Motivation and objectives

Although there are many studies tried to explain the mechanism of ischemia-reperfusion injury and perform real-time detection to assess the degree of organ damage by different ways during ischemia and reperfusion periods, but most of those methods provide just only spatial or temporal information at a time. It still lacks sufficient spatial and temporal information of ischemia-reperfusion injury at the same time to evaluate the state of organ damage.

Herein we want to develop an intravital autofluorescence microscopy to image

organ ischemia-reperfusion in real-time to provide spatial and temporal information at same time to help understanding and assessing the damage events during ischemia and reperfusion. For this purpose, rat liver and mouse hepatocytes were used for in vivo and in vitro studies to develop a novel method to detect ischemia-reperfusion injury in real-time. We used a bulldog clamp to interrupt blood flow to induce ischemia-reperfusion of rat liver and a homemade perfusion chamber with hypoxia or normoxia medium to simulate ischemia-reperfusion of mouse hepatocytes. Both of them were excited by 445 nm laser with suitable power which induced minimum negative effect in control group to pursue the change of autofluorescence images during ischemia-reperfusion. Furthermore, the fluorescent spectra and inhibitors of flavoproteins and electron transport chain were employed to illustrate the change of autofluorescence was attributed to the change of the redox states of flavin molecules in mitochondria. All of these results indicate that the autofluorescence of flavins can be employed in detection of ischemia-reperfusion injury and may provide subcellular information to assess the degree of damage.

Parenchymal cell Endothelial cell

Oxygenated erythrocyte Deoxygenated erythrocyte

Leukocyte Reactive species

Normal

Normal IschemiaIschemia ReperfusionReperfusion

NoNo--reflow areareflow area

Parenchymal cell Endothelial cell

Oxygenated erythrocyte Deoxygenated erythrocyte

Leukocyte Reactive species

Normal

Normal IschemiaIschemia ReperfusionReperfusion

NoNo--reflow areareflow area NoNo--reflow areareflow area

Figure 1.1: Schematic of ischemia-reperfusion injury.

During ischemia, the interruption of blood flow induces lacking of necessary substances to maintain cell survival and alterations of ion homeostasis to narrow capillaries by cell swelling. In the following reperfusion, the blood reflow area is damaged by reflow-paradox and the no-reflow area is damaged by continuous ischemic injury.

N

Figure 1.2: The change of main structure of flavin in the redox reaction.

In the redox reaction, flavin will switch between fluorescent oxidized form and non-fluorescent reduced form.

Chapter 2

Materials and Methods

2.1 Reagents

Sucrose, Tris, EGTA, antimycin A, diphenyleneiodonium chloride (DPI), dimethyl sulfoxide(DMSO), and flavin adenine dinucleotide (FAD) were purchased (Sigma Aldrich, U.S.A.).

2.2 Animals and feeding protocols

This study was approved by the Animal Investigation Committee of the National Chiao Tung University. Three male Wistar rats (The Animal Center of the National Taiwan University College of Medicine, Taiwan) aged between 7 and 9 weeks and weighing 200-250 g were used in this study. Throughout the study, all animals had free access to laboratory rodent diet (MF-18, Oriental Yeast Co., Japan) and water, and were maintained at room temperature under a cycle with 12 h light, 12 h dark.

2.3 Rat model of hepatic ischemia and reperfusion

The ischemia and reperfusion process was revised from the method reported by Abe et al.⁶⁰ All animals were fasted for 24 h before anesthetized by intraperitoneal injection with 2.5 % pentothal (70 mg/kg body wt, Abbott, Italy). After anesthesia, a midline laparotomy was performed to dissect the abdominal wall and all ligamentous attachments between liver and the diaphragm in order to expose the liver. To induce hepatic ischemia, the hepatic artery and portal vein were clamped by a bulldog clamp for designated durations of time as illustrated in the cartoon in Figure 2.1(D). After ischemia, the appearance of the liver changes from blood red to grey red as clearly shown in the two photographs displayed in Figure 2.1(B) and Figure 2.1(C). After a designated duration of time (20, 60, or 120 min) of ischemia, the bulldog clamp was removed to follow reperfusion. Throughout the experiments, the abdominal cavity of

the animal was covered with moisturized gauze rinsed with saline to prevent drying, and the animal was placed in a side lying position with heating pads to maintain body temperature. For microscopy experiments, the rat was placed on a home-made sample stage of an inverted optical microscope with the left lobe of the liver positioned on a glass window as shown in the photograph displayed in Figure 2.2. At the end of the experiments, the rat was euthanized by injection of excessive pentothal.

2.4 Culturing of mouse hepatocytes

Mouse hepatocytes (FL83B) were grown in F12 medium supplemented with thermally inactivated fetal bovine serum (10 %, Invitrogen, USA) and penicillin/streptomycin mixture (100 U/mL, Invitrogen, USA) at 37 °C and 5 % CO2. Cells were grown to 80-90 % confluence and were split 1:4 every two days to maintain healthy. Cells were plated in glass-bottomed Petri dishes and cultured for another two days before further measurements.

2.5 Simulated ischemia and reperfusion of mouse hepatocytes

For simulated ischemia-reperfusion experiments on hepatocytes, the Petri dish that contained cells was mounted in a homemade, air-tight perfusion chamber as shown in the photograph displayed in Figure 2.3. Flow of the medium through the perfusion chamber was controlled by a peristaltic pump (MP-1000, EYELA, Japan).

To simulate an ischemic condition on cells, the chamber was filled and perfused with a hypoxia F12 medium with the concentration of O2 below 1 %. The hypoxia medium was prepared by continuously bubbling of the medium with N₂ for over 30 min to ensure the concentration of O2 in the medium reached below 1 %. During the process of bubbling, we used the dissolved oxygen meter (OM-51, Horiba, Japan) to measure the concentration of O2. Restoration from the ischemic condition was achieved by perfusing the chamber with a normoxia F12 medium. The control sham group was perfused with a normoxia F12 medium at different stages.

2.6 Inhibitory assays of mitochondrial electron transport chain and flavoproteins

To inhibit the mitochondrial electron transport chain, cells were incubated in the medium that contained antimycin A (10 μg/mL, in 0.5 % DMSO). To inhibit flavoproteins, cells were incubated in the medium that contained DPI (50 μM, in 0.5

% DMSO). The control groups were incubated in the medium mixed with 0.5 % DMSO. The incubation time was 60 min for treatments.

2.7 Isolation of mitochondria from rat liver

Mitochondria of rat liver were isolated with isolation buffer (250 mM Sucrose, 5 mM Tris, 1 mM EGTA, pH 7.4) by differential centrifugation from the livers of male Wistar rats weighing 200-250 g. All subsequent procedures were performed at 4 °C.

The liver was minced and homogenized by Dounce homogenizer (Kontes Glass Co., USA) with ice-cold isolation buffer. And then the homogenized liver was centrifuged with ice-cold isolation buffer at 1000 × g 10 min to remove the chunks and at 12000 × g 15 min twice to collect the mitochondria.

2.8 Intravital autofluorescence imaging and spectroscopy

The experimental setup employed in this study was modified from a confocal laser scanning optical microscope (FV300, Olympus, Japan) to allow autofluorescence and bright-field imaging of living animals and the hepatocytes, and measurements of emission spectra at selected regions of liver tissues, hepatocytes, mitochondria of liver, and FAD such as shown in Figure 2.4.

To image living animals and the hepatocytes, a 445-nm diode laser (PhoxX445, Omicron, Germany) was employed as excitation and the autofluorescence emission was detected between 458 and 630 nm by employing a 458 nm long-wave pass edge filter (LP02-458RS-25, Semrock, U.S.A.) and a 630 long-wave pass dichroic mirror (DM630, Olympus, Japan) in front of the photomultiplier tube (PMT) detector

(R3896, HAMAMATSU, Japan). For the experiments of ischemia-reperfusion of rat liver, the image was acquired at 512 × 512 pixels (707.1 × 707.1 μm) of scan area with a laser power about 10 μW which measured before entering the microscope and the scan time per image was about 2.71 s. The autofluorescence emission was collected by a 20X objective (N.A. 0.75, UPLSAPO, Olympus, Japan), passed through a 150 μm confocal pinhole and then detected by PMT. For the experiments of simulated ischemia-reperfusion of hepatocytes, the parameters was modified to scan area of 512 × 512 pixels (157.1 × 157.1 μm), time interval of images of 5.42 s with 2 frame Kalman filter, and the autofluorescence emission was collected by a 60X water-immersion objective (N.A. 1.20, UPLSAPO, Olympus, Japan) then passed through a 300 μm confocal pinhole. For the experiments of inhibitory assays, the parameters was modified to scan area of 512 × 512 pixels (235.7 × 235.7 μm), laser power of 50 μW, and the autofluorescence emission was collected by a 60X water-immersion objective (N.A. 1.20, UPLSAPO, Olympus, Japan) then passed through a 100 μm confocal pinhole. The transmitted light was collected by a long working distance universal condenser (IX2-LWUCD, N.A. 0.55, Olympus, Japan) and then directed to the PMT (R7400U-02, HAMAMATSU, Japan), The signal detected by PMT was transmitted to Fluoview program (Olympus, Japan) to form autofluorescence images and bright-field images.

To measure spectra of autofluorescence produced from FAD, hepatocytes, mitochondria of liver, or liver tissues under 445 nm laser excitation, a spectrometer (Shamrock SR-303i, Andor Technology, U.K.) that equipped with a grating (300 groves/mm), the silt width of 100 μm, and a thermoelectrically cooled CCD detector (iXon; Andor, U.K.) was employed. The laser power used in this experiment was 1 mW and the fluorescence emission was collected by a 60X water-immersion objective (N.A. 1.20, UPLSAPO, Olympus, Japan) then directed to the spectrometer with the

detection range between 483 nm to 764 nm and an exposure time of 0.0005 s, 5 s, 10 s, and 5 s for FAD, hepatocytes, mitochondria of liver, and liver tissues, respectively.

2.9 Image analysis

All autofluorescence images were analyzed with ImageJ program (Rasband, U.S.A.). The intensity of each image of animal was obtained by summing the intensity of each pixel after subtracting background and averaging over pixels with intensity exceeding the background. To remove the effect of photobleaching caused by laser, we fitted the decrease function of photobleaching of control group to find a correction function. Herein, we used the function: I = R + 0.0229t to correct the decrease of photobleaching of raw data and normalized by each control group such as shown in Figure 2.5, where I is the new intensity, R is the raw intensity, and t is the time from the start of laser exposure. The intensity of hepatocytes was obtained by removing the area of nucleus and subtracting background.

2.10 Statistical methods

Data were expressed as mean ± SEM. Comparison between the means of two groups was made using the two-tailed Student’s t test. The levels of statistical significance were set at P < 0.05, P < 0.01, and P < 0.001, respectively.

(D)

Portal vein Hepatic artery

Bulldog clamp

Portal vein Hepatic artery

Bulldog clamp

Figure 2.1: Changes in rat liver before and after ischemia.

(A) The normal rat liver with the median and left lateral lobe was reflected back to expose the hepatic artery and portal vein after laparotomy. (B, C) The change of rat liver before (B) and after (C) hepatic artery and portal vein were clamped by bulldog clamp. (D) Schematic of the ventral view of the rat liver before and after ischemia induced by bulldog clamp. After the hepatic artery and portal vein was clamped by bulldog clamp, the median and left lateral lobe was ischemic and the color of the median and left lateral lobe was from blood red to grey red. LLL: left lateral lobe, ML:

median lobe, RLL: right lateral lobe, HA: hepatic artery, PV: portal vein, BC: bulldog clamp.

Figure 2.2: The actual situation of the rat during experiment process.

In the experiment, the rat was placed in a sidelying position with heating pads on the microscopy, the abdominal cavity was covered with moist gauze with saline, and the left lateral lobe of liver was placed on glass slide for measurement.

Figure 2.3: Installation of homemade perfusion chamber.

The Petri dish was mounted in the perfusion chamber and the perfused medium flowed trough the flexible tube which connected with the peristaltic pump. The direction of the arrow is the flow direction of perfusion medium.

XY galvanometer

Figure 2.4: Schematic of the experimental apparatus for autofluorescence and bright-field imaging, and measurements of emission spectra of living animals and the hepatocytes.

The autofluorescence which emitted from sample was acquired by PMT or spectrometer with suitable filters to obtain images or spectra.

0 20 40 60 80 100 120

Figure 2.5: Correction of the photobleaching induced decrease of hepatic autofluorescence intensity.

The raw data (left) was fitted for the decrease function of photobleaching (red) to get the correction function: I = R + 0.0229t, and then corrected by the correction function to get a corrected data (right).

Chapter 3

Results and Discussion

In this study, we want to develop a label free method to do real-time observation for the change and damage of organ during ischemia and reperfusion period. To achieve this goal, we tried to use the fluorescence which emitted from intrinsic organelle of hepatocytes to observe the change of rat liver during ischemia and reperfusion period. As mentioned earlier, we tried to use flavins as a target to probe ischemia-reperfusion, so we chose 445 nm laser to be a source of excitation in this work. Figure 3.1 shows clearly that the range of fluorescence spectrum of rat liver is between 450 nm and 700 nm under 445 nm excitation. According to the spectral range of rat live, we acquired fluorescent images between 458 nm and 630 nm with excitation at 445 nm such as shown in Figure 3.2.

3.1 Time-lapse autofluorescence images of rat liver during rat hepatic warm ischemia and reperfusion

Before performing ischemia and reperfusion, we used the 445 nm laser with 10 μW of power to observe the effect of laser on normal liver of rat during 120 min measurement. The morphology, autofluorescence images, and statistics of autofluorescence intensity of normal rat liver shown in Figure 3.3(A) and Figure 3.4(A) did not have any intense change during 120 min measurement. It may mean that the laser power we used make a minimal influence and harm to the rat liver during long measurement time. Figure 3.3(B), Figure 3.3(C), and Figure 3.3(D) show the immediate change of autofluorescence. In the period of 20 min, 60 min, and 120 min ischemia, the autofluorescence images had been dimmer at first few minutes and kept stable in the following time. In the period of following reperfusion, all autofluorescence images showed a little uneven at the end of reperfusion and they

showed different behavior and phenomena between different ischemic times during following reperfusion. After 20 min ischemia, the autofluorescence had a rapid and overall restoration at first 10 min of reperfusion and kept steady in the following 80 min. On the contrary, the autofluorescence images of following reperfusion after 60 min ischemia showed a suddenly brightening on different area and then became dimmer quickly at first few minutes, but it still showed a restoration on autofluorescence images during reperfusion. The changes of autofluorescence images of following reperfusion after 120 min ischemia also showed similar behavior as after 60 min ischemia, but it showed a more serious no-reflow phenomenon on some area, it made that area look darker than the other area which had blood reflow. On the other hand, the intensity statistics of these autofluorescence images such as shown in Figure 3.4(B), Figure 3.4(C), and Figure 3.4(D), it exhibited that the autofluorescence intensity had rapid decrease about 40-50 % compared to control group at first 10 min and then kept stably in following time at all ischemic time. But during reperfusion, although not only short time ischemia (20 min) but also long time ischemia (60 min and 120 min) showed a gradual restoration of autofluorescence intensity, but the rate of restoration and the final intensity of autofluorescence was lower as the ischemic time was longer such as shown in Figure 3.4(E). The changes of autofluorescence is similar to some optical spectroscopy methods measured during ischemia and reperfusion period,^{39, 53, 61} the intensity of autofluorescence had suddenly increase or decrease in the periods of organic ischemia and reperfusion. But in those studies, it just provided overall or single point change on the observation area. On the contrary, according to the autofluorescence images that we acquired, we can provide not only overall change but also the regional change on the observation area. The difference between control, short ischemic time, and long ischemic time may based on the degree of lacking of nutrition, oxygen, and other necessary substances, alterations of

ion homeostasis and metabolisms, acidosis during ischemia, and the degree of blood reflow, reoxygenation, oxidative burst, and changes in physiological conditions after reperfusion, all of these adversely factors during ischemia and reperfusion induce different degree of cell damage to show a different response on autofluorescence images.

3.2 Time-lapse autofluorescence images of mouse hepatocytes subject to simulated ischemia-reperfusion

In the experiment of rat liver ischemia-reperfusion, it showed clearly that the autofluorescence change over time. To confer the origin of autofluorescence and to learn how the origin of autofluorescence varying with ischemia and reperfusion, we choose the normal mouse hepatocytes cell line FL83B for this study, not hepatocellular carcinoma cell line such as Hep G2 to prevent from the possible change of metabolic characteristics caused by carcinogenesis.⁶²

Under excitation by 445 nm laser with 10 μW of power and detection between 458 nm and 630 nm, the autofluorescence was appearing in the cytoplasm except for nucleus such as shown in Figure 3.5.

To understand whether it has same behavior as rat liver on cellular level during hypoxia and reoxygenation, we designed a perfusion chamber and a process to simulate the condition of ischemia and reperfusion. Before simulated ischemia and reperfusion was performed with hypoxia medium, the sham process was done with normoxia medium to simulate the process of ischemia and reperfusion to prevent from any acute changes to influence the observation during the perfusion process of medium in simulated ischemia and reperfusion period. The change of hepatocellular autofluorescence images during the period of 20 min sham ischemia and following 60 min reperfusion is shown in Figure 3.6(A). Such as shown as the hepatocellular autofluorescence images, the morphology and autofluorescence images did not have

radical and obvious change in a short time after perfusing with normoxia medium at different stages. Figure 3.7(A) shows the variation of hepatocellular autofluorescence intensity during the period of 20 min sham ischemia and following 60 min reperfusion, it also did not show obvious change after perfusing with normoxia medium in sham simulated ischemia and following reperfusion, it meant that the laser power we used did not influence and harm the cell during long measurement time.

To produce the hypoxia medium, we used nitrogen to bubble the F12 medium to remove the oxygen from medium and tested how long it can keep in low oxygen concentration when it was exposed to air after bubbled. As shown in Figure 3.8, the concentration of O2 dropped rapidly and reached below 1 % within 3 min after bubbling with N₂, and the concentration of O₂ remained below 1 % for over 1h even after exposing the medium to air owing to the relatively low desolation rate of O2 in medium. The rate of reoxygenation was slow enough for us to transfer the medium to work. Figure 3.6(B) shows the variation of hepatocellular autofluorescence images with the advance of time during the period of 20 min simulated ischemia and following 60 min reperfusion. After perfusing with hypoxia medium which was bubbled with nitrogen more than 30 min to simulate 20 min ischemia, the hepatocellular autofluorescence had been dimmer. But in the following period of reperfusion which perfused with normoxia medium, the autofluorescence had a rapid restoration. In the statistics of autofluorescence intensity such as shown in Figure 3.7(B), it showed a rapid decrease at the first 10 min after perfusing with hypoxia medium and kept about 70 % of the original intensity stably in the following 10 min.

In the following reperfusion, the intensity of autofluorescence increased rapidly at the first few minutes and then the increase rate of autofluorescence intensity has been slow down in the following observation. The final autofluorescence intensity at the end of observation was about 10 % increase compared with the initial intensity which

before ischemia. To compare the variation during simulated ischemia-reperfusion with the variation during sham simulated ischemia-reperfusion under the same experimental condition, they showed very different response to the stage of ischemia and reperfusion. The intensity of hepatocellular autofluorescence kept stable for the sham group which only perfused with normoxia medium at different stage, but for the simulated ischemia-reperfusion, it had a radical decrease in ischemic stage which perfused with hypoxia medium and had radical increase in following stage of reperfusion which perfused with normoxia medium. Despite the intensity of hepatocellular autofluorescence had a little perturbation in sham simulated ischemia and reperfusion, but the perturbation which may caused by fluctuation of focus was less than the variation during simulated ischemia and reperfusion, it did not affect our observation. In other words, the effects of perfusion or perturbation do not influence to observe the change caused by the different concentration of oxygen during simulated ischemia and reperfusion. The radical decrease or increase of

before ischemia. To compare the variation during simulated ischemia-reperfusion with the variation during sham simulated ischemia-reperfusion under the same experimental condition, they showed very different response to the stage of ischemia and reperfusion. The intensity of hepatocellular autofluorescence kept stable for the sham group which only perfused with normoxia medium at different stage, but for the simulated ischemia-reperfusion, it had a radical decrease in ischemic stage which perfused with hypoxia medium and had radical increase in following stage of reperfusion which perfused with normoxia medium. Despite the intensity of hepatocellular autofluorescence had a little perturbation in sham simulated ischemia and reperfusion, but the perturbation which may caused by fluctuation of focus was less than the variation during simulated ischemia and reperfusion, it did not affect our observation. In other words, the effects of perfusion or perturbation do not influence to observe the change caused by the different concentration of oxygen during simulated ischemia and reperfusion. The radical decrease or increase of