During the last decade, theranostics [1, 2], which integrates both diagnostics and therapy, has become more and more popular in personalized medicine because it offers a platform to monitor and track the therapeutic progress by diagnostic agents instantaneously. Various systems including biocompatible inorganic nanoparticles (mesoporous silica nanoparticles [3], quantum dots [4], gold nanoparticles [5], etc.) and organic/polymeric nanoparticles [6] have been widely used as theranostic agents to improve the therapeutic efficacy and safety during cancer treatment. Among all the imaging techniques, near-infrared (NIR) fluorescence imaging [7, 8] has received considerable attention due to its deep penetration, low auto-fluorescence, and minimum photo-damage. While for therapeutic agents, Pt-based drugs [9] are still the mostwidely used chemotherapeutics for sarcoma, carcinoma, etc. Therefore, the combination of near-infrared fluorescence imaging and Pt-based drugs represents an effective way to prepare theranostic agents [10-14].

Supramolecular coordination complexes (SCCs) [15-24] are a type of kinetically reversible and thermodynamically favored twodimensional (metallacycles [25-30]) or three-dimensional (metallacages [31-40]) structures formed by metal-coordination-driven self-assembly. So far, various shapes of SCCs have been developed and the interest on SCCs has gradually shifted from their appealing structures to their advanced functions and applications. The incorporation of fluorescence to SCCs to prepare emissive SCCs [41-47] opens a new window for their biological applications in bio-sensors, cell imaging, controlled release and so on. Moreover, due to the potential anticancer activity of platinum ligands, platinum-based emissive SCCs could be used as theranostic agents for cancer treatment [48-52]. For example, Stang et al. reported a type of theranostic supramolecular nanoparticles formed by covalent bond-linked fluorescent metallacycle-cored polymers [48]. Based on the chemotherapy of platinum ligands and fluorescence imaging and photodynamic therapy of BODIPY units, Cook, Huang and coworkers prepared two emissive trigonal metallacycles which exhibited synergistic anticancer effect and superior anticancer activity towards cisplatin-resistant cells [49]. However, almost all these fluorophores emit in the visible region (the maximum emission is normally less than 700 nm), which is far from real applications. Therefore, NIR emissive SCCs are urgently needed to be developed for cancer theranostics.

Herein, by the introduction of two morpholine groups as electron donors to a cyanostilbene-based dipyridyl ligand, the formation of D-π-A structures [53-55] will drive the emission of the derived ligand 6 to the NIR region (λ_em =735 nm). Three Pt(Ⅱ) metallacycles 1, 2 and 3 were prepared by the metal-coordinationdriven self-assembly of ligand 6 with different platinum(Ⅱ) acceptors and corresponding carboxylic ligands. These metallacycles show NIR emission derived from their organic precursor, making them good contrast agents for cell imaging. Moreover, due to the potential anti-cancer activity of platinum(Ⅱ) acceptors, they also exhibit cytotoxicity towards seven different cancer cell lines. It is worth mentioning that metallacycle 1 shows the better anti-cancer activities compared with the other two metallacycles for all the test cell lines, and the IC₅₀ values of metallacycle 1 are even lower than those of cisplatin, indicating that metallacycle 1 may serve as a broad-spectrum theranostic agents for cancer treatment.

The synthetic procedures for ligand 6 are showed in Scheme 1. The oxidation of 1, 4-dibromo-2, 5-dimethylbenzene offered 2, 5- dibromoterephthalaldehyde 4, which was further used to react with 4-pyridineacetonitrile hydrochloride under basic condition to give intermediate 5. Ligand 6 was prepared via a classical Buchwald–Hartwig coupling reaction of 5 with morpholine in 48% isolated yield. Single crystal of ligand 6 (CCDC No. 1918663) obtained by slow evaporation in methanol/dichloromethane (1:2, v/v) clearly confirmed its structure. The pyridyl rings twisted away from the central phenyl plane, and the dihedral angles between the pyridyl rings and the phenyl plane are 23.16°. The dihedral angles between the middle phenyl plane and the morpholine rings are 49.12°, in order to reduce the steric hindrance. Although twisted, the nearly linear configuration of the two pyridyl rings in 6 makes it act as 180° dipyridyl ligands for metal-coordination-driven selfassembly.

Metallacycles 1, 2 and 3 were prepared by standard procedures of metal-coordination-driven self-assembly, which was depicted in Supporting information in detail. The formation of metallacycles 1, 2 and 3 was supported by multinuclear NMR (³¹P NMR and ¹H NMR) analysis and electrospray ionization time-of-flight mass spectrometry (ESI-TOF-MS). Partial ³¹P{¹H} NMR spectra of 1 and 2 exhibited sharp singlets with concomitant ¹⁹⁵Pt satellites at 13.00 ppm for 1, and -0.52 ppm for 2, consistent with a single phosphorous environment (Figs. 1a and b). While metallacycle 3 showed two coupled doublets at 5.66 ppm and 0.16 ppm of approximately similar intensities with concomitant ¹⁹⁵Pt satellites (Fig. 1c), because of the two distinct phosphorus environments. Correspondingly, obvious downfield chemical shift changes were found for α-pyridyl protons H_a and β-pyridyl protons H_b for all the metallacycles due to the coordination of nitrogen atoms to Pt(Ⅱ) ions. All the results are consistent with previously reported results[42-45] and confirm the successful formation of discrete metallacycles.

ESI-TOF-MS supported the formation of metallacycles 1, 2 and 3 based on their stoichiometries. In the mass spectra, different distinct peaks with charge states (from 3⁺ to 5⁺ for 1, from 4⁺ to 6⁺ for 2, from 2⁺ to 3⁺ for 3) were found because of the loss of triflate counterions (OTf^-), and the pattern of the peaks matches well with their calculated results, indicating the formation of metallacycles 1, 2 and 3. It is worth noting that the self-assembly of ligand 6 and cis-Pt(PEt₃)₂(OTf)₂ gives a [6 + 6] hexagonal structure, different from the previously reported [3 + 3] assemblies formed by linear dipyridyl ligands and cis-Pt(PEt₃)₂(OTf)₂ [49], suggesting that the flexibility of cyanostilbene-based dipyridyl ligand plays an important role on the shapes of the formed metallacycles.

The UV–vis absorption and emission spectra of ligand 6 and metallacycles 1, 2 and 3 in 1% DMSO/H₂O are shown in Fig. 2. Ligand 6 exhibits two broad absorption peaks centered at 362 nm and 519 nm with molar absorption coefficients (ε) of 1.08 × 10⁴ and 0.52 ×10⁴ L mol^-1 cm^-1, respectively. Metallacycle 1 shows three intense peaks centered at 260 nm, 356 nm and 473 nm, with ε = 1.82 × 10⁵, 8.66 × 10⁴ and 2.50 × 10⁴ L mol^-1 cm^-1, respectively, due to the integration of the absorption of 60° diplatinum(Ⅱ) acceptor and ligand 6. Two distinct peaks at 365 nm and 530 nm were observed for metallacycle 2 and at 364 nm and 526 nm for metallacycle 3, with ε = 9.65 ×10⁴, 5.59 × 10⁴ and 2.73 × 10⁴, 1.55 ×10⁴ L mol ^-1 cm^-1, respectively.

Ligand 6 and the metallacycles 1, 2 and 3 display intense emission bands centered at 735 nm, 771 nm, 744 nm and 748 nm in 1% DMSO/H₂O, respectively. The formation of metallacycles planarizes the structure of ligand 6, leading to the red shifts from the ligand to the metallacycles. In the solid state, ligand 6 shows one intense emission peak centered at 697 nm, while metallacycles 1, 2 and 3 exhibit moderate emission centered at 794 nm, 795 nm and 821 nm, respectively. Red shifts were also observed from ligand 6 to the metallacycles in the solid state. The quantum yields of the four compounds in solution (1% DMSO/H₂O) and in the solid state were also determined. It is worth noting that the quantum yield of ligand 6 increases from 6.90% in solution to 11.63% in the solid state, owing to its aggregation-induced enhanced emission (AIEE) property [56]. The quantum yields of metallacycles 1, 2 and 3 are less than 0.5% both in solution and in the solid state, indicating that the metal-coordination plays an important role on the emission of the fluorophores. All the emission bands locate at the NIR region, indicating that these metallacycles can be used as contrast agents for NIR fluorescence imaging.

The stability of the ligand 6 and metallacycles 1, 2 and 3 in RPMI 1640 cell culture medium was studied by dynamic light scattering (DLS) before the biological experiments (Fig. S21 in Supporting information). The size of nanoparticles formed by all the compounds remained unchanged within two days, indicating that these compounds are stable in cell culture medium. Then the use of ligand 6 and metallacycles 1, 2 and 3 for cell imaging was explored. Inverted fluorescence microscopy (IFM) was used to study the cellular uptake and intracellular localization of the compounds in cells. Cells were stained by DAPI, FITC and our compounds simultaneously and the images were taken after 12 h of incubation (Fig. 3). Bright red fluorescence derived from the synthesized compounds was observed, indicating that both ligand 6 and metallacycles 1, 2 and 3 can be used for cell imaging. Moreover, the fluorescence intensity of metallacycle 1 in cytoplasm and the cell nucleus is much more brighter than that of metallacycles 2 and 3, which is consistent with the flow cytometry (FCM) results, indicating that cells show better uptake efficiency for metallacycle 1 compared with metallacycles 2 and 3. The cytotoxicity of all the compounds towards seven different cancer cell lines was also evaluated by MTT assay. Ligand 6 shows negligible toxicity for all the test cells, indicating that it is biocompatible. The IC50 values of metallacycles 1, 2 and 3 and cisplatin were collected in Table 1. All the metallacycles show cytotoxicity towards the tested cancer cells. Metallacycle 1 shows better anticancer activities than its analogous metallacycles 2 and 3 and its IC₅₀ values are even lower than that of cisplatin for all the test cells. This was also evidenced by Annexin V-FITC/PI assay to distinguish viable cells from dead cells using flow cytometry. For example, the populations of the early apoptotic, late apoptotic, and necrotic A549 cells were 58.0%, 36.1%, and 0.23%, respectively, after the treatment by metallacycle 1 for 48 h, compared with 1.69%, 21.2% and 0.08% after the treatment by cisplatin for 48 h. The cell apoptosis rate increased significantly from 23.0% by cisplatin to 94.3% by metallacycle 1. These results are consistent with the better cellular uptake efficiency of metallacycle 1, indicating that it is much more toxic than cisplatin and may be used as a therapeutic agent for cancer treatment.

In conclusion, a cyanostilbene-based dipyridyl ligand with NIR emission was prepared by the incorporation of D-π-A structures. It was further used to prepare three platinum(Ⅱ) metallacycles via metal-coordination-driven self-assembly. Based on the NIR emission of the dipyridyl ligands and the anti-cancer activities of the platinum(Ⅱ) ligands, these metallacycles can be used as theranostic agents for NIR fluorescence imaging and cancer therapy. Among all the metallacycles, trigonal metallacycle 1 exhibits the best anti-cancer activities, which are even better than the clinically used cisplatin. This study offers a new type of NIR emissive theranostic agents, which will pave the way of imagingguided therapy, drug delivery, and cancer theranostics.

Acknowledgments

This work was supported by the National Natural Science Foundation of China (No. 21801203 to M. Zhang) and The Key Research and Development Program of Shaanxi Province (No. 2019KW-019 to M. Zhang, No. 2019KW-066 to W. Shi). M. Zhang is thankful for start-up funds from Xi'an Jiaotong University. We thank Prof. Weifeng Bu at Lanzhou University for fluorescence measurements and Prof. Xiaopeng Li at University of South Florida for ESI-TOF-MS measurements. We also acknowledge Dr. Gang Chang and Yu Wang at Instrument Analysis Center and Dr. Aqun Zheng and Junjie Zhang at the Provincial Demonstration Center for Experimental Chemistry Education of Xi'an Jiaotong University for NMR experiments.

Appendix A. Supplementary data

Supplementary material related to this article canbefound, in the online version, at doi:https://doi.org/10.1016/j.cclet.2019.07.043.