A Photoacoustic Contrast Nanoagent with a Distinct Spectral Signature for Ovarian Cancer Management

Abstract

Photoacoustic imaging (PAI) has tremendous potential for improving ovarian cancer detection. However, the lack of effective exogenous contrast agents that can improve PAI diagnosis accuracy significantly limits this application. This study presents a novel contrast nanoagent with a specific spectral signature that can be easily distinguished from endogenous chromophores in cancer tissue, allowing for high-contrast tumor visualization. Constructed as a 40 nm biocompatible polymeric nanoparticle loaded with two naphthalocyanine dyes, this agent is capable of efficient ovarian tumor accumulation after intravenous injection. The developed nanoagent displays a spectral signature with two well-separated photoacoustic peaks of comparable PA intensities in the near-infrared (NIR) region at 770 nm and 860 nm, which remain unaffected in cancer tissue following systemic delivery. In vivo experiments in mice with subcutaneous and intraperitoneal ovarian cancer xenografts validate that this specific spectral signature allows for accurate spectral unmixing of the nanoagent signal from endogenous contrast in cancer tissue, allowing for sensitive non-invasive cancer diagnosis. In addition, this nanoagent can selectively eradicate ovarian cancer tissue with a single dose of photothermal therapy by elevating the intratumoral temperature to ~49 °C upon exposure to NIR light within the 700–900 nm range.

Graphical Abstract

A nanoagent with a distinct photoacoustic spectral signature identifies ovarian cancer tumors via photoacoustic imaging and guides the subsequent application of phototherapy.

1. Introduction

Ovarian cancer has a high mortality rate because most patients are diagnosed at an advanced stage, when cancer cells have already spread to the abdominal cavity. This cancer is particularly challenging to diagnose because it is often asymptomatic or presents with symptoms similar to those of other disorders.^[1] There are currently no reliable tests for early ovarian cancer diagnosis. Therefore, accurate and sensitive imaging techniques are needed to improve ovarian cancer detection at early stages and evaluation of treatment outcomes.

Photoacoustic imaging (PAI) is an emergent technique that combines the superior spatial resolution of ultrasound imaging with the high contrast of optical imaging.^[2] As a result, it has tremendous potential for improving ovarian cancer detection. PAI operates by detecting sound waves generated by thermal expansion when photo-absorbers within the tissue are excited by pulsed laser light. Ultrasound waves exhibit significantly less scatter than the light utilized in traditional optical imaging techniques, allowing for superior resolution and sensitivity in deeper tissues compared to other optical modalities. PAI also avoids the use of harmful ionizing radiation, relying instead on photoacoustic signals from tissue components such as melanin, hemoglobin, and collagen (endogenous contrasts).^[3] Zhu et al. previously demonstrated that by using the photoacoustic characteristics of oxygenated and deoxygenated hemoglobin, PAI can provide more accurate detection of ovarian cancer tissue than conventional imaging techniques.^[3–6]

The diagnostic accuracy and sensitivity of PAI can be further improved by using exogenous contrast agents that absorb light and generate photoacoustic signals in the near-infrared (NIR) optical imaging window (700–2,500 nm), where light scattering and tissue attenuation are at their lowest.^{[4, 5]} To effectively distinguish themselves from endogenous contrast agents in cancer tissue, exogenous agents should exhibit distinct spectral characteristics (e.g., one or multiple sharp peaks), high molar extinction coefficients, photostability, and efficient conversion of light energy to heat.^[6] The photoacoustic signal intensity and distinct spectral features of an exogenous contrast agent are critical for the accuracy of PAI, especially as tissue imaging depth increases. The clinical potential of PAI has not yet been completely exploited owing to a lack of efficient exogenous agents capable of providing real-time photoacoustic contrast clearly distinguishable from background signal. Only a few FDA-approved dyes with sub-optimal photoacoustic properties, such as indocyanine green and methylene blue, have been utilized as PAI contrast agents in clinical studies.^{[6, 7]} To address this, a variety of exogenous contrast agents have been developed.^{[6, 8–10]} NIR dye-loaded liposomes,^[11] mesoporous silica nanoparticles,^[12] and iron oxide nanoparticles^{[13, 14]} are among the few exogenous contrast agents utilizing single-peak photoacoustic spectra of NIR dyes for ovarian cancer detection.^[15–17] One of the main challenges in obtaining reliable PAI data is the spectral overlap of the contrast agents with endogenous PA signals (poor signal-to-background ratio), which can produce false negative or false positive imaging outcomes, complicating target identification.^[18] Previous research suggests that contrast agents with multiple sharp peaks that are well separated are ideal for unmixing their photoacoustic signal from endogenous chromophores prevalent in tissue, resulting in a high signal-to-noise ratio.^{[19, 20]} For example, Peters et al. reported that bacteriochlorophyll α in purple bacteria with two well-defined peaks at 800 and 860 nm provides an excellent target spectrum for unmixing of bacterial signal from endogenous absorbers in cancer tissue.^[20] Thus, PAI contrast agents with distinct spectral signatures could be a promising solution for precise target identification, which is crucial for early detection of ovarian cancer.

Herein, we present a novel PAI contrast nanoagent with two well-separated peaks of comparable high intensities at 780 nm and 860 nm. This nanoagent efficiently accumulates in subcutaneous and intra-abdominal ovarian cancer tumors following systemic administration, and the distinct spectral signature allows for accurate unmixing of its photoacoustic signal from endogenous contrast in malignant tissue. In addition, the developed nanoagent can efficiently elevate intratumoral temperature up to 49 °C upon exposure to NIR light and treat the cancer tissue with photothermal therapy.

2. Results and Discussion

Development and characterization of a nanoagent for photoacoustic imaging and photothermal therapy

Contrast agents with distinct spectral signatures can offer accurate unmixing of their PA signals from endogenous photo-absorbers found in tumors, thereby improving cancer detection. We examined several naphthalocyanine (Nc) derivatives (Figure 1A and Figure S1) in order to identify building blocks for the development of a PA imaging agent with a distinct dual-peak spectral signature. Nc derivatives are promising PAI candidates due to their strong NIR light absorption, photostability, and ability to efficiently convert light energy to heat.^{[21, 22]} The identified silicon naphthalocyanine (SiNc, Figure 1A) and nickel naphthalocyanine (NiNc) exhibit distinctly narrow and well-separated absorption peaks in tetrahydrofuran (THF) at 780 nm and 860 nm, respectively (Figure S2). A polymeric nanoparticle was used to encapsulate the identified Nc derivatives in order to enhance their aqueous solubility, preserve photophysical properties, and improve delivery to cancer tissues. The employed amphiphilic PEG-PCL polymers self-assemble in aqueous solution upon evaporation of the organic solvent, forming nanoparticles with a hydrophilic PEG outer shell and hydrophobic PCL core (Figure S3) for encapsulation of naphthalocyanine molecules with negligible aqueous solubility (<1 ng mL⁻¹).^{[23, 24]} Previous reports suggest that PEG-PCL nanoparticles are safe and biocompatible, and provide efficient delivery of hydrophobic payloads to cancer tumors after intravenous injection while preventing their leaching in systemic circulation.^[25–28]

Figure 1. Imaging and photothermal properties in aqueous solution.

(A) Schematic of the naphthalocyanine-based nanoparticles (Nc-NPs) used in this study: PEG-PCL polymeric nanoparticles loaded with SiNc (Si-NP) and NiNc (Ni-NP) individually, or both dyes co-encapsulated within the same nanoparticle (SiNi-NP). Normalized NIR absorption (B) and photoacoustic (PA) spectra (C) of Si-NP, Ni-NP and SiNi-NPin saline. (D) Cryo-TEM image of SiNi-NP. (E) Overlaid ultrasound (grey) and unmixed PA (cyan) image, obtained when the overall PA signals are spectrally unmixed for the dual-peak spectral signature of SiNi-NP recorded for tube phantoms containing SiNi-NP (cyan) and saline as a control (black), within layer of chicken breast tissue (3 mm). (F) Corresponding PA spectra of tube phantoms placed within chicken breast tissue. (G) Absorption spectra of SiNi-NP indicating the application of laser illumination at 780, 808, and 830 nm. (H) Temperature recorded for SiNi-NP aqueous solutions and saline (as control) following exposure to NIR light (10 min, 1.0 W cm⁻²) at 780, 808, and 830 nm. (I) Photostability of SiNi-NP aqueous solution demonstrated by absorption spectra before and after irradiation with NIR light (15 min, 1.0 W cm⁻²).

When SiNc and NiNc dyes were individually encapsulated in PEG-PCL nanoparticles, the resulting SiNc- and NiNc-loaded nanoparticles (Si-NP and Ni-NP) maintained their absorption peaks (~780 nm and ~860 nm), with minimal red spectral shifts (Figures S1A,B). This could be explained by the presence of bulky organic substituents at or near the core of SiNc and NiNc (Figure 1A), which provide steric hindrance to these molecules, limiting their intermolecular interactions and thus minimizing aggregate formation inside nanoparticles.^{[24, 29]} In contrast, broadened absorption peaks were observed when vanadyl phthalocyanine or naphthalocyanine derivatives with peripheral substituents were encapsulated in PEG-PCL nanoparticles, which is not a desirable feature for PAI contrast agents (Figure S1C-D). Therefore, the identified SiNc and NiNc were co-encapsulated within PEG-PCL nanoparticles to prepare a PA contrast nanoagent with two well-separated narrow absorption peaks (~780 nm and ~860 nm) of comparable intensities (Figure 1B, brown solid line). Due to the fact that the extinction coefficient of NiNc (472891 M⁻¹cm⁻¹) is half that of SiNc (239935 M⁻¹cm⁻¹), while both have similar molecular weights, a two-fold greater concentration of NiNc was used (0.07 mg mL⁻¹ SiNc, 0.14 mg mL⁻¹ NiNc) to produce a nanoparticle with a ~1:1 absorption intensity dye peak ratio (780nm:860nm).

All studied nanoparticles, including dual-dye loaded SiNi-NP, and single-dye loaded Si-NP and Ni-NP, exhibited similar characteristics, including hydrodynamic size of ~ 40 nm, polydispersity index (PDI) <0.2, and near-neutral surface charge (Figure S4, Table S1). The representative cryo-transmission electron microscope (cryo-TEM) image reveals a spherical shape of SiNi-NP (Figure 1D).

The recorded PA spectrum of SiNi-NP in saline reveals two well-separated peaks of comparable intensities at 780 and 860 nm (Figure 1C, brown solid line) that are congruent with PA peaks obtained for the individual dye loaded nanoparticles (Si-NP and Ni-NP, green and red lines). To evaluate the potential of SiNi-NP for tissue imaging, capillary tubes containing SiNi-NP and saline were positioned underneath a layer of chicken breast tissue and imaged with the Vevo LAZR system. The overall PA signal recorded for chicken breast phantom was unmixed using the dual-peak spectral signature (Figure 1C) of SiNi-NP to generate a representative image (Figure 1E). The obtained image has a distinct PA signal (cyan) in the region with the SiNi-NP-loaded tube. The generated PA spectrum of this region contains two distinct peaks at ~780 and 860 nm (Figure 1F), consistent with the spectrum of SiNi-NP in solution (Figure 1C).

Previously, we reported that Nc-loaded nanoparticles could efficiently generate heat in the presence of NIR light.^[24] To assess photothermal properties of the developed SiNi-NP, their heating efficiency was evaluated upon exposure to NIR light (1.0 W cm⁻²) at three different wavelengths (780 nm for SiNc, 830 nm for NiNc, and 808 nm capable of exciting both dyes, Figure 1G). SiNi-NP were capable of elevating the temperature of an aqueous solution within 5 min up to 51 °C, 55 °C and 52 °C at 780 nm, 808 nm and 830 nm, respectively (Figure 1H). This property enhances the translational potential of the developed nanoagent because it can be activated by various clinically available NIR laser systems, emitting in the broad NIR spectral region. Importantly, the absorption spectrum of SiNi-NP was not affected by NIR light exposure (1 W cm⁻²), indicating high photostability of SiNi-NP (Figure 1I). Along with high heating efficiency, photostability is another important parameter for photoacoustic and photothermal agents, because photobleaching compromises their ability to efficiently convert light energy into heat, diminishing their imaging and therapeutic potential.

In-Vitro evaluation of PA imaging and photothermal properties of SiNi-NP

To evaluate the PA imaging properties of SiNi-NP with a distinct dual-peak PA signature in-vitro, two ovarian cancer cell lines were used: A2780N and ES2. Cell suspensions incubated for 24 h with Si-NP, Ni-NP and SiNi-NP were imaged with the Vevo LAZR system after being embedded within an agarose mold to reduce air-associated acoustic reflection (Figure 2A and B). Ultrasound images (Figure 2A) show the difference in density of the cell suspensions (white color) as compared to the surrounding gel (dark grey). PA imaging of cell-containing agarose molds at various wavelengths within the NIR region (650–970 nm) revealed pronounced overall signal at 780 and 860 nm, which corresponds to SiNc and NiNc photoacoustic signal, respectively (Figure 2A, rainbow). In contrast, only the background level PA signal was detected at 680 nm, wavelength distant from main SiNc and NiNc photoacoustic peaks, demonstrating only background spectra often associated with trapped air bubbles without any distinguished peaks. The generated PA spectra of agarose molds containing nanoparticle-treated cells are in agreement with spectra of Si-NP, Ni-NP, and SiNi-NP recorded in solution (Figure 1C).

Figure 2. In Vitro Evaluation of Photoacoustic imaging and PTT properties of Nc-NPs.

(A) Overlaid ultrasound (grey) and PA (rainbow) images, and corresponding PA spectra of A2780N cells treated with Si-NP, Ni-NP, SiNi-NPs for 24 h, acquired at 680, 780, and 860 nm. (B) Schematic illustration of PA imaging of cells treated with Nc-NPs, embedded in agarose gel. (C) Ultrasound images overlaid with PA images of ES2 and A2780N cells treated with SiNi-NP, unmixed individually for Si-NP (green) and Ni-NP (yellow), Si-NP+Ni-NP (green and yellow overlay), and for the unique dual-peak spectral signature of SiNi-NP at 10 μg mL⁻¹ of SiNc dye (cyan). Note: PA spectra confirm the presence of dual-dye signature within ROI. (D) Photothermal therapeutic (PTT) effect on ES2 cells incubated with SiNi-NP (10 μg mL⁻¹ of SiNc) for 24 h and exposed to NIR laser light for 10 min (1.0 W cm⁻², 808 nm) versus three controls: cells only (Cells), cells exposed to NIR light only (Light), and cells treated with SiNi-NP only (NP). Cell viability values are expressed as mean ±SD (n=3). ****P<0.0001 vs. untreated cells (control).

Next, we confirmed imaging properties of SiNi-NP in ES2 and A2780N ovarian cancer cells using PAI following spectral unmixing for SiNc (green), NiNc (yellow), individually for SiNc and NiNc (green & yellow), and dual-peak signature of SiNi-NP (cyan) (Figure 2C). Strong signals for both SiNc and NiNc dyes, distributed throughout the sample, can be detected when unmixed for Si-NP (green) and Ni-NP (yellow) separately. It appears that unmixing for individual dye signals (Si-NP+Ni-NP) gives the false impression that NiNc (yellow) is faint and overpowered by SiNc (green). However, unmixing for the dual-peak signature (cyan) demonstrates a more accurate representation with a clear signal equally distributed throughout the cell suspensions, as is also confirmed by the dual-peak PA spectra for both ES2 and A2780N cells (Figure 2C, spectra). Hence, in vitro studies in ovarian cancer cells treated with SiNi-NP confirm that its specific PA signature is preserved after cellular internalization and this dual-peak unmixing approach offers a clear advantage over individual dye unmixing for accurate target detection.

Finally, we evaluated whether the dual-dye co-encapsulated SiNi-NP possesses photothermal properties sufficient to generate enough heat to destroy cancer cells. After incubating ES2 cells with SiNi-NP for 24 h, loosely pelleted cells were irradiated with 808 nm NIR light, capable of simultaneously exciting both SiNc and NiNc, for 10 min. During NIR light exposure, the internal temperature within PTT-treated cell pellets increased from ~37 to ~46 °C resulting in approximately 92% cell death (Figure 2D, PTT). This increase in temperature was achieved during the initial 2 minutes of NIR exposure and sustained for the duration of treatment (10 min). In contrast, treatment with either SiNi-NP alone or NIR light alone, resulted in a negligible increase in temperature (<0.5°C) without compromising cell viability (Figure 2D, Light & NP). These results confirmed that SiNi-NP, upon exposure to NIR light, rapidly increases intracellular temperature and causes cell death. Notably, the individual components of this applied photothermal therapeutic modality (SiNi-NP and NIR light) are non-toxic when administered separately, and only produce the desired photothermal effect when combined (PTT). This phototherapeutic effect was also qualitatively confirmed, as the Calcein AM dye used to identify live cells is clearly visible via fluorescence microscopy (Figure S5). These data confirm that, in addition to its photoacoustic imaging properties, SiNi-NP combined with NIR light is capable of generating intracellular heat sufficient to exert a phototherapeutic effect.

In-Vivo SiNi-NP-mediated photoacoustic imaging of subcutaneous and intraperitoneal tumors

Obtained in-vitro data justified further evaluation of the PAI properties of SiNi-NP in athymic nude mouse models bearing either subcutaneous or intraperitoneal ovarian cancer tumors. A subset of overall PA signal images of the subcutaneous ES2 tumor following intravenous (IV) administration of SiNi-NP, taken over the 680–970 nm NIR spectra (Figure 3A, rainbow), shows a significantly higher signal at the wavelengths 780 and 860 nm that corresponds to the signal peaks of SiNc (780 nm) and NiNc dyes (860 nm) (Figure 3A). As demonstrated in Figure 3B, the PA image of the SiNi-NP-treated tumor revealed a distinct PAI signal not only visually, but the presence of dual-peak signature associated with SiNi-NP was confirmed with the corresponding PA spectrum (green) as compared to the untreated tumor where only the PA spectrum of blood (red) was detected. Next, we evaluated the capability of SiNi-NP to generate accurate unmixed PA images. The different regions of SiNi-NP-treated tumor were probed as indicated by color-coded ROIs (Figure 3C), and the corresponding PA spectra were generated (Figure 3D). Thus, predominantly the spectral signature of SiNi-NP was detected within ROI 1 (green), whereas ROI 2 (red) and ROI 3 (purple) showed the presence of a mixture of SiNi-NP and blood signals. Notably, the SiNi-NP and blood spectral signals were distinguishable, providing efficient identification of SiNi-NP within the background. Subsequently, the unmixed images of the same tumor (recorded 48 h post nanoagent IV injection, from Figure 3C) were generated where images were produced after unmixing for the dual-peak signature of SiNi-NP (cyan), or individually for SiNc and NiNc dyes (green+yellow). Unmixing for the dual-peak signature of SiNi-NP demonstrated a distribution of the nanoagent in the tumor (Figure 3E, cyan) that agrees with the spectra indicating the presence of SiNi-NP within corresponding ROIs (Figure 3C). In a scenario where one or the other of the two peaks is obscured by endogenous photoabsorbers such as hemoglobin, melanin, or other macromolecules, the remaining visible peak would still confirm the presence of SiNi-NP. However, when images were produced by unmixing individually for SiNc and NiNc, it resulted in misallocation of the dyes by preferentially assigning dye-related colors to one dye over the other (Figure 3D, SiNc and NiNc, green+yellow), which does not correspond with the detected spectra. According to the results of the unmixing experiment, utilizing a dual-peak signature allows for the generation of more accurate unmixed PA images and hence more accurately portrays the biodistribution of the nanoparticles throughout the tumor.

Figure 3. SiNi-NP-mediated PA imaging of subcutaneous mouse tumor.

(A) Overall PA signal at specific wavelengths (680, 780, 860, and 970 nm) superimposed on ultrasound images of athymic nu/nu mice with subcutaneous ES2 ovarian cancer tumor xenograft, 48 h after IV injection with SiNi-NP (0.5 mg kg⁻¹ of SiNc). (B) PA spectra generated from PA images of SiNi-NP-treated tumor and compared to the untreated control tumor, confirming the presence of double-peak signature (green) vs. blood (red). (C) Regions of interest (ROI) established as noted on the PA image of SiNi-NP-treated tumor, and (D) the resulting corresponding PA spectra taken and normalized from the ROIs, as indicated by corresponding colors. (E) PA images of ES2 subcutaneous tumor (from C) unmixed for dual-peak signature SiNi-NP (cyan) and individually for SiNc and NiNc (green+yellow), demonstrating the difference in unmixing accuracy.

Next, we assessed whether the developed SiNi-NP could be used as a PAI agent to locate intraperitoneal (IP) ovarian tumors in the mouse abdominal cavity. Thus, the mice were IP inoculated with luciferase-expressing ES-2 (ES2/Luc) ovarian cancer cells, and the development of ES2/Luc ovarian IP tumors was confirmed with bioluminescence signal (Figure 4A, ROI). Encouragingly, SiNi-NP was detected by unmixing the overall PA signal for the dual-peak spectral signature (Figure 4B, cyan) within the bioluminescent region of the abdomen (Figure 4A, ROI) 48 h following IV-injection of SiNi-NP. Furthermore, performing PAI of the same region in 3D mode, following accumulation of SiNi-NP within the abnormal growth, allowed visualization of the suspected tumor (cyan, Video S1) by differentiating it from surrounding tissue, e.g., oxygenated (Oxy-Hb, red) and deoxygenated (Deoxy-Hb, blue) hemoglobin (Figure 4B). In addition, as demonstrated in Figure 4C, the presence of dual-peak signature associated with SiNi-NP was confirmed with the corresponding PA spectrum (green) as compared to abdominal tissue background (red) prior to treatment. Further confirming the presence of SiNi-NP within the tissue identified by PA imaging, the mass was resected, and fluorescence microscopy of thin sections revealed SiNc signal at 775 nm (Figure S6). Supporting the assumption that the identified tissue was cancerous, hematoxylin and eosin (H&E) staining of thin sections depicted features consistent with ovarian clear cell carcinoma: high degree of cellularity, large nuclei, disorganized tissue structure, glandular spaces, and large hyaline globules (Figure 4D,E). This proof-of-concept study demonstrates that PA imaging, employing the unique dual-peak PA spectral signature of SiNi-NP, is sufficient for real-time non-invasive detection of lesions as small as ~1mm, which could easily be overlooked by currently available imaging technologies in clinic. Such a PAI contrast agent that is clearly distinguishable from endogenous background signal has the potential to be an effective tool for accurate target identification, especially crucial for early detection of ovarian cancer.

Figure 4. SiNi-NP-mediated PA imaging of intraperitoneal mouse tumor.

(A) Whole-body bioluminescence image of athymic nude mouse with IP xenograft of luciferase(Luc)-labeled ES2 ovarian tumor, 48 h after IV injection with SiNi-NP (0.5 mg kg⁻¹ SiNc). (B) PA image superimposed on ultrasound image of mouse abdomen bearing ES2-Luc IP tumor, unmixed for SiNi-NP (cyan), Oxy-Hb (red), and Deoxy-Hb (blue). (C) PA spectra generated from PA of SiNi-NP-treated IP tumor (from Figure 4B) and compared to the control prior to treatment, confirming the presence of double-peak signature (green) of SiNi-NP vs. control tissue background (red). (D) H&E stained section of the suspected tumor tissue collected 48 h post-IV-injection of SiNi-NP (0.5 mg kg⁻¹ SiNc). (E) Magnified H&E stained section from Figure 4D, where arrow indicates hyaline globule and asterisk indicates gland.

In-Vivo image-guided photothermal (PTT) therapy

Finally, as a proof-of-concept, we demonstrated that SiNi-NP could be used to photoacoustically guide the localized application of photothermal therapy (PTT, tissue eradication using heat) while simultaneously acting as the PTT agent. For this evaluation (Figure 5A), mice with subcutaneous ES2 tumors were IV injected with SiNi-NP (0.5 mg kg⁻¹ SiNc), and after 48 h accumulation of nanoparticles within tumors was confirmed by the presence of the PA signal of SiNi-NP (Figure 5B), and by the dual-peak PA spectral signature (Figure 5C). Thus, nanoagent uptake by the tumor, confirmed by photoacoustic imaging within the tissue of interest, enables controlled, localized application of PTT. After confirming the accumulation of SiNi-NP, PTT was performed by illuminating the tumor with NIR light (808 nm, 1.0 W cm⁻², 15 min). The PTT treatment generated temperatures of ~49°C as demonstrated by thermal imaging (FLIR) at the tumor surface that were sustained for the duration of exposure (Figure 5D). Tumors of treated and control mice were monitored for 10 days following PTT treatment and demonstrated complete tumor eradication after 4 days post-PTT, whereas continuous tumor growth was observed in control mice (Figure 5E). These results demonstrate the theranostic capabilities of SiNi-NP to act as both imaging and phototherapeutic agent.

Figure 5. Image-guided photothermal therapy in-vivo.

(A) Schematic timeline of PTT experiment performed in mice: 10 days following subcutaneous tumor inoculation, SiNi-NP was intravenously injected and PTT was performed 48 hours later, followed by tumor growth monitoring for 10 days (n=3). (B) PA image of tumor demonstrating PA signal produced by SiNi-NP, 48 h post-injection of SiNi-NP (0.5 mg kg⁻¹ SiNc), and (C) corresponding spectrum depicting dual-peak PA signature of SiNi-NP within the tumor. (D) Thermal image of the mouse, bearing the subcutaneous ES2 tumor, injected with SiNi-NP (0.5 mg kg⁻¹ SiNc), captured with FLIR infrared camera at 15 min of PTT treatment (post 48 h IV injection, 808 nm laser light, 1.0 W cm⁻²). (E) Growth profiles of subcutaneous ES2 tumors along with untreated controls after the mice were injected with SiNi-NP (0.5 mg kg⁻¹ SiNc) and exposed to 808 nm laser light (1.0 W cm⁻²) for 15 min, at 48 h post-injection.

3. Experimental Section

Synthesis and characterization of Nc-based nanoparticles (Nc-NPs)

Nanoparticles were produced by encapsulation of silicon 2,3-naphthalocyanine bis(trihexylsilyloxide) (SiNc) or nickel(II) 5,9,14,18,23,27,32,36-octabutoxy-2,3-naphthalocyanine (NiNc) dyes (Sigma-Aldrich, Milwaukee, WI) individually, or co-encapsulation of the two dyes (SiNc and NiNc), within poly(ethylene glycol)-block-poly(ε-caprolactone) methyl ether, PEG(5k)-PCL(10k) co-polymer. The solvent evaporation method was used,^{[23, 24]} where PEG-PCL (Advanced Polymer Materials Inc., Montreal, Canada) and corresponding dye(s) were solubilized in tetrahydrofuran (THF). Saline was added under constant stirring conditions, and THF was removed via rotovap (Heidolph North America, IL). Any unincorporated hydrophobic dyes were removed via microcentrifugation at 10,000 rpm for 3 minutes, and the resulting aqueous nanoparticle solutions were filtered (0.2 μm pore size membrane). The size and surface charge of the produced nanoparticles were characterized using a Zetasizer Nano ZEN3600 (Malvern, UK). The morphology and size of the nanoclusters were accessed by cryogenic transmission electron microscopy (cryo-TEM) on a Tecnai Spirit TEM (Tecnai, Hillsboro, OR). Absorbance spectra of the Nc-NP solutions were recorded using a Shimadzu spectrophotometer (UV-1800, Carlsbad, CA). The actual dye-loading concentrations were determined by absorption of the dye at ~780 (SiNc) and ~850 nm (NiNc) and calculated using an extinction coefficient of 472891 and 239935 M⁻¹cm⁻¹, respectively, in THF.

Photoacoustic imaging in solution and tissue

For PAI studies, the co-encapsulated SiNi-NP (~0.07 mg mL⁻¹ SiNc, ~0.14 mg mL⁻¹ NiNc) and the mixture of Si-NP and Ni-NP were formulated to produce samples with a 1:1 PA intensity peak ratio of SiNc:NiNc. Photoacoustic properties of the Nc-NPs were assessed by injecting their solutions into polyethylene tubes (inner diameter: 0.38 mm, BD Intramedic™ PE tubing, Franklin Lakes, NJ) and sealing the ends. The tubes containing Nc-NP were imaged using the Vevo LAZR imaging system (FUJIFILM VisualSonics, Toronto, Canada) with the LZ550 probe (axial resolution 40 μm, transducer center operating frequency of 40 MHz). In addition, tubes with SiNc-NP vs. saline as control were imaged within chicken breast tissue (purchased at local grocery store). The “Spectro“ sub-mode was used to record overall PA signal across the 680–970 nm operating region. The photoacoustic spectra were generated from PA signal within the region of interest (ROI) of the acquired PA image. “Nanostepper“ PA sub-mode was then employed to obtain 2D information and generate spectrally unmixed images.

Heat generation under NIR light exposure in solution

The SiNc-NP aqueous solution (200 μL, ~0.07 mg mL⁻¹ SiNc, ~0.14 mg mL⁻¹ NiNc) was exposed for 10 min to the NIR laser at 785, 808, and 830 nm (1.0 W cm⁻², continuous-wave (CW) diode laser, Wavespectrum Laser Group Limited, China). To evaluate heat production under NIR exposure, a fiber optic thermal probe (Neoptix, Québec, Canada) was used to record temperature in SiNc-NP solution and compared to saline as a control (samples were maintained at 37°C prior to NIR laser exposure). To assess photostability, absorption spectra of SiNc-NP solution were recorded prior to and after 15 min NIR light treatment.

Photoacoustic imaging of Nc-NPs in cell pellets

The ovarian cancer cell lines ES2 and A2780N, cultured in RPMI 1640 media with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (pen/strep), were seeded in 6-well plates at 0.3 ×10⁶ cells per well and allowed to grow for 24 h, then treated for 24 h with the studied formulations: Si-NP, Ni-NP, and SiNi-NP (10 μg mL⁻¹ SiNc). The cells were then washed with DPBS, trypsinized, and pelleted via centrifugation. The pellet was suspended in 100 μL of DPBS and mixed at a 1:1 volumetric ratio with 2% warm agarose solution, then inserted into a cavity cast within a 2% agarose mold and left to sit for 30 min. Background noise often appears in the images due to air bubbles caught in the gel during the casting phase and/or air bubbles trapped in ultrasound gel. The agarose mold containing the cell suspensions was inverted to minimize background noise associated with imperfections in the mold surface (Figure 2B), and PAI was performed with the LZ550 probe. The Spectro and, subsequently, Nanostepper PA sub-modes were then employed to obtain 2D information and generate spectrally unmixed images. Based on the employed unmixing algorithms, the system automatically selects the specific PA spectra previously established for common endogenous agents (Deoxy-Hb, Oxy-Hb, etc.) or newly acquired PA spectra of developed exogenous agents to produce unmixed PA images.

SiNi-NP-based photothermal therapy (PTT) in vitro

ES2 ovarian cancer cells were cultured in RPMI (10% FBS, 1%pen/strep) and incubated overnight with SiNi-NP (10 μg mL⁻¹ SiNc) in T175 culture flasks (3.2×10⁶ cells/well). Twenty four hours later, cells were washed, trypsinized and loosely pelleted. The same number of cells was used for all treatment groups: cells incubated with SiNi-NP and exposed to NIR laser (10 min, 808 nm, 1.0 W cm⁻²) and three controls (non-treated cells, cells exposed to 808 nm light only, and cells incubated with SiNi-NP only). Cells were then seeded at 10,000/well in a 96-well plate, incubated for 24 h, and a Calcein AM assay was performed to determine cell viability in each treatment group. The temperature within cell pellets was recorded using a fiber optic thermal probe as previously described.^[24]

In-vivo PA imaging of subcutaneous and intraperitoneal tumors in mice

Athymic nude mice were injected subcutaneously in the flank with 3×10⁶ cells (ES2) mixed at a 1:1 ratio with Matrigel (Corning, USA). To grow intraperitoneal (IP) ovarian cancer tumors, ES2-Luc (luciferase-expressing) cells cultured in RPMI media (10% FBS, 1% pen/strep) were seeded at 200 cells/well in a spheroid 96-well microplate (Corning, USA) for approximately 48 h, forming spheroids of ~150 μm. Media was removed, and the spheroids were dislodged with saline and injected into the mouse abdominal cavity using a 21G needle. The growth of IP tumors was monitored via bioluminescence imaging using the IVIS Lumina XRMS In Vivo Imaging System (PerkinElmer, USA), 10 min after IP injection of D-luciferin (200 μL, 15 mg mL⁻¹). Only cancer cells expressing luciferase (Luc-labeled) generate a bioluminescence signal following injection of D-luciferin substrate. After ~2 weeks, mice with subcutaneous tumors (~150–250 mm³), or intraperitoneal (IP) tumors as confirmed with bioluminescence imaging, were injected intravenously with the studied nanoformulations and imaged via the Pearl Impulse Small Animal Imaging System (LI-COR, USA) to monitor nanoparticle distribution via SiNc fluorescence at 800 nm.

48 h following IV delivery of nanoparticles, PAI was performed using Vevo LAZR with a LZ550 transducer. During PA imaging, mice were anesthetized using inhaled 2% isoflurane administered in 100% oxygen (1–3 L min⁻¹) on a 37°C heated platform which allowed for monitoring of respiration, heart rate, and body temperature of the mouse. Pre-warmed ultrasound gel was applied to the skin to provide an interface between the transducer and the mouse skin. In Spectro mode, overall PA signal was recorded across the 680–970 nm region, enabling PA spectra generation from corresponding ROIs. Nanostepper mode was then employed to obtain 2D information and generate a spectrally unmixed PA image which can be overlaid with the corresponding ultrasound image. In addition, 3D scan was performed to generate 3D video of the ultrasound and PA image overlay, or a series of individual frames of unmixed 2D images. The corresponding PA images were recorded prior to, and 48 h after, injection of studied nanoformulations.

Fluorescence imaging and histology of IP tumor

Suspected tumor tissue was collected at necropsy and incubated first in 10% formalin, followed by 70% ethanol; each step for 24 h at 4°C with shaking. Tissue was placed in paraffin embedding cassettes and dehydrated in two changes of 95% ethanol, 100% ethanol, and 100% xylene for 30 min each. Dehydrated tissue was then incubated overnight in paraffin wax at 60°C and embedded in a paraffin block the following day. The paraffin block was sectioned at 10 μm using a Leica RM2255 microtome (Leica, Germany). Sections were transferred to positively charged histology slides, air-dried for 30 min, and then baked overnight at 50 °C. Thin sections were deparaffinized in two changes of Xylene. At this point, thin tumor sections were imaged using a BZ-X710 fluorescence microscope (Keyence, Japan) at ex./em. 710/810 nm (SiNc). Tissue morphology was then evaluated using hematoxylin and eosin (H&E) stain. Sections were rehydrated using two changes of 100% ethanol for 5 min each, 95% ethanol for 5 min, 70% ethanol for 5 min, then PBS for 5 min. Slides were then immersed in Gill’s hematoxylin for 1min to stain cell nuclei and rinsed with tap water until clear, followed by 10X dip in 0.02% glacial acetic acid, 10X dip in tap water, 2 min in bluing solution, 10X dip in tap water, and 5 min in 70% ethanol. Incubation in eosin solution for 1min was used to stain cell cytoplasm, followed by dehydration with 5X dip in 95% ethanol, 5 min in 100% ethanol, and 2 changes of 100% xylene for 5 min each. H&E stained slides were cover slipped and imaged using the brightfield setting on the BZ-X710 Fluorescence Microscope. Individual fluorescence and brightfield images were merged using Keyence BZ-X Analyzer software to create whole images of thin tumor sections.

Image-guided photothermal therapy (PTT) in vivo

Athymic nude mice bearing subcutaneous ES2 tumors of ~150 mm³ (n=3 per group) were injected intravenously with either 1) SiNi-NP (0.5 mg kg⁻¹ SiNc) or saline (control group). 48 h following IV injection of SiNi-NP, after the presence of nanoparticles was confirmed via fluorescence and PA imaging, mice were anesthetized and the tumors were exposed to 808 nm laser (1.0 W cm⁻², 15 min). To evaluate heat production in the tumor during PTT, an infrared camera (FLIR Systems, Inc., Wilsonville, OR) was used to record temperature each min for the duration of treatment (15 min). Length and width of tumors were measured using calipers for 10 days following treatment and used to calculate tumor volume as V=W(2)×L/2, where V, W, and L are volume, width, and length of tumor, respectively.

Statistical Analysis

Data are presented as mean values ± standard deviation (mean ± SD) from three separate measurements. The sample size for each statistical analysis (n) is indicated in the figure legends. GraphPad Prism v9 (GraphPad Software, CA, USA) was used for statistical analyses. One-way analysis of variance (ANOVA) was used to test statistical significance for more than two groups. Differences between groups were considered statistically significant at p< 0.05. Statistical significance is denoted as ***P<0.0001 vs. untreated control)

4. Conclusion

In this study, we demonstrate an approach that is focused on the development of a nanoformulation co-encapsulating two naphthalocyanine dyes with distinctive PA features, producing a unique dual-peak PA spectral signature that is easily identifiable and distinguishable from the endogenous background. By evaluating the unique PA signature of the constructed dual-dye nanoagent in both subcutaneous and intraperitoneal mouse models of ovarian cancer, we demonstrated a more accurate in vivo PA signal identification than when using either dye individually, as verified by corresponding PAI spectra. Notably, the developed PAI nanoagent enabled location and delineation of ovarian cancer lesions as small as ~1 mm in the mouse abdomen. As a proof-of-concept, the PA image-guided photothermal therapy mediated by the developed nanoagent was performed in mice bearing subcutaneous ovarian cancer tumors, resulting in intratumoral heat production of 42–49°C and subsequent complete tumor eradication. These reported findings demonstrate the potential to enhance the performance of PAI by utilizing exogenous contrast agent with specifically distinct PA spectral signatures, with the goal of improving cancer diagnostics and guiding the application of photo-hyperthermia.

Abbreviations

NIR

near infrared

PAI

photoacoustic imaging

SiNc

silicon naphthalocyanine

NiNc

nickel naphthalocyanine

NP

nanoparticles

PEG-PCL

(poly (ethylene glycol)-b-poly(ɛ-caprolactone)

FDA

Food & Drug Administration

THF

tetrahydrofuran

DLS

dynamic light scattering

cryoTEM

cryogenic transmission electron microscopy

IACUC

the Institutional Animal Care and Use Committee

ROI

regions of interest

IV

intravenous

Data Availability Statement

The data that support the findings of this study are available from the cor-responding author upon reasonable request