Urine‐Stable Aptamer‐Conjugated Gold Nanorods for the Early Detection of High‐Grade Bladder Cancer Residual Disease

Abstract

The limitations of current diagnostic imaging techniques and therapies for bladder cancer are associated with and responsible for the recurrence and progression of residual disease, with an impact on social costs and quality of life of patients. This study delivers a cost‐effective solution for the detection of bladder cancer residual disease, which is represented by the early detection of bladder cancer lesions < 1 mm. Urine‐stable 34‐mer SH‐terminated 2′F‐Py‐RNA aptamer that recognizes the integrin α5β1, expressed by 81% of human high‐grade non‐muscle invasive bladder cancer, is developed. The intravesical instillation of aptamer‐conjugated gold nanorods as contrast agent for photoacoustic imaging is validated in a preclinical model of orthotopic bladder cancer expressing the integrin α5β1. The photoacoustic signal of gold nanorods remains on the tumor surface for 3 h and allows early detection of cancer lesions < 1 mm. The aptamer is internalized into lysosomes, an opportunity that paves the way for lysosomal‐mediated drug release in tumor cells. This study highlights the potential of urine‐stable aptamer for the delivery of a solution to target the residual high‐grade bladder cancer disease.

Aptamer (A) conjugated gold nanorods (B) for the recognition of bladder cancer cells expressing the integrin α5β1 (C) and the visualization of the targeted GNRs@Chit‐Apt‐Itg by photoacoustic imaging at the surface of the orthotopic bladder cancer in the preclinical model (D).

1. Introduction

Bladder cancer is the 10^th most common malignancy worldwide,^{[ 1 ]} but the most expensive to be treated. The annual cost of bladder cancer is $5 billion in Europe and $6.5 billion in the USA, accounting for 5% of the overall cancer costs in Europe and USA.^{[ 2 ]} The reason of such high costs is associated i) to the multiple interventions,^{[ 3 ]} cost of treatments,^{[ 4 ]} and routine follow‐up that are associated with the presence of residual disease in non‐muscle invasive bladder cancer (NMIBC) patients, ii) neoadjuvant therapy and radical cystectomy for those patients that from NMIBC progresses to muscle invasive bladder cancer (MIBC),^{[ 4} , ^{5 ]} and iii) immunotherapy treatments for metastatic bladder cancer.^{[ 6 ]}

A solution to reduce the social costs of bladder cancer, while improving the quality of life of patients, is to deliver an approach that tackles the residual disease in NMIBC patients, which they represent 75% of new diagnosis of bladder cancer, and in particular for the high‐grade NMIBC that is characterized by the highest frequency of relapse and risk of progression.^{[ 7 ]} Technological limitations of the current diagnostic imaging techniques are associated with the residual disease in NMIBC patients. Even the blue‐light cystoscopy that carries out the photodynamic diagnosis is ineffective in detecting tumors smaller than 1 mm, with a specificity that is lower than white light endoscopy (63% vs 81%). This is because photodynamic diagnosis can detect areas of inflammation following trans‐urethral resection of bladder tumor (TURBT) or intravesical instillation of the Bacillus of Calmette‐Guerin (BCG) in patients with a diagnosis of high‐grade NMIBC.^{[ 8 ]} Other recent studies reported that micro‐ultrasound using a 29‐MHz side‐fire transducer and MRI based imaging (VI‐RADS, dynamic contrast‐enhanced magnetic resolution imaging‐DCE MRI, and intravesical contrast‐enhanced MRI that is ICE‐MRI) enables accurate prediction of muscle‐wall invasion of bladder cancer (MIBC).^{[ 9 ]} Likewise, recent reports have shown that contrast‐enhanced ultrasound imaging provides quick, unexpensive, non‐ionizing radiation, non‐invasiveness detection of MIBC, but unable to detect NMIBC or lesions that are flat or <5 mm width.^{[ 10 ]}

We have recently demonstrated that contrast‐enhanced photoacoustic imaging allows the detection of bladder cancer lesions of < 1 mm. This solution is based on the delivery of gold nanorods as contrast agent for photoacoustic imaging and the use of a cyclic peptide as targeting moiety for tumor cells.^{[ 11 ]} The use of the cyclic peptide is very efficient and valid for preclinical research, but the racemization process that occurs with cyclic peptides, in particular when they contain the amino acid phenylglycine,^{[ 12 ]} gives rise to mixed formulations of the D and L isoforms. For purposes of clinical application, their purification is therefore necessary to have a repeatable product, but the purification process and the required quality control tests for the cyclic peptide makes this product very expensive. Our goal is to make the use of nanoparticles acceptable in clinical use. To this aim, in addition to demonstrating their effectiveness, their economic cost must also be competitive with current clinical and therapeutic approaches. To improve the acceptability of our solution, in this study we aimed to refine the formulation of the targeted nanoparticles. We here investigated the use of a nuclease‐resistant RNA aptamer as delivering agent for gold nanoparticles, since aptamers have target recognition similar to that of antibodies but, being synthesized chemically, they are easier and cheaper to produce and scale up without batch‐to‐batch variations, have greater versatility in chemical modification, higher stability, longer shelf life and are not immunogenic.^{[ 13} , ^{14 ]} This study delivers a novel nanosystem that maintains chemical‐physical properties in the presence of urine and binding capacity to bladder tumor cells positive for the integrin α5β1, which is expressed in the preclinical model of orthotopic bladder cancer and 81% of high‐grade human NMIBC.^{[ 11 ]} In addition to the development of the nanosystem and the detailed characterization, we have investigated its stability in the presence of urine and the strength as a contrast agent for targeted photoacoustic imaging for application in the clinical scenario.

2. Results

2.1. Optimization of the Anti‐Integrin α5β1 Aptamer

A 68‐mer unmodified RNA aptamer (H02) targeting integrin α5β1 was recently selected.^{[ 15 ]} However, while exploring the scope of the aptamer for our research, we reasoned that its high size and low resistance to nucleases degradation will have impeded its conjugation to gold‐nanorods and further clinical translation. Thus, to address these limitations, a new 34‐mer aptamer was derived from the reported H02 by removing nucleotides from 1 to 17 and 52 to 68 in the fixed flanks of the full‐length molecule. The secondary structure and free energy change (ΔG) of the new and shorter aptamer (Apt‐Itg) was predicted by using Mfold web server;^{[ 16 ]} the ΔG value of ‐14.4 Kcal mol⁻¹ indicates the high structural stability of the Apt‐Itg (Figure  1A). To have a better understanding of the interaction between Apt‐Itg and integrin α5β1, we led molecular docking analysis using AlphaFold 3 Server prediction.^{[ 17 ]} AlphaFold 3 is an artificial intelligence‐based system that can model protein–nucleic acid interactions at high accuracy.^{[ 18 ]} The potential binding region of Apt‐Itg on integrin α5 within the integrin α5β1 heterodimer is shown in Figure 1B. Interestingly, D154 and Y208, which are key residues for the interaction of integrin α5 with the synergy site of its natural ligand fibronectin, are predicted to be involved in the interaction with Apt‐Itg (Figure 1B, frame on the side).^{[ 19 ]}

Figure 1.

Features of the aptamer and tested bladder cancer cell lines. A) The secondary structure of H02 full‐length and Apt‐Itg (nucleotides from 18 to 51, boxed); structures of the H02 full‐length and Apt‐Itg were predicted by using Mfold web server, and the calculated change in Gibbs free energy (ΔG) values are reported. B) AlphaFold 3 molecular docking result of integrin α5β1‐Apt‐Itg binding complex; Apt‐Itg, integrin α5 and integrin β1 predicted structures are shown in yellow, light blue and red, respectively. The predicted interaction of Apt‐Itg with integrin α5, involving D154 and Y208 residues, is framed on the side. C) The expression of integrin α5β1 by human and murine bladder cancer cell lines evaluated by western blot analysis of the α5 and β1 chain of the integrin. The expression of the housekeeping protein α‐tubulin was used to estimate the relative fold of expression of integrin chains D). MB49‐Luc; murine bladder cancer cell line.

In the present study, all the pyrimidine residues in Apt‐Itg sequence are modified to 2′F‐Py to enhance RNA stability, and the 5′‐end is modified by adding a Thiol‐C6 group for functionalization of GNRs@Chit.

The SH‐terminated 2′F‐Py RNA with no affinity for integrin α5β1, consisting in the scrambled sequence of a previously generated anti‐EGFR aptamer was here used as negative control,^{[ 20 ]} and is indicated as “Unrelated”.

To confer active targeting abilities to GNRs@Chit toward integrin α5β1‐positive MB49 cells, we used the nuclease‐resistant 2′F‐Py RNA Apt‐Itg aptamer as the targeting agent.

Apt‐Itg has the following advantages over the original 68‐mer H02 longer sequence: 1) easier chemical synthesis at lower costs and higher yield; 2) easier conjugation to GNRs@Chit nanovector; 3) increased resistance toward nucleases due to the modified pyrimidine residues.^{[ 21 ]}

2.2. Identification of the Integrin α5β1 on Bladder Cancer Cell Lines

Murine and human bladder cancer cell lines were searched for the expression of the integrin α5β1. We confirmed recently published data showing that human bladder cancer cell lines express the integrin α5β1,^{[ 11 ]} and here further demonstrated that the integrin α5β1 is mainly expressed by bladder cancer cells of grade 2/3 (RT4 = G1; T24 = G3; 5637 = G2; HT‐1376 = G3; RT112 = G2), as the human in situ carcinoma that always appears as high grade/grade 3.^{[ 22 ]} The murine bladder cancer cell line MB49‐Luc also expressed the integrin α5β1, similarly to the human bladder cancer cell line of grade 2/3 (Figure 1C).

2.3. Apt‐Itg Binds Human and Murine Bladder Cancer Cell Lines with Nanomolar Affinity and is Actively Internalized into Target Cells

Since H02 was selected to bind to the human integrin α5,^{[ 15 ]} we first checked whether Apt‐Itg contains the H02 active site and binds to human integrin α5β1‐positive T24 and RT112 bladder cancer cell lines, expressing high and moderate levels of integrin α5, respectively, and comparable levels of integrin β1 (Figure 1D). Binding was assessed through flow cytometric assays using internally Alexa 647‐labeled aptamer. As shown (Figure  2A), Apt‐Itg binds to T24 cells with a Kd value of 225.6 ± 30.8 nm which is somewhat comparable to that observed for the full‐length H02 aptamer on human glioblastoma U87MG cells genetically modified to overexpress integrin α5 (277.8 ± 51.8 nm). A higher Kd value (467.4 ± 94.0 nm) was calculated on RT112 cells (Figure 2B), thus reflecting the lower expression level of α5 on these cells compared to T24 cells. By contrast, the unrelated control aptamer gave no appreciable binding on the two cell lines, even at concentrations as high as 4 µm.

Figure 2.

Binding assay of Alexa 647‐Apt‐Itg to human and murine bladder cancer cells using flow cytometry. A–C) Representative binding profiles of increasing concentrations of Alexa 647‐Apt‐Itg and Unrelated aptamer incubated with T24 (A), RT112 (B) and MB49‐Luc (C) cells. Apparent dissociation curves of aptamer‐cell interaction and Kd value of Apt‐Itg is shown. Data are presented as the mean ± SD. D) Competition binding assay by flow cytometry (left) and quantification of the gMFI (right) using 1 µm Alexa 647‐Apt‐Itg in the absence (blue) or in the presence (red) of 30‐fold excess of unlabeled unrelated aptamer.

Next, to test GNRs@Chit conjugated to Apt‐Itg in mice bearing MB49‐Luc‐derived tumors, we checked the capability of Apt‐Itg to bind to murine MB49‐Luc cells. Even if some species specificity has been shown in aptamers recognition ability,^{[ 23 ]} we hypothesized a high affinity binding for Apt‐Itg to murine integrin α5, which shares ≈90% sequence identity, according to BLAST analysis, in the extracellular region with the human receptor. As expected, Apt‐Itg efficiently binds to MB49‐Luc cells, with a Kd value of 244.7 ± 58.1 nm (Figure 2C), even in the presence of pre‐treatment of cells with a 30‐fold excess of unrelated aptamer that shows almost undetectable nonspecific binding on cells (Figure 2D).

Integrins contain binding sites for divalent cations (such as calcium and magnesium) that regulate their ligand‐binding affinity.^{[ 24 ]} Kd values of Apt‐Itg in the presence of physiological concentrations of Ca²⁺ and Mg²⁺ cations on human (232.4 ± 29.7 nm) and murine (264.3 ± 30.3 nm) cancer cell lines (Figure S2, Supporting Information) did not statistically differ from those calculated in the absence of cations (P = 0.7 and 0.2, respectively), thus indicating that the integrin α5 recognition by the aptamer is not affected by the used ions concentration.

The integrin α5β1 H02 aptamer has been reported to actively internalize into target cells.^{[ 15 ]} We verified by confocal microscopy imaging whether our Apt‐Itg maintains the cell uptake ability of the full‐length aptamer. To this aim, bladder cancer cell lines of human and murine origin, T24 and MB49, respectively, were used. Cells were incubated with Alexa 647‐labeled Apt‐Itg for 40 min at 4 °C and 37 °C, and co‐localization experiments with WGA or LysoTracker were performed to detect aptamer localization on cell surface or into late endosomes/lysosomes compartments, respectively. The fluorescent signal from Apt‐Itg and WGA colocalized on the cell membrane at 4 °C, whereas punctuate aptamer signal colocalized with LysoTracker at 37 °C, indicating aptamer internalization within the target cells (Figure  3A). Prolonging the incubation at 37 °C up to 2 h, the Alexa 647 fluorescence intensity in the cell increased, suggesting continuous active Apt‐Itg uptake (Figure 3A,B). Apt‐Itg internalization was also confirmed within murine target MB49‐Luc cells, as shown by its accumulation in compartments positive for LysoTracker following 2 h incubation (Figure S3, Supporting Information). As expected, no signal was observed in the presence of the non‐targeting Alexa 647‐Unrelated aptamer (Figure 3A; Figure S3, Supporting Information).

Figure 3.

Apt‐Itg actively internalizes into T24 bladder cancer cells. A) Representative confocal images of T24 cells incubated for indicated times at 4 °C or 37 °C with 4 µm Alexa 647‐Apt‐Itg or Alexa 647‐Unrelated aptamer as a negative control. Aptamer, WGA‐488 (cell surface), LysoTracker (lysosomes), and DAPI (nuclei) are visualized in red, green, blue, and gray, respectively. All digital images were captured at the same setting to allow direct comparison of staining patterns. Magnification 63x, 1.0x digital zoom, scale bar = 10 µm. B) Plot of MFI of Alexa 647‐Apt‐Itg normalized to cell number. Bars depict mean ± SD. ***p < 0.001.

Together, these results demonstrate that optimized Apt‐Itg preserves the cell targeting and uptake of parental H02 and provides a striking tool for both bladder cancer cells detection and delivery of secondary cargos inside the target cells.

2.4. Silencing of the Integrin α5 Reduced the Binding of the Apt‐Itg

To definitely confirm the specificity of Apt‐Itg aptamer for integrin α5, we used siRNA to reduce integrin α5 expression in human T24 and RT112 cells, and murine MB49‐Luc. Compared with siRNA ctrl, the tested anti‐α5 siRNAs specifically reduced the expression of α5 (Figure  4A). Importantly, Apt‐Itg binding to either human and murine cells transfected with HS_ITGA5_7 and MM_ ITGA5_4, respectively, was significantly reduced compared to binding to siRNA ctrl treated cells (Figure 4B). Moreover, when integrin α5 was knocked down, Apt‐Itg binding to the cells was reduced by a similar fraction as integrin α5 antibody binding, as assessed with T24 representative cell line (Figure 4C).

Figure 4.

Integrin α5 silencing results in reduced Apt‐Itg binding. A) T24, RT112, and MB49‐Luc cells were left untreated (NT) or transfected with the indicated Integrin α5 siRNAs or siRNA ctrl. At 24 h post‐transfection, cells were harvested, and cell lysates prepared and immunoblotted with anti‐integrin α5 or anti‐integrin β1 antibody. Equal loading was confirmed by immunoblot with anti‐α‐tubulin antibody. The histograms indicate the integrin chains/α‐tubulin ratio of the densitometric signals. Values are shown relative to siRNA ctrl, arbitrarily set to 1. B) Binding of 1 µm Alexa647‐Apt‐Itg to human and murine bladder cancer cells following 24 h‐transfection with si‐ITGα5 (blue) and siRNA ctrl (red). C) Binding of Alexa488‐Ab‐Itg to T24 cells 24 h post‐transfection. (B, C) The histogram indicates gMFI of aptamer‐ or antibody‐treated cells normalized to the gMFI of siRNA ctrl untreated cells (black), arbitrarily set to 1. Bars depict mean ± SD of two independent experiments. *p < 0.005, **p < 0.001, and ***p < 0.0001.

Overall, these results indicate Apt‐Itg as a viable integrin α5‐targeting candidate for nanovectors functionalization.

2.5. Quantification of Apt‐Itg Bound to GNRs@Chit

We exploited a heterobifunctional cross‐linker reagent, composed of an ethylene oxide spacer (‐OCH₂CH₂‐ in a PEG₁₂) bearing a N‐hydroxysuccinimidyl (NHS) ester to its extremities and a maleimide functional group (NHS‐PEG₁₂‐maleimide). The NHS ester terminus reacts with the free amino groups on chitosan ensuring the binding of the linker to the chitosan, while the maleimide group is available to react in a further step with the sulfhydryl group of the aptamer, expressly designed with the thiol at the 5′ end (Figure  5A).

Figure 5.

Characterization of the final nanosystem, GNRs@Chit‐Apt‐Itg. A) Scheme representative of the aptamer Apt‐Itg conjugated to the GNRs@Chit via PEG₁₂ linker. B) TEM images of nanoparticles. C) Left, representative PCR calibration curve with unconjugated Apt‐Itg (red); samples from TOP, 1/2 and 1/10 GNRs@Chit‐Apt‐Itg preparations are shown in blue, green, and black, respectively. Right, quantification of the Apt‐Itg and unrelated aptamer loaded onto GNRs@Chitosan; data shown mean ± SD of three independent experiments. D) UV–vis spectra of GNRs@CTAB, GNRs@Chit, and GNRs@Chit‐Apt‐Itg/SCR showing no significant differences in the absorption spectrum of GNRs after surface modification; FAAS analysis of GNRs@Chit‐Apt‐Itg with calibration line; thermogravimetric analysis of GNRs@Chit in N₂ and air atmosphere (above 600 °C). GNRs@Chit‐Apt‐Itg stability in human urine was investigated by E) VIS‐NIR spectra of GNRs@Chit‐Apt‐Itg compared to the VIS‐NIR spectra of GNRs@CTAB and GNRs@Chit, and F) VIS‐NIR spectra and absorption intensity of GNRs@Chit‐Apt‐Itg in the presence of human urine over time. G) Binding of GNRs@Chit‐Apt‐Itg to MB49 cells after 15 min of incubation, followed by cell internalization of GNRs and accumulation into lysosome after 4 h. SCR: unrelated aptamer.

Transmission electron microscopy (TEM) confirmed effective synthesis of monocrystalline GNRs^{[ 25 ]} with the appropriate aspect ratio (3.43 ± 0.52) showing the presence of cylindrical gold nanostructures with width (25.7 ± 2.0 nm) and length (88.2 ± 6.4 nm), compatible with the NIR optical behavior of the nanosystem (Figure 5B).

To evaluate the efficiency and reproducibility of the conjugation of the Apt‐Itg to the nanoparticles, GNRs@Chit was left reacting with three different amounts of Apt‐Itg, such as 100, 50, and 10 pmol. The resulting product was called GNRs@Chit‐Apt‐Itg “TOP”, 1/2 sample and 1/10 sample. For the unrelated aptamer, only the “TOP” concentration was tested. From the three preparations of GNRs@Chit‐Apt‐Itg we quantified by RT‐qPCR analysis 4.0, 1.9, and 0.7 pmol of Apt‐Itg‐SH aptamer for the “TOP”, 1/2 and 1/10 preparation, respectively (Figure 5C). This data shows that the recovery of the bound Apt‐Itg was 4% for the TOP and 1/2 batches of nanoparticles, and 7% for the 1/10 preparation. For the “TOP” preparation we estimated a conjugation of 1.14 pmol Apt‐Itg‐SH per mg of dry content, which means 45 ng of aptamer for 0.1 mg of gold. Overall, the efficiency of conjugation of the Apt‐Itg ranged between 4% and 7% with a mean ± SEM equal to 5 ± 1%.

Overall, this information shows that the use of Apt‐Itg allows a reproducible preparation of targeted GNRs, with low batch‐to‐batch variation. For all subsequent tests the GNRs@Chit loaded with the TOP amount of Apt‐Itg or unrelated aptamer were used.

2.6. The Conjugation of Apt‐Itg does not Modify the Chemical‐Physical Properties of the GNRs@Chit

Next, we assessed the physical‐chemical properties of the metal core of GNRs@Chit‐Apt‐Itg/unrelated aptamer, determining the longitudinal‐LSPR to peak in any synthesized batch between 780 and 840 nm, with the specific analyzed batch peaking at 802 nm and having an aspect ratio of 3.62 ± 0.67 (length 90.2 ± 7.2 nm, width 24.9 ± 2.6 nm). These findings show that conjugation of aptamers does not modify the physical‐chemical properties of GNRs, with only small and gradual variation of the longitudinal‐LSPR peak versus GNRs@Chit (800 nm for this specific batch) and GNR@CTAB (798 nm for this specific batch) that is representative of the variation of the chemical environment surrounding the GNRs (Figure 5D).

Flame Atomic Absorption Spectroscopy (FAAS) and thermogravimetric analysis (TGA) were then used to establish the relative amount of gold and chitosan in the final nanosystem (Figure 5D; Figure S4, Supporting Information). TGA estimated a total dry matter of 7.0 mg ml⁻¹. FAAS estimated a gold amount of 0.2 mg mL⁻¹ of gold. The rheology in terms of viscosity of the solution was estimated by varying the temperature from +4 °C to +40 °C: the result is a constant decrease of the viscosity with increasing temperature. These findings indicate that GNRs@Chit were made of 2% Au and 98% chitosan, and that this solution is suitable for intravesical instillation.

Both the Apt‐Itg and the unrelated‐conjugated nanosystems displayed a reduction in the Z‐potential (+10.8 mV and +10.3 mV respectively, versus +40.0 mV measured for GNRs@Chit), according to the reduced amount of free amino groups on chitosan due to the presence of the PEG linker.

Moreover, the crystallinity and the composition of the GNRs@Chit‐Apt‐Itg/unrelated was proven to be conserved after the consecutive conjugation steps, additionally revealing no detectable Br atoms from CTAB. GNRs@Chit‐Apt‐Itg/unrelated was therefore aliquoted in vials, freeze‐dried and vacuum sealed to achieve sterile single use vials containing the lyophilized product.

2.7. GNRs@Chit‐Apt‐Itg is Stable in Human Urine

Urine represents a harsh environment, characterized by a complex mixture of metabolites,^{[ 26 ]} and bacteria,^{[ 27 ]} that could interact with the nanoparticles. Therefore, we estimated the potential impact of urine on the light absorption of the three formulations of GNRs (GNRs@CTAB, GNRs@Chit, and GNRs@Chit‐Apt‐Itg), and the stability of the GNRs@Chit‐Apt‐Itg in urine in a time frame of 2 h.

In human urine, the longitudinal LSPR peak of GNRs@Chit‐Apt‐Itg was observed at 834 nm, consistently with the LSPR peak of the unconjugated GNRs@Chit that was 836 nm for the specifically used batch (Figure 5E; and Table S1, Supporting Information). The stability of the optical properties of GNRs@Chit‐Apt‐Itg dispersed in urine has been studied over time, showing 96% stability up to 1 h and 7% decrease of light absorbance after 2 h (Figure 5F; and Table S2, Supporting Information).

2.8. GNRs@Chit‐Apt‐Itg are Monodispersed

We compared the properties of GNRs@Chit‐Apt‐Itg versus our previously produced GNRs@Chit‐Iso4. The magnitude of the zeta potential is an indication of the potential stability of the colloidal system, as particles with large negative or positive zeta potential tend to repel each other and there will be no tendency for the particles to come together.^{[ 28 ]} We observed a more positive Z‐potential for GNRs@Chit‐Apt‐Itg versus GNRs@Chit‐Iso4, suggesting better colloidal property: the difference in Z‐potential is probably due to the minor amount of aptamer that needs to be conjugated in comparison to Iso4 (1.14 pmol of aptamer, this study, versus 34 µmol of Iso4,^{[ 11 ]}) to get a proper biological targeting that leaves more numerous protonated amino groups free onto the chitosan, which ensure higher Zeta potential and a consequent higher colloidal stability. In agreement, the polydispersity index (PDI, used to describe the degree of “non‐uniformity” of a distribution) was lower for GNRs@Chit‐Apt‐Itg (Table  1 ). Likewise, also the histogram representation showed more uniform distribution of the size of the GNRs@Chit‐Apt‐Itg versus GNRs@Chit‐Iso4 that was composed by several clusters of size distribution (Figure S5, Supporting Information).

Table 1.

Analysis of the potential and of the hydrodynamic range of GNRs.

	Z potential [mV]	Electrophoretic mobility [mmcm/Vs]	Z‐average [nm]	Polydispersity index (PdI)
GNRs@Chit‐Apt‐Itg	+32.5	+2.54	881.2	0.489
GNRs@Chit‐Iso4	+11.1	+0,87	150.7	0.709

Electrophoretic light scattering and dynamic light scattering were performed by an accredited laboratory, according to ISO‐13099‐2:2012 and ISO 22 412:2017 (www.alfatestlab.com).

2.9. The Apt‐Itg Delivers the GNRs into Lysosomes

The integrin α5β1 is targeting the lysosomal compartment after binding to its natural ligands.^{[ 29 ]} By transmission electron microscopy we followed the pathway of GNRs delivered by the Apt‐Itg, by incubating integrin α5β1+ MB49 cells with 100 µm GNRs@Chit‐Apt‐Itg. Fifteen minutes after treatment the nanoparticles were located on the cell membrane, were internalized within 2 h and accumulated in lysosomes after 4 h (Figure 5G). Noteworthy, despite being in the lysosomal acidic compartment the GNRs still maintained their characteristic shape, in agreement with in vitro data that supports the stability of GNRs in low pH conditions (Figure S6, Supporting Information).

2.10. The Apt‐Itg does not Modify the PA Spectra of GNRs

To assess the PA signal of GNRs@Chit after the conjugation with the Apt‐Itg, the the PA properties of the GNRs were investigated using agar drops containing GNRs@Chit‐Apt‐Itg. The GNRs@Chit‐Apt‐Itg were detectable with PA imaging and had a bell‐shape PA spectrum in the NIR‐I with a peak at 836 nm (Figure  6A). These data agree with the information obtained in the VIS‐NIR spectra (Figure 5E–F), thus demonstrating that the functionalization of the GNRs@Chit with the Apt‐Itg does not modify the optical and photoacoustic properties of the GNRs@Chit in the NIR‐I.

Figure 6.

In vivo specificity and stability of the PA signal of GNRs@Chit‐Apt‐Itg. A) Photoacoustic imaging of agar drop containing GNR@Chit‐Apt‐ITG (6 mg Au; 30 nmol Au, red), with the light blue region of interest that delineates the PA signal of GNRs from which the photoacoustic spectra of GNRs@Chit‐Apt‐Itg was derived. B) Axial frame of a murine bladder without tumor, with a region of interest (red circle) in which a photoacoustic spectrum has been analysed 15 min after the intravesical instillation of GNRs@Chit‐Apt‐Itg (100 µl of 100 µm Au); the identified PA spectra is ascribed to that of the oxygenated blood. C) Axial frame of a murine bladder with a US recognizable tumor nine days after intravesical instillation of MB49‐Luc cells; 15 min after the intravesical instillation of untargeted GNRs@Chit (100 µl of 100 µm Au) the photoacoustic signal of GNRs was searched on the surface of the tumor mass, and the photoacoustic spectra from the region of interest compared with the photoacoustic spectra of GNRs. D) Axial frame of a murine bladder with a US recognizable tumor nine days after intravesical instillation of MB49‐Luc cells, with the region of interest on the surface of the tumor mass in which a photoacoustic spectrum has been analyzed, showing an unspecific PA spectrum. The PA imaging, spectra and signal were also analyzed from the surface of the tumor mass at different time points after the intravesical instillation of GNR@Chit‐Apt‐Itg (100 µl at concentration of 100 µm Au) (panels E‐I). Data from one representative animal are shown. J) Quantification of the percentage of the photoacoustic signal of GNR@Chit‐Apt‐Itg (100 µl at concentration of 100 µm Au) calculated on the tumor volume over time, from five independent animals. Data are expressed as fold difference versus the percentage of the photoacoustic signal quantified 15 min after intravesical instillation and two intravesical washes of the bladder with saline solution. K) Quantification of gold measured in three animals with different tumor mass, all visible by US imaging, 15 min after intravesical instillation of GNR@Chit‐Apt‐Itg (100 µl at concentration of 100 µm Au), and the amount of gold expressed as ratio versus the tumor surface calculated on the PAUS 3D acquisition. Each dot represents a single animal, lines and bars show means and SEM. PAUS imaging of the three animals are shown in Figure S8 (Supporting Information). L) Histological analysis of murine bladder with orthotopic tumor 24 h after the intravesical instillation of GNR@Chit‐Apt‐Itg (100 µl at concentration of 100 µm Au), with a magnification of the non‐neoplastic tissue (dashed box). The ultrasound and PA images shown in panels A‐I were acquired with a step size of 0.2 mm.

2.11. Specificity and Stability of the Photoacoustic Signal of GNRs@Chit‐Apt‐Itg for the Recognition of Orthotopic Bladder Cancer

To assess the specificity of GNRs@Chit‐Apt‐Itg for the detection of the orthotopic bladder tumor the nanoparticles were intravesically instilled in the preclinical model without and with the orthotopic tumor. When the luminal part of the bladder was investigated, in the tumor‐free bladder the PA signal of GNRs@Chit‐Apt‐Itg was not present, while vascularization of the bladder wall was detectable through the identification of the PA signal of oxygenated blood (Figure 6B). Likewise, absence of PA signal ascribed to the GNRs was detected upon the intravesical instillation of untargeted GNRs@Chit in tumor bearing animals (Figure 6C).

2.12. Stability of the Photoacoustic Signal of GNRs@Chit‐Apt‐Itg for the Recognition of Orthotopic Bladder Cancer

When the tumor mass was present and detectable by US imaging before the intravesical instillation of targeted GNRs@Chit‐Apt‐Itg, the periphery of the tumor mass was negative for the spectra of GNRs, and only an unspecific PA signal was present on the surface of the tumor (Figure 6D), with a PA spectra superimposable to that observed in the tumor bearing animal treated with the untargeted GNRs@Chit reported in Figure 6C. Fifteen minutes after the intravesical instillation of the targeted GNRs, followed by two intravesical washes with saline solution to remove the unbound fraction of nanoparticles, the PA spectra of GNRs@Chit‐Apt‐Itg was located only on the surface of the tumor mass, but not on the surrounding non‐neoplastic urothelium (Figure 6E).

The PA signal of GNRs@Chit‐Apt‐Itg was monitored up to 24 h after their intravesical instillation. Specific PA signal and spectra of the targeted GNRs were always present only on the surface of the tumor mass until 3 h post instillation. At 24 h post instillation the PA signal of GNRs was lost, and the same PA spectra that was detected before the intravesical instillation of GNRs was monitored (Figure 6F–I). This information was reproduced and quantified in five animals: the PA signal of GNRs on the tumor surface was detectable up to 3 h after the intravesical installation of the targeted GNRs, despite a reduction at 3 h post instillation (Figure 6J).Of note, to prevent the risk of death of the animal due to 3 h of anesthesia, the animal was recovered from the anesthesia and placed in the cage after the initial 15 min, and at the following three time points; during this time frame the animal was moving and able to urinate, indicating that during the 3 h the binding and chemical‐physical properties of the targeted GNRs was not influenced by the accumulation of urine in the bladder.

2.13. The Amount of Gold Bound to the Tumor Mass is Proportional to the Bladder Cancer Surface

To assess the reproducibility of this solution, we quantified the amount of gold bound onto the tumor surface. Three animals with tumor mass of different size were identified through US imaging, then treated with the same amount of GNRs@Chit‐Apt‐Itg corresponding to 1970 ng of gold. The amount of gold bound to the tumor mass was proportional to the tumor surface (Figure S7, Supporting Information), and with a density of gold per mm² of tumor surface equal to 1.8 ± 0.17 ng mm⁻² (Figure 6K).

2.14. Absence of Cytotoxicity by GNRs@Chit‐Itg

In vivo and in vitro tests were carried out to assess the safety of the treatment with GNRs@Chit‐Apt‐Itg. The day after intravesical instillation of targeted GNRs in mouse bladders containing the tumor the histological analysis showed the presence of an intact bladder wall with no signs of inflammation, and the urothelium composed by 2–3 layers of urothelial cells with no signs of cell proliferation (Figure 6L). Cytotoxicity test was also carried out on the integrin α5β1+ target cells that are the murine bladder cancer cell line MB49‐Luc, used to induce the orthotopic tumor. The concentrations tested were up to 100 µm Au, because this concentration already represents the condition in which an excess of gold is intravesically instilled (1970 ng of gold instilled, and density of gold per mm² of tumor surface equal to 1.8 ± 0.17 ng mm⁻², reported in Figure 6K). Gold concentrations above 100 µm Au would represent a more expensive condition and no more effective than 100 µm Au. The concentration of 100 µm Au represents the condition that will be tested in the clinic in the future. Cytotoxicity test was carried out with the GNRs@Chit‐Apt‐Itg and the diluent and evaluated 15 min and 24 h after treatment, and both conditions did not have impact on cellular metabolic activity (Figure S8, Supporting Information).

2.15. Early Detection of Orthotopic Bladder Cancer

Using the clearly visible tumors reported above, we demonstrated both the specificity of GNRs@Chit‐Apt‐Itg to detect only the neoplastic cells and the in vivo feasibility of the approach. Tests were repeated to investigate the use of GNRs@Chit‐Apt‐Itg for the detection of tumor lesions at the very early stage, when not yet visible by US imaging (Figure  7 ). Fifteen minutes after the intravesical instillation of the nanoparticles two tissue spots smaller than 1 mm were detected by searching for the specific PA signal of GNRs@Chit‐Apt‐Itg. One hour later the specific PA signal remained in the same location, from which a tumor mass of 2.4 µl was evident the day after and of 20 µl 4‐day after the instillation of GNRs.

Figure 7.

Early detection of bladder cancer with GNRs@Chit‐Apt‐Itg. GNR@Chit‐Apt‐Itg (100 µl at concentration of 100 µm Au) were instilled intravesically in mouse seven days after intravesical instillation of MB49‐Luc cells. PAUS imaging was acquired 15 min and 1 h after the instillation of GNRs and 2 intravesical washes with saline buffer A,B). US imaging was acquired 1 and 4 days after the intravesical instillation of GNRs C,D). The arrows indicated the PA signal (green spots) specific for the GNRs@Chit‐Apt‐Itg after unmixing the total PA signal for the PA signal of melanin, deoxy‐ and oxy‐genated blood and GNRs; the spots indicated by the arrow are those with the PA spectra ascribed to the GNRs. One representative test out of five is shown. The ultrasound and PA images were acquired with a step size of 0.2 mm.

The preclinical model of murine orthotopic bladder cancer based of the intravesical instillation of the syngeneic murine bladder cancer cell line MB49 model is characterized by the early establishment of clusters of a few MB49 cells attached to the luminal side of the urothelium, followed by the onset of sessile carcinoma that develops toward the bladder lumen, and pedunculated tumors that also infiltrate the lamina propria.^{[ 30 ]} Our data agree with the biology of tumor development in this preclinical model, showing that GNRs@Chit‐Apt‐Itg detects tumor lesions of few hundred microns that give rise to neoplasia in the following few days.

3. Discussion

This study aimed to identify and validate a reproducible and cost‐effective solution for the intravesical delivery of targeted GNRs for the recognition of orthotopic bladder tumor expressing the integrin α5β1. To this aim, by rationale truncation and chemical modification of the existing 68‐mer RNA aptamer against integrin α5β1,^{[ 15 ]} a new 34‐mer 2′‐FPy‐containing RNA aptamer (Apt‐Itg) was delivered, and its cell targeting and uptake were extended to human and murine bladder cancer cells. The optimized aptamer was further modified at the 5′‐end by adding a Thiol‐C6 group to allow efficient binding on the gold surface of GNRs@Chit. Indeed, dose‐dependent amount of Apt‐Itg conjugated to the GNRs@Chit was obtained when three different concentrations of ligand were used during the process of synthesis of the GNRs@Chit‐Apt‐Itg. Using three different concentrations of Apt‐Itg during the synthesis process, we demonstrated an average conjugation efficiency of 5%, in agreement with previous report,^{[ 31 ]} and a deeply reduced batch‐to‐batch variation.

The feasibility of using intravesical instillation of nano‐ and micro‐particles for the photoacoustic imaging of bladder cancer has been previously demonstrated, such as in our recent report that used GNRs@Chit‐Iso4,^{[ 11 ]} and silicon dioxide microparticle,^{[ 32 ]} respectively. Indeed, both systems are biased by settling down to the bottom side of the bladder, and thus they require solutions to keep them in suspension in the bladder.

By delivering nanoparticles that have positive charge and uniform size distribution, the nanosystem GNRs@Chit‐Apt‐Itg has overcome the above bias, since does not require external solutions to keep the nanoparticles in suspension inside the bladder. Furthermore, the system we are proposing has the advantage of using an aptamer for the selective recognition of integrins. The aptamers are obtainable through a consolidated, highly reproducible, large‐scale synthesis and very economical prices compared to monoclonal antibodies and certain complex peptides such as the cyclic‐one. All these characteristics are very important considering future scale‐ups to obtain quantities suitable for clinical trials.

Another innovative aspect of this study is the stability of the nanosystem and its binding to the target cells in the urinary environment. The optical properties of the GNRs@Chit‐Apt‐Itg are stable and maintained for at least 2 h in the presence of urine, which allow for the detection of the neoplastic lesions up to 3 h after their intravesical instillation. The implication of these features is that the in vivo PA signal of the GNRs could be followed over time using the parameters identified during their production, and in the clinical scenario allow sufficient time for their identification during intravesical surgery.

The integrin α5β1 is targeting the lysosomal compartment after binding to its natural ligands, such as fibronectin.^{[ 29 ]} As it occurs with fibronectin, our Apt‐Itg that binds the α5β1 integrin was also found in the lysosomes, indicating that it follows the same path for integrin recycling induced by natural ligands. The internalization of the fluorescent cargo into lysosomes provided by the Apt‐Itg also paves the way for delivering intravesical therapy targeted against integrin α5β1⁺ high‐grade bladder cancer cells. Current intravesical therapies for bladder cancer, either immunotherapy (i.e., Bacillus of Calemtte Guerin) or chemotherapy (i.e., Mitomycin C), are not targeted to tumor cells, and can have side effects.^{[ 33 ]} The nucleotide sequence of Apt‐Itg offers the possibility to conjugate drugs through an organic linker that is activated by the acidic environment of the lysosomes.^{[ 34 ]} The Apt‐Itg offers the possibility to deliver a therapeutic cargo directly into high‐grade bladder cancer cells expressing the integrin α5β1, which is expressed by 81% of human high‐grade NMIBC.^{[ 11 ]} Likewise, the Apt‐Itg also delivers the GNRs into lysosomes of integrin α5β1+ cells, indicating that the conjugation with the nanoparticle did not influence its functionality. It has been reported that accumulation of gold nanoparticles in lysosomes causes impairment of lysosome degradation capacity.^{[ 35 ]} This pathway could therefore induce toxicity in stromal cells expressing integrin α5β1, for example following intravenous instillation of GNRs targeted against integrin α5β1. Instead, intravesical delivery of targeted nanoparticles is overcoming this potential side effect because high‐grade bladder cancer cells but not normal urothelial cells express α5β1 integrin.^{[ 36 ]} The integrin α5β1 was reported to be expressed also in other neoplasia of epithelial origin such as ovarian and cervical tumors in which 80% and 84% of tumors, respectively, are positive for the expression of integrin α5β1 that significantly correlates with higher clinical stage.^{[ 37 ]} The application of targeted GNRs, including GNRs@Chit‐Apt‐Itg, in hollow organs would target tumor cells expressing the α5β1 integrin, while not recognizing normal epithelial cells; this approach should overcome the occurrence of side‐effect due to accumulation of gold in lysosomes of non‐neoplastic α5β1+ cells. Furthermore, concerning bladder cancer the intravesical instillation of GNRs@Chit‐Apt‐Itg would highlight the presence of small tumors and provide the opportunity to the surgeon to resect the lesion.

GNRs functionalized with an aptamer for the application in the urinary environment have never been reported. The nanosystem GNRs@Chit‐Apt‐Itg represents a novel contrast agent for the detection of α5β1⁺ bladder cancer with high specificity. The same GNRs that we have employed here can also be used for the application of photothermal therapy (PTT),^{[ 11} , ^{38 ]} which is obtained by using continuous laser at the same wavelength used by the pulsed laser to carry out the photoacoustic imaging. The wavelength of irradiation of GNRs@Chit‐Apt‐Itg is 836 nm, which is in the biological optical window (near infrared region‐I at a wavelength of 680–970 nm) and is approved for clinical practice by the American National Standards Institute.^{[ 39 ]}

The intravesical delivery of the nanosystem GNRs@Chit‐Apt‐Itg offers unmet theranostic opportunities for bladder cancer, and being unrelated to cellular metabolism, it will overcome bias in the gender medicine,^{[ 40 ]} will be equally effective in male and female bladder cancer patients,^{[ 41 ]} is not impaired by the use of antibiotic,^{[ 42 ]} and is also suitable for frail patients who may have troubles in tolerating the standard of care for high‐grade NMIBC that is the intravesical instillation of Bacillus Calmette‐Guerin.^{[ 43 ]} The possibility of detecting small high‐grade bladder cancer lesions that express the integrin α5β1 allows to identify and treat the residual high‐grade disease found after the first trans urethral resection of the bladder tumor (TURBT),^{[ 44 ]} and is associated with high or very high risk of relapse and progression.^{[ 45 ]} Our solution is applicable during the TURBT procedure, to detect small lesions that are not recognized by TURBT. This solution can be applied to a variety of NMIBC patients, such as i) at the first TURBT to detect the smaller lesions and thus to provide a more accurate diagnosis to drive the follow‐up: ii) in a second‐look TURBT to identify the residual disease: iii) in relapsing patients; and iv) to investigate the presence/absence of residual disease in patients enrolled in bladder‐sparing protocols.^{[ 46 ]} The acceptability of our solution in the urological clinical scenario is supported by several facts, as it does not use ionizing agents, but uses gold as biocompatible metal that is delivered locally into the bladder through intravesical instillation and an incubation period of the contrast agent followed by intravesical washes, as it is carried out for another agent of current use that is Hexvix.^{[ 8 ]}

This study delivers a solution for the recognition of bladder tumors expressing the integrin α5β1 that exploits the exquisite selectivity of Apt‐Itg for these tumors. The use of the aptamer as functionalization agent of the nanoparticles only requires chemical synthesis of the aptamer and PCR as quality control test on the functionalized GNRs, making the production of GNRs@Chit‐Apt‐Itg cheaper than using other ligands. Being the GNRs@Chit‐Apt‐Itg used as contrast agent for photoacoustic imaging that provides spatial resolution in the range of 50–100 microns^{[ 47 ]} this novel, cost‐effective, highly urine‐stable and reproducible nanosystem allows for the detection of small tumor areas that are currently undetectable.

4. Experimental Section

Ethics Approval

All procedures and studies involving mice were approved by the Institutional Animal Care and Use Committee of San Raffaele Scientific Institute and performed according to the prescribed guidelines (IACUC, approval number 1397).

Tumor Cell Lines

All human bladder cancer cell lines were from American Type Culture Collection (ATCC, Manassas, VA) and cultured according to instructions. The murine bladder cancer cell line MB49‐Luciferated (MB49‐Luc) were kindly provided by Prof. Carla Molthoff (VU University Medical Center, The Netherlands) and used as recently reported.^{[ 11 ]}

Murine Orthotopic Bladder Tumor Model

The orthotopic bladder tumor in mice develops after the intravesical instillation of murine bioluminescent MB49‐Luc bladder cancer cells, as recently reported.^{[ 11 ]} Briefly, female albino C57BL/6 J mice (9 weeks old, weighing ≈20 g, Charles River Laboratories, Italy) were anesthetized with ketamine (80 mg kg⁻¹) and xylazine (15 mg kg⁻¹) and kept in dorsal position. Using a 24‐gauge catheter, the bladder of each mouse was emptied and instilled with or without MB49‐Luc cells (50.000 cells per 100 µl in Dulbecco's phosphate‐buffered saline, DPBS, containing 1 mm MgCl₂ and 0.5 mm CaCl₂). Thirty minutes later, the catheter was removed, and mice were allowed to recover and return to their cage. The sex of the animals was chosen based on the principles of the 3Rs (Replacement, Reduction, and Refinement). Compared to the male, the female has a shorter urethra, does not present urethral stricture due to the absence of the prostate gland and the conformation of the urethra is more linear and less convoluted. The use of female animals simplifies the procedure and minimizes mechanical trauma and the potential risk of infections. The use of female animals allows for less troublesome intravesical instillation and greater success in tumor implantation and reproducibility of the treatment, therefore a reduction in the number of animals used.

Aptamers

The SH‐terminated 2′Fluoro‐Pyrimidine (2′F‐Py)‐RNA aptamer targeting integrin α5β1 (Apt‐Itg) and unrelated aptamer were purchased from Metabion (Martinsried, Germany). The sequences of aptamers modified with Thiol‐C6 group were as follows:

Apt‐Itg: 5′‐Thiol‐C6‐G(2′F‐C)GGA(2′F‐C)GGA(2′F‐C)AGAGAG(2′F‐U)G(2′F‐C)AA(2′F‐C)(2′F‐C)(2′F‐U)G(2′F‐C)(2′F‐C)G(2′F‐U)G(2′F‐C)(2′F‐C)G(2′F‐C)‐3′.

Unrelated: 5′‐Thiol‐C6‐(2′F‐U)(2′F‐U)(2′F‐C)G(2′F‐U)A(2′F‐C)(2′F‐C)GGG(2′F‐U)AGG(2′F‐U)(2′F‐U)GG(2′F‐C)(2′F‐U)(2′F‐U)G(2′F‐C)A(2′F‐C)A(2′F‐U)AGAA(2′F‐C)G(2′F‐U)G(2′F‐U)(2′F‐C)A‐3′.

Protein Extraction and Immunoblot

Cells were lysed in a buffer containing 50 mm HEPES (pH 7.5), 150 mm NaCl, 1% glycerol, 1% Triton X‐100, 1.5 mm MgCl₂, 5 mm EGTA, supplemented with 1 mm Na₃VO₄ and protease inhibitors (Roche Diagnostics, Indianapolis, USA). After incubation on ice for 20 min, the extracts were clarified by centrifugation at 13200 rpm for 10 min at 4 °C. Protein concentration was determined by the Bradford colorimetric assay (Bio‐Rad, Hercules, CA) using bovine serum albumin as the standard. SDS‐PAGE was carried out according to Laemmli. Gels were electroblotted into polyvinylidene difluoride membranes (Amersham BioSciences, UK) and filters were probed with anti‐Integrin beta 1 (clone EPR16895, Abcam, Cambridge, UK), anti‐Integrin alpha 5 (clone EPR7854, Abcam) or anti‐α‐tubulin (DM1A, Cell Signaling Technology, Danvers, MA, USA) primary antibodies, followed by the HRP‐conjugated secondary antibodies. Densitometric analysis was performed on at least two different expositions to assure the linearity of each acquisition using ImageJ (v1.46r).

Integrin α5 Silencing

Human T24 and RT112, and murine MB49‐Luc cells (1.8 × 10⁵) were seeded in 6‐well plates and after 24 h were overlaid with the transfection mixtures containing 60 nm small interfering RNAs (siRNAs) targeting murine and human integrin α5, respectively, and Lipofectamine RNAiMAX Reagent (ThermoFisher Scientific) in Opti‐MEM I reduced serum medium (Invitrogen), according to the manufacturer's instructions of transfection reagent. siRNAs (Qiagen, Hilden, Germany) tested were: HS_ITGA5_7, HS_ITGA5_5, HS_ITGA5_4 and HS_ITGA5_2 (T24 cells); HS_ITGA5_7 and HS_ITGA5_5 (RT112 cells); MM_ITGA5_1, MM_ITGA5_2, MM_ITGA5_3 and MM_ITGA5_4 (MB49‐Luc cells). Scrambled non‐targeting siRNA (siRNA ctrl, Qiagen, Hilden, Germany) was used as a negative control. After 5 h incubation, complete culture medium was added to the cells and incubation was prolonged up to 24 h. Silencing efficiency was assessed by immunoblotting.

Cell Binding Assay

Binding of Apt‐Itg to human T24 and RT112, and murine MB49‐Luc cells was assessed by flow cytometry. Apt‐Itg and unrelated aptamer were internally labelled with Alexa Fluor 647 fluorescent probe by using the Ulysis Nucleic Acid Labeling Kit (Invitrogen), which provides a non‐enzymatic method for chemically labeling purine bases in the entire sequence with the fluorescent dye.

Cells were detached from culture plates with 0.02% EDTA and washed twice with 500 µl DPBS. Cells (2.0 × 10⁵) were left untreated or incubated with increasing concentrations of Alexa Fluor 647‐Apt‐Itg or unrelated aptamer in binding buffer (DPBS supplemented with 1 mg ml⁻¹ yeast tRNA and 1 mg ml⁻¹ ultrapure salmon sperm DNA as nonspecific competitors), with or without 1 mm MgCl₂ and 0.5 mm CaCl₂. After 10 min at room temperature (RT), cells were washed three times with 100 µL DPBS, suspended in 100 µl DPBS and analyzed by flow cytometry (BD Accuri™ C6). Data analysis was performed using FlowJo software (version 10.0.7). The geometric mean fluorescence intensity (gMFI) of aptamer‐treated samples was normalized to the gMFI of the cells alone and the apparent dissociation constant (Kd) of the aptamer‐cell interaction was obtained by fitting the dependence of relative gMFI on the concentration of aptamer to the one‐site saturation equation Y = Bmax × X/(Kd + X), using GraphPad Prism version 6.00, as previously described.^{[ 48 ]} Binding curves represent data from at least two independent experiments.

Binding of 1 µm Alexa 647‐Apt‐Itg to human and murine bladder cancer cell lines transfected with HS_ITGA5_7 and MM_ITGA5_4 (both referred to as si‐ITGα5), respectively, was performed after 24 h from transfection using flow cytometry, as described above.

For cell binding of anti‐integrin α5 antibody (Ab‐Itg), 2.0 × 10⁵ T24 cells were detached from culture plates as described above and left untreated or incubated with Human Integrin alpha 5/CD49e Alexa Fluor488‐conjugated Antibody (dilution 1:10, R&D system, Minneapolis, MN) in DPBS supplemented with 0.1% bovine serum albumin. After 20 min at RT, cells were washed three times in DPBS, suspended in 100 µl DPBS and analyzed by flow cytometry, as described above.

Cellular Uptake of Apt‐Itg

Uptake of Apt‐Itg into human T24 and murine MB49‐Luc cells was assessed by confocal microscopy. Cells (5.0 × 10⁴ cells per well in 24‐well) were seeded onto glass coverslips and after 24 h were left untreated or incubated for 1 h at 37 °C with LysoTracker Red DND‐99 (1:1000, Invitrogen, Carlsbad, CA) in the complete growth medium. Cells were then incubated at 4 °C or 37 °C with 4 µm Alexa 647‐Apt‐Itg or Unrelated aptamer, washed three times with DPBS, fixed with 4% paraformaldehyde in DPBS for 30 min at RT, and incubated with WGA‐Alexa Fluor 488 conjugate (WGA‐488, Invitrogen, Carlsbad, CA, USA) for 30 min at RT. Cells were then washed three times with DPBS and then subjected to nuclear staining with 1.5 µm 40,6‐Diamidino‐2‐phenylindole (DAPI, D9542, Sigma‐Aldrich, Milan, Italy) and mounted with glycerol/DPBS. Samples were visualized by Zeiss LSM 700 META confocal microscopy equipped with a Plan‐Apochromat 63x/1.4 Oil DIC objective. Images were taken with the same parameters and MFI was evaluated by Zeiss software on a minimum of 80 cells for each time point (40 and 120 min) in seven randomly picked fields.

Transmission Electron Microscopy (TEM)

MB49‐Luc cells incubated with 100 µm GNRs@Chit‐Apt‐Itg were prepared for TEM imaging as follows. After exposure to nanoparticles for 15 min, 2 h, and 4 h, the cells were fixed in a monolayer using 2.5% glutaraldehyde in 0.1 m cacodylate buffer (pH 7.4) for 30 min at RT, then scraped and pelleted. The pellets were subsequently fixed for 24 h at 4 °C and post‐fixed for 1 h in a reduced osmium solution (1.5% potassium ferricyanide and 1% aqueous osmium tetroxide in 0.1 m cacodylate buffer). Following the initial incubation with the heavy metal‐based solutions, the pellets were washed with bi‐distilled water at room temperature, then immersed in a 0.5% uranyl acetate solution and left overnight at 4 °C. The samples were dehydrated using a graded ethanol series (70%, 80%, 90%, and 100%) and acetone for 10 min each, before being embedded in Epon resin. After curing at 60 °C for 48 h, thin sections were cut using a Leica UC7 ultramicrotome (Leica Microsystems, Vienna, Austria). The sections were mounted onto 300‐mesh copper grids and imaged using a Talos L120C G2 transmission electron microscope (Thermo Fisher Scientific Inc., Waltham, MA, USA) at an acceleration voltage of 120 kV.

Cytotoxicity Assay

MB49‐Luc cells were plated in 96 well culture plates (30.000 cells in 200 µl of culture medium each well) and incubated in CO₂ incubator at 37  C. The day after the culture supernatant was removed and cells were treated with culture medium added of GNR@Chit‐Apt‐Itg or the diluent (GNRs@Chit‐Apt‐Itg were suspended with 1% acetic acid /PBS to obtain a stock of 1 mm Au). Fifteen minutes after the treatment was removed, the cells were washed twice with culture medium and assayed for cytotoxicity or returned to culture for an additional 24 h. MTT assay was carried out as detailed by the producer (SIGMA‐Aldrich) and the absorbance was read at 570 nm on a Multimode microplate reader (Mithras LB 940, Berthold Technologies).

Synthesis of GNRs@CTAB with Maximal Absorption in the NIR Range

Cetyltrimethylammonium bromide (CTAB)‐coated GNRs (GNRs@CTAB) with a maximum absorption at 780–840 nm, were prepared according to a previously published procedure with minor modifications and scaled to a volume of 2 liters. The initial solution for the growth of GNRs was prepared by dissolving CTAB (31.14 g, 85.4 mmol) and sodium oleate (4.28 g, 14.0 mmol) in 1.8 L of warm water (≈50 °C) in a thermostatically controlled 2‐L jacketed reactor, equipped with mechanical stirring. After the dissolution of the solids, the solution was let to cool down to 30 °C. When the solution reached 30 °C, 810 µL of AgNO₃ solution (0.4 m in ultrapure water) were added and the mixture was then incubated for 15 min without stirring. Then, under continuous stirring (700 rpm), 8.652 mL of HAuCl₄ solution (0.1 m in ultrapure water) were added. Au(III) was reduced to Au(I) thanks to sodium oleate for 90 min, after which 3.56 mL of hydrochloric acid (37%) and 3.6 mL of ascorbic acid (0.079 M) were added to adjust the pH to 1.0 and to ensure the complete reduction of the gold precursor. Separately, a seed solution was prepared by dissolving 364 mg (1.0 mmol) of CTAB in 10 mL of warm water in a 50 mL round‐bottomed flask. After cooling down to room temperature 25 µL of HAuCl₄ solution (0.1 m) were added while stirring. The seeds were formed by quickly injecting 600 µL of ice‐cold sodium borohydride (0.01 m) into the solution, causing the color of the solution to change from yellow to brown, indicating the formation of ultra‐small gold seeds. Finally, after ageing the seed solution for 30 min at room temperature, 690 µL of it were added to the growth solution, which was vigorously stirred for 30 s then left undisturbed overnight at 30 °C to allow for GNRs growth. Purification of GNRs@CTAB was performed by i) centrifugation in 50 mL Falcon tubes (6000 rpm for 100 min), ii) removal of 40 mL of the supernatant and iii) re‐dispersion in 40 mL of ultrapure water; the process was repeated 3 times. The final product was then collected in 200 ml of ultrapure water. The gold concentration in GNRs@CTAB was 2.09 mm, as determined by atomic absorption spectroscopy.

GNRs were designed for displaying peak absorption in the NIR biological window, placed between 780 and 840 nm. The synthesis of GNRs@CTAB performed in a 2L jacketed reactor was able to produce 200 mL of 2.09 mm GNRs@CTAB (Yield = 49%), determined by Flame Atomic Absorption Spectroscopy (FAAS), and a total dry matter of 0.8 mg mL⁻¹, measured by gravimetric analysis. This means that the gold content was a ≈52% of the nanosystem mass, in agreement with the fact that the CTAB completely covers the gold core and an excess of surfactant in solution was required to keep the nanosystem stable. The surface zeta potential was +29.8 mV, as expected for the model of CTAB‐capped gold nanosystem.

Coating GNRs with Chitosan (GNRs@Chit)

For this system, chitosan was used for coating the nanostructures: this biopolymer is nowadays extensively used in biomedical applications thanks to its biocompatibility and biodegradability; chitosan is also employed in the field of nanoparticles due to its capacity to stabilize GNRs during the synthesis process, not including any toxic reagent. Moreover, chitosan is full of NH₂ groups that allows the functionalization with amine‐reactive linkers to perform the conjugation with several drugs, antibodies or peptides. GNRs@CTAB were covered with chitosan by means of a thiolated‐chitosan conjugate. This thiol insertion happens on the single chitosan monomer and is performed through carbodiimide‐assisted coupling between amino groups of chitosan and the carboxylic acid moiety of thioglycolic acid. The effective modification of chitosan with thiol groups was confirmed by the means of ¹H‐NMR, as previously reported.^{[ 25 ]}

Medical grade chitosan (500 mg, 3.1 mmol of monomer, 88.3% deacetylated) was dissolved in 50 mL of 1 vol. % acetic acid and mixed with 0.5 mL (7.2 mmol) of thioglycolic acid under moderate stirring. N‐(3‐dimethylaminopropyl)‐N′‐ethylcarbodiimide hydrochloride (500 mg, 2.6 mmol) was then added to activate the carboxylic group of thioglycolic acid and to promote the coupling to the amino groups of chitosan. The reaction was left to incubate for 6 h at room temperature under continuous stirring. The product was then dialyzed using a 3.5 kDa cut‐off dialysis tube overnight against ultrapure water and the resulting thiolated‐chitosan was diluted to 500 mL with water. At this point, 30 mL of GNRs@CTAB were added dropwise under mild stirring and the resulting solution was incubated (48 h, room temperature) to allow for the coupling of thiolated‐chitosan and GNRs. The product was subsequently concentrated using an Amicon stirred cell equipped with PES membranes (100 kDa cut‐off, using 4 bar nitrogen pressure) to remove CTAB, and the final product (60 mL), called GNRs@Chit, was then split in aliquots, freeze dried and stored at +4 °C until the subsequent step.

The replacement of CTAB with chitosan first and the binding does not modify the shape, morphology, or optical properties of the GNRs, as they result perfectly preserved. The whole replacement of CTAB was assessed by ¹H‐NMR and by means of surface zeta potential measurement of purified GNRs that ranged from +35 to +45 mV. This increased zeta potential suggests that the exchange CTAB/Chit happened, being that the residual non‐functionalized amino groups of chitosan preserve their cationic nature in water.

Functionalization of GNRs@Chit with Aptamers

GNRs@Chit was functionalized with Apt‐Itg and unrelated aptamer. A heterobifunctional cross‐linker reagent was exploited, composed of an ethylene oxide spacer (PEG₁₂) bearing a N‐hydroxysuccinimidyl (NHS) ester to its extremities and a maleimide functional group (Maleimide‐PEG12‐NHS). The NHS ester terminus reacts with the free amino groups on chitosan ensuring the binding of the linker to the chitosan, while the maleimide group was available to react in a further step with the sulfhydryl group of the aptamer, expressly designed with the thiol at the 5′ end. An excess of Maleimide‐PEG₁₂‐NHS (10 mL, 1 mg ml⁻¹ in ultrapure water, 11.55 µmol, 10 mg) were mixed with 10 ml of GNR@Chit (1 mm of Au, 10 µmol, 1.969 mg Au) under stirring to achieve a final weight ratio of Maleimide‐PEG₁₂‐NHS:Au = 5:1. It was decided to calculate the amount of linker in ratio to gold and not to chitosan since it was more precise to determine gold (via AAS analysis) instead of chitosan (via TGA analysis). The mixture was then left to incubate overnight at RT and then dialyzed against ultrapure water using a 3.5 kDa cut‐off dialysis tube for 24 h, at RT, to remove the excess of the cross‐linker, so the maximum conjugation efficacy could be ensured for the aptamer. The product (25 ml), consisting of activated GNRs@Chit‐PEG₁₂‐maleimide (Figure S1, Supporting Information), was then mixed with pre‐established appropriate amount of aptamer (Apt‐Itg or Unrelated), previously dissolved in 1 ml of water and activated by three cycles at 85°‐ 0°‐ 37 °C. The mixture was left to react for 24 h. The day after, the solutions of GNRs@Chit‐Apt‐Itg were washed and concentrated by using Slide‐A‐Lyzer Dialysis Cassettes (3.5 kDa MWCO), freeze dried in sterile conditions and stored at +4 °C until shipping. Each vial, at the end of the fabrication, contains 98.48 µg of Au and has to be redissolved in 500 µL of water or saline to reach 1 mm [Au] final concentration. The dry content for the freeze‐dried vial is ≈3.5 mg.

Quantification of the Apt‐Itg Bound to the GNRs@Chit‐Apt‐Itg

To determine the amount of aptamers bound to GNRs@Chit‐Apt‐Itg, 5 out of 500 µL nanoparticle sample 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 95 °C for 5 min. The product from the reverse transcription reaction was PCR amplified with iQ SYBR Green Supermix (Bio‐Rad, Hercules, CA) in the presence of the Apt‐Itg or unrelated aptamer 5′ and 3′ primers. The qPCR protocol was as follows: the reverse transcription product was heated at 95 °C for 2 min, followed by 40 cycles of heating at 95 °C for 30 s, annealing at 55 °C for 30 s, and extending at 60 °C for 30 s. A melt curve stage by heating at 60–95 °C was performed. Reactions were all done in 25 µL volume in triplicate.

The primers used were:

Apt‐Itg: 5′ primer, 5′‐ AGCGGACGGACAGAG‐3′; 3′ primer, 5′‐GCGGCACGGCAGGTT‐3′.

Unrelated aptamer: 5′ primer, 5′‐TAATACGACTCACTATAGGGTTCGTACCGGGT‐3′; 3′ primer, 5′‐ TGACACGTTCTATGTGCA‐3′.

The quantity of the amplified product was extrapolated from a standard amplification curve obtained with unconjugated aptamers. Three independent experiments were performed.

Analysis of GNRs@Chit‐Apt‐Itg Stability in Human Urine

Half mL of water was added to a lyophilized vial of functionalized gold‐nanorods (GNRs@Chit‐Apt‐Itg), obtaining a concentration of 1 mm of gold.

GNRs@Chit‐Apt‐Itg (0.1 mL) were added to 3 mL of human urine from a healthy volunteer and subjected to VIS‐NIR spectroscopy (Cary 3500 UV–VIS–NIR modular spectrometer, Agilent Technologies, Santa Clara, USA) using a 1 cm path‐length plastic cuvette, collecting absorbance values between 500 and 1100 nm at various time points at 25 °C in order to check the non‐modification of maximum absorption wavelength and absorbance value over time. Undiluted human urine solution was used as reference for the analysis.

In Vitro PAI of GNRs

The PA spectra of GNRs@Chit‐Apt‐Itg was evaluated in vitro, using agar drop, as recently reported.^{[ 11 ]} Briefly, GNRs@Chit‐Apt‐Itg was dissolved to 1 mm as described for their in vivo use, then 30 µl of this solution were mixed with 30 µl of 1% agar solution), and the mixture was then poured on Parafilm M (Sigma) and left to solidify in a humidified chamber. The solidified product (called “agar drop”) was then placed in the ultrasound gel (Aquasonic 100, Parker). GNRs may undergo morphological changes when irradiated with a high threshold of pulsed laser light. After remodeling, GNRs lose their rod shape and become spherical, thus losing their PA property. To avoid reshaping when using pulsed light to obtain the PA signal of GNRs, light attenuators were used to reduce the laser fluence, as recently reported.^{[ 11} , ^{36 ]} The PAI was performed using the two optical fiber bearing light attenuators, with the US transducer placed perpendicularly on the object under investigation. Axial sections were acquired using the following settings for B‐Mode: 2D Power 100%; 2D Gain 13 dB; for PA‐mode: Pa Power 100%; PA gain 44 dB; TGC and depth were maintained identical for all drops. PA and US data were analyzed using VevoLab 5.8.1 software.

Intravesical Instillation of GNRs@Chit‐Apt‐Itg

Freeze‐dried GNRs@Chit‐Apt‐Itg were dissolved in 500 µl of 1% acetic acid in PBS (10 mm sodium phosphate buffer, pH 7.4, 138 mm sodium chloride, 2.7 mm potassium chloride, Sigma, P‐3813) containing 1 mm MgCl₂ and 0.5 mm CaCl₂ to obtain 1 mm solution of gold, then diluted to 100 µ of gold using PBS containing 1 mm MgCl₂ and 0.5 mm CaCl₂. Intravesical instillation of 100 µl at 100 µm of gold was carried out through a catheter, as reported.^{[ 11 ]} The only reported protocol modification was that, instead of using ultrasound‐mediated shaking, GNRs instilled into the bladder lumen were kept in suspension by flushing the bladder 3 times through a catheter that was maintained in the urethra for 15 min, a procedure that agrees with the current operations made in the surgery room where the surgeon flushes the bladder with saline solution.

In Vivo Ultrasound (US) and Photoacoustic Imaging (PAI) of Mice Instilled with GNRs@Chit‐Apt‐Itg

Photoacoustic (PA) and ultrasound (US) imaging was carried out on female albino C57BL/6 J mice bearing orthotopic bladder tumor following intravesical instillation with GNRs@Chit‐Apt‐Itg. High‐resolution US and PA imaging had been acquired using the Vevo F2 LAZR‐X platform (FUJIFILM VisualSonics, Inc.,Toronto, ON, Canada). The imaging platform includes a US system that operates in the frequency range of 1‐71 MHz (Vevo F2) combined with an Nd:YAG nano‐second pulsed laser with repetition rate of 20 Hz. The linear US transducer array UHF57x consists of 256 elements with a nominal center frequency of 40 MHz (57–25 MHz bandwidth) and a spatial resolution of 40 µm with a maximum imaging depth of18 mm. Light from the laser in the NIR 1 spectral range (680–970 nm) was delivered to the tissue through optical fibers mounted on either side of the transducer; since the GNRs were susceptible to change in shape at the higher laser threshold, light attenuators were prepared to reduce the laser fluence to avoid GNRs reshaping.^{[ 11 ]} During volumetric US‐PA acquisitions, a stepper motor was used for the linear translation of the US transducer and optical fibers along the sample, with a step size of 0.2 mm. The linear stepper motor moves in steps of a minimum of 0.1 mm while capturing 2‐D parallel images, for a maximum 3D range distance of 6.4 cm.

3D B mode scan was carried out for mouse bladder. The photoacoustic spectra between 680 and 970 nm were scanned with a step size of 5 nm; for the in vivo studies (murine bladder) the 3D multispectral PA scans were acquired selecting PA spectral curve of tissue components melanin, deoxy‐ and oxygenated blood, and GNRs; the processed wavelengths (680; 722; 764; 810; 924; 970 nm) were automatically selected from the spectral curve used to spectrally unmix the GNRs signal from other endogenous tissue chromophore signals such as oxy, deoxy hemoglobin. The algorithm reported by Luke et al.^{[ 49 ]} was used to select these wavelengths which were ideal for separating the signal from GNRs from other endogenous absorbers. Data analyses were conducted using the VevoLab software; volumes of interest were obtained by manually drawing Volumes of Interest (VOIs) on 3D B‐mode images. GNRs, melanin, oxy‐, and deoxy‐hemoglobin content were estimated through spectral unmixing analyses of the spectroscopic data.

Statistical Analysis

Statistical values were defined by unpaired two‐tail t‐test, using GraphPad Prism version 6.00. A p‐value < 0.05 was considered significant for all analysis.

Inclusion and Ethics Statement

All collaborators of this study had fulfilled the criteria for authorship were included as authors, as their participation was essential for the design and implementation of the study. Roles and responsibilities were agreed among collaborators ahead of the research. This research was not severely restricted or prohibited in the setting of the researchers, and does not result in stigmatization, incrimination, discrimination or personal risk to participants. Research relevant to this study were considered in citations.

Conflict of Interest

Alfano Massimo, Irene Locatelli, Chiara Venegoni, Erica Locatelli, Silvia Tortorella, and Mauro Comes Franchini have filed a patent application regarding the nanoparticles reported in this study (GNRs@Chit‐Apt‐Itg).

Supporting information

Supporting Information

Venegoni C., Tortorella S., Caliendo A., Locatelli I., Coste A. D., Locatelli E., Capancioni F., Bua E., Camorani S., Salonia A., Montorsi F., Jose J., Moschini M., Cerchia L., Franchini M. C., Alfano M., Urine‐Stable Aptamer‐Conjugated Gold Nanorods for the Early Detection of High‐Grade Bladder Cancer Residual Disease. Adv. Healthcare Mater. 2025, 14, 2403314. 10.1002/adhm.202403314

Data Availability Statement

All data and detailed methods to reproduce the information are reported in the manuscript and Supporting Information.