Light‐Up Probe for Targeted and Activatable Photodynamic Therapy with Real‐Time In Situ Reporting of Sensitizer Activation and Therapeutic Responses

Integrated systems that offer traceable cancer therapy are highly desirable for personalized medicine. Herein, a probe is reported that is composed of a red‐emissive photosensitizer (PS) with aggregation‐induced emission characteristics and a built‐in apoptosis sensor with activatable green emission for targeted cancer cell ablation and real‐time monitoring of PS activation and therapeutic response. The probe is nonemissive in aqueous media and can be selectively uptaken by αvβ3 integrin overexpressed cancer cells. Cleavage of the probe by intracellular glutathione leads to release of the apoptosis sensor and red fluorescence turn‐on to report the PS activation. Upon light irradiation, the PS can generate reactive oxygen species to induce cell apoptosis and activate caspase‐3/‐7, which will cleave the apoptosis sensor to yield intense green fluorescence. Both the red and green emission can be obtained through a single wavelength excitation, which makes the probe very convenient for therapeutic protocol development.


Introduction
Targeted delivery, accurate assessment of active therapeutic agents, and in situ reporting of their therapeutic effects are of high importance in cancer therapy. [ 1 ] The capability to visualize the whole process will undoubtedly provide more useful information to further optimize and advance the therapy. However, our capability is still limited in answering precisely when, where and how the therapeutic agents are delivered as well as what is their therapeutic effect. To realize targeted delivery, drug conjugates or drug encapsulated carriers with surface targeting moieties have been developed. [ 2 ] Fluorescent dyes and quantum dots and small radius of action (<20 nm) of ROS. [ 19 ] There is thus very limited strategies reported to provide in situ monitoring of the therapeutic effect of PDT. [ 20 ] In addition, the currently available PSs frequently cause unwanted normal cell death due to their intrinsic phototoxicity and lack of selectivity for cancer cells. [ 14 ] As a consequence, targeted and/or activatable PSs with high therapeutic effi ciency to cancer cells but minimized side effects to normal cells are subsequently reported. [21][22][23][24][25] However, the integration of all the desired properties into a single molecular probe that is able to simultaneously target and image cancer cells, report the PS activation and predict the therapeutic response in situ remains a grand challenge.
To visualize multiple processes in cancer therapy, fl uorophores exhibiting different fl uorescence upon a single wavelength excitation are highly desirable as they can minimize the complexity of fl uorescence imaging by offering simultaneous detection. The design principle of traditional fl uorescent probes often requires a pair of fl uorescent dyes (energy donor/ acceptor or donor/quencher) or relies on photoinduced electron transfer (PET) mechanism to visualize a single biological event, which makes it diffi cult to develop one molecular probe for monitoring multiple processes. [ 26 ] This calls for the development of multifunctional materials, where a single molecule can be used to report an event or even also born with therapeutic functions. It is fortunate that the recent development of fl uorogens with aggregation-induced emission characteristics (AIEgens) offers such an opportunity. [ 11,[27][28][29][30][31][32][33] AIEgens show almost no fl uorescence as molecular species, but bright fl uorescence in aggregate state due to restriction of intramolecular motions and prohibition of energy dissipation via non radiative channels. [ 34 ] Taking advantages of the unique properties of AIEgens, we have developed light-up probes without incorporation of any quencher moiety or energy transfer pairs. [ 27,28 ] More recently, we and other groups also reported functional AIEgens with tunable emissions, and demonstrated that they could be used as potential chemo-drugs or PSs for chemotherapy and PDT. [35][36][37] In this contribution, we developed an AIEgen based probe, namely TPETP-SS-DEVD-TPS-cRGD ( Figure 1 ), for targeted imaging, activatable and traceable cancer cell PDT with additional capacities of real-time self-reporting of PS activation and therapeutic responses upon a single wavelength excitation. Tetraphenylethenethiophene (TPETP) is a red-emissive AIEgen with photosensitization property, while tetraphenylsilole (TPS) is an iconic green emissive AIEgen [ 27 ] and both could be excited with a 405 nm laser to yield distinct emission peaks. As ROS could cause cancer cell death through apoptosis pathway within a short period of time, [38][39][40] and caspases-3/7 are key mediators of cell apoptosis, [ 41 ] a caspase-3/-7 responsive peptide substrate DEVD based apoptosis sensor is incorporated into the probe. In addition, c-RGD and -S-S-are used as the targeting ligand and linker to realize dual-targeted activatable PDT. Upon c-RGD mediated cellular uptake, the -S-S-bond can be cleaved by glutathione (GSH) to yield red fl uorescence turn-on of the PS and release the apoptosis sensor. Consequently, the activated caspases cleave the DEVD substrate on the probe and release hydrophobic TPS residue with intense green fl uorescence to report the therapeutic effect. As compared to the recent report of a lysotracker and PS coencapsulated nanoparticles for monitoring of the therapeutic responses of PDT via a fl uorescenceoff signal, [ 20 ] our distinct fl uorescence turn-on signal is more desirable in complicated biological systems. (2) cleavage of the disulfi de bond through intracellular reduction by glutathione to release the nonemissive apoptosis sensor and the PS with red fl uorescence; (3) upon light irradiation, ROS generation to induce cell apoptosis and activate caspase enzymes; (4) cleavage of the DEVD substrate by the activated caspases on the apoptosis sensor to result in green fl uorescence turn-on. The red fl uorescence of the TPETP residue can be used for image-guided photodynamic therapy while the green fl uorescence of TPS can be used for the therapeutic response imaging. The structures of TPETP and TPS residues are shown in the Scheme S1, Supporting Information.

Design Principle of the AIEgen PS and the Probe
The design principle of the AIEgen PS is based on the following two considerations: AIE characteristics with effi cient ROS generation, and the same excitation but distinct emission wavelengths from those of TPS. Build upon tetraphenylethylene (TPE), one of the most commonly used AIEgens with UV absorption and weak ROS generation, [ 42 ] introducing dicyanovinyl group as an electron acceptor (A) and methoxy group as an electron donor (D) into a TPE core forms a D-π-A structure, which is expected to yield a new AIEgen with red shifted absorption/emission maxima. In addition, the D-π-A structure can lead to small Δ E S1-T1 (energy gap between the lowest singlet excited state (S 1 ) and the lowest triplet excited state (T 1 ) states), which is favorable for intersystem crossing (ISC) to yield efficient ROS generation. [ 43 ] The molecular structures, HOMO and LUMO distribution and Δ E S1-T1 values of TPE (1.22 eV), TPS (1.27 eV) and TPETP (0.29 eV) are summarized in Figure 2 . TPETP exhibits a much smaller Δ E S1-T1 compared with TPE, suggesting a potentially high ISC rate and thus possibly efficient ROS generation upon light irradiation. [ 44 ] The probe design principle is illustrated in Figure 1 . The probe is composed of fi ve components: (1) a red emissive TPETP to serve as a visualizing agent and a PS; (2) a disulfi de linker, which can be cleaved by GSH overexpressed in cancer cells; (3) a highly hydrophilic DEVD substrate that can be specifi cally cleaved by caspase-3/-7; (4) a green emissive TPS molecule to serve as the signal reporter for the apoptosis sensor; and (5) a hydrophilic cyclic arginine−glycine−aspartic acid (cRGD) peptide for targeting cancer cells with overexpressed α v β 3 integrin. The probe is water-soluble and shows very weak fl uorescence in aqueous media. Once the probe is selectively taken up by α v β 3 integrin overexpressed cancer cells through receptor mediated endocytosis, the disulfi de bond will be cleaved by intracellular GSH, which will release the apoptosis sensor and turn on the red fl uorescence of the TPETP residue (Scheme S1, Supporting Information) as an indication of PS activation. Upon light illumination, the PS will generate ROS, which induces the cell apoptosis and activates caspase-3/-7. The caspases will then cleave the DEVD substrate to result in green fl uorescence turnon of the TPS residue (Scheme S1, Supporting Information), which is able to real-time report the therapeutic effi ciency of PDT. Considering that the overall process can be monitored with a single wavelength excitation, the probe thus offers a unique opportunity for targeted and activated cancer therapy with real-time reporting of PS activation and therapeutic responses.

Synthesis of TPETP-NH 2 and Identifi cation of the Isomer
The two isomers of TPETP-NH 2 ( cis -TPETP-NH 2 and trans -TPETP-NH 2 ) were synthesized as depicted in Figure 3 . Briefl y, compound 1 was treated with n -butyllithium, trimethyl borate, and diluted HCl to yield compound 2 with boronic acid functionality, which underwent palladium-catalyzed coupling with 2-thiophenecarbonyl chloride to yield compound 3 . Compound 3 was further treated with TiCl 4 and malanonitirile to generate compound 4 with a dicyanovinyl group, which was subsequently treated with BBr 3 to yield compound 5 with one free hydroxyl group. The reaction between compound 5 and 3-(boc-amino) propyl bromide furnished compound 6 , which was reacted with trifl uoroacetic acid (TFA) to deprotect the Boc group and yield the mixture of cis and trans isomers of TPETP-NH 2 . The compounds 1 -6 were characterized by NMR and mass spectroscopies (Figures S1-S4, Supporting Information). The two isomers of TPETP-NH 2 were separated by preparative high-performance liquid chromatography (HPLC) as red powders after freeze drying ( Figures S5 and S6, Supporting Information). After measuring 1 H COSY and NOESY NMR for the major isomer, all the peaks with different chemical shifts were assigned (Figures S7-S9, Supporting Information). As shown in the NOESY spectrum ( Figure 4 ), the H-3&5 have strong correlations with the H-21&25, which suggest that the rings A and C are on the same side. Since there was no signifi cant difference in fl uorescence and ROS generation ability between the two isomers ( Figure S10A,B, Supporting Information), the major isomer of cis-TPETP-NH 2 was used for the subsequent modifi cation.

Optical Properties of the Probe and its Activation by GSH
The absorption and emission spectra of cis -TPETP-NH 2 in DMSO/water mixtures (v/v = 1/199) are shown in Figure 5 A. The UV-vis absorption of cis -TPETP-NH 2 covers the range of 400-500 nm with a shoulder peak at 430 nm. The PL spectrum is ranged from 550 to 850 nm with an emission maximum at 650 nm and the quantum yield (Φ) is determined to be 0.13 using 4-(dicyanomethylene)-2-methyl-6-(4-dimethylaminostyryl)-4 H -pyran (DCM) as the reference (Φ = 0.43 in water). To study whether cis -TPETP-NH 2 retains the AIE property, its PL spectra in DMSO/water mixtures were monitored at the same compound concentration but with different water fractions ( f w ). As shown in Figure 5 B, cis -TPETP-NH 2 is almost nonemissive in its benign solvent DMSO. This is attributed to the free motions of cis -TPETP-NH 2 in molecularly dissolved state. The fl uorescence of cis -TPETP-NH 2 increases quickly when the f w is higher than 50 vol%. At f w = 99 vol%, the fl uorescence intensity of cis -TPETP-NH 2 is 120-fold higher than its intrinsic fl uorescence in pure DMSO. The PL results demonstrate that cis -TPETP-NH 2 is AIE-active.
The PL spectra of cis -TPETP-NH 2 and the probe in DMSO/ PBS mixtures (v/v = 1/199) are shown in Figure 5 C. cis -TPETP-NH 2 shows intense red fl uorescence. However, the probe TPETP-SS-DEVD-TPS-cRGD is almost nonemissive because of its good water-solubility due to the attached hydrophilic DEVD and cRGD peptides. The signifi cant difference in the PL spectra of the probe and cis -TPETP-NH 2 inspired us to develop a cancer cell specifi c light-up probe based on cleavage of the disulfi de bond by intracellular thiols. The fl uorescence response of the probe toward GSH was studied fi rst. As shown in Figure 5 D, the red fl uorescence of TPETP residue is increased steadily over 2 h upon addition of GSH to the probe solution. The fl uorescence intensity reaches a plateau after 90 min incubation, which is 14-fold higher than the original intensity of the probe. The gradual enhancement of red fl uorescence is ascribed to the increased amount of TPETP residue formed by cleavage of disulfi de bond of the probe. Furthermore, the formation of TPETP aggregates was confi rmed by laser light scattering (LLS) measurements. No LLS signals could be detected from the probe initially, while aggregates with an average diameter of 148 ± 12 nm were detected after GSH treatment ( Figure S15A, Supporting Information). Subsequently, the probe at different concentrations was incubated with GSH for 90 min and the corresponding fl uorescence changes were recorded. As shown Adv. Funct. Mater. 2015, 25, 6586-6595 www.afm-journal.de www.MaterialsViews.com  in Figure 5 E, the plot of the PL intensities at 650 nm shows a linear relationship with the probe concentrations ( R 2 = 0.98), indicating that the red fl uorescence intensity could be used to quantify the amount of activated PS. The selectivity of the probe was subsequently studied, and the PL results revealed that the red fl uorescence is only turned-on in the presence of reducing agents such as GSH and ascorbic acid while other bioacids and proteins show negligible fl uorescence changes (Figure 5 F). These results further support that the increase of red fl uorescence in Figure 5 D is due to the cleavage of the disulfi de bond by the reducing agents.
The ROS generation of PSs upon light irradiation is a critical step for effi cient PDT. The ROS quantum yield (Φ) of cis -TPETP-NH 2 in DMSO/PBS (v/v = 1/199) was determined to be 0.68 using Rose Bengal (RB) as the standard photo sensitizer (Φ RB = 0.75 in water), which is much higher than the clinically used PSs such as Photofrin (Φ = 0.28) or Laserphyrin (Φ = 0.48). [ 46 ] The ROS generation from the TPETP residue was also evaluated by measuring the absorbance of the solution containing the GSH pretreated probe and a ROS indicator, 9,10-anthracenediyl-bis(methylene)dimalonic acid (ABDA), in DMSO/PBS mixtures (v/v = 1/199) under light irradiation. As depicted in Figure S10C, Supporting Information, the absorption peaks of ABDA at 358, 378, and 400 nm decrease gradually upon light irradiation, as a result of fast reaction between the ROS generated by the TPETP residue and the anthracene moiety in ABDA. After 12 min irradiation, the absorbance at 400 nm is decreased to 22% of its original value. When vitamin C (a well-known ROS scavenger) is added, only 6% of the original ABDA absorbance is decreased under the same treatment, further confi rming that ROS is generated by the TPETP residue upon light irradiation.

Caspase-3/-7 Activation of the Released Apoptosis Sensor
The absorption and emission spectra of TPS-2N 3  green fl uorescence (the broad emission above 600 nm is attributed to the TPETP residue), indicating that the release of the apoptosis sensor alone will not yield TPS fl uorescence. However, a quick green fl uorescence increase is observed when the GSH-pretreated probe (10 × 10 −6 M ) is further treated with caspase-3/-7 due to the release and aggregation of the hydrophobic TPS residue ( [ 47 ] It is important to note that the fl uorescence of GSH-pretreated probe at 480 nm intensifi es upon incubation with increased concentrations of caspase-3 (Figure 6 E). The plot of the PL intensities at 480 nm against caspase-3 concentrations gives a linear fi t ( R 2 = 0.97), suggesting that it is possible to semiquantify the concentration of caspase-3 through PL intensity changes. The detection limit of caspase-3 is 2.3 × 10 −12 M estimated by three times standard deviation (3σ) method. The selectivity of the apoptosis sensor DEVD-TPS-cRGD was evaluated by incubating the GSH-pretreated probe with pepsin, lysozyme, bovine serum albumin (BSA), and other caspases. As shown in Figure 6 F, only caspase-3/-7 treated groups trigger fl uorescence of the apoptosis sensor, indicating that the DEVD substrate is specifi cally cleaved by caspase-3/-7 when the cells undergo apoptosis. Since there are many kinds of enzymes in the cells, we also incubated the probe with cellular lysates of untreated and apoptotic (induced with 2 × 10 −6 M staurosporine, STS) MDA-MB-231 cancer cells. [ 48 ] The fl uorescence intensities at 480 and 650 nm were monitored after different periods of incubation. As shown in Figure S16, Supporting Information, the fl uorescence intensities at 650 nm increase quickly in both the untreated and apoptotic cells confi rming that the intracellular GSH can cleave the disulfi de bond. On the other hand, the fl uorescence enhancement at 480 nm is only observed when the probe is incubated with the apoptotic cell lysate.

Monitoring of Intracellular PS Release
To demonstrate the light-up imaging of specifi c α v β 3 integrin overexpressed cancer cells, the probe was incubated with α v β 3 integrin overexpressed MDA-MB-231 breast cancer cells and low α v β 3 integrin expressed MCF-7 breast cancer cells as well as 293T normal cells. [ 49 ] As shown in Figure 7 , the red fl uorescence from TPETP residues in MDA-MB-231 cells intensifi es gradually with the increase of incubation time. In addition, the red fl uorescence in MDA-MB-231 is much stronger than those in MCF-7 and 293T cells under the identical conditions. It is important to note that free TPETP can also be uptaken by Adv. Funct. Mater. 2015, 25, 6586-6595 www.afm-journal.de www.MaterialsViews.com MDA-MB-231 cells to yield strong red fl uorescence (Figure 7 H), and the fl uorescence pattern is similar to that in Figure 7 E. When MDA-MB-231 cells are pretreated with free cRGD prior to probe incubation, the red fl uorescence in the cells is dramatically reduced. Similarly, the MDA-MB-231 cells show much weaker red fl uorescence when the cells are pretreated with g -glutamylcysteine synthetase buthionine sulfoximine (BSO), which could inhibit the cells to produce GSH. [ 50 ] Moreover, fl ow cytometry analysis ( Figure S17, Supporting Information) is used to further evaluate the probe specifi city and the results are in well accordance with the confocal images shown in Figure 7 . These results clearly demonstrate that the probe is specifi cally internalized by MDA-MB-231 cells through receptor mediated endocytosis and the red fl uorescence is turned on by intracellular GSH, which can be used for PS activation monitoring and specifi c cancer cell imaging.

Imaging of Cell Apoptosis Induced by PDT
The intracellular ROS generation upon light irradiation in MDA-MB-231 cells was fi rst evaluated by using 2′,7′-dichlorodihydrofl uorescein diacetate (DCF-DA) as the indicator ( Figure S18, Supporting Information). To explore the capability of using DEVD-TPS-cRGD as an apoptosis sensor, the MDA-MB-231 cells were pretreated with cis-TPETP-NH 2 (10 × 10 −6 M ) with light irradiation (1 min) or STS (1 × 10 −6 M ), then the cells were further incubated with DEVD-TPS-cRGD for 4 h. As shown in Figure S19, Supporting Information, strong green fl uorescence of the TPS residue is observed in both STS and cis -TPETP-NH 2 treated MDA-MB-231 cells with light irradiation for the latter, indicating that apoptosis occurs in both cases. It should be noted that the DEVD-TPS-cRGD itself is of low cytotoxicity ( Figure S20, Supporting Information), which does not induce cell apoptosis.
To further explore the possibility of the probe for PDT and in situ apoptosis imaging in α v β 3 integrin rich cancer cells, we incubated the probe with MDA-MB-231, MCF-7, and 293T cells for 4 h with light irradiation. The fl uorescence changes of TPS in cells were monitored by confocal images. As shown in Figure 8 , the green fl uorescence of TPS residue in probe incubated MDA-MB-231 cells increases gradually with time when the cells are exposed to light irradiation. Together with that shown in Figure S16, Supporting Information, these results clearly indicate that the ROS generation of TPETP residue upon light irradiation can cause cell apoptosis and activate caspase-3/-7 to cleave the DEVD substrate. This is further confi rmed by TPETP and cRGD-TPS-DEVD coincubated MDA-MB-231 cells, which show strong green fl uorescence after light irradiation. On the contrary, only a weak fl uorescent signal is detected in MCF-7 or 293T cells for more than 4 h of light irradiation. When MDA-MB-231 cells were pretreated with free cRGD or vitamin C prior to the probe incubation, the cells showed much weaker green fl uorescence. This is because the free cRGD competes with the α v β 3 integrin receptor to the probe while vitamin C serves as a ROS scavenger to consume the ROS generated by the activated PS.

Imaging of Cell Apoptosis
To further confi rm that the probe can image cell apoptosis, the probe incubated MDA-MB-231 cells after light irradiation were costained with anticaspase-3 primary antibody and a Texas Redlabeled secondary antibody. As shown in Figure 9 A, the green fl uorescence of the TPS residue and the red immunofl uorescence signals from Texas Red are overlapped with high coefficient. When the cells are incubated with a caspase-3/7 inhibitor of 5-[( S )-(+)-2-(methoxymethyl)pyrrolidino]sulfonylisatin, followed by light irradiation, the green fl uorescence of TPS residue is greatly reduced and only red fl uorescence of Texas Red remains (Figure 9 B). Overall, these results demonstrate that the ROS generation of TPETP residue can induce cell apoptosis and the activation of the caspase-3/7 has led to the green fl uorescence turn-on of TPS residue, which can be used for selfreporting of cell apoptosis in situ.

Cell Death Prediction
The CytoTox-Fluor cytotoxicity assay was used to evaluate the correlation between cell viability and apoptosis after cell Adv. Funct. Mater. 2015, 25, 6586-6595 www.afm-journal.de www.MaterialsViews.com www.afm-journal.de www.MaterialsViews.com incubation with the probe and light irradiation. The CytoTox-Fluor cytotoxicity assay can evaluate the cell viability using a fl uorogenic peptide substrate (bis-alanyl-alanyl-phenylanlanylrhodamine 110; bis-AAF-R110). We studied the correlation between the R110 fl uorescence intensity evaluated by the assay and the TPS residue fl uorescence of the probe in MDA-MB-231 cells after light irradiation for different period of time. After the cells were incubated with the probe for 4 h and exposed to light irradiation for different time durations, the fl uorescence intensities of R110 and the TPS residue were recorded by a microplate reader. As shown in Figure 10 , the fl uorescence intensity of R110 after different time of light irradiation correlates well with the apoptosis induced TPS residue fl uorescence change, suggesting that the probe is capable of predicting the PDTinduced cell viability by measuring the TPS residue fl uorescence intensity noninvasively. Therefore, the probe can be used to report its therapeutic response in situ.

Conclusions
We developed a multifunctional probe based on two AIEgens with distinguishable emissions for targeted cancer cell ablation and simultaneous real-time reporting of PS activation and therapeutic responses upon a single wavelength excitation. The probe itself is almost nonemissive in aqueous media due to the attachment of hydrophilic peptides, and the molecularly dissolved state facilitates free rotation of the phenyl rings contained in both AIEgens. Upon cRGD-mediated cellular uptake, the red fl uorescence of TPETP residue is turned on as a result of the cleavage of the disulfi de group by intracellular GSH, which has been used for cellular imaging and monitoring of the PS activation. Upon image-guided light illumination, the green fl uorescence of TPS residue is generated by apoptosis induced caspase-3/-7 activation, which is used for monitoring of the cell apoptosis and evaluation of the therapeutic responses. As compared with the  previously reported molecular probes that can only monitor a single process of drug activation or therapeutic response, simultaneous monitoring of the whole processes is able to provide a clearer picture for further improved cancer therapy. We anticipate that the integrated probe design will offer new opportunities for cancer therapy and early evaluation of the therapeutic response will help to forecast specifi c therapeutic regimes for individual patient to help realize the ultimate goal of personalized medicine. Based on the proof-of-concept probe design in this work, further development of new AIEgens through molecular design with longer absorption and emission wavelengths for direct in vivo application is our next goal.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.