An activatable near-infrared fluorescent probe for methylglyoxal imaging in Alzheimer's disease mice

Yijing Dang a, Fengyang Wang a, Lingling Li a, Yi Lai a, Zhiai Xu *a, Zhi Xiong b, Ao Zhang b, Yang Tian a, Chunyong Ding *b and Wen Zhang *a
aSchool of Chemistry and Molecular Engineering, East China Normal University, Shanghai 200241, China. E-mail:;
bCAS Key Laboratory of Receptor Research, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai 201203, China. E-mail:

Received 22nd October 2019 , Accepted 9th December 2019

First published on 18th December 2019

Visual detection of the methylglyoxal (MGO) level in the brain is critical for understanding its role in the onset and progression of AD. Herein, we disclosed a NIR fluorescent probe, DBTPP, for detecting MGO by utilizing a thiadiazole-fused o-phenylenediamine moiety as a MGO-specific sensing unit. DBTPP exhibits a series of distinct advantages, such as NIR emission, high selectivity and sensitivity, excellent acid-stability, and a huge off–on ratio. The probe could accurately monitor both exogenous and endogenous MGO variations in SH-SY5Y cells. Besides, it was able to image the endogenous MGO in a transgenic AD mouse model successfully, suggesting the great potential of MGO as a biomarker for early AD diagnosis.

Alzheimer's disease (AD) is a neurodegenerative disease characterized by cognitive dysfunction and progressive memory loss. Although the detailed pathogenesis of AD is not yet clear, it is believed that this disease possesses the pathological features of deposition of amyloid-β peptide (Aβ) and neurofibrillary tangles (NFTs), as well as massive neuronal cell death.1 Accumulating evidence suggests that glycation is closely related to AD, and may contribute to both extensive protein cross-linking and oxidative stress in AD.2 Advanced glycation end products (AGEs) generated in the endogenous glycation process are also found to be associated with the formation of Aβ plaques and NFTs in AD.3 As the most potent glycation agent to form AGEs, methylglyoxal (MGO) is a highly reactive dicarbonyl by-product of the cellular metabolism, which has been demonstrated to be positively correlated with an increase of oxidative stress in AD.4 It was reported that the MGO levels of AD patients were twofold higher than those in controls and numerous studies have indicated that MGO and MGO-derived AGEs play key roles in the etiopathogenesis of AD.5,6 Accordingly, detection of the cellular MGO level in AD patients is critical for understanding the roles of MGO and MGO-derived AGEs in the onset and progression of AD as well as related anti-AD therapy and diagnosis.

Over the past several decades, liquid chromatography-mass spectrometry, high performance liquid chromatography and gas chromatography have been commonly used for detecting MGO in biological samples.7 However, these methods suffer from a series of tedious processing steps, such as cell lysis, which gives a poor accuracy of MGO detection. Therefore, the development of efficient methods for real-time detection of on-site MGO in living systems, especially in the brain, is urgently required but challenging. Owing to a series of advantages such as high sensitivity and selectivity, non-invasive visualization, and high spatiotemporal resolution, fluorescent imaging methods have been attracting great attention and are applied to biological MGO detection. To this end, several MGO fluorescent probes have been developed by attaching the o-phenylenediamine (OPD) moiety to various fluorophores.8–13 Nevertheless, these reported probes emit in the green region with the wavelength less than 550 nm, which may be disturbed by endogenous biomolecules, resulting in a poor collection efficiency of the fluorescence signal in deep imaging. In addition, most of the MGO probes designed require an electron-rich OPD moiety to capture MGO. However, this highly reactive moiety also reacts with other reactive carbonyl/oxygen species to perturb the detection of MGO in a pathological setting. Actually, all of these reported MGO fluorescent probes are limited to detecting MGO in a diabetic pathological model. In situ imaging of MGO in the AD model has never been reported as yet. Therefore, it will be challenging but highly demanding to develop NIR fluorescent probes with high specificity and sensitivity to monitor endogenous MGO in the AD animal models.

Inspired by the excellent optical properties of the donor–acceptor–donor (D–A–D) type architecture,14 we herein disclosed a novel near-infrared (NIR) fluorescent probe, namely 3,3′-(((5,6-diaminobenzo[c][1,2,5]thiadiazole-4,7-diyl)bis(4,1-phenylene))bis(oxy))bis(propan-1-ol) (DBTPP) for MGO detection. The electron-deficient thiadiazole-fused OPD (TOPD) skeleton was concurrently utilized as the MGO-specific sensor and the weak electron acceptor of D–A–D type fluorophores, and two hydroxyl alkyoxyl groups designated as the electron donors were linked to both terminals of TOPD via a benzene unit as a π electron bridge. When DBTPP reacts with MGO, the TOPD unit would transform into the tricyclic thiadiazole-fused quinoxaline as a stronger electron acceptor of the product 3,3′-(((6-methyl-[1,2,5]thiadiazolo[3,4-g]quinoxaline-4,9-diyl)bis(4,1-phenylene))bis(oxy))bis(propan-1-ol) (MTQPP). The resulting probe achieves high selectivity and sensitivity for MGO with maximum NIR fluorescence emission at around 650 nm. It could accurately monitor both exogenous and endogenous MGO variations in SH-SY5Y cells. Besides, it successfully visualizes the endogenous MGO in a transgenic AD mouse model. To the best of our knowledge, DBTPP is the first reported NIR fluorescent probe capable of specifically imaging MGO in an AD animal model ex vivo and in vivo.

The nucleophilic substitution reaction of 1 with 3-bromopropan-1-ol under K2CO3 produced the precursor 2, which underwent Fe/NH4Cl-mediated nitroreduction to afford DBTPP. Further reaction of DBTPP with MGO created the fluorescent product MTQPP (Scheme 1B). The structures of DBTPP and MTQPP were characterized by 1H-NMR, 13C-NMR, FT-IR and HR-ESI mass spectrometries (Fig. S1–S7, ESI). The fluorescence quantum yields of DBTPP and MTQPP were 6.8% and 12.4%, respectively (Table S1 and Fig. S8–S10, ESI). Their molar extinction coefficients were 14[thin space (1/6-em)]280 (375 nm) and 5462 L mol−1 cm−1 (500 nm), respectively (Table S1 and Fig. S11, ESI).

image file: c9cc08265d-s1.tif
Scheme 1 (A) Schematic illustration of D–A–D type fluorescent probe DBTPP for the detection of MGO in APP/PS1 and wild-type mouse brain. (B) Synthetic route of D–A–D type fluorescent probe DBTPP. (C) Photos of the solutions under sunlight and UV light, top and bottom, respectively. From left to right: DBTPP, DBTPP + MGO, and MTQPP.

The optical properties of DBTPP were inspected next through UV-vis and fluorescence spectral measurement. As shown in Fig. 1, DBTPP exhibited a single absorbance peak centred at 375 nm. Upon addition of MGO, the peak wavelength red-shifted to 500 nm with distinct color changes. When excited at 500 nm, negligible fluorescence emission was detected for DBTPP because of the weak electron push–pull strength, while the fluorescence intensity at the wavelength range from 550 nm to 850 nm was remarkably enhanced upon addition of MGO with color changes from pale green to red under UV irradiation, which was likely attributed to the MGO-induced formation of a large and more electron-deficient thiadiazole-fused quinoxaline motif as an enhanced electron acceptor. The absorption and fluorescence emission characteristics of the synthesized MTQPP were found to match essentially with those of DBTPP buffer solution treated with MGO. In addition, HPLC analysis showed that treating DBTPP in PBS buffer solution with MGO almost quantitatively provided a high-yield product, which had the same retention time as that of the synthetic MTQPP (Fig. S12, ESI). Collectively, these results suggest that the reaction of DBTPP with MGO in PBS buffer produces the quinoxaline product MTQPP.

image file: c9cc08265d-f1.tif
Fig. 1 (A) UV-vis absorbance and (B) fluorescence spectra of the DBTPP probe (25 μM) upon the addition of 0–60 μM MGO, insets: photographs of the solution under sunlight and UV light. (C) F.I. values at 650 nm as a function of various concentrations of MGO. Inset: Linear fitting plot of F.I.650nm and MGO concentration. (D) Fluorescence lifetime decay curves of DBTPP, DBTPP + MGO, and MTQPP. Buffer: 10 mM PBS (pH = 7.4, and 40% DMAC).

As depicted in Fig. 1C, a gradual enhancement of fluorescence intensity was observed along with the final concentration of MGO up to 60 μM. The fluorescence intensity at 650 nm increased linearly with MGO concentration in the range of 4 to 16 μM (R2 = 0.99952) with a detection limit of 262 nM (S/N = 3). The pH effect on the fluorescence response of DBTPP towards MGO was examined. The weak fluorescence intensity of DBTPP was stable in the range of pH 4.0–9.0 (Fig. S13, ESI). Upon the addition of MGO, the fluorescence intensity was significantly increased with slight variation throughout the tested pH range, indicating the wide pH tolerance with low background fluorescence. DBTPP showed a very short lifetime of 1.05 ns, while both DBTPP + MGO and MTQPP uniformly had a longer lifetime of 4.59 ns (Fig. 1D). DBTPP and MTQPP also showed good photostabilities (Fig. S14, ESI). To investigate the specificity of DBTPP towards MGO, we inspected the fluorescence response of DBTPP to MGO as well as a wide array of the potentially interfering substances, which were classified as reactive oxygen/nitrogen species (ROS/RNS: ˙OH, ONOO, NO, H2O2, O2˙ and 1O2), metal ions (Fe3+, Fe2+, and Cu2+), active biomolecules (Cys, Lys, and GSH), and active carbonyls (BA, FA, GO, MDA and OPA). All these substances showed negligible influences on the fluorescence of DBTPP, while only MGO induced significant fluorescence response (Fig. 2). Moreover, the probe still has good selectivity under the coexistence of interferents and MGO. These results demonstrate that DBTPP shows a high MGO selectivity, which could likely be attributed to the relatively low reactivity of electron-withdrawing TOPD. It is worth noting that TOPD responds to NO in acidic surroundings only, and no fluorescence response was observed in the neutral buffer solution,14 which may account for its selectivity for MGO versus NO. In addition, the intracellular level of MGO is at least 2 orders of magnitude larger than that of NO, which is another reason to distinguish these two molecules.10

image file: c9cc08265d-f2.tif
Fig. 2 Fluorescence changes (F650nm) of DBTPP (25 μM) before (−) and after (+) incubation with MGO in the presence of ROS/RNS, metal ions, active biomolecules and aldehydes, a–q (˙OH, ONOO, NO, H2O2, O2˙, 1O2, Fe3+, Fe2+, Cu2+, Cys, Lys, GSH, BA, FA, GO, MDA, and OPA). The concentration of the interferents was 500 μM, except that for NO it was 50 nM.

Both DBTPP and MTQPP exhibited more than 70% cell viabilities toward SH-SY5Y cells at 15 μM (Fig. S15, ESI). We next assessed the ability of DBTPP to image exogenous MGO in living SH-SY5Y cells by confocal laser scanning microscopy (CLSM) examination. SH-SY5Y cells were pre-incubated with 6 μM of the probe for 1.5 h, and then treated with different concentrations of MGO ranging from 0 to 8 μM for 1.5 h. In the absence of exogenous MGO, the CLSM image of the probe-loaded SH-SY5Y cells showed no obvious red fluorescence under excitation at 488 nm (Fig. 3). Upon incubation with 2 μM MGO, red fluorescence was observed in the cells, and the fluorescence intensity was gradually enhanced up to 5-fold with the increase of the MGO concentrations from 2 to 8 μM (Fig. 3 & Fig. S16, ESI). Encouraged by its excellent performance for imaging MGO, we next sought to testify whether DBTPP was capable of measuring endogenous MGO. The most important enzymatic detoxification pathway for MGO is the glyoxalase system, including glyoxalase 1 (GLO1) and glyoxalase 2 (GLO2), which is responsible for converting MGO to non-toxic D-lactate. It has been reported that inhibition of enzyme activity would increase the intracellular concentration of MGO.7 Therefore, endogenous MGO in SH-SY5Y cells could be induced with GLO1 inhibitor BHGD. As shown in Fig. S17 (ESI), in the absence of BHGD, no obvious red fluorescence was observed in SH-SY5Y cells co-incubated with DBTPP. In contrast, when pretreated with GLO1 inhibitor BHGD for 6 h, SH-SY5Y cells stained with DBTPP displayed significantly enhanced red fluorescence, which gradually increased along with the increase of BHGD concentration (15, 30 and 60 μM).

image file: c9cc08265d-f3.tif
Fig. 3 Confocal images of SH-SY5Y cells incubated with 6 μM DBTPP at different MGO concentrations (0, 2, 4, and 8 μM) for 1.5 h. Red channel: λex = 488 nm, and λem = 600–800 nm. Scale bar = 25 μm.

After verifying the capability of DBTPP to track exogenous and endogenous MGO in vitro, we further examined its potential for in situ detection of MGO in an AD mouse model with modified genotypes of APPSWe and PSEN1D9 (APP/PS1). Around six-month-old male APP/PS1 mice were selected as the AD group, and wild-type male mice at the same age were used as the control group (wild-type). All these mice were sacrificed by cervical dislocation without any introduction of interferents. The brain tissues were frozen, sectioned, and subsequently stained with 10 μM of DBTPP and 5 μg mL−1 of DAPI for 2 h and 5 min, respectively. As illustrated in Fig. 4A, DBTPP stained brain slices of the wild-type mice group displayed marginally detectable red fluorescence. The image of the brain slices from APP/PS1 mice stained with DBTPP exhibited a notably brighter red fluorescence, which was about 4-fold boosted compared to that in the wild-type mice group (Fig. 4B). To further determine if the red fluorescence enhancement was ascribed to MGO endogenously generated in the brain of AD mice, the brain slices from APP/PS1 mice were pretreated with MGO scavenger GSH before staining with DBTPP (APP/PS1 + GSH). As expected, a dramatic decrease of red fluorescence intensity was observed in the GSH-pretreated APP/PS1 mice group. These results suggested that the AD mice with APP/PS1 genotypes possessed higher MGO concentration than wild-type mice. To explore the application of DBTPP for in vivo imaging, MTQPP was intracranially injected and fluorescence pictures were acquired at 3, 6, 9, 12, and 24 h after administration. The fluorescence intensity in the brain of nude mice reached a maximum at 12 h after injection. Meanwhile, MTQPP gradually spread to the liver and kidneys for metabolization and elimination, as demonstrated in ex vivo fluorescence images (Fig. S18, ESI). Then, probe DBTPP was injected into the brain hippocampus of APP/PS1 and wild-type mice through intracranial injection. As shown in Fig. 4C and Fig. S19 (ESI), compared to the wild-type group, the fluorescence intensity was obviously higher in the brains of the AD mice when in vivo monitored at 1.5 and 4.5 h after injection, while the fluorescence intensity in the AD group decreased to the level of the wild-type group at 6 h probably due to MTQPP being cleared out from the brain. All these results indicated the high potential of probe DBTPP for in vivo MGO imaging in AD mouse models.

image file: c9cc08265d-f4.tif
Fig. 4 (A) Fluorescence imaging demonstration of brain slices through staining with DAPI and DBTPP in wild-type, APP/PS1, and APP/PS1 + GSH groups. Scale bar = 3000 μm. Blue channel: λex = 377 ± 50 nm, and λem = 447 ± 60 nm. Red channel: λex = 469 ± 35 nm, and λem = 647 ± 57 nm. (B) The ratio of fluorescence intensities of the red channel and blue channel in 10 sections for wild-type and APP/PS1 groups. (C) In vivo fluorescence MGO imaging with DBTPP at 1.5 h. λex = 488 nm, and λem = 650–800 nm.

In general, we have developed an activatable fluorescent probe DBTPP for imaging biological MGO by utilizing the electron-deficient TOPD moiety as a MGO-specific sensing unit. DBTPP is the first reported NIR fluorescent probe with high selectivity for MGO over other interfering substances. Besides, it also displays several other advantages including a large Stokes shift, a large off–on ratio with minimal auto-fluorescence, and excellent acid-stability. The probe could accurately monitor both exogenous and endogenous MGO variations in SH-SY5Y cells. More importantly, it successfully visualizes the endogenous MGO in a transgenic AD mouse model in vivo, suggesting the great potential of MGO as a biomarker for early AD diagnosis. Taken together, DBTPP might be a promising MGO probe for understanding the role of MGO in the onset and progression of AD as well as anti-AD therapy and diagnosis.

The financial support by the National Natural Science Foundation of China (No. 21675055, 81402790, 21775046, and 21635003) is greatly appreciated. Also, the financial support by ‘‘Personalized Medicines-Molecular Signature-based Drug Discovery and Development” of Strategic Priority Research Program of the Chinese Academy of Sciences (No. XDA12020320), the National Science & Techonolgy Major Project “Key New Drug Creation and Manufacturing Program”, China (2018ZX09711002), and the State Key Laboratory of Drug Research (SIMM1903KF-13) of the Chinese Academy of Sciences is appreciated.

Conflicts of interest

There are no conflicts to declare.

Notes and references

  1. B. De Strooper and E. Karran, Cell, 2016, 164, 603–615 CrossRef CAS PubMed.
  2. (a) R. J. Castellani, P. L. R. Harris and L. M. Sayre, Free Radicals Biol. Med., 2001, 31, 175–180 CrossRef CAS; (b) Y. Koriyama, A. Furukawa, M. Muramatsu, J. I. Takino and M. Takeuchi, Sci. Rep., 2015, 5, 13313 CrossRef CAS PubMed.
  3. H. J. Lüth, V. Ogunlade, B. Kuhla, R. Kientsch-Engel, P. Stahl, J. Webster and G. Münch, Cereb. Cortex, 2004, 15, 211–220 CrossRef PubMed.
  4. S. Koike, C. Ando, Y. Usui, Y. Kibune, S. Nishimoto, T. Suzuki and Y. Ogasawara, Brain Res. Bull., 2019, 144, 164–170 CrossRef CAS PubMed.
  5. (a) S. S. More, A. P. Vartak and R. Vince, ACS Chem. Neurosci., 2012, 4, 330–338 CrossRef PubMed; (b) A. Wong and H. J. Lüth, Brain Res., 2001, 920, 32–40 CrossRef CAS PubMed.
  6. B. Kuhla, H. J. Lüth, D. Haferburg, K. Boeck, T. Arendt and G. Münch, Ann. N. Y. Acad. Sci., 2005, 1043, 211–216 CrossRef CAS PubMed.
  7. (a) A. C. McLellan, S. A. Phillips and P. J. Thornalley, Anal. Biochem., 1992, 206, 17–23 CrossRef CAS PubMed; (b) F. W. Chaplen, W. E. Fahl and D. C. Cameron, Anal. Biochem., 1996, 238, 171–178 CrossRef CAS PubMed; (c) I. Nemet, L. Varga-Defterdarovic and Z. Turk, Clin. Biochem., 2004, 37, 875–881 CrossRef CAS PubMed; (d) A. Moraru, J. Wiederstein, D. Pfaff, T. Fleming, A. K. Miller, P. Nawroth and A. A. Teleman, Cell Metab., 2018, 27, 926–934 CrossRef CAS PubMed.
  8. H. Wang, Y. Xu, L. Rao, C. T. Yang, H. Yuan, T. Gao and C. Liu, Anal. Chem., 2019, 91, 5646–5653 CrossRef CAS PubMed.
  9. S. T. Wang, Y. Lin, C. D. Spicer and M. M. Stevens, Chem. Commun., 2015, 51, 11026–11029 RSC.
  10. T. Wang, E. F. Douglass Jr, K. J. Fitzgerald and D. A. Spiegel, J. Am. Chem. Soc., 2013, 135, 12429–12433 CrossRef CAS PubMed.
  11. T. Tang, Y. Zhou, Y. Chen, M. Li, Y. Feng, C. Wang and X. Zhou, Anal. Methods, 2015, 7, 2386–2390 RSC.
  12. M. Yang, J. Fan, J. Zhang, J. Du and X. Peng, Chem. Sci., 2018, 9, 6758–6764 RSC.
  13. W. Zhang, F. Zhang, Y. L. Wang, B. Song, R. Zhang and J. Yuan, Inorg. Chem., 2017, 56, 1309–1318 CrossRef CAS PubMed.
  14. (a) F. Wang, S. Yu, Z. Xu, L. Li, Y. Dang, X. Xu, Y. Luo, Z. Cheng, H. Yu, W. Zhang, A. Zhang and C. Ding, Anal. Chem., 2018, 90, 7953–7962 CrossRef CAS PubMed; (b) L. Teng, G. Song, Y. Liu, X. Han, Z. Li, Y. Wang and W. Tan, J. Am. Chem. Soc., 2019, 141, 13572–13581 CrossRef CAS PubMed.


Electronic supplementary information (ESI) available. See DOI: 10.1039/c9cc08265d

This journal is © The Royal Society of Chemistry 2020