Contrast-enhanced ultrasound imaging for monitoring the efficacy of near-infrared photoimmunotherapy

Summary

Background

Near-infrared photoimmunotherapy (NIR-PIT) is a promising cancer therapy combining NIR-light irradiation with an antibody and IR700DX, a light-sensitive substance, to destroy tumours. However, homogeneous irradiation is difficult because the light varies depending on the distance and tissue environment. Therefore, markers that indicate sufficient irradiation are necessary. Nanoparticles sized 10∼200 nm show enhanced permeation and retention within tumours, which is further enhanced via NIR-PIT (super enhanced permeability and retention, SUPR). We aimed to monitor the effectiveness of NIR-PIT by measuring SUPR.

Methods

A xenograft mouse tumour model was established by inoculating human cancer cells in both buttocks of Balb/C-nu/nu mice, and NIR-PIT was performed on only one side. To evaluate SUPR, fluorescent signal examination was performed using QD800-fluorescent nanoparticles and NIR-fluorescent poly (d,l-lactide-co-glycolic acid) (NIR-PLGA) microparticles. Harmonic signals were evaluated using micro-bubbles of the contrast agent Sonazoid and contrast-enhanced ultrasound (CEUS) imaging. The correlation between SUPR immediately after treatment and NIR-PIT effectiveness on the day after treatment was evaluated.

Findings

QD800 fluorescent signals persisted only in the treated tumours, and the intensity of remaining signals showed high positive correlation with the therapeutic effect. NIR-PLGA fluorescent signals and Sonazoid-derived harmonic signals remained for a longer time in the treated tumours than in the controls, and the k_E value of the two-compartment model correlated with NIR-PIT effectiveness.

Interpretation

SUPR measurement using Sonazoid and CEUS imaging could be easily adapted for clinical use as a therapeutic image-based biomarker for monitoring and confirming of NIR-PIT efficacy.

Funding

This research was supported by ARIM JAPAN of MEXT, the Program for Developing Next-generation Researchers (Japan Science and Technology Agency), KAKEN (18K15923, 21K07217) (JSPS), CREST (JPMJCR19H2, JST), and FOREST-Souhatsu (JST). Mochida Memorial Foundation for Medical and Pharmaceutical Research; Takeda Science Foundation; The Japan Health Foundation; and Princess Takamatsu Cancer Research Fund.

Funders only provided financial support and had no role in the study design, data collection, data analysis, interpretation, and writing of the report.

Research in context.

Evidence before this study

Near-infrared photoimmunotherapy (NIR-PIT) is a cancer therapy that combines NIR-light irradiation with an antibody and IR700DX, a light-absorber substance, to destroy target cancer cells. It is thought to be a promising fifth cancer treatment following the four major conventional treatments: surgery, chemotherapy, radiotherapy, and immunotherapy. However, homogeneous irradiation is difficult to achieve because the light varies and attenuates according to the distance to target and tumour environment. To confirm sufficient NIR-light irradiation in NIR-PIT-treated patients, methods that can be easily adapted for clinical use are highly needed.

Added value of this study

We demonstrated that NIR-PIT-treated tumour showed super enhanced permeability and retention (SUPR) with micro-sized particles. We also proved that the assessment of SUPR via contrast-enhanced ultrasound (CEUS) imaging could monitor and confirm NIR-PIT therapeutic effect.

Micro-bubble (Sonazoid)-derived harmonic signals remained longer in the treated tumours, and the calculated k_E value from the two-compartment model correlated with the therapeutic effect.

Implications of all the available evidence

This study provides a discovery that SUPR enhances the intra-tumour persistence of microparticles with a diameter of up to 5 μm, approximately, which we named as the micro-sized SUPR effect. Moreover, an image-based marker for monitoring and confirming the effectiveness of NIR-PIT using micro-bubble and CEUS could be realised and the method could be easily adapted for clinical use.

Introduction

Near-infrared photoimmunotherapy (NIR-PIT) is a recently developed multidisciplinary cancer therapy that uses antibodies and light-sensitive substances, such as IR700, and NIR-light; it has attracted much attention owing to its high specificity via two-step targeting. Cell death associated with NIR-PIT is unique. IR700 has a silica-phthalocyanine structure. NIR-light of sufficient intensity liberates the silanol group in the structure under the conditions of electron donor abundance. Subsequently, the structure rapidly becomes hydrophobic and aggregates, causing damage in the target cell membrane.¹ This technique has received remarkable clinical approvement for previously treated head and neck cancers in Japan, and clinical trials are underway worldwide.²

Overall, NIR-PIT is thought to be a promising new modality for cancer treatment, and various potential applications have been studied.3, 4, 5, 6, 7, 8 However, a therapeutic biomarker to confirm and monitor sufficient NIR-light irradiation during NIR-PIT in the operating room has not yet been established.⁹ Since NIR-PIT uses NIR-light to trigger an anti-tumour effect via a photo-chemical reaction, optimizing the NIR-light irradiation is essential. However, NIR-light is subject to distance-dependent attenuation, and the living tissue scatters light, resulting in undesired light-dose attenuation.¹⁰ In addition, obstructive barriers such as bones or vessels may block the light from reaching the intended area. Therefore, therapeutic biomarkers, which allow us to confirm whether the desired NIR-PIT therapeutic effect has been achieved, are highly desirable. If biomarkers indicate that the amount of NIR-light irradiation is not sufficient at the NIR-PIT, additional NIR-light irradiation could be added in a flexible manner. This might increase the effectiveness of NIR-PIT.

When nanoparticles are administered intravenously, nanosized drugs have long circulation times and are typically not excreted by the urinary tract but tend to be taken up in the reticuloendothelial system.¹¹ Thus, nano-sized drugs with long circulation times preferentially leak from the abundant and permeable tumour vessels into the tumour tissue and remain in the tumour due to limited drainage by the underdeveloped lymphatic vessels. This is called the enhanced permeability and retention (EPR) effect.¹² The EPR effect has been used in delivery systems to maintain intra-tumour concentrations of anticancer drugs, and some are clinically used, including antibody and drug conjugate,¹¹ micelle drugs,¹³ and liposome drugs,¹⁴ among others. In tumours treated with NIR-PIT, the EPR effect of nanoparticles is approximately 24-fold greater owing to the increased permeability caused by rapid cell death of perivascular tumours. This enhanced EPR effect with NIR-PIT is known as the super enhanced permeability and retention (SUPR) effect.¹²

Ultrasonography is widely used for diagnosis, therapeutic guidance, and follow-up given its continuous imaging capability, non-invasiveness, low cost, portability, easy-accessibility (no protection required) compared to other imaging modalities.¹⁵ In addition, contrast-enhanced ultrasound (CEUS) using ultrasound contrast agents significantly improved the detection of liver and breast tumours.^16,17 The scatter signals from micro-bubble contrast agents have both the same frequency as the transmitted pulse and multiples of the transmitted frequency (harmonics) due to non-linear acoustic effects. By subtracting the outgoing signal from the received signal, the harmonics signal can be selectively extracted. CEUS is a method of drawing an ultrasonographic image from these signals.^18,19 It is used to identify disease lesions such as hepatocellular carcinoma, breast cancer, or renal malignancies and to provide guidance for percutaneous biopsy because it can noninvasively delineate differences in properties within parenchymal organs.¹⁹

Currently, rapid therapeutic biomarkers of NIR-PIT potentially presented for clinical use can detect decreased IR700-fluorescence, analyse released ligands of IR700DX (silanols) in urine,¹ and measure the SUPR effect via the accumulation of contrast agent. Methods for quantifying the SUPR effect have been reported using ICG-fluorescence and MRI contrast agents.^20,21 However, fluorescence measurement does not provide comprehensive information, and MRI measurement is difficult to access (patients have to move to the MRI room) and requires a long imaging time. Therefore, easier methods to serve as therapeutic biomarkers for NIR-PIT are highly required.

In this study, we establish a new concept, which we labelled as ‘micro-sized SUPR’, and propose a new technological method to evaluate SUPR via CEUS imaging, allowing the monitoring and confirming NIR-PIT therapeutic efficacy.

Methods

Study design

Our primary objective was to evaluate the SUPR effect associated with NIR-PIT to establish a monitoring imaging biomarker for in vivo NIR-PIT effect via ultrasound, which has already been equipped in the clinic. We demonstrated a new prediction method for NIR-PIT using a series of controlled and approved laboratory experiments.

Animals were assigned to each experimental group to achieve the highest possible consistency in tumour luciferase activity among all groups. The number of mice is indicated in the figure legends; otherwise, each group comprised at least three mice.

Ethics statement

All animal experiments were conducted in compliance with the Guide for the Care and Use of Laboratory Animal Resources of the Nagoya University Animal Care and Use Committee (Nagoya, Japan; approval numbers 2021: M210770-001 and 2022: M220370-003).

Animals were housed in accordance with the Act on Welfare and Management of Animals, Standards relating to the Care and Keeping and Reducing Pain of Laboratory Animals, and Fundamental Guidelines for Proper Conduct of Animal Experiment and Related Activities in Academic Research Institutions.

Reagents

IRDye 700DX-NHS ester was purchased from LI-COR Biosciences (Lincoln, NE, USA) (cat # 929–70,010). Panitumumab (Vectibix) (KEGG Drug; D05350), a fully humanised IgG2 mAb directed against EGFR, was purchased from Amgen (Thousand Oaks, California, USA). Qdot 800 ITK™ Amino (PEG) Quantum Dots were purchased from Thermo Fisher Scientific (Waltham, MA, USA). Near-IR Fluorescent PLGA nanoparticles were purchased from Sigma-Aldrich (St. Louis, MO, USA).

Cell culture

All cell lines were obtained from the American Type Culture Collection (Manassas, VA, USA) (A431 (RRID: CVCL_0037), MDAMB468 (RRID: CVCL_0419), NIH 3T3 (RRID: CVCL_0594, PC9 (RRID: CVCL_B260)). All cell lines were authenticated by Short Tandem Repeat (STR) tests. Recent mycoplasma testing has been performed. A431-luc-GFP cells (human epidermoid cancer cells) with genes encoding firefly luciferase and green fluorescence protein (GFP),^22,23 MDAMB468-luc-GFP cells (human breast cancer cells) with genes encoding firefly luciferase and GFP,^24,25 and 3T3-luc (murine fibroblast, NIH-3T3), PC9-luc cells (human lung cancer cells) (RRID: CVCL_B260)with genes encoding firefly luciferase were cultured in RPMI-1640 (Thermo Fisher Scientific)²⁶ supplemented with 10% fetal bovine serum and penicillin (100 IU/mL)–streptomycin (100 mg/mL) (Thermo Fisher Scientific).

Conjugation of IR700 to panitumumab

Panitumumab (Vectibix) (KEGG Drug; D05350) (6.8 nmol) was incubated with IR700 NHS-ester (30.8 nmol, LI-COR Biosciences) in 0.1 mol/L Na₂HPO₄ (pH 8.6) at 25 °C for 1 h.²⁷ The mixture was separated and purified using a Sephadex G50 column (PD-10; GE Healthcare, Piscataway, NJ, USA; cat # 17,085,101).⁴ Protein concentration was confirmed using a Coomassie Plus Protein Assay Kit (Thermo Fisher Scientific) by measuring absorption at 595 nm via spectroscopy (UV-1900; Shimadzu, Kyoto, Japan).^28,29 IR700 concentration was measured via absorption at 689 nm using the spectrophotometer UV-1900 to confirm the number of fluorophore molecules conjugated to the mAb (dye-mAb ratio),29, 30, 31 Protein concentration was further quantified via SDS-PAGE.³² The dye to mab ratio is 2.5. Panitumumab-IR700 is referred to as pan-IR700 thereafter. Bioactivity of the conjugated products was determined by testing its binding on A431-luc-GFP, MDAMB468-luc-GFP, and PC9-luc cells, which overexpress EGFR proteins, using flowcytometry. For flowcytometry, the cells (1 × 10⁵) were incubated with pan-IR700 (10 μg/mL) in the medium for 6 h at 37 °C. To confirm the binding specificity of the new conjugates, a competition assay was performed by adding excess untreated panitumumab (1 μg). The cells were analysed using a flow cytometer (Gallios; Beckman Coulter, Inc., Fullerton, CA, USA, cat # B43618) and the results were processed using the Kaluza software version 2.1 (Beckman Coulter).

Fluorescence microscopy

To detect the antigen-specific localization of IR700 conjugates, fluorescence microscopy was performed (A1Rsi; Nikon Instech, Tokyo, Japan). A431-luc-GFP, MDAMB468-luc-GFP, or PC9-luc cells (2 × 10⁴) were seeded on glass-bottom dishes and incubated for 24 h. Then, 10 μg/mL pan-IR700 was added onto the culture medium, and the cells were incubated at 37 °C for 6 h. Next, the cells were washed twice with phosphate-buffered saline (PBS). Propidium iodide (PI, final concentration 2 μg/mL; Thermo Fisher Scientific; cat # P1304MP) or SYTOX blue (final concentration: 2 μg/mL; Thermo Fisher Scientific; cat # S34857) were added 20 min before microscopic observation.³³ The cells were exposed to NIR light (20 J/cm²); serial microscopic images were captured.

In vitro NIR-PIT

For EGFR-targeted-NIR-PIT, A431-luc-GFP, MDAMB468-luc-GFP, and PC9-luc (1 × 10⁵) were seeded onto 12-well plates and incubated in a medium containing pan-IR700 (10 μg/mL) for 12 h at 37 °C. After washing twice with PBS, the cells were irradiated using an NIR light-emitting diode (L690-66-60; Ushio-Epitex, Kyoto, Japan) at 690 nm wavelength or 690 nm NIR light-laser (MLL-III-690; Changchun New Industries Optoelectronics Tech., Co., Ltd., Changchun, China).³⁴ The actual power density (mW/cm²) in the experiments was measured using an optical power meter (PM100; Thorlabs, Newton, NJ, USA).³⁵

The photocytotoxic effects of NIR-PIT were assessed by measuring luciferase activity. To monitor luciferase activity, 150 μg/mL D-luciferin-containing medium (Goryo Chemical, Sapporo, Japan) was administered to PBS-washed cells at 1 h after NIR-PIT, and the cells were analysed using a plate reader to detect their bioluminescence (Cytation 5; BioTek, Winooski, VT, USA).³⁶

Animals and tumour models

All in vivo procedures were conducted in compliance with the institutional guidelines of the Animal Care and Use Committee of Nagoya University. All mice were purchased from SLC Japan, Inc., Tokyo. During all experimental procedures, the mice were anaesthetised using isoflurane (Wako, Osaka, Japan). Approximately 9–15-week-old Balb/C-Slc-nu/nu mice were inoculated with A431-luc-GFP (5 × 10⁶), MDAMB468-luc-GFP (1 × 10⁷), or PC9-luc cells (5 × 10⁶) into the right, left, or both dorsa. The xenografted mice tumour model mirror human tumorigenesis and are the standard method used to assess EPR efficacy.

In vivo NIR-PIT

In vivo NIR-PIT was performed at 6 days after tumour cell inoculation. Mice were intravenously injected with 100 μg pan-IR700; on the next day following the injection, they were irradiated with NIR light (690 nm) at 0∼100 J/cm² using an NIR light-emitting diode (L690-66-60; Ushio-Epitex) or 690 nm NIR light-laser (MLL-III-690; Changchun New Industries Optoelectronics Tech., Co., Ltd., Changchun, China) (setting power was 1.0 W at 690 nm), unless otherwise specified, applied on the right tumour.37, 38, 39 For correlation between the SUPR effect and NIR-PIT, NIR-light doses were varying (20–100 J/cm²).

In vivo nano or micro drug delivery after NIR-PIT

After the NIR-PIT, 32 pmol of pegylated nontargeted quantum dots (Qdot 800 ITK™ Amino (PEG) Quantum Dots; Thermo Fisher Scientific) (QD800), 1 mg of NIR-poly (d,l-lactide-co-glycolic acid) (PLGA) (Near-IR Fluorescent PLGA particles; Sigma Aldrich, St. Louis, USA), or 0.4 μL MB of Sonazoid (KEGG DRUG; D05440) (GE Healthcare, Milwaukee, WI, USA) were injected intravenously, and in vivo imaging was performed for each.

In vivo IR700, QD800, and NIR-PLGA imaging

IR700, QD800-fluorescence, and NIR-PLGA-fluorescence were detected before and after therapy using a fluorescence imager (Pearl Imager; LI-COR Biosciences, cat # 9430–00).40, 41, 42

In vivo bioluminescence imaging

For bioluminescence imaging (BLI), 2.5 mg of D-luciferin (CAS# 115144-35-9) (Cat #7903) (Biovision, CA, USA) was injected intraperitoneally, and 10 min after the injection, the mice were captured on a bioluminescence imager (IVIS, PerkinElmer, Waltham, MA, USA).^43,44

CEUS imaging

Echocardiography was performed using a MX250 ultra-high frequency linear array transducer (15–30 MHz, centre transmit: 21 MHz, axial resolution: 75 μm) together with a Vevo® 3100 high-resolution Imaging System (FUJIFILM VisualSonics, Toronto, ON, Canada). Mice were sedated using 3% isoflurane and kept in the dorsal position on a heated pad at 37 °C (FUJIFILM VisualSonics). Isoflurane concentration was reduced to the minimum (1–2%) to achieve constant and comparable heart rates during examination.

Optical 3D tissue imaging

After the experiment, the mice were euthanised and the tumours were harvested. The extracted tissue was soaked into CUBIC-HL for approximately 7–14 days to remove blood components and lipid, among others. Then, the tissue was stained with a CD31 (PECAM-1) monoclonal antibody (390), eFluor™ 450 (RRID; AB_10598807) (Thermo Fisher Scientific, Catalog # 48-0311-82), and soaked into 50% CUBIC-R (Wako, Catalog # 298-85101) or 1 day for tissue clearing followed by 100% CUBIC-R for approximately 1–2 days for refractive index matching.⁴⁵ After tissue clearing, the measurement was performed using a light sheet microscope (Lightsheet 7; Zeiss, Göttingen, Germany).

Fluorescence immunostaining of the tumour tissue

To evaluate the SUPR effect after NIR-PIT, the tumours of euthanised mice were harvested and frozen in an optimal cutting temperature compound (Sakura Finetechnical, Tokyo, Japan). Sliced frozen tumours (10 μm thickness) were subjected to platelet/endothelial cell adhesion molecule-1 (PECAM-1 conjugated to PE) (Catalog # 3528S) (Cell Signaling Technology, Danvers, MA, USA) immunohistochemical staining. For immunostaining, HistoVT One (Nacalai Tesque, Kyoto, Japana) was added on to specimen and the samples were incubated 70 °C for 20 min. After washing twice, Blocking One Histo (Catalog # 06349-64) (Nacalai Tesque) was dripped on the samples and incubated for 10 min. After washing twice, PECAM-PE diluted solution was dripped on the sample and incubated for 24 h. After washing twice, the samples were mounted and enclosed for microscopic observation. The slides were imaged, and the fluorescence was assessed using a fluorescence digital microscope (BZ-X810; Keyence Corporation, Osaka, Japan).

Macro-sized (whole tumour slice imaging) imaging was performed with PEARL-fluorescence imager.

Regression model

To maximize the use of the continuous signal data, regression with a pharmacological model was used. A two-compartment model, comprising a compartment that represents intra-tumour vessels and another that represents uptake into tumour tissue, was established to evaluate signal retention in solid tumours. The model was fitted using temporal rate constants that represent the exchange between the intra-tumour vessels and tumour tissue compartments (k₁₂ and k₂₁) and the efflux from the intra-tumour vessels (k_E). The following two-compartment model was used for drug efflux from the central compartment into the tissue:

$$\mathrm{C}1\left(\mathrm{t}\right)=\mathrm{A}{\mathrm{e}}^{-\mathrm{\alpha }\mathrm{t}}+\mathrm{B}{\mathrm{e}}^{-\mathrm{\beta }\mathrm{t}}$$

$$\mathrm{C}2\left(\mathrm{t}\right)=\mathrm{D}({\mathrm{e}}^{-\mathrm{\alpha }\mathrm{t}}-{\mathrm{e}}^{-\mathrm{\beta }\mathrm{t}})$$

The harmonic signal measured in this study is the sum of the central (intra-tumour vessels) and peripheral (tumour tissue) compartments, and the discharge and exchange coefficients were expressed as follows:

$$\mathrm{C}1\left(\mathrm{t}\right)+\mathrm{C}2\left(\mathrm{t}\right)=\left(\mathrm{A}+\mathrm{D}\right){\mathrm{e}}^{-\mathrm{\alpha }\mathrm{t}}+\left(\mathrm{A}-\mathrm{D}\right){\mathrm{e}}^{-\mathrm{\beta }\mathrm{t}}$$

$${{\mathrm{K}}_{21}=\frac{\mathrm{B}\mathrm{\alpha }+\mathrm{A}\mathrm{\beta }}{\mathrm{A}+\mathrm{B}}\mathrm{K}}_{\mathrm{E}}=\frac{\mathrm{\alpha }\mathrm{\beta }}{{\mathrm{K}}_{21}}$$

Regression was performed using this model, and k_E was regarded as the emission factor for SUPR.

Statistical analysis

Data are expressed as means ± SEM as determined by a minimum of three experiments, unless otherwise indicated. Statistical analyses were performed using GraphPad Prism (GraphPad Software, San Diego, California, USA) and Python (https://www.python.org/). For two group comparisons, unpaired t test was used. Correlation was determined based on Pearson’s product-moment correlation coefficient. For liner regression, p < 0.05 was considered to indicate a statistically significant difference.

Role of the funding source

This research was supported by “Advanced Research Infrastructure for Materials and Nanotechnology in Japan”; ARIM JAPAN of MEXT, the Program for Developing Next-generation Researchers (Japan Science and Technology Agency), KAKEN (18K15923, 21K07217) (JSPS), CREST (JPMJCR19H2, JST), and FOREST-Souhatsu (JST). Mochida Memorial Foundation for Medical and Pharmaceutical Research; Takeda Science Foundation; The Japan Health Foundation; and Princess Takamatsu Cancer Research Fund.

Funders only provided financial support and had no role in the study design, data collection, data analysis, interpretation, and writing of the report.

Data and materials availability

All data associated with this study are presented in the paper or the Supplementary Materials.

Results

Production of pan-IR700

The pan-IR700 conjugate was synthesised for EGFR-targeted photoimmunotherapy. In the SDS-PAGE, each band of pan-IR700, or panitumumab, indicated a molecular weight of around 150 kDa, which is the specific molecular weight of panitumumab, and the band of the pan-IR700 complex showed NIR-fluorescence in the same SDS-PAGE gel, representing the signal of IR700 (Fig. 1a).

Fig. 1.

Production of panitumumab-IR700 (pan-IR700) as an EGFR-targeting conjugate for NIR-PIT, and evaluation of in vitro NIR-PIT effect. (a) Quality control of panitumumab-IR700 with SDS-PAGE (upper: Colloidal Blue staining, lower: IR700-fluorescence). (b) Confirmation of binding of pan-IR700 on EGFR-expressing cells with a flow cytometer. pan-IR700 could bind on A431-luc-GFP cells, mdamb468-luc-GFP cells, or PC9-luc cells (EGFR overexpressing cancer cells), while did not bind on 3T3-luc cells (human EGFR negative mouse cells, negative control to see unspecific binding). Pan-IR700 was inhibited binding on the EGFR-expressing cells with excess panitumumab (panitumumab blocking). (c) Fluorescence microscopic evaluation of in vitro NIR-PIT (2 J/cm²) targeting EGFR; A431-luc-GFP cells binding with pan-IR700 rapidly ruptured upon NIR-light-irradiation (2 J/cm²), which were stained with PI (Propidium Iodide, necrotic cell staining). Bar = 50 μm. In vitro NIR-PIT targeting EGFR on mdamb469-luc-GFP or PC-9 cells were also evaluated in Supplementary Figs. S1a and S2a. (d) Fluorescence microscopic observation of in vitro NIR-PIT (2 J/cm²) under co-culture of EGFR-positive (A431-luc-GFP) and EGFR-negative (3T3-RFP) cells; NIR-PIT (2 J/cm²) did not cause toxicity to the adjacent 3T3-RFP and only destroyed A431-luc-GFP cells which was stained Sytox Blue dead cell staining. Bar = 50 μm. (e) Quantitative cytotoxicity evaluation of in vitro NIR-PIT varying NIR-light doses with luciferase activities (n = 4, p = 0.5 J: 0.034026436, 1 J: 0.020806296, 2 J: 0.010444204 < 0.05, student’s t test).

Next, the binding of pan-IR700 to the EGFR antigen was confirmed using several cell lines including a murine cell line (3T3-luc) which is a negative control (no human EGFR expressing) and human cancer cell lines (A431-luc-GFP, MDAMB468-luc-GFP, and PC9-luc) which showed different EGFR expression. pan-IR700 showed no binding on the 3T3-luc murine cell line, suggesting pan-IR700 did not bind on no EGFR expressing cell. The EGFR-overexpressing human cancer cell lines A431-luc-GFP, MDAMB468-luc-GFP, and PC9-luc showed higher IR700-fluorescence than that of controls, suggesting that pan-IR700 can bind to these cells. Blocking study with excess naked panitumumab inhibited the binding of pan-IR700. These flowcytometric data indicate that pan-IR700 can specifically bind to the EGFR antigen and target EGFR-positive cells (Fig. 1b).

Effectiveness of in vitro NIR-PIT

To confirm the in vitro NIR-PIT effect, phase contrast fluorescence microscopy was performed before and after the NIR-light irradiation. Before NIR-light irradiation, A431-luc-GFP, an EGFR-positive skin cancer cell line, showed GFP-derived green fluorescence and NIR-fluorescence due to pan-IR700 binding. NIR-light irradiation (2J/cm²) reduced NIR-fluorescence via the NIR-PIT cell death mechanism and green-fluorescence due to extracellular leakage of the GFP protein, which is consistent with previous reports.^1,46,47

The necrotic cell death dye PI flowed into the cells when the cell membranes ruptured, resulting in stronger red fluorescence (Fig. 1c). The other cancer cell lines, MDAMB468-luc-GFP (human breast cancer) and PC-9-luc (human lung cancer), which highly expressed EGFR, also showed similar NIR-PIT effect in vitro (Supplementary Figs. S1a and S2a).

Next, to confirm selective cell rupture with EGFR-targeted NIR-PIT, microscopic examination was performed using Sytox Blue dead cell staining, with co-cultured A431-luc-GFP and 3T3-RFP. In vitro EGFR-targeted NIR-PIT (2J/cm²) induced cell death only in A431-luc-GFP cells without affecting the adjacent non-targeted 3T3-RFP (Fig. 1d).

Next, to quantitatively assess the effect of in vitro NIR-PIT with varying NIR-light doses, we evaluated the luciferase activity in A431-luc-GFP. NIR-PIT destroyed EGFR-positive cells in a light-dose dependent manner (n = 4, p = 0.5 J: 0.034026436, 1 J: 0.020806296, 2 J: 0.010444204 < 0.05, Student’s t test) (Fig. 1e). NIR-PIT also induced photocytotoxicity in the other cancer cell lines, MDAMB468-luc-GFP and PC-9-luc cells, in a light-dose dependent manner (Supplementary Figs. S1b and S2b).

Collectively, these results indicated that EGFR-targeted NIR-PIT could induce selective cytotoxicity in EGFR-expressing cancer cell lines, and that the effects were NIR-light dose-dependent.

Evaluation of the SUPR effect after NIR-PIT using QD800 nanoparticles

First, QD800, a quantum dot based fluorescent nanoparticle with a mean diameter of 20 nm, was used to evaluate the correlation between the SUPR effect and therapeutic efficacy. After NIR-PIT was performed on a A431-luc-GFP xenografted bilateral tumour mouse model, QD800 was intravenously injected. One tumour was irradiated and the contralateral tumour was shielded; SUPR was assessed by externally detecting the fluorescent signal of QD800. On the next day after NIR-PIT, decreased bioluminescence in tumour cells, which indicates tumour cell activity, was assessed via BLI (Fig. 2a).

Fig. 2.

Evaluation of Super Enhanced EPR effect (SUPR) after NIR-PIT using QD800 nanoparticles, and the correlation with in vivo NIR-PIT effect. (a) Scheme of the experiments. Pan-IR700 was intravenously injected at 1 day before NIR-light irradiation. In vivo NIR-PIT (100 J/cm²) was performed on a mouse tumour model and QD800 was injected intravenously after the treatment. The sequential QD800-fluorescence images were taken with a 800 nm-fluorescence pearl imager. At the next day of the NIR-PIT therapy, bioluminescence images (BLI) were taken for the NIR-PIT effect evaluation. (b) The regimen of SUPR-evaluation. One hour after in vivo NIR-PIT, 32 pmol QD800 was injected intravenously and QD800-fluorescence in the tumour area was assessed. (c) Representative in vivo BLIs. Evaluation of luciferase activity was decreased only in the tumour on the side irradiated with NIR-light (100 J/cm²). Luciferase activity in the other side (non-NIR-light irradiated tumour) was increased along with the tumour progression. (d) Quantification of luciferase activity. In vivo NIR-PIT (100 J/cm²) decreases luciferase activity on the treated tumour (n = 6, p = 0.000980195 < 0.05, vs. untreated side tumour, student's t test). (e) Representative in vivo IR700-fluorescence images. IR700-fluorescence decreased at immediately after NIR-light irradiation (100 J/cm²). (f) Correlation between IR700-fluorescence decrease and in vivo NIR-PIT effect with varying NIR-light doses. The decrease of IR700-fluorescence after NIR-light irradiation positively correlated with the decrease of tumour activity measured as luciferase activities (n = 16, r = 0.56161, p = 0.023584 < 0.05). (g) Sequential QD800-fluorescence images after in vivo NIR-PIT (100 J/cm²). On the NIR-light irradiated side tumour, QD800-fluorescence had more remained until 60 min than on the non-treated tumour. Ex vivo treated tumours were also confirmed the more 800 nm-fluoresce retention in NIR-PIT treated tumour than the non-NIR-light irradiated tumour. Bar = 5 mm. (h) Quantification of QD800-fluorescence-signals. Signal retained more predominantly increased in the treated tumours (NIR-PIT 100 J/cm²) than non-irradiated ones (n = 6, p = 30 min: 0.012807878 60 min: 0.006226658 < 0.05, vs. untreated side tumour, student's t test). Video images were in Supplementary Video S1a. (i) Correlation analysis between QD800-fluorescence signal and NIR-PIT therapeutic effect with varying NIR-light doses at each time point (n = 20, 10 min: r = −0.51077, p = 0.015136 < 0.05, 30 min: r = −0.63794, p = 0.001402 < 0.005, 60 min: r = −0.65118, p = 0.001191 < 0.005). The increase of QD800-fluorescence at each time point represented the enhancement of SUPR, and the decrease in luciferase represents the NIR-PIT therapeutic effect.

Intravenous QD800 was administered at 1 h after treatment, according to a previous study.¹⁰ Changes in the fluorescent signal of QD800 in the tumour were followed for 60 min after QD800 intravenous infusion (Fig. 2b).

Although the luciferase activity of A431-luc-GFP control tumours usually increases exponentially according to the rapid tumour growth, the NIR-PIT-treated (100 J/cm²) luciferase activities in A431-luc-GFP tumours decreased significantly (Fig. 2c).

Quantification of the luciferase activities by determining relative light units (RLU) showed a significant decrease (n = 6, p = 0.000980195 < 0.001, vs. untreated side tumour, by Student’s t test) in the NIR-PIT (100 J/cm²) group mice tumours on the day after treatment (Fig. 2d).

For analysis of IR700-fluorescence, at next day after the injection of pan-IR700, pan-IR700 specifically accumulated in the tumour, which was confirmed by increased IR700-fluorescence. Immediately after NIR-light irradiation (100 J/cm²), IR700-fluorescence decayed, suggesting that a photo-chemical reaction occurred in the treated tumours (Fig. 2e). Furthermore, this decrease in IR700-fluorescence positively correlated with the therapeutic effect with varying NIR-light doses (n = 16, r = 0.56161, p = 0.023584 < 0.05) (Fig. 2f), and the changes were consistent with those in previous reports.¹

In vivo fluorescence imaging of QD800 administered at 1 h after NIR-light irradiation showed that the QD800-fluorescent signal was uniformly distributed in the body immediately after administration and was gradually eliminated from the body. However, the signal intensity was retained for a longer time in the NIR-PIT-treated tumours (100 J/cm²) than in the non-irradiated tumours (Fig. 2g), suggesting that NIR-PIT enhances the retention and permeability of injected QD800 nanoparticles in the tumour (Supplementary Video S1a). Ex vivo imaging of tumours confirmed the higher QD800-fluorescence in the treated tumours than in the controls. This phenomenon in known as the super enhanced retention and permeability (SUPR) effect (Fig. 2g). Quantification of the QD800-fluorescent signals showed a significant increase (four-fold) in the relative fluorescent signals in the NIR-PIT-treated tumours (100 J/cm²) (n = 6, p = 30 min: 0.012807878, 60 min: 0.006226658 < 0.05, vs. untreated side tumour, Student’s t test) (Fig. 2h). The other tumour types, MDAMB468-luc-GFP and PC-9-luc, also showed SUPR of QD800 with NIR-PIT (Supplementary Figs S1 and S2). Analysis between relative QD800 signals and treatment effectiveness with varying the NIR-light doses showed positive correlation between the QD800 signals and therapeutic effects at each time point (n = 20, 10 min: r = −0.51077, p = 0.015136 < 0.05, 30 min: r = −0.63794, p = 0.001402 < 0.005, 60 min: r = −0.65118, p = 0.001191 < 0.005) (Fig. 2i). Altogether, these results demonstrate that the quantification of QD800-SUPR effect with NIR-PIT could correlate with the therapeutic effect at one day after therapy, which could serve as a predictive therapeutic biomarker.

Next, frozen sections of the tumours were prepared to confirm the accumulation of QD800 particles in the tumours, which showed that QD800 particles were more superimposed in the NIR-PIT-treated tumours than in the controls (Supplementary Fig. S3). Quantification of the QD800 particles fluorescence in the frozen sections showed significantly higher fluorescence in the NIR-PIT-treated tumours than in the controls (n = 8, p = 0.0006 < 0.001, vs. untreated side tumour, student's t test) (Supplementary Fig. S3b).

Then, pathological analysis of detailed observation of the immuno-stained frozen tumour sections confirmed that pan-IR700 accumulates higher around vasculature in tumours with pan-IR700 injection without NIR-light. The fluorescence analysis of the specimen demonstrated that NIR-PIT (100 J/cm²) significantly damaged the tumour cells, especially around the blood vessels, as indicated by PECAM-1 staining, and that QD800 was more superimposed around and along these vessels in the treated tumours than in the controls (Fig. 3a). Under the same conditions, the control tumours and tumours with pan-IR700 injection without NIR-light showed almost no QD800-fluorescent signals (Fig. 3a). The fluorescence intensity observed in a line perpendicular to the blood vessels showed that the peak of QD800-fluorescence was similar to that of PECAM-1 fluorescence staining, while that of the control tumours and tumours with pan-IR700 injection without NIR-light showed no remarkable peaks of QD800 and PECAM-1 (Fig. 3b).

Fig. 3.

Histological analyses of SUPR effect with QD800 after in vivo NIR-PIT. (a) Fluorescence microscopic observation of the tumours. Immunostainings were performed on the frozen tumour specimen; blood vessels: red (PECAM-1), panitumumab-IR700: magenta, QD800: green. Bar = 200 μm. (b) Fluorescence intensity of PECAM-1 (red) and QD800 (green) over a cross-section of cells along the white lines in the frozen tumour specimen. (c) Quantification of QD800-fluorescence-signal along the intratumoral vessels. The QD800-fluorescence-signal besides the vascular was predominantly and significantly more increased in the NIR-PIT (100 J/cm²) group (n = 6, p = 0.00000108 < 0.05, vs. untreated side tumour, student's t test). (d) A representative image of tumours before and after tissue clearing with CUBIC-R. Bar = 1 cm. (e) 3D imaging of the tumours with a light sheet fluorescence microscopy. Immunostainings of cleared 3D tumour are showed. Blood vessels: red (CD34), QD800: green. Bar = 300 μm. (f) Quantification of microvascular diameter in cleared tumours. The microvascular in the NIR-PIT (100 J/cm²) group inflated around 4-folds (n = 5, p = 0.0000000061 < 0.05, vs. untreated side tumour, student's t test), compared to control. (g) Quantification of co-localization between microvascular and QD800-fluorescence. About 4-fold increase in the amount of QD800-fluorescence consistent along with the presence of microvascular (n = 5, p = 0.011954 < 0.05, vs. untreated side tumour, student's t test).

In the quantitative analysis, the QD800-fluorescence around the vessel wall area was significantly more increased (about 2.5-fold increase) in the NIR-PIT-treated tumours (100 J/cm²) than in the other tumours (n = 6, p = 0.00000108 < 0.05, vs. untreated side tumour, Student’s t test) (Fig. 3c).

We also performed tumour tissue clearing for detailed analysis. 3D tissue imaging was performed on the treated explanted tumour to confirm the distribution of blood vessels and QD800 nanoparticles. The tumour tissues were cleared with the CUBIC-R method and the blood vessels were simultaneously detected with CD34 staining (Fig. 3d). In the NIR-PIT-treated tumours (100 J/cm²), the vascular diameter was increased and strong CD34-staining was observed, suggesting that NIR-PIT not only destroyed the perivascular endothelial tumour cells but also promoted intense vascular dilatation. Furthermore, considerable accumulation of QD800 nanoparticles was detected in the dilated areas (Fig. 3e). Quantification of the vascular diameter from the cross-sectional slides showed significant dilation of approximately 4-fold in the NIR-PIT-treated tumours (100 J/cm²) (average; 46.3406 μm) compared to the non-NIR-light irradiated control tumours (average; 11.3192 μm) (n = 5, p = 0.0000000061 < 0.05, vs. untreated side tumour, Student’s t test) (Fig. 3f). In addition, overlap quantification of blood vessels and QD800-fluorescent signals showed a significant increase (n = 5, p = 0.011954 < 0.05, vs. untreated side tumour, Student’s t test) (Fig. 3g).

These observations indicate that, QD800-nanoparticles were remarkably permeabilized and retained in NIR-PIT-treated tumours, mainly along with the dilated peripheral of blood vessels, at 1 h after NIR-PIT.

Taken together all the histologic results, the nano-sized QD800 particles were significantly accumulated, permeabilised, and retained in NIR-PIT-treated tumours, and the NIR-PIT dilated the tumour blood vessels since the closer the distance from the tumour vessels, the larger the pan-IR700 distribution, thereby increasing the space around to the vessels, as consistent to previously reported.^48,49

Evaluation of the SUPR effect after NIR-PIT using NIR-PLGA microparticles

Conventional EPR and previous reported SUPR effects with NIR-PIT have been thought to be exerted by nanoparticles sized 10∼200 nm. However, the 3D-cleared tissue images suggested that dilated tumour vessels were stretched around inside the NIR-PIT-treated tumours (Fig. 3e and f). These observations urged us to test whether the micro-sized particles could be permeabilised, retained, and accumulated inside NIR-PIT-treated tumours.

To evaluate the micro-sized SUPR effect, 2 μm-sized NIR-PLGA particles were used (Fig. 4a). The sequence of changes in the fluorescence of 2 μm-sized NIR-PLGA particles in the tumours was followed for 5 min after the intravenous infusion (Fig. 4b and c), which showed that NIR-PIT further enhances the retention and permeability of injectable 2 μm-sized NIR-PLGA particles inside the treated tumours compared to the untreated ones (Supplementary Video S1b).

Fig. 4.

Evaluation of micro-sized-SUPR effect after NIR-PIT using 2 μm-sized NIR-Poly(lactic-co-glycolic acid) (NIR-PLGA). (a) A representative image of 2 μm-sized NIR-PLGA (catalog sized 2 μm). The diameter of NIR-PLGA is around 2 μm. Bar = 50 μm. 5 μm NIR-PLGA is showed in Supplementary Fig. S4a. (b) Scheme of the experiments. Pan-IR700 was intravenously injected at 1 day before NIR-light irradiation. In vivo NIR-PIT (100 J/cm²) was performed on a mouse tumour model and NIR-PLGA was injected intravenously after the treatment. The sequential NIR-PLGA-fluorescence images were taken with a 800 nm-fluorescence pearl imager. At the next day of the NIR-PIT therapy, bioluminescence images (BLI) were taken for NIR-PIT effect evaluation via luciferase activity. (c) Regimen of this study was shown. 2 μm-sized NIR-PLGA was injected intravenously immediately after NIR-PIT treatment (100 J/cm²), and 2 μm-sized NIR-PLGA-fluorescence (800 nm) in the tumour area was assessed. (d) BLI of A431-luc-GFP xenograft models before and after NIR-PIT (100 J/cm²). Luciferase activity was decreased only in tumours irradiated with NIR-light (n = 3, p = 0.004542163 < 0.05, vs. untreated side tumour, student's t test). (e) Sequential 2 μm-sized NIR-PLGA-fluorescence (800 nm) images after in vivo NIR-PIT (100 J/cm²). On the NIR-light irradiated side tumour, 2 μm-sized NIR-PLGA-fluorescence had more increased and retained until 5 min than the non-treated tumour. Ex vivo 2 μm-sized NIR-PLGA-fluorescence of extracted tumour also showed higher fluorescence, consistent with in vivo 2 μm-sized NIR-PLGA-fluorescence. Bar = 5 mm. (f) Quantification of 2 μm-sized NIR-PLGA-fluorescence signals. Signal retained more predominantly increased in the NIR-PIT (100 J/cm²) treated tumour than control (n = 4, p = 4 min: 0.034319917, 5 min: 0.017826908 < 0.05, vs. untreated side tumour, student's t test). (g) 2 μm-sized NIR-PLGA-fluorescence macro-images of ex vivo frozen tumour slice. On the NIR-light irradiated side tumour (NIR-PIT (100 J/cm²)), 2 μm-sized NIR-PLGA-fluorescence signal was higher than non-irradiated tumour slice. Bar = 5 mm. (h) 2 μm-sized NIR-PLGA-fluorescence quantification of frozen tumour slice. The fluorescence signal was predominantly and significantly increased in the NIR-PIT (100 J/cm²) group (n = 4, p = 0.002937117 < 0.05, vs. untreated side tumour, student's t test). Video images were showed Supplementary Video S1b. (i) Fluorescence microscopic observation of the tumour frozen sections with immunostainings. Blood vessels: red (PECAM-1), panitumumab-IR700: magenta, PLGA (2 μm): green. Bar = 100 μm. (j) Quantification of 2 μm-sized NIR-PLGA-fluorescence signal along the intratumoral vessels. The 2 μm-sized NIR-PLGA-fluorescence signal besides the vascular was predominantly and significantly increased in the NIR-PIT (100 J/cm²) group (n = 3, p = 0.036197743 < 0.05, vs. untreated side tumour, student's t test).

As in the QD800-SUPR experiments, the therapeutic effect was evaluated via BLI. Quantification of the luciferase activities in RLU showed a significant decrease (n = 3, p = 0.004542163 < 0.05, vs. untreated side tumour, Student’s t test) in the NIR-PIT group mice (100 J/cm²) tumours on next day after treatment (Fig. 4d).

In vivo 800 nm-fluorescence imaging of 2 μm-sized NIR-PLGA was performed at 1 h after NIR-light irradiation (100 J/cm²). The NIR-imaging showed that the 2 μm-sized NIR-PLGA-fluorescent signals at 800 nm was uniformly distributed in the body immediately after administration and was gradually eliminated. However, the signal intensity was more accumulated and retained in the NIR-PIT-treated tumours than in non-treated tumours (Fig. 4e), suggesting that NIR-PIT increases the retention and permeability of injectable 2 μm-sized NIR-PLGA inside the tumour. Ex vivo imaging of tumours confirmed the higher fluorescence of 2 μm-sized NIR-PLGA particles in the treated tumours than in the controls. Quantification of the 2 μm-sized NIR-PLGA-fluorescent signals showed significantly increased (1.4 fold) relative fluorescence in the NIR-PIT-treated tumours compared to non-NIR-light irradiated control tumours (n = 4, p = 4 min: 0.034319917, 5 min: 0.017826908 < 0.05, vs. untreated side tumour, Student’s t test) (Fig. 4f).

Next, frozen sections of the tumours were prepared to confirm the accumulation of the 2 μm-sized NIR-PLGA particles in the tumours, which showed that the 2 μm-sized NIR-PLGA particles were more superimposed in the NIR-PIT-treated tumours than in the controls (Fig. 4g). Quantification of the 2 μm-sized NIR-PLGA particles fluorescence in the frozen sections showed significantly higher fluorescence in the NIR-PIT-treated tumours than in the controls (n = 4, p = 0.002937117 < 0.05, vs. untreated side tumour, Student’s t test) (Fig. 4h). Detailed observation of the immuno-stained frozen tumour sections indicated that the 2 μm-sized NIR-PLGA particles were more superimposed around the dilated blood vessels in the NIR-PIT-treated tumours than in tumours with pan-IR700 injection without NIR-light (Fig. 4i). Quantitative analysis showed that the 2 μm-sized NIR-PLGA particles fluorescent signals along the vessel wall area was significantly increased (approximately 2-fold increase) in the NIR PIT-treated tumours more than in the other tumours (n = 3, p = 0.036197743 < 0.05, vs. untreated side tumour, Student’s t test) (Fig. 4j).

In addition, a similar experiment using 5 μm-sized NIR-PLGA (Supplementary Fig. S4a) was performed to explore the particle size-upper limit. In vivo fluorescence imaging showed a limited effect of NIR-PIT on the accumulation of 5 μm-sized NIR-PLGA particles (Supplementary Fig. S4b and S4c, Supplementary Video S1c); however, direct ex vivo quantification of fluorescence in tumours showed a significantly higher fluorescence of accumulated particles in the NIR-PIT-treated tumours than that in the untreated tumours (n = 6, p = 0.040993178 < 0.05, vs. untreated side tumour, Student’s t test) (Supplementary Fig. S4d).

The 5 μm-sized NIR-PLGA particles were particularly supposed to be captured by the lung microvasculature (Supplementary Video S1c). Hence, the particle size-upper limit of the NIR-PIT-induced enhancement of particle accumulation and retention was thought to be approximately 5 μm.

The results presented herein demonstrate that besides the conventional nano-sized SUPR effect, NIR-PIT also enhances the permeability and retention of micro-sized particles in tumours, which we named as the “micro-sized” SUPR effect, to distinguish it from the conventional nano-sized SUPR effect. These results also proved that the particle size-upper limit for NIR-PLGA particles with NIR-PIT was approximately 5 μm.

Ultrasound harmonic imaging with micro-bubbles predicts NIR-PIT therapeutic effectiveness in vivo

The positive correlation between the nano-sized SUPR effect evaluated using QD800 nanoparticles and the NIR-PIT antitumour effect suggested that the evaluation of SUPR effect could predict NIR-PIT effectiveness. We also proved that the micro-sized SUPR effect can be exerted using micro-sized NIR-PLGA particles. Because neither QD800 nor NIR-PLGA can currently be used in clinical practice, we were urged to evaluate the micro-sized SUPR effect using micro-bubbles (Sonazoid) and harmonic ultrasound imaging. The ultrasound contrast agent Sonazoid was selected because it is already approved in human and used for the evaluation of liver tumours (hepatic cell carcinoma). Moreover, when NIR-PIT was performed for recurrent HNCC (Head and Neck Cancers) in the clinic, the tumour is usually evaluated with ultrasound imaging. After the evaluation of the tumour site via ultrasound, cylindrical optic diffusers were inserted into the tumour along the guide of ultrasound, and then NIR-light was irradiated. The ultrasound imaging is already essential in clinical NIR-PIT, and evaluation with harmonic ultrasound imaging with using micro-bubbles is thought to be easy. Sonazoid micro-bubbles are sized approximately 2 μm (Fig. 5a, Supplementary Video S2) and are metabolised and excreted in the exhaled breath within a few minutes after intravenous injection.⁵⁰ This rapid metabolism enabled us to evaluate its permeability and retention in the tumour before and after NIR-PIT-treatment in the same tumour and the same mouse.

Fig. 5.

Evaluation of micro-sized-SUPR effect after NIR-PIT using a micro-sized ultrasound contrast agent (micro-bubbles). (a) A representative image of the ultrasound contrast agent micro-bubbles, Sonazoid. The diameter of Sonazoid is around 2 μm. Bar = 25 μm. (b) Scheme of the study with an ultrasound contrast agent; Sonazoid. Pan-IR700 was intravenously injected at 1 day before NIR-light irradiation. To evaluate SUPR via micro-bubbles, intravenous injection of Sonazoid and harmonic ultrasound sonography was performed before and after the NIR-PIT with same mouse. At the next day of the NIR-PIT, the treated tumour was evaluated by BLIs and analyzed the correlation with micro-sized SUPR effect. (c) Regimen of this study was shown. Sonazoid was injected intravenously before and immediately after treatment, and serial harmonic echogenic signal in the tumour area was assessed. The ultrasonographic analysis before and after the NIR-PIT was evaluated. (d) Quantification of luciferase activity at 1 day after treatment. NIR-PIT (0, 30, 60 J/cm²) suppressed the tumour luciferase activity (n = 5, p = 0.011181862 < 0.05, vs. untreated control tumour, student's t test). (e) Representative sequential harmonic echogenic images of the tumours after in vivo NIR-PIT (60 J/cm²). In the NIR-PIT group, the harmonic echogenic signal flowed in the tumour more rapidly after intravenous injection and disappeared more slowly than in the control group, which showed the more retaining harmonic echogenic signals. (f) Two-compartment model with peripheral and tumour tissue compartments. C₁(t) and C₂(t) represent the drug concentration in blood (peripheral compartment) and tumour tissue (tumour tissue compartment), respectively. The first order rate constant k_E describes all elimination pathways, including clearance from tumour tissue compartment. The first-order rate constants k₁₂ and k₂₁ describe the exchange between the two compartments. (g) The regression of harmonic echogenic signals in the tumour assessed by a two-compartment model. Blue: raw data from the serial harmonic echogenic signal, green: moving average of raw data, red: regression line from two-compartment model. (Please see the method). (h) Change in k_E values before and after NIR-PIT (0, 30, 60 J/cm²). As NIR-light was irradiated, k_E values Ratio (k_E after/before) decreased (n = control: 5, 30 J: 6, 60 J: 16, p = 0.023022825 < 0.05, vs. untreated side (without NIR-irradiation) tumour, student's t test). (i) Correlation analysis between k_E Values (k_E after/before) (varying NIR-light dose) and anti-tumour effect at 1 day after NIR-PIT (r = 0.74292, p = 0.0000137998 < 0.005). The decrease of k_E rationally meant enhancement of micro-sized-SUPR, and the decrease of RLU represented anti-tumour effect at 1 day after NIR-PIT.

To evaluate the micro-sized SUPR effect with micro-bubbles, an image-based study was performed (Fig. 5b). Micro-sized SUPR was evaluated by externally detecting the harmonic signals of the micro-bubbles with an ultrasound echo device (Supplementary Fig. S5a). To facilitate imaging the same site of the tumour before and after NIR-light irradiation, the echo probe was fixed at the same position during the procedure (Supplementary Fig. S5b) and an indwelling needle was placed in the tail vein of the mouse for the intravenous injection (Supplementary Fig. S5c). The surface of the echo gel was adjusted at a right angle to the NIR-light to reduce the refraction of the NIR-light, and we assessed the actual NIR-light dose using a power meter with the same situation throughout the experiment (Supplementary Fig. S6). The therapeutic effect of NIR-PIT was determined by quantitatively measuring luciferase activities via in vivo BLI (Fig. 5b).

To evaluate the micro-sized SUPR effect with micro-bubble, we first administered the intravenous infusion of micro-bubble immediately after NIR-PIT. The residual harmonic signals were continuously followed and recorded for approximately 2.5 min after the micro-bubble injection (Fig. 5c). With EGFR-targeted NIR-PIT, the luciferase activities determined in RLU in the A431-luc-GFP tumours decreased in a dose-dependent manner (n = 5 p = 0.011181862 < 0.05, vs. untreated control tumour, Student’s t test) (Fig. 5d). The sequential in vivo contrast harmonic echogenic imaging with Sonazoid showed that the harmonic signals immediately after the administration were more distributed in the treated tumours than in the untreated ones (Fig. 5e, Supplementary Video S3).

Since micro-bubbles are distributed in both the blood and tissue fractions inside the body, the two-compartmental model could be used for analysing the harmonic ultrasound signals of the micro-bubbles.⁵¹ C₁(t) and C₂(t) represented the drug concentration in blood (peripheral compartment) and tumour tissue (tumour tissue pathways, including clearance from tumour tissue compartment), respectively. The first-order rate constants k₁₂ and k₂₁ reflected the substrate exchange between the two compartments (Fig. 5f). The harmonic signal intensity of detected by contrast imaging fits the regression curve well (Fig. 5g). When the harmonic signals (derived from injected micro-bubbles in the tumour) were detected before NIR-PIT, the signals in the tumour decreased steeply after the influx, showing almost the same curve as the control, which suggested that the micro-bubbles just pass-through the vessels inside tumours. Intriguingly, the decrease in harmonic signals after NIR-PIT was more gradual than that before NIR-PIT, indicating enhanced permeability and retention of particles in the tumour after NIR-PIT. In contrast, the harmonic signal curve did not change between before and after NIR-PIT in the control tumours with NIR-light irradiation without pan-IR700.

Taken together, these results suggested that NIR-PIT induced a micro-sized SUPR effect, and that the effect could be evaluated using the micro-bubble contrast agent Sonazoid with harmonic echo imaging.

The emission coefficient k_E values were calculated from the harmonic signals according to the compartmental model in Fig. 5f. The comparison of k_E obtained from this regression revealed that the k_E value decreased in an NIR-light-dose-dependent manner in the NIR-PIT-treated tumours (n = control: 5, 30 J: 6, 60 J: 16, p = 0.023022825 < 0.05, vs. untreated side (without NIR-irradiation) tumour, Student’s t test) (Fig. 5h). To clarify the relationship between k_E values and the therapeutic effect at 1 day after NIR-PIT, a correlation analysis was performed. A positive correlation between the k_E values and NIR-PIT effect at 1 day after irradiation was detected (r = 0.74292, p = 0.0000137998 < 0.005) (Fig. 5i).

These findings suggested that the k_E value ratios calculated from the harmonic ultrasound signals detected via micro-bubbles before and after NIR-PIT could confirm and predict the NIR-PIT effect.

Discussion

In this study, we demonstrated that the evaluation of micro-sized SUPR via CEUS imaging with micro-bubbles could be exploited for the monitoring of NIR-PIT therapeutic effect. We named this procedure MS-CEUS (Micro-SUPR evaluation via CEUS imaging), which could serve as an image-based marker for NIR-PIT therapeutic effect (Fig. 6).

Fig. 6.

Micro-sized SUPR via CEUS imaging with micro-bubbles could be exploited for the monitoring of NIR-PIT therapeutic effect. Serial harmonic signal by Sonazoid persists in well-treated tumours and does not retain in poorly treated tumours. By assessing the increasing microbubbles’ permeability and retention via the harmonic signals comparing the before NIR-light irradiation, the future efficacy of NIR-PIT can be predicted.

First, we confirmed the targeting of EGFR via NIR-PIT in vitro. Next, we performed fluorescence monitoring for to detect the SUPR effect of NIR-PIT in vivo using nano-sized QD800 particles (20 nm). As a result, we found and demonstrated that the SUPR effect could be used as a marker to confirm and predict the therapeutic effect of NIR-PIT. Histologic analysis of the SUPR effect using QD800 nanoparticles showed accumulation of particles around and along the vessels inside the treated tumours. Intriguingly, immunostaining of 3D-cleared tumours revealed that NIR-PIT caused vasodilation, and that the diameter of the vessels four-fold larger than that of the vessels in the control tumours. These observations opened the door for examining whether micro-size particles could be permeated and retained in the NIR-PIT-treated tumours. Using 2 μm-sized and 5 μm-sized NIR-PLGA particles, we demonstrated that the SUPR effect can be measured not only with nanoparticles but also with microparticles. To investigate whether the monitoring of micro-sized SUPR could be clinically applied, we used clinically-approved ultrasound micro-bubbles, demonstrating that the micro-sized SUPR-effect could be measured via CEUS-based technologies. Moreover, we proved the hypothesis that the micro-sized SUPR-effect determined via CEUS imaging is positively correlated with NIR-PIT therapeutic effect. Therefore, evaluation of micro-sized SUPR-effect with CEUS has the potential to be used as a biomarker for NIR-PIT.

The particle size of the conventional EPR and SUPR effect ranged between 10 and 200 nm, which can easily circulate in the body; particles with a diameter of 200 nm are not easily excreted in the liver, and those with a diameter of 10 nm are less susceptible to renal excretion.¹¹ Therefore, micro-sized particles were thought to be impossible to accumulate in the tumour. However, this study has revealed that the SUPR effect allows the accumulation of particles with diameters greater than 200 nm, or even 2 μm∼5 μm. Although 5 μm-sized particles tend to be captured by lung capillary vessels, the ex vivo imaging of tumours showed increased accumulation of particles in NIR-PIT-treated tumours. Moreover, micro-bubbles can adapt to the diameter and shape of the blood vessels. Therefore, micro-bubbles such as Sonazoid can be used for the evaluation of SUPR effect after NIR-PIT. We named this new concept as the micro-sized SUPR effect.

The MS-CEUS method presented in this study has advantages over other treatment effect-monitoring modalities. First, it has the advantage of Sonazoid's rapid metabolism, which allows pre- and post-treatment SUPR assessment for the same tumour. Because the SUPR effect varies from a tumour to another and from a patient to another, enabling the monitoring of differences in the EPR effect before and after treatment is crucial for precise evaluation of NIR-PIT therapeutic effect. With this method, a tailored and suitable image-based biomarker for each patient can be realised. Second, because the micro-bubble ultrasound contrast agent Sonazoid has already been clinically approved for detecting hepatic cellular carcinomas,⁵² we believe this method could be easily translated into clinical use. Third, harmonic ultrasonography devices are widely used in hospitals; hence, our method does not require additional installations. At last, ultrasound diagnostic devices are portable and can therefore be placed beside the patient during the therapeutic procedure.

To adjust the evaluation method proposed in this study for clinical use, both NIR-PIT and SUPR evaluation for deep tumours should be optimised and developed. Methods of NIR-light irradiation, such as subcutaneous insertion of an optical fibre diffuser,⁵³ transvacuolar access sources,⁵⁴ and implantable light sources,⁵⁵ are being developed. In addition to these light sources, there are other options that can be used as detectors in this evaluation method, such as endoscopic ultrasonography (EUS), endobronchial ultrasonography (EBUS), and the ultrasound probes for usual abdominal echo. As for now, since NIR-PIT has been conditionally approved for previously treated recurrent head and neck carcinoma, we can use the ordinary probes for ultrasound detection. Our concept can be applied to oesophageal, gastric, and lung cancers, among others, and the devices such as EUS, EBUS, and new probes could be exploited. Moreover, recently combination of ultrasound and photoacoustic imaging devices have been developed,⁵⁶ MS-CEUS could be used with such new devices.

In this study, a two-compartment model was used to evaluate the SUPR effect in CEUS images. However, rather than comparing k_E values, it is more user-friendly and easier to evaluate the contrast average power from Frame 100 to Frame 2000 (Fig. 5g). In a clinical setting, for example, evaluating the ratio of contrast-averaged power between t100 Frame and t2000 Frame would be more easily utilized as an image-based therapeutic biomarker. In any case, based on the results of this study, there is a need to evaluate in clinical studies to more easily assess the increase in retention when treatment is completed.

This study has some limitations. First, the pharmacological model used in our study does not include clutter filtering and ignores coefficients of variation such as heart rate, which might trivially affect Sonazoid signals. When this method is applied clinically, a more realistic model should be used. We need to optimise this in future clinical studies. Second, MS-CEUS requires imaging before and after therapy for comparing, and the position might be changed to monitor the tumour. However, intraoperative marks can be used to decide the precise position for evaluation. Thirdly, MS-CEUS provides only a 2D evaluation, limiting the 3D nature of CEUS. Finally, it is necessary to establish a measurement system that does not rely on individual techniques, perhaps via the automation of k_E value calculation.

In conclusion, we demonstrated that NIR-PIT can induce micro-sized SUPR, and its evaluation via CEUS imaging could be used as a monitoring marker for NIR-PIT therapeutic effect.

Contributors

KM and KS have equally contributed to this work. KM mainly conducted the experiments and KS interpreted the experimental results. KM, MY, MS, NF, KG, HY, YB, MS, and KS analysed the data. KG, HY, and YB supported and did the tumour clearing experiments. MS, AK, and YK supported the animal studies. KM and KS wrote the manuscript. KS designed and supervised all the study. KM, KS, and MS verified the underlying data. All authors have agreed on the final version of the manuscript.

Data sharing statement

Data can be made available upon request to the corresponding author.

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.