Electron Transfer based Ultra-bright Organic Afterglow Nanoprobe for Accurate Molecular Imaging in Mice

Afterglow luminescence can greatly improve the signal-to-background ratio (SBR) of molecule imaging in living animal owing to the no need of real-time light excitation. However, the relatively low luminescence of afterglow nanoprobe and attenuation of maximum intensity (afterglow photobleaching) usually lead to the insucient sensitivity and the inaccurate quantication for repeated molecular imaging. Furthermore, the requirement of high power of light excitation (up to 1 W/cm 2 ) may result in the inevitable phototoxicity, and the long acquisition time (up to 1 min) make it dicult to detect the rapid biological events. Herein, we design electron-rich trianthracene derivatives (TA)-based organic afterglow nanoparticles (TA-NPs) for high-sensitive, safe, lossless and longitudinal molecular imaging. Notably, a great enhancement of afterglow luminescence performance over the previous reported afterglow nanoparticles is achieved though electron transfer engineering (Table 1): Specically, TA-NPs can be excited by room light with ultra-low power (58 µW/cm 2 ) and with ultra-short acquisition time (0.01 s). The luminescent intensity of TA-NPs is ~ 500-fold of commonly used organic MEHPPV-based nanoparticles. Negligible afterglow photobleaching in mice is observed even after re-excitation for more than 15 cycles. Such ultra-bright afterglow enables the deep-tissue imaging (up to 6.0 cm) and the ultra-fast afterglow imaging of freely-moving mice in waken state. Moreover, TA-NPs can dynamically and accurately visualize subcutaneous tumor, orthotopic glioma and distinguish the plaque in carotid atherosclerosis. Finally, we develop an afterglow nanoprobe (TA-BHQ), activated only in the presence of Granzyme B, for tracking the time-sensitive Granzyme B activity as a direct way to monitor immunotherapeutic responses.


Introduction
Molecular imaging technology enables the dynamical, quantitative visualization of speci c biochemical activity without trauma in vivo at cellular or molecular level, so as to study biological process, diagnose disease or monitor therapy [1][2][3][4][5][6][7] . High sensitivity together with novel methods to reliably quantitate imaging signal are strongly needed for accurate molecular imaging, which should allow not only the detection of molecular events that occur either with a low mass of substance or at low concentration, but also the repeated studies in the same animal to collect longitudinal data and reduce the number of animals and cost [8][9][10][11] . However, there are several intrinsic limitations for current imaging modalities, such as poor sensitivity (for MRI), ionizing radiation (for PET and CT), or strong background signals (for uorescence imaging) 12 . Interestingly, afterglow luminescence, called long-persistent luminescence, is an intrinsic luminescent process that occurs after the cessation of light excitation. Owing to no requirement of realtime light excitation, afterglow luminescence can eliminate auto uorescence of biological tissues, which enables to greatly improve signal-to-background ratio (SBR), in contrast to uorescent imaging [13][14][15][16][17] .
Thereby, it holds great promise for various biomedical applications in vivo such as imaging orthotopic liver tumors or tiny peritoneal metastatic tumors, mapping lymph node, early detecting drug-induced hepatotoxicity, or afterglow imaging-guided surgery or therapy in living animals [18][19][20][21] .
To date, a few inorganic nanoparticles (e.g., ZnGa 1.995 Cr 0.005 O 4 , lanthanide-doped NaYbF 4 upconversion nanoparticles) have been developed to produce afterglow luminescence 15,16 . However, those may be hampered by systemic toxicity concerns related to the potential leakage of heavy metal ions. Compared with inorganic nanoparticles, organic afterglow nanoparticles (e.g., poly[2-methoxy-5-(2-ethylhexyloxy)-1,4-phenylenevinylene] (MEHPPV) or its analogs) are more suitable for biomedical applications in terms of biocompatibility and biodegradability 13,[22][23][24][25][26] . Unfortunately, current organic afterglow nanoprobes are usually limited by the poor "photo-afterglow conversion performance", the relatively low afterglow luminescent intensity, the requirement of high power density of laser excitation, the long acquisition time, and even the attenuation in maximum intensity (termed "afterglow photobleaching") for repeated imaging. Speci cally, for the typical organic afterglow systems (MEHPPV-based nanoparticles), an 808 nm of laser with power density of as high as 1 W/cm 2 was required to excite adequate afterglow luminescence, which is ∼3 times higher than the conservative limit (~ 0.33 W/cm 2 , the American National Standard for the Safe Use of Lasers) for 808 nm laser power density setting for human skin exposure 13,14,27 . The afterglow imaging conducted at such excessive illumination may lead to phototoxicity and thereby tissue injury. Moreover, the low luminescence intensity (~ 0.01 % -1 % of uorescent intensity) and long acquisition time (up to 1 min) may result in the low level of key biomolecules and rapid biological events to be missed during lengthy and insensitivity imaging 28 . Furthermore, the serious attenuation in maximum intensity was observed from MEHPPV-based nanoparticles (60 % of decrease after 10-cycle irradiation) 13 . Such "afterglow photobleaching" due to light-induced degradation of luminophore (e.g., phenylenevinylene) can render the failure of accurate quanti cation for repeated imaging during hours-long experiment. Thence, efforts of afterglow imaging should focus on not only the design of sensitive organic afterglow molecules with high "photo-afterglow conversion effect" that support ultra-bright afterglow luminescence and safe excitation power, but also the development of negligible photobleaching of afterglow allowing for lossless and accurate imaging over repeated irradiation. To date, technical hurdles continue to hamper accurate, safe, high-sensitivity, and longitudinal afterglow imaging in vivo, the reason is that the ultra-bright afterglow luminescent, ultra-low excitation power, as well as well-recharged nanoparticles remains unexplored.
Herein, we introduce the electron-rich anthracene into the afterglow molecule design and synthesize trianthracene derivatives (TA)-based afterglow nanoparticles (TA-NPs) (Fig. 1a) 30 , herein we exploit the electron transfer mechanism to boost the production e cacy of intermediates and thereby afterglow intensity. We introduce the electron-rich anthracene into the afterglow molecule design to synthesize trianthracene derivatives (TA), which was substituted with alkoxy groups at orthoand bay-positions to enhance the electro-rich properties. Its isomer dianthraphenanthrene (DAP) was also synthesized as a comparison ( Fig. 1a In order to construct biocompatible afterglow luminescent nanoparticles, we transformed TA or DAP molecules into water-soluble nanoparticles through one-step nanoprecipitation in the presence of dual polymers (poly (styrene-co-maleic anhydride) (PSMA) and 1,2-distearoyl-sn-glycero-3phosphoethanolamine-N-[methoxy(poly(ethylene glycol))] (DSPE-PEG)) ( Fig. 1a, Supplementary Table 4).
In particular, transmission electron microscope (TEM) image further revealed that TA-based nanoparticles (TA-NPs) appeared as mono-disperse nanoparticles, with a mean hydrodynamic size of ~ 40 nm (Fig. 1b). The hydrodynamic diameters of the nanoparticles measured by dynamic light scattering (DLS) were about 60 nm (Fig. 1b, Supplementary Fig. 1). Figure 1c and 1d unveiled the characteristic absorption bands and the uorescence emission spectra of TA-NPs and DAP-based nanoparticles (DAP-NPs). Subsequently, the afterglow luminescence of TA-NPs and DAP-NPs was recorded with the IVIS Lumina XR imaging system under bioluminescence (without excitation) modes to identify the capability of afterglow luminescence. Notably, TA-NPs showed stronger luminescence than DAP-NPs after cessation of white light irradiation (Fig. 1e, f). Besides, the afterglow spectra of TA-NPs were similar to its uorescence spectra from 600-750 nm (Fig. 1g).

Mechanism of afterglow luminescence
To study the underlying mechanism governing afterglow luminescence through Type I pathway, we Interestingly, TA showed lower rst oxidation potential (-0.15 V) compared with the isomer DAP (-0.12 V), according to the cyclic voltammogram measurement, indicating that TA was easier to give electrons to form free radical cation (Fig. 2c, d). Next, SOD was added into TA-NPs solutions, to further validate that the enhanced afterglow signal was indeed related to the generated O 2 •-, as expected, the afterglow luminescence showed 26.8 % of decrease (Fig. 2e). Furthermore, we have measured 1 O 2 generation from TA-NPs via the uorescence intensities of SOSG at 535 nm and found that a considerable amount of 1 O 2 was produced during irradiation of TA-NPs (Fig. 2f). Subsequently, the incubation of 1 O 2 (Na 2 MoO 4 + H 2 O 2 ) with TA-NPs resulted in higher luminescence intensity of TA-NPs, than that of PBS, H 2 O 2 or Na 2 MoO 4 alone (Fig. 2g), which con rmed that 1 O 2 was also responsible for afterglow luminescence via Type II pathway. Figure 2h showed that the afterglow of TA-NPs in O 2 -saturated solutions was 1.9fold that in N 2 -saturated solutions, highlighting the key role of O 2 in afterglow luminescence.
To further understand the mechanism on EPOs processes, density functional theory (DFT) calculations were applied to predict the relative energies of EPOs intermediates brie y, and gave one of possible reaction paths (Fig. 2i, j). The reaction process of O 2 •or 1 O 2 cycloaddition in anthracene units to anthracene-dioxetane intermediate was determined, based on the corresponding intermediate EPOs found, con rmed by MALDI-TOF MS measurement ( Fig. 2i and Supplementary Fig. 3). When exposed to light and air, TA molecules undergo a serial of cycloaddition with O 2 or VI). Given that the energy of EPOs (II and IV) was slightly lower than that of EPOs (III and V) (Fig. 2j), it might be prone to generate EPOs (II and IV) during gradual cycloaddition reaction. After long-term light irradiation of TA, the typical UV-vis absorption bonds (500 -600 nm) and typical uorescent emission peak at 650 nm were signi cantly decreased, indicating the decomposition of chemical structures for TA molecules ( Supplementary Fig. 4). Besides, the obvious carbonyl characteristic peak of oxidized TA fragments was detected in Fourier transform infrared spectroscopy (FTIR) at 1,735 cm -1 , which con rmed the oxidation of TA into various aromatic-carbonyl compounds (Fig. 2k).
Based on the above results, we created a possible picture of underlying mechanism for TA-NPs afterglow luminescence. In the photooxidation, TA-NPs underwent two different photochemical reaction pathways to generate EPOs and thereby afterglow luminescence: electron transfer pathway (Type I) and energy transfer pathway (Type II) 30 , which may increase the path of Type I afterglow reaction (Fig. 2c). We also calculated the process of addition of 1 O 2 to DAP using DFT calculations, and found that the energy of EPOs (II-VI) from TA were lower than those corresponding EPOs from DAP, suggesting the larger possibility of forming EPOs for TA ( Supplementary Fig. 6).
Optimization and performance study of afterglow luminescence We subsequently investigated the effect of amphiphilic polymers on TA-NPs afterglow luminescence, and found that the dual polymers (PSMA + DSPE-PEG) co-coating resulted in higher afterglow luminescence, in contrast to other kinds of polymer coating groups (e.g., pluronic-F127 or PSMA only) ( Supplementary  Fig. 7). Therefore, the polymers (PSMA + DSPE-PEG) co-coated nanoparticles were used for the subsequent in vitro/vivo experiments.
Currently, the semiconducting polymer (MEHPPV) is the most commonly used organic afterglow material, reported in the previous articles 13,14,19 . We further compared the afterglow performance of TA-NPs with MEHPPV-based nanoparticles (MEHPPV-NPs) and other organic molecules-based nanoparticles such as anthracene, perylene, naphthalene, respectively (Fig. 3a, Supplementary Table 5). Then, we characterized those nanoparticles including UV-vis absorption spectra, uorescence emission spectra and DLS sizes ( Supplementary Figs. 8 -11). We found that TA-NPs showed extremely higher afterglow intensity than that of MEHPPV-, anthracene-, perylene-, and naphthalene-based nanoparticles.
Through the continuous acquisition of afterglow images, the long-lasting afterglow luminescence could persist for over 60 mins for TA-NPs after light cessation, with a prolonged half-life of ~ 16 min at biologically relevant conditions (pH = 7.4 at 37 °C; Fig. 3d), compared with half-life of ~ 6 min for MEHPPV-NPs. Moreover, as the power density of light irradiation increasing from 1.4 to 15 mW/cm 2 , TA-NPs showed gradually brighter afterglow luminescence (Fig. 3f, Supplementary Fig. 13). The afterglow intensities of TA-NPs were continuously enhanced with the irradiation time increasing, and reached the maximum at 8 s of irradiation (15 mW/cm 2 ) (Fig. 3g, Supplementary Fig. 14). Furthermore, afterglow luminescence of TA-NPs became brighter as the concentration increasing, and showed a linear correlation between afterglow intensities and concentrations (0 -40 μg/mL) (Fig. 3h, Supplementary Fig.  15). As a result, under various power density, irradiation time or concentration, TA-NPs showed extremely higher afterglow intensities than that of MEHPPV-NPs (Fig. 3f, g, h and Supplementary Figs. 13 -15). Notably, we found TA-NPs could be effectively excited by the white light source with ultra-low power density (room light, 58 μW/cm 2 ) and was able to exhibit bright afterglow (Fig. 3i). This ultra-bright afterglow luminescence of TA-NPs facilitated to shorten the signal acquisition time. Especially, when the acquisition time was reduced to 0.01s, TA-NPs could still produce strong afterglow luminescence. However, no obvious afterglow signals were detected from MEHPPV-NPs under the same conditions (Fig.   3i, Supplementary Fig. 16). Notably, the ultra-low power density (58 μW/cm 2 ) that is far below the intensity of direct sunshine (0.13 W/cm 2 ) can provide safe excitation condition and thereby avoid photoinduced tissue damage.
Moreover, the ultra-low power of excitation light leads to the negligible chemical decomposition of afterglow molecules. Thereby, the afterglow signal could be recharged by more than 15cycle light irradiation (30 mW/cm 2 , 3s) or 30-cycle light irradiation (6.6 mW/cm 2 , 1s), with no obvious decay of maximum afterglow intensity (Fig. 3j, Supplementary Fig. 17). Such negligible afterglow photobleaching may be attributed to the e cient photo-afterglow conversion and ultra-low power density of excitation light, which enables to guarantee the accurate quanti cation of afterglow signals during hours-long and repeated molecular imaging.
Prior to afterglow luminescence imaging in vivo, we studied the tissue penetration depth of afterglow luminescence emitted from TA-NPs. Although afterglow and uorescence signals were declined with the increase of tissues thickness (Fig. 3k, Supplementary Fig. 18), the afterglow luminescence exerted signi cantly higher signal-to-background ratio (SBR) and tissue penetration (up to 6.0 cm), compared with that of uorescence, owing to the no auto uorescence from tissues during afterglow imaging.
Besides, the tissue penetration and SBR of afterglow luminescence for TA-NPs was much higher than that for MEHPPV-NPs ( Supplementary Fig. 19).  Fig. 20a). Secondly, for Type II afterglow, TA-NPs display the stronger ability to generate 1 O 2 , compared to MEHPPV-NPs ( Supplementary Fig. 20b and c), which could promote the subsequent generation of EPOs via Type II pathway. Thus, TA-NPs emitted an extremely high afterglow luminescence than MEHPPV-NPs, through both electron transfer path (Type I afterglow) and the enhanced energy transfer path (Type II afterglow) (Fig. 3e).

Afterglow imaging of tumors and plaque-bearing carotid
The features of TA-NPs, such as extremely high afterglow luminescence and prolonged luminescence lifetime, offer it noteworthy advantages in designing afterglow molecular imaging in vivo. Prior to afterglow luminescence imaging using TA-NPs in vivo, the stability and biocompatibility of TA-NPs were examined. The hydrodynamic size and luminescence intensity of TA-NPs remained highly stable after incubating those in different buffer solution (cell culture medium, PBS, water) for various time Inspired by the excellent afterglow performance of TA-NPs, we applied TA-NPs for afterglow luminescence imaging of subcutaneous CT-26 tumor-bearing mice after intratumoral injection, through IVIS Lumina XR imaging system (Fig. 4a). In dark condition, there was no afterglow signal at the tumor site at the beginning. Then, those mice were taken out of dark box and accepted the irradiation of smart mobile phone (1.0 mW/cm 2 ) or electric torch (4.0 mW/cm 2 ). The afterglow luminescence of mice was collected immediately after irradiation, and the tumor areas showed strong afterglow luminescence.
Importantly, mice could be even excited by room light with ultra-low power density (58 μW/cm 2 ) (Fig. 4a). Moreover, the afterglow luminescence of tumor could well be recharged by multiple irradiation (more than 3 cycles) with electric torch, mobile phones or room light irradiation (Fig. 4b, Supplementary Fig. 23), as well as showed no obvious attenuation in maximum intensity during 15-cycle light irradiation (30 mW/cm 2 ) (Fig. 4c, Supplementary Fig. 23). The above data demonstrated TA-NPs in mice could be e ciently re-excited via low power density of portable light sources and displayed no notable afterglow photobleaching.
Highly e cient afterglow luminescence of TA-NPs can signi cantly reduce the required dose for imaging in vivo. Under white light at low-power density of 6.6 mW/cm 2 , TA-NPs (20×) showed afterglow signal as large as ~ 6×10 7 p/sec/cm 2 /sr, which is signi cantly higher than that of MEHPPV-NPs (20×). When the lower dose (1×) was used for TA-NPs, the signal intensity was still ~ 4 time that of MEHPPV-NPs (20×) (Fig. 4d, Supplementary Fig. 24). Afterward, we applied TA-NPs for afterglow luminescence imaging and uorescence imaging of subcutaneous CT-26 tumors after intravenous injection. Afterglow and uorescence images showed TA-NPs could effectively target tumor areas via enhanced permeability and retention (EPR) effect ( Supplementary Fig. 25).
Tightly controlled behavior (mice in anesthetization) is more experimentally tractable, but it provides a limited view or information of the mice 32 . The ability to image a freely moving animal would represent a critical step towards understanding the relationship between the brain function and the established learning paradigms (fear conditioning, novel object recognition or social interactions), or the pathophysiology of conditions [33][34][35] . However, the real-time imaging of those unrestrained animals is greatly challenging. Mice movements would cause motion artifacts and distorted images, when the rapid movements of mice body were faster than the frame acquisition on a time scale, which is manageable in anesthetized animals but is exacerbated in awaken state 36,37 . Thereby, the visualizing study of freely moving animals requires the imaging system with ultra-high signal intensity of probe and ultra-short acquisition time. Fortunately, the Type I and Type II afterglow of TA-NPs guarantee extremely large luminescence intensity and ultra-short acquisition time (0.01 s), thereby we tried to investigate the applicability of TA-NPs for ultra-fast afterglow imaging of naturally behaving animals (Fig.  4e). Amazingly, TA-NPs allowed the observation of afterglow luminescence in a freely moving mouse for more than 10 min, through subcutaneous injection (Fig. 4f, Supplementary Fig. 26). Especially, we were able to monitor the development of afterglow signal in subcutaneous CT-26 tumors in the freely moving mouse via intravenous injection of TA-NPs (Fig. 4g, Supplementary Fig. 28), while no afterglow signals were detected from MEHPPV-NPs -injected mice under same condition (acquisition time: 0.01 s; irradiation conditions: 30 mW/cm 2 ) (Supplementary Figs. 27 and 29). By providing a moment-by-moment record of behavior, TA-NPs with high luminescence could create a serial of frameworks for relating natural behaviors to the afterglow signals, so as to assess sensitive behavioral changes in the freely moving animals.
Glioblastoma multiforme (GBM) is the most common and aggressive malignant tumor found in the central nervous system, with remarkably poor prognosis showing 5-year survival rate of 4 -5 % in clinic 38 . Thereby, the precise imaging of GBM is bene cial to early diagnosis, and of great value in assessment of GBM aggressiveness, treatment response, prognosis and recurrence [39][40][41] . Owing to the non-invasion, high sensitivity and bright luminescence, we promoted TA-NPs for imaging orthotopic GBMbearing mice (Fig. 4h). The orthotopic GBM was prepared via implantation of human glioma cells (U87) into nude mice by stereotactic implantation and con rmed via hematoxylineosin (H&E) staining of brain slice (Fig. 4k). After intravenous injection of TA-NPs, the orthotopic GBM-bearing mice showed the dynamically enhanced afterglow signals in head areas, while healthy mice displayed no obvious afterglow signals of mice head (Fig. 4i, j, Supplementary Fig. 30). Thus, TA-NPs could diffuse into the orthotopic glioma tumor via EPR effect through disrupted blood-brain barrier, and exerted high-sensitivity and accurate imaging of glioma tumor model.
The vascular system dysfunction due to atherosclerosis is one of leading causes for severe health problems around the world 42,43 . The plaque buildup in the walls of carotid arteries subsequently leads to luminal stenosis, low blood supply to brain, and even acute stroke incidents with plaque rupture 44 . Next, TA-NPs were applied to image the plaque in carotid atherosclerosis-bearing mice through intravenous injection of TA-NPs (Fig. 4l). From afterglow and uorescence images of mice (Fig. 4m, n, Supplementary   Figs. 31 -32), the atherosclerotic plaque-bearing carotid (right side) showed signi cantly higher signal intensity than the normal carotid (left side). Moreover, ex vivo afterglow and uorescence images displayed the plaque-bearing carotid with stronger signal than normal carotid (Fig. 4o, Supplementary  Fig. 33), which was consistent with the histological analysis of Oil red O staining and H&E staining . Those results suggested that TA-NPs were able to target atherosclerotic lesions and exert sensitive and non-invasive afterglow imaging of atherosclerotic plaque in living mice.

Afterglow imaging of immune responses after chemotherapy in vivo
Dynamic imaging of immune cell activity (cytotoxic T lymphocytes (T cell) activity) during immunotherapy could enable to directly monitor immune response of tumor toward immunotherapy 45 .
Granzyme B is a serine-protease released by CD8 + T cells during the cellular immune response and represents one of the two dominant mechanisms by which T cells mediate cancer cell death 46-50 . To visualize granzyme B during immunotherapy, probe should possess the su cient signal intensity, high imaging penetration depth, high SBR and no obvious afterglow photobleaching, which allows for accurate quanti cation over the repeated imaging. Encouraged by the excellent afterglow performance of TA-NPs, we further promoted TA-NPs for afterglow imaging cytotoxic T lymphocytes activity within tumors after chemotherapy or radiotherapy. To achieve this, we designed, synthesized and characterized a functional activity-based Granzyme B afterglow nanoprobe (TA-BHQ), where a Granzyme B-cleavable peptide sequence (Ile-Glu-Phe-Asp, IEFD) was conjugated with TA-NPs as luminophores and BHQ-3 as quenchers ( Fig. 5a; Supplementary Methods: Preparation of TA-BHQ and Supplementary Fig. 36). These donor-acceptor pairs with excellent spectral overlap resulted in the maximum e cacy of quenching luminescence ("off" state) through afterglow resonance energy transfer (ARET), when spatially constrained (less than 10 nm) (Fig. 5b). In the presence of Granzyme B, the IEFD peptide between TA-NPs and BHQ-3 was cleaved and ARET was reduced, leading to the recovery of afterglow luminescence ("on" state) (Fig. 5a). As a result, the gradual increase in afterglow intensity was observed over concentrations of Granzyme B (Fig. 5c). Signal quanti cation from afterglow or uorescent images showed a linear correlation between luminescence intensities and Granzyme B concentrations (Fig. 5d, Supplementary   Fig. 37). Besides, the luminescence signals remained nearly unchanged for other enzymes and biomolecules, demonstrating the selective response of TA-NPs probes toward Granzyme B (Supplementary Fig. 38).
The capabilities of TA-BHQ probes for in vivo real-time afterglow luminescence imaging of Granzyme B were validated against subcutaneous CT-26 tumor-bearing mice model. To activate the immune response, three immunotherapeutic agents including NLG919, BMS-1 and BEC were administered into the living mice (Fig. 5e, Supplementary Fig. 39). These immunotherapeutic agents mainly acted in tumor microenvironment to facilitate the in ltration of CD8 + T cells into tumor. Using BEC-treated mice as an example, the afterglow and uorescence signals of tumors were increased gradually after injection of TA-BHQ ( Supplementary Fig. 40). At 6 h post-injection of TA-BHQ, strong afterglow signals were observed for the tumor of immunotherapeutic-treated mice, which was higher than that of PBS group. Moreover, multiple-time injection of BEC showed stronger afterglow signals than that of one-time injection, indicating the dose-dependent immune response (Fig. 5f, g). Although the uorescence imaging of Granzyme B via TA-BHQ were consistent with that of those afterglow imaging, the SBR of afterglow was higher than that of uorescence ( Supplementary Figs. 41 -42).
Next, immuno uorescence staining and ow cytometry analysis were performed for measuring CD8 + T cells in ltration and Granzyme B level within tumors after various treatments (Fig. 5h, Supplementary   Figs. 43 -45). Those tumor tissues excised from BEC-treated mice displayed higher in ltration of CD8 + T cells and higher expression of Granzyme B, as compared to other groups, which revealed a higher level of immunoactivation after BEC-treatment. As summarized in heat map (Fig. 5i), the ex vivo quantitative analyses of immuno uorescence staining were coincided with the signals of afterglow, which validated that the activated signals of TA-BHQ were well correlated with the immunoactivation. Afterglow imaging of radiotherapy or/and immune checkpoint blockade therapy Radiotherapy (RT), referring to the external-beam X-ray irradiation, has gained acceptance for treating over 50 % of cancer patients in clinic 51,52 . Ionizing radiation can trigger immune system response through multiple ways, such as augmenting the expression of MHC-I in the affected cells, attenuating PD-1/ CTLA-4 in T lymphocytes, or inducing immunogenic cell death 53, 54 . Thereby, there is a growing consensus that the combination of immune checkpoint inhibitors (such as anti-PD-L1) with RT can increase immune response rates 55, 56 . Unfortunately, studies have also provided the intriguing outcomes of immune response to different doses and fractionation schedules of RT 54 .
Considering the lack of validated tools for predicting the immune response to RT or/and immune checkpoint blockade therapy, we favored to promote TA-BHQ for evaluating the immune response in mice receiving RT combined with anti-PD-L1, by monitoring Granzyme B level. The subcutaneous CT-26 tumorbearing mice received RT or/and anti-PD-L1 treatment, followed by injection of TA-BHQ and afterglow luminescence imaging (Fig. 6a). Notably, mice treated with anti-PD-L1 + RT showed dynamic enhancement of afterglow signals from tumor sites post injection TA-BHQ (Fig. 6b, c, Supplementary Fig.  46). Speci cally, compared with mice treated with PBS, those tumors treated with anti-PD-L1, RT or anti-PD-L1 + RT exhibited higher afterglow signals (Fig. 6d, e), which indicated TA-BHQ could monitor Granzyme B level during RT or/and immune checkpoint blockade therapy. Besides, those uorescence imaging results were well matched with that of afterglow imaging (Supplementary Fig. 47).
Next, the expression level of Granzyme B and population of T cell in tumors were measured by immuno uorescence staining and ow cytometry analysis (Fig. 6f, g, h, Supplementary Fig. 48). Those results demonstrated that the combination treatment (anti-PD-L1 + RT) led to more in ltration of CD8 + T cells and higher level of Granzyme B in subcutaneous CT-26 tumor, compared with single treatment (RT or anti-PD-L1) or PBS treatment, which validated that the activation of TA-BHQ probe at tumor site was owing to the increased level of Granzyme B released from activated CD8 + T cells in ltrated in the tumors.
Moreover, the heat map (Fig. 6i) displayed the afterglow signals were well correlated with both CD8 + T cells in ltration and Granzyme B levels within tumors after various treatments. Therefore, TA-BHQ was able to assess the real-time immunoactivation in living mice during chemotherapy, RT or/and immune checkpoint blockade therapy.

Conclusion
In summary, we have designed trianthracene derivatives (TA)-based nanoparticles as a new type of afterglow luminescence nanoprobe for high-sensitive, safe, accurate and dynamic imaging. TA-NPs emitted an extremely high afterglow luminescence than MEHPPV-NPs, through both electron transfer path (Type I afterglow) and the enhanced energy transfer path (Type II afterglow). Excited by light irradiation with ultra-low power density, TA-based nanoparticles showed no obvious attenuation of maximum intensity in mice even after 15-cycle excitation. Such ultra-bright afterglow and negligible photobleaching endowed TA-NPs to realize rechargeable, accurate and longitudinal imaging of subcutaneous tumor, orthotopic glioma and plaque-bearing carotid. Notably, the ultra-high afterglow signals of TA-based nanoparticles were captured under as short as 0.01 s of acquisition time, resulting in their ability for ultra-fast afterglow imaging of unrestrained subcutaneous tumors-bearing mice, and thereby the potential for gaining further insight into pathophysiology mechanism of neurologic diseases (such as epilepsy, stroke, Alzheimer disease, Parkinson). Furthermore, the engineering of TA-NPs facilitates the development of smart activatable afterglow nanoprobe (TA-BHQ) to directly monitor the e cacy of immunotherapies by imaging Granzyme B. Speci cally, the afterglow luminescence of TA-BHQ were activated by Granzyme B during chemotherapy, radiotherapy or/and immune checkpoint blockade therapy, enabling the dynamic and accurate imaging of T cell-based immunotherapies in mice. In the future, TA-BHQ probe may be extended to evaluate the immune activation of distant tumors, or even estimate the role of RT dose/ fractionation toward immunomodulation, so as to guide the individualized immunotherapy.