1.

Introduction

As a major metabolic organ, the liver is susceptible to toxic effects from drug molecules or environment toxins. In the United States, drug-induced hepatotoxicity is a major clinical problem and accounts for more than 50% of the cases of acute liver failure.¹ Acetaminophen and temozolomide are some of the medications contributing to hepatotoxicity.^2–5 Clinically, chronic hepatitis can lead to fibrosis and cirrhosis of the liver. To study the hepatotoxic effects of drugs or toxins, acetaminophen and carbon tetrachloride (${\mathrm{CCl}}_4$) are popular models.^6–8 In the case of ${\mathrm{CCl}}_4$, detoxification involves converting of the molecule into free radicals through the oxidase pathway. Eventually, the products of carbon dioxide (${\mathrm{CO}}_2$), carbon monoxide (CO), and hydrogen chloride (HCl) are produced.^9,10

Toxicology, similar to many other biomedical disciplines, relies largely on histological techniques for studying the effects of toxins on hepatic physiology. Therefore, the lack of in vivo imaging data prevents the understanding of the mechanism of toxin-induced hepatotoxicity and dynamics of toxin-induced effects on hepatobiliary metabolism. In this study, we aim to address these issues by performing intravital multiphoton microscopy on mouse models that have been treated with ${\mathrm{CCl}}_4$. By imaging at different time points following ${\mathrm{CCl}}_4$ induction and performing analysis through an image-based kinetic model, we can derive image maps of rate constants depicting spatial variations of metabolic capabilities of the mouse liver.

2.

Materials and Methods

The multiphoton fluorescence microscope used in this study is a home-built laser scanning unit based on an inverted microscope (Nikon TE-2000, Tokyo, Japan). The excitation source was a titanium-sapphire laser (Tsunami, Spectra-Physics, Mountain View, California) pumped by a diode-pumped solid-state laser (Millennia X, Spectra-Physics). The 780-nm laser beam was scanned in the focal plane by an x-y mirror scanning system (Model 6220, Cambridge Technology, Cambridge, Massachusetts). An oil immersion objective ($20\times /\mathrm{NA}0.75$, Plan Fluor, Nikon, Japan) was used to focus the laser beam onto the sample. With an on-sample power of around 50 mW, second harmonic generation (SHG) and fluorescence emission can be effectively excited and collected in the epi-illuminated geometry and separated into four simultaneous detection channels by dichroic mirrors (435DCXR, 495DCXR, 555DCLP, Chroma Technology, Rockingham, Vermont) and band-pass filters (HQ390/20, HQ460/50, HQ525/50, HQ610/75, Chroma Technology). In this configuration, SHG at 390 nm can be used to locate the surface of the liver by imaging the capsule (HQ390/20). Furthermore, the blue fluorescence channel can be used to identify autofluorescent hepatic stellate cells (HQ460/50). Finally, green (HQ525/50) and red (HQ610/75) fluorescence can be used to image the distribution of 6-CF and RITC-dextran labeled sinusoids. Each acquired multiphoton image is $200\times 200{\mu \mathrm{m}}^2$ in area and composed of $512\times 512\text{pixels}$. Following the acquisition of each image, the mouse, mounted on a specimen translation stage (Prior Scientific, United Kingdom), was translated to the next position of the liver. Sequential multiphoton images are then acquired to generate a large-area image composed of $3\times 3$ optically scanned images. In this manner, $600\times 600{\mu \mathrm{m}}^2$ of the mouse liver can be imaged in 1 min.

Prior to the acquisition of multiphoton images, wide-field fluorescence microscopy was used to identify the hepatic acinus structure. The 514-nm output of an argon laser was used for excitation. The fluorescence emission passed through a combination of a dichroic mirror (525DCLP, Chroma Technology) and band-pass filter (HQ590/80, Chroma Technology), before being imaged by a high sensitivity EM-CCD camera (iXon DV887DCS-BV, Andor Technology, Ireland).

The mice model used in this study were 7 to 8 weeks old C57BL/6 male mice and were maintained in a temperature and humidity controlled facility ($22\pm 1°\mathrm{C}$, $60\%\pm 5\%$), where the lighting is controlled in a 12-h on/12-h off cycle. For direct imaging of the mouse liver, an intravital hepatic imaging chamber (IHIC), made of 6Al/4V medical grade titanium alloy and a cover glass, was installed on the mouse abdomen. For IHIC installation, anesthetic, such as Zolitle 50, was applied. For ${\mathrm{CCl}}_4$ induction, $100\mu \mathrm{l}$ of ${\mathrm{CCl}}_4$ [40% (vol/vol)] was injected into the mice twice per week. To visualize the sinusoids, rhodamine B isothicyanate (RITC)-dextran with molecular weight of 70,000 was intravenously injected into the mouse. Finally, the metabolism of 6-carboxyfluorescein diacetate (6-CFDA) into 6-carboxyfluorescein (6-CF) was used for quantifying hepatobiliary metabolism. Five mice were used at each time point. The procedures for mouse treatment are similar to those described previously.¹¹ In order to minimize animal suffering, results of normal mice were obtained from a previous study and not repeated for this work.¹² Care and experimental protocols of the mouse model were in accordance to the regulations of the National Taiwan University Institutional Animal Care and Use Committee.

After multiphoton imaging had been performed, we used a first-order kinetic model [Eq. (1)] to fit the time-lapse images:

In this model, $k_1$ is the rate constant used to describe uptake and intracellular esterase processing of 6-CFDA into 6-CF. $k_2$ is the rate constant, which accounts for the process of 6-CF excretion into the bile canaliculi. ${[6-\mathrm{CFDA}]}_0$ is initial 6-CFDA sinusoidal concentration. ${[6-\mathrm{CF}]}_{\mathrm{H}}$ is hepatocyte 6-CF concentration. $V_{\mathrm{B}}$ and $V_{\mathrm{H}}$ are the volume of the mouse blood and a hepatocyte, respectively.¹²

3.

Results and Discussion

Representative time-lapse multiphoton images 1 day post-${\mathrm{CCl}}_4$-induced hepatotoxicity are shown in Fig. 1. As shown in the images, 6-CF fluorescence appeared at 1 min after 6-CFDA injection. The sinusoids are well marked by the high molecular weight RITC-dextran. Since we know the difference in time between sequentially acquired images, Eq. (1) can be used to fit for the kinetic rate constants $k_1$ and $k_2$ at each position. After processing of the time-lapse images, rate constant-resolved images can be obtained (Fig. 2). The rate constant-resolved maps show that in response to ${\mathrm{CCl}}_4$ induction, different regions of the liver exhibit different metabolic response.

Fig. 1

Multiphoton imaging of 6-CFDA metabolism in day 1 post-${\mathrm{CCl}}_4$-induced hepatic injury. Time-lapse images at (a) 0 min, (b) 1 min, (c) 5 min, (d) 10 min, (e) 30 min, and (f) 60 min following 6-CFDA injection. Green: 6-CF fluorescence and red: rhodamine B-dextran labeled sinusoids. Scale bar: 100 μm.

Fig. 2

(a) and (b) Kinetic rate constants ($k_1$ and $k_2$) resolve map of hepatobiliary metabolism in mouse liver (day 1 post-${\mathrm{CCl}}_4$ injection).

In addition to the day 1 results, we also obtained time-lapse images of mice at days 4, 7, and 14 following ${\mathrm{CCl}}_4$ injection. Representative time-lapse images are shown in Fig. 3.

Fig. 3

In vivo multiphoton imaging of hepatobiliary metabolism in ${\mathrm{CCl}}_4$-induced mice. (a) Normal, (b) day 1, (c) day 4, (d) day 7, and (e) day 14 following ${\mathrm{CCl}}_4$ injection. Green: 6-CF fluorescence and red: rhodamine B-dextran labeled sinusoids. Scale bar: $100\mu \mathrm{m}$.

Similarly, point-by-point analysis of the time-lapse images allows $k_1$ and $k_2$ resolved images to be derived. Representative results are shown in Fig. 4. Bar graphs of rate constants of hepatobiliary metabolism ($k_1$ and $k_2$) at normal, days 1, 4, 7, and 14 following ${\mathrm{CCl}}_4$-induced hepatotoxicity are shown in Fig. 5.

Fig. 4

Kinetic map in mice following ${\mathrm{CCl}}_4$-induced hepatotoxicity. (a), (b), (c), (g), and (h) $k_1$-resolved images; (d), (e), (f), (i), and (j) $k_2$-resolved images of normal, days 1, 4, 7, and 14 post-${\mathrm{CCl}}_4$-induction mice, respectively.

Fig. 5

Rate constants of hepatobiliary metabolism. (a) $k_1$ and (b) $k_2$ coefficients of normal, days 1, 4, 7, and 14 following ${\mathrm{CCl}}_4$-induced hepatotoxicity.

The rate-constant resolved images in Fig. 4 clearly show that there is a heterogeneous distribution of hepatobiliary metabolic capabilities across the acinus. In the case of normal mouse, zone 3 of the liver acinus is most active in the uptake/processing and excretion of 6-CFDA. However, with ${\mathrm{CCl}}_4$-induced hepatotoxicity, the zonal distribution of $k_1$ and $k_2$ becomes less apparent. This observation is most apparent in day 4 and is supported by the fact that the variances of the $k_1$ and $k_2$ bar graphs are the smallest for day 4 data. Interestingly, $k_1$ and $k_2$ show different trends from ${\mathrm{CCl}}_4$-induced hepatotoxicity. $k_1$ first decreased to a minimum value at day 4 and then increased toward that of the normal mouse by day 14. On the other hand, $k_2$ consistently increased once ${\mathrm{CCl}}_4$ induction has occurred (Fig. 5). We also acquired a histological image of the mouse liver at different time points (Fig. 6). Liver damage is apparent at days 2 and 4. Near normal physiological features of the liver resume at days 7 and 14.

Fig. 6

Representative histological images (hematoxylin and eosin labeled) of ${\mathrm{CCl}}_4$-induced liver injury. (a) Normal, (b) day 1, (c) day 2, (d) day 4, (e) day 7, and (f) day 14 following ${\mathrm{CCl}}_4$ induction. Scale bar: $50\mu \mathrm{m}$.

We also examined the possible mechanism of ${\mathrm{CCl}}_4$-induced hepatotoxicity. As shown in the time-lapse images in Fig. 7, we would observe an increase in intracellular 6-CF fluorescence. Unlike normal hepatocytes, the intracellular 6-CF fluorescence of these hepatocytes decreases at a significantly slower rate. This observation suggests that although the uptake and processing of 6-CFDA by intracellular esterase appear normal, membrane function of the excretion of metabolites to the bile canaliculi is compromised. As a result, the processed 6-CF molecules persist in the hepatocytes. This conclusion can only be achieved through intravital microscopic examination.

Fig. 7

Time-lapse multiphoton imaging of from ${\mathrm{CCl}}_4$-induced membrane damage in hepatocytes (white arrow). Time points are (a) 0 min, (b) 1 min, (c) 11 min, (d) 27 min, (e) 39 min, and (f) 56 min. Scale bar: $50\mu \mathrm{m}$.

4.

Conclusion

In this study, we used intravital multiphoton microscopy to study the recovery of mouse liver from ${\mathrm{CCl}}_4$-induced hepatotoxicity. In the mouse model, we found that with progression of hepatotoxicity, the zonal differences of hepatobiliary metabolism subside. Furthermore, the uptake/processing and excretion of 6-CFDA followed different kinetic pathways during the recovery process. In addition, we found that ${\mathrm{CCl}}_4$ can hinder membrane function and hinder the excretion of the processed 6-CF molecules into the bile canaliculi. By day 14, a restoration of the mouse hepatobiliary function and physiology is complete. Although multiphoton imaging can lead to the formation of cyclobutane pyrimidine dimmers,^13,14 the lack of apparent photodamage enables our approach to be used in studying the effect of chemical agents on hepatobiliary metabolism in real time and may be applied to investigating pharmacokinetics of therapeutic drug molecules in patients under laparatomic procedures.