Near-infrared imaging for visualizing the synergistic relationship between autophagy and NFS1 protein during multidrug resistance using an ICT–TICT integrated platform

Drug resistance is a major challenge for cancer treatment, and its identification is crucial for medical research. However, since drug resistance is a multi-faceted phenomenon, it is important to simultaneously evaluate multiple target fluctuations. Recently developed fluorescence-based probes that can simultaneously respond to multiple targets offer many advantages for real-time and in situ monitoring of cellular metabolism, including ease of operation, rapid reporting, and their non-invasive nature. As such we developed a dual-response platform (Vis-H2S) with integrated ICT–TICT to image H2S and viscosity in mitochondria, which could simultaneously track fluctuations in cysteine desulfurase (NFS1 protein and H2S inducer) and autophagy during chemotherapy-induced multidrug resistance. This platform could monitor multiple endogenous metabolites and the synergistic relationship between autophagy and NFS1 protein during multidrug resistance induced by chemotherapy. The results indicated that chemotherapeutic drugs simultaneously up-regulate the levels of NFS1 protein and autophagy. It was also found that the NFS1 protein was linked with autophagy, which eventually led to multidrug resistance. As such, this platform could serve as an effective tool for the in-depth exploration of drug resistance mechanisms.


Introduction
Chemotherapy, the main method for treating malignant tumours, usually produces signicant drug resistance during the later stages of treatment. 1This reduces therapeutic effects and even results in resistance to alternative drugs that use different mechanisms, which is termed multidrug resistance (MDR). 2 MDR can interfere with the intake and accumulation of drugs in tumour cells, reducing the ability of anticancer drugs to induce cell death, resulting in drug inactivation and degradation, and the activation of anti-apoptosis and antioxidant defence, as well as DNA repair and replication.These factors are the main causes of the failure of chemotherapy. 3Common tumour MDR detection methods include MTT assays, drug sensitivity tests, MDR gene and pathway detection, and highthroughput screening technologies. 4However, using these methods makes it impossible to achieve in situ, real-time measurements and accurate detection of MDR in tumour cells.The methods are also expensive and time-consuming, which signicantly limits their application for the evaluation of MDR mechanisms.
Confocal uorescence imaging requires simple operation and exhibits high sensitivity and temporal and spatial resolution and can provide in situ and real-time analysis. 5However, uorescence-based sensors based on traditional sensing mechanisms have only one recognition site and can only recognize one analyte due to their single mode sensing mechanism. 6These are unsuitable for evaluating multiple interconnected biomolecules in complex living systems.To overcome this deciency, uorescence-based sensors have emerged that use two or more sensing mechanisms. 7,8hemotherapeutic drugs which increase the amounts of reactive oxygen species (ROS) (such as doxorubicin and cisplatin) can kill tumour cells.However, in addition they induce many factors that can result in MDR, including the activation of autophagy 9 (caused by excessive ROS, which then prevents cancer cell apoptosis by degrading the ROS) and upregulating the levels of cysteine desulfurase 10 (NFS1 protein, which can catalyse cysteine to produce endogenous hydrogen sulde and consume ROS, which prevents cancer cell apoptosis).However, the synergistic relationship between autophagy and NFS1 protein overexpression via ROS-based chemotherapeutic drugs has not yet been reported.
To achieve the in situ imaging of the synergistic relationship between autophagy and reductive protein overexpression, a probe with mitochondrial anchoring ability with a dualresponse toward viscosity (changes in mitochondrial viscosity can indicate autophagy 11 ) and H 2 S (mitochondrial H 2 S uctuations may indicate the NFS1 protein level 12 ) is required.With this research, a cyanine dye QCy7 with mitochondrial anchoring ability was used as the parent uorophore 13 and 2-iodobenzoate was used as the specic recognition site for H 2 S (Vis-H 2 S).When Vis-H 2 S reacts with H 2 S, intramolecular charge transfer (ICT) is activated and the uorescence emission wavelength red-shis from 492 nm to 687 nm, allowing it to be used to monitor changes in the H 2 S concentration, while the double bond contained in QCy7 can undergo twisted intramolecular charge transfer (TICT), resulting in a signicant decrease in the uorescence intensity, but this can be inhibited by a high-viscosity environment.Therefore, when the viscosity increased from 0.903 cp to 965 cp, the uorescence intensity at 492 nm increased 110-fold.This enabled the monitoring of environmental viscosity.Using the ICT-TICT dual-response strategy, Vis-H 2 S could simultaneously respond to mitochondrial viscosity and H 2 S uctuations.Thus, it can be used to explore the synergistic relationship between autophagy and the NFS1 protein.

Design and synthesis of Vis-H 2 S
As mentioned above, the "AND" logic gate-based "ICT-TICT dual-response mechanism" probe (Vis-H 2 S) requires the coexistence of two analytes (viscosity and H 2 S) and as such can monitor the synergistic relationship between autophagy and NFS1 protein.As shown in Scheme 1a and b, Vis-H 2 S generated NIR uorophores under the simultaneous activation by two agents, i.e., H 2 S and viscosity.In a low viscosity environment, weak NIR uorescence of Vis-H 2 S was observed in the presence of H 2 S, while in the presence of high viscosity (autophagyregulates mitochondrial viscosity) and the absence of H 2 S, Vis-H 2 S emits remarkable green uorescence.
If, however, both high viscosity and H 2 S are present, it emits strong near-infrared uorescence.For normal cells, H 2 S is present at a low level, and mitochondria have a high viscosity due to the presence of many proteins (resulting in strong green uorescence), but H 2 S and mitochondrial viscosity are easily destroyed by hypoxia and autophagy, respectively.Increased cell hypoxia can lead to the overexpression of reducing proteins such as NFS1 protein which results in a strong NIR uorescence response.On the other hand, autophagy reduces the viscosity of the mitochondria, resulting in minimal uorescence from either the red or green channel.As such, MDR produced by chemotherapy can lead to a simultaneous increase in autophagy and hypoxia, resulting in only weak NIR uorescence from the system.The response mechanism is provided in Scheme 1c and d.Vis-H 2 S, with a molecular rotor structure (two symmetric vinyl-coupled indolium moieties), enhanced the probe uorescence in a high-viscosity environment due to restricted intramolecular rotation inhibiting nonradiative twisted intramolecular charge transfer (TICT [14][15][16] ).The nucleophilic reaction of the ester with HS − releases QCy7 in an anionic form, and the negative charge is delocalized to form quinone QCy7a or QCy7-b.
Vis-H 2 S was synthesized according to Scheme S1, † and characterization data are provided in Fig. S1-S4.† The reaction mechanism between Vis-H 2 S and H 2 S was veried by measuring the relative molecular mass before and aer the reaction using HR-MS.As shown in Fig. S5, † when Vis-H 2 S was reacted with H 2 S, a peak appeared at m/z = 461.2098,representing the reaction product and the peak at m/z = 346.1212(representing the reactant) disappeared.We concurrently employed 1 H-NMR spectroscopy to conduct a mechanistic analysis of Vis-H 2 S's response to H 2 S. The results are illustrated in Fig. S6, † and exhibit signicant chemical shi variations in the methyl hydrogens (H a , H b , H c and H d ) on the indole moiety of the probe.Simultaneously, the signals corresponding to the hydrogens from the iodinated benzene disappear, substantiating the validity of the predicted mechanism underlying the response of Vis-H 2 S's to H 2 S. To further investigate the response of Vis-H 2 S and QCy7 to viscosity, a PBS-glycerol solvent system was employed to monitor the spectral response of Vis-H 2 S to viscosity changes (Fig. S7 and S8 †).Vis-H 2 S and QCy7 exhibited a nearly 110-fold and 22-fold increase in uorescence, respectively, in response to the increase in solution viscosity from 0.903 cP to 965 cP.The relative uorescence intensity of Vis-H 2 S and QCy7 exhibited a linear relationship with viscosity (R 2 = 0.985 for Vis-H 2 S and R 2 = 0.984 for QCy7).The above results conrmed that Vis-H 2 S and QCy7 exhibit excellent responses to viscosity and can be used to detect viscosity changes during in vivo and in vitro experiments.Additionally, we delved into the polarity responsiveness of QCy7 by examining its behavior in PBS-THF mixtures with different volume fractions (f w ) of PBS, spanning the range from 10% to 100%.As shown in Fig. S9, † there was minimal change in uorescence intensity, indicating that QCy7 is not sensitive to polarity.
2-Iodobenzoate was used to regulate the "push-pull" electronic effects (intramolecular charge transfer, ICT) of Vis-H 2 S so that it could specically respond to changes in H 2 S. We initially examined the absorption and uorescence spectra of Vis-H 2 S in response to H 2 S. The results revealed that Vis-H 2 S initially exhibited low levels of absorption and uorescence intensity, but a signicant increase was observed upon the addition of H 2 S (Fig. S10a and b †).We also investigated the response of Vis-H 2 S to H 2 S in environments with different viscosities (PBS, pH = 7.4, containing 0%, 50%, and 70% glycerol).The ability of Vis-H 2 S to simultaneously respond to H 2 S and viscosity was also evaluated.As shown in Fig. S10a † and 1, upon increasing the H 2 S concentration in the absence of glycerol, Vis-H 2 S vs. H 2 S exhibited a "turn-on" response, with an approximately 20-fold increase in the uorescence intensity at 687 nm.Upon increasing the glycerol concentration, a ratiometric sign was found with increasing concentrations of H 2 S (Vis-H 2 S exhibited a decreasing intensity at l em = 492 nm and an increasing intensity at l em = 687 nm).Upon increasing the glycerol concentration, I green and I red both increased signicantly.The ratio signal also exhibited a good linear relationship at 50% and 70% glycerol when the H 2 S concentration was lower than 75 mM.From Fig. 1e and f the correlation coefficients were 0.982 and 0.989, respectively, and the limit of detection (LOD) values were calculated to be 45 nM and 50 nM, respectively, according to 3s/ k (where s is the standard deviation of the blank (n = 11) and k is the slope for the range of linearity).The above experiments conrmed the dual-response of Vis-H 2 S to viscosity and H 2 S. Kinetic experiments indicated that when using 70% glycerol, the reaction between Vis-H 2 S and H 2 S (100 mM) reached a plateau within 60 min, indicating that Vis-H 2 S could rapidly recognize H 2 S (Fig. S10f †).Eighteen different interferents were selected to evaluate the selectivity of Vis-H 2 S. As shown in Fig. S10g, † these interferents did not change the uorescence intensity of Vis-H 2 S in either a high-viscosity (glycerol) or lowviscosity (PBS) environment.These results clearly indicate that Vis-H 2 S exhibits minimal interference and can be used for in situ imaging of cellular viscosity.Finally, we evaluated the effect of pH on the response of Vis-H 2 S to H 2 S. As shown in Fig. S10c, † for pH over a range from 4-8, there was no obvious change in the uorescence signal before or aer the simultaneous action of Vis-H 2 S with H 2 S and glycerol.In addition, the interference of cell viscosity stimulators (temperature and monensin 17 ) on the uorescence intensity of Vis-H 2 S was investigated.The results in Fig. S10d † show that neither temperature nor monensin interfered with the uorescence intensity of Vis-H 2 S, which provides an excellent basis for subsequent cell experiments.According to the literature, 18 acylphenol can serve as a recognition moiety for carboxylesterase-2, which is structurally similar to our recognition domain.To exclude the potential interference of carboxylesterase-2 in this study, we conducted photostability experiments of Vis-H 2 S in response to carboxylesterase-2, as shown in Fig. S10f.† The results indicate that over time, there was no change in uorescence intensity, suggesting that carboxylesterase-2 has no interference with H 2 S.
With the above results in hand, we evaluated the in situ imaging capability of Vis-H 2 S for H 2 S in cells.First, different concentrations of Vis-H 2 S were incubated with HepG2 cells for 24 h.The MTT assay results indicated that the cell viability remained above 80% and the IC 50 was 63 mM, which indicates that Vis-H 2 S was not cytotoxic and could be used for cell imaging (Fig. S11 †).To accurately image autophagy and NFS1 protein simultaneously, the probe needs to exhibit mitochondrial-anchoring ability.According to the literature, as an organelle with a double-membrane structure, the negative charge of the inner mitochondrial membrane enables positively charged molecules to be specically enriched and anchored in the mitochondria. 19Since the indole group of the Vis-H 2 S probe is positively charged, the intracellular distribution should be targeted to the mitochondria.To this end, Vis-H 2 S was mixed with a commercial mitochondrial localization dye, 100 nM Mito-Tracker Red, a commercial lysosome localization dye, 100 nM Lyso-Tracker Red, and a commercial endoplasmic reticulum localization dye, 100 nM ER-Tracker Red.These were then co-incubated with HepG2 cells for 30 min and imaged using confocal microscopy.As shown in Fig. S12 †, Vis-H 2 S was mainly distributed in the mitochondria (Pearson colocalization coefficient was 0.94), while the Pearson colocalization coefficients in the lysosome and endoplasmic reticulum were 0.41 and 0.33, respectively, indicating that Vis-H 2 S exhibits good mitochondrial localization and could detect the response to viscosity and H 2 S in the mitochondria during cellular experiments.
We then investigated the ability of Vis-H 2 S (10 mM) to respond to exogenous and endogenous H 2 S and cell viscosity.The results are given in Fig. S13.† First cells were treated with sulydryl scavenger N-ethylmaleimide 20 (NEM, 0.5 mM) and then incubated with different concentrations of H 2 S for 40 min.The experimental results indicated that upon increasing the NaHS (H 2 S donor) concentration (0, 5, 10, or 20 mM), I green gradually decreased and I red gradually increased, indicating that Vis-H 2 S could detect exogenous H 2 S in cells.Compared with the control group, the uorescence intensity was signicantly reduced for cells in the presence of NEM (0.5 mM).This was because of the removal of endogenous H 2 S by NEM.Then, in the absence of H 2 S due to the addition of NEM (reduced interference by H 2 S), intracellular viscosity levels were regulated using monensin and different cell incubation temperatures (25 °C and 4 °C; a lower temperature increases the intracellular viscosity).As shown in Fig. S14, † upon decreasing the temperature and adding monensin, I green increased signicantly, while I red was almost negligible.These experiments conrm that Vis-H 2 S can readily respond to cell viscosity.
Next, we evaluated the dual-response of Vis-H 2 S to H 2 S and autophagy at the cellular level.Previous studies have shown that during autophagy, mitochondria are wrapped by vesicles to form autophagosomes that then fuse with lysosomes to form autolysosomes. 21 This allows mitochondria and their contents to be degraded, resulting in reduced mitochondrial viscosity. 22s shown in Fig. S15, † compared with the control group and NEM group, aer cells were treated with autophagy inducer rapamycin (100 nM, mTORC1 complex inhibitor 23 ), I green was negligible, indicating that autophagy signicantly reduced the mitochondrial viscosity.In addition, different concentrations of NaHS (5, 10, or 20 mM) did not affect the I green , but I red gradually increased with an increase in H 2 S concentration, indicating that Vis-H 2 S could generate a dual response to autophagy and H 2 S.
To further evaluate the ability of Vis-H 2 S to monitor autophagy and H 2 S in real-time, confocal real-time imaging analysis was performed on HepG2 cells stained with Vis-H 2 S at different times and under cell culture conditions.The results are shown in Fig. 2a-e.The uorescence of the red and green channels of the control group (HepG2 cells were untreated) remained almost constant for HepG2 cells loaded with Vis-H 2 S with a 5 min interval for a total of 60 min, indicating that Vis-H 2 S possesses good photostability.The cells were then starved to induce autophagy, 24 and the results indicated that both I green and I red were signicantly reduced.From the above experimental results, a decrease in I green occurred due to a decrease in mitochondrial viscosity during autophagy.In addition, the gradual decrease in I red is attributed to the role of NFS1 protein as a reductive protein that facilitates the release of H 2 S through desulfuration of L-cysteine.During the autophagy process, lysosomes degrade NFS1 protein within the mitochondria, resulting in a reduction of endogenous H 2 S and subsequently weakening of I red .Aer the cells were starved and the autophagy inhibitor 3-methyladenine 25 (3-MA, 5 mM) was added, both I green and I red were signicantly enhanced and the uorescence intensity remained almost unchanged for the next 60 min.This again shows that autophagy can simultaneously reduce mitochondrial viscosity and H 2 S levels.When cells were preincubated with NEM, regardless of whether the cells were incubated with 3-MA, changes in I green were the same as those without NEM, while I red was almost negligible.Due to the ability of ow cytometry to rapidly analyze a large number of individual cells, achieving high-throughput data collection is crucial for comprehensive assessment of uorescence signal changes.In this regard, our study employed ow cytometry to investigate uorescence intensity alterations under the aforementioned conditions, as illustrated in Fig. S16.† In comparison to the control group, cells subjected to starvation exhibited a signicant decrease in uorescence intensity in both the green and red channels, with the addition of 3-MA inhibiting this process.Concurrent starvation and NEM incubation resulted in a notable reduction in uorescence intensity in both channels, with 3-MA incubation during this process leading to an increase in green channel uorescence intensity and minimal interference with red channel uorescence intensity.These ndings align with real-time imaging results, conrming the dualresponsive capability of Vis-H 2 S to simultaneously monitor cellular viscosity and H 2 S levels.
Finally, we evaluated the source of MDR during tumor chemotherapy.Adriamycin, a widely used chemotherapeutic drug, works by consuming oxygen to produce ROS, which in turn triggers autophagy (ROS has been conrmed as the main cause of upregulating cellular autophagy) and overexpression of NFS1 protein due to increased hypoxia in tumor cells. 26This combination of events leads to MDR but the synergistic effect between cellular autophagy and NFS1 protein overexpression in causing MDR has not been previously reported.To investigate the combined impact of autophagy and NFS1 protein on multidrug resistance (MDR) during tumor chemotherapy, we employed Vis-H 2 S for imaging.Before conducting imaging experiments, we initially employed western blot analysis to assess the NFS1 protein levels.This step aimed to validate our hypothesis that the upregulation of NFS1 protein, induced by Adriamycin, leads to a substantial increase in H 2 S production.The mechanism suggests that elevated NFS1 levels act as a protective response in tumor cells, preventing the cytotoxic effects of Adriamycin and thereby contributing to the development of multidrug resistance.The results are depicted in Fig. S18, † where the addition of Adriamycin led to a signicant increase in NFS1 protein content, while NFS1KD rendered the NFS1 protein content negligible.As shown in Fig. 3a-f, 3h and i, the control group experiment indicates that Vis-H 2 S retained excellent photostability (Fig. 3a).Aer adding Adriamycin, I green was signicantly reduced, while I red increased rst and then decreased (Fig. 3b).This occurred because the addition of Adriamycin signicantly increased the overexpression of NFS1 protein.Then upon prolonged incubation time, autophagy decreased the mitochondrial viscosity resulting in decreased I red .To validate this result, cells incubated with Adriamycin were also treated with the autophagy inhibitor 3-MA (10 mM) or the NFS1 gene of HepG2 cells was eliminated (NFS1KD, operational approach according to the reported method 27 ).The results indicated that I green gradually decreased and I red increased gradually in cells pretreated with 3-MA (Fig. 3c).This conrms that the uorescence intensity was only regulated by NFS1-induced H 2 S. In the Adriamycin + NFS1KD group, I green gradually decreased, and I red was negligible, indicating that the uorescence intensity was only regulated by autophagy.Compared with the Adriamycin group, the Adriamycin + NFS1KD group exhibited a more pronounced decrease in I green .We believe that the knockout of the NFS1 gene led to an increase in the cellular ROS level, which increased the autophagy level.In addition, the results of the negative control (performed according to the previous report 28 ) group were similar to those of the Adriamycin group, indicating that the NFS1KD group results were caused by the reduction of H 2 S levels via the knockout of the NFS1 protein.Finally, when autophagy and NFS1 protein were simultaneously inhibited (as shown in Fig. 3f), unsurprisingly, I green remained almost unchanged, while I red was negligible.We then further validated the accuracy of the results using ow cytometric analysis, as illustrated in Fig. S17.† The addition of Adriamycin resulted in a signicant decrease in uorescence intensity in both the green and red channels.Under these conditions, inhibiting autophagy (3-MA) led to a signicant increase in red channel uorescence intensity, while inhibiting NFS1 protein levels (NFS1KD) resulted in a signicant decrease in red channel uorescence intensity.This aligns with the real-time imaging results, demonstrating the probe's capability to detect cellular autophagy and NFS1 level uctuations during the development of Adriamycin-induced drug resistance.
Having ascertained the autophagy and NFS1 protein levels under the given conditions, this study also employed Annexin V-FITC/PI apoptosis detection agent assay and the CCK8 assay to comprehensively evaluate the Adriamycin drug resistance under different conditions.As shown in Fig. 3g, 3j and k, the cells in the control group maintained a high survival rate, while the survival rate of the cells in the Adriamycin group decreased to 60.5%.Interestingly, when autophagy was inhibited or when the NFS1 protein was knocked out, the apoptosis rate was signicantly increased to 49.6% and 31.1%,respectively.In addition, the lowest cell survival rate (9.7%) was achieved by the simultaneous inhibition of autophagy and knockdown of the NFS1 protein.The above experiments indicate that autophagy and the NFS1 protein are responsible for Adriamycin drug resistance.

Conclusions
A near-infrared uorescence-based probe, Vis-H 2 S, which could simultaneously respond to H 2 S and viscosity, was constructed based on the QCy7 parent uorophore.The probe could simultaneously monitor autophagy and H 2 S during the treatment of cells with Adriamycin and further conrmed the direct relationship between autophagy and NFS1 protein overexpression with ROS-based chemotherapeutic drugs.The results indicated that Adriamycin up-regulated the degree of autophagy and the NFS1 protein levels during tumour treatment, which resulted in MDR.Furthermore, reducing the level of NFS1 protein results in a small increase in autophagy levels, which may be because the NFS1 protein is unable to fully clear the ROS.Hence, using probe Vis-H 2 S we could conrm that both autophagy and NFS1 protein are closely related to the MDR of tumour cells.Therefore, Vis-H 2 S is a powerful analytical tool for the in-depth exploration of the drug resistance mechanism of Adriamycin.

Scheme 1
Scheme 1 Schematic illustration (a) and "AND" logic gate (b) of the dual-response platform; (c) the response mechanism of Vis-H 2 S toward H 2 S and viscosity; (d) the sensing mechanism of H 2 S using Vis-H 2 S.

Fig. 1
Fig. 1 Fluorescence intensity changes of Vis-H 2 S (10 mM) in response to H 2 S and viscosity.(a) 0% glycerol, (b) 50% glycerol, and (c) 70% glycerol; fluorescence titration curves of the probes Vis-H 2 S and H 2 S (0-500 mM); (d)-(f) the fluorescence intensity responses of Vis-H 2 S for (a)-(c) to different concentrations of H 2 S (0-500 mM) at the ratio signal (I 687 nm /I 492 nm ), respectively.The inset shows the linear relationship of the Vis-H 2 S response to a low concentration of H 2 S in different glycerol systems.The excitation wavelength was 400 nm, and the test medium was phosphate buffer (10 mM; pH = 7.4).

Fig. 2
Fig. 2 Real-time fluorescence imaging images of the HepG2 cells in: (a) control group, (b) starvation treatment, (c) starvation treatment + 5 mM 3-MA, (d) starvation treatment + 0.5 mM NEM, and (e) starvation treatment + 5 mM 3-MA + 0.5 mM NEM incubated with Vis-H 2 S (10 mM).(f) Fluorescence trend graph of intracellular I green as a function of time.(g) Fluorescence trend graph of intracellular I red as a function of time.Light collection range: green channel 450-520 nm, red channel 650-720 nm, the excitation light source is 552 nm, and the scale bar is 20 mm.