Cerenkov radiation-activated probes for deep cancer theranostics: a review

Abstract

Cerenkov radiation (CR) from radionuclides and megavoltage X-ray radiation can act as an in situ light source for deep cancer theranostics, overcoming the limitations of external light sources. Despite the blue-weighted emission and low quantum yield of CR, activatable probes-mediated CR can enhance the in-vivo diagnostic signals by Cerenkov resonance energy transfer and also can produce therapeutic effects by reactive species generation/drug release, greatly promoting the biomedical applications of CR. In this review, we describe the principles and sources of CR, construction of CR-activated probes and their application to tumor optical imaging and therapy. Finally, future prospects for the design and biomedical application of CR-activated probes are discussed.

1. Introduction

Light-based diagnosis and treatment methods have developed rapidly: examples include fluorescence imaging, optoacoustic imaging, photothermal therapy, and photodynamic therapy (PDT) ^1-5. However, these methods are less effective for deep tumors because of the rapid attenuation of light as it passes through tissue ⁶. To overcome this limitation, Cerenkov radiation (CR) from radionuclides and megavoltage X-ray radiation can serve as an in situ excitation source within deep tissues ⁷. Blue-weighted CR occurs when a charged particle travels faster than the speed of light in a given medium ^{8, 9}. Improvements in optical imaging technology and the increasing sensitivity of CCD cameras allow the detection of blue-weighted CR photons for functional molecular imaging ^10-11. Preclinical and clinical explorations of CR imaging have demonstrated its potential for high throughput, high sensitivity, and rapid imaging ^12-17. Nevertheless, current CR-based imaging lacks the necessary depth because the blue-weighted CR photons do not penetrate deep enough to image certain tumors ^{10, 18}.

Activatable probes-mediated CR provides a way to enhance the diagnostic signals or produce therapeutic effects ^{19, 20}. These optical probes participate in Cerenkov resonance energy transfer (CRET) to red-shift emission into the near-infrared (NIR) region or red region of the visible spectrum, enabling sensitive detection from deeper inside the tissue. Besides, some activatable probes can produce therapeutic effects at greater depths: for example, ultraviolet (UV)-responsive probes can be activated by CR to produce therapeutic reactive species or to release drugs, referred to as CR-induced therapy (CRIT) ²¹. The combination of activatable probes and CR expands the biomedical applications of CR.

In this review, we summarize CR-activated probes that have been reported to enhance optical imaging and therapy. CR efficiently triggers these optical probes with red-shifted emission, which include quantum dots, gold nanoclusters, lanthanide-based downconversion probes, persistent luminescence nanoparticles (PLNPs), and organic dye-associated NPs. The CR-induced PDT, photoimmunotherapy, drug release and even combinations of these therapies are also discussed in detail.

2. Cerenkov radiation

2.1 Principles of CR

CR is generated through the interaction of charged particles and matter ⁹. It can be produced from radionuclides, megavoltage X-ray radiation, cosmic events and nuclear reactors; clinical facilities may have access to radionuclides and linear accelerators ¹⁹. The radionuclides release α, β, or γ particles during the decay, and the subatomic particles are charged when polarizing the medium. A similar process occurs in megavoltage X-ray radiation, except that the kinetic energy of the resulting electrons is 2-3 orders of magnitude greater when they pass through tissue ¹⁵.

When a charged particle travels faster than the speed of light through a dielectric medium, the surrounding molecules in the medium become locally polarized. After the particles pass through the medium, the molecules return to their ground state and release blue-weighted light called CR (Figure 1) ^{15, 19}. CR intensity is related to the velocity and energy of the charged particles (υ), as well as the refractive index of the medium (c). The threshold of input energy needed to generate CR can be described as [19]:

Figure 1.

CR is produced by the medium through which charged particles propagate. Reproduced with permission ¹⁹. Copyright 2017, Nature Publishing Group.

Additionally, the photon flux of CR at a given frequency can be calculated using the Frank-Tamm equation:

which demonstrates that CR is blue-weighted and decreases with 1/λ² from the UV to infrared regions ¹⁹.

2.2 Sources of CR

2.2.1 Radionuclides

Radionuclides undergo decay through capture of positrons (β⁺) or electrons (β^-), followed by an isomeric transition. CR arises when charged particles travel faster through the surrounding medium than the phase velocity of light ²². Table 1 lists radioisotopes that emit energy over the CR threshold, such that they induce CR emission ²³. Positron emitters in clinical use of positron emission tomography (PET) include ¹⁸F, ⁶⁴Cu, ⁶⁸Ga, ⁷⁴As, ⁸⁹Zr, and ¹²⁴I. Electron emitters in clinical use include ³²P, ⁹⁰Y, ¹³¹I, ¹⁷⁷Lu, and ¹⁹⁸Au, which can also be used for radiotherapy of cancer.

Table 1.

The properties of radionuclides used for CR applications ²³, ²⁴

Radionuclide	Half-life	Decay mode	Energy (keV)	Mean photo yield per disintegration	Clinical application
18F	109.8 min	β+ (97%)	633.5	1.32	PET
64Cu	12.7 h	β+ (17.8%)	653	0.557	PET
68Ga	67.6 min	β+ (88%)	1899	33.9	PET
89Zr	78.4 h	β+ (23%)	2400	2.29	PET
124I	4.17 d	β+ (11.7%)	1534	8.97	PET
32P	14.26 d	β- (100%)	1710	28.1	Radiotherapy
90Y	64 h	β- (99.9%)	2280	47.3	Radiotherapy
131I	8.02 d	β- (89.9%)	606	0.669	Radiotherapy
		γ (81.7%)	364
177Lu	6.73 d	β- (78.6%)	498.3	0.141	Radiotherapy
198Au	2.69 d	β- (98.9%)	960.7	Not reported	Radiotherapy

The radioactivity of each radionuclide strongly influences the intensity of the produced CR. In CR-based imaging, radionuclides with a relatively short half-life, such as ¹⁸F and ⁶⁴Cu, are preferred for building theranostic nanomaterials. In CRIT, radionuclides with a long half-life, such as ⁹⁰Y, ¹³¹I and ¹⁷⁷Lu, are preferred because they can generate CR in the long term for prolonged treatment, and they can generate ionizing radiation that provides radiotherapy.

2.2.2 Megavoltage X-ray radiation

The emission of CR needs the irradiation energy of X-rays to meet the in the high kilovoltage or megavoltage range while only 1% of secondary electrons can be delivered ^{25, 26}. Using X-rays to generate CR and in turn induce luminescence allows imaging at sub-millimeter resolution with nanomolar sensitivity, such as during real-time monitoring during radiotherapy in vivo ^{25, 26}. When X-rays pass through tissues, soft collisions during energy deposition lead to de-excitation of primary or secondary electrons, generating Cerenkov emission (Figure 2A) ⁷. Megavoltage X-ray radiation is an efficient CR source because the X-ray beams contain far more Cerenkov photons per electron than what traditional radionuclides generate (Figure 2B) ⁷. An experimental scanning imaging system to detect CR-excited luminescence is shown in Figure 2C. The system includes an accelerator beam and an intensified charge-coupled device. The CR shows minimal fluence loss as it propagates from deep within tissue to the detector (Figure 2D) ²⁷.

Figure 2.

(A) X-rays induce CR through soft collisions of electrons. (B) Numbers of Cerenkov photons per electron generated from radionuclides or a linear accelerator. Reproduced with permission ⁷. Copyright 2021, Springer-Verlag Wien. (C) Experimental setup for scanning imaging based on Cerenkov-excited luminescence. (D) Fluence of X-ray induced CR from different depths. Reproduced with permission ²⁷. Copyright 2018, Nature Publishing Group.

3. Construction of CR-activated probes

CR can be combined in space and time with activatable probes in two ways. One way is the unbound CR emitter with probes. Megavoltage X-ray radiation or clinical radiopharmaceuticals such as ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) can be used, which requires co-localizing the CR and probes in time and space in vivo. The other way is to use probes already labeled with radionuclides, mainly through (1) extrinsic chelation or (2) intrinsic radiolabelling (Figure 3) ²⁸. The most appropriate radiolabeling chelator depends on the coordination chemistry of the radiometal ion and the surface chemistry of the probe ²⁹. Commonly used chelators include diethylene triamine pentaacetic acid (DTPA), desferrioxamine (Df), 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA), and 1,4,7-triazacyclononane-1,4,7-triacetic acid (NOTA). The chelation should be carefully designed in order to avoid altering the probe's pharmacokinetics. This approach carries risk of radionuclide detachment, leading to nonspecific biodistribution that leads to “noise” in the images. Intrinsic radiolabeling means incorporating radionuclides directly into nanoprobes, through such methods as “hot-plus-cold precursors”, cation exchange, and specific trapping ²⁹. For example, radioactive and non-radioactive precursor components can be mixed during NPs synthesis, or the radiochemical can be generated after nanoformulations through surface elemental exchange or radionuclide deposition. Intrinsic radiolabeling ensures a highly stable label that does not alter the probes' intrinsic behavior.

Figure 3.

Interactions between a CR emitter and activatable probe in the case of (A) an unbound Cerenkov emitter, (B) chelator-bound CR emitter, or (C) intrinsic radiolabelling. Reproduced with permission ²⁸. Copyright 2015, American Chemical Society.

4. Probes for imaging based on CRET

A series of optical probes have been developed to minimize signal attenuation as it passes through tissue, yet tissue autofluorescence still generates sufficient background to severely degrade image quality ³⁰. CR can act as an in situ light source to activate fluorescent probes with large Stokes shifts. In this way, CR overcomes the limited penetration depth of external light, and it avoids tissue autofluorescence. At the same time, the excited fluorescent probe shifts blue-weighted CR to the NIR region or the red visible region, resulting in deeper tissue penetration. Below, we summarize different types of CR-excited fluorescent probes for optical imaging of tumors (Table 2).

Table 2.

Summary of CR-activated optical probes for imaging applications

Activatable probe	CR emitter	Combination way	CRET emission	Imaging subject	Ref.
Qtracker705	64Cu and 18F	Unbound	705nm	Subcutaneous pseudotumors	33
three CdSe/ZnS QDs	131I	Unbound	665nm, 705nm, 800nm	Subcutaneous pseudotumors	34
cRGD-QD605	89Zr-labeled deferoxamine-trastuzumab	Unbound	605nm	HER2/neu-expressing xenografts	35
CdSeTe/CdSe/CdZnS	89Zr	Extrinsic chelation	710nm	sentinel lymph nodes and prostate cancer	36
CdSe/ZnS	64Cu	Intrinsic radiolabelling	526nm, 580nm, 636nm	U87MG tumor	37
CuInS/ZnS	64Cu	Intrinsic radiolabelling	680nm	U87MG tumor	38
PdSe QDs	Megavoltage X-ray radiation	Unbound	1030nm	Subcutaneous pseudotumors	39
Ba0.55Y0.3F2:Eu3+	18F-FDG	Unbound	597nm, 615nm, 692nn	Subcutaneous pseudotumors	43
Y2O2S:Eu3+	68Ga	Unbound	660nm	Subcutaneous pseudotumors	44
Gd2O3:Eu3+	18F-FDG	Unbound	620nm, 700nm	4T1 tumor	45
Eu2O3	18F-FDG	Unbound	620nm, 700nm	Subcutaneous 4T1 tumor and orthotropic HCC tumor	46-49
Eu2O3	89Zr	Intrinsic radiolabelling	620nm, 700nm	Lymph node and CT26-tumor	50
Tb/Eu complexes	18F, 89Zr	Extrinsic chelation	490, 545nm/ 620nm	Phantoms	51
Tb(DO2Apic)-DUPA[Eu(DO2Aphen)-DUPA]	18F-FDG	Unbound	490, 545nm/ 620nm	PC3-PIP tumor (intratumoral treatment, IT)	52
Gold nanoclusters	18F-FDG	Unbound	680-700 nm	subcutaneous breast carcinomas	54
Gold nanoclusters	64Cu	Intrinsic radiolabelling	667nm	U87MG tumor	55
ZnGa2O4:Cr3+	18F-FDG	Unbound	695nm	4T1 tumor	61
Fluorescein sodium	18F-FDG, 11C-CHO	Unbound	520-530nm	subcutaneous 4T1 tumor and orthotropic HCC tumor	64
Fluorescein, rhodamine 6G, rhodamine 101, gresyl violet, cyanine 5, and indocyanine green	90YCl3	Unbound	710nm	EMT6-tumor (IT)	65
Phthalocyanine-pyranine conjugates	18F-FDG	Unbound	700nm	Subcutaneous pseudotumors	67
Naphthofluorescein derivatives	18F	Intrinsic radiolabelling	pH-sensitive	Normal mice	68
aluminum phthalocyanine, platinum II G4 (PtG4)	Megavoltage X-ray radiation	Unbound	780nm	Various tumor	69-72
[Ir(pq)2(bpy)]Cl liposome	18F-FDG	Unbound	570nm	4T1 tumor (IT)	74
Pluronic five different dyes-doped F127 silica nanoparticles	32P-ATP	Unbound	840nm	Phantoms	75

4.1 Quantum dots

Quantum dots (QDs) are nanoscale crystalline clusters (1-10 nm) with high quantum yields, large spectral shifts, and tunable emission colors. They have attracted substantial attention for in vivo fluorescence imaging ^{31, 32}. They absorb broadly throughout the UV-visible range, overlapping with blue-weighted CR. ⁶⁴Cu and ¹⁸F have been used to stimulate Qtracker705 QDs, and high CRET from the enhanced photons in the >590 nm filter was achieved during imaging of phantoms and tissues in vivo ³³. Na¹³¹I has been used as an unbound CR emitter to activate three CdSe/ZnS QDs with emission peaks of 665, 705, and 800 nm, thereby achieving multiplexed optical imaging ³⁴. In an in vivo study, ⁸⁹Zr-labeled deferoxamine-trastuzumab actively targeted BT-474 tumors, and intravenously injected cRGD-QD605 accumulated in the tumor region, allowing tumors to be imaged through secondary Cerenkov-induced fluorescence ³⁵. Nevertheless, that work found that the radionuclides did not completely co-localize with the QDs. To improve co-localization, CdSeTe, CdSe, and CdZnS were modified with ⁸⁹Zr-labeled chelator, and the resulting self-illuminating systems enabled good mapping of sentinel lymph nodes and prostate cancer ³⁶.

To avoid radionuclide detachment, we developed two chelator-free methods to generate ⁶⁴Cu-labeled QDs ^{37, 38}. One is to introduce ⁶⁴Cu into QDs by cation exchange ³⁷, and the other is to use ⁶⁴CuCl₂ as a precursor when synthesizing CuInS/ZnS (Figure 4) ³⁸. These two methods lead to high radiostability. The two types of ⁶⁴Cu-labeled QDs exhibited higher tumor uptake and signal-to-background ratio in CRET imaging in vivo. CR was induced by X-rays to excite PdSe QDs, which emitted in the short-wave infrared region and CRET allowed sensitive depth imaging ³⁹.

Figure 4.

(A) Schematic of chelator-free synthesis of radiolabeled QDs and application to PET and CRET imaging. (B) Photon flux of different samples obtained with different filters. (C) CRET images of U87MG tumor-bearing mice after 6 h of different treatments. (D,E) Total photon fluxes and red filter ratios in the tumor region. Reproduced with permission ³⁸. Copyright 2015, American Chemical Society.

4.2 Lanthanide-based downconversion probes

Lanthanide-based downconversion probes possess good characteristics for optical imaging, such as narrow emission bandwidths, large Stokes shifts, and good photostability ^{40, 41}. They are usually synthesized by doping Tb³⁺, Eu³⁺, and Dy³⁺ to achieve the traditional Stokes luminescence, and they are activated by UV light to generate visible-NIR emission ⁴². Therefore, lanthanide-based downconversion probes may be good CRET mediators to harness CR energy. Below we summarize the two groups of probes, lanthanide nanoparticles (NPs) and lanthanide complexes.

Lanthanide NPs with hard-core structures are usually synthesized by high-temperature chemical methods. Eu³⁺-doped Ba_0.55Y_0.3F₂ nanophosphors with emission peaks at 597, 615, 692 nm have been prepared by a solvent-thermal method. Through CRET, excitation by ¹⁸F-FDG enhanced nanophosphor radiance at 700 nm for phantom and animal imaging ⁴³. Eu³⁺-doped Y₂O₃ NPs have also used as CR mediators for CRET imaging ⁴⁴; CR, rather than γ rays from ⁶⁸Ga, was confirmed to enhance luminescence ⁴⁴. CR from ¹⁸F-FDG as well as β and γ radiation efficiently activated Eu³⁺-doped Gd₂O₃ NPs, enhancing optical imaging and intraoperatively guiding tumor surgery ⁴⁵. To improve the efficiency of photon leap in lanthanide NPs, Eu₂O₃ NPs, which have multiple absorption peaks in the UV range, have been proposed to show more efficient photon leap in lanthanide NPs after activation by ionizing radiation including CR, β, and γ scintillation. This reflects the fact that β interactions significantly contribute to activation of Eu₂O₃ ^46-49. Indeed, this approach allowed radiopharmaceutical-excited fluorescence imaging of phantoms as well as Bcap-37, U87MG, and 4T1-luc2 tumor models in vivo with high signal-to-background ratio. Also, Ultra-small ⁸⁹Zr-labelled Eu₂O₃ NPs have been developed for enhanced Cerenkov imaging, allowing clear imaging of lymph nodes and tumors (Figure 5) ⁵⁰.

Figure 5.

(A, B) TEM image, excitation and emission spectra of ultrasmall Eu₂O₃ NPs. (C, D) CRET images of CT26 tumor-bearing mice and ex vivo organ imaging after 48h's injection of ⁸⁹Zr- Eu₂O₃. Red arrow, tumor area. Reproduced with permission ⁵⁰. Copyright 2021, American Chemical Society.

Other lanthanide complexes have been designed to be excited in situ by radionuclides. ¹⁸F- or ⁸⁹Zr-excited terbium(III) complexes of a macrocyclic polyaminocarboxylate ligand achieved a detection limit of 2.5 nmol in CRET luminescence ⁵¹. Peptide-functionalized Tb(III) and Eu(III) complexes have been constructed for radiopharmaceutical-activated imaging in vivo. Combining these two probes with ¹⁸F-FDG led to emission at 570 and 620 nm, allowing multiplexed optical imaging of tumors ⁵².

4.3 Gold nanoclusters

Gold nanoclusters (AuNCs) have been extensively developed as fluorescent probes for biomedical applications owing to their advantages of high compatibility, strong fluorescence, superior photostability, and excellent water solubility ⁵³. AuNCs are usually biomineralized with naturally functional macromolecules, which provide good colloidal stability and facilitate surface functionalization. AuNCs can fluoresce across a broad region extending from visible to near-infrared regions with a large Stokes shift. ⁶⁴Cu-doped AuNCs with a diameter of 2.54 nm have been synthesized using an approach directed by human serum albumin, leading to AuNCs with an absorption peak at 280 nm and an emission peak at 667 nm (Figure 6A, B) ⁵⁴. During imaging of phantoms and a U87MG tumor model, ⁶⁴Cu-doped AuNCs showed 4.3-fold greater optical intensity than free ⁶⁴CuCl₂ with an optical filter of 695-770 nm but lower intensity below 510 nm, demonstrating CRET from ⁶⁴Cu to AuNCs (Figure 6C, D, E).

Figure 6.

(A) TEM image of Cu-doped AuNCs (1%Cu). (B) The absorption and emission spectrum of Cu-doped AuNCs (1% Cu). (C) Intensity changes of ⁶⁴CuCl₂ and ⁶⁴Cu-doped AuNCs at different optical filter sets. (D, E) Representative CRET images and photon flux of ⁶⁴Cu-doped AuNCs on the U87MG tumor model. Reproduced with permission ⁵⁴. Copyright 2014, Elsevier.

To verify the mechanism of efficient radioisotope energy transfer by AuNCs, they were excited using ¹⁸F-FDG, ⁹⁰Y, or ^99mTc (γ-emitter). Activation of ^99mTc did not induce optical signal, while activation of ¹⁸F-FDG and ⁹⁰Y enhanced optical emission ⁵⁵. The fact that the signal enhancement was not proportional to CR implies that AuNCs were excited by CR and direct Coulombic interaction ⁵⁵.

4.4 PLNPs

PLNPs maintain their luminescence after excitation has stopped, so they have been exploited for autofluorescence-free optical imaging with different excitation sources ^56-58. Cr³⁺-doped zinc gallates (ZGCs) are NIR-emitting PLNPs commonly used for tumor diagnosis, with three main excitation bands with peaks at 260, 465, and 570 nm ^{59, 60}. For the first time, we report that both CR and γ radiation from ¹⁸F-FDG can efficiently activate ZGCs to emit in the NIR region and persistently release photons for long-term imaging (Figure 7A) ⁶¹. In fact, ¹⁸F-FDG excited ZGCs in 4T1 tumors in mice, maintaining luminescence for longer than 3 h (Figure 7B). Multiple injections of ¹⁸F-FDG also allowed for long-term real-time observation of tumor status, which may facilitate image-guided surgery in the future.

Figure 7.

(A) An illustration of in vivo ¹⁸F-FDG activated persistent luminescence imaging of ZGCs. (B) Representative luminescence images of 4T1 tumor-bearing mice after administration of only 200 μCi ¹⁸F-FDG or 200 μg ZGCs injection before 24 h and following with 200 µCi ¹⁸F-FDG. Reproduced with permission ⁶¹. Copyright 2020, Wiley-VCH.

4.5 Organic dye or associated NPs

Inorganic fluorescent probes have excellent optical properties and can take full advantage of CR to undergo highly efficient CRET with large Stokes shifts. However, their potential toxicity and non-biodegradability have prevented their extensive development for clinical use. Organic compounds, in contrast, show good biodegradability and low toxicity ⁶². For example, fluorescein sodium is clinically used to image retinal blood vessels; it absorbs strongly at 465-490 nm and emits at 520-530 nm ^{63, 64}. Despite its short Stokes shift, fluorescein sodium can be activated by ¹⁸F-FDG or ¹¹C-CHO to show efficient CRET, as demonstrated in a subcutaneous 4T1 tumor model and orthotopic hepatocellular carcinoma tumor model (Figure 8A, B,C) ⁶⁴. In that study, tumors were precisely and completely resected through intraoperative CRET and confocal laser endomicroscopy imaging ⁶⁴.

Figure 8.

(A) CR and CRET imaging of subcutaneous 4T1 tumor model after injection of 300µCi ¹⁸F-FDG (CRET group was treated with different doses of fluorescein). (B) The signal intensities and TNR of CLI and CRET. (C) Intraoperative CRI and CRET imaging of orthotropic HCC tumor model, and the process of image-guided surgery. Reproduced with permission ⁶⁴. Copyright 2019, Wiley-VCH. (D, E) pO2 images and the corresponding pO2 changes during each day of radiation (6MV X-ray beam with 5 Gy/fraction), PtG4 was iv injected 24h before the first treatment. Reproduced with permission ⁷³. Copyright 2019, Nature Publishing Group.

Various fluorophores have been used to convert CR to NIR emission. Excitation of a mixture of fluorophores (fluorescein, rhodamine 6G, rhodamine 101, gresyl violet, cyanine 5, and indocyanine green) with ⁹⁰YCl₃ led to a CRET process, followed by several rounds of Förster resonance energy transfer (FRET), culminating in emission at 710 nm ^{65, 66}. All these fluorophores overlapped well in spectra, emitting twice as many photons as the radionuclide on its own. Featuring larger Stokes shifts, water-soluble phthalocyanine-pyranine conjugates have been developed with absorption at 250-450 nm and emission at 700 nm. They can be activated by ¹⁸F-FDG, leading to strong CRET using a filter of 710 nm as well as a photodynamic effect ⁶⁷. Interestingly, ¹⁸F-labeled naphthofluorescein derivatives have been developed that show selective bandwidth quenching under alkaline conditions but full-spectrum CR under acidic conditions ⁶⁸.

Comparison of different CR-excited fluorophores induced by X-rays ^69-72 suggests that aluminum phthalocyanine may be better for molecular luminescence ⁶⁹, while platinum II G4 (PtG4), with absorption peaks at 435 and 624 nm and an emission peak at 780 nm, may be better for in vivo sensing ⁷⁰. PtG4 and X-ray induced CR have been used to detect spatial distribution of oxygen in tumors based on phosphorescence lifetimes ^{27, 73}. This approach revealed much higher partial pressure of oxygen (pO₂) in MDA-MB-231 tumors than in FaDu tumors in animal models, suggesting that MDA-MB-231 tumors are more susceptible to radiotherapy (Figure 8D, E).

Organic dye-based NPs can enhance the accumulation of contrast agents in tumors while preserving the optical property of those agents. [Ir(pq)₂(bpy)]Cl has been encapsulated into liposomes to formulate Ir@liposome, then irradiated with ¹⁸F-FDG, leading to imaging of deep tissue with a high signal-to-noise ratio ⁷⁴. Pluronic F127 silica NPs doped with five dyes have been constructed to absorb strongly in the visible range, leading to the efficient shift of CR towards NIR emission with a quantum yield of 0.12 ⁷⁵. The highly efficient energy transfer allowed the NPs to be detected from muscle as thick as 1.0 cm.

5. Probes for CR-induced therapy

CRIT refers to therapeutic effects that occur when CR-activated probes directly or indirectly produce reactive molecules that kill tumor cells. The efficiency of CRIT depends on delivery of adequate CR to target tissue, and proper stimulation of the probe. Therefore, the CR emitter and activatable probe must be selected carefully. Of course, effective tumor suppression also requires co-localization of emitter and probe. In addition, toxic effects that arise when the two inevitably co-localize in normal tissues must also be considered.

Below we provide an overview of probe-based CRIT, including CR-induced PDT, CR-induced photoimmunotherapy, CR-triggered drug release, and combination therapy (Table 3).

Table 3.

Summary of CR-activated probes for therapeutic applications

Activatable probe	CR emitter	Combination way	CRIT	Treated subject	Ref.
TiO2-Tf-Tc	18F-FDG	Unbound	PDT	HT1080 tumor	79
TiO2-Tf	89Zr	Intrinsic radiolabelling	PDT	Multiple myeloma	80
Titanocene-loaded nanoparticles	18F-FDG	Unbound	PDT	Metastatic breast cancer and disseminated multiple myeloma	81
TiO2	68Ga-BSA	Unbound	PDT	4T1 tumor (IT)	82
Chlorin e6 loaded hollow mesoporous silica NPs	89Zr	Intrinsic radiolabelling	PDT	4T1 tumor (IT)	83
Porphyrin surface-modified magnetic NPs	89Zr	Intrinsic radiolabelling	PDT	4T1 tumor	84
Df-PPN	89Zr-labeled Df-PPN	Unbound	PDT	4T1 tumor	85
IRDye700DX	18F-FDG	Unbound	Photoimmunotherapy	A431-luc tumor	88
Folate-modified DOX/nanomicelles	Megavoltage X-ray radiation	Unbound	Chemotherapy	Hela tumor	92
ZGCs-ZnPcC4	131I	Extrinsic chelation	PDT-radiotherapy	4T1 tumor (IT)	95
sPS NPs	131I	Extrinsic chelation	PDT-radiotherapy	4T1 tumor-bearing mice and VX2 liver tumor-bearing rabbit	96
EM@ALA	131I	Extrinsic chelation	PDT-radiotherapy	4T1 tumor-bearing mice	97
Titanium-Oxo nanoclusters	18F-FDG	Unbound	Photo/chemodynamic therapy	HepG2 subcutaneous tumor	98
Single-layer 2D nanosheets	32P	Intrinsic radiolabelling	Radioisotope-immunotherapy	4T1 tumor (IT)	100

5.1. CR-induced PDT

PDT has been widely used to treat many solid tumors because of its non-invasiveness and dual selectivity ⁷⁶. In PDT, external light activates photosensitizers to generate reactive oxygen species (ROS), which damage cancer cells. A major drawback is that PDT requires direct irradiation of tissue with visible or even UV light in order to stimulate photosensitizers. This limits penetration depth, making the approach ineffective against large or deep-seated tumors ^{6, 77}. Blue-weighted CR can overcome this problem by acting as an in situ light source to activate photosensitizers continuously ²¹.

For example, TiO₂ NPs act as photocatalysts that can be regenerated to efficiently absorb UV light at 275-390 nm and to produce cytotoxic superoxide radicals and hydroxyl ⁷⁸. A TiO₂-based CRIT-nanoplatform has been achieved for deep PDT, in which ¹⁸F-FDG excites TiO₂ to continuously generate OH^- by electron-hole pair generation (Figure 9A) ⁷⁹. Of course, it is worth noting that β and γ scintillation produced from radionuclides also play a role in TiO₂'s activation ⁴⁹. Intratumoral administration of TiO₂-PEG and ⁶⁴Cu led to remarkable tumor regression within 3 days and complete inhibition within 30 days. In another approach, apo-transferrin, a ligand for overexpressed Tf receptors, and titanocene, a photogenerator of peroxyl radicals, were integrated into TiO₂ (TiO₂-Tf-Tc) to enhance CRIT with an intravenous platform (Figure 9B,C,E). This system accumulated in tumors to a greater extent than TiO₂-PEG within 24 h, based on ex vivo fluorescence imaging (Figure 9D). Intravenous injection of TiO₂-Tf-Tc followed by ¹⁸F-FDG at 24 h later significantly inhibited tumor growth (Figure 9E,F). This work paves the way to treating deep-seated tumors using CR-induced PDT.

Figure 9.

(A, B) An illumination of CR-induced PDT during the CR-excited TiO₂ NPs and Tc. (C) TEM image of TiO₂-Tf-Tc. (D) In vivo biodistribution of TiO₂-PEG and TiO₂-AlexaTf at 24h postinjection in HT1080 tumor-bearing mice. (E) Cell-viability assays on HT1080 cells with various treatments. (F) Tumor volumes of HT1080 tumor-bearing mice with various treatments. Reproduced with permission ⁷⁹. Copyright 2015, Nature Publishing Group.

To compensate for low uptake of ¹⁸F-FDG in bone, ⁸⁹Zr-labeled TiO₂-Tf has been developed to target bone marrow with high selectivity, leading to accumulation of 70% of injected radioactivity there. This accumulation allowed continuous ROS production that completely eliminated tumors from a multiple myeloma model and that doubled survival time ⁸⁰. Tf-mediated transport of titanocene into various types of tumor has been improved by using high-affinity VLA-4-bonded LLP2A conjugated phospholipid micelles or human serum albumin NPs ⁸¹. Activation of the titanocene in situ by intravenously injected ¹⁸F-FDG inhibited breast cancer metastasis and growth of disseminated multiple myeloma. ⁶⁸Ga-bovine serum albumin has been used instead of ¹⁸F-FDG to improve PDT efficiency with CR-excited TiO₂, leading to strong tumor inhibition in vitro and in vivo ⁸².

CR-induced PDT has been achieved using the photosensitizer dyes, such as chlorin e6 and tetrakis(4-carboxyphenyl)porphyrin (TCPP) ^83-85. Hollow mesoporous silica NPs were loaded with ⁸⁹Zr-labeled chlorin e6, and the ⁸⁹Zr provided CR to activate chlorin e6, which generated tumor-inhibiting ROS ⁸³. Given that intravenously injected theranostic agents do not strongly accumulate in tumors, ⁸⁹Zr-labeled magnetic NPs with a modified porphyrin surface were constructed to enable accumulation of the NPs in tumors (15% ID/g) after application of a magnetic field; the accumulation was five times higher than without the magnetic field ⁸⁴. TCPP-loaded magnetic NPs significantly inhibited tumor growth through CR-induced PDT during two weeks. While promising, this strategy led to substantial NPs accumulation also in healthy organs, suggesting the potential for adverse effects. To enhance targeting in CRIT, a “missile-detonation” strategy has been developed: high-dose p-SCN-Bn-deferoxamine-porphyrin-PEG nanocomplex (Df-PPN) was administered as a CR energy receiver, and it passively targeted tumors; then, at the optimal time, low-dose ⁸⁹Zr-labeled Df-PPN was injected as a CR emitter (Figure 10A) ⁸⁵. This strategy effectively mitigated toxic side effects of CRIT due to the relatively low accumulation of PPN in normal tissues. The nanocomplex persisted in the tumor in the long term, based on CR-excited fluorescence imaging (Figure 10B,C,D), and it drove continuous ROS production that significantly suppressed tumor growth (Figure 10E).

Figure 10.

(A) Schematic of the “missile-detonation” strategy. (B,C) Representative CR-excited optical images and the corresponding photon fluxes in a 4T1 tumor model on different days after treatment. (D) Biodistribution of ⁸⁹Zr-Df-PPNs at 14 days after injection, based on measurement of radioactivity in each organ. (E) Relative volume of 4T1 tumors in mice after various treatments. Reproduced with permission ⁸⁵. Copyright 2019, Wiley-VCH.

5.2 CR-induced photoimmunotherapy

NIR photoimmunotherapy is a newly targeted cancer treatment that uses antibody-photoabsorbers (IRDye700DX) to actively bind to cancer cells and then induce the selective immunogenic cell death under NIR light irradiation, resulting in the activation of anti-cancer immune system locally in the tumor microenvironment ^{86, 87}. However, NIR laser light cannot penetrate deep in tissue, but the CR emitter ¹⁸F-FDG can be delivered deep into tissue to deliver radiation at the absorbance peak of IRDye700DX at 350 nm, allowing photoimmunotherapy of deep-seated tumors ⁸⁸. The CR-induced photoimmunotherapy in that study inhibited tumor growth to some degree, but the effect was not consistent during the observation period. This may be because the lower CR energy stimulated a weak immune response. Megavoltage X-ray radiation may be a better CR source for activation of IRDye700DX.

5.3 CR-triggered drug release

Although chemotherapeutics can be effective against cancer, delivering them to target tissue without causing toxicity in normal tissue is challenging ⁸⁹. Therefore, drug delivery systems have been developed to release drugs at specific sites in response to certain stimuli, especially light ^{89, 90}. However, the limited penetration of light through tissue means that external light sources cannot efficiently trigger drug release in deep tissue. Using CR as in situ light can overcome this shortcoming. UV-responsive drug delivery systems may be most appropriate for CR-activated drug release. For example, phenacyl bis-azide crosslinkers make dextran-based hydrogels photosensitive, and the co-loaded doxorubicin (DOX), BSA, and IgG are efficiently released after UV light irradiation ⁹¹. In another approach, DOX has been caged using a photocleavable o-nitrobenzyl ester derivative and then encapsulated into nanomicelles for CR-triggered drug release and chemotherapy (Figure 11) ⁹². Irradiation with X-rays induced CR, which converted the hydrophobic DOC into hydrophilic DOX, which was rapidly released and killed cancer cells. This strategy may pave the way to a new chemoradiotherapy with minimal systemic side effects.

Figure 11.

The CR-responsive folate-modified DOX/nanomicelles for folate cellular targeting and nuclear localization toxicity. Reproduced with permission ⁹². Copyright 2020, American Chemical Society.

5.4 Combination therapy

The low efficiency of CR limits its therapeutic efficacy ⁹³, which combination therapy may improve ⁹⁴. We have developed a ZGC nanoplatform conjugated to ¹³¹I-labeled ZnPc(COOH)₄ (¹³¹I-ZGCs-ZnPcC4) that can support both radiotherapy as well as continuous, radiation-induced PDT through energy transfer from the ionizing radiation to ZGCs and then to ZnPc(COOH)₄ ⁹⁵. In another approach, we engineered pyropheophorbide-a containing PEG and a diisopropylamino group with an ¹³¹I-labeled tyrosine, which self-assembled into pH-sensitive NPs ⁹⁶. These NPs exhibited minimal phototoxicity in normal tissue because of the quenched photodynamic effect, and under acidic conditions, they disassembled to promote the generation of ROS to achieve PDT and radiotherapy against deep-seated tumors (Figure 12A). Their strong accumulation in tumors makes the NPs effective against 4T1 tumors in mice and VX2 liver tumors in rabbits (Figure 12B). Thus, this strategy shows promise for deep tumor therapy. Combination therapy also has been developed in which the precursor of protoporphyrin IX called 5-aminolevulinic acid (ALA) and ¹³¹I are co-loaded into biomimetic exosomes. In the tumor microenvironment, where mitochondria are abundant, the ALA is converted to protoporphyrin IX, minimizing CRIT side effects on normal tissues ⁹⁷.

Figure 12.

(A) Schematic of pH-activatable CRET-PDT. (B) Representative imaging of VX2 liver tumor-bearing rabbits using PET/CT before and at seven days after the indicated treatments. Reproduced with permission ⁹⁶. Copyright 2021, Wiley-VCH. (C) Schematic of ³²P-triggered NO release for enhanced radioisotope-immunotherapy. (D) Concentration of NO released from ZnNO(³²P) nanosheets at various time points. (E) Growth curves of CT26 tumors in mice after various treatments. Reproduced with permission ⁹⁹. Copyright 2019, Elsevier.

Functionalized inorganic materials can also achieve multiple therapeutic effects. Titanium-oxo nanoclusters have been reported for CR-induced photo/chemodynamic therapy of tumors ⁹⁸. CR can drive efficient production of hydroxyl radical (·OH) from titanium-oxo nanoclusters due to the enhanced separation of hole (h⁺)-electron (e^-) pairs. The reaction of h⁺ and H₂O provides type I PDT, while the transferred e^- augmented Ti³⁺ provides chemodynamic therapy. Besides, ³²P-labeled single-layer 2D nanosheets have been constructed by mixing Zn²⁺ ions and sodium nitroprusside [Na₂Fe(CN)₅NO]. CR from ³²P can persistently stimulate nanosheets to release NO, which can modulate the tumor microenvironment to induce anti-tumor immunotherapy and improve the efficacy of radionuclide therapy (Figure 12C,D) ^99-100. This nanosystem has been extended to include immune-checkpoint blockade therapy against anti-programmed cell death protein 1 in order to achieve sustainably strong immune responses that effectively suppress tumor growth (Figure 12E).

6. Conclusion and perspectives

CR-activated probes take advantage of the ability of CR to pass through tissue without attenuation, enabling detection and treatment of deep-seated tumors. This review has focused on probes that can be activated by CR and that provide theranostics via CRET or production of reactive species/drug release. While these activatable probes can also be excited by ionizing radiation from radiotracers, especially lanthanide-based downconversion probes, the present review demonstrates the promise of approaches based on activation by CR. Further development of CRET will depend on the design of biocompatible optical probes with large Stokes shifts in the “deep NIR” to NIR-II regions. Future work should aim to overcome several limitations of CRIT, such as inefficient delivery of CR to therapeutic drugs in tumors, short duration of CR excitation, and damage to normal tissue during CRIT. In any case, suitable CR emitters and CR-responsive photosensitizers or drugs must always be chosen carefully in order to maximize tumor targeting and tumor cytotoxicity. Radionuclides with long half-lives are particularly promising as CR emitters, because they continuously generate CR to activate CR-responsive photosensitizers or drugs while also providing radiotherapy. Although CR remains in early stages of preclinical development, we believe that its use as an in situ light source will broaden the possibilities for imaging-based theranostics of cancer.