Bispecific aptamer-decorated and light-triggered nanoparticles targeting tumor and stromal cells in breast cancer derived organoids: implications for precision phototherapies

4-(1,1,3,3-Tetramethylbutyl)phenyl-polyethylene glycol (Triton X-100), n-hexanol, cyclohexane, ammonium hydroxide solution (28% w/w), hydrogen tetrachloroaurate (III) trihydrate (HAuCl₄·3H₂O), sodium 2-mercaptoethanesulfonate (Mesna), sodium borohydride (NaBH₄), (3-aminopropyl)triethoxysilane (APTES), tetraethoxysilane (TEOS), N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC), N-hydroxysuccinimide (NHS) and phosphate buffer saline tablets were purchased from Sigma-Aldrich (Saint Louis, MO, USA). 11-Triethoxysilylundecanoic acid (95%) and N-(3-triethoxysilyl) propylsuccinic anhydride (94%) were purchased from ABCR (Karlsruhe, Germany). Ultrapure water (Milli-Q, 18 MΩ·cm) was used for the preparation of the aqueous solutions and for all rinses. Phosphate buffer solution (PBS, pH 7.4) was prepared by dissolving one phosphate buffer saline table in 200 mL of Milli-Q water. All other solvents used were of analytical grade.

NH₂-terminated 2′F-Py-containing RNA, CL4, its scrambled sequence (Scr) used as a negative control, and Gint4.T were synthesized by LGC Biosearch Technologies (Risskov-Denmark).

The sequences are as follows:

Before usage, aptamers were heated at 85 °C for 5 min, snap-cooled on ice for 3 min, and allowed to warm up for 10 min to 37 °C, before incubation with the cells for checking appropriate folding, or to room temperature (RT) for nanoparticles conjugation.

Multifunctional gold-core/silica-shell nanoparticles embedding the photosensitizing and luminescent molecule Ir_en [41] were synthesized following a previously reported protocol [17, 18] with slight modifications. Briefly, quaternary water/oil (W/O) microemulsion was prepared by mixing 3.6 mL of Triton X-100, 3.6 mL of n-hexanol, 15 mL of cyclohexane and a water solution consisting of 0.9 mL HAuCl₄·3H₂O (12.75 mM), 0.9 mL Mesna (36.5 mM) and 0.3 mL NaBH₄ (423 mM). Then, Ir_en (7 mg/0.1 mL water) was added to the microemulsion, followed by the addition of 0.010 mL of APTES and 0.150 mL of TEOS. After 30 min, 0.080 mL of ammonium hydroxide solution was added. The mixture was stirred overnight at RT. Afterward, the functionalization of the nanoparticle surface with carboxyl-terminated aliphatic chains, was carried out by addition after 24, 24 + 3 and 48 h of 0.015 mL of 11-Triethoxysilylundecanoic acid. Finally, in order to improve the colloidal stability, the silane coupling agent N-(3-triethoxysilyl) propylsuccinic anhydride (0.015 mL) was added after 48 + 3 h. The mixture was stirred overnight at RT. Then, the microemulsion was broken by addition of isopropanol and water in a volume ratio 1:1:1. Purification steps by centrifugal ultrafiltration (Vivaspin 20 PES, 100,000 MWCO, Sartorius, Gottingen, Germany) allowed the complete removal of all unreacted species. The obtained nanoparticles (Ir_en-AuSiO₂_COOH) were finally dispersed in water to a final volume of 20 mL and filtered by a 200 nm nylon membrane (Sartorius).

Ir_en-AuSiO₂_COOH were conjugated with the selected aptamers through a covalent bond between the carboxyl group (− COOH) of the nanoparticle surface coating agent and the amino group (− NH₂) on the 5′-end of RNA scaffold. 1 mL of EDC (26 mM) was mixed under stirring to 1 mL of Ir_en-AuSiO₂_COOH nanoparticles solution. Then, 1 mL of NHS (24.3 mM) was added to the reaction solution. After 50 min, 3 mL of PBS were added, followed by the addition of 1 mL of an aqueous solution of CL4 (0.280 μM), Scr (0.280 μM) and Gint4.T (0.280 μM), respectively, and of 0.5 mL of CL4 (0.280 μM) and 0.5 mL of Gint4.T (0.280 μM) to get the dual aptamer-decorated preparation. The reaction continued for 24 h. Afterward, aptamer-nanoconjugates Ir_en-AuSiO₂_CL4, Ir_en-AuSiO₂_Scr, Ir_en-AuSiO₂_Gint4.T and Ir_en-AuSiO₂_CL4_Gint4.T were washed by centrifugal filter devices (Vivaspin 20 PES, 100,000 MWCO, Sartorius) to eliminate unconjugated aptamers and unbound reaction components, and concentrated to a final volume of 1 mL, then stored at 4 °C until use. The same procedure, without aptamers addition, was followed to obtain Ir_en-AuSiO₂_COOH/NHS nanoparticles in order to use them as control.

The synthesized Ir_en-AuSiO₂_COOH nanoparticles were characterized by Transmission Electron Microscopy (TEM), Dynamic Light Scattering (DLS) and UV–Vis spectroscopy.

The morphology was observed using a JEOL 2010F transmission electron microscope. The sample was prepared by depositing a drop of a diluted colloidal solution on 200 mesh carbon-coated copper grids. After evaporation of the solvent in air at RT, the nanoparticles were observed at an operating voltage of 80 kV. The hydrodynamic sizes and ζ-potential values were measured with a Zetasizer Nano ZS (Malvern) instrument (632.8 nm, 4 mW HeNe gas laser, avalanche photodiode detector, 173° detection), using glass cuvettes (1 × 1 cm) and disposable folded capillary zeta cells, respectively, and the results expressed as average of three measurements. Extinction and excitation/emission spectra were recorded with a PerkinElmer Lambda 900 spectrophotometer and Perkin-Elmer LS-50B spectrofluorometer, using quartz cuvettes with a light path 1 × 0.4 cm. The calculation of the nanoparticles concentration (number of nanoparticles per mL) and the yield of encapsulation of Ir_en (number of Ir_en molecules per nanoparticle) were carried out according to the procedure previously reported [42]. Absorption spectra were acquired over time (over one month) to monitor the stability of the nanostructures in the aqueous medium.

To validate the presence of the target-specific aptamers on the nanoparticles surface, Ir_en-AuSiO₂_CL4, Ir_en-AuSiO₂_Scr, Ir_en-AuSiO₂_Gint4.T and Ir_en-AuSiO₂_CL4_Gint4.T were characterized by DLS and UV–Vis spectroscopy techniques.

To determine the amount of aptamer conjugated to the nanoparticles, reverse transcription-quantitative polymerase chain reaction (RT-qPCR) based assay was performed as previously described [40]. Briefly, Ir_en-AuSiO₂_CL4, Ir_en-AuSiO₂_Scr, Ir_en-AuSiO₂_Gint4.T and Ir_en-AuSiO₂_CL4_Gint4.T were incubated with aptamer specific 3′ primers, heated at 65 °C for 5 min and annealed at 22 °C for 5 min. RNA was reverse transcribed using Tetro Reverse Transcriptase (Bioline London, UK) at 42 °C for 15 min followed by an extension at 50 °C for 30 min and enzyme inactivation at 85 °C for 5 min. The products from the reverse transcription reaction were subjected to qPCR amplification. The sequences of aptamer-specific 5′ and 3′ primers for Gint4.T and Scr, and CL4 were reported in [40, 43], respectively. The conjugation efficiency was calculated as pmoles of aptamer_conjugated/ aptamer_total (%).

Human MES-TNBC MDA-MB-231 and BT-549, luminal B/HER2-positive breast cancer BT-474, luminal A/estrogen receptor and progesterone receptor-positive breast cancer MCF7, and epidermoid carcinoma A431 cell lines were purchased from the American Type Culture Collection (ATCC, Manassas, VA) and grown as previously reported [44]. Green fluorescent protein (GFP)-labeled BT-549 cells (BT-549-GFP) was produced as previously described [30] and grown in Roswell Park Memorial Institute-1640 medium (Sigma-Aldrich, Milan, Italy) supplemented with 10% fetal bovine serum (Sigma-Aldrich). Human adipose MSCs were purchased from Sigma-Aldrich (SCC038) and grown in Human Mesenchymal-XF Expansion Medium (Sigma-Aldrich) in a humidified incubator in 5% CO2 at 37 °C.

For 3D heterotypic tumor spheroids, 2 × 10³ cancer cells were mixed with 8 × 10³ MSCs (ratio 1:4) and seeded in 24-ultralow attachment plates (Corning Incorporate, Corning, NY) in Dulbecco's Modified Eagle Medium (DMEM)/F-12 medium (D8437 Sigma-Aldrich), supplemented with 2% Matrigel Basement Membrane Matrix Growth Factor Reduced (Corning Incorporate), B27 (1:50, Gibco™ by Invitrogen, Carlsbad, CA), 20 ng/ml basic fibroblast growth factor (Sigma-Aldrich) and 10 ng/ml epidermal growth factor (Sigma-Aldrich). For homotypic cultures, 2 × 10³ cancer cells or 8 × 10³ MSCs were seeded alone. Spheroid formation was checked daily using a phase-contrast microscopy (Leica DMI3000 B apparatus), images were captured and the diameter of spheroids was measured.

Breast cancer samples from three patients who underwent surgery at the National Cancer Institute ″Fondazione Giovanni Pascale″ of Naples, Italy, were enrolled in this study. This study was approved by the ethics committee of INT Pascale (Prot. CEI/390/15) and all the patients provided written informed consent. The tissue samples were collected for histopathological diagnosis and an aliquot was stored in the Institutional Biobank (BBI). Immunohistochemical staining was done on biobank histological formalin-fixated and paraffin-embedded tissue samples slides (4 μm), as previously reported [39], by using primary antibodies against PDGFRβ (dilution 1:50, rabbit monoclonal antibody, clone 28E1, Cell Signaling Technology Inc., Danvers, MA), and EGFR (dilution 1:100, rabbit monoclonal antibody, clone D38B1, Cell Signaling Technology Inc.). Results were interpreted using a light microscope. Ten fields on each of two cores and at least > 500 cells were analyzed for each sample. Two pathologists independently evaluated the intensity, extent and subcellular distribution of the immunostaining. Receptor expression was interpreted as positive when membranous and/or cytoplasmic (PDGFRβ) or membranous (EGFR) staining was observed.

Organoid development methods followed a previous reported procedure study [45]. Briefly, fresh tumor specimens were collected and transported in working medium (WM), consisting of DMEM-F/12 medium (Sigma-Aldrich) supplemented with 1 × Antibiotic–Antimycotic (Gibco™ by Invitrogen) and 10 μg/mL gentamycin (Euroclone, Milan, Italy), minced mechanically and then digested in WM (10 mL/g of tissue) supplemented with 2 mg/mL Collagenase Type II (Gibco™ by Invitrogen), for 16 h at 37 °C under shaking. The digested tissue was filtered through a 100 μm cell strainers (Corning Incorporate) followed by a 40 μm cell strainer (Corning Incorporate) to separate organoids from single cells, and then tumor fragments retained by the cell strainer were washed in WM. The suspension containing tumor organoids was centrifuged 5 min at 500 × g and the organoid pellet was plated in 24-ultralow attachment plates (Corning Incorporate) in the culture medium previously reported [45]. After 24 h, the organoids were centrifuged again at 500 × g and finally resuspended in 90% Matrigel Basement Membrane Matrix Growth Factor Reduced (Corning) diluted in WM and 35 μL drops were allowed to solidify in the inverted 24-well plates for 20 min at 37 °C. Organoids-Matrigel drops were then covered with 500 μL of culture medium and transferred into incubator for culturing. The culture medium was replaced every 3–4 days and organoids were passaged every 2–3 weeks at a split ratio of 1:2–1:3 using mechanical dissociation by pipetting or enzymatic digestion using TripLE Express (Gibco™ by Invitrogen) for 5–15 min at 37 °C.

Cell lysates’ preparation and immunoblot analyses were performed as previously reported [35]. The filters were probed with the indicated primary antibodies: anti-EGFR, anti-PDGFRβ, anti-vinculin and anti-α-tubulin (Cell Signaling Technology Inc.). The blots shown are representative of at least three independent experiments.

PDCOs were mechanically and enzymatically disaggregated into a single-cell suspension and then incubated with anti-EGFR or anti-PDGFRβ (dilution 1:50, R&D system, Minneapolis, MN) primary antibody diluted in Dulbecco’s phosphate buffered saline (DPBS)/BlockAid™ blocking solution (Invitrogen), for 15 min at RT. After three washes with DPBS, cells were incubated with Alexa Fluor 488 Anti-Goat (Invitrogen), washed three times in DPBS, suspended in 500 μl DPBS and analyzed by flow cytometry (BD Accuri™ C6). Data analysis was performed using FlowJo software (version 10.0.7).

To evaluate uptake of nanoparticles in 2D cell systems, MDA-MB-231, BT-549, BT-474, A431, MCF7 cells (5.0 × 10⁴ cells/well in 24-well) or MSCs (4.0 × 10⁴ cells/well in 24-well) were seeded on the coverslip and, after 24 h, were incubated for 30 or 60 min at 37 °C with Ir_en-AuSiO₂_COOH/NHS or Ir_en-AuSiO₂_Aptamer nanoparticles, diluted at 5 μM Ir_en concentration in culture medium with 0.1 mg/mL yeast tRNA and 0.1 mg/mL ultrapure™ salmon sperm DNA (Invitrogen), as nonspecific competitors. After three washes with DPBS, cells were fixed with 4% paraformaldehyde in DPBS for 30 min at RT, permeabilized with 0.5% Triton X-100/DPBS for 5 min, subjected to nuclear staining with the NucRed 647 (Invitrogen), following the provider indications, and mounted with glycerol/DPBS. Wheat Germ Agglutinin-Alexa Fluor 488 conjugate (WGA-488) was used to visualize BT-549 cell membrane.

To test the ability of Ir_en-AuSiO₂_Aptamer nanoplatforms to enter 3D models, heterotypic spheroids composed of BT-549-GFP or BT-474 tumor cells mixed with MSCs, and PDCOs (~ 100–200 μm diameter) were collected, centrifuged at 500 × g for 5 min, suspended in 90% Matrigel and 15-μL drops were deposited in prewarmed 8-well Chamber Slide (1 drop/well, Ibidi GmbH, Gräfelfing, Germany). Upon completed gelation, 200 μL of culture medium was added to each well. After 24 h, spheroids or PDCOs were incubated for 24 h at 37 °C with Ir_en-AuSiO₂_COOH/NHS or Ir_en-AuSiO₂_Aptamer nanoparticles, diluted at 5 μM Ir_en concentration in culture medium with nonspecific competitors. After three washes with DPBS, they were fixed, permeabilized and stained with NucRed 647, as described above. For α-smooth muscle actin (α-SMA) and fibroblast activation protein (FAP) staining, spheroids, fixed and permeabilized as described above, were subjected to blocking in 3% bovine serum albumin (BSA)/DPBS for 20 min at RT, incubated with anti-α-SMA (1:100, D4K9N, Cell Signaling Technology Inc.) or anti-FAP (1:50, SS-13, Santa Cruz Biotechnology Inc.) primary antibody diluted in 3% BSA/DPBS for 1 h at 37 °C, washed three times in DPBS, and incubated with Alexa Fluor 568 anti-rabbit or Alexa Fluor 568 anti-mouse (Invitrogen), respectively. Finally, glycerol/DPBS was added to each well.

Samples were visualized by Zeiss LSM 700 META confocal microscopy and imaging was performed using the following excitation wavelengths: 405 nm (Ir_en), 488 nm (GFP; WGA-488), 555 nm (α-SMA; FAP), 639 nm (NucRed 647).

MDA-MB-231, BT-549, A431, MCF7 and BT-474 (7.0 × 10³ cells/well) and MSCs (4.0 × 10³ cells/well) were plated in 96-well microplates (Corning Incorporate) and, after 16 h at 37 °C, were either left untreated or treated for 1 h with 5 μM free Ir_en or Ir_en-loaded nanoformulations (Ir_en-AuSiO₂_COOH/NHS, Ir_en-AuSiO₂_CL4, Ir_en-AuSiO₂_Gint4.T, Ir_en-AuSiO₂_CL4_Gint4.T or Ir_en-AuSiO₂_Scr), diluted in cell culture medium at 5 μM Ir_en concentration, in the presence of 0.1 mg/mL yeast tRNA and 0.1 mg/mL ultrapure™ salmon sperm DNA (Invitrogen), as nonspecific competitors. After two washes with DPBS, fresh medium was added to the plate and the cells were kept in the dark or exposed to 254 nm light irradiation (maximal irradiance 4 W m − 2) for 1 h. Cell viability was evaluated 24 h after PDT treatment by Thiazolyl Blue Tetrazolium Bromide (MTT, AppliChem GmbH, Darmstadt, Germany), according to the manufacturer’s protocol.

Heterotypic spheroids grown in 24-ultralow attachment plates up to reach a diameter of approximately 150–180 µm, were left untreated or treated for 24 h with Ir_en-AuSiO₂_CL4, Iren-AuSiO₂_Gint4.T, Ir_en-AuSiO₂_CL4_Gint4.T or Ir_en-AuSiO₂_Scr (5 µM Ir_en concentration), diluted in culture medium in the presence of nonspecific competitors (0.1 mg/mL yeast tRNA and 0.1 mg/mL ultrapure™ salmon sperm DNA, Invitrogen). After two washes with DPBS, fresh culture medium was added to the plates and spheroids, either untreated and treated with nanoparticles, were exposed to 254 nm light irradiation (maximal irradiance 4 W m − 2) for 1 h. After 24 h, spheroids with a diameter greater than 50 µm were counted in at least 5 fields per condition, to monitor spheroid destruction. For cell viability assays, heterotypic spheroids or PDCOs (~ 100–200 µm diameter) were transferred (20 μL drop plus 90% Matrigel) into black clear bottom 96-well plates (Corning Incorporate) that were filled with culture medium. After 24 h, spheroids or PDCOs were left untreated or treated with each nano-formulation and irradiated as reported above. Cell viability was assessed by CellTiter-Glo® 3D (ATP) luminescent assay (Promega BioSciences Inc., San Luis Obispo, CA), according to the manufacturer’s protocol, using TECAN Infinite 200Pro microplate reader.

Multifunctional light-responsive nanoplatforms were implemented by incorporating an Iridium-based metal complex, Ir_en, with photosensitizing and luminescent properties [41], within the polysiloxane matrix of gold-core/silica-shell nanoparticles (Ir_en-AuSiO₂_COOH).

The preparation of Ir_en-AuSiO₂_COOH was performed according to the well-known reverse microemulsion method [46] and a schematic representation of the synthetic procedure is illustrated in Fig. 1A. In a W/O microemulsion, water nanodroplets are stabilized by surfactant molecules and dispersed in a continuous oil phase. These reverse micelles act as nanoreactors, within which the homogenous and highly reproducible synthesis of nanoparticles takes place, minimizing the batch-to-batch variability. In the first step, the reduction of tetrachloroaurate(III) to Au⁰ leads to the formation of the gold core. Then, the addition of the silane precursors in alkaline environment gives rise, through hydrolysis and condensation processes, to the formation of the silica shell. The addition of the photosensitizing and luminescent Ir_en before the start of the polymerization process, ensures its physical incorporation into the polysiloxane matrix. Finally, the nanoparticles surface was functionalized with a coating agent containing a siloxy alkyl chain with a terminal carboxyl group. The full characterization of the Ir_en-AuSiO₂_COOH nanoparticles, morphological features, surface charge, optical properties, photosensitizing and thermoplasmonic abilities, was reported in the Supplementary Information. TEM images revealed a homogeneous population of spherical gold-core/silica-shell particles, with an average size of 49.13 ± 4.28 nm and a gold core of 6.31 ± 0.81 nm (Supplementary Fig. 1A). Ir_en-AuSiO₂_COOH are characterized by a hydrodynamic diameter of 85.51 ± 1.17 nm, Polydispersity Index (PDI) of 0.148 and a negative ζ-potential value of -23.5 ± 3.95 mV (Supplementary Fig. 1B). Similar ζ-potential values are reported for silica surfaces functionalized with terminal carboxyl groups in neutral aqueous medium [47, 48]. The successful loading of Ir_en within the nanoparticle silica shell was confirmed by UV–Vis spectroscopy (Supplementary Fig. 1C,D). In particular, the extinction spectrum of the colloidal solution of Ir_en-AuSiO₂_COOH shows a broad band centered at 520 nm, corresponding to the characteristic Localized Surface Plasmon Resonance (LSPR) of the spherical gold core [49], and a more intense absorption band at lower wavelengths (250–300 nm), due to electronic transitions involving the Ir_en molecule [41]. Under light excitation, the emission spectrum of Ir_en-AuSiO₂_COOH presents the characteristic band of the Iridium compound, peaked at 520 nm. The goodness of the emission spectrum was confirmed by the excitation one (Supplementary Fig. 1D).

In order to validate the photo-triggered properties of the obtained nanoplatforms, light-induced singlet oxygen generation and photothermal effects were assessed (see Supplementary Information for details). In particular, the singlet oxygen generation capability was evaluated by chemical method using 9,10-Anthracenediyl-bis(methylene)dimalonic acid (ABDA) as detection probe [50]. The photooxidation of ABDA in presence of Ir_en-AuSiO₂_COOH was monitored by measuring its absorbance at 378 nm; while the absorbance attenuation of ABDA was negligible for the control solution (see Supplementary Information and Supplementary Fig. 2A), in presence of Ir_en-AuSiO₂_COOH, the absorption decreased significantly with the extension of the irradiation time (Supplementary Fig. 2B). For an immediate comparison, the ABDA absorbance values as function of the irradiation time were plotted (Supplementary Fig. 2C) and for both solutions (control and sample) a good linear relationship was observed.

The heat generation of the nanoplatforms under continuous illumination was investigated (Supplementary Information). As clearly demonstrated by the thermal images acquired after an irradiation time of 90 min (Supplementary Fig. 3), in the case of the control solution (see Supplementary Information for its description), a non-significant temperature variation was observed, whereas in the case of Ir_en-AuSiO₂_COOH, a photothermal heating of the irradiated solution as well as the surrounding environment was highlighted. The observed thermoplasmonic effect can be explained as a result of a photothermal conversion of the energy absorbed by Ir_en molecules and partially transferred to the metal nanoparticle. In particular, the evidenced spectral overlap between the emission band of the Iridium compound and the LSPR of the gold core, allows donor–acceptor energy transfer processes, from the Iridium-based molecules to the gold core, which in turn converts the received energy into heat [51]. Therefore, using a single excitation wavelength in the Ir_en absorption region, Ir_en-AuSiO₂_COOH nanoplatforms can act as luminescent probes, photosensitizing agents, and heat nanosources.

In order to specifically address Ir_en-AuSiO₂_COOH to target tumor/stromal cells, conjugation with the EGFR CL4 (Ir_en-AuSiO₂_CL4) or PDGFRβ Gint4.T (Ir_en-AuSiO₂_Gint4.T) aptamers, and dual functionalization with both CL4 and Gint4.T (Ir_en-AuSiO₂_CL4_Gint4.T) were performed. Ir_en-AuSiO₂_Scr, decorated with a non-targeting scrambled aptamer, were used as a negative control. The schematic representation of the synthetic steps involved in the development of the aptamers-conjugated gold–silica nanoplatforms and a key illustration of all the prepared samples is shown in Fig. 1B,C. Specifically, to obtain aptamers-nanoplatforms conjugates (Ir_en-AuSiO₂_Aptamer), the free –COOH groups of Ir_en-AuSiO₂_COOH were activated—through EDC/NHS chemistry [52, 53]—to react with the 5' NH₂-aptamers, with consequent formation of amide bonds (− CO–NH −). In particular, EDC reacts with a carboxylic group on nanoparticles surface, resulting in an amine-reactive O-acylisourea intermediate. The addition of NHS stabilizes the amine-reactive intermediate by converting it to an amine-reactive NHS ester. Then, through a nucleophilic attack the NHS ester is easily displaced by the amine group on the 5′-end of aptamer to yield a stable amide bond (Fig. 1B). Extinction spectra of Ir_en-AuSiO₂_COOH/NHS and all conjugated samples are displayed in Fig. 2A. Where the spectral profile of Ir_en-AuSiO₂_COOH/NHS results superimposable on the extinction spectrum of Ir_en-AuSiO₂_COOH (Supplementary Fig. 1C), in the case of Ir_en-AuSiO₂_CL4, Ir_en-AuSiO₂_Scr, Ir_en-AuSiO₂_Gint4.T and Ir_en-AuSiO₂_CL4_Gint4.T, a shoulder at 260 nm—characteristic of the maximum absorption peak of RNA sequences [54]—was observed. This absorption feature proves the successful RNA functionalization. The stability of aptamers-conjugated nanoparticles was monitored by recording extinction spectra over 7, 14, 21 and 28 days. As shown in the Supplementary Fig. 4A-D, no significant variation of the spectral profiles – decrease in optical density or shift of absorption maxima—was observed over time, highlighting the stability of the nanoplatforms in the aqueous environment.

The DLS results (Fig. 2B,C, Supplementary Table 1) reported that the Ir_en-AuSiO₂_COOH/NHS nanoplatforms were characterized by a hydrodynamic diameter equal to 102.0 ± 0.15 nm (PDI of 0.172) and a negative ζ-potential value of -26.0 mV. Ir_en-AuSiO₂_CL4 and Ir_en-AuSiO₂_Scr were characterized by a hydrodynamic diameter of 102.1 ± 0.93 nm (PDI = 0.163) and 101.4 ± 0.20 nm (PDI = 0.162), respectively, and a negative ζ-potential value of -22.3 and -22.9 mV each. Ir_en-AuSiO₂_Gint4.T and Ir_en-AuSiO₂_CL4_Gint4.T were characterized by a hydrodynamic diameter equal to 104.4 ± 0.47 nm (PDI = 0.168) and 103.7 ± 0.82 nm (PDI = 0.162), respectively, and a negative ζ-potential value of -28.3 and -26.4 mV each. Therefore, Ir_en-AuSiO₂_COOH/NHS as well as all Ir_en-AuSiO₂_Aptamer samples have a hydrodynamic diameter very similar to each other. However, these values are larger than that of Ir_en-AuSiO₂_COOH, plausibly due to variations in the surface coating and/or hydration sphere. Conversely, no significant changes were observed in the surface charge values of all the nanoplatforms (Ir_en-AuSiO₂_COOH, Ir_en-AuSiO₂_COOH/NHS, Ir_en-AuSiO₂_Aptamer), as result in all cases of an extensive presence of free carboxyl groups on the surface of the nanoplatforms. The amount of each aptamer conjugated to the nanoplatforms was evaluated by RT-qPCR analysis on Ir_en-AuSiO₂_Aptamer (Supplementary Fig. 5). We quantified approximately 3.0 pmol aptamer per 16.0 pmol of Ir_en-AuSiO₂_ Aptamer nanoplatform and the efficiency of conjugation ranged between 2 and 6% with a mean ± SEM equal to 3.3 ± 0.7%.

To assess the cell targeting/uptake ability of nanoparticles conjugated to a single aptamer, CL4 or Gint4.T, or functionalized to both aptamers, we took advantage of different human cell lines expressing only EGFR, only PDGFRβ, both, or neither of the receptors. Specifically, human MES-TNBC MDA-MB-231 and BT-549 cells were chosen as double-positive cells as they express both EGFR and PDGFRβ ([23] and Supplementary Fig. 6) and, consequently, are specifically targeted by the two aptamers either when grown in classical 2D cultures and 3D Matrigel-embedded cultures, or implanted in mice [23, 30, 39]. Furthermore, we previously proved that CL4 strongly improves the uptake of drug-loaded and aptamer-decorated poly(lactic-co-glycolic)-poly ethylene glycol-based nanoparticles into both cell lines in vitro and in vivo [24]. As models of EGFR⁺/PDGFRβ⁻ cell lines, we used breast cancer BT-474 and epidermoid carcinoma A431 cell lines, which express moderate and high levels of EGFR, respectively, without expressing PDGFRβ ([23, 44] and Supplementary Fig. 6), and are thus recognized by CL4 [19] but not Gint4.T [35, 39]. Moreover, stromal MSCs that express high levels of PDGFRβ were selected as EGFR⁻/PDGFRβ⁺ cells ([30, 55] and Supplementary Fig. 6). Importantly, we previously demonstrated that Gint4.T, by blocking PDGFRβ, efficiently inhibits MSCs recruitment into TNBC thus preventing their pro-metastatic function [30]. Finally, EGFR⁻/PDGFRβ⁻ breast cancer MCF7 cells ([23] and Supplementary Fig. 6) were used as a negative control.

The intrinsic Iridium-compound associated fluorescence emitted (Em = 520 nm) from the unconjugated Ir_en-AuSiO₂_COOH/NHS or Ir_en-AuSiO₂_Aptamer nanoplatforms (5 μM Ir_en concentration) was collected by confocal microscopy after incubation of the cells with the nanoparticles at 37 °C in the presence of yeast tRNA and salmon sperm DNA competitors to hinder any non-specific interactions. As shown (Fig. 3A), the signal associated with Ir_en-AuSiO₂_Aptamer (Ir_en-AuSiO₂_CL4, Ir_en-AuSiO₂_Gint4.T and Ir_en-AuSiO₂_CL4_Gint4.T) nanoparticles was clearly visible in the cytoplasm of MDA-MB-231 cells at 30 min and further increased at 60 min of incubation. Conversely, an almost undetectable signal was obtained with unconjugated Ir_en-AuSiO₂_COOH/NHS or scrambled decorated Ir_en-AuSiO₂_Scr nanoparticles at the two incubation times. Importantly, MDA-MB-231 cells treated with the dual-targeting EGFR/PDGFRβ nanoparticles had significantly higher fluorescence intensity than that treated with single-targeting EGFR or PDGFRβ nanoparticles (Fig. 3A), indicating that CL4 and Gint4.T, simultaneously attached to the nanoparticles, confer improved cellular uptake. Similar results were observed on EGFR⁺/PDGFRβ⁺ BT-549 cells (Fig. 3A). Labeling cells with WGA to visualize cell membrane confirmed that the CL4 and Gint4.T aptamers properly drive the nanoparticles in the cytoplasm (Supplementary Fig. 7A). As expected, no signal was observed in double negative MCF7 control cells (Fig. 3A).

As a next step, the nanoparticle’ formulations were incubated for 60 min onto only EGFR⁺ A431 (Fig. 3B) and BT-474 (Supplementary Fig. 7B) cell lines, and only PDGFRβ⁺ MSCs (Fig. 3C), and confocal microscopy analyses revealed that the internalization ability of the nanoparticles strictly depends on the aptamer present on the nanoparticle’s surface and the expression of its receptor partner on the target cell.

Overall, these results show efficient capability of Ir_en-AuSiO₂_Aptamer nanoparticles to selectively enter into the cell through EGFR and/or PDGFRβ recognition and indicate that the carriers’ parameters, including composition, size, shape and surface chemistry, do not affect the interactions of CL4 and Gint4.T aptamers to their proper receptor on membranes of target cells.

Next, we wondered whether multifunctional nanoparticles retain their targeting ability in more relevant in vitro cancer models by using 3D multicellular spheroids obtained by co-culturing tumor and stromal cells on non-adhesive culture dishes, which resemble the organization and properties of a native tumor, as a key factor of translational medicine [56, 57]. Even if these models have been successfully used to study tumor-MSC interaction [56, 57], to date a still limited number of studies have employed 3D tumor spheroids for evaluating the functionality of nanomedicine [58].

In order to distinguish cancer cells from stromal cells we established tumor spheroids consisting of GFP-labeled BT-549 and unlabeled MSCs. In agreement with previous findings, both BT-549 cells [59, 60] and MSCs [61, 62] were able to form spheroids alone in non-attached culture (Supplementary Fig. 8A). When cancer cells were co-cultured with stromal cells at a 1:4 ratio, respectively [63], consistent heterotypic spheroids (BT-549 + MSCs), were obtained, reaching approximately 180 µm in diameter, in 13 days (Supplementary Fig. 8A and Fig. 4A). In order to better visualize formed spheroids, they were embedded in Matrigel and observed by confocal microscopy (Fig. 4B). NucRed 647 nuclear stain was used to visualize both cancer and stromal cells, while the GFP to visualize cancer cells. As shown (Fig. 4B), the presence of NucRed 647 nuclear stain (visualized in blue) either associated to the GFP signal (BT-549) or not (MSCs) (see arrows in the merged image), indicates the mixed composition of the spheroids. Moreover, immunofluorescence with α-SMA and FAP markers confirmed the presence of tumor-associated MSCs in the heterotypic spheroids (Supplementary Fig. 8B). To examine the penetration of the different nanoformulations into the spheroids, BT-549/MSCs spheroids were exposed to Ir_en-AuSiO₂_Aptamer and unconjugated nanoparticles, at an Ir_en concentration of 5 μM, for 24 h at 37 °C and visualized by confocal microscopy (Fig. 4C). Importantly, the presence of the EGFR and PDGFRβ targeting aptamers on the surface of the nanoparticles, either single-targeted (Ir_en-AuSiO₂_CL4 and Ir_en-AuSiO₂_Gint4.T) or dual-targeted (Ir_en-AuSiO₂_CL4_Gint4.T), allowed them to penetrate the mixed tumor/stromal spheroids as qualitatively displayed by the nanoplatform-associated fluorescent signal (visualized in red) that was visible throughout the spheroids. 3D images clearly indicated the accumulation of the aptamer-functionalized nanoparticles inside the spheroid mass. Conversely, no signal was detected with unconjugated Ir_en-AuSiO₂ or Ir_en-AuSiO₂_Scr negative control (Fig. 4C), thus indicating that passive infiltration of the spheroids by untargeted nanoparticles could not occur at least under the experimental conditions used.

We wondered whether the nanoparticles conjugated to CL4, Gint4.T or both the aptamers, once penetrated into the spheroids, kill cancer and stromal cells upon light irradiation. Thus, we first validated that AuSiO₂_COOH/NHS, AuSiO₂_Scr, AuSiO₂_CL4 or AuSiO₂_Gint4.T nanoparticles, which had no load of photosensitizing and luminescent molecule Ir_en, when incubated for 24 h at 37 °C on MDA-MB-231 and BT-549 cells have no adverse effects on cell viability (Supplementary Fig. 9A,B).

Next, before moving on to more complex 3D cell systems, we verified the therapeutic efficacy of Ir_en-loaded nanoformulations in 2D cell cultures. To this aim, BT-549 and MDA-MB-231, positive for EGFR and PDGFRβ, BT-474 and A431, positive for EGFR, MSCs, positive for PDGFRβ, and MCF7 cells, negative for both receptors, were incubated with the different nanoformulations containing Ir_en (5 µM) and decorated or not with the aptamers, for 1 h, given the rapid cell uptake of aptamer-decorated nanoparticles, washed to remove not-internalized nanoparticles, exposed to 1-h light irradiation and analyzed for their viability after 24 h. As shown (Supplementary Fig. 9C-H), Ir_en-AuSiO₂_Aptamer nanoplatforms demonstrated efficient capability to selectively kill the cells, through EGFR and/or PDGFRβ recognition, in comparison with untargeted NPs (unconjugated Ir_en-AuSiO₂_ COOH/NHS or scrambled-conjugated Ir_en-AuSiO₂_Scr). No toxicity of aptamer-decorated nanoplatforms was observed on each cell line under dark conditions, thus indicating their safety behavior at least in the concentrations tested in the PDT experiments. We verified that free Ir_en is nontoxic and does not contribute to the photoinduced killing activity effects, at least at the concentration and exposure time used in the encapsulated form (Supplementary Fig. 9C-H).

Importantly, when incubated for 24 h with heterotypic BT-549 + MSCs spheroids, Ir_en-AuSiO₂_CL4, Ir_en-AuSiO₂_Gint4.T and Ir_en-AuSiO₂_CL4_Gint4.T (5 µM Ir_en concentration; 1-h light irradiation), disrupted spheroid structure (Fig. 5A,B) and inhibited cell viability (Fig. 5C), as determined by spheroids number counting and CellTiter-Glo 3D cell viability assay, respectively, with the dual targeted nanoparticles being more effective than the single-targeted ones (approximately 80% inhibition, Ir_en-AuSiO₂_CL4_Gint4.T vs 40% inhibition, Ir_en-AuSiO₂_CL4, and 50% inhibition, Ir_en-AuSiO₂_Gint4.T). Conversely, no effect was observed on controls consisting of untreated or treated with scrambled-decorated Ir_en-AuSiO₂ nanoformulations spheroids (Fig. 5A-C). Similarly, a higher effect on cell viability inhibition was observed upon treatment of heterotypic MDA-MB-231 + MSCs spheroids with Ir_en-AuSiO₂_CL4_Gint4.T compared to single-targeted nanoparticles (Fig. 5D,E).

Because MES-TNBC cells express both EGFR and PDGFRβ, in order to confirm the efficacy of the dual-targeting nanovectors on both cancer and stromal cells, we generated tumor spheroids consisting of cancer BT-474 cells (only EGFR⁺) and MSCs (only PDGFRβ⁺) that grew as colonies with approximately a 150 μm diameter after 13 days in culture (Supplementary Fig. 10A and Fig. 6A). As shown, uptake into BT-474/MSCs spheroids of either single-targeted Ir_en-AuSiO₂_CL4 and Ir_en-AuSiO₂_Gint4.T, as well as dual-targeted Ir_en-AuSiO₂_CL4_Gint4.T nanoparticles after 24 h of incubation was clearly detected by confocal microscopy (Fig. 6B) and the cytotoxic effect of Ir_en-AuSiO₂_CL4_Gint4.T nanoparticles on both the spheroid disruption (Fig. 6C,D) and cell viability inhibition (Fig. 6E) was higher than that of either Ir_en-AuSiO₂_CL4 or Ir_en-AuSiO₂_Gint4.T (approximately 70% inhibition for Ir_en-AuSiO₂_CL4_Gint4.T vs 40% inhibition for Ir_en-AuSiO₂_CL4 and Ir_en-AuSiO₂_Gint4.T, relative to untreated cultures), thus confirming the ability of dual-decorated nanoparticles to kill both cancer and stromal cells through EGFR and PDGFRβ targeting, respectively.

Similarly, tumor spheroids consisting of A431 cells (only EGFR⁺) grown with MSCs (only PDGFRβ⁺) up to 13 days (Supplementary Fig. 10B and Fig. 6F) were higher affected by EGFR/PDGFRβ bispecific nanoformulations than those single-targeted, with a reduction of cell viability of approximately 90% for Ir_en-AuSiO2_CL4_Gint4.T, 40% for Ir_en-AuSiO₂_CL4 and 30% for Ir_en-AuSiO₂_Gint4.T relative to untreated cultures (Fig. 6G).

To further prove the targeting efficacy and the photoinduced killing activity of dual aptamer-decorated nanoparticles, through selective recognition of EGFR-positive tumor cells and PDGFRβ-positive stromal component in the entire tumor bulk, we employed 3D patient organoids from human surgical specimens collected from three patients with diagnosis of breast cancer. Tumor samples, henceforth named M23, M41 and M43, were chosen for the presence of EGFR only in tumor cells (Fig. 7A, blue arrows) and PDGFRβ only in the stromal component (Fig. 7B), specifically in vascular endothelial cells (green arrows) and/or mesenchymal stromal cells (orange arrows), as assessed by immunohistochemical analyses.

The clinical pathological characteristics of the three tumors are summarized in the Supplementary Table 2. PDCOs at the first passage were cultured up to day 10 (Fig. 7C, upper panels), disaggregated into a single-cell suspension and analyzed by flow cytometry to confirm the expression of EGFR on epithelial tumor cells and PDGFRβ on stromal cells (Fig. 7C, lower panels). Confirming results obtained in 3D multicellular tumor spheroids, Ir_en-AuSiO₂_Aptamer nanoplatforms efficiently spread into the organoid mass (Fig. 7D) and exert excellent anticancer photokilling activity, which was superior with the dual-aptamer-decorated nanoparticles over single-aptamer-conjugated (Fig. 7E). Conversely, no effect was observed in untreated organoids or those treated with scrambled-aptamer-conjugated nanoparticles (Fig. 7E).

Taken together, our results show the striking antitumor potential of our bispecific light-triggered nanoplatforms targeting tumor and stromal cells and highlight the potential translational value of integrative research combining patient-derived organoids and cancer nanomedicine.