Open Access
Add to BibSonomyAbstract
The potential of NADH autofluorescence as an in vivo intrinsic optical signature to monitor tissue metabolism is well recognized and supported by experimental results mainly in animal models. In this work, we propose a non-contact implementation of this method using large area excitation and employing a normalization method to account for non-metabolic signal changes. Proof of principle in vivo experiments were carried out using an autofluorescence imaging experimental system and a rat renal ischemia model. A hand-held fiber-optic probe was utilized to test the ability of the signal normalization method to address operational conditions associated with the translation of this method to a clinical setting. Preliminary pre-clinical in vivo test of the probe system was carried out using the same rat model.
©2009 Optical Society of America
1. Introduction
Assessing the degree of ischemic injury in a tissue or organ in a timely manner in a clinical setting remains an unsolved problem. Histopathology is invasive, requiring cutting, fixing, and staining of the tissue which can have a turnaround of several days and introduces additional trauma to the organ. Microdialysis is a real-time method which samples the extracellular environment for certain metabolic byproducts but is also invasive, requiring direct insertion of a 2 mm diameter needle into the tissue. When dealing with tissue of critical or unknown status such as in clinical transplantation, a non-invasive method is preferred.
The autofluorescence signal from NADH under ultraviolet excitation has been shown to be sensitive to conditions of alteration in oxygen supply [1] including ischemia and reperfusion [2,3]. In most studies, a fiber optic probe is placed in contact with the kidney surface to interrogate a small region between the emitting and receiving fibers. This method, although highly promising, may have limitations arising from the fact that it samples the local optical properties of the tissue which can change during the measurement for two main reasons. First, pressure applied by the probe can alter local hemodynamics which can directly influence local metabolic activity by interfering with normal oxygen supply and saturation [4]. Since UV light penetrates only ~200 microns in tissue, measurements of autofluorescence intensity by UV excitation may be particularly sensitive to surface pressure [5]. Second, the inability to monitor precisely the same local region when multiple measurements are made over time renders the measurement sensitive to variations in local microstructure. In addition, when making a judgment of function of tissue as a whole, as is frequently done in transplantation, it may be beneficial to interrogate a larger tissue area than has been previously performed with conventional microfluorimeters or fiber probes requiring contact with the tissue. These issues indicate that measurement of tissue NADH fluorescence may be preferably done using a non-contact method and a large excitation area for signal acquisition.
In a recent study we reported on the use of such a non-contact method to monitor the kidney response to ischemia and reperfusion using NADH autofluorescence [6]. Specifically, we utilized imaging of the entire exposed organ surface under 355 nm excitation for signal acquisition. As large a surface of the organ was measured in order to average any heterogeneous tissue response and to assess the status of the kidney as a whole. The time-dependent average autofluorescence intensity showed characteristic responses during reperfusion based on duration of injury. A double exponential model was adopted to quantify these changes. Time constants derived from fits of the data to this model yielded longer time constants for longer exposures to ischemic injury times. This experimental arrangement provided initial proof-of-principle results but requires modifications and improvements in order to become suitable for translation into a clinical setting.
A clinically more suitable approach is a hand-held fiber probe that can display spectroscopic information in real-time. However, signal artifacts independent of changes in tissue metabolism can be introduced and interfere with the injury time-dependent signal. Specifically, while in our previous work the camera and illumination fiber were in a fixed position with respect to the target organ (kidneys), this cannot be expected using a portable non-contact probe utilized intermittently during a surgical procedure; the distance between the fiber tip and the kidney surface, as well as the angle between the illumination direction and the kidney surface, can change and thus significantly affect the measured signal intensity. In addition, tissue hydration state can change during a surgical procedure when normal supply of blood flow may be interrupted and the tissue is exposed to air for extended periods while intermittently rewetted with saline (as is done in the clinic). This in turn affects the scattering properties of the tissue and the way and amount of the excitation light that is reflected, especially at the air-tissue interface. Consequently, a hand-held fiber optic approach requires a normalization method to account for these non-metabolic signal changes.
The objective of this work is two-fold: a) propose a dual UV excitation method for signal normalization and demonstrate that it preserves the quality of the data and b) design and test a non-contact, large illumination area hand-held probe to obtain preliminary measurements that demonstrate that the proposed normalization method can account for changes in kidney hydration state, distance between fiber tip and kidney surface, and angle between the illumination direction and the kidney.
2. Experimental design
2.1 Normalization method
A commonly used approach to account for a change in the scattering properties of the tissue or the changing illumination-collection geometry is to monitor the reflected excitation intensity [1-3]. However, this is not a sufficient solution in our application because it cannot account for the angular distribution of the reflected light (specular and diffuse components) in a non-contact configuration. These light scattering components can also change during the course of the measurement in a pre-clinical or clinical setting when tissue may dehydrate and rehydrate with the application of saline during surgery. Consequently another signal normalization solution is required.
Using the emission of other fluorophores can be considered as a more precise normalization method. Among these fluorophores, collagen and elastin are structural components comprising the extracellular matrix in tissue that fluoresce under the primary 355 nm excitation. These structural components are not expected to deteriorate or accumulate appreciably in the time scale of minutes to a couple of hours of ischemia, and should in principle provide a signal suitable for normalization. However, their signal intensity in renal tissue is impractically low compared to that from NADH [7]. Another fluorophore that has been considered is flavins, with excitation peak at 460 nm and peak of emission at 525 nm. Most significantly, however, flavins (like NADH) are electron carriers in the transport chain and thus their concentration is not independent of metabolic state. It is also important to note that excitation light for flavins penetrates deeper into the kidney, where metabolic activity and hemodynamics are known to differ [8]. In addition, flavin fluorescence efficiency is very weak compared to NADH at physiological temperature [9]. After considering these alternatives, we choose 266 nm as the secondary excitation aiming at using the emission by tryptophan for normalization. Tryptophan is an amino acid not directly involved in cellular metabolic function, and so its concentration should not depend on duration of ischemia. Also, tryptophan has a relatively high quantum yield and can provide sufficient emission intensity up to about 500 nm. In addition, previous work has suggested that emission longer than 395 nm under excitation near 266 nm was minimally sensitive to ischemia and reperfusion in anesthetized rats [10].
Based on the above discussion, we employ in this work the autofluorescence intensity under the primary excitation at 355 nm (arising mainly from NADH and referred to henceforth as the NADH signal), as well as the autofluorescence intensity under a secondary excitation at 266 nm, delivered through the same fiber as that of the primary, for normalization to account for geometrical and hydration signal changes. The normalized signal represents the ratio of autofluorescence under 355 nm divided by that under 266 nm (henceforth referred to as the signal ratio). To test the hypothesis that this dual UV excitation method can provide the means for signal normalization, we first examined if the temporal profile (as a function of reperfusion time, see ref. Raman et al.) of the signal ratio was characterized by similar fitting constants to those obtained by monitoring the change of the NADH signal (Section 2.2.3).
2.2 Normalization validation in imaging arrangement
2.2.1 Renal ischemia model
All rat procedures were approved by the University of California, Davis Animal Use and Care Administrative Advisory Committee. These rats were the same experimental animals that were described in a previous report, where rat preparation has been described in detail [6]. In brief, adult male Wistar-Furth rats were anesthetized with 2% isoflurane inspiratory concentration delivered with O2 at 1L/min. A midline laparatomy was performed and the left renal pedicle was exposed. Saline was applied topically to the kidneys every 5 minutes during imaging to keep them moist. A clamp was applied to the exposed vascular pedicle of the left kidney resulting in unilateral ischemic injury, while the contralateral kidney was not clamped. Injury periods of either 20 minutes (n=12 rats), 50 minutes (n=12), or 150 minutes (n=15) were followed by clamp release and reperfusion for at least 40, 60, or 60 minutes, respectively.
2.2.2 Optical imaging setup
The experimental imaging setup is similar to that used previously [6] except for the inclusion of a diode-pumped solid state (DPSS) 266 nm laser source operating at 7 kHz repetition rate, 3 μJ/pulse (Power Technology, Little Rock, AR, DTL-382QT) whose beam was aligned to co-propagate with the primary 355 nm excitation (10kHz, 5 μJ/pulse, 0.22 mJ/cm2) and coupled into the same delivery fiber with a plano-convex fused silica lens (f=10 cm, Ø=2.54 cm, Edmund Optics, Barrington, NJ). The experimental layout is depicted in Fig. 1. Both beams were allowed to expand to ~6 cm in diameter to fully illuminate both kidneys 45 cm away from the fiber bundle tip, corresponding to dosages of 0.22 mJ/cm2 (355 nm) and 0.30 mJ/cm2 (266 nm) for an exposure time of 5 seconds each (concurrent with image acquisition) at 15-second intervals. Dual sources were operated on a time sharing basis by computer-controlled electronic shutters (NM Laser Products, Inc., Sunnyvale,CA). The fluorescence from a small piece of plastic tape placed near the kidneys was used to monitor the excitation laser power in each image. A large collection mirror redirected emitted light from the kidneys through a 420-640 band-pass filter, after which a lens focused the image onto a liquid nitrogen-cooled charge-coupled device (CCD) (Roper Scientific, Trenton, NJ). A first image was acquired prior to clamping to establish the baseline intensity under either excitation. Image acquisition continued immediately after clamping and lasted through the injury and reperfusion phases. The average autofluorescence intensity under either wavelength was recorded over as large an area of the kidney as possible (~1.5 cm2).
Fig. 1. Schematic of imaging experimental arrangement. DC=dichroic mirror, F=filter, CM=collection mirror.
2.2.3 Signal analysis
The signal ratio profile was modeled using a double exponential function developed in our previous study to model the direct (not normalized) NADH signal profile under 355 nm [6]:
where RN is the contribution from the first component to the detected signal ratio corresponding to NADH concentration, RE is the contribution to the second component representing environmental influence on the signal ratio, and RN0 and RE0 are the values of the first signal ratio data point upon reperfusion. ΔRN and ΔRE represent the magnitude of decrease and increase, respectively, of the first and second component. Δτ is the time delay of the first component and τN and τE are the relaxation times of components 1 and 2, respectively. Fits of the model to the signal ratio profile data were performed using custom code written in MATLAB (The MathWorks, Natick, MA), and best-fit values of the time constants were extracted. The results for the time constants derived under 355 nm excitation alone (originally discussed in Ref. 6) are included in Table 1 for comparison under the sub-heading “355 nm excitation.” The time constants within each time group but under different methods (NADH signal vs. signal ratio) were compared using a two-sample t-test to determine if using the signal ratio significantly influenced the values of the time constants derived from the NADH signal under 355 nm excitation. Then, one-way analysis of variance (ANOVA) was employed to determine any significant association of derived time constants with injury time, followed by a post-hoc Tukey test. These ANOVA results were compared to those previously published using the NADH signal under 355 nm alone [6].
Table 1. Time constants and ANOVA results for three injury times (N=12 rats for each group) derived from the autofluorescence intensity under 355 nm excitation vs. the signal ratio.
2.3 Normalization validation in probe system
2.3.1 Preliminary probe system design
This fiber optic probe system was designed to incorporate the signal ratio method with the following features that would be relevant in a clinical setting: large illumination area, non-contact excitation/signal collection of the kidney surface, common excitation fiber for both laser excitation sources, and real-time display of the signal ratio. The experimental layout and probe design are depicted in Fig. 2. The intensity of either laser beam during signal acquisition is monitored using the reflection from a quartz slide to direct a portion of the beam onto a UV-sensitive photodiode (Thorlabs, Newton, NJ, FGAP71). The photodiode current is amplified with a transistor circuit in a common-collector configuration and generates the voltage output across a load resistor. The intensities of the 355 nm and 266 nm lasers entering the excitation fiber are controlled with neutral density (ND) filters.
The laser sources, shutters, dichroic mirrors, and coupling lens are the same as those in the imaging configuration (Section 2.2.2). Both beams are coupled into the low-solarization silica excitation fiber (400 μm diameter core, numerical aperture (NA)=0.22) of a custom 2.5 m-long fiber optic probe (Romack, Williamsburg, WV). A low NA was chosen for the excitation fiber to avoid overfilling the target tissue (a human kidney in future applications) with excitation light at clinically relevant distances between the tip of the probe and the tissue surface. The probe tip is 5 mm in diameter and terminates with a 2 mm-thick 400 nm long-wavelengths-pass (LP) filter (Chroma Technology, Rockingham, VT). The excitation fiber is located in the middle and is flush to the outer surface of the LP filter after passing through a 500 μm-diameter hole drilled through the center of the filter.
The emission is collected by the ~100 collection fibers of the probe (hard clad silica, 400 μm diameter core, NA=0.39) surrounding the excitation fiber after passing through the 400 nm LP filter. The 400 nm LP was chosen to select the spectral region covering NADH emission while discriminating against emission from other chromophores (collagen , elastin) under 355 nm excitation that emit at wavelengths lower than does NADH. In addition, emission under 260 nm excitation in this spectral range was previously found to be minimally sensitive to ischemia and reperfusion [10]. This filter also allows for rejection of the excitation light entering the emission fibers which can cause secondary emission signal (due to fluorescence from the collection optics) that would be proportional to the scattered light [11].
Fig. 2. Schematic of fiber probe experimental arrangement. NDF=neutral density filter, BS=beam splitter, CF=collection fibers, PD=photodetector, PM=photomultiplier.
The autofluorescence signal at the output of the collection fibers passes through a second 400 LP filter (to reject any remaining excitation light) and a 10% transmission ND filter (Spindler-Hoyer, Germany, 371152) in order to prevent saturation of the channel photomultiplier detector (Perkin Elmer Optoelectronics, GmbH, Germany, CPMMD1344) and maintain operation with linear response. The entire system including the lasers, optics, and electronics fit onto a portable optical breadboard with a footprint of 2’×1’.
Synchronization of the laser shutters with the photodiode and photomultiplier readout under each excitation wavelength is controlled by LABVIEW software (National Instruments, Austin, TX) using custom code. A custom user interface allows the user to set 1) the exposure time during which the photomultiplier voltage is sampled under either excitation wavelength (3 seconds for our experiments), 2) the sampling rate (10 kHz for our experiments), 3) the readout time to compute the average of the sampled voltages (2 seconds for our experiments) until switching open the 266 nm shutter, 4) the dead time following the 266 nm exposure before beginning the next cycle, and 5) the number of cycles determining the duration of the experiment. The interface displays the individual photomultiplier voltages and generates the real-time signal ratio following each cycle.
2.3.2 System characterization
Probe characterization measurements were performed on metabolically inactive grocery-grade porcine kidney samples. These samples were chosen in order to avoid signal variation due to metabolic change of the kidney in a live rat as well as to avoid overfilling the smaller rat kidney with excitation light at the operational probe-kidney distances used for normalization validation. Specifically, the signal ratio in kidney was monitored over a range of separation distances d (10-55 mm) from the fiber tip to the kidney surface. This range of distances corresponds to a range of intensities from 0.06-0.84 mW/cm2 for 355 nm and 0.37-4.7 mW/cm2 for 266 nm, the range in which system linear response was maintained. At d=25 mm, average incident intensities were 0.20 mW/cm2 and 1.1 mW/cm2 for 355 nm and 266 nm laser sources, respectively.
2.3.3 In vivo testing of the probe
In vivo testing of the probe was performed using 4 rats prepared identically as described in Section 2.2.1. Ischemia was induced by clamping the left renal pedicle for 30 minutes, another relatively short injury time from which the rat would be likely to recover [12,13], followed by clamp release and 45 minutes of reperfusion. Probe was placed d=15 mm from the kidney surface (corresponding to intensities of 0.55 mW/cm2 and 3.1 mW/cm2 for 355 nm and 266 nm sources, respectively) and exposure time was 10 seconds for each excitation. Saline was applied topically to each kidney every 5 minutes to keep the tissue moist.
3. Results
3.1. Normalization validation using imaging arrangement
Figure 3(a) shows typical temporal response of the average autofluorescence intensity under 355 nm and 266 nm excitation extracted from the corresponding images from a rat kidney undergoing 50 minutes of injury followed by 60 minutes of reperfusion. The resultant signal ratio is also shown. All profiles are displayed in arbitrary units with the value immediately prior to clamping always set to 1 to facilitate direct comparison of these profiles. The NADH signal increased slightly within the first minute of clamping, followed by a decrease (as discussed in our previous study [6]) that settled to a minimum value of about 0.66 before the end of the injury phase. Upon unclamping, the NADH signal increased in two phases as reported in our previous study [6]. The autofluorescence signal under 266 nm (henceforth referred to as the 266 nm excitation signal) appeared to display a similar temporal profile though the changes were smaller: the decrease during injury was only to about 0.88, and the magnitude of the reperfusion inflection from peak to valley was relatively smaller. The signal ratio exhibited dynamics similar to that of the NADH signal with initial decrease reaching to about 0.75. In the reperfusion phase, while the signals under either excitation (355 nm and 266 nm) overshot baseline at the end of reperfusion to about 1.06, the normalized signal ratio stays at baseline.
Figure 3(b) illustrates one example of best-fit of the signal ratio to the model described in section 2.2.3 during 60 min of reperfusion following 50 min of ischemia. The profiles of the two components and their characteristic time constants are displayed. Figures 3(c) and 3(d) summarize the results for Δτ, τE, and τN obtained from all rat kidneys imaged, indicating that longer injuries tended to yield longer time constants. The profiles of 3 rats did not exhibit any return to baseline and were characterized with infinitely long time constants.
The fitting parameters obtained from all experimental animals using the reperfusion components of the temporal profile of the NADH signal as well as that of the signal ratio are summarized in Table 1. Specifically, the mean values and standard deviation of each time constant (Δτ, τN, and τE) are provided on the left side of the table. Comparison of the corresponding fitting values obtained from the NADH signal and the signal ratio indicates that they are similar. For all 3 time constants and for all injury times, the difference between the corresponding time constants obtained from each method was not significant (2 sample t-test, p>0.1). These results suggest that the signal normalization process did not significantly change the values of the time constants obtained using the NADH autofluorescence under 355 nm excitation.
Fig. 3. (a) Typical temporal responses of autofluorescence intensities (under 355 nm excitation and 266 nm excitation), and the resultant signal ratio for rat kidney that underwent 50 min ischemia followed by 60 min reperfusion. (b) Fit of signal ratio in 3(a) during reperfusion to the model. Values of extracted time constants are displayed. (c) The values of τE vs Δτ and (d) τE vs τN obtained from fitting the reperfusion signal ratio profile from all rats.
One-way ANOVA showed that the time constants obtained from the signal ratio reperfusion profiles were significantly different with respect to injury time, as also found using the NADH signal. F-values for this study were 15.65, 16.42, and 25.19 for Δτ, τN, and τE, respectively (p<0.0001 for all), compared to p<0.0001, p<0.01, and p<0.0001, respectively, as found using the direct NADH signal. Tukey post-hoc test, which was performed pairwise between all injury time groups for each time constant (Table 1, right side), also lead to the same conclusion, namely that significant differences existed between the time constants of the 150 min group vs. those of either of the shorter injury groups, but not between those of the 20 min and 50 min groups. Tukey score q>3.47 signifies p<0.05 and q>4.43 signifies p<0.01.
3.2. Normalization validation using the fiber probe
The results discussed in the previous section suggest that the normalization process does not significantly alter the time constants from data obtained in vivo using the imaging configuration. The next step is to validate the normalization process against the scattering properties of the tissue or the changing illumination-collection geometry using the fiber probe configuration. First, the linear response of the probe system was tested. The measurements described below were made in the linear regime of these detectors. Background photomultiplier voltage was measured for each laser by turning on each source, pointing the probe in the air, and subtracting from this voltage that due to ambient light. This background voltage was found to be ≤1mV under either excitation source and is believed to be due to fluorescence of the optics as well as photomultiplier dark current. All experiments were performed with minimal ambient light, and all background noise (including that from ambient light) was subtracted from the voltages in the following measurements.
Each of the three criteria, namely signal ratio sensitivity to kidney hydration state, distance between probe distal tip and kidney surface, and angle between illumination direction and kidney surface, was tested. Test measurements were made on grocery-grade porcine kidney samples. To observe the effect on NADH signal and signal ratio arising from changes in tissue hydration, experiments were performed using dry kidney as obtained from a local store. The kidneys were submerged in saline containing ice, and measurements were immediately taken. Typical experimental results are shown in Fig. 4. The NADH signal and the 266 nm excitation signal (with initial emission intensity in all cases set to 1) began to increase at a similar rate before leveling off, while the signal ratio was relatively constant. This experiment was thereafter repeated in different locations but no additional changes were observed, indicating that the observed behavior is related to changes in the optical properties of the dehydrated kidney tissue after the initial interaction with the saline solution.
Fig. 4. Responses of ex vivo porcine kidney tissue autofluorescence and signal ratio immediately submerged in saline bath.
Next, the signal ratio response to probe-tissue distance was tested. Figure 5(a) shows the NADH signal, the 266 nm excitation signal, and the signal ratio for distances between the probe tip and the kidney surface (at normal incidence) between 10 and 55 mm. In all profiles, the intensity value at d=10 mm was set to 1 to better allow comparison. To eliminate emission intensity dependence on kidney hydration state, the kidney samples in this test were submerged in saline for 2 hours prior to and during measurement. The signal ratio remained nearly constant (varied less than 4%) over these distances. At d less than 10 mm, the tissue exhibited signs of photobleaching under either excitation in which the photomultiplier voltage was observed to progressively decrease with exposure time.
The next experiment was to test the signal ratio response to the angle between the probe (excitation) direction and the surface of the kidney. Representative experimental results from this set of measurements are shown in Fig. 5(b). The probe was rotated with respect to the tissue surface to cover excitation angles from 0 to 45°. During rotation, the fiber tip-to-kidney distance was kept at ~25 mm. The results demonstrate that as the excitation incidence angle increased, the emission collection decreased but the signal ratio was maintained constant within 5%.
Fig. 5. Sensitivity of autofluorescence intensities and signal ratio to (a) probe-tissue separation and (b) probe-tissue angle for grocery-grade porcine kidney immersed in saline.
3.3. Probe testing in vivo
Following these normalization method and probe characterization results, the probe system was tested for its ability to reproduce the optical response described in the imaging results. Figure 6 shows a typical result depicting the temporal profile of the NADH signal, the 266 nm excitation signal, and the signal ratio of a kidney undergoing 30 minutes of ischemia followed by 45 minutes of reperfusion in an anesthetized rat, scaled to the value prior to clamping. Measurements were taken every 5 minutes. The NADH signal progressively decreased during injury and increased during reperfusion with a similar response to that observed in the imaging configuration for short injury times. This signal ended above baseline at around 1.2. The 266 nm excitation signal did not change much during the injury phase but increased rapidly during the reperfusion phase to reach ~1.35. The signal ratio decreased similarly to the NADH signal during injury and increased during reperfusion. While both signals overshot baseline, the signal ratio ended close to baseline at about 0.9.
4. Discussion
The example profiles in Fig. 3(a) reveal that the emission under 266 nm excitation measured with a 420 nm long-pass filter was not completely insensitive to ischemia and reperfusion, but still considerably less sensitive than that under 355 nm. This may in part be due to 1) the fact that NADH fluorescence is mixed with that of tryptophan under excitation at 266 nm since our collection is not at the peak of tryptophan emission (340 nm) but is at > 420 nm, 2) effects arising from changes in blood content and optical properties manifested via absorption of the excitation and emission light, 3) a change in the scattering properties of the tissue during ischemia and reperfusion, and 4) changes in temperature and pH during injury and reperfusion which can affect the efficiency of tissue fluorescence.
Fig. 6. In vivo responses of autofluorescence intensities and the resultant signal ratio using the fiber probe system on a rat kidney subjected to 30 min ischemia followed by 45 min reperfusion.
Comparison of the time constants obtained using the signal ratio versus those obtained using the NADH signal demonstrates that the proposed normalization process does not significantly change the values of the time constants. The main difference seems to be the smaller mean value and tighter standard deviation for τN in the 20 min injury group obtained using the signal ratio. Perhaps the normalization process (signal ratio) is removing artifacts that occur upon removal of the clamp, the time period during which τN begins to significantly change in the case of the shorter injuries when there exists little to no delay time Δτ of the NADH concentration (first) component.
The post-hoc finding that significant differences exist between time constants of the 150 min group vs. those of either of the shorter injury groups but not between those of the 20 min and 50 min group is the same as that arrived at under 355 nm alone and provides additional evidence in support of the conclusions that 1) the results are not significantly altered when dividing by the autofluorescence under 266 nm and that 2) prolonged injury yields longer time constants. It is interesting to note that the post-hoc test revealed that the significance level in the case of the signal ratio was equal to or greater than that found under 355 nm alone (signal ratio Tukey score ≥ NADH signal Tukey score). Specifically, the signal ratio yielded more distinct values of the time constants between the 20 min vs. 150 min groups than did the signal under 355 nm alone and may again be related to the kidney movement artifact immediately following unclamping and the behavior of the time constants of the shorter injury during this period. The signal ratio does not affect the quality of the results derived from fits to the imaging data and in some cases even reduced the standard deviation of these measurements.
We hypothesized that the most suitable approach to implement this method, which may provide sensitivity to tissue injury, in the clinical setting is through the use of a hand-held probe configuration. The use of the signal ratio enables this implementation by providing a normalization method to account for signal changes independent of metabolic activity. Specifically, Figs. 4 and 5 show that probe-tissue distance and angle, as well as tissue hydration state (all which are factors independent of kidney metabolic state), affect the autofluorescence signal under 355 nm excitation of a non-contact fiber optic probe system. However, the results indicate that the signal ratio normalization method can adequately resolve this problem and isolate the changes related to metabolism. We hypothesize that the similar response under either excitation as the tissue hydrates (Fig. 4) may be an indication of a change in the scattering properties of the tissue. The ratio method may potentially partially compensate for NADH signal change due to change in tissue temperature, such as occurs during organ hypothermic preservation and rewarming, since fluorophores in general increase in quantum efficiency and therefore in intensity at lower temperatures. Similarly, change in the pH known to take place during tissue injury can affect the fluorescence efficiency of tissue fluorophores. However, further study is necessary to measure whether the effect of changing temperature and pH influences the measured signal under each excitation wavelength. In the signal response shown in Fig. 6, the initial increase in NADH signal was not recorded due to low sampling frequency.
The main concern of using 266 nm excitation on living human tissue is its phototoxicity. However, recent experimental results by our group [unpublished] suggest that in a human kidney, the 266 nm excitation light is largely confined to the renal capsule, a ~100 μm protective membrane surrounding a human kidney, consisting primarily of structural components such as collagen and elastin and not metabolites as found in the cortex. This confinement is expected to advance the use of 266 nm to correct for non-metabolic changes in signal in human kidneys. By keeping exposure times and total power low, in particular reducing the intensity at 266 nm to a minimal level, accepted dose limits can be met in this design [14]. In addition, a modification can be made in a future clinical design to allow probe use in vivo in the presence of ambient light and background noise. By modulating the excitation at a frequency different from that of the ambient noise, a lock-in-amplifier stage can be included in signal detection where the amplifier separates and amplifies the desired signal. Also, a larger number of collection fibers can be used to increase signal collection efficiency.
Acknowledgments
This research is supported by funding from the Center for Biophotonics, an NSF Science and Technology Center, managed by the University of California, Davis, under Cooperative Agreement No. PHY 0120999. This work was performed in part under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
References and links
1. B. Chance, P. Cohen, F. Jobsis, and B. Schoener, “Intracellular oxidation-reduction states in vivo,” Science 137, 499–508 (1962). [CrossRef] [PubMed]
2. A. Mayevsky and B. Chance, “Intracellular oxidation-reduction state measured in situ by a multichannel fiber-optic surface fluorometer,” Science 217, 537–540 (1982). [CrossRef] [PubMed]
3. J. M. C. C. Coremans, M. Van Aken, D. C. W. H Naus, M. F. Van Velthuysen, H. A. Bruining, and G. J. Puppels, “Pretransplantation assessment of renal viability with NADH fluorimetry,” Kidney Int. 57, 671–683 (2000). [CrossRef] [PubMed]
4. R. Reif, M. S. Arnorosino, K. W. Calabro, O. A’Amar, S. K. Singh, and I. J. Bigio, “Analysis of changes in reflectance measurements on biological tissues subjected to different probe pressures,” J. Biomed. Opt. 13, 07424LR (2008). [CrossRef]
5. Valery Tuchin, Tissue Optics: Light Scattering Methods and Instruments for Medical Diagnosis (SPIE Press, 2000).
6. R. N. Raman, C. D. Pivetti, D. L. Matthews, C. Troppmann, and S. G. Demos, “Quantification of in vivo autofluorescence dynamics during renal ischemia and reperfusion under 355 nm excitation,” Opt. Express 16, 4930–4944 (2008). [CrossRef] [PubMed]
7. A. Hansch, D. Sauner, I. Hilger, J. Bottcher, A. Malich, O. Frey, R. Brauer, and W. A. Kaiser, “Autofluorescence spectroscopy in whole organs with a mobile detector system,” Acad. Radiol. 11, 1229–1236 (2004). [CrossRef] [PubMed]
8. R. F. Pitts, Physiology of the Kidney and Body Fluids: An Introductory Text, 3rd ed. (Year Book Medical Publishers, 1974).
9. B. Chance, B. Schoener, R. Oshino, F. Itshak, and Y. Nakase, “Oxidation-reduction ratio studies of mitochondria in freeze-trapped samples,” J. Biol. Chem. 254, 4764–4771 (1979). [PubMed]
10. J. T. Fitzgerald, A. P. Michalopoulou, C. D. Pivetti, R. N. Raman, C. Troppmann, and S. G. Demos, “Realtime assessment of in vivo renal ischemia using laser autofluorescence imaging,” J. Biomed. Opt. 10, 044018 (2005). [CrossRef]
11. A. Weber and J. S. Schulz, “Fiber-optic fluorimetry in biosensors: comparison between evanescent wave generation and distal-face generation of fluorescent light,” Biosens. Bioelectron. 7, 193–197 (1992). [CrossRef]
12. K. C. Calman, R. O. Quinn, and P. R. Bell, “Metabolic aspects of organ storage and the prediction of viability,” in Organ Preservation, D. E. Pegg, ed. (Churchill Press, 1973), pp. 225–240.
13. P. Jablonski, B. O. Howden, D. A. Rae, C. S. Birrell, V. C. Marshall, and J. Tange, “An experimental model for assessment of renal recovery from warm ischemia,” Tranplantation 35, 198–204 (1983). [CrossRef]
14. American National Standards Institute. Z136.1 American National Standards Institute (1993).
