- Email: [email protected]
High-Resolution Microscopic Determination of Hepatic NADH Fluorescence forin VivoMonitoring of Tissue Oxygenation during Hemorrhagic Shock and Resuscitation
Microvascular Research 54, 164 – 173 (1997) Article No. MR972028
TECHNICAL REPORT High-Resolution Microscopic Determination of Hepatic NADH Fluorescence for in Vivo Monitoring of Tissue Oxygenation during Hemorrhagic Shock and Resuscitation Brigitte Vollmar,* Markus Burkhardt,* Thomas Minor,† Hiltrud Klauke,† and Michael D. Menger* *Institute for Clinical & Experimental Surgery, University of Saarland, 66421 Homburg/Saar, Germany; and †Institute for Experimental Surgery, University of Cologne, 50931 Cologne, Germany Received February 3, 1997
Impaired microvascular oxygen supply reduces oxidative phosphorylation and causes an increase in cellular NADH, which was monitored densitometrically in vivo by high-resolution fluorescence microscopy (330 – 390/ ú430 nm excitation/emission wavelengths) in rat livers (n Å 8) subjected to hemorrhagic shock and resuscitation. At each time point, NADH fluorescence was recorded from 10 different observation fields of the left liver lobe. Withdrawal of a total of 4.5 ml arterial blood for induction of volume-controlled hemorrhagic shock resulted in an increase in NADH fluorescence by Ç31% from 45.1 { 3.9 to 59.2 { 4.2 aU, which was associated with a fall of arterial blood pressure from 110 { 3 to 51 { 8 mmHg, a decrease in hepatic tissue oxygenation (flexible polarographic surface electrode) from 18 { 2 to 2 { 1 mmHg, and a restriction of hepatic bile flow from 1.7 { 0.1 to 0.5 { 0.2 ml/min 1 g. Normovolemic resuscitation with 10% hydroxyethylstarch failed to completely restore the metabolic state of liver tissue (NADH fluorescence 49.9 { 3.1 aU), arterial blood pressure (83 { 8 mmHg), hepatic tissue oxygenation (7.4 { 1.5 mmHg), and hepatocellular excretory func-
tion (1.3 { 0.1 ml/min 1 g). During both shock and resuscitation, the ratio between pericentral and periportal NADH fluorescence intensities slightly increased, but calculation of coefficients of variance of interlobular NADH fluorescence did not reveal an increase in heterogeneity of tissue metabolic state. Significant correlations were found between NADH fluorescence and both hepatic tissue oxygenation (r2 Å 0.78, P õ 0.01) and hepatic bile flow (r2 Å 0.85, P õ 0.01), indicating that high-resolution intravital microscopic assessment of NADH fluorescence reflects appropriately the relation between local oxygen supply and demand in hepatic tissue in vivo. q 1997 Academic Press
INTRODUCTION In hemorrhagic shock, ischemic hepatocellular injury is a major determinant for the manifestation of liver dysfunction and/or hepatic organ failure. In addition to loss of circulating blood volume and reduction of perfusion pressure, microcirculatory events, i.e., eleva0026-2862/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
164
AID
MVR 2028
/
6z11$$$$$1
09-03-97 11:45:15
mvrg
AP: MVR
165
Technical Report
tion of hydraulic resistance due to hypoxia-induced cell swelling, rise of microvascular permeability due to endothelial disintegration with interstitial edema formation, and an increase in microvascular hematocrit and plasma viscosity, denominate the nature and magnitude of hepatic nutritive perfusion failure following hemorrhagic shock (Vollmar et al., 1994a). Pathological consequence of shock-induced impairment of hepatic microcirculation is the progressive deprivation of oxygen and, thus, energy supply to liver parenchyma, finally leading to hepatocellular death. Therefore, techniques providing information on the metabolic state of liver tissue are welcomed. To visualize the extent of tissue hypoxia, optical spectroscopy can be used. Chance and co-workers pioneered the fluorescence property of NADH as an indicator of the mitochondrial redox state and, in the presence of sufficient substrate and phosphate, as an indicator of cellular oxygen (Chance et al., 1962, 1976). NADH is a naturally occurring intracellular fluorophore and one of the main means to transfer energy from the tricarboxylic acid cycle to the respiratory chain in the mitochondria (Ince et al., 1992). Inhibition of the respiratory chain due to inadequate oxygen supply is reflected by increased intracellular NADH levels. Upon ultraviolet epiillumination of tissue, NADH, unlike NAD/, fluoresces in blue (Renault et al., 1982; Ince et al., 1992). In this way, NADH fluorescence provides direct information about the activity of mitochondrial respiration and enables imaging of its spatial distribution in organ surfaces in vivo. In almost all organs, NADH fluorimetry has been commonly used for investigations into pathophysiology of ischemia/hypoxia (Ince et al., 1992). Experiments of acute hypoxia to solid liver tissue (Chance and Schoener, 1965; Chance et al., 1965) and hepatic parenchymal cell suspensions (Obi-Tabot et al., 1993) demonstrated the association between the hypoxia-induced decrease in aerobic metabolism and the increase in NADH fluorescence. However, the major drawback of the techniques available is the low spatial resolution, which does not allow determination of the association of distinct NADH fluorescence intensities with individual microvascular perfusion conditions. Moreover, until now there is no validation as to whether changes of NADH fluorescence linearly correlate with changes in tissue oxygenation. Using a hemorrhagic shock/resus-
citation model in the rat, we herein report for the first time on the application of high-resolution intravital fluorescence microscopy for the assessment of hepatic NADH fluorescence with special emphasis on its interand intralobular distribution as well as its relation to tissue oxygenation and liver excretory function. Our study further addresses the interference of hepatic NADH fluorescence assessment with Ito cell-associated vitamin A autofluorescence.
METHODS
Animal Model Eight Sprague – Dawley rats were anesthetized with pentobarbital (50 mg/kg body wt ip) and placed in supine position on a heating pad. After tracheotomy the right carotid artery and jugular vein were cannulated with polyethylene catheters (PE-50; ID 0.58 mm; Portex, Hythe, UK) for monitoring of heart rate and blood pressure as well as for volume replacement. An additional catheter (PE-50) in the tail artery served for induction of hemorrhagic shock by arterial blood withdrawal. After transverse laparotomy and cannulation of the ductus choledochus (PE-50) for measurement of bile flow as an indicator of liver function, the rats were positioned on their left side and the livers were prepared for intravital fluorescence microscopy by placing the left lobe on a plasticine disk held by an adjustable stage that was attached to the heating pad (Vollmar et al., 1994a,b). Thereby, the lower surface of the liver was placed horizontal to the microscope, which guaranteed adequate focus level for the microscopic procedure on the area of liver surface under investigation. In addition, the adjustment of the plasticine disk allowed us to avoid the mechanical obstruction of feeding and draining macrovessels and to minimize respiratory movements of the lobe. The exposed area of the left liver lobe was immediately covered with an oxygenimpermeable transparent foil (LICOX Oxyblock foil; GMS, Kiel-Mielkendorf, Germany) to prevent evaporation and the influence of the ambient air.
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$1
09-03-97 11:45:15
mvrg
AP: MVR
166
Technical Report
FIG. 1. Schematic representation of the exteriorized left liver lobe, displaying area 1 (prebleached and allowing 20-min NADH recovery) and area 2 (nonprebleached) with the five observation fields studied at the different experimental time points (B, baseline; S1, S2, and S3, shock induction by withdrawal of 1.5 ml of arterial blood thrice; R1, R2, and R3, 15, 30, and 60 min after normovolemic resuscitation with 10% hydroxyethylstarch).
Intravital Fluorescence Microscopy For visualization of NADH fluorescence, a modified epiillumination Zeiss microscope (Zeiss Axiotech, Oberkochen, Germany) with a dichroic filter ((09), 330 – 390 nm/ú430 nm excitation/emission wavelengths) was used. Microfluorographic images were recorded by a CCD video camera (FK 6990; Prospective Measurements Inc., San Diego, CA) and transferred to a video system (VO-5800 PS; Sony GmbH, Munich, FRG). With the use of a 10X objective (10X/0.3, Zeiss) epiillumination and, thus, excitation was applied to a field of 1150 mm 1 850 mm (one frame). Magnification achieved on the video screen (PVM-1442 QM, diagonal: 330 mm, Sony) was 1235. The camera setting was identical in all experiments; i.e., black level automatics were excluded and no adjustments on gain, contrast, or brightness were performed. Half of the surface of the left liver lobe (area 1; Fig.
1) was exposed to epiillumination at 360 nm for a period of Ç20 sec per field. By this procedure, intrinsic vitamin A-derived autofluorescence located in fat deposits of Ito cells was completely eliminated because of its rapid photobleaching property (Suematsu et al., 1993a). This was followed by a 20-min period of normoperfusion for recovery of parenchymal autofluorescence up to the control level before epiillumination (Suematsu et al., 1993b). At each experimental time point, NADH fluorescence was recorded in five randomly selected lobuli from five different observation fields within the prebleached area by epiillumination for a controlled exposure time of 2 sec per single fluorograph (Fig. 1). To test whether prebleaching and parenchymal autofluorescence recovery is an essential prerequisite for appropriate in vivo determination of NADH fluorescence, we additionally epiilluminated five randomly selected lobuli per experimental time point within the nonprebleached area of the liver lobe (area 2; Fig. 1) for a controlled exposure time of 60 sec per fluorograph. For each time point, measurements were made from different lobuli/observation fields within the respective area to exclude the potential interference of bleaching of NADH upon repetitive light exposure of the tissue under investigation. By off-line analysis, NADH fluorescence in periportal and pericentral regions (classification according to Rappaport (1973)) was densitometrically assessed by computerassisted gray level determination (Cap-Image; Zeintl, Heidelberg, Germany; Zeintl et al., 1986) at the end of either the 2-sec exposure of the prebleached liver lobuli or after 30 sec and at the end of the 60-sec exposure time of the nonprebleached liver lobuli (Fig. 2). Microvessels on the liver surface, in particular postsinusoidal venules, were excluded from densitometric analysis to limit the interference of gray levels of nonparenchymal structures with the estimation of parenchymal NADH fluorescence. Inhomogeneity (heterogeneity) of NADH fluorescence between hepatic lobuli was analyzed by calculation of the coefficient of variance (relative dispersion) as standard deviation/mean of lobular NADH fluorescence.
Polarographic Tissue PO2 Measurements Hepatic tissue oxygenation was assessed by means of a flexible polyethylene microcatheter Clark type PO2
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$2
09-03-97 11:45:15
mvrg
AP: MVR
167
Technical Report
FIG. 2. Intravital microscopic images of identical liver lobules of area 1 (left) and area 2 (right). The liver lobule of area 1 (left) is shown before prebleaching (top), 1 min after prebleaching (middle) and after at least a 20-min NADH recovery period at the experimental time point ‘‘baseline’’ (bottom), i.e., when the NADH measurement was performed. The liver lobule of area 2 (right) is shown at the experimental time point ‘‘baseline’’ at the beginning of light exposure (top), as well as after 30 sec (middle) and 60 sec (bottom) of light exposure, i.e., when the NADH measurements were performed. Magnification, 471.
probe (diameter 470 mm, length 300 mm; LICOX System, GMS, Kiel-Mielkendorf, Germany), which was positioned beneath the oxygen-impermeable foil upon the liver surface. This allowed the LICOX probe to integrate local tissue PO2 values over the tissue area in contact with the 5-mm-long PO2-sensitive area near the catheter tip without interference with the ambient air (polarization voltage of 795 mV). On-line temperature compensation was performed by a temperature probe (type K thermocouple probe; LICOX System, GMS) which was also positioned between the hepatic surface and the oxygen-impermeable foil. Average tissue oxygenation
was continuously monitored during baseline, shock, and resuscitation, respectively.
Bile Flow Measurements For measurement of bile flow, bile was continuously collected via the cannula into a 1-ml syringe. With the assumption of a specific weight of bile of 1 g/ml (Sumimoto et al., 1988), the collected bile was weighed and standardized in microliters per minute and gram liver wet weight.
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$2
09-03-97 11:45:15
mvrg
AP: MVR
168
Technical Report
Experimental Protocol After the surgical procedure and epiillumination of the liver surface for elimination of vitamin A autofluorescence followed by a subsequent 20-min period for recovery of parenchymal autofluorescence, baseline recordings (B) were obtained, including macrohemodynamics, hepatic NADH fluorescence microscopy, hepatic tissue PO2 , bile flow, and arterial blood sampling. Hemorrhagic shock was induced by withdrawal of 1.5 ml arterial blood thrice and recordings were repeated (S1, S2, and S3) when arterial blood pressure was found to be stable at a lower level. In accordance to the protocol of previous studies (Vollmar et al., 1994a, 1996a), the animals received 10% hydroxyethylstarch 200/0.6 (identical volume of shed blood) intravenously over 5 min for resuscitation. Repeated analysis of hemodynamics, NADH fluorescence, hepatic tissue PO2 , and bile flow were then performed after 15, 30, and 60 min of resuscitation (R1, R2, and R3). The animals were killed by exsanguination for sampling of blood, and livers were removed for determination of wet weight.
Statistics To test for time effects, multivariate ANOVA for repeated measures was applied, followed by paired Student’s t test, including corrections of the a-error according to the Bonferroni probabilities for repeated measurements. To assess the correlation between NADH fluorescence and hepatic tissue oxygenation and liver function, linear regression analysis was performed. All data are given as mean { SEM and overall statistical significance was set at P õ 0.05.
RESULTS
Withdrawal of 1.5 ml arterial blood thrice (S1, S2, and S3) caused a progredient reduction of mean arterial blood pressure with a concomitant decrease of heart
rate (Table 1). During early resuscitation (R1), mean arterial blood pressure and heart rate increased, but did not completely return to baseline values. One hour after volume replacement, mean arterial blood pressure recovered to 74% of baseline values (Table 1). In addition to systemic hemodynamic changes, normovolemic resuscitation following shock caused a significant fall of systemic hematocrit and leukocyte count (Table 1). Arterial blood gas parameters except those of marked base deficit values, remained unaffected following shock and resuscitation (Table 1).
NADH Fluorescence, Hepatic Tissue Oxygenation, and Bile Flow During baseline conditions, analysis of NADH fluorescence of the hepatic parenchymal surface in area 1 and area 2 (30 sec epiillumination) revealed mean values of 47.5 { 4.8 aU and 45.1 { 3.9 aU, respectively (Table 2). Ratios of NADH fluorescence between pericentral and periportal regions of the liver lobuli in area 1 and area 2 were 0.84 { 0.02 and 0.81 { 0.03, respectively. Lowering of perfusion pressure by withdrawal of arterial blood caused a progredient (P õ 0.05) increase in NADH fluorescence to 134 – 139% (S3) of baseline in observation fields of area 1 and area 2 with a slight rise of ratios of pericentral to periportal NADH fluorescence to 0.90 { 0.05 and 0.90 { 0.03, respectively. Resuscitation failed to return NADH fluorescence to preshock conditions with values in the range of 116 – 118% at 60 min after volume replacement (Fig. 3 and Table 2). During resuscitation, pericentral to periportal ratios of NADH fluorescence ranged between 0.91 and 0.97. Calculation of coefficients of variance did not reveal an increase in heterogeneity of interlobular NADH fluorescence during shock and resuscitation (Table 2). Analysis of NADH fluorescence after 60 sec of epiillumination in observation fields of area 2 revealed absolute values of NADH fluorescence which were constantly lower compared to those values assessed at 30 sec of epiillumination (Table 2). Values of NADH fluorescence assessed in area 1 did significantly correlate with values of NADH fluorescence assessed in area 2 after both 30 sec (P õ 0.01; r2 Å 0.84; Fig. 4) and 60 sec of epiillumination (P õ 0.01; r2 Å 0.81). In parallel to shock- and resuscitation-induced
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$2
09-03-97 11:45:15
mvrg
AP: MVR
169
Technical Report
TABLE 1 Changes in Macrohemodynamics, Hematocrit, Systemic Leukocyte Count, and Arterial Blood Gases during Shock and Resuscitation
MAP (mmHg) HR (min01) Hct (%) LCsyst (103/ml) PaO2 (mmHg) O2-sat (%) BE (mmol/liter)
Baseline
S1
S2
S3
R1
R2
R3
110 { 3 374 { 11 52 { 1 12.2 { 1.2 88 { 6 96.7 { 0.6 02.7 { 0.5
83 { 6 357 { 14 n.d. n.d. n.d. n.d. n.d.
58 { 8* 328 { 19 n.d. n.d. n.d. n.d. n.d.
51.2 { 7.5* 334 { 36 47 { 1 11.1 { 1.3 105 { 6 97.9 { 0.3 06.0 { 1.5*
104 { 3 346 { 12 n.d. n.d. n.d. n.d. n.d.
107 { 4 353 { 12 n.d. n.d. n.d. n.d. n.d.
83 { 8 344 { 16 38 { 3* 6.9 { 1.8* 106 { 6 97.8 { 0.4 04.4 { 1.6*
Note. S1, S2, and S3, shock induction by withdrawal of 1.5 ml of arterial blood thrice; R1, R2, and R3, 15, 30, and 60 min after normovolemic resuscitation with 10% hydroxyethylstarch; MAP, mean arterial blood pressure; HR, heart rate; Hct, hematocrit; LCsyst , systemic leukocyte count; PaO2 , arterial oxygen partial pressure; O2-sat, arterial hemoglobin saturation; BE, base excess; n.d., not determined. Mean { SEM; *P õ 0.05 vs baseline.
changes of NADH fluorescence, hepatic tissue PO2 values and hepatocellular excretory function were found significantly reduced during the shock period without complete restoration to preshock baseline values after colloid volume replacement (Table 2). Regression analysis revealed significant correlations between mean values of both hepatic tissue PO2 and bile flow and mean values of NADH fluorescence (Figs. 5 and 6).
demonstrates applicability of the noninvasive and continuous monitoring of NADH fluorescence during hemorrhagic shock and resuscitation. Linear relationships between NADH fluorescence and both hepatic tissue oxygenation and hepatic bile flow indicate that intravital microscopic assessment of NADH fluorescence reflects appropriately the relation between local oxygen supply and demand in hepatic tissue and, thus, liver function in vivo.
DISCUSSION
Intravital Fluorescence Microscopy for the Determination of NADH Fluorescence
Using high-resolution intravital fluorescence microscopy and densitometric techniques, the present study
In contrast to the common use of microlight guides (Ji et al., 1979) in numerous studies on NADH fluores-
TABLE 2 NADH Fluorescence, Coefficient of Variance of Interlobular NADH Fluorescence, Hepatic Tissue PO2, and Bile Flow during Shock and Resuscitation Baseline
S1
S2
S3
R1
47.5 { 4.8 45.1 { 3.9 40.4 { 4.2
53.1 { 5.6 50.2 { 4.1 44.8 { 4.2
56.5 { 4.1* 54.3 { 3.8* 48.2 { 4.5*
59.7 { 7.1* 59.2 { 4.2* 50.6 { 3.1*
0.08 { 0.01 0.06 { 0.02 0.08 { 0.02
0.08 { 0.01 0.06 { 0.01 0.05 { 0.01
0.06 { 0.01 0.05 { 0.01 0.05 { 0.01
7.7 { 1.9
5.4 { 1.5*
2.1 { 0.6*
1.29 { 0.04
1.03 { 0.09*
0.51 { 0.15*
R2
R3
NADH fluorescence (aU) Area 1 Area 2 (30 sec) Area 2 (60 sec)
49.4 { 4.0 49.3 { 4.5 45.1 { 4.9
50.7 { 4.6 47.6 { 4.2 43.8 { 4.4
52.0 { 1.9 49.9 { 3.1 44.1 { 3.6
Coefficient of variance of interlobular NADH fluorescence Area 1 Area 2 (30 sec) Area 2 (60 sec)
0.06 { 0.01 0.08 { 0.02 0.07 { 0.01
Hepatic tissue PO2 (mmHg) Bile flow (ml/min 1 g)
17.9 { 1.8 1.67 { 0.11
0.08 { 0.02 0.04 { 0.01 0.05 { 0.01 11.9 { 1.7 1.03 { 0.01*
0.07 { 0.01 0.06 { 0.03 0.04 { 0.01
0.06 { 0.01 0.05 { 0.01 0.05 { 0.01
8.9 { 1.2
7.4 { 1.5*
1.16 { 0.21*
1.34 { 0.09
Note. S1, S2 and S3, shock induction by withdrawal of 1.5 ml of arterial blood thrice; R1, R2, and R3, 15, 30, and 60 min after normovolemic resuscitation with 10% hydroxyethylstarch. Area 1, hepatic tissue surface area which was epiilluminated for elimination of vitamin A autofluorescence and allowed to recover for 20 min prior to onset of the experiment. Area 2, the hepatic surface tissue area where NADH fluorescence was assessed after 30 and 60 sec of epiillumination without preceding elimination of vitamin A autofluorescence prior to onset of the experiment. Mean { SEM; *P õ 0.05 vs baseline.
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$3
09-03-97 11:45:15
mvrg
AP: MVR
170
Technical Report
FIG. 3. Changes of NADH fluorescence of the hepatic surface (in percentage of baseline) of rats subjected to hemorrhagic shock and resuscitation. B, baseline preshock conditions; S1, S2, and S3, shock induction by withdrawal of 1.5 ml arterial blood thrice; R1, R2, and R3, 15, 30, and 60 min after normovolemic resuscitation with 10% hydroxyethylstarch 200/0.6 iv. Circles, measurements obtained in observation fields of area 1 (epiilluminated and recovered hepatic tissue area prior to the experiment); squares and triangles, measurements obtained at either 30 or 60 sec, respectively, of epiillumination of observation fields in area 2 (non epiilluminated area prior to the experiment). Mean { SEM.
cence for assessment of hepatic ketogenesis and alcohol metabolism (Kashiwagi et al., 1983; Olson and Thurman, 1987), the present study first reports on the use of high-resolution intravital fluorescence microscopy for in vivo assessment of hepatic surface fluorescence of NADH during shock/resuscitation. In contrast to microlight guides with tips of 80 – 170 mm in diameter, whose accurate localization on periportal or pericentral areas is only guaranteed in trained hands (Kashiwagi et al., 1983), intravital fluorescence microscopy using a 10X objective allows for direct visualization and quantitative analysis of NADH fluorescence within a field of 1150 1 850 mm, including both periportal and pericentral segments of two to three liver lobuli. The high resolution of the microscopic technique (Ç1.2 mm) and the visualization of the individual hepatic lobular structures, which can be achieved by standard epiillumination techniques (McCuskey, 1986; Menger et al., 1993; Vollmar et al., 1994b), allow for the calculation of the
ratio of NADH fluorescence intensities between pericentral and periportal regions and the deduction of coefficients of variance of intra- and interlobular NADH fluorescence from data of the identical observation field without potential errors due to epiillumination of nonrepresentative or inadequate regions. Thus, both intraand interlobular heterogeneities of parenchymal tissue oxygenation can easily be detected using the microscopic procedure. For reduction of optical fiber motion artifacts, fluorescence recordings using microlight guides might require synchronization of the detection system with organ surface movements due to breathing or heart activity (Ince et al., 1992). The herein presented approach of intravital fluorescence microscopy of the left liver lobe of the rat, which is exteriorized and placed horizontal to the microscope on an adjustable stage, readily compensates for motion artifacts. In addition to movement, oximetric effects, i.e., variations in the oxygenation of hemoglobin, which might affect organ surface fluorescence due to changes of the hemoglobin spectrum (Ince et al., 1992), should necessarily be considered in the use
FIG. 4. Regression analysis between values of NADH fluorescence assessed in observation fields of area 1 and values assessed at 30 sec of epiillumination in observation fields of area 2. All values obtained during baseline conditions, shock (S1, S2, and S3), and resuscitation (R1, R2, and R3) were included.
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$3
09-03-97 11:45:15
mvrg
AP: MVR
171
Technical Report
FIG. 5. Regression analysis between values of NADH fluorescence assessed at 30 sec of epiillumination in observation fields of area 2 and values of hepatic tissue PO2 (flexible polyethylene microcatheter Clark type PO2 probe). Mean values { SEM obtained during baseline conditions, shock (S1, S2, and S3), and resuscitation (R1, R2, and R3) were included.
of NADH fluorescence for determination of the metabolic state of hepatic tissue. In the present study, however, this aspect can be ignored, since arterial hemoglobin saturation remained unchanged during shock and resuscitation. Finally, changes in blood volume, i.e., a varying number of blood corpuscles, may interfere with fluorescence measurements, since the spectral characteristics of blood pigments make absorption of both ultraviolet excitation light and blue fluorescence dependent on perfusion of the organ surface (Ince et al., 1992). In contrast to the application of microlight guides, direct visualization of the hepatic surface by means of intravital microscopy allows us to restrain quantitative densitometry of only parenchymal fluorescence by excluding the darkly visualized nonparenchymal structures, such as sinusoids and postsinusoidal venules, which would interfere with the accurate estimation of NADH fluorescence in the parenchymal regions. The lower density of parenchymal tissue in pericentral than periportal regions of the lobuli due to the merging of sinusoids into each other and their increase in diameter
might be the reason for slightly lower NADH fluorescence intensities in pericentral than periportal areas and thus ratios õ1, which is in accordance with reports of others (Olson and Thurman, 1987). Assessment of hepatic surface NADH fluorescence is complicated by questions concerning the relative contribution of Ito cell-associated vitamin A autofluorescence, which cannot be distinguished from each other using ultraviolet epiilluminated microfluorographs. However, vitamin A autofluorescence can easily be eliminated due to its rapid photobleaching property. In line with in vitro studies on liver perfusion reported by Suematsu et al. (1993b), we therefore assessed NADH fluorescence in liver surface regions, which were epiilluminated for elimination of vitamin A autofluorescence and were allowed to recover for 20 min prior to onset of the experiment (area 1). To assess whether this prebleaching and recovery is necessary for accurate estimation of NADH fluorescence, we additionally studied NADH fluorescence in nonprebleached surface regions (area 2) after a 30- and 60-sec epiillumination
FIG. 6. Regression analysis between values of NADH fluorescence assessed at 30 sec of epiillumination in observation fields of area 2 and values of hepatic bile flow serving as an indicator of hepatocellular excretory function. Mean values { SEM obtained during baseline conditions, shock (S1, S2, and S3), and resuscitation (R1, R2, and R3) were included.
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$3
09-03-97 11:45:15
mvrg
AP: MVR
172
Technical Report
period. Epiillumination of these hepatic surface regions for more than 20 sec guaranteed also complete elimination of vitamin A autofluorescence (Vollmar et al., 1996b) and further allowed us to calculate the subsequent loss of NADH fluorescence due to a potential photobleaching effect after 30 and 60 sec of continuous light exposure. Corresponding data of the present study reveal a constant loss of NADH fluorescence of approximately 8 – 14% due to its photobleaching property over the time period from 30 to 60 sec of epiillumination. However, the linear correlation of values of NADH fluorescence assessed in area 1 (epiilluminated and recovered area) and 2 (area not epiilluminated prior to onset of the experiment) reveals that the procedure for elimination of vitamin A autofluorescence, including ultraviolet epiillumination with a subsequent NADH recovery period of at least 20 min, is not a prerequisite for the reliable assessment of intraindividual changes of hepatic NADH fluorescence in vivo.
Shock- and Resuscitation-Induced Changes of Hepatic Surface NADH Fluorescence Shed blood volume-controlled lowering of arterial blood pressure followed by isovolemic resuscitation with hydroxyethylstarch caused the well-known pattern of host response, including the decrease in heart rate, systemic hematocrit and leukocyte count, as well as the deterioration of arterial acid – base balance (Vollmar et al., 1994a, 1996a). As shown previously (Vollmar et al., 1994a), hemorrhagic shock is associated with significant nutritive perfusion failure which persists despite resuscitation and accounts for insufficient energy supply to tissue, as indicated in the present study by the lowered tissue oxygenation during shock with inadequate restoration following resuscitation. Restriction of bile flow is considered a direct consequence of insufficient oxygen supply, since hepatocellular excretory function involves ATP-consuming steps and has been repeatedly shown to correlate with hepatocellular ATP levels (Kamiike et al., 1985; Bowers et al., 1987; Karwinski et al., 1989). In parallel to changes of hepatic tissue oxygenation and bile flow, hepatic surface NADH fluorescence increased, with most pronounced changes at the end of hemorrhagic shock and uncomplete restoration during normovolemic resuscitation. The linear cor-
relation between NADH fluorescence and data of hepatic tissue PO2 and bile flow indicates that NADH fluorescence reliably resembles the metabolic state of hepatic tissue in vivo. In conclusion, high-resolution intravital fluorescence microscopic assessment of hepatic surface NADH fluorescence represents an accurate measure for hepatocellular energy metabolism/tissue oxygenation and a promising tool due to the noninvasive and continuously applicable technical approach.
REFERENCES
Bowers, B. A., Branum, G. D., Rotolo, F. S., Watters, C. R., and Meyers, W. C. (1987). Bile flow: An index of ischemic injury. J. Surg. Res. 42, 565 – 569. Chance, B. (1976). Pyridine nucleotide as an indicator of the oxygen requirements for energy-linked functions of mitochondria. Circ. Res. 38(Suppl. 1), I31 – I38. Chance, B., Cohen, P., Jo¨bsis, F., and Schoener, B. (1962). Intracellular oxidation-reduction states in vivo. The microfluorometry of pyridine nucleotide gives a continuous measurement of the oxidation state. Science 137, 499 – 508. Chance, B., and Schoener, B. (1965). A correlation of absorption and fluorescence changes in ischemia of the rat liver, in vivo. Biochem. Z. 341, 340 – 345. Chance, B., Schoener, B., Krejci, K., Russmann, W., Wesemann, W., Schnitger, H., and Bu¨cher, Th. (1965). Kinetics of fluorescence and metabolite changes in rat liver during a cycle of ischemia. Biochem. Z. 341, 325 – 333. Ince, C., Coremans, J. M. C. C., and Bruining, H. A. (1992). In vivo NADH fluorescence. Adv. Exp. Med. Biol. 317, 277 – 296. Ji, S., Chance, B., Nishiki, K., Smith, T., and Rich, T. (1979). Microlight guides: A new method for measuring tissue fluorescence and reflectance. Am. J. Physiol. 236, C144 – C156. Kamiike, W., Nakahara, M., Nakao, K., Koseki, M., Nishida, T., Kawashima, Y., Watanabe, F., and Tagawa, K. (1985). Correlation between cellular ATP level and bile excretion in the rat liver. Transplantation 39, 50 – 55. Karwinski, W., Husoy, A.-M., Farstad, M., and Soreide, O. (1989). Sixty minutes of normothermic ischemia in the rat liver: Correlation between adenine nucleotides and bile excretion. J. Surg. Res. 46, 99 – 103. Kashiwagi, T., Lindros, K. O., Ji, S., and Thurman, R. G. (1983). Aldehyde dehydrogenase-dependent acetaldehyde metabolism in periportal and pericentral regions of the perfused rat liver. J. Pharmacol. Exp. Therap. 224, 538 – 542.
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$3
09-03-97 11:45:15
mvrg
AP: MVR
173
Technical Report
McCuskey, R. S. (1986). Microscopic methods for studying the microvasculature of internal organs. In Physical Techniques in Biology and Medicine Microvascular Technology (C. H. Baker and W. F. Nastuk, Eds.), pp. 247 – 264. Academic Press, New York. Menger, M. D., Vollmar, B., Glasz, J., Mu¨ller, M. J., Post, S., and Messmer, K. (1993). Hepatic ischemia/reperfusion. Prog. Appl. Microcirc. 19, 106 – 124. Obi-Tabot, E. T., Hanrahan, L. M., Cachecho, R., Beer, E. R., Hopkins, S. R., Chan, J. C. K., Shapiro, J. M., and LaMorte, W. W. (1993). Changes in hepatocyte NADH fluorescence during prolonged hypoxia. J. Surg. Res. 55, 575 – 580. Olson, M. J., and Thurman, R. G. (1987). Quantitation of ketogenesis in periportal and pericentral regions of the liver lobule. Arch. Biochem. Biophys. 253, 26 – 37. Rappaport, A. M. (1973). The microcirculatory hepatic unit. Microvasc. Res. 6, 212 – 228. Renault, G., Raynal, E., Sinet, M., Berthier, J. P., Godard, B., and Cornillault, J. (1982). A laser fluorimeter for direct cardiac metabolism investigation. Optics Laser Technol. 14, 143 – 148. Suematsu, M., Oda, M., Suzuki, H., Kaneko, H., Watanabe, N., Furusho, T., Masushige, S., and Tsuchiya, M. (1993a). Intravital and electron microscopic observation of Ito cells in rat hepatic microcirculation. Microvasc. Res. 46, 28 – 42. Suematsu, M., Suzuki, H., Ishii, H., and Tsuchiya, M. (1993b). Paradox
of oxygen-radical-dependent cell injury in the hypoxic liver microcirculation. Prog. Appl. Microcirc. 19, 127 – 138. Sumimoto, K., Inagaki, K., Yamada, K., Kawasaki, T., and Dohi, K. (1988). Reliable indices for the determination of viability of grafted liver immediately after orthotopic transplantation. Bile flow rate and cellular adenosine triphosphate level. Transplantation 46, 506 – 509. Vollmar, B., Lang, G., Menger, M. D., and Messmer, K. (1994a). Hypertonic hydroxyethylstarch restores hepatic microvascular perfusion in hemorrhagic shock. Am. J. Physiol. 266, H1927 – H1934. Vollmar, B., Glasz, J., Leiderer, R., Post, S., and Menger, M. D. (1994b). Hepatic microcirculatory perfusion failure is a determinant of liver dysfunction in warm ischemia-reperfusion. Am. J. Pathol. 145, 1421 – 1431. Vollmar, B., Preissler, G., and Menger, M. D. (1996a). Small-volume resuscitation restores hemorrhage-induced microcirculatory disorders in rat pancreas. Crit. Care Med. 24, 445 – 450. Vollmar, B., Ru¨cker, M., and Menger, M. D. (1996b). A new method for the intravital microscopic quantification of hepatic sinusoidal perfusion failure using the dye bisbenzamide H33342. Microvasc. Res. 51, 250 – 259. Zeintl, H., Tompkins, W. R., Messmer, K., and Intaglietta, M. (1986). Static and dynamic microcirculatory video image analysis applied to clinical investigations. Prog. Appl. Microcirc. 11, 1 – 10.
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$4
09-03-97 11:45:15
mvrg
AP: MVR
TECHNICAL REPORT High-Resolution Microscopic Determination of Hepatic NADH Fluorescence for in Vivo Monitoring of Tissue Oxygenation during Hemorrhagic Shock and Resuscitation Brigitte Vollmar,* Markus Burkhardt,* Thomas Minor,† Hiltrud Klauke,† and Michael D. Menger* *Institute for Clinical & Experimental Surgery, University of Saarland, 66421 Homburg/Saar, Germany; and †Institute for Experimental Surgery, University of Cologne, 50931 Cologne, Germany Received February 3, 1997
Impaired microvascular oxygen supply reduces oxidative phosphorylation and causes an increase in cellular NADH, which was monitored densitometrically in vivo by high-resolution fluorescence microscopy (330 – 390/ ú430 nm excitation/emission wavelengths) in rat livers (n Å 8) subjected to hemorrhagic shock and resuscitation. At each time point, NADH fluorescence was recorded from 10 different observation fields of the left liver lobe. Withdrawal of a total of 4.5 ml arterial blood for induction of volume-controlled hemorrhagic shock resulted in an increase in NADH fluorescence by Ç31% from 45.1 { 3.9 to 59.2 { 4.2 aU, which was associated with a fall of arterial blood pressure from 110 { 3 to 51 { 8 mmHg, a decrease in hepatic tissue oxygenation (flexible polarographic surface electrode) from 18 { 2 to 2 { 1 mmHg, and a restriction of hepatic bile flow from 1.7 { 0.1 to 0.5 { 0.2 ml/min 1 g. Normovolemic resuscitation with 10% hydroxyethylstarch failed to completely restore the metabolic state of liver tissue (NADH fluorescence 49.9 { 3.1 aU), arterial blood pressure (83 { 8 mmHg), hepatic tissue oxygenation (7.4 { 1.5 mmHg), and hepatocellular excretory func-
tion (1.3 { 0.1 ml/min 1 g). During both shock and resuscitation, the ratio between pericentral and periportal NADH fluorescence intensities slightly increased, but calculation of coefficients of variance of interlobular NADH fluorescence did not reveal an increase in heterogeneity of tissue metabolic state. Significant correlations were found between NADH fluorescence and both hepatic tissue oxygenation (r2 Å 0.78, P õ 0.01) and hepatic bile flow (r2 Å 0.85, P õ 0.01), indicating that high-resolution intravital microscopic assessment of NADH fluorescence reflects appropriately the relation between local oxygen supply and demand in hepatic tissue in vivo. q 1997 Academic Press
INTRODUCTION In hemorrhagic shock, ischemic hepatocellular injury is a major determinant for the manifestation of liver dysfunction and/or hepatic organ failure. In addition to loss of circulating blood volume and reduction of perfusion pressure, microcirculatory events, i.e., eleva0026-2862/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
164
AID
MVR 2028
/
6z11$$$$$1
09-03-97 11:45:15
mvrg
AP: MVR
165
Technical Report
tion of hydraulic resistance due to hypoxia-induced cell swelling, rise of microvascular permeability due to endothelial disintegration with interstitial edema formation, and an increase in microvascular hematocrit and plasma viscosity, denominate the nature and magnitude of hepatic nutritive perfusion failure following hemorrhagic shock (Vollmar et al., 1994a). Pathological consequence of shock-induced impairment of hepatic microcirculation is the progressive deprivation of oxygen and, thus, energy supply to liver parenchyma, finally leading to hepatocellular death. Therefore, techniques providing information on the metabolic state of liver tissue are welcomed. To visualize the extent of tissue hypoxia, optical spectroscopy can be used. Chance and co-workers pioneered the fluorescence property of NADH as an indicator of the mitochondrial redox state and, in the presence of sufficient substrate and phosphate, as an indicator of cellular oxygen (Chance et al., 1962, 1976). NADH is a naturally occurring intracellular fluorophore and one of the main means to transfer energy from the tricarboxylic acid cycle to the respiratory chain in the mitochondria (Ince et al., 1992). Inhibition of the respiratory chain due to inadequate oxygen supply is reflected by increased intracellular NADH levels. Upon ultraviolet epiillumination of tissue, NADH, unlike NAD/, fluoresces in blue (Renault et al., 1982; Ince et al., 1992). In this way, NADH fluorescence provides direct information about the activity of mitochondrial respiration and enables imaging of its spatial distribution in organ surfaces in vivo. In almost all organs, NADH fluorimetry has been commonly used for investigations into pathophysiology of ischemia/hypoxia (Ince et al., 1992). Experiments of acute hypoxia to solid liver tissue (Chance and Schoener, 1965; Chance et al., 1965) and hepatic parenchymal cell suspensions (Obi-Tabot et al., 1993) demonstrated the association between the hypoxia-induced decrease in aerobic metabolism and the increase in NADH fluorescence. However, the major drawback of the techniques available is the low spatial resolution, which does not allow determination of the association of distinct NADH fluorescence intensities with individual microvascular perfusion conditions. Moreover, until now there is no validation as to whether changes of NADH fluorescence linearly correlate with changes in tissue oxygenation. Using a hemorrhagic shock/resus-
citation model in the rat, we herein report for the first time on the application of high-resolution intravital fluorescence microscopy for the assessment of hepatic NADH fluorescence with special emphasis on its interand intralobular distribution as well as its relation to tissue oxygenation and liver excretory function. Our study further addresses the interference of hepatic NADH fluorescence assessment with Ito cell-associated vitamin A autofluorescence.
METHODS
Animal Model Eight Sprague – Dawley rats were anesthetized with pentobarbital (50 mg/kg body wt ip) and placed in supine position on a heating pad. After tracheotomy the right carotid artery and jugular vein were cannulated with polyethylene catheters (PE-50; ID 0.58 mm; Portex, Hythe, UK) for monitoring of heart rate and blood pressure as well as for volume replacement. An additional catheter (PE-50) in the tail artery served for induction of hemorrhagic shock by arterial blood withdrawal. After transverse laparotomy and cannulation of the ductus choledochus (PE-50) for measurement of bile flow as an indicator of liver function, the rats were positioned on their left side and the livers were prepared for intravital fluorescence microscopy by placing the left lobe on a plasticine disk held by an adjustable stage that was attached to the heating pad (Vollmar et al., 1994a,b). Thereby, the lower surface of the liver was placed horizontal to the microscope, which guaranteed adequate focus level for the microscopic procedure on the area of liver surface under investigation. In addition, the adjustment of the plasticine disk allowed us to avoid the mechanical obstruction of feeding and draining macrovessels and to minimize respiratory movements of the lobe. The exposed area of the left liver lobe was immediately covered with an oxygenimpermeable transparent foil (LICOX Oxyblock foil; GMS, Kiel-Mielkendorf, Germany) to prevent evaporation and the influence of the ambient air.
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$1
09-03-97 11:45:15
mvrg
AP: MVR
166
Technical Report
FIG. 1. Schematic representation of the exteriorized left liver lobe, displaying area 1 (prebleached and allowing 20-min NADH recovery) and area 2 (nonprebleached) with the five observation fields studied at the different experimental time points (B, baseline; S1, S2, and S3, shock induction by withdrawal of 1.5 ml of arterial blood thrice; R1, R2, and R3, 15, 30, and 60 min after normovolemic resuscitation with 10% hydroxyethylstarch).
Intravital Fluorescence Microscopy For visualization of NADH fluorescence, a modified epiillumination Zeiss microscope (Zeiss Axiotech, Oberkochen, Germany) with a dichroic filter ((09), 330 – 390 nm/ú430 nm excitation/emission wavelengths) was used. Microfluorographic images were recorded by a CCD video camera (FK 6990; Prospective Measurements Inc., San Diego, CA) and transferred to a video system (VO-5800 PS; Sony GmbH, Munich, FRG). With the use of a 10X objective (10X/0.3, Zeiss) epiillumination and, thus, excitation was applied to a field of 1150 mm 1 850 mm (one frame). Magnification achieved on the video screen (PVM-1442 QM, diagonal: 330 mm, Sony) was 1235. The camera setting was identical in all experiments; i.e., black level automatics were excluded and no adjustments on gain, contrast, or brightness were performed. Half of the surface of the left liver lobe (area 1; Fig.
1) was exposed to epiillumination at 360 nm for a period of Ç20 sec per field. By this procedure, intrinsic vitamin A-derived autofluorescence located in fat deposits of Ito cells was completely eliminated because of its rapid photobleaching property (Suematsu et al., 1993a). This was followed by a 20-min period of normoperfusion for recovery of parenchymal autofluorescence up to the control level before epiillumination (Suematsu et al., 1993b). At each experimental time point, NADH fluorescence was recorded in five randomly selected lobuli from five different observation fields within the prebleached area by epiillumination for a controlled exposure time of 2 sec per single fluorograph (Fig. 1). To test whether prebleaching and parenchymal autofluorescence recovery is an essential prerequisite for appropriate in vivo determination of NADH fluorescence, we additionally epiilluminated five randomly selected lobuli per experimental time point within the nonprebleached area of the liver lobe (area 2; Fig. 1) for a controlled exposure time of 60 sec per fluorograph. For each time point, measurements were made from different lobuli/observation fields within the respective area to exclude the potential interference of bleaching of NADH upon repetitive light exposure of the tissue under investigation. By off-line analysis, NADH fluorescence in periportal and pericentral regions (classification according to Rappaport (1973)) was densitometrically assessed by computerassisted gray level determination (Cap-Image; Zeintl, Heidelberg, Germany; Zeintl et al., 1986) at the end of either the 2-sec exposure of the prebleached liver lobuli or after 30 sec and at the end of the 60-sec exposure time of the nonprebleached liver lobuli (Fig. 2). Microvessels on the liver surface, in particular postsinusoidal venules, were excluded from densitometric analysis to limit the interference of gray levels of nonparenchymal structures with the estimation of parenchymal NADH fluorescence. Inhomogeneity (heterogeneity) of NADH fluorescence between hepatic lobuli was analyzed by calculation of the coefficient of variance (relative dispersion) as standard deviation/mean of lobular NADH fluorescence.
Polarographic Tissue PO2 Measurements Hepatic tissue oxygenation was assessed by means of a flexible polyethylene microcatheter Clark type PO2
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$2
09-03-97 11:45:15
mvrg
AP: MVR
167
Technical Report
FIG. 2. Intravital microscopic images of identical liver lobules of area 1 (left) and area 2 (right). The liver lobule of area 1 (left) is shown before prebleaching (top), 1 min after prebleaching (middle) and after at least a 20-min NADH recovery period at the experimental time point ‘‘baseline’’ (bottom), i.e., when the NADH measurement was performed. The liver lobule of area 2 (right) is shown at the experimental time point ‘‘baseline’’ at the beginning of light exposure (top), as well as after 30 sec (middle) and 60 sec (bottom) of light exposure, i.e., when the NADH measurements were performed. Magnification, 471.
probe (diameter 470 mm, length 300 mm; LICOX System, GMS, Kiel-Mielkendorf, Germany), which was positioned beneath the oxygen-impermeable foil upon the liver surface. This allowed the LICOX probe to integrate local tissue PO2 values over the tissue area in contact with the 5-mm-long PO2-sensitive area near the catheter tip without interference with the ambient air (polarization voltage of 795 mV). On-line temperature compensation was performed by a temperature probe (type K thermocouple probe; LICOX System, GMS) which was also positioned between the hepatic surface and the oxygen-impermeable foil. Average tissue oxygenation
was continuously monitored during baseline, shock, and resuscitation, respectively.
Bile Flow Measurements For measurement of bile flow, bile was continuously collected via the cannula into a 1-ml syringe. With the assumption of a specific weight of bile of 1 g/ml (Sumimoto et al., 1988), the collected bile was weighed and standardized in microliters per minute and gram liver wet weight.
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$2
09-03-97 11:45:15
mvrg
AP: MVR
168
Technical Report
Experimental Protocol After the surgical procedure and epiillumination of the liver surface for elimination of vitamin A autofluorescence followed by a subsequent 20-min period for recovery of parenchymal autofluorescence, baseline recordings (B) were obtained, including macrohemodynamics, hepatic NADH fluorescence microscopy, hepatic tissue PO2 , bile flow, and arterial blood sampling. Hemorrhagic shock was induced by withdrawal of 1.5 ml arterial blood thrice and recordings were repeated (S1, S2, and S3) when arterial blood pressure was found to be stable at a lower level. In accordance to the protocol of previous studies (Vollmar et al., 1994a, 1996a), the animals received 10% hydroxyethylstarch 200/0.6 (identical volume of shed blood) intravenously over 5 min for resuscitation. Repeated analysis of hemodynamics, NADH fluorescence, hepatic tissue PO2 , and bile flow were then performed after 15, 30, and 60 min of resuscitation (R1, R2, and R3). The animals were killed by exsanguination for sampling of blood, and livers were removed for determination of wet weight.
Statistics To test for time effects, multivariate ANOVA for repeated measures was applied, followed by paired Student’s t test, including corrections of the a-error according to the Bonferroni probabilities for repeated measurements. To assess the correlation between NADH fluorescence and hepatic tissue oxygenation and liver function, linear regression analysis was performed. All data are given as mean { SEM and overall statistical significance was set at P õ 0.05.
RESULTS
Withdrawal of 1.5 ml arterial blood thrice (S1, S2, and S3) caused a progredient reduction of mean arterial blood pressure with a concomitant decrease of heart
rate (Table 1). During early resuscitation (R1), mean arterial blood pressure and heart rate increased, but did not completely return to baseline values. One hour after volume replacement, mean arterial blood pressure recovered to 74% of baseline values (Table 1). In addition to systemic hemodynamic changes, normovolemic resuscitation following shock caused a significant fall of systemic hematocrit and leukocyte count (Table 1). Arterial blood gas parameters except those of marked base deficit values, remained unaffected following shock and resuscitation (Table 1).
NADH Fluorescence, Hepatic Tissue Oxygenation, and Bile Flow During baseline conditions, analysis of NADH fluorescence of the hepatic parenchymal surface in area 1 and area 2 (30 sec epiillumination) revealed mean values of 47.5 { 4.8 aU and 45.1 { 3.9 aU, respectively (Table 2). Ratios of NADH fluorescence between pericentral and periportal regions of the liver lobuli in area 1 and area 2 were 0.84 { 0.02 and 0.81 { 0.03, respectively. Lowering of perfusion pressure by withdrawal of arterial blood caused a progredient (P õ 0.05) increase in NADH fluorescence to 134 – 139% (S3) of baseline in observation fields of area 1 and area 2 with a slight rise of ratios of pericentral to periportal NADH fluorescence to 0.90 { 0.05 and 0.90 { 0.03, respectively. Resuscitation failed to return NADH fluorescence to preshock conditions with values in the range of 116 – 118% at 60 min after volume replacement (Fig. 3 and Table 2). During resuscitation, pericentral to periportal ratios of NADH fluorescence ranged between 0.91 and 0.97. Calculation of coefficients of variance did not reveal an increase in heterogeneity of interlobular NADH fluorescence during shock and resuscitation (Table 2). Analysis of NADH fluorescence after 60 sec of epiillumination in observation fields of area 2 revealed absolute values of NADH fluorescence which were constantly lower compared to those values assessed at 30 sec of epiillumination (Table 2). Values of NADH fluorescence assessed in area 1 did significantly correlate with values of NADH fluorescence assessed in area 2 after both 30 sec (P õ 0.01; r2 Å 0.84; Fig. 4) and 60 sec of epiillumination (P õ 0.01; r2 Å 0.81). In parallel to shock- and resuscitation-induced
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$2
09-03-97 11:45:15
mvrg
AP: MVR
169
Technical Report
TABLE 1 Changes in Macrohemodynamics, Hematocrit, Systemic Leukocyte Count, and Arterial Blood Gases during Shock and Resuscitation
MAP (mmHg) HR (min01) Hct (%) LCsyst (103/ml) PaO2 (mmHg) O2-sat (%) BE (mmol/liter)
Baseline
S1
S2
S3
R1
R2
R3
110 { 3 374 { 11 52 { 1 12.2 { 1.2 88 { 6 96.7 { 0.6 02.7 { 0.5
83 { 6 357 { 14 n.d. n.d. n.d. n.d. n.d.
58 { 8* 328 { 19 n.d. n.d. n.d. n.d. n.d.
51.2 { 7.5* 334 { 36 47 { 1 11.1 { 1.3 105 { 6 97.9 { 0.3 06.0 { 1.5*
104 { 3 346 { 12 n.d. n.d. n.d. n.d. n.d.
107 { 4 353 { 12 n.d. n.d. n.d. n.d. n.d.
83 { 8 344 { 16 38 { 3* 6.9 { 1.8* 106 { 6 97.8 { 0.4 04.4 { 1.6*
Note. S1, S2, and S3, shock induction by withdrawal of 1.5 ml of arterial blood thrice; R1, R2, and R3, 15, 30, and 60 min after normovolemic resuscitation with 10% hydroxyethylstarch; MAP, mean arterial blood pressure; HR, heart rate; Hct, hematocrit; LCsyst , systemic leukocyte count; PaO2 , arterial oxygen partial pressure; O2-sat, arterial hemoglobin saturation; BE, base excess; n.d., not determined. Mean { SEM; *P õ 0.05 vs baseline.
changes of NADH fluorescence, hepatic tissue PO2 values and hepatocellular excretory function were found significantly reduced during the shock period without complete restoration to preshock baseline values after colloid volume replacement (Table 2). Regression analysis revealed significant correlations between mean values of both hepatic tissue PO2 and bile flow and mean values of NADH fluorescence (Figs. 5 and 6).
demonstrates applicability of the noninvasive and continuous monitoring of NADH fluorescence during hemorrhagic shock and resuscitation. Linear relationships between NADH fluorescence and both hepatic tissue oxygenation and hepatic bile flow indicate that intravital microscopic assessment of NADH fluorescence reflects appropriately the relation between local oxygen supply and demand in hepatic tissue and, thus, liver function in vivo.
DISCUSSION
Intravital Fluorescence Microscopy for the Determination of NADH Fluorescence
Using high-resolution intravital fluorescence microscopy and densitometric techniques, the present study
In contrast to the common use of microlight guides (Ji et al., 1979) in numerous studies on NADH fluores-
TABLE 2 NADH Fluorescence, Coefficient of Variance of Interlobular NADH Fluorescence, Hepatic Tissue PO2, and Bile Flow during Shock and Resuscitation Baseline
S1
S2
S3
R1
47.5 { 4.8 45.1 { 3.9 40.4 { 4.2
53.1 { 5.6 50.2 { 4.1 44.8 { 4.2
56.5 { 4.1* 54.3 { 3.8* 48.2 { 4.5*
59.7 { 7.1* 59.2 { 4.2* 50.6 { 3.1*
0.08 { 0.01 0.06 { 0.02 0.08 { 0.02
0.08 { 0.01 0.06 { 0.01 0.05 { 0.01
0.06 { 0.01 0.05 { 0.01 0.05 { 0.01
7.7 { 1.9
5.4 { 1.5*
2.1 { 0.6*
1.29 { 0.04
1.03 { 0.09*
0.51 { 0.15*
R2
R3
NADH fluorescence (aU) Area 1 Area 2 (30 sec) Area 2 (60 sec)
49.4 { 4.0 49.3 { 4.5 45.1 { 4.9
50.7 { 4.6 47.6 { 4.2 43.8 { 4.4
52.0 { 1.9 49.9 { 3.1 44.1 { 3.6
Coefficient of variance of interlobular NADH fluorescence Area 1 Area 2 (30 sec) Area 2 (60 sec)
0.06 { 0.01 0.08 { 0.02 0.07 { 0.01
Hepatic tissue PO2 (mmHg) Bile flow (ml/min 1 g)
17.9 { 1.8 1.67 { 0.11
0.08 { 0.02 0.04 { 0.01 0.05 { 0.01 11.9 { 1.7 1.03 { 0.01*
0.07 { 0.01 0.06 { 0.03 0.04 { 0.01
0.06 { 0.01 0.05 { 0.01 0.05 { 0.01
8.9 { 1.2
7.4 { 1.5*
1.16 { 0.21*
1.34 { 0.09
Note. S1, S2 and S3, shock induction by withdrawal of 1.5 ml of arterial blood thrice; R1, R2, and R3, 15, 30, and 60 min after normovolemic resuscitation with 10% hydroxyethylstarch. Area 1, hepatic tissue surface area which was epiilluminated for elimination of vitamin A autofluorescence and allowed to recover for 20 min prior to onset of the experiment. Area 2, the hepatic surface tissue area where NADH fluorescence was assessed after 30 and 60 sec of epiillumination without preceding elimination of vitamin A autofluorescence prior to onset of the experiment. Mean { SEM; *P õ 0.05 vs baseline.
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$3
09-03-97 11:45:15
mvrg
AP: MVR
170
Technical Report
FIG. 3. Changes of NADH fluorescence of the hepatic surface (in percentage of baseline) of rats subjected to hemorrhagic shock and resuscitation. B, baseline preshock conditions; S1, S2, and S3, shock induction by withdrawal of 1.5 ml arterial blood thrice; R1, R2, and R3, 15, 30, and 60 min after normovolemic resuscitation with 10% hydroxyethylstarch 200/0.6 iv. Circles, measurements obtained in observation fields of area 1 (epiilluminated and recovered hepatic tissue area prior to the experiment); squares and triangles, measurements obtained at either 30 or 60 sec, respectively, of epiillumination of observation fields in area 2 (non epiilluminated area prior to the experiment). Mean { SEM.
cence for assessment of hepatic ketogenesis and alcohol metabolism (Kashiwagi et al., 1983; Olson and Thurman, 1987), the present study first reports on the use of high-resolution intravital fluorescence microscopy for in vivo assessment of hepatic surface fluorescence of NADH during shock/resuscitation. In contrast to microlight guides with tips of 80 – 170 mm in diameter, whose accurate localization on periportal or pericentral areas is only guaranteed in trained hands (Kashiwagi et al., 1983), intravital fluorescence microscopy using a 10X objective allows for direct visualization and quantitative analysis of NADH fluorescence within a field of 1150 1 850 mm, including both periportal and pericentral segments of two to three liver lobuli. The high resolution of the microscopic technique (Ç1.2 mm) and the visualization of the individual hepatic lobular structures, which can be achieved by standard epiillumination techniques (McCuskey, 1986; Menger et al., 1993; Vollmar et al., 1994b), allow for the calculation of the
ratio of NADH fluorescence intensities between pericentral and periportal regions and the deduction of coefficients of variance of intra- and interlobular NADH fluorescence from data of the identical observation field without potential errors due to epiillumination of nonrepresentative or inadequate regions. Thus, both intraand interlobular heterogeneities of parenchymal tissue oxygenation can easily be detected using the microscopic procedure. For reduction of optical fiber motion artifacts, fluorescence recordings using microlight guides might require synchronization of the detection system with organ surface movements due to breathing or heart activity (Ince et al., 1992). The herein presented approach of intravital fluorescence microscopy of the left liver lobe of the rat, which is exteriorized and placed horizontal to the microscope on an adjustable stage, readily compensates for motion artifacts. In addition to movement, oximetric effects, i.e., variations in the oxygenation of hemoglobin, which might affect organ surface fluorescence due to changes of the hemoglobin spectrum (Ince et al., 1992), should necessarily be considered in the use
FIG. 4. Regression analysis between values of NADH fluorescence assessed in observation fields of area 1 and values assessed at 30 sec of epiillumination in observation fields of area 2. All values obtained during baseline conditions, shock (S1, S2, and S3), and resuscitation (R1, R2, and R3) were included.
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$3
09-03-97 11:45:15
mvrg
AP: MVR
171
Technical Report
FIG. 5. Regression analysis between values of NADH fluorescence assessed at 30 sec of epiillumination in observation fields of area 2 and values of hepatic tissue PO2 (flexible polyethylene microcatheter Clark type PO2 probe). Mean values { SEM obtained during baseline conditions, shock (S1, S2, and S3), and resuscitation (R1, R2, and R3) were included.
of NADH fluorescence for determination of the metabolic state of hepatic tissue. In the present study, however, this aspect can be ignored, since arterial hemoglobin saturation remained unchanged during shock and resuscitation. Finally, changes in blood volume, i.e., a varying number of blood corpuscles, may interfere with fluorescence measurements, since the spectral characteristics of blood pigments make absorption of both ultraviolet excitation light and blue fluorescence dependent on perfusion of the organ surface (Ince et al., 1992). In contrast to the application of microlight guides, direct visualization of the hepatic surface by means of intravital microscopy allows us to restrain quantitative densitometry of only parenchymal fluorescence by excluding the darkly visualized nonparenchymal structures, such as sinusoids and postsinusoidal venules, which would interfere with the accurate estimation of NADH fluorescence in the parenchymal regions. The lower density of parenchymal tissue in pericentral than periportal regions of the lobuli due to the merging of sinusoids into each other and their increase in diameter
might be the reason for slightly lower NADH fluorescence intensities in pericentral than periportal areas and thus ratios õ1, which is in accordance with reports of others (Olson and Thurman, 1987). Assessment of hepatic surface NADH fluorescence is complicated by questions concerning the relative contribution of Ito cell-associated vitamin A autofluorescence, which cannot be distinguished from each other using ultraviolet epiilluminated microfluorographs. However, vitamin A autofluorescence can easily be eliminated due to its rapid photobleaching property. In line with in vitro studies on liver perfusion reported by Suematsu et al. (1993b), we therefore assessed NADH fluorescence in liver surface regions, which were epiilluminated for elimination of vitamin A autofluorescence and were allowed to recover for 20 min prior to onset of the experiment (area 1). To assess whether this prebleaching and recovery is necessary for accurate estimation of NADH fluorescence, we additionally studied NADH fluorescence in nonprebleached surface regions (area 2) after a 30- and 60-sec epiillumination
FIG. 6. Regression analysis between values of NADH fluorescence assessed at 30 sec of epiillumination in observation fields of area 2 and values of hepatic bile flow serving as an indicator of hepatocellular excretory function. Mean values { SEM obtained during baseline conditions, shock (S1, S2, and S3), and resuscitation (R1, R2, and R3) were included.
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$3
09-03-97 11:45:15
mvrg
AP: MVR
172
Technical Report
period. Epiillumination of these hepatic surface regions for more than 20 sec guaranteed also complete elimination of vitamin A autofluorescence (Vollmar et al., 1996b) and further allowed us to calculate the subsequent loss of NADH fluorescence due to a potential photobleaching effect after 30 and 60 sec of continuous light exposure. Corresponding data of the present study reveal a constant loss of NADH fluorescence of approximately 8 – 14% due to its photobleaching property over the time period from 30 to 60 sec of epiillumination. However, the linear correlation of values of NADH fluorescence assessed in area 1 (epiilluminated and recovered area) and 2 (area not epiilluminated prior to onset of the experiment) reveals that the procedure for elimination of vitamin A autofluorescence, including ultraviolet epiillumination with a subsequent NADH recovery period of at least 20 min, is not a prerequisite for the reliable assessment of intraindividual changes of hepatic NADH fluorescence in vivo.
Shock- and Resuscitation-Induced Changes of Hepatic Surface NADH Fluorescence Shed blood volume-controlled lowering of arterial blood pressure followed by isovolemic resuscitation with hydroxyethylstarch caused the well-known pattern of host response, including the decrease in heart rate, systemic hematocrit and leukocyte count, as well as the deterioration of arterial acid – base balance (Vollmar et al., 1994a, 1996a). As shown previously (Vollmar et al., 1994a), hemorrhagic shock is associated with significant nutritive perfusion failure which persists despite resuscitation and accounts for insufficient energy supply to tissue, as indicated in the present study by the lowered tissue oxygenation during shock with inadequate restoration following resuscitation. Restriction of bile flow is considered a direct consequence of insufficient oxygen supply, since hepatocellular excretory function involves ATP-consuming steps and has been repeatedly shown to correlate with hepatocellular ATP levels (Kamiike et al., 1985; Bowers et al., 1987; Karwinski et al., 1989). In parallel to changes of hepatic tissue oxygenation and bile flow, hepatic surface NADH fluorescence increased, with most pronounced changes at the end of hemorrhagic shock and uncomplete restoration during normovolemic resuscitation. The linear cor-
relation between NADH fluorescence and data of hepatic tissue PO2 and bile flow indicates that NADH fluorescence reliably resembles the metabolic state of hepatic tissue in vivo. In conclusion, high-resolution intravital fluorescence microscopic assessment of hepatic surface NADH fluorescence represents an accurate measure for hepatocellular energy metabolism/tissue oxygenation and a promising tool due to the noninvasive and continuously applicable technical approach.
REFERENCES
Bowers, B. A., Branum, G. D., Rotolo, F. S., Watters, C. R., and Meyers, W. C. (1987). Bile flow: An index of ischemic injury. J. Surg. Res. 42, 565 – 569. Chance, B. (1976). Pyridine nucleotide as an indicator of the oxygen requirements for energy-linked functions of mitochondria. Circ. Res. 38(Suppl. 1), I31 – I38. Chance, B., Cohen, P., Jo¨bsis, F., and Schoener, B. (1962). Intracellular oxidation-reduction states in vivo. The microfluorometry of pyridine nucleotide gives a continuous measurement of the oxidation state. Science 137, 499 – 508. Chance, B., and Schoener, B. (1965). A correlation of absorption and fluorescence changes in ischemia of the rat liver, in vivo. Biochem. Z. 341, 340 – 345. Chance, B., Schoener, B., Krejci, K., Russmann, W., Wesemann, W., Schnitger, H., and Bu¨cher, Th. (1965). Kinetics of fluorescence and metabolite changes in rat liver during a cycle of ischemia. Biochem. Z. 341, 325 – 333. Ince, C., Coremans, J. M. C. C., and Bruining, H. A. (1992). In vivo NADH fluorescence. Adv. Exp. Med. Biol. 317, 277 – 296. Ji, S., Chance, B., Nishiki, K., Smith, T., and Rich, T. (1979). Microlight guides: A new method for measuring tissue fluorescence and reflectance. Am. J. Physiol. 236, C144 – C156. Kamiike, W., Nakahara, M., Nakao, K., Koseki, M., Nishida, T., Kawashima, Y., Watanabe, F., and Tagawa, K. (1985). Correlation between cellular ATP level and bile excretion in the rat liver. Transplantation 39, 50 – 55. Karwinski, W., Husoy, A.-M., Farstad, M., and Soreide, O. (1989). Sixty minutes of normothermic ischemia in the rat liver: Correlation between adenine nucleotides and bile excretion. J. Surg. Res. 46, 99 – 103. Kashiwagi, T., Lindros, K. O., Ji, S., and Thurman, R. G. (1983). Aldehyde dehydrogenase-dependent acetaldehyde metabolism in periportal and pericentral regions of the perfused rat liver. J. Pharmacol. Exp. Therap. 224, 538 – 542.
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$3
09-03-97 11:45:15
mvrg
AP: MVR
173
Technical Report
McCuskey, R. S. (1986). Microscopic methods for studying the microvasculature of internal organs. In Physical Techniques in Biology and Medicine Microvascular Technology (C. H. Baker and W. F. Nastuk, Eds.), pp. 247 – 264. Academic Press, New York. Menger, M. D., Vollmar, B., Glasz, J., Mu¨ller, M. J., Post, S., and Messmer, K. (1993). Hepatic ischemia/reperfusion. Prog. Appl. Microcirc. 19, 106 – 124. Obi-Tabot, E. T., Hanrahan, L. M., Cachecho, R., Beer, E. R., Hopkins, S. R., Chan, J. C. K., Shapiro, J. M., and LaMorte, W. W. (1993). Changes in hepatocyte NADH fluorescence during prolonged hypoxia. J. Surg. Res. 55, 575 – 580. Olson, M. J., and Thurman, R. G. (1987). Quantitation of ketogenesis in periportal and pericentral regions of the liver lobule. Arch. Biochem. Biophys. 253, 26 – 37. Rappaport, A. M. (1973). The microcirculatory hepatic unit. Microvasc. Res. 6, 212 – 228. Renault, G., Raynal, E., Sinet, M., Berthier, J. P., Godard, B., and Cornillault, J. (1982). A laser fluorimeter for direct cardiac metabolism investigation. Optics Laser Technol. 14, 143 – 148. Suematsu, M., Oda, M., Suzuki, H., Kaneko, H., Watanabe, N., Furusho, T., Masushige, S., and Tsuchiya, M. (1993a). Intravital and electron microscopic observation of Ito cells in rat hepatic microcirculation. Microvasc. Res. 46, 28 – 42. Suematsu, M., Suzuki, H., Ishii, H., and Tsuchiya, M. (1993b). Paradox
of oxygen-radical-dependent cell injury in the hypoxic liver microcirculation. Prog. Appl. Microcirc. 19, 127 – 138. Sumimoto, K., Inagaki, K., Yamada, K., Kawasaki, T., and Dohi, K. (1988). Reliable indices for the determination of viability of grafted liver immediately after orthotopic transplantation. Bile flow rate and cellular adenosine triphosphate level. Transplantation 46, 506 – 509. Vollmar, B., Lang, G., Menger, M. D., and Messmer, K. (1994a). Hypertonic hydroxyethylstarch restores hepatic microvascular perfusion in hemorrhagic shock. Am. J. Physiol. 266, H1927 – H1934. Vollmar, B., Glasz, J., Leiderer, R., Post, S., and Menger, M. D. (1994b). Hepatic microcirculatory perfusion failure is a determinant of liver dysfunction in warm ischemia-reperfusion. Am. J. Pathol. 145, 1421 – 1431. Vollmar, B., Preissler, G., and Menger, M. D. (1996a). Small-volume resuscitation restores hemorrhage-induced microcirculatory disorders in rat pancreas. Crit. Care Med. 24, 445 – 450. Vollmar, B., Ru¨cker, M., and Menger, M. D. (1996b). A new method for the intravital microscopic quantification of hepatic sinusoidal perfusion failure using the dye bisbenzamide H33342. Microvasc. Res. 51, 250 – 259. Zeintl, H., Tompkins, W. R., Messmer, K., and Intaglietta, M. (1986). Static and dynamic microcirculatory video image analysis applied to clinical investigations. Prog. Appl. Microcirc. 11, 1 – 10.
Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
AID
MVR 2028
/
6z11$$$$$4
09-03-97 11:45:15
mvrg
AP: MVR