Commercially available chemicals and enzymes were purchased and used without further purification. In thin-layer chromatography (TLC) and column chromatographic separation, 300–400 mesh silica gel was purchased from Qingdao Haiyang Chemical Co., Ltd. (China). The cell lines used in this study were obtained from the American Type Culture Collection (ATCC) and stored in the central laboratory of the Jinhua Institute of Zhejiang University. The solution system, including phosphate-buffered saline (PBS), was prepared with pure water produced by the Milli-Q Reference Water Purification System (Merck, Darmstadt, Germany). Characterization was performed by nuclear magnetic resonance (NMR) spectroscopy on a Bruker DRX-600 Spectrometer (Germany) and high-resolution mass spectrometry (HRMS) on an AB SCIEX Triple-TOF 4600 System (United States). The UV-VIS spectra tests were conducted on a Shimadzu UV-2550 Spectrophotometer (Shimazu, Kyoto, Japan), while the fluorescence signals were measured on a Hitachi F-7000 Fluorescence Spectrophotometer (Japan). The confocal imaging experiments were performed on a Leica Mai Tai SP8 Microscope (Germany).

The probe VanPI-CarE was stored as a 1 mM stock solution, with dimethyl sulfoxide (DMSO) as the solvent. The solution system at a total volume of 200 µL for the detection consisted of 20 µL DMSO (containing the probe), 80 µL PBS (final concentration 10 mM, for preparing CarE), and 100 µL pure water (containing the aqueous analytes). Unless the condition was being tested, the working conditions were set as pH 7.4, incubation time 20 min, incubation temperature 37 °C, photomultiplier voltage 600 V, excitation/emission slit width 5 nm * 5 nm, and excitation wavelength 355 nm. The signal collection range in the confocal imaging experiments was 450 nm–600 nm in the green channel.

The chemical synthesis process of the probe VanPI-CarE is depicted in Figure 2. There were two main steps. At first, 15 mL of acetic acid was added to a 50-mL round-bottom flask to dissolve the reagents phenyl(pyridin-2-yl)methanone (0.27 g, 1.5 mmol), ammonium acetate (0.15 g, 2 mmol), and vanillin (0.23 g, 1.5 mmol). The reaction was carried out under reflux for 5 h, and its completion was monitored by TLC. Subsequently, ice water was added to the reaction system, and the precipitate was collected. After column chromatography (petroleum ether: ethyl acetate = 5:1), the fluorophore VanPI-OH was acquired as a yellow solid (yield 75.2%). The ¹H NMR spectrum (600 MHz, CDCl₃) showed signals at δ 9.79 (s, 1H), 8.18 (d, J = 7.3 Hz, 1H), 7.92 (d, J = 7.2 Hz, 2H), 7.81 (d, J = 9.3 Hz, 1H), 7.45 (t, J = 7.7 Hz, 2H), 7.35 (d, J = 1.8 Hz, 1H), 7.29 (t, J = 7.4 Hz, 1H), 7.28–7.24 (m, 1H), 7.01 (d, J = 8.0 Hz, 1H), 6.76 (dd, J = 9.7, 6.3 Hz, 1H), 6.55 (t, J = 7.1 Hz, 1H), and 3.91 (s, 3H). The ¹³C NMR spectrum (151 MHz, CDCl₃) showed peaks at δ 147.35, 146.66, 138.27, 134.73, 131.40, 128.73, 127.33, 126.86, 126.56, 121.87, 120.91, 119.62, 119.03, 114.68. 114.53, 113.20, 111.90, and 56.08. HRMS (Q-TOF m/z) provided a calculated value of 317.1290 for [C₂₀H₁₇N₂O₂]⁺ and a found value of 317.1279.

Furthermore, 20 mL of dichloromethane (DCM) was added to a 50-mL round-bottom flask to dissolve VanPI-OH (0.5 mmol), N,N-dimethyl carbamic acid (0.5 mmol), 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDCI, 1 mmol), and 1-hydroxybenzotriazole (HOBt, 0.5 mmol). The reaction was carried out at room temperature for 5 h, and its completion was monitored by TLC. The solvent was vaporized, and following that, column chromatography (petroleum ether: ethyl acetate = 16:1) was conducted. The probe VanPI-CarE was acquired as a yellow solid (yield 48.6%). The ¹H NMR spectrum (600 MHz, CDCl₃) showed signals at δ 8.27 (d, J = 7.3 Hz, 1H), 7.96 (d, J = 7.3 Hz, 2H), 7.85 (d, J = 9.2 Hz, 1H), 7.50–7.48 (m, 3H), 7.38 (dd, J = 8.2, 1.9 Hz, 1H), 7.32 (t, J = 7.4 Hz, 1H), 7.26 (d, J = 8.1 Hz, 1H), 6.80 (dd, J = 9.2, 6.3 Hz, 1H), 6.59 (t, J = 6.5 Hz, 1H), 3.95 (s, 3H), 3.18 (s, 3H), and 3.06 (s, 3H). The ¹³C NMR spectrum (151 MHz, CDCl₃) showed peaks at δ 154.57, 152.39, 141.16, 137.67, 134.86, 131.87, 128.75, 128.26, 127.73, 126.86, 126.61, 123.65, 121.92, 119.97, 119.81, 119.11, 113.35, 113.26, 56.28, 36.86, and 36.65. HRMS (Q-TOF m/z) provided a calculated value of 388.1661 for [C₂₃H₂₂N₃O₃]⁺ and a found value of 388.1643.

The reference method using an ethanol solution of rhodamine B (10 μM, Φ = 0.69, λ_ex = 365 nm) was employed to calculate the fluorescence quantum yield (FQY) values. In this study, the FQY values of the probe VanPI-CarE and the detection product VanPI-OH were 0.15 and 0.71, respectively.

In this work, the limit of detection (LOD) was calculated using the formula LOD = 3σ/k, where the background noise σ was obtained from 25 independent tests of the solution system containing only the probe and the slope k was determined from the linear regression equation. Thus, σ = 0.4985, k = 88.52, and LOD = 0.017 U/mL.

Cell viability was tested on RAW264.7 (mouse monocyte macrophage leukemia cells) and MC3T3-E1 (mouse embryonic osteoblast precursor cells) using the thiazolyl blue (MTT) assay (Buranaamnuay, 2021). Absorbance at 570 nm was measured.

Moreover, RAW264.7 cells were incubated in Dulbecco’s modified Eagle’s medium (DMEM) with hydroxyethyl piperazine ethane sulfonic acid buffer (HEPES, pH 7.2), 10% fetal bovine serum (FBS), and 1% penicillin–streptomycin at 37 °C under a 5% CO₂ atmosphere. HEPES was added to improve the solubility of receptor activator of nuclear factor-κB ligand (RANKL). The RAW264.7 cell line was chosen for its relevance to the study of macrophage polarization in bone homeostasis. The cells were divided into five groups. The first group served as the original condition, incubated with HEPES for 30 min, followed by incubation with the probe VanPI-CarE (10 μM) for 30 min, and then imaged. The second group served as the inhibited condition, incubated with the CarE inhibitor bis(4-nitrophenyl)phosphate (BNPP) at 1 μM for 30 min, followed by incubation with VanPI-CarE (10 μM) for 30 min, and then imaged. The third group served as the stimulated condition, which was pre-treated with oxidized low-density lipoprotein (ox-LDL) at 20 μg/mL during the last 12 h of culturing to induce the CarE level, incubated with VanPI-CarE (10 μM) for 30 min, and then imaged. The fourth group served as the bone homeostasis-related macrophage polarization condition, which was induced by RANKL (100 ng/mL) during the last 12 h of culturing (Elango et al., 2021), incubated with HEPES for 30 min, followed by incubation with VanPI-CarE (10 μM) for 30 min, and then imaged. The fifth group served as the macrophage polarization-inhibited condition, which was induced by RANKL, treated with the RANKL inhibitor denosumab (1 μg/mL) during the last 1 h of culturing (Deligiorgi and Trafalis, 2021), incubated with HEPES for 30 min, followed by incubation with VanPI-CarE (10 μM) for 30 min, and then imaged. The fluorescence signals in the green channel of 450 nm–600 nm were collected when the excitation wavelength was set to 355 nm.

The general synthetic route of the probe VanPI-CarE is depicted in Figure 2 as comprising two main steps. The initial step was the formation of the fluorophore VanPI-CarE from the cyclization of vanillin, phenyl(pyridin-2-yl)methanone, and ammonium acetate (Shi et al., 2025; Yuan et al., 2020). The vanillin–pyridine–imidazole core structure was constructed thereby. The second step was anchoring the carbamate recognition onto the fluorophore to yield the probe VanPI-CarE. The structures of the probe and the fluorophore were confirmed by satisfactory characterization data (¹H NMR, ¹³C NMR, and HRMS, Supplementary Figures S1–S6 in Supporting Information). The recognition mechanism was supported by previous reports and variations in the HRMS data.

When the optical performance was studied, the UV–VIS absorption and fluorescence spectra were examined to provide the initial information. In this work, the probe VanPI-CarE (10 μM) exhibited a visible peak at 530 nm (due to the frequency-doubling effect), while recognition with CarE (20 U/mL) for 20 min at 37 °C resulted in a decrease in the signal (Supplementary Figures S7a). More importantly, for the aspect of the fluorescence reporting signal, when the excitation wavelength was set to 355 nm, the probe VanPI-CarE (10 μM) was almost non-fluorescent, while recognition with CarE (20 U/mL) for 20 min at 37 °C caused a remarkable enhancement in the peak at 490 nm (Supplementary Figures S7b). The response scale referred an over 35-fold fluorescence enhancement, which was suitable for establishing the system of turning-on recognition. Based on the collection of both the absorbance and fluorescence data, the FQY values of the probe VanPI-CarE and the detection product VanPI-OH were 0.15 and 0.71, respectively. Since the basic signal variation during CarE recognition had been studied, the following experiments were carried out to examine the working conditions, including pH, incubation time, and temperature. Recognition of the enzymatic indicator is commonly affected by the pH condition. In this study, the probe VanPI-CarE showed no obvious fluorescence signal within the whole tested range of 3.0–12.0, while the fluorescence reporting signal with a certain intensity after recognition between VanPI-CarE and CarE remained stable in the range of 7.0–9.0 (Figure 3a). This result indicated the considerable potential for detection in physiological and pathological procedures. Meanwhile, since the recognition time is usually a significant factor, it was also tested by setting different checking points. Recognition of VanPI-CarE toward CarE was completed within 20 min, which is a shorter period than that in similar reports (Figure 3b). The time-dependent response followed the Michaelis–Menten model, with parameters including V_max = 1,011 min⁻¹ and K_m = 1.185 U/mL, consistent with typical values of carboxylesterase (Nagaoka et al., 2024). For the condition of incubation temperature, VanPI-CarE itself showed no obvious fluorescence signal in the tested range of 25 °C–45 °C, while the fluorescence reporting signal remained stable in the range of 35 °C–40 °C (Supplementary Figures S8). This result was also consistent with the requirements of the physiological micro-environment.

After the working conditions were investigated, the correlation between the reporting signal intensity at 490 nm and the CarE level (0–20 U/mL) in the solution system containing VanPI-CarE (10 μM) was established. The upper limit of the CarE level was set at 20 U/mL because this level ensured a relatively transparent solution and fulfilled the requirements for physiological detection. As the CarE level increased gradually, the fluorescence reporting signal correspondingly enhanced, reaching a saturated value at a CarE level of 15 U/mL (Figures 3c, d). A linear correlation was found in the range of 0 U/mL–10 U/mL, with a Pearson’s r value of 0.9997 (Figure 3d Inner). Using the formula 3σ/k, the LOD value was determined to be 0.017 U/mL, indicating relatively high sensitivity. Both the linear range and the LOD value are suitable for the potential research scenarios in this work. Therefore, in the solution system, the probe VanPI-CarE showed potential optical capabilities for CarE detection.

In the next step, the selectivity of the probe VanPI-CarE (10 µM) toward CarE (20 U/mL) was investigated. The most concerned species were the competing enzymes, including alkaline phosphatase (ALP), alanine aminotransferase (ALT), aspartate aminotransferase (AST), β-glucosidases (β-Glu), xanthine oxidase (XO), tyrosinase, trypsin, monoamine oxidase-A (MAO-A), monoamine oxidase-B (MAO-B), human serum albumin (HSA), and bovine serum albumin (BSA) from the similar physiological micro-environment of CarE (Figure 4a). In particular, the inhibition and induction agents in intracellular imaging, including BNPP, ox-LDL, RANKL, and denosumab, were involved. None of the tested species, except CarE, led to a remarkable enhancement of the fluorescence reporting signal. In consideration of their activity in physiological events, the reactive oxygen/nitrogen species (ROS/RNS), including NO, ¹O₂, ONOO⁻, HClO, OH, H₂O₂, and O₂ ⁻, and anions, including Br⁻, F⁻, CO₃ ²⁻, HCO₃ ⁻, SO₄ ²⁻, SO₃ ²⁻, and NO₃ ⁻, were also tested in this section (Figure 4b). None of them caused any detectable fluorescence reporting signal. In further steps, the tests covered more analytes, including the usual amino acids (Ala, Arg, Asp, Asn, Gln, Gly, Glu, His, Ile, Leu, Lys, Met, Pro, Ser, Thr, Tyr, and Val; Figure 4c) and cations (Al³⁺, Ca²⁺, Cu²⁺, Fe³⁺, Fe²⁺, K⁺, Li⁺, Mg²⁺, Mn²⁺, Na⁺, Pb²⁺, Ti⁴⁺, and Zn²⁺; Figure 4d). None of the tested analytes produced a notable fluorescence reporting signal. Therefore, in the solution system, the high selectivity of the probe VanPI-CarE toward CarE was guaranteed.

This work focused on carboxylesterase detection in macrophage polarization during bone homeostasis. Thus, bone homeostasis-related macrophages and precursor cells, including the RAW264.7 (Supplementary Figures S9a) and MC3T3-E1 (Supplementary Figures S9a) cell lines, were tested for cell viability using a standard MTT assay. After 24 h of incubation, both the cell lines retained over 90% cell viability when the working concentration of the probe gradually increased to 50 µM. Thus, VanPI-CarE inferred low cytotoxicity for imaging in living macrophages.

RAW264.7 cells were maintained in an uninduced state before the confocal experiments because bone homeostasis-related macrophage polarization requires induction during culturing. The cells were divided into five groups according to the different treatment conditions. The first group, which represented the original condition, was incubated with HEPES for 30 min, followed by incubation with the probe VanPI-CarE (10 μM) for 30 min, and then imaged (Figures 5a–c). Since the living RAW264.7 cells bore a certain level of CarE, the fluorescence reporting signal was observed in the green channel. In the second group, in which CarE was inhibited by BNPP (1 μM), the following incubation with VanPI-CarE resulted in a remarkable decrease in the fluorescence reporting signal (Figures 5d–f). On the contrary, the third group was pre-treated with ox-LDL (20 μg/mL) during the last 12 h of culturing before being incubated with VanPI-CarE for 30 min and then imaged (Figures 5g–i). In this group, the CarE level was stimulated, and the fluorescence reporting signal was notably enhanced. The results from the initial three groups suggested that VanPI-CarE was capable of visualizing the CarE levels in living macrophages under both inhibition and activation conditions. Then, the following two groups were associated with macrophage polarization during bone homeostasis. The fourth group, which served as the bone homeostasis-related macrophage polarization condition, was induced by RANKL (100 ng/mL) during the last 12 h of culturing before being incubated with HEPES for 30 min, followed by incubation with VanPI-CarE (10 μM) for 30 min, and then imaged (Figures 5j–l). Correspondingly, the fluorescence reporting signal in the green channel exhibited a remarkable decrease, which was consistent with the fact that the M1-type macrophage polarization process induced by ox-LDL caused inflammation and affected the metabolism of fatty acids (Yang et al., 2022). Finally, the fifth group was established on the basis of the fourth condition. After induction with RANKL, the cells were treated with the RANKL inhibitor denosumab (1 μg/mL) during the last 1 h of culturing before incubation with HEPES and VanPI-CarE (Figures 5m–o). The fluorescence reporting signal subsequently indicated a recovery close to the original uninduced condition in the first group, suggesting a corresponding restoration of the CarE level. Therefore, VanPI-CarE achieved reflection of macrophage polarization in bone homeostasis, regardless of induction or inhibition, by visualizing the CarE level.