane anesthesia to assess long-term immunological memory and protective immunity against tumor rechallenge. Tumor growth was monitored by caliper measurements biweekly over an additional 2 weeks. At the end of the study, the mice were sacrificed and their spleens were harvested, minced, and washed with PBS. Single-cell suspensions were stained with fluorescein-conjugated antibodies targeting T memory cell markers and analyzed via flow cytometry using FlowJo v10 software.

Whole-blood samples were obtained via retro-orbital bleeding 48 h following the final treatment. Blood cell counts, including RBCs, white blood cells (WBCs), and platelets (PLTs), were quantified using a hematology analyzer (Sysmex XT-1800iv; Sysmex, Kobe, Japan). Serum was separated by centrifugation (1500 rpm, 20 min) and analyzed for biomarkers of hepatic (glutamic-pyruvic transaminase, GPT)), cardiac (creatine kinase-MB, CK-MB), and renal (blood urea nitrogen, BUN) function using specific assay kits (Fujifilm, Tokyo, Japan) and a Fuji Dri-Chem 7000V clinical chemistry analyzer (Fujifilm Corp., Tokyo, Japan).

To quantify the tissue distribution of CB at 24 h post-injection, mice (n = 3 per group) were euthanized by CO₂ asphyxiation, and tissue samples (tumors, liver, spleen, kidneys, and heart) were harvested, rinsed briefly in cold PBS to remove blood, blotted dry on filter paper, weighed (50–200 mg), and immediately homogenized with ice-cold methanol (~1 mL per 100 mg tissue) and distilled water (at a ratio of methanol:water:tissue = 4:1:1 v/v/w) using a tissue homogenizer, and then incubated on ice for 15 min to allow for complete protein precipitation and CB extraction. Following centrifugation of tissue homogenates at 12,000× g for 15 min at 4 °C, the supernatant containing CB was harvested and appropriately diluted with methanol. CB levels were then quantified spectrophotometrically at 280 nm using the Ultrospec 8000 PC spectrophotometer (Biochrom, Holliston, MA, USA) at a wavelength of 280 nm after the appropriate dilution with methanol.

Comparative statistical analyses were performed using Student's t-test, with results expressed as means ± standard deviations (SDs). Differences were considered statistically significant at * p < 0.05, with increasing levels of significance denoted by ** p < 0.01 and *** p < 0.001.

¹H NMR spectroscopy confirmed the successful conjugation of PEG–PGA, with characteristic methylene proton peaks at 3.62 ppm and methyl proton peaks at 3.35 ppm, alongside methylene and methylidyne signals from PGA at 2.30 ppm and 4.15 ppm, respectively (Figure 1A). DSPE-PEG-peptides were synthesized via thiol–maleimide coupling of DSPE-PEG-maleimide with cysteine-containing C, S, or W peptides. Mass spectrometry verified molecular weights consistent with predicted values, confirming peptide conjugation (Figure S1A–C). Dynamic light scattering (DLS) revealed hydrodynamic diameters ≤ 200 nm for CB/SLN-CSW, CB + miR/SLN-CSW, and CB + miR + R/SLN-CSW (Figure 1B, Table 1), falling within the range considered optimal for passive and active accumulation in immune-excluded, MSS CRC tumors [7]. All formulations of SLN-CSW maintained narrow PDI values between 0.14 and 0.16, reflecting consistent homogeneity across batches. According to Table 1, the EE% remained high, ranging from 82.31–83.65% for CB, 89.54–89.56% for miR, and 85.18% for R. The corresponding DL% values were 19.79–20.99% for CB, 19.11–19.23% for miR, and 18.92% for R. Spherical morphology and uniform particle size distribution were confirmed by TEM, with no observable aggregation (Figure 1B). Each batch demonstrated high reproducibility (<5% variation) and excellent long-term stability at 4 °C over 24 weeks, showing negligible changes in size, zeta potential, or PDI (Figure 1C,D). These characteristics support the suitability and reproducibility of the formulations for preclinical MSS CRC models requiring extended formulation integrity.

The hydrodynamic diameter of CB + miR + R/PGA-SLN-CSW decreased from 195.26 ± 1.45 nm (pH 7.4) to 173.34 ± 3.89 nm (pH 6.0) and showed a zeta potential shift from −3.49 ± 0.23 mV to 19.78 ± 1.21 mV, consistent with PGA–PEG de-shielding in acidic environments (Figure 2A,B). This acid-triggered de-shielding leads to a more compact nanoparticle surface structure, which explains the reduction in hydrodynamic size and the marked increase in positive surface charge. TEM images showing uniform, non-aggregated spheres further support the stability and structural integrity of the nanoparticles under both pH conditions. Flow cytometry indicated enhanced DiI-CB uptake in CT-26 cells with SLN-CSW compared to unmodified SLNs or free drug, and further uptake enhancement at pH 6.0 for PGA-coated particles (Figure 2C). Similar trends were observed for FAM-miR delivery (Figure 2D). Uptake in IEC-6 cells at pH 7.4 was minimal (Figure S2A,B), suggesting tumor selectivity. Hemolysis assays confirmed non-lytic behavior for SLNs in contrast to free CB (11% hemolysis) (Figure S2C). Given that CT-26 cells recapitulate the limited immune infiltration characteristic of MSS CRC, with elevated PD-L1 and EGFR expression (Figure S3A), CLSM revealed that DiI-CB + miR/SLN-CSW rapidly associated with PD-L1 and EGFR within 1 h in CT-26 cells, mediated by W and C peptides. Subsequent trafficking to the ER was observed at 3 h via the S peptide, with intracellular retention sustained for up to 24 h (Figure 2E,F). Both PD-L1 and EGFR engagement were evident within 1 h, while FAM-miR + CB + R/SLN-CSW exhibited cytoplasmic localization by 8 h and remained detectable up to 24 h in CT-26 cells (Figure S3B,C). In IEC-6 cells, the PGA coating reduced CB-induced toxicity (Figure 2G). In CT-26 cells, CB/SLN-CSW increased cytotoxicity relative to free CB, and CB + miR-142 showed a combination index (CI) = 0.38. Treatment with triple-loaded SLN-CSW significantly reduced CT-26 cell viability to 46.68 ± 4.94%, with a further decline observed under acidic conditions (pH 6.0; Figure 2H and Figure S4A,B), aligning with the enhanced drug delivery capacity required to surmount the drug penetration barriers typically associated with MSS tumors.

VCP is frequently overexpressed in CRC and associated with metastasis and progression [11], including in immune-cold MSS subtypes where proteostasis reliance is heightened. Its inhibition disrupts proteostasis by promoting misfolded protein accumulation in the ER, triggering ERS as indicated by Bip upregulation [12]. This activates the IRE1, ATF6, and PERK pathways, which regulate downstream proteins such as CHOP, Casp-3, Casp-9, and autophagy-related mediators, contributing to apoptotic cell death [7,13] (Scheme 2). Given the association of VCP overexpression with CRC progression [12], Western blot analyses showed that CB-containing formulations reduced VCP levels and elevated Bip, ATF6, and K48-linked ubiquitin, particularly in the CB + miR/SLN-CSW and CB + miR + R/SLN-CSW groups (Figure 3A). IRE1α signaling promoted autophagic responses (Atg5, Beclin, LC3II) through two branches: one involving XBP1s and LC3II, and the other mediated by JNK and Beclin (Figure 3B,C). Separately, the PERK pathway was distinctly activated, leading to phosphorylation of eIF2α (Figure 3C). This activation increased the translation of ATF4, which subsequently induced the expression of CHOP (Figure 3C). This pathway contributed to the upregulation of pro-apoptotic proteins including Bim, Casp-3, and Casp-9 (Figure 3C). These molecular events were also accompanied by a reduction in Bcl-2 expression (Figure 3B), reinforcing the dual induction of apoptosis and autophagy (Figure 3B,C). Annexin V/PI staining confirmed maximal apoptosis with CB + miR + R/SLN-CSW (Figure 3D). These findings indicated that targeted VCP inhibition not only disrupted proteostasis in MSS CRC cells but also initiated ERS-driven death pathways capable of reshaping the immune-resistant TME.

CT26 cells displayed elevated EGFR expression (Figure S3A) within an immune-inert microenvironment. A recent study also reported upregulation of EGFR and principal effectors of the PI3K/AKT/mTOR signaling cascade, consistent with hyperproliferative activity and enhanced survival signaling [1]. Treatment with CB-loaded SLNs (Figure S5) suppressed the oncogenic PI3K, p-AKT, and p-mTOR signaling cascade (Figure S5). The combination SLNs containing CB, miR, and R further downregulated key cell cycle regulators (Figure 3E). Notably, mitotic progression factors including securin, FBXO5, and CDC20 were significantly reduced, disrupting mitotic checkpoint control (Figure 3E). In parallel, the expression of proliferation drivers such as survivin, c-Myc, and cyclin D1 also decreased (Figure 3E), contributing to the induction of G0/G1 cell cycle arrest (Figure 3F). The most pronounced effects were observed in CB + miR + R/SLN-CSW-treated cells (Figure 3F). EMT reversal was evidenced by increased E-cadherin and reduced Smad2/3, Snail, and N-cadherin levels (Figure 3G), correlating with the impaired migration observed in the wound-healing assays (Figure 3H).

This study provided mechanistic insight into CB as a type II ICD inducer. In CT-26 cells, which recapitulate these MSS-associated immune barriers, CB directly activated ERS to initiate ICD, as confirmed by Western blot analysis (Figure 3A–C). Consistent with the stress-prone phenotype of MSS CRC, VCP ablation triggered ROS production, validated by DCFH-DA staining quantification via flow cytometry and CLSM (Figure S6A,B). This led to the accumulation and release of DAMPs—CRT, HMGB1, and ATP—from tumor cells (Figure 4A–C and Figure S7), events that can help breach the immune-exclusive TME. These alterations coincided with a decrease in the secretion and/or expression of immunosuppressive molecules including MSR1, CD206, IL-10, TGF-β, Arg1, and PD-L1 (Figure 4D–F), and increased proinflammatory mediators (TNF-α, IFN-γ, IL-6, iNOS, COX-2) (Figure 4G–I), indicating restoration of an inflamed phenotype. Notably, CB + miR + R/SLN-CSW markedly suppressed IL-10 and TGF-β (Figure 4E,F) while increasing IFN-γ and TNF-α (Figure 4H,I), changes particularly relevant to reversing the MSS CRC immunosuppressive milieu.

Considering the frequent dysfunction of dendritic cell activation and limited infiltration of effector T cells in MSS CRC, the therapeutic impact of formulations containing CB alone or in combination with immunomodulatory miR-142 and R was examined in CT-26 tumor-bearing mice (Figure 5 and Figure S8–S11). All SLN-CSW-based treatments, especially CB + miR + R, promoted DC maturation and expansion of CD11c⁺MHCII⁺ populations, with the PGA-coated formulation yielding the most substantial effect (Figure 5A–D), consistent with ICD-driven antigen presentation (Figure 4A–C). CB + miR + R/SLN-CSW, with or without PGA, markedly increased intratumoral infiltration of CD4⁺ helper and CD8⁺ cytotoxic T cells (Figure 5E–H and Figure S8), a hallmark typically diminished in MSS CRC, while concomitantly reducing regulatory T cells expressing CD25 and FoxP3, and TAMs marked by F4/80 and CD206 (Figures S9–S11). The concurrent increase in CD8⁺ and CD4⁺ T cell populations and reduction in PD-L1 and TAM markers in tumor and spleen tissues (Figure 5I–M and Figure S11A–D) indicates reversal of the TAM-enriched TME. Cytokine analysis demonstrated a transition favoring a proinflammatory milieu, characterized by increased levels of IL-1β, IL-2, IL-12, IFN-γ, and TNF-α, alongside a reduction in immunosuppressive mediators such as TGF-β, IL-4, IL-5, IL-10, and CXCL-1, particularly in the group receiving the triple combination within PGA-SLN-CSW (Figure 5N).

In light of the intrinsic immunoresistance of MSS CRC, achieving tumor regression in CT26-bearing mice sets a demanding benchmark for therapeutic performance. Treatment with the CB, miR, and R formulation delivered via the PGA-SLN-CSW platform significantly reduced tumor volume compared with all other groups (Figure 6A–D), which was attributed to selective activation of the PGA coating in the acidic TME, which enhances drug release and promotes immunomodulatory effects within MSS CRC lesions. PET/MRI demonstrated reduced ¹⁸F-FDG uptake (Figure 6B), and optical imaging confirmed tumor shrinkage (Figure 6C). Histopathology showed extensive TUNEL positivity and pyknosis (Figure 6E), features indicative of apoptosis induction in immune-refractory MSS tumors.

Post-surgical re-challenge in CT-26-bearing mice (Figure 7A) revealed that CB + miR + R/PGA-SLN-CSW limited secondary tumor outgrowth, maintaining the smallest volumes compared to controls (Figure 7B–D). The T cell subset with effector memory function in the spleen (CD62L⁻CD44⁺) progressively expanded following treatment with CB, miR, and R, reaching their highest levels within the PGA-coated formulation group (Figure 7E). In MSS CRC, where immune memory formation is typically compromised by a suppressive TME, these findings highlight the platform's potential to induce durable antitumor immunity capable of preventing recurrence.

The SLN-based formulations exhibited favorable systemic safety. Over the 42-day period, they consistently preserved body weight (Figure 8A), maintained hematological indices near baseline levels (Figure 8B and Figure S12), and mitigated biochemical signs of organ toxicity evident in the free CB-treated group (Figure 8C). Histopathological analysis demonstrated reduced off-target tissue damage compared to free drug administration (Figure S13). Biodistribution studies revealed maximal tumor-specific accumulation alongside minimal hepatic and splenic deposition for PGA-coated SLNs co-loaded with CB, miR, and R (CSW) (Figure 8D), suggesting effective targeted delivery to the immune-cold MSS CRC microenvironment.