Research ArticleHEALTH AND MEDICINE

Noninvasive monitoring of hepatic glutathione depletion through fluorescence imaging and blood testing

See allHide authors and affiliations

Science Advances  19 Feb 2021:
Vol. 7, no. 8, eabd9847
DOI: 10.1126/sciadv.abd9847

Abstract

Hepatic glutathione plays a key role in regulating redox potential of the entire body, and its depletion is known to increase susceptibility to oxidative stress involved in many diseases. However, this crucial pathophysiological event can only be detected noninvasively with high-end instrumentation or invasively with surgical biopsy, limiting both preclinical research and clinical prevention of oxidative stress–related diseases. Here, we report that both in vivo fluorescence imaging and blood testing (the first-line detection in the clinics) can be used for noninvasive and consecutive monitoring of hepatic glutathione depletion at high specificity and accuracy with assistance of a body-clearable nanoprobe, of which emission and surface chemistries are selectively activated and transformed by hepatic glutathione in the liver sinusoids. These findings open a new avenue to designing exogenous blood markers that can carry information of local disease through specific nanobiochemical interactions back to the bloodstream for facile and rapid disease detection.

INTRODUCTION

As the master antioxidant in the liver, hepatic glutathione (GSH) is the key to maintaining the redox environment of the liver (1, 2). Constant efflux of GSH from the hepatocytes into the bloodstream allows the redox potential of the entire body to be precisely regulated (3, 4), and depletion of hepatic GSH has been found to strongly correlate with an increased susceptibility to oxidative stress and high risk of many liver diseases ranging from drug-induced liver injury (5, 6) to alcoholic/nonalcoholic fatty liver disease (1, 79), steatohepatitis (10, 11), liver fibrosis, and cirrhosis (1214). However, unlike conventional liver serum markers, such as alanine aminotransferase (ALT) and aspartate aminotransferase (AST), hepatic GSH depletion in the liver is very hard to be noninvasively monitored since hepatic GSH (~7 mM) is diluted by more than two orders of magnitude after entering into the blood circulation (~30 μM in blood plasma), followed by consumption in other organs and rapid clearance through the kidneys (15, 16). As a result, in the preclinical and clinical research, surgical or needle liver biopsies have usually been required for ex vivo measurement of hepatic GSH concentrations (1719). These invasive methods essentially preclude monitoring the kinetics of GSH depletion of individual animals or patients at higher temporal resolution. In more recent studies, nuclear magnetic resonance spectroscopy has been used to monitor 13C-GSH synthesis in the liver after injection of [2-13C]-glycine (20). While this technique has advanced noninvasive but indirect monitoring of GSH depletion in preclinical animals and even in humans (21), it is very difficult for many research laboratories and clinics with limited resources to access the high-end instrumentation and conduct the monitoring. On the other hand, fluorescence imaging and blood testing are considered both low cost and easily accessible for preclinical research and family clinics if specific and accurate markers were to become available.

Here, we report that by taking advantage of hepatic GSH–mediated biotransformation of gold nanoparticles in the liver sinusoids (22), we successfully used a GSH-activatable and body-clearable nanoprobe to monitor hepatic GSH depletion in the liver sinusoids with either in vivo fluorescence imaging or blood testing (Fig. 1). This nanoprobe is indocyanine green (ICG)–conjugated GSH-protected Au25 nanocluster (ICG4-GS-Au25), of which the ICG can be released and its near-infrared (NIR) emission can be reactivated by GSH. After it was intravenously injected into mice, we first found that ICG fluorescence activation kinetics in the liver were linearly dependent on hepatic GSH concentrations in the mice with controlled hepatic GSH level. Furthermore, this linear GSH-dependent activation of nanoprobe in the liver was successfully translated to monitor hepatic GSH depletion in mice after acetaminophen (APAP)–induced liver injury, the most frequently encountered cause of drug-induced liver injury in the Western world (23). Because ICG was specifically released in the liver sinusoids but not in the peripheral blood circulation, surface chemistry of ICG4-GS-Au25 collected from the peripheral blood was also found to strongly correlate with the hepatic GSH level, which further allowed us to noninvasively and consecutively monitor hepatic GSH depletion and recovery in the mice with liver injuries at high accuracy through a simple blood test. Because hepatic GSH depletion usually occurs before the death of liver cells as well as the elevation of common liver function biomarkers (5, 24), early detection and noninvasive monitoring of hepatic GSH depletion through simple preclinical imaging tool and blood testing are expected to greatly advance understandings of oxidative stress–related liver diseases and make it possible for early disease management in outpatient setting.

Fig. 1 Schematic illustration of the ICG4-GS-Au25–enabled noninvasive monitoring of hepatic GSH depletion through both in vivo fluorescence imaging and blood test.

After intravenous administration, ICG4-GS-Au25 nanoprobes bind to serum proteins and are efficiently transported to the liver, where their surface chemistries undergo biotransformation in the extracellular space specifically by liver sinusoidal GSH efflux, displacing the ICG-GS moiety from the surface of Au25 and activating the NIR fluorescence (FL) of ICG, thus allowing for reporting of liver GSH through in vivo fluorescence imaging. The biotransformed ICG4-GS-Au25 derivatives with reduced number of surface ICG molecules will carry the hepatic GSH information back to blood circulation, where they are fairly stable and therefore enable detecting hepatic GSH level from peripheral blood after detaching the remaining ICG-GS from Au25 with dithiothreitol (DTT). Compared with normal condition, hepatic GSH depletion results in decreased ICG fluorescence activation kinetics in the liver and increased ICG fluorescence on/off ratios in the peripheral blood because ICG molecules are less efficiently released in the liver, and thus, more ICG molecules will remain on the Au25 surface after going through the liver sinusoids. Last, the separated ICG-GS and Au25 will be cleared through the hepatobiliary and renal pathways, respectively, without being retained in the body.

RESULTS

Fluorescence activation kinetics of ICG4-GS-Au25 depends linearly on the GSH level both in vitro and in vivo

To successfully detect hepatic GSH depletion,, the signals of probes need to specifically correlate with the hepatic GSH level. Activatable NIR emission of ICG4-GS-Au25 in the presence of GSH made it possible to use fluorescence measurement to investigate this process at both the in vitro and in vivo levels. ICG4-GS-Au25 was synthesized by conjugating ICG to GSH-coated Au25 nanocluster GS-Au25 [Au25(SG)18] according to procedures reported in our previous study (22). The monodispersity and purity of ICG4-GS-Au25 after conjugation was confirmed by transmission electron microscopy (TEM) and high-performance liquid chromatography (fig. S1). The conjugation of ICG onto GS-Au25 led to a substantial blue shift of ICG absorption peak from 795 to 710 nm, which was due to the H coupling effect of multiple ICG molecules on the same Au25 surface (Fig. 2A) (25). Because of the close proximity between Au25 and the conjugated ICG, the NIR fluorescence of ICG was nearly completely (99.2%) quenched due to the efficient photoinduced electron transfer process (26). However, once the ICG-GS was displaced from the Au25 surface by free GSH, the NIR fluorescence of ICG was instantaneously recovered, resulting in over 100-fold enhancement in the NIR emission (fig. S2). In vitro incubation of ICG4-GS-Au25 in phosphate-buffered saline (PBS) solution containing 0.2 to 1 mM GSH revealed that the activation of ICG fluorescence as a function of incubation time exhibited a quasi-linear relationship within the first 180 s of incubation (Fig. 2B). Moreover, the fluorescence activation kinetics was found to linearly correlate with both the GSH concentration (Pearson’s r = 0.99; Fig. 2C) and ICG4-GS-Au25 concentration (r = 1.00; fig. S3) in vitro, indicating that GSH-mediated activation kinetics of ICG fluorescence is a second-order reaction at the in vitro level. These results suggest that the fluorescence activation kinetics of ICG4-GS-Au25 could be used to measure the GSH concentration.

Fig. 2 Fluorescence activation kinetics of ICG4-GS-Au25 depends linearly on GSH level both in vitro and in vivo.

(A) Absorption spectra of ICG4-GS-Au25, free ICG, and GS-Au25 nanocluster. Inserted is a schematic of the ICG4-GS-Au25 structure. (B) Time-dependent ICG fluorescence intensity of ICG4-GS-Au25 incubated in PBS containing different concentrations of GSH. (C) Correlation between GSH concentration and ICG fluorescence recovery kinetics of ICG4-GS-Au25. The degree of correlation is quantified by Pearson’s correlation coefficient (Pearson’s r). (D) Representative noninvasive in vivo fluorescence images show the activation of ICG fluorescence in the liver (pointed by arrows) following intravenous injection of the same dose of ICG4-GS-Au25 in mice pretreated with different doses of diethyl maleate (DEM) 30 min in advance. (E) Time-dependent liver ICG fluorescence curves of DEM-treated mice after injection of the same dose of ICG4-GS-Au25. (F) Correlation between the liver GSH level and liver ICG fluorescence kinetics of individual mouse. N = 3 mice for 0.3 and 0.6 ml/kg DEM-treated groups, and N = 4 for 0 mg/kg DEM-treated group. Statistical significance is evaluated by two-sample equal variance t test (P < 0.05; statistically significant). Data points are presented as means and SD.

Because of the high affinity of ICG to serum proteins, ICG4-GS-Au25 strongly bound to serum proteins as well (fig. S4), which induced rapid transport to the liver once it entered the bloodstream. To test whether the fluorescence activation kinetics of ICG4-GS-Au25 in the liver was also linearly dependent on hepatic GSH level in vivo, we treated BALB/c mice with different doses of diethyl maleate (DEM; 0, 0.3, and 0.6 mg/kg body weight; intraperitoneal injection), a GSH-selective depletion agent that can temporarily deplete liver GSH and control hepatic GSH level (27). At 30 min after DEM treatment, we intravenously injected ICG4-GS-Au25 into the mice and then conducted the noninvasive dynamic in vivo fluorescence imaging of the liver within 180 s, followed by immediately collecting the liver tissue samples and directly measuring the GSH level of the extracted liver ex vivo. As shown in Fig. 2 (D and E), in the mice without DEM treatment, intravenous injection of ICG4-GS-Au25 resulted in an immediate activation of ICG fluorescence in the liver because sinusoidal GSH efflux from the hepatocytes generated high local concentrations of GSH as well as cysteine by GSH-mediated reduction of plasma cystine, which together displaced ICG from GS-Au25 and activated ICG emission in vivo (22). In DEM-treated mice, the ICG fluorescence activation kinetics in the liver decreased accordingly with the increased DEM doses and demonstrated quasi-linear characteristics (Fig. 2, D and E, and fig. S5). By directly measuring the GSH level of the extracted liver ex vivo, we further confirmed that DEM treatment led to a marked decrease in liver GSH level from ~6.6 μmol/g liver tissue in the control mice to ~3.1 and ~ 0.6 μmol/g liver tissue in the mice receiving 0.3 and 0.6 mg/kg body weight DEM, respectively. By plotting the liver GSH levels against the liver fluorescence activation kinetics of ICG4-GS-Au25, we found that the corresponding activation kinetics of the liver ICG fluorescence decreased from 71 s−1 to 33 and 12 s−1 and exhibited a strong linear correlation with the decrease in liver GSH level (Pearson’s r = 0.95; Fig. 2F), consistent with the observation at the in vitro level. To further explore whether liver enzymes such as γ-glutamyl transferase (GGT) could also activate ICG emission, we incubated ICG4-GS-Au25 with GGT, the only enzyme that is known to cleave the γ-glutamyl moiety of GSH, in PBS at 37°C and monitored the activation of ICG fluorescence. The results show that fluorescence activation kinetics of ICG4-GS-Au25 in 50 U/liter GGT (normal GGT level in blood) was 0.04% of that in 5 mM GSH (0.0026 ± 0.002 s−1 versus 5.97 ± 0.17 s−1), and even in 5000 U/liter GGT, the activation kinetics (0.31 ± 0.02 s−1) is only ~5% of that in 5 mM GSH, indicating that GGT had little effect on the displacement of ICG from Au25 (fig. S6). Combining these results suggests the displacement of ICG from Au25 is mainly dictated by hepatic GSH in vivo and the linear correlation of liver fluorescence activation kinetics with hepatic GSH level in vivo lays down a foundation for further understanding of activation process in the diseased mouse model.

Fluorescence monitoring of APAP-induced hepatic GSH depletion

Previous ex vivo studies have shown that dose-dependent toxins such as APAP induce hepatic GSH depletion after overdose (5, 6, 28); however, hepatic GSH depletion has never been noninvasively detected with in vivo fluorescence imaging technique (2830). To answer this question, we established the well-known APAP-induced liver injury mouse model by intraperitoneally administering the mice with an overdose of APAP (300 mg/kg body weight) (31), a normal dose of APAP (60 mg/kg body weight), or saline as control. At 30 min after APAP administration, in vivo fluorescence imaging of the mice was conducted after intravenous injection of the same amount of ICG4-GS-Au25 (~3 nmol). As shown in Fig. 3 (A and B), the liver ICG fluorescence activation in APAP-overdosed mice was significantly slower than that of the control and normally dosed mice (see fig. S7 for data of individual mouse). Quantitative analysis showed that the liver fluorescence activation kinetics of the overdosed mice was 20 ± 1.6 s−1, ~50% of those of the control mice (41.6 ± 7.9 s−1) and the normally dosed mice (40.4 ± 7.1 s−1) (Fig. 3C). On the other hand, the Au accumulation in the liver of the APAP-overdosed mice [7.5 ± 0.59% injected dose (ID)/g] was even slightly higher than that of the control mice (6.1 ± 0.16% ID/g) at 5 min after ICG4-GS-Au25 administration (fig. S8). Thus, the marked decrease in hepatic fluorescence activation kinetics was not a result of lower liver accumulation of the nanoprobes. To further determine the origin of the decreased activation kinetics, we ex vivo quantified GSH levels of the extracted liver of the control mice and the mice receiving normally dosed APAP and overdosed APAP immediately after the in vivo fluorescence imaging. As shown in Fig. 3D, the hepatic GSH level of overdosed mice was 2.1 ± 1.2 μmol/g tissue, only ~32 to 35% of those of the control mice (6.0 ± 0.1 μmol/g tissue) and the normally dosed mice (6.6 ± 0.8 μmol/g tissue). Combining these results clearly indicates that the significant decrease in hepatic fluorescence activation kinetics was due to the deficiency of hepatic GSH in APAP-overdosed mice.

Fig. 3 Monitoring of drug-induced hepatic GSH depletion with ICG4-GS-Au25 through fluorescence imaging.

(A) Representative noninvasive in vivo fluorescence images of mice show the activation of ICG fluorescence in the liver (pointed by arrows) following intravenous injection of the same dose of ICG4-GS-Au25 in mouse model of APAP-induced liver injury. Mice were intraperitoneally injected with an overdose of APAP (300 mg/kg), a normal dose of APAP (60 mg/kg), or saline (0 mg/kg) as control at 30 min before the injection of ICG4-GS-Au25. (B) Time-dependent liver ICG fluorescence curves of APAP-treated mice after injection of the same ICG4-GS-Au25. (C) Comparison of liver ICG fluorescence kinetics of mice treated with different doses of APAP. N = 3 mice for control and 60 mg/kg groups, and N = 4 mice for 300 mg/kg group. (D) Liver GSH level measured ex vivo immediately after in vivo fluorescence imaging. (E) Consecutively monitor liver GSH level in the same group of APAP-overdosed mice through repeated administration of ICG4-GS-Au25. APAP (300 mg/kg body weight) was intraperitoneally administered to mice at 0 hour, and the liver fluorescence imaging was performed at 4 hours before APAP injection (normal status) and at 0.5, 4, 8, and 18 hours after APAP injection. N = 3 mice for consecutive liver fluorescence imaging. The liver GSH levels at different time points before and after APAP injection were directly measured ex vivo using liver tissues of five groups of mice (N = 3 for each group). Statistical significance is evaluated by two-sample equal variance t test (P < 0.05; statistically significant). Data points are presented as means and SD.

Linear correlation of fluorescence activation kinetics of ICG4-GS-Au25 in the liver with hepatic GSH level also allowed consecutive imaging of hepatic GSH depletion and recovery in individual APAP-overdosed mouse at high temporal resolution. Because of the efficient biotransformation and hepatobiliary clearance of ICG-GS, the ICG fluorescence signals in the blood (half-life ~11 min) and liver (half-life ~33 min) decayed rapidly after injection of ICG4-GS-Au25 (fig. S9), making it feasible to readminister the nanoprobe for consecutive imaging of hepatic GSH level during the progression of liver injury. We repeatedly administered the nanoprobes to the same mice at 4 hours before APAP injection (normal status) and at 0.5, 4, 8, and 18 hours after APAP injection and monitored their liver ICG fluorescence activation kinetics, which were found to correlate well (Pearson’s r = 0.87; fig. S10) with the liver GSH level at the respective time points (Fig. 3E). The liver GSH level was detected to drop rapidly from 6.74 ± 0.32 μmol/g tissue (normal status) to 2.88 ± 0.34 μmol/g tissue [4 hours postinjection (p.i.)] after APAP overdose and recovered to 6.40 ± 0.78 μmol/g tissue (18 hours p.i.) gradually afterward, consistent with the previous reports (30).

Monitoring APAP-induced hepatic GSH depletion through blood testing

While in vivo liver fluorescence imaging has been successfully used to image APAP-induced GSH depletion in the mouse model at high specificity, it is challenging to translate this fluorescence imaging to large animals and even humans due to the limited tissue penetration depth of light. In addition, we also monitored the serum GSH level after APAP overdose and found that serum GSH level did not change significantly even after 2 hours of APAP administration (fig. S11), suggesting that serum GSH level was much less sensitive to drug-induced liver injury than hepatic GSH level, likely due to the fact that serum GSH is more than two orders of magnitude less abundant than that in the liver and is also consumed by other organs. Since ICG4-GS-Au25 returned to blood circulation after GSH-mediated biotransformation in the liver sinusoids, the hepatic GSH depletion might potentially be also monitored in the blood. To validate this feasibility, we first investigated the stability of ICG4-GS-Au25 in the blood plasma. As shown in Fig. 4A, very little (less than 5%) activation of the ICG4-GS-Au25 was observed after 20 min of incubation with fresh blood due to the high activation threshold of ICG4-GS-Au25. Since fluorescence activation of ICG on the Au25 is almost exclusively dependent on hepatic GSH level, more ICG molecules are expected to remain on the Au25 surface after going through hepatic GSH–mediated biotransformation and returning to blood circulation when liver GSH level is significantly reduced due to APAP overdose. Thus, ex vivo activation of the nanoprobe in peripheral blood with dithiothreitol (DTT) will lead to much higher blood fluorescence on/off ratios in the mice with overdosed APAP relative to those of the normal mice (Fig. 4B). As shown in Fig. 4C, at 30 min after APAP administration, ICG4-GS-Au25 was intravenously injected, and two peripheral blood samples (~10 μl) were acquired at 1 and 15 min afterward. The blood samples of mice injected with overdose APAP, normal dose APAP, and saline (control) all exhibited weak ICG fluorescence when being measured directly after the withdrawal; however, after incubation with DTT, the blood ICG fluorescence of APAP-overdosed mice was much stronger than that of control and normally dosed mice (Fig. 4D), allowing us to quantitatively compare the blood fluorescence on/off ratio before and after DTT treatment. Shown in Fig. 4E, the blood fluorescence on/off ratios of APAP-overdosed mice at 1 (15.6 ± 1.4) and 15 min p.i. (12.1 ± 1.1) were ~2- and ~2.6-fold higher than those (7.6 ± 0.6 and 4.6 ± 0.7) of the control mice, respectively, because of the significantly lowered hepatic GSH level in APAP-overdosed mice. On the other hand, the on/off ratios of normally dosed mice at 1 min p.i. (8.4 ± 0.2) and 15 min p.i. (4.6 ± 1.0) were comparable to those of the control mice at both time points. To evaluate the accuracy of the nanoprobe-based blood test in the detection of hepatic GSH depletion, we used ex vivo–measured liver GSH level as the “gold standard” and conducted a blinded receiver operating characteristics (ROC) analysis in a group of 20 mice. At 30 min p.i. of saline (normal control) or various doses of APAP (60, 150, and 300 mg/kg body weight), we intravenously administered ICG4-GS-Au25 into the mice and then collected the blood sample at 1 min after probe injection, followed by immediately collecting liver tissues. For each mouse, we obtained the ICG fluorescence on/off ratio of the blood samples and directly measured the GSH level of the extracted liver ex vivo. On the basis of the liver GSH level of the normal control mice (5.34 to 6.96 μmol/g tissue), the APAP-injected mice were divided into two groups, with and without hepatic GSH depletion, using the lowest liver GSH level (5.34 μmol/g tissue) of the normal control mice as the threshold (fig. S12). Our results show that the area under the ROC curve of the blood test reached 0.93 (Fig. 4F), clearly indicating high accuracy and specificity of ICG4-GS-Au25 in noninvasive detection of hepatic GSH depletion through blood testing. Moreover, by readministration of the ICG4-GS-Au25 nanoprobe to the same group of mice and multiple blood samplings, we were also able to consecutively monitor the rapid reduction and subsequent recovery of hepatic GSH level during the progression of APAP-induced liver injury (Fig. 4G), consistent with the previous finding obtained by fluorescence imaging (Fig. 3E).

Fig. 4 Monitoring APAP-induced hepatic GSH depletion with ICG4-GS-Au25 through blood test.

(A) ICG fluorescence of the same ICG4-GS-Au25 after 20 min of incubation in fresh mouse serum or mouse serum containing 10 mM DTT. (B) Schematic of the rationale for blood sampling–based detection method. (C) Timeline of the experiment. IV, intravenous. (D) Typical ICG fluorescence of peripheral blood samples from mice treated with 0, 60, and 300 mg/kg body weight APAP at 1 min after the injection of ICG4-GS-Au25 nanoprobe. DTT was used to release ICG-GS from the surface of Au25 nanocluster. (E) Peripheral blood ICG fluorescence on/off ratio at 1 and 15 min after the injection of ICG4-GS-Au25 nanoprobe at 30 min after APAP treatment. N = 3 mice for control and 60 mg/kg groups, and N = 4 mice for 300 mg/kg group. (F) ROC curves of ICG4-GS-Au25–based blood test for detecting APAP-induced hepatic GSH depletion. Mice (N = 5 for each group) were randomly dosed with 0, 60, 150, or 300 mg/kg body weight APAP; and blinded blood tests were performed 30 min afterward. The results were verified using ex vivo–measured liver GSH level as the gold standard. (G) Consecutively monitor liver GSH level via blood test in the same APAP overdosed mice (N = 3) through repeated administration of ICG4-GS-Au25. APAP (300 mg/kg body weight) was intraperitoneally (IP) administered to mice at 0 hours (indicated by the arrow), and the peripheral blood was sampled at 1 and 15 min after the injection of ICG4-GS-Au25 at multiple time points before and after APAP treatment. Statistical significance is evaluated by two-sample equal variance t test (P < 0.05; statistically significant). Data points are presented as means and SD.

Guiding early remedy of APAP-induced liver injury with ICG4-GS-Au25

While serum ALT has been used as the gold standard liver injury biomarker to guide the treatment of liver injuries, its low sensitivity and delayed response make it very challenging for early remedy of liver injuries. As shown in Fig. 5A, serum ALT level did not significantly increase even 4 hours after overdosing APAP and only exhibited a less than twofold increase over the normal value at 5 hours after APAP administration, which is often overlooked in the clinics. However, if being left untreated, then liver damage was found to progress rapidly, and the serum ALT level would exponentially increase by over 100-fold at 12 hours after APAP administration. Such low sensitivity of ALT to the liver injury could delay the remedy and reduce the efficacy of N-acetyl cysteine (NAC), a precursor for hepatic GSH biosynthesis used clinically as an antidote to APAP poisoning. As shown in Fig. 5B, if NAC was used to treat APAP-induced liver injury 5.5 hours after the mice were being overdosed with APAP, then the serum ALT level (3907 ± 330 U/liter) was just slightly below that of the mice receiving the same APAP dose but without NAC treatment (4950 ± 327 U/liter) at 12 hours after APAP administration, and severe hemorrhage on the liver tissue as well as the widespread death of liver cells were still observed. On the other hand, if the NAC treatment was conducted just 3.5 or even 1.5 hours earlier, then the ALT levels were substantially lowered by 35-fold (111 ± 7 U/liter) or 14-fold (279 ± 71 U/liter), respectively, indicating that the NAC therapeutic efficacy was significantly improved when being administered earlier. These findings were consistent with the cell death analysis of liver tissue via the terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL) assay (Fig. 5C), highlighting the importance of early remedy of the liver injury.

Fig. 5 Monitoring hepatic GSH depletion in APAP-induced liver injury allows early treatment and improves prognosis.

(A) Serum ALT levels of mice (N = 3) at multiple time points after receiving 300 mg/kg body weight APAP. Inserted figure shows the serum ALT levels in the first 5 hours. (B) Serum ALT levels of mice (N = 3) at 12 hours p.i. of overdose APAP with NAC treatment at different time points after APAP administration. (C) Representative color pictures of the liver (12 hours after APAP administration) and fluorescence microscope images of cell death in liver tissues (detected using TUNEL assay) from APAP overdosed mice without NAC treatment or receiving NAC treatment at 30 min (early treatment) and 5.5 hours (delayed treatment) after APAP administration. Cell nuclei were counterstained with DAPI. Scale bars, 20 μm. (D) Serum ALT levels (12 hours p.i. of APAP or saline) of mice (N = 3 for each group) with or without NAC treatment after diagnosis of hepatic GSH depletion based solely on ICG4-GS-Au25 blood test. The blinded blood tests were performed at 0.5 hour after the mice were dosed with saline or 300 mg/kg body weight APAP, and NAC treatment was conducted after analysis of the test results (~1 hour after administration of APAP). Statistical significance is evaluated by two-sample equal variance t test (P < 0.05; statistically significant). Data points are presented as means and SD. Photo credit: X. Jiang, The University of Texas at Dallas.

Since depletion of hepatic GSH induced by overdosed hepatotoxic drugs such as APAP is known to occur earlier than the liver cell death and the elevation of liver biomarkers (e.g., ALT) (23, 32), early detection of hepatic GSH depletion is expected to help treat the liver injury more effectively. To demonstrate the application of ICG4-GS-Au25 in guiding the early remedy of drug-induced liver injury, we conducted a blinded test by using ICG4-GS-Au25–based blood test at 30 min after APAP injection to differentiate a mixed group of nine mice injected with either PBS or overdosed APAP according to their liver GSH depletion status, followed by giving NAC treatment to half of the mice diagnosed with depleted hepatic GSH. With blood fluorescence on/off ratios (8.2 ± 1.1) of normal control mice in the ROC study serving as the normal reference range, we were able to successfully identify all six mice injected with overdosed APAP (on/off ratios, 10.3 to 14.9) and the rest of the mice injected with PBS (on/off ratios, 7.6 to 8.0) based solely on their blood fluorescence on/off ratios (fig. S13). Furthermore, our results show that the serum ALT levels of mice that were diagnosed with liver GSH depletion and received NAC treatment (122.4 ± 43.9 U/liter) were 25-fold lower than those without NAC treatment (3102.3 ± 864.3 U/liter) after diagnosis (Fig. 5D). These findings validate the feasibility of using ICG4-GS-Au25 and blood testing to guide early remedy of liver injuries induced by the overdosed APAP.

Evaluating the biocompatibility of ICG4-GS-Au25

The ICG4-GS-Au25 nanoprobe also demonstrated excellent biocompatibility. The Au25 nanoclusters were efficiently cleared through the urine (27.4 ± 4.9% ID) with minimal accumulation in the liver (5.0 ± 0.1% ID) and other major organs at 24 hours after administration of ICG4-GS-Au25 to APAP-overdosed mice (fig. S14). With this rapid clearance, repeated administration of the nanoprobe did not induce hepatotoxicity, which was confirmed by the serum ALT test showing statistically comparable levels of ALT between repeatedly dosed mice (41.8 ± 7.38 U/liter) and the control ones (38.8 ± 5.52 U/liter) (fig. S15). In addition, liver endothelial integrity was also uncompromised after repeated dosing, as evaluated by a liver pathologist through histology (fig. S16) and CD34 immunostaining (fig. S17). Compared with the control, there was no sinusoidal dilation, sinusoidal congestion, perivenular hepatocytes necrosis, or centrilobular fibrosis in the ICG4-GS-Au25–injected group to suggest endothelial injuries. Meanwhile, CD34 stain patterns showed no notable differences between the control and the ICG4-GS-Au25–injected group and were restricted to periportal vasculatures. No CD34 loss or capillarization was present to suggest endothelial injures. These findings lay down a foundation for the future clinical translation of ICG4-GS-Au25.

Early detection of other drug-induced liver injury with ICG4-GS-Au25

Not limited to APAP, many other hepatotoxic drugs (6, 28, 33), heavy metals (34, 35), and chemicals (17, 29, 36) can also injure the liver cells through a similar GSH depletion mechanism. Considering that hepatic GSH depletion is a common phenomenon that precedes many types of liver injuries and diseases, this strategy should also be applied to other drugs as well. For example, we further showed that this strategy could also be used to prognosticate liver injuries induced by chlorpromazine (CPZ; Fig. 6A), a widely used antipsychotic drug that is hepatotoxic if overdosed. We intravenously administered ICG4-GS-Au25 to mice at 0.5 hour after they were overdosed with CPZ. As shown in Fig. 6 (B to D), the liver fluorescence activation in CPZ-treated mice was significantly decreased compared with that of the control mice, with activation kinetics of CPZ-overdosed mice (20.7 ± 0.1 s−1) only ~45% of that of the control mice (45.6 ± 1.5 s−1). Meanwhile, the blood fluorescence on/off ratio in CPZ-overdosed mice (9.6 ± 1.1) was found to be 1.8-fold higher than that of the control mice (5.5 ± 0.5) (Fig. 6E). All these results indicated a significant loss of hepatic GSH in CPZ-overdosed mice, which was confirmed by direct measurement of the liver GSH level ex vivo (Fig. 6F). While substantial hepatic GSH depletion can be detected as early as 0.5 hour after a CPZ overdose, the serum ALT level remained unchanged at that time but increased over fivefold at 12 hours (Fig. 6G), suggesting that the occurrence of liver injury after hepatic GSH was depleted by the overdosed CPZ.

Fig. 6 Early detection of CPZ-induced liver injury by ICG4-GS-Au25–enabled fluorescence imaging and blood test of hepatic GSH depletion.

(A) Structure of the antipsychotic drug chlorpromazine (CPZ). CPZ was administered to the mice intraperitoneally at a dose of 100 mg/kg body weight. (B) Representative in vivo fluorescence imaging of control (PBS-treated) and CPZ-treated mice at 0.5 hour after drug administration with intravenous injection of the same ICG4-GS-Au25. (C) Time-dependent liver ICG fluorescence following the injection of ICG4-GS-Au25 (N = 3 mice for each group). (D) Liver ICG fluorescence activation kinetics of control and CPZ-treated mice. (E) The ICG fluorescence on/off ratio of the peripheral blood at 1 min after the injection of ICG4-GS-Au25 at 0.5 hour post CPZ treatment. (F) Measured liver GSH level of the control and CPZ-treated mice at 0.5 hour after CPZ administration confirms the decrease in hepatic GSH level by CPZ. (G) Serum ALT levels of mice before and at 0.5 and 12 hours after the administration of 100 mg/kg body weight CPZ. Statistical significance is evaluated by two-sample equal variance t test (P < 0.05; statistically significant).

DISCUSSION

As the major detoxification organ in the body, the liver is frequently injured by various xenobiotics such as overdosed drugs, heavy metals, and alcohol. In particular, drug-induced liver injury has been the primary reason for nonapproval of new drugs and the market withdrawal of existing drugs, as well as the leading cause of acute liver failure in many countries (37). Thus, early detection of liver injuries is crucial to both preclinical drug development and reducing the liver failure incidents in the clinics. To date, the detection of liver injuries still relies heavily on the endogenous serum biomarkers (e.g., ALT) released into the blood after the death of liver cells because of their reliability and easiness to perform in clinical settings (38); however, these serum biomarkers are not sensitive enough for early detection of liver injuries and preclude early remedy of liver injuries. Despite that more sensitive exogenous probes have been developed to optically image liver injuries through detecting reactive or oxidative species (31, 39, 40), none of them can be used for blood testing, and their future clinical application is severely limited by the shallow tissue penetration depth of light.

Hepatic GSH depletion is a key pathophysiological event involved in many liver injuries and occurs before liver cell death and elevation of conventional liver biomarkers such as ALT. However, this key pathophysiological event has never been monitored using either simple fluorescence imaging or blood testing. By leveraging hepatic GSH–mediated biotransformation in the liver sinusoids, we demonstrated that ultrasmall GSH-activatable ICG4-GS-Au25 can serve as a powerful fluorescent probe and an exogenous blood marker that enable hepatic GSH depletion and recovery in the mice with drug-induced liver injuries to be noninvasively and consecutively monitored with fluorescence imaging and blood testing at high specificity and accuracy, respectively. The observed high specificity and accuracy fundamentally originate from the unique liver transport of ICG4-GS-Au25 nanoprobes and their selective hepatic GSH–mediated biotransformation in the liver sinusoids, so that their fluorescence activation kinetics in the liver is linearly dependent on hepatic GSH level. The early detection of hepatic GSH depletion with this unique probe also made it possible to precisely guide early remedy of the liver injuries before the elevation of conventional liver biomarkers, so that therapeutic efficacy in the treatment of liver injury can be significantly improved. Ultrasmall size of Au25 and hepatic GSH–mediated dissociation of ICG from the surface of Au25 eventually lead to efficient body clearance of the nanoprobe and minimized nonspecific accumulation, which greatly enhances its feasibility in future clinical translation. These findings not only address a long-standing challenge in noninvasive monitoring of hepatic GSH depletion involved in many diseases beyond liver injuries but also highlight importance of fundamental understandings of specific nanobiochemical interactions in vivo and physiology at the nanoscale, which are expected to broaden biomedical applications and clinical translation of nanomedicines.

MATERIALS AND METHODS

Materials and equipment

ICG–N-hydroxysuccinimide (NHS) was purchased from Intrace Medical (Switzerland), while all the other chemicals were obtained from Sigma-Aldrich (USA) and used as received unless specified. The absorption spectra were measured with a Varian 50 Bio ultraviolet-visible spectrophotometer. Fluorescence spectra were acquired with a Photon Technology International (PTI) QuantaMaster 30 fluorometer. The size of nanoparticles was measured by a 200-kV JEOL 2100 transmission electron microscope. In vivo fluorescence images were recorded using a Carestream In Vivo FX Pro imaging system. Optical imaging of liver tissue slides was conducted with an Olympus IX-71 inverted fluorescence microscope coupled with PhotonMAX 512 charge-coupled device camera (Princeton Instruments). Animal studies were performed according to the guidelines of the University of Texas System Institutional Animal Care and Use Committee. BALB/c mice (strain code 047, 6 to 8 weeks old, weighing 20 to 25 g) were purchased from Envigo. All mice were randomly allocated and housed under standard environmental conditions (23 ± 1°C, 50 ± 5% humidity, and 12/12-hour light/dark cycle) with free access to water and standard laboratory food.

Synthesis of ICG4-GS-Au25 nanoprobes

Atomically monodisperse Au25(SG)18 (GS-Au25) nanoclusters were first synthesized according to the reported method (41). For the synthesis of ICG4-GS-Au25, 4 mg of ICG-NHS (dissolved in dimethyl sulfoxide) was added into 6 mg of GS-Au25 aqueous solution, and the mixture was vortexed for 3 hours. Then, ICG-GS-Au25 conjugates were purified after removing the unconjugated ICG dye through centrifugation in the presence of ethanol. The conjugates were again redispersed in 1× PBS buffer and purified by 30-kDa Amicon Ultra Centrifugal filters to remove any unconjugated GS-Au25 nanoclusters. The resulting ICG4-GS-Au25 inside the centrifuge filter was resuspended in ultrapure water and lyophilized for future usage.

Depletion of the liver GSH by DEM and quantification of the liver GSH level

The variation of liver GSH level was achieved by a single injection of different doses of DEM. DEM was intraperitoneally administered into BALB/c mice at doses of 0, 0.3, and 0.6 ml/kg body weight ~30 min before the injection of ICG4-GS-Au25 nanoprobes. Immediately after the in vivo liver fluorescence imaging, mice were euthanized, and their livers were harvested and promptly snap frozen in liquid nitrogen for the determination of the liver GSH level. Quantification of the liver GSH level was carried out by a modified Tietze enzymatic recycling assay reported previously (42).

In vivo liver fluorescence imaging with ICG4-GS-Au25

Hair-removed BALB/c mice were pretreated with saline (control), DEM, or APAP. Then, under 3% isoflurane anesthesia, mouse was tail vein catheterized and prone positioned on the imaging stage. ICG4-GS-Au25 (20 μM in PBS) (150 μl) was then tail vein injected under sequential time series imaging collection for ~4 min. The fluorescence imaging parameters were set as follows: excitation, 760 nm; emission, 830 nm; 7-s exposure time; 2 × 2 binning.

APAP-induced liver injury animal model

Healthy male BALB/c mice were prefasted for 5 hours and then intraperitoneally injected with APAP saline solution (~28 mg/ml) at an overdose of 300 mg/kg body weight. As control, the prefasted mice were intraperitoneally injected with APAP at a normal dose of 60 mg/kg body weight or an equivalent volume of saline. The normal APAP dose of 60 mg/kg body weight was derived from the maximum recommended daily dose (3 to 4 mg) of a healthy adult human (~70 kg). Food was immediately retrieved after the injection of APAP or saline.

Serum ALT level test

Blood was withdrawn retro-orbitally from mice at specific time points and placed on ice for 30 min to allow for coagulation. Then, the blood samples were centrifuged at 10,000g at 4°C to separate serum from blood cells. The serum samples were collected and stored at −20°C until ALT test. The serum ALT activity was measured colorimetrically with the ALT Activity Assay Kit (Sigma-Aldrich, catalog no. MAK052).

Measurement of the ICG fluorescence on/off ratio in blood sample

At specific time points following the intravenous injection of ICG4-GS-Au25, ~10 μl of blood was withdrawn retro-orbitally from the mouse and immediately placed in a glass vial containing 500 μl of 2% EDTA PBS solution. After this, the off state blood ICG fluorescence in the vial was promptly recorded using the in vivo fluorescence imaging system. Then, 500 μl of DTT solution (40 μM, pH 7.4) was added to the vial to release ICG-GS from the Au25 surface, and the on state blood ICG fluorescence in the vial was recorded at the same imaging condition after 10 min of incubation. The on/off ratio of the blood ICG fluorescence was then calculated on the basis of these two measurements.

TUNEL staining of liver tissue

Liver tissues were excised promptly from euthanized mice and immediately fixed with 4% freshly prepared paraformaldehyde PBS solution for 2 hours. Then, the fixed liver tissues were immersed in 20% sucrose PBS solution overnight at 4°C before being embedded in optimal cutting temperature (OCT) freezing compound. The embedded liver tissues were sectioned into 5-μm-thick slides in cryostat and placed on Superfrost Plus Microscope Slides (Fisherbrand) for TUNEL staining. The TUNEL staining was accomplished using the ApopTag Fluorescein In Situ Apoptosis Detection Kit (MilliporeSigma, catalog no. S7110). Cell nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) before slides were mounted and subject to fluorescence microscopy.

ROC analysis of ICG4-GS-Au25–based blood test

Twenty mice were randomly intraperitoneally injected with saline and 60, 150, and 300 mg/kg body weight APAP (n = 5 mice for each dose) in a blinded fashion, and ICG4-GS-Au25 (150 μl 20 μM per mouse) was administrated intravenously at 30 min p.i. of APAP or saline. The blood samples were retro-orbitally collected at 1 min p.i. of ICG4-GS-Au25, and the ICG fluorescence on/off ratio was measured for ROC analysis. The liver tissues were immediately collected after blood sampling for ex vivo quantification of hepatic GSH concentration, which served as the gold standard to compare with the blood test result. Mice with liver GSH level lower than the lowest liver GSH level of the control mice were classified as hepatic GSH depletion; otherwise, the mice were classified as with normal hepatic GSH level.

NAC treatment of mice with APAP-induced liver GSH depletion diagnosed by ICG4-GS-Au25

Mice were randomly intraperitoneally injected with saline (n = 3 mice) or 300 mg/kg APAP (n = 6 mice), and the ICG4-GS-Au25–based blood test was carried out 0.5 hours afterward in a blinded fashion. The blood samples were collected at 1 min p.i. of ICG4-GS-Au25 (150 μl 20 μM per mouse), and the ICG fluorescence on/off ratio was measured for diagnosis of liver GSH depletion. Blood fluorescence on/off ratios of the normal control mice in the ROC study were used as the normal reference. The mice with blood fluorescence on/off ratios that fall within the means ± SD of the normal reference were identified as those that had normal liver GSH; otherwise, the mice were identified as having depleted liver GSH. After analysis of the blood test results (~1 hour after APAP injection), half of the liver GSH–depleted mice were treated with intraperitoneal injection of 900 mg/kg body weight NAC, while the other half received no treatment. The serum ALT levels of all mice were measured at 12 hours after AAP injection to evaluate liver injury.

CPZ-induced liver injury animal model

Healthy male BALB/c mice were prefasted for 5 hours and then intraperitoneally injected with CPZ saline solution (~10 mg/ml) at an overdose of 100 mg/kg body weight. As control, the prefasted mice were intraperitoneally injected with equivalent volume of saline. Food was immediately retrieved after the injection of CPZ or saline.

SUPPLEMENTARY MATERIALS

Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/7/8/eabd9847/DC1

https://creativecommons.org/licenses/by-nc/4.0/

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.

REFERENCES AND NOTES

Acknowledgments: Funding: We acknowledge the financial support, in part, from the National Institutes of Health (NIH) (R01DK103363 and R01DK115986), Cancer Prevention Research Institute of Texas (CPRIT) (RP200233), Welch Research Foundation (AT-1974-20180324), and Cecil H. and Ida Green Professorship in Systems Biology Science from the University of Texas at Dallas (to J.Z.). Author contributions: J.Z. conceived the idea and designed the experiments with X.J.; X.J. and Q.Z. conducted the experiments with the assistance of B.D., S.L., Y.H., M.Y., and Z.C.; X.J., Q.Z., W.M.L., M.Y., and J.Z. discussed and analyzed the results; and J.Z., X.J., and Q.Z. composed the manuscript. All authors commented on the manuscript. Competing interests: A patent application derived from the studies has been filed. The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.

Stay Connected to Science Advances

Navigate This Article