The probe was dissolved in DMSO to produce a stock solution (1.0 × 10⁻⁴ M). To obtain good optical performance of the probe, we used 5% DMSO as co-solvent. Since phosphate could inhibit the activity of ALP, we used 10 mM tris-buffered saline (TBS, pH 8.0) for in vitro experiments. Both the UV-Vis absorption and fluorescent spectrum experiments were carried out in 10 mM TBS (pH 8.0). The fluorescent emission spectra were tested using a Hitachi F7000 spectrophotometer, with the excitation wavelength at 680 nm and the emission wavelength ranging from 700 to 900 nm. The test solutions were kept at 37°C for 30 min and then analyzed using spectrophotometry.

Cells were co-incubated with the probe in 10 mM TBS (pH 7.4). Cytotoxicity was assessed using the MTS assay. Cells were seeded in a 25-mm glass-bottom dish and grown for 24 h to approximately 85% confluency, followed by the live-cell confocal imaging experiments, for which the cells were divided into three groups as follows. The first group included HeLa cells that were incubated with 5 μM CS-ALP for 30 min. The cells were then washed with TBS buffer before imaging. In the second group, after pretreatment with Na₃VO₄ for 4 h, HeLa cells were incubated with 5 μM CS-ALP for 0.5 h and then washed with TBS prior to imaging. The third group included control cells without CS-ALP. For the cisplatin-induced nephrotoxicity model in HK-2 cells, the experiment was divided into four groups. The first group included HK-2 cells that were treated with 5 μM CS-ALP. The second and third groups included HK-2 cells that were incubated with cisplatin, 500 μM or 1000 μM, for 4 h, respectively, and then co-incubated with 5 μM CS-ALP for 0.5 h. The fourth group included cells that were pre-incubated with 50 μM Na₃VO₄ in the presence of cisplatin (1 mM) for 4 h and then treated with 5 μM CS-ALP for 0.5 h. Confocal fluorescence imaging of cells was performed using an Olympus FV1000 laser confocal microscope (Japan).

The drug-induced AKI mouse model was constructed following methods detailed in our previous report. Liu et al. (2020a) cisplatin was dissolved in 0.9% saline. BALB/c mice were intraperitoneally preinjected with cisplatin (20 mg/kg) for 24 h or 48 h. Live cells and live mouse experiments were in agreement with institutional animal care and use regulations, according to protocol No. SYXK (Xiang) 2022-0007, approved by Laboratory Animal Center of Hunan.

Drug-induced AKI is accompanied by oxidative stress. Many studies have indicated that high levels of ROS are generated by cells under oxidative stress (Mao et al., 2017; Mao et al., 2022). Thus, the probe used for imaging of AKI must be chemically stable in the presence of ROS. As proof of concept, we first studied the reactivity of the CS dye with typical oxidants and nucleophiles, including H₂S, Cys, GSH, H₂O₂, O₂ ^•−, •OH, ONOO⁻, and ClO⁻. As exhibited in Figure 1, the absorption intensity of CS at 685 nm remained unchanged with the addition of test species at physiological concentrations. In contrast, seven other kinds of NIR dyes, five of which (dyes 1–5) have been used for constructing NIR ALP fluorescent probes (Li et al., 2012; Li et al., 2017; Zhang P. et al., 2019; Zhang Q. et al., 2019; Gao et al., 2019; Wang et al., 2021b), were also tested. The results indicated that dyes 1–7 were not stable in the presence of O₂ ^•−, •OH, ONOO⁻, and ClO⁻. These ROS have been shown to increase in concentration during drug-induced AKI (Lv et al., 2018; Huang et al., 2020). The results demonstrated that the existing NIRF ALP probes are not suitable for in vivo imaging and measurement of ALP levels in the drug-induced AKI model. In addition, for AKI in vivo imaging, probes must be water soluble at low concentration to allow enrichment in the kidney. We generated the probe through phosphorylation of the hydroxyl group of the CS fluorophore. The log P values of the probe and CS dye were 2.238 and 3.644, respectively, under pH 8.0 buffer conditions; as a control, the log P value of rhodamine B, a well know water-soluble dye, was calculated to be 2.032 under the experimental condition. The log P values indicated that CS-ALP has good water solubility and the potential for use in imaging to assess renal metabolism. To our knowledge, CS-ALP is the first molecular probe suitable for NIRF/PA bimodal in vivo detection of ALP in drug-induced AKI. The synthesis methodology for CS-ALP is shown in Supplementary Scheme S1, with structures assessed by ¹H NMR and MS (ESI).

After synthesizing the probe, we conducted a spectral study of the reaction between CS-ALP and ALP. CS-ALP showed a maximum absorption peak at 600 nm. After addition of ALP, a new absorption peak appeared at 685 nm. The increased absorption intensity displayed a good linear relationship with the activity of ALP at 0–10 U/L (Figure 2A, and Supplementary Figure S1, ESI). Concomitantly, the fluorescence emission peak at 716 nm grew dramatically upon increased ALP activity (Figure 2B). According to the fluorescence titration experiment, the fluorescence intensity at 716 nm exhibited a good linear relationship with ALP at 0–10 U/L (Figure 2C), with a calculated detection limit of 0.26 U/L. The catalytic activity of the probe in response to ALP was then assessed. Figure 2D depicts the fluorescence dynamics of the reaction of the probe to ALP at different concentrations (0, 2, 20, and 40 U/L). The results indicate that high-activity ALP allows the probe to crack quickly, resulting in a dramatic increase in fluorescence intensity. The fluorescence intensity of the reaction system is essentially balanced at about 26 min. At the same time, in the absence of ALP, there was no significant change in the fluorescence intensity of the reaction system (Supplementary Figure S2, ESI), and the photo-stability of the probe and the CS dye is good (Supplementary Figure S3). Together, these results indicate that CS-ALP is stable at physiological pH and could be applied in bioimaging.

To further demonstrate the specificity of the reaction between CS-ALP and the enzyme, the specificity of the probe for various potentially biologically relevant substances was assessed. As shown in Supplementary Figure S4, the signal intensity is not significantly affected by other biological interference. Na₃VO₄, a commonly used and competitive phosphatase inhibitor (Liu et al., 2017), was then co-incubated with ALP for 30 min before addition of the probe to the test solution. As shown in Supplementary Figure S5, addition of 50 μM Na₃VO₄ resulted in a significant reduction in fluorescence intensity. These results confirm that ALP-induced phosphate group-specific cleavage of the probe contributes to the enhancement of fluorescence intensity. Next, the effect of pH on CS-ALP and the enzymatic reaction was also evaluated, with results demonstrating that CS-ALP could perform well at physiological pH (Supplementary Figure S6).

The detectability of CS-ALP using PA imaging was then tested. We first tested the PA spectra for CS-ALP in the presence of varied ALP activity. The results showed that it was generally consistent with its absorption spectra. As displayed in Figure 3A, the PA signal intensity at 680 nm (PA₆₈₀) exhibited 1.8-fold enhancement at the saturation point. PA₆₈₀ images of CS-ALP were obtained (Figure 3B) by treating the probe with increasing concentrations of ALP (0–30 U/L). The PA₆₈₀ signals displayed linear responses to ALP in the 0–30 U/L concentration range, with a calculated detection limit of 2.5 U/L (Figure 3C).

The cytotoxicity of CS-ALP was first assessed using the MTS assay. The results showed that high cell viability was maintained after co-incubation with 2–10 μM probe, indicating that CS-ALP had good biocompatibility (Supplementary Figure S7). Next, we used HeLa cells for co-incubation with the probe for live-cell imaging experiments because ALP is overexpressed in HeLa cells. As shown in Supplementary Figure S8, when excited at 635 nm, HeLa cells showed bright fluorescence. However, when pretreated with Na₃VO₄ before co-incubation with CS-ALP, HeLa cells showed negligible fluorescence. The results demonstrate that the signal enhancement in HeLa cells is due to ALP activity. We then constructed an AKI cell model using cisplatin, a drug known to induce AKI. HK2 cells treated with CS-ALP alone were excited at 635 nm and showed virtually no fluorescence (Figure 4A). At the same time, we observed that, when pretreating HK2 cells with 0.5 mM or 1 mM cisplatin, signal variations indicated cisplatin dose dependence (Figures 4B,C) and the fluorescence intensity ratio increased by 2.6 and 3.4 times, respectively (Figure 4E). Furthermore, when the cisplatin-treated HK2 cells were co-incubated with Na₃VO₄ for 30 min before addition of CS-ALP, there was almost no fluorescence (Figure 4D). These results suggest that CS-ALP could be used for measurement of endogenous ALP in cisplatin-treated HK2 cells, and that the activity of ALP is directly related to the process of cisplatin-induced AKI.

The AKI mouse model was established by intraperitoneal injection of 20 mg/kg cisplatin or normal saline as control. We first tested the urine samples of mice and found that the signal intensity from mice treated with cisplatin for 24 h and 48 h was 1.51-fold and 2.23-fold higher, respectively, than that of the control group. Blood sample analysis showed limited signal increase in the experimental groups compared to control. This result is consistent with that observed when using the commercially available ALP probe 4-MUP (Figure 5). These results are also consistent with a previous study showing that ALP was discharged into the urine following cellular injury due to early stage AKI (Charlton et al., 2014). According to the fluorescence correction curve for CS-ALP to ALP, the calculated ALP activity levels in the urine of AKI mice were 30.1 U/L (cisplatin-treated 24 h) and 45.6 U/L (cisplatin-treated 48 h), respectively. Thus, the CS-ALP probe may be an effective tool for early detection of AKI.

After successful NIRF imaging of ALP in the AKI cell model, we applied CS-ALP to in vivo bimodal NIRF/PA imaging using commercial in vivo imaging systems. First, we constructed an AKI mouse model according to our previous protocol (Liu H.-W. et al., 2020). The biodistribution of CS-ALP after intravenous (i.v.) injection showed that CS-ALP remained inactive in the control mouse but was activated in the AKI model mouse (Supplementary Figure S9A). In another group, imaging of isolated organs of the AKI mouse at 1.5 h post injection showed that CS-ALP accumulated primarily in the kidneys and liver (Supplementary Figure S9B), indicating that cisplatin may induce liver injury and kidney injury at the same time (Palipoch and Punsawad, 2013). As mentioned earlier, the log P value of CS-ALP is 2.238, which indicates that the probe has good water solubility and shows the potential for use in imaging to assess renal metabolism. Subsequently, 0.15 mg/kg CS-ALP was intravenously administered at different cisplatin post-treatment time points (24, 48, and 72 h). Real-time in vivo imaging showed that CS-ALP could be rapidly activated in the kidneys at 10 min post injection and that the signal intensity was stable at 10–30 min, followed by an obvious decrease at 120 min post injection (Supplementary Figure S10). Therefore, we timed the in vivo NIRF imaging to be at 30 min after injection of the probe. As displayed in Figures 6A,B, the NIRF signal intensity gradually enhanced, mainly in the kidneys of mice treated with cisplatin for 24 h or 48 h. The signal ratios for the 24 h and 48 h groups were 3.52-fold and 3.92-fold higher, respectively, than that of control group, indicating that ALP was upregulated in the kidneys of cisplatin-treated mice. We also carried out PA imaging, and the results are shown in Figure 6C and Supplementary Figure S11. The PA imaging data indicate that the kidney regions of the control group (treated with normal saline) were low contrast at 1 h post injection of CS-ALP, while the PA signal in the kidney regions of cisplatin-treated mice gradually increased. The average signal intensity is displayed in Figure 6D, which indicates that the probe could be used for semi-quantitative analysis of ALP in drug-induced AKI. The PA imaging results were consistent with the NIRF imaging results, but the in vivo PA imaging required a longer time to obtain high contrast results compared to NIRF imaging due to the high background signal with PA imaging in this region, hence the advantage of dual-mode imaging for acquisition of more reliable in vivo imaging results (Luo et al., 2020; Zhang et al., 2020). Hematoxylin–eosin (H&E) staining revealed normal tubular morphology in the control group, compared to widely spaced tubules, damage to the brush border, and formation of hyaline casts within the damaged tubules in the experimental group 24 h after treatment with cisplatin (Supplementary Figure S12). Considering these results, we conclude that upregulation of ALP expression will occur at the early stages of AKI and will continue to be expressed with the progression of AKI. Additionally, we demonstrated the development and validation of use of CS-ALP in NIRF/PA bimodal imaging as a tool for evaluating drug-induced AKI in response to ALP.