Open Access Article
Dong
Wang
ad,
Huifang
Su
c,
Ryan T. K.
Kwok
a,
Xianglong
Hu
*ab,
Hang
Zou
a,
Qianxin
Luo
a,
Michelle M. S.
Lee
a,
Wenhan
Xu
a,
Jacky W. Y.
Lam
a and
Ben Zhong
Tang
*a
aHong Kong Branch of Chinese National Engineering Research Center for Tissue Restoration and Reconstruction, Department of Chemistry, Institute of Molecular Functional Materials, State Key Laboratory of Neuroscience, Division of Biomedical Engineering, Division of Life Science, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China. E-mail: tangbenz@ust.hk
bMOE Key Laboratory of Laser Life Science & Institute of Laser Life Science, College of Biophotonics, South China Normal University, Guangzhou, 510631, China. E-mail: xlhu@ust.hk; xlhu@scnu.edu.cn
cDepartment of Osteology, The First Affiliated Hospital of Zhengzhou University, Zhengzhou University, Zhengzhou 450000, PR China
dCenter for AIE Research, College of Materials Science and Engineering, Shenzhen University, Shenzhen 518060, China
First published on 13th March 2018
The synthesis of water-soluble near-infrared (NIR)-emissive fluorescent molecules with aggregation-induced emission (AIE) characteristics and theranostic functions is highly desirable but remains challenging. In this work, we designed and readily prepared for the first time such a molecule with AIE features, good water-solubility and intense emission in the NIR region. This AIE luminogen (AIEgen) is able to specifically “light up” the cell membrane without the involvement of a washing procedure. Interestingly, the staining process can be performed by simply shaking the culture with cells at room temperature for only a few seconds after the addition of the AIEgen, indicating an ultrafast and easy-to-operate staining protocol. This is the first fluorescent “light-up” probe for cell-imaging that allows the combination of a short staining period (at the second-level) with a wash-free process. Additionally, the presented AIEgen has also been developed to serve as an excellent phototherapeutic agent for high efficiency generation of reactive oxygen species (ROS) upon visible light irradiation, which allows its effective application in the photodynamic ablation of cancer cells, demonstrating its dual role as an imaging and phototherapeutic agent.
The current development of NIR AIEgens is far from ideal, and until now, only a handful of AIEgens exhibiting high-performance NIR emission have been developed and used in biological studies.16–18 Furthermore, considering that biological research studies are conducted in physiological environments or aqueous media, the utilization of water-soluble AIEgens holds an intrinsic advantage. Although some water-soluble AIEgens with short-wavelength emissions have been prepared and employed as powerful bioprobes,19–23 to the best of our knowledge, there have been no previous reports on water-soluble NIR AIEgens. Developing NIR AIEgens with good water-solubility remains an important and challenging task, even though considerable efforts have been devoted by the scientific community.
In vitro cellular imaging, that is one of the most widely used applications of fluorescence bioimaging techniques, has become indispensable for biological analysis and clinical diagnosis. As an important cell organelle, the plasma membrane that consists of the phospholipid bilayer is a protective two-dimensional boundary between a living cell and its surroundings. The plasma membrane has been demonstrated to be involved in various cellular processes and bio-functions, such as cell signaling, cell adhesion, endocytosis, exocytosis and selective permeation of substances.24,25 The abnormality of the cell plasma membrane is a critical biomarker for cell status and many diseases. Therefore, visualizing the plasma membrane using fluorescent bioprobes must be significantly important and useful. However, previously developed plasma membrane-specific fluorophores (such as DiO, DiI, and CellMask) have their respective and collective drawbacks including short emission wavelengths, small Stokes shifts, requirement of hazardous organic solvents for preparing stock solution, long incubation periods and tedious washing procedures after cell staining.26–29 In particular, the latter two shortcomings have been long-term unresolved issues in cellular fluorescence imaging. Long incubation is time-consuming, and often causes nonspecific illumination of cellular components.30 Aiming to improve the signal-to-noise (S/N) ratio of cell imaging, a washing process after cell staining is usually required for eliminating the strong residual signal from the free dyes. The post-washing process could result in some problems, for instance, delaying the acquisition of microscopic data, decreasing the accuracy of cell-imaging results due to both the altered cellular environment and the loss of cells. Moreover, the washing procedure is incompatible with continuous sensing or monitoring of biological processes.31,32 Additionally, the plasma membrane is considered to be a wonderful cellular targeting site for implementing therapeutic applications33 because the plasma membrane is strongly related to various cellular processes, and is the outermost protection layer of cells, due to which its destruction is fatal to cells. However, almost all of the previously reported plasma membrane-staining fluorophores can only be used as imaging probes instead of dual applications in simultaneous imaging and therapy. Evidently, developing a novel fluorescent plasma membrane probe that overcomes the above-mentioned deficiencies would be remarkably important and urgently needed.
In this contribution, we report for the first time the design and facile synthesis of a water-soluble AIEgen (named TTVP) with emission in the NIR region. The plasma membrane can be specifically targeted by TTVP through a wash-free and ultrafast staining procedure, and furthermore, image-guided photodynamic cancer cell ablation was successfully achieved upon visible light irradiation.
![]() | ||
| Scheme 1 (a) Synthetic route to TVP. (b) Design rationale and synthetic route to water-soluble NIR AIEgen TTVP. | ||
As depicted in Scheme 1, TVP was facilely synthesized by a one-step reaction of 4-(diphenylamino)benzaldehyde with 4-methyl-1-(3-(trimethylammonio)propyl)pyridin-1-ium bromide, with a yield of 72%. TTVP was produced by a two-step reaction. The Suzuki–Miyaura coupling reaction of 4-bromo-N,N-diphenylaniline with (5-formylthiophen-2-yl)boronic acid smoothly proceeded and generated 5-(4-(diphenylamino)phenyl)thiophene-2-carbaldehyde, which then underwent a condensation reaction with the pyridinium salt, facilely giving TTVP with a total yield of 65%.
070 M−1 cm−1molar extinction coefficient) and 480 nm (with a 33
517 M−1 cm−1molar extinction coefficient), respectively (Fig. 1a and Table S1†). The longer absorption wavelength of TTVP is ascribed to its smaller HOMO–LOMO energy gap than TVP (Fig. 2) due to the stronger electron donating–accepting interaction of the emitting center of TTVP, resulting from the existence of the thiophene fragment with electron-donating properties. Density functional theory (DFT) calculations also indicate that the electron density in the highest occupied molecular orbital (HOMO) is delocalized at the triphenylamine segment, while the pyridinium unit dominates the lowest unoccupied molecular orbital (LUMO).
The investigation of AIE features in water/THF mixtures with different THF fractions (fT) demonstrates that both TVP and TTVP are typical AIE-active molecules. They are almost non-emissive in aqueous solutions in the single-molecule state, mainly because the rotational motions of molecular rotors (such as the phenyl rings of the triphenylamine moiety, carbon–carbon double bonds, and pyridinium and thiophene units) consume exciton energy and increase nonradiative decay rates, leading to non-emission. And the photoluminescence (PL) intensities gradually increase with increasing the fraction of THF due to the formation of nanoaggregates (Fig. 1b). The strongest PL intensities were observed at a 90% fraction of THF upon aggregation, in which the PL intensities were enhanced to about 41.7- and 97.3-fold compared with that of aqueous solutions (Fig. 1c). The significantly enhanced emissions in aggregates could be attributed to the fact that the restriction of the rotor motions activates radiative decay; meanwhile, the twisted conformation of the triphenylamine segment can extend the intermolecular distance and prevent emission quenching by the reduction of the intermolecular π–π interaction, thus switching on the luminescence process in the aggregation state. Their maximum emissions in the aggregation state are located at 629 and 708 nm, respectively, with 4.4% and 1.7% quantum yields, indicating both their red/NIR-emission properties and large Stokes shifts. Moreover, dynamic light scattering (DLS) analysis and/or transmission electron microscopy (TEM) measurements were performed to confirm the formation of aggregates upon the addition of THF into the aqueous solutions. DLS reveals that the average hydrodynamic diameters of these nanoaggregates that formed in the suspension containing a 90% fraction of THF are around 46 for TVP, and 43 nm for TTVP (Fig. S8† and 1d), while their spherical morphology was observed by TEM analysis. In the solid state, the PL spectra of TVP and TTVP peaked at 657 and 705 nm, respectively (Fig. 1e), while their fluorescence decay curves reveal that their lifetimes are 5.75 and 0.92 ns (Fig. 1f). In addition, the solvatochromism study shows that with the increase of solvent polarity, the emission maximum of TTVP largely red shifts while the emission intensity was considerably reduced (Fig. S9†), suggesting a strong twisted intramolecular charge transfer (TICT) effect.34
The specificity of TTVP to the plasma membrane was evaluated by co-staining with DiO, which is a commercially available bioprobe for the plasma membrane. In this co-localization experiment, after incubating HeLa cells with DiO for 10 min, TTVP was added into the culture followed by culture shaking for a few seconds at room temperature. In order to accommodate the staining protocol of DiO, postwashing after cell staining was carried out. As shown in Fig. 4, TTVP can selectively accumulate at the plasma membrane and emit strong red fluorescence. The well-merged image between TTVP and DiO indicates good specificity for the plasma membrane, and the Pearson correlation coefficient is determined to be 89%. It was observed that TVP can also specifically stain the plasma membrane within a short incubation period (Fig. S11†). Moreover, aiming to assess the photostability of TTVP and DiO parallelly, continuous excitation and sequential scanning with a confocal microscope were performed. The result shows that the emission intensity of TTVP slightly decreased within 15 min irradiation (Fig. 4e and f), and the fluorescence loss of DiO is very obvious upon irradiation under the same conditions (Fig. 4g and h), demonstrating the superior photostability of TTVP to that of DiO. Furthermore, encouraged by the distinct advantages of TTVP for membrane-specific imaging, this ultrafast staining and wash-free cellular imaging protocol was further employed for staining other cell lines, including 293T, HCC827, HCT116, and MDCK2. In all tested cases, the plasma membrane was clearly visualized with a high S/N ratio of cell imaging with intense red emission (Fig. S12†), suggesting the good tolerance of TTVP to cell types.
The strong absorption of TTVP in the visible light region could lead to the utilization of visible light as the excitation light source for photodynamic therapy (PDT) application, which is an appropriate and gentle approach for cancer therapy, and has been clinically approved for eliminating malignant tumor cells with minimal invasion and precise controllability.36–39 Visible light causes less damage to the biological system than UV light. The ROS generation efficiency of TTVP was initially determined by the use of H2DCF-DA as an indicator, which emits fluorescence with a “turn on” process triggered by ROS. As depicted in Fig. 5a, TTVP or H2DCF-DA alone was non-emissive or weakly emissive, and each fluorescence intensity remains almost constant during 60 s white light irradiation. In contrast, in the presence of TTVP, the emission intensity of H2DCF-DA gradually enhanced with increasing exposure time to white light, reaching 87-fold within 60 s. In addition, a high 1O2 quantum yield (80.16%) for TTVP was determined using a commercial 1O2 probe, 9,10-anthracenediyl-bis(methylene) dimalonic acid (ABDA), as the indicator, and employing Rose Bengal (RB) as the standard photosensitizer (Fig. S13†). The efficient ROS generation of TTVP could be attributed to both its small singlet–triplet energy gap (0.47 eV) and excellent monodispersity. The former favors the yield improvement of the triplet excited state, and the latter can enlarge the contact area between TTVP and oxygen. Moreover, it was observed that the ROS generation efficiency of TTVP was enhanced upon aggregate formation (Fig. S14†). The enhancement could be attributed to the increased intersystem crossing (ISC) rate and the improved yield of the triplet excitons, which result from the smaller singlet–triplet energy gap in the aggregation state.39,40 Furthermore, the 1O2 quantum yield (49.22%) of TVP is lower than that of TTVP due to the larger singlet–triplet energy gap (0.56 eV) of TVP. The effective ROS generation endows TTVP with prominent potential for PDT application, which was quantitatively evaluated on HeLa cells by a standard MTT assay. A dose-dependent toxicity was determined in both the absence and presence of white light irradiation. The results demonstrate that TTVP exhibits low cytotoxicity under dark conditions, which is one of the essential features of photosensitizers for PDT application. The HeLa cell viability dropped rapidly to 15% with a concentration of 500 nM, and 1 μM TTVP causes almost complete cell death under white light irradiation (Fig. 5b), indicating its remarkable efficiency for cancer cell ablation in the PDT pathway. In comparison, HeLa cells maintained 90% viability when they were incubated with 1 μM of TTVP under dark conditions. Furthermore, flow cytometric analysis using Annexin V-FITC/propidium iodide (PI) double staining was utilized to determine cell apoptosis (Fig. 5c and d).41 It was observed that light irradiation caused cancer cell necrosis in a short time, and the ratio of necrotic cells significantly increased with prolonging the irradiation time, which suggested the high efficiency of TTVP in photodynamic ablation of cancer cells. It is worth noting that continuous light irradiation with strong power (around 18.5-fold higher than that for cell imaging depicted in Fig. 3) led to some changes of cells, for instance, TTVP can gradually enter the cells (Fig. 5e–i); the cell membrane morphology changed, and the formation of blebs on the plasma membrane was clearly observed (Fig. S15†), which is a sign of cell death.42 These changes can be attributed to the fact that the ROS generated from TTVP considerably disrupt the rigidity and permeability of the plasma membrane, and induce cancer cell death. In addition, COS-7 normal cells were also employed for both cellular imaging and PDT application. It was found that TTVP can stain normal cells, but a higher concentration (5 μM) is needed for achieving clear cellular imaging (Fig. S16†). In comparison, when HeLa cancer cells were tested for cellular imaging, 0.5 μM of TTVP is able to provide very good quality of cellular imaging (see Fig. 3). The difference is perhaps caused by the more negatively charged surface of cancer cells.35 The positive charges of TTVP enable it to rapidly and efficiently bind with cancer cells through electrostatic interactions. The results of the PDT study showed that COS-7 normal cells maintained 33% cell viability when 20 μM of TTVP was used (Fig. S17†). In comparison, 0.5 μM of TTVP almost completely killed HeLa cancer cells. The big difference could be attributed to the much less efficient staining of TTVP with normal cells.
The in vivo imaging experiment and cytotoxicity test have also been carried out. As depicted in Fig. 6, TTVP clearly provided tumor imaging after intratumoral injection of TTVP aqueous solution in HeLa tumor-bearing mice. In order to evaluate the tumor retention potential of TTVP, the observation of tumor imaging for the duration from 10 min to 24 h after injection was performed. It was observed that the tumor site was continuously imaged upon intratumoral injection with TTVP. At 24 h post-injection, the tumor fluorescence was still significant for observation, suggesting the outstanding tumor retention properties of TTVP, possibly benefiting from the persistence of membrane insertion. In addition, TTVP can accumulate at tumors with good specificity at 24 h post-injection (Fig. 6b), and the images of H&E-stained organ slices showed that there were no obvious pathological changes of the main organs in TTVP treated mice, indicating its undetectable systemic toxicity.
This study provides a new insight into the design of water-soluble AIEgens with long emission wavelengths for the development of efficient and easy-to-operate fluorescent bioprobes with “light up” nature. Our findings would also promote new strategies for the construction of efficient phototherapeutic molecules for the ablation of cancer cells, and facilitate the exploration of bioprobes for theranostic applications.
1H spectra were measured on Bruker ARX 400 NMR spectrometers using CD3OD as the deuterated solvent. High-resolution mass spectra (HRMS) were recorded on a Finnegan MAT TSQ 7000 Mass Spectrometer system operating in the MALDI-TOF mode. UV absorption spectra were recorded on a Milton Ray Spectronic 3000 array spectrophotometer. Steady-state fluorescence spectra were recorded on a Perkin Elmer LS 55 spectrometer. Fluorescence images were collected on an Olympus BX 41 fluorescence microscope. Laser confocal scanning microscopy images were collected on a Zeiss laser scanning confocal microscope (LSM7 DUO) and analyzed using ZEN 2009 software (Carl Zeiss).
:
2 v/v) as the eluting solvent to give a red brown powder of TVP (92 mg, 66% of yield). 1H NMR (400 MHz, DMSO-D6), δ (ppm): 9.04 (d, J = 6.4 Hz, 2H), 8.23 (d, J = 6.8 Hz, 2H), 8.03 (d, J = 16.4 Hz, 1H), 7.64 (d, J = 8.8 Hz, 2H), 7.34–7.40 (m, 5H), 7.10–7.18 (m, 6H), 6.94 (d, J = 8.4 Hz, 2H), 4.61 (t, J = 7.2 Hz, 2H), 3.47 (t, J = 8.2 Hz, 2H), 3.12 (s, 9H), 2.46 (t, J = 7.4 Hz, 2H). 13C NMR (100 MHz, DMSO-D6), δ (ppm): 153.58, 149.54, 146.17, 144.22, 141.05, 129.87, 129.79, 127.97, 125.43, 124.55, 123.24, 120.64, 120.47, 61.77, 56.32, 52.42, 24.18. ESI HRMS: calcd for C31H35N3 [M − 2Br]+: 449.2831, found: 449.2823.
A solution of 5-(4-(diphenylamino)phenyl)thiophene-2-carbaldehyde (71 mg, 0.2 mmol) and 1-(3-trimethylammoniopropyl)-4-methylpyridinium dibromide (71 mg, 0.2 mmol) was refluxed under nitrogen in dry ethanol catalyzed by a few drops of piperidine overnight. After cooling to room temperature, the solvent was removed by evaporation under reduced pressure. The residue was purified using a neutral aluminum oxide column using a DCM and methanol mixture (98
:
2 v/v) as the eluting solvent to give a red brown powder of TTVP (98 mg, 71% of yield). 1H NMR (400 MHz, CD3OD), δ (ppm): 8.78 (d, J = 6.8 Hz, 2H), 8.13–8.17 (m, 3H), 7.58–7.60 (m, 2H), 7.48 (d, J = 4.0 Hz, 1H), 7.40 (d, J = 4.0 Hz, 1H), 7.30–7.34 (m, 4H), 7.02–7.12 (m, 9H), 4.60 (t, J = 7.8 Hz, 2H), 3.52–3.56 (m, 2H), 3.20 (s, 9H), 2.51–2.59 (m, 2H). 13C NMR (100 MHz, CD3OD), δ (ppm): 155.99, 150.81, 150.19, 148.72, 145.26, 140.33, 136.72, 135.83, 130.81, 128.12, 126.40, 125.16, 125.04, 124.90, 123.80, 121.71, 64.04, 58.04, 54.13, 26.27. ESI HRMS: calcd for C35H37N3S [M − 2Br]+: 531.2708, found: 531.2693.
Footnote |
| † Electronic supplementary information (ESI) available: Compound characterization, PL spectra of TTVP in solvents with different polarities, plasma membrane-imaging of other cells, and morphological changes of the plasma membrane upon light irradiation. See DOI: 10.1039/c7sc04963c |
| This journal is © The Royal Society of Chemistry 2018 |