1. Introduction

End stage liver disease is a leading cause of death in the USA. Liver transplantation (LT) is the only cure, but liver organs qualified for LT are in shortage. One of the reasons for this shortage is ischemia and reperfusion injury (IRI) which has a significant negative impact on the organ functionality. IRI occurs both during hepatic surgery with clamping of the vascular pedicle of the porta hepatis (Pringle maneuver) and in LT [1]. Hepatic IRI has a profound clinical impact on graft function after LT with organs from marginal or extended criteria donors because its deleterious effects are augmented in these grafts. IRI causes early organ failure in up to 12% of patients, and 15% to 25% of patients experience long-term graft dysfunction [2]. Post-reperfusion syndrome, with an incidence rate of up to 30%, causes acute cardiovascular collapse that could lead to patient death [3]. Poor graft function after LT contributes to the need for a retransplantation of the liver and results in an increase in resource utilization.

Hepatic IRI begins with an interruption of the blood flow to the liver (ischemia) that leads to acidosis; a depletion of energy substrates and oxygen; impaired adenosine triphosphate (ATP) regeneration; and decreased levels of both the endogenous antioxidant glutathione (GSH) and the reduced form of nicotine adenine dinucleotide (NADH), a key enzyme in the electron transport chain. Paradoxically, the return of blood flow after a period of ischemia (reperfusion) results in further hepatocyte injury due to induced oxidative stress. This occurs through a complex cascade of events that includes mitochondrial dysfunction and systemic release of inflammatory mediators, reactive oxygen species (ROS), and proteases [1, 4]. These processes lead to failure in the ATP production mechanisms, causing cell death [5].

Determining the metabolic state of hepatocytes after IRI could potentially aid in the development of interventions that would mitigate the adverse effects of IRI and facilitate the recovery of liver function. At present, there is no quantitative test to evaluate the hepatocyte metabolic state after IRI. Optical fluorescence cryoimaging methods can provide a quantitative tool by capturing 3D images from fluorophores involved in cellular metabolism and use these signals to obtain information about bioenergetics [6]. The fluorophores can either come from exogenous probes used to stain the tissue or from the intrinsic autofluorescence of certain biomolecules. Previously, using optical imaging techniques IRI was quantified in different organs [7, 8], while hepatic IRI was investigated in animal models using near-infrared spectroscopy of indocyanine green [9] and molecular imaging of flavins [10].

The mitochondrial metabolic coenzymes NADH and flavin adenine dinucleotide (FAD) are autofluorescent and can be monitored using optical techniques without additional stains [11]. NADH and FAD are the two primary electron carriers in the electron transport chain that is responsible for producing ATP, the energy source of cells. NADH is only fluorescent in its reduced form while FAD is only fluorescent in its oxidized from. For this reason, the ratio of these two (NADH/FAD), called the redox ratio (RR), can be used as a quantitative marker of the oxidation state of the tissue mitochondria [12]. Our custom-designed cryoimager can provide the metabolic state of the tissue after snap freezing a sample. A frozen sample improves the fluorescence quantum yield, and cryoimaging can provide a 3D volumetric redox image to evaluate the tissue metabolic state. This unique optical imaging technique is advantageous since no exogenous probes are needed for fluorescence imaging. We used cryoimager to investigate the use of the metabolic redox ratio as a quantitative marker of hepatocyte viability after IRI.

2. Methods

2.1 Sample preparation

This study was performed using Sprague-Dawley rats (Charles River Laboratories International Inc.) that were 57 ± 3 days old with an average weight of 233 ± 30 g. All experiments complied with the standards presented in the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee (IACUC).

A total of five groups of rats containing five per group were studied. Animals were anesthetized by isoflurane inhalation. The first group consisted of the control (ctrl) which received no treatment, and a representative sample from this group of livers is shown in Fig. 1(a). The second group consisted of rats subjected to 60 minutes of ischemia (Isc60), which was achieved by temporarily occluding the blood vessel pedicles to the median and lateral lobes of the liver with a clamp. While the clamping was applied, the abdomen was temporarily closed during the ischemia to avoid dehydration. A liver subjected to this ischemia can be seen in Fig. 1(b), where the change in color is evident. The third group received a treatment similar to the second group, but with 90 minutes of ischemia injury (Isc90). The fourth and fifth groups had the same ischemia treatments as Isc60 and Isc90, but instead of harvesting the liver after ischemia, the clamp was removed and the abdomen was closed to allow reperfusion for 24 hours for each ischemia injury group (IRI60 and IRI90). An example of a liver in the reperfusion phase can be seen in Fig. 1(c). Two different ischemia durations were used to investigate the effect of longer ischemia time followed by reperfusion injury.

 

Fig. 1 Sample images of livers for each treatment: a) control (Ctrl), b) ischemia for 60 or 90 minutes (Isc) by blocking the blood flow with a clamp (as shown with the red circle) c) 24h reperfusion after ischemia (IRI).

Download Full Size | PDF

At the end of each protocol, the rats were sacrificed and their livers were harvested. The liver was flushed with normal saline (0.9% NaCl) to remove blood. The lateral lobe was snap frozen in a serial manner using isopentane followed by liquid nitrogen to avoid crack formation for high quality cryoimaging. Additionally, tissue samples from the median lobe of the liver were collected for ATP measurements, and blood samples were taken from the inferior vena cava when the animal was sacrificed.

2.2 Blood assays and ATP measurements

To assess the degree of hepatocyte injury, serum aspartate transaminase (AST) and alanine transaminase (ALT) levels, which provide an overview of hepatocyte damage, were measured. The whole blood collected from the inferior vena cava was sent to a local veterinary laboratory (Marshfield Labs, WI, USA) in serum-separating tubes (BD Medical Technology, NJ, USA).

Tissue samples from the median lobe of the liver were snap frozen in liquid nitrogen and stored at −80°C for ATP analysis. The frozen tissue was homogenized in a generic 20 × lysis buffer containing 200 mM Tris (pH 7.5), 2 M NaCl, 0.2% Triton X-100, and 20 mM EDTA. ATP was quantified using the Molecular Probes ATP Determination Kit, adhering to the supplied protocol. The microplate reader was set to the following parameters: luminescence, endpoint, integration time = 1.0 seconds, emission = 528/20, optics position = top, sensitivity = 135, and top probe vertical offset = 1 mm.

2.3 Cryoimaging

The cryo-preserved lateral lobes of the livers were imaged in our custom-made cryoimager in Biophotonics Lab at the University of Wisconsin-Milwaukee. To keep the tissue in place for imaging, the frozen liver lobes were first embedded in a customized black mounting medium for slicing and imaging. The imaging system consists of a freezer, kept at cryogenic temperature, containing an automated acquisition system that sequentially slices the tissue and acquires images of up to five fluorophores. The NADH and FAD metabolic co-enzymes can each be excited by their respective wavelength and will emit light at different wavelengths, which can then be detected by a camera.

A mercury arc lamp (200 W lamp, Oriel, Irvine, CA, with light source from Ushio Inc., Japan) was used as light source, and the light was selectively filtered by wavelength to excite each specific fluorophore from the surface of the tissue. The excitation filter for NADH was set to 350 nm (80 nm bandwidth, UV Pass Blacklite, HD Dichroic, Los Angeles, CA), and the FAD filter was set to 437 nm (20 nm bandwidth, 440QV21, Omega Optical, Brattleboro, VT). The emitted fluorophores were again filtered at wavelengths of 460 nm (50 nm bandwidth, D460/50M, Chroma, Bellows Falls, VT) and 537 nm (50 nm bandwidth, QMAX EM 510-560, Omega Optical) for NADH and FAD, respectively. All filters were controlled using two motorized filter wheels (Oriental Motor Vexta Step Motor PK268-01B). The images were captured using a charge-coupled device camera (QImaging, Rolera EM-C2, 14 bit) with a 1,004 × 1,002 pixel array. The z-sampling was 30 µm, defined by slicing thickness. Freezing, embedding and imaging procedures have been previously described [12, 13].

The captured NADH and FAD autofluorescence images from the liver slices were analyzed using Matlab. Calibration was performed to minimize day-to-day variations in light intensity and non-uniformity of the illumination pattern. The samples were detected from the background with thresholding and setting the background pixels to zero. From the captured images, a 3D reconstruction of the redox ratio of the whole sample can be created by calculating voxel-by-voxel according to Eq. (1):

The RR marker were used both to make a quantitative comparison between different treatment groups and for statistical analyses. The data were compared by one-way ANOVA. When a significant difference (p < 0.05) was found, Tukey’s post hoc analysis was performed to determine where the differences were located. Linear correlation among the variables was evaluated using Spearman’s rho.

3. Results

Figure 2(a) shows redox ratio cryoimaging results for all the samples in each group. The results are shown as the maximum projection of the redox ratio represented by a pseudo-color scale, with red indicating high signal level and blue indicating low signal level. It can be seen, that both ischemia groups have higher redox ratio values, which indicate a more reduced metabolic state, whereas both reperfusion groups show lower redox ratio values and therefore a more oxidized state. Furthermore, there is little variation within each group (SEM 0.15–0.54). To avoid system influences and other variations, only the mean of redox ratio (RR) was considered in the following analysis. Some of the samples during snap freezing got cracks with a visible zero-value line inside the liver images (see Fig. 2(a) row 4). The corresponding voxels were excluded in the RR calculation.

 

Fig. 2 a) Maximum projection of the redox ratio for n = 5 samples in each group. Each image scale is 4 cm × 4 cm. b) Relative voxel frequency distribution of redox ratio for a representative sample in each group.

Download Full Size | PDF

From the full 3D reconstruction of the redox ratio images, the relative voxel frequency distribution of redox ratio for each sample was constructed. These can be seen in Fig. 2(b) for a representative sample from each group. The ischemia groups have slightly higher redox ratio levels than the control, and the reperfusion groups show two levels of lower RR. A 3D rendering of the representative samples in each group can be seen in the Appendix.

A bar plot of the RR values for each group of samples can be seen in Fig. 3. This revealed no significant difference between the control and both ischemia groups. However, there were significant differences among the control and reperfusion groups (p = 0.0168 for IRI60 and p = 2.79·10⁻⁷ for IRI90). It was found that the IRI60 group had a 29% decrease in RR, whereas the IRI90 group had a 71% decrease in RR.

 

Fig. 3 Statistical analysis of mean values of RR with n = 5 samples in each group and significant difference marked with *.

Download Full Size | PDF

The results from the blood analyses are in Table 1. It can be seen, that IRI leads to higher AST and ALT values. After a one-way ANOVA test followed by a Tukey’s post hoc analysis, it can be seen, that although IRI60 had an increase in both ALT and AST levels, only the IRI90 group demonstrated a significant difference compared to the control. On the other hand, the same analysis performed on the ATP measurements shows that ATP levels dropped significantly in all ischemia and reperfusion groups compared to controls.

Table 1. Overview of AST, ALT, and ATP measurements for each sample group. Stars show a significant difference (p < 0.05) from the control group. Unit of AST and ALT is enzyme unit/L (U/L) and for ATP it is Luminescence/g.

View Table

The correlations between injury markers AST, ALT, and ATP and the measured mean RR values were calculated and can be seen in Fig. 4. There was a significant negative linear correlation between the values calculated for mean RR and both liver enzymes, which implies that RR increases as liver damage progresses (r = −0.896 and p < 0.001 in AST, and r = −0.839 and p < 0.001 in ALT). The strong positive linear correlation between mean RR and tissue ATP supports this observation (r = 0.714 and p = 0.004).

 

Fig. 4 Correlation of mean RR with a) AST, b) ALT, and c) ATP measurements from fifteen animals in control, IRI60, and IRI90 groups. RR had a significant negative correlation with serum liver enzyme levels (AST and ALT), and a significant positive correlation with tissue ATP levels. These results imply that RR decreases as a hepatocyte injury progresses.

Download Full Size | PDF

4. Discussion

This study investigated the potential use of optical imaging for the quantitative measurement of IRI in liver transplantation. Using optical 3D cryoimaging of tissue autofluorescence, the mitochondrial redox state (NADH/FAD) was investigated. The results revealed that: (i) the measured RR value, corresponding to the NADH/FAD ratio, did not change during ischemia alone, while it decreased in reperfusion phase, (ii) RR marker correlated with the degree of hepatic IRI as the ischemia period increased.

3D optical cryoimaging was used to quantify the mitochondrial redox state of liver during ischemia without reperfusion (Isc) and during reperfusion following ischemia (IRI). The major source of the redox state is mitochondria, with low contribution from cytoplasmic sources and NADPH [14]. FAD autofluorescence signal originates only from mitochondria [15]. Our study showed that ischemia only groups had statistically no difference from the controls regardless of the duration of the ischemia (60 or 90 minutes). The liver is an organ that can tolerate ischemia to a much higher degree than other organs, such as the heart. Typically, damage from ischemia is evident in the liver after only 30 minutes [1], though, most of this damage occurs in the reperfusion phase. This is evident in the ischemia-only groups in this experiment, where statistically there is no difference from the controls.

The main mechanism of injury in hepatic IRI is due to an immune-mediated response during the reperfusion phase [4]. This also matches well with our results because we see a significant difference in the severity of damage in the reperfused livers according to the duration of ischemia before reperfusion. The reperfusion groups both show lower redox ratios than the control group, indicating a more oxidized redox state. The change in coenzyme levels during hepatic IRI has been known to be incongruent and it is difficult to interpret. The exact details of NADH and FAD metabolism during hepatic IRI remain elusive, though a report on human liver perfusion models indicated that the NADH:NAD ratio was unchanged after reperfusion, but the estimated FAD level was increased [16]. This means that RR, which is NADH/FAD, decreased and it agrees with our findings that the RR decreased as the severity of the IRI increased.

Mitochondria are the lynchpin of hepatitic tissue energy and redox homeostasis, and hence, deterioration in mitochondrial function leads to hepatocyte and endothelial cell death and subsequent liver dysfunction. The abrupt cessation of oxidative phosphorylation (OxPhos) during ischemia rapidly leads to cellular ATP depletion. IRI causes oxidative/nitrosative stress [17] that leads to damage of mitochondrial proteins (tricarboxylic acid cycle enzymes) and impaired electron transfer (defective electron transport chain complexes) [18]. IRI could also lead to damaging mitochondria that contribute to reducing equivalents (NADH and FADH2) for OxPhos. The resultant injuries are reflected in the increasing serum level of hepatic aminotransferase enzymes. We observed a strong correlation between those quantitative markers of injury and the RR, and we believe it is evidence that the RR can offer a quantitative measurement of hepatic IRI.

IRI is a critical hepatic complication that occurs during several clinical settings for treatment of severe liver diseases. Quantitative assessment of liver viability after IRI allows surgeons to decide organ transplant suitability. While liver transplantation is the main life-saving treatment for severe liver diseases, organ shortage is the principal limiting factor. The mechanism of hepatic IRI has been explored in numerous animal models and mitochondria have been implicated as critical in the etiology of the disease. Our quantitative method for assessing the viability of the hepatic tissue might help determine if a liver can potentially be used for transplant.

Tissue samples obtained by biopsy can be used for various molecular analysis. However, it offers information only from a localized area. The underlying condition of the liver and microcirculation may vary inside the liver [19]. In contrast, optical imaging can monitor the whole organ for a research and clinical purpose. Online real-time fluorometer is also under investigation to obtain a real-time information of RR in the tissue [13]. It may offer a stronger utility tool to monitor the liver function at the time of liver transplantation.

In conclusion, the mitochondrial redox state can be successfully measured using optical cryoimaging as a quantitative marker of hepatic IR injury. This optical method provides information about this redox state through 3D measurement of the RR and subsequent volumetric quantification of the tissue damage. It has also provided information useful for understanding the metabolic mechanisms behind liver IR injury. In the future, this could be applied clinically through real-time optical measurement methods that can provide online RR values.

Disclosures

The authors declare that there are no conflicts of interest related to this article.

Appendix

In Fig. 2 the RR maximum projection of all samples was shown. A full 3D rendering of the representative samples in each group can be seen in Fig. 5 below.

 

Fig. 5 3D rendering of redox ratio for a representative sample in each group.

Download Full Size | PDF

Funding

This work is supported by UWM RGI 101x290 to MR.

References and links

1. C. García-Ruíz, A. Morales, and J. C. Fernández-Checa, “Oxidative Stress and Liver Ischemia–Reperfusion Injury,” in Studies on Hepatic Disorders, 149–170 (2015).

2. I. Hilmi, C. N. Horton, R. M. Planinsic, T. Sakai, R. Nicolau-Raducu, D. Damian, S. Gligor, and A. Marcos, “The impact of postreperfusion syndrome on short-term patient and liver allograft outcome in patients undergoing orthotopic liver transplantation,” Liver Transpl. 14(4), 504–508 (2008). [CrossRef]   [PubMed]  

3. S.-M. Jeong, “Postreperfusion syndrome during liver transplantation,” Korean J. Anesthesiol. 68(6), 527–539 (2015). [CrossRef]   [PubMed]  

4. K. M. Quesnelle, P. V. Bystrom, and L. H. Toledo-Pereyra, “Molecular responses to ischemia and reperfusion in the liver,” Arch. Toxicol. 89(5), 651–657 (2015). [CrossRef]   [PubMed]  

5. M. Abu-Amara, S. Y. Yang, N. Tapuria, B. Fuller, B. Davidson, and A. Seifalian, “Liver ischemia/reperfusion injury: processes in inflammatory networks-a review,” Liver Transpl. 16(9), 1016–1032 (2010). [CrossRef]   [PubMed]  

6. T. Vo-Dinh, Biomedical Photonics Handbook: Therapeutics and Advanced Biophotonics (CRC press, 2014).

7. A. P. Michalopoulou, J. T. Fitzgerald, C. Troppmann, and S. G. Demos, “Spectroscopic imaging for detection of ischemic injury in rat kidneys by use of changes in intrinsic optical properties,” Appl. Opt. 44(11), 2024–2032 (2005). [CrossRef]   [PubMed]  

8. C. H. Barlow, W. R. Harden 3rd, A. H. Harken, M. B. Simson, J. C. Haselgrove, B. Chance, M. O’Connor, and G. Austin, “Fluorescence mapping of mitochondrial redox changes in heart and brain,” Crit. Care Med. 7(9), 402–406 (1979). [CrossRef]   [PubMed]  

9. A. M. Seifalian, H. El-Desoky, D. T. Delpy, and B. R. Davidson, “Effects of hepatic ischaemia/reperfusion injury in a rabbit model of Indocyanine green clearance,” Clin. Sci. 102(5), 579–586 (2002). [CrossRef]   [PubMed]  

10. H.-H. Lu, Y.-M. Wu, W.-T. Chang, T. Luo, Y.-C. Yang, H.-D. Cho, and I. Liau, “Molecular imaging of ischemia and reperfusion in vivo with mitochondrial autofluorescence,” Anal. Chem. 86(10), 5024–5031 (2014). [CrossRef]   [PubMed]  

11. B. Chance, B. Schoener, R. Oshino, F. Itshak, and Y. Nakase, “Oxidation-reduction ratio studies of mitochondria in freeze-trapped samples. NADH and flavoprotein fluorescence signals,” J. Biol. Chem. 254(11), 4764–4771 (1979). [PubMed]  

12. R. Sepehr, K. Staniszewski, S. Maleki, E. R. Jacobs, S. Audi, and M. Ranji, “Optical imaging of tissue mitochondrial redox state in intact rat lungs in two models of pulmonary oxidative stress,” J. Biomed. Opt. 17(4), 046010 (2012). [CrossRef]   [PubMed]  

13. M. Ranji, M. M. Motlagh, F. Salehpour, R. Sepehr, J. S. Heisner, R. K. Dash, and A. K. S. Camara, “Optical Cryoimaging Reveals a Heterogeneous Distribution of Mitochondrial Redox State in ex vivo Guinea Pig Hearts and Its Alteration During Ischemia and Reperfusion,” IEEE J. Transl. Eng. Health Med. 4, 1800210 (2016). [PubMed]  

14. R. W. Estabrook, “Fluorometric measurement of reduced pyridine nucleotide in cellular and subcellular particles,” Anal. Biochem. 4(3), 231–245 (1962). [CrossRef]   [PubMed]  

15. E. M. Nuutinen, “Subcellular origin of the surface fluorescence of reduced nicotinamide nucleotides in the isolated perfused rat heart,” Basic Res. Cardiol. 79(1), 49–58 (1984). [CrossRef]   [PubMed]  

16. B. G. Bruinsma, G. V. Sridharan, P. D. Weeder, J. H. Avruch, N. Saeidi, S. Özer, S. Geerts, R. J. Porte, M. Heger, T. M. van Gulik, P. N. Martins, J. F. Markmann, H. Yeh, and K. Uygun, “Metabolic profiling during ex vivo machine perfusion of the human liver,” Sci. Rep. 6(1), 22415 (2016). [CrossRef]   [PubMed]  

17. A. K. Camara, M. Bienengraeber, and D. F. Stowe, “Mitochondrial approaches to protect against cardiac ischemia and reperfusion injury,” Front. Physiol. 2, 13 (2011). [CrossRef]   [PubMed]  

18. A. K. Gadicherla, D. F. Stowe, W. E. Antholine, M. Yang, and A. K. Camara, “Damage to mitochondrial complex I during cardiac ischemia reperfusion injury is reduced indirectly by anti-anginal drug ranolazine,” Biochimica et Biophysica Acta (BBA)-Bioenergetics 1817(3), 419–429 (2012). [CrossRef]  

19. O. W. Hamer, D. A. Aguirre, G. Casola, J. E. Lavine, M. Woenckhaus, and C. B. Sirlin, “Fatty liver: Imaging patterns and pitfalls 1,” Radiographics 26(6), 1637–1653 (2006). [CrossRef]   [PubMed]