Characterization and Predictive Value of Near Infrared 2-Deoxyglucose Optical Imaging in Severe Acute Pancreatitis

Background Studying the uptake of 2-deoxy glucose (2-DG) analogs such as 2-Deoxy-2-[18F] fluoroglucose (FDG) is a common approach to identify and monitor malignancies and more recently chronic inflammation. While pancreatitis is a common cause for false positive results in human studies on pancreatic cancer using FDG, the relevance of these findings to acute pancreatitis (AP) is unknown. FDG has a short half-life. Thus, with an aim to accurately characterize the metabolic demand of the pancreas during AP in real-time, we studied the uptake of the non-radioactive, near infrared fluorescence labelled 2-deoxyglucose analog, IRDye® 800CW 2-DG probe (NIR 2-DG; Li-Cor) during mild and severe biliary AP. Methods Wistar rats (300 g; 8–12/group) were administered NIR 2-DG (10 nM; I.V.). Mild and severe biliary AP were respectively induced by biliopancreatic duct ligation (DL) alone or along with infusing glyceryl trilinoleate (GTL; 50 μL/100 g) within 10 minutes of giving NIR 2-DG. Controls (CON) only received NIR 2-DG. Imaging was done every 5–10 minutes over 3 hrs. Average Radiant Efficiency [p/s/cm²/sr]/[μW/cm²] was measured over the pancreas using the IVIS 200 in-vivo imaging system (PerkinElmer) using the Living Image® software and verified in ex vivo pancreata. Blood amylase, lipase and pancreatic edema, necrosis were measured over the course of AP. Results NIR 2-DG uptake over the first hour was not influenced by AP induction. However, while the signal declined in controls and rats with mild AP, there was significantly higher retention of NIR 2-DG in the pancreas after 1 hour in those with GTL pancreatitis. The increase was > 3 fold over controls in the GTL group and was verified to be in the pancreas ex vivo. In vitro, pancreatic acini exposed to GTL had a similar increase in NIR 2-DG uptake which was followed by progressively worse acinar necrosis. Greater retention of NIR 2-DG in vivo was associated with worse pancreatic necrosis, reduced ATP concentrations and mortality, which were not predicted by the blood parameters. Conclusion In-vivo fluorescent imaging of a non-radioactive near infrared 2-DG optical probe can predict the AP severity early during the disease.

associated with worse pancreatic necrosis, reduced ATP concentrations and mortality, which were not predicted by the blood parameters.

Conclusion
In-vivo fluorescent imaging of a non-radioactive near infrared 2-DG optical probe can predict the AP severity early during the disease.

Background
While the utility of 2-deoxy glucose (2-DG) analogs is established in diagnosis and prognostication of malignant diseases, the relevance of such studies in inflammatory diseases is emerging [1,2,3,4,5]. Acute inflammatory diseases such as acute pancreatitis (AP) have a sudden onset with a rapid, variable and unpredictable course ranging from resolution with minimal care over a few days, to a severe course (severe acute pancreatitis; SAP) progressing to extensive pancreatic necrosis, requiring intensive care, a prolonged hospitalization with high costs and sometimes resulting in death. However we currently lack tools which can reliably predict the course of AP takes early on in the disease.
There are reports of increased 2-Deoxy-2-[18F] fluoroglucose (FDG) uptake in human pancreatitis detected by positron emission tomography (PET) imaging [6,7,8,9,10,11]. These are typically done with intent to study pancreatic cancer, and pancreatitis is an incidental finding contributing to the "false positives" seen. However, the relevance of the uptake of 2-DG tracers in predicting the AP severity early in the course of the disease is unknown and may need further exploration. Such evidence, if present in preclinical studies, can potentially guide studies in human pancreatitis.
The current project was designed to explore the behavior of, and study whether a 2-DG probe can be used to predict the severity of AP early in its course. For this we chose the nonradioactive, near-infrared probe (IRDye 1 800CW 2-DG; NIR 2-DG; Li-Cor) and studied its uptake into the pancreas during mild AP and SAP. NIR 2-DG has been previously used as a 2-DG mimetic for both in vivo and in vitro studies [12,13]. Compared to FDG it offers the additional advantage of not being limited by a short half-life, and since it is non-radioactive, it is amenable to high throughput imaging requiring less specialized instrumentation. The emission spectra of NIR probes range between 650-900 nm which allows for their in vivo use without interference by either absorbance or non-specific emission from neighboring tissue. The severe and mild AP models used in this study were chosen based on their relevance to human disease [14]. The mild model involves bilio-pancreatic duct ligation, which is sufficient to fulfil the criteria of biliary acute pancreatitis, but does not result in severe necrosis as shown recently [15], this is similar to mild biliary pancreatitis as may occur in humans due to a mass in the head of the pancreas causing duct obstruction. The severe model simulates severe biliary pancreatitis complicated by fat necrosis, as may occur in humans [16,17,18,19,20,21]. In this, the triglyceride precursor of the second most abundant unsaturated fatty acid in pancreatic fat (linoleic acid), i.e. glyceryl trilinoleate (GTL) is infused into the pancreatic duct prior to the ligation. The amount of GTL is equivalent or lower than the percentage of adipocytes in the obese human pancreas [14,22,23,24,25,26]. Using the above tools we set out to study whether the course of NIR 2-DG uptake in the pancreas in controls, mild and severe AP can be used to predict AP outcomes.

Materials and Methods Reagents
IRDye 800CW 2-DG (NIR 2-DG, Li-Cor Biosciences) was dissolved in 1 ml sterile saline at a final concentration of 100 nM. Glyceryl trilinoleate (GTL; TCI America) for in vitro studies was dissolved and sonicated in HEPES containing buffer; for in vivo studies was directly injected into the pancreatic duct of the rats. Ketamine Hydrochloride (Ketaset, Fort Dodge Animal health), Xylazine (Anased, Lloyd Laboratories) and Isoflurane USP (Piramal Healthcare).

Animal work
All experiments were approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh (Pittsburgh, PA), the Mayo Clinic (Scottsdale, AZ) and the Animal Care and Use Review Office (ACURO), a component of the USAMRMC Office of Research Protections. Mice: 6-8 week old male CD-1/ICR mice (Charles River Laboratories, Wilmington, MA) housed with a 12-hour light/dark cycle at temperatures ranging from 21-25°C, were fed standard laboratory chow and allowed to drink ad libitum. Rats: 250-350 g male Wistar rats (Charles Rivers Laboratories, Wilmington, MA) were housed with a 12 hour light-dark cycle, fed standard laboratory chow and were allowed to drink ad libitum. A Jugular venous catheter was placed in all rats under anesthesia (ketamine-Xylazine, 90 mg/kg, 8 mg/kg respectively, I.P.; or isoflurane, 3% + 0.8-1 L/min induction; 1.5% + 0.8-1 L/min maintenance). The rats were left to recover for at least 3 days prior experiments. Catheters were flushed with 4U/ mL heparin/saline solution every other day and the patency of catheter verified.

Acinar Harvest and 2-DG in vitro assay
For NIR 2-DG in vitro studies, pancreatic acini were harvested from mice and used in oxygenated HEPES buffer pH 7.4 as described previously [27,28,29]. Viability was confirmed by trypan blue exclusion (>95%) at the start of each experiment. All studies were done in a shaking water bath (80 RPM) at 37°C at an ambient atmosphere. Cells were treated with either glyceryl trilinoleate (GTL, 300 μM) or saline as a control. Cell injury was quantified by measuring LDH leakage (Roche Diagnostics, Indianapolis, IN) into the medium as described previously [22,30,31]. This was done at 0, 1, 2, 4 hours by taking a 2% (40 μl out of 2 ml) aliquot of medium, centrifuging it (200 g, 5 minutes) and measuring LDH in the supernatant. At the end of the incubation the cells were lyzed by incubating them for one hour in Triton x-100 at final concentration of 1%. LDH activity in this lysate was taken as representing total (100%). The activity in the aliquots calculated as a percentage of the total amount was depicted as % LDH leakage for each time point. For measurement of NIR 2-DG uptake, IRDye 800CW 2-DG (NIR 2-DG) was added at a final concentration of 1 μM to the medium containing the acini. These were then stimulated with GTL or saline. The reaction was terminated at 1 hour by putting cells on ice followed by 3 washes in ice cold HEPES buffer to remove unbound dye. The pellet was suspended in HEPES buffer and platted in triplicate in 96-well black plate. The plate was scanned at 800 nm for the targeted NIR 2-DG fluorescence signal (arbitrary units; AU) using the Odyssey Imaging System (Li-Cor) and analyzed using the Image Studio Software. Each condition in an experiment had a minimum of 3 replicates. Signals were averaged and the relative fluorescence was determined by normalizing to protein levels (mg/ml) per well measured using the BCA Protein Assay Kit (Thermo Fisher Scientific). Ratios were calculated by normalizing to the saline treated control group. 4 or more experiments were performed for each parameter.

In vivo 2-DG uptake and AP studies
In vivo 2-DG fluorescence imaging was performed with an IVIS Spectrum in vivo imaging system (PerkinElmer). All fluorescent images were acquired using identical illumination settings (EX/EM ICG filter, f/stop 2, field of views 22, binning 8), using 0.5 second of exposure time and analyzed using the Living Image 1 software. Fluorescence emission was normalized to photons per second per centimeter square per steradian and expressed as average radiant efficiency [p/s/cm²/sr] / [μW/cm²]. All NIR fluorescent images were displayed in the same scale of fluorescent intensity. At the day of experiments, rats under ketamine-Xylazine anesthesia had the fur overlying the abdomen and back completely shaved, the catheter patency checked and intravenous bolus injected with 100 μl of 2-DG (10 nM) followed by 100 μl of saline to wash the catheter. Preliminary studies were done to study early intensity changes. In the abdomen, the only changes noted were rapid accumulation in the bladder after 1 minute in the supine position and after 10 minutes as a diffuse uptake in the posterior superior abdomen in the right lateral position. After 10 minutes of 2-DG signal stabilization, severe AP was induced by intra pancreatic duct injection of GTL (50 μl/100 g of body weight) followed by ligating of the biliopancreatic duct just proximal to its entry into the duodenum (GTL group). Mild AP was induced by bilio-pancreatic duct ligation alone (DL group). Control groups just received 2-DG. For imaging, sedated rats were placed inside the 37°C warmed IVIS chamber and images were taken every 5-10 min after NIR 2-DG administration before AP induction and over the 3 hours following this. When rats were euthanized and pancreas harvested for ex vivo imaging or gauging the severity of pancreatitis. Animals were treated in a humane fashion with close attention to pain control. This was via keeping them anesthetized with ketamine-Xylazine, (90 mg/ kg, 8 mg/kg respectively, repeated as needed when animal could be aroused by toe pinch) over the period imaging studies were done and electively sacrificing them 1 or 3 hours post AP induction. The sedation prevented interference from movement artefact. For post-AP induction survival studies, animals were administered cefazolin (WG critical care, NJ, 100 mg/kg, intramuscular) for infection prevention, buprenorphine SR (ZooPharm, at 0.6 mg/Kg, S.C.) for pain management and normal saline (10 ml, subcutaneously) immediately after AP induction and allowed to recover. These animals were monitored continuously for signs of distress. Those noted to become moribund after AP induction (between 5 and 8 hours in GTL group, median 6 hours) were sacrificed with carbon dioxide anesthesia. Animals were sacrificed electively at 8 hours in the DL group. Serum, tissue samples were collected to study severity. Blood and pancreas tissue were harvested to analyze plasma lipase, amylase, alanine aminotransferase (ALT), bilirubin, and pancreas tissue edema and ATP levels and for pancreatic necrosis analysis as described previously [14,22,32]. Blood glucose levels were measured before and after 2-DG injection. There were 8-12 rats in each group for all quantified data.

Biochemical assays
LDH (Roche Diagnostics, Indianapolis, IN), plasma Amylase, Lipase, ALT (Pointe Scientific Inc) were measured following the manufacturer's instructions [14,22,32] on the FlexStation 3 Multi-Mode Microplate Reader (Molecular Devices) and total bilirubin (Pointe Scientific Inc) on the Eppendorf Biophotometer (Eppendorf). Blood glucose concentrations were measured by the glucose oxidase method using a FreeStyle glucose meter (Abbott Laboratories).

Pancreas water content
Pancreata was weighed on an analytic balance (wet weight) and dehydrated by heating at 37°C overnight (dry weight). Water content was calculated according to the formula: ½ðwet weight À dry weightÞ=wet weight x 100 and expressed in total weight percentage: Pancreatic tissue ATP level determination This was done as described previously [31,33]. A bioluminescent kit was used to measure ATP levels (Sigma-Aldrich, Saint Louis, MO) following manufacturer's instructions. Briefly, pancreatic tissue was disrupted in tri-chloroacetic acid and EDTA containing buffer followed by appropriate dilution in Tris-EDTA buffer and application of a luminescent substrate. Luminescence was measured on a Promega Glomax 20/20 Luminometer and was normalized per milligram of protein determined using the BCA Protein Assay Kit (Thermo Fisher Scientific).

Pancreatic necrosis
Whole pancreas paraffin section (5 micron) slides stained by hematoxylin & eosin were examined by a trained morphologist (CDO) blinded to the sample as described previously [14,22,31,32,33]. In brief, all pancreatic parenchymal area was imaged using the PathScan Enabler IV slide scanner (Meyer Instruments, Huston, TX) and images were evaluated for acinar necrosis as described previously [14,22,31,32,33]. Necrotic area and total acinar area were measured in pixels for each pancreas. Percentage necrosis was reported as a percentage of total area for each pancreas.

Statistical analysis
This was done using the SigmaStat statistical package integrated into the graphic program Sig-maPlot 11 (Systat Software Inc, San Jose, CA). All values, unless otherwise specified, are reported as mean ± SEM. Groups were compared by one way analysis of variance (ANOVA) versus controls. Pairs were compared using a two tailed Students t-test when the distribution was normal (Shaprio-Wilk test) or a Mann-Whitney test when failing the Shapiro-Wilk test. Differences were considered significant at a p value <0.05

NIR 2-DG uptake in the pancreas is not affected early in AP induction
We started with intent to understand the dynamics of NIR 2-DG uptake into the pancreas and whether this could reliably discriminate between mild and severe AP. Since rats are commonly used in surgical models of biliary pancreatitis, we first chose to learn if NIR 2-DG could be reliably visualized in an area overlying the pancreas in 250-350 g rats during the initial 3 hours over which AP progressively gets worse. The blood glucose in these rats after overnight fasting and prior to infusion of NIR 2-DG was 177±9.1 mg/dl and did not change over the duration of the study in controls (182±25 mg/dl). 3 groups were studied: controls, severe AP; i.e. rats with GTL infusion which develop severe pancreatic necrosis [14], and mild AP; with duct ligation alone who have a much milder course [15]. We chose a lateral position since the stomach overlies the pancreas and to avoid the confounding effect of a variable amount of food present in the stomach of rats despite fasting. A right side down position was chosen since the pancreas tail extends to the left side and thus would not be covered by the liver.
Prior to NIR 2-DG administration, there was no fluorescence detectable in any group of sedated rats (Fig 1A). NIR 2-DG administration without AP resulted in an initial generalized increase in fluorescence which was similar in all groups (Fig 1B and 1K). This synchronously increased over the pancreas (dashed polygon fig 1A-1I) and plateaued by 30-40 minutes to values that were no different in the control (black dots Fig 1J and 1K), GTL and DL groups (red and green dots). This increase remained similar in all three groups for the hour following AP induction (Fig 1K) for a total of 90 minutes after NIR 2-DG administration. These results show that manipulation of the pancreas or induction of AP does not affect the initial uptake or equilibration of the NIR 2-DG.

NIR 2-DG is retained in the pancreas during severe AP
We then analyzed the pattern of NIR 2-DG retention during the subsequent 2 hours. It was notable that during this time the control and DL groups progressively showed lesser retention of NIR 2-DG (Fig 1J). We then averaged the values for every 30 minute interval (i.e. time ± 15 minutes) and normalized it to the first 30 minute value when NIR 2-DG accumulation peaks or starts to plateau. As seen in Fig 2A the NIR 2-DG uptake in the GTL group remained elevated and equivalent to the peak 30 minute value for the subsequent 2 hours. This was significantly elevated over controls after 90 minutes (p<0.05, Ã , on ANOVA) and higher than the DL group at or after 120 minutes (P<0.03, #, Students t-test). To compare the fold increase in average radiance efficiency (ARE) in an AP group over the controls at any particular time point, the baseline ARE values i.e. those measured immediately after NIR-2DG administration (average 5 minutes) were subtracted from the ARE at that time point, and divided by the ARE change in controls. For e.g. the fold change in ARE at n minutes after NIR 2-DG in the GTL group was calculated as: fARE in GTL at n min: À ARE in GTL at 5 min:g Ä fARE in control at n min: À ARE in control at 5 min:g: As can be seen in Fig 2B, this progressively increased, peaking at 3-14 folds between 2-3 hours of AP induction, while the increase in the DL group was insignificant. To verify that the increase was indeed in the pancreas, we removed the pancreas at 3 hours and measured the ARE overlying it. As seen in Fig 2C and quantified in Fig 2D, there was a large increase in the NIR-2DG accumulation in the pancreas of the GTL group compared to the DL group.

Higher retention of NIR 2-DG may be helpful in predicting severe AP early on in the disease
We then compared the uptake of NIR 2-DG to parameters of AP commonly measured in blood. As can be seen in Fig 3A and 3B the parameters of AP induction i.e. plasma amylase and lipase were similar in the GTL and DL groups whether measured early or later in the disease course. Similarly the markers of a biliary etiology, i.e. plasma ALT and bilirubin were not higher in the in the GTL group. These findings correlate well with human data showing that criteria used to diagnose AP or its etiology have no bearing on its severity [34].
We then turned our attention to parameters of pancreatitis severity, focusing on local injury and mortality. Pancreatic necrosis in the GTL groups progressively increased from a median of 18% at 1 hour (Fig 4A-4C), when the necrosis in DL group was insignificant to a median of 54% (range 40-70%) at 8 hours with associated 100% mortality (Fig 4D). Duct ligation alone caused mild pancreatic necrosis at the time of sacrifice i.e. 8 hours (median 10%), at which time the pancreas had a yellowish hue, consistent with bile staining induced by the duct ligation. The worsening necrosis in the GTL group compared to the DL group correlates well with the fluorescence in the GTL group starting to increase between 60 and 90 minutes of AP induction (Fig 2A and 2B) and remaining increased over 3 hours. The in vivo findings of increased NIR 2-DG signal over the pancreas early in the course of AP were paralleled by increased NIR 2-DG binding to acinar cells within 1 hour of exposure to 300 μM GTL in vitro (Fig 5A and  5B). The fluorescence (arbitrary units/ well) /protein (mg/ml) was 54323 ± 18317 (AU/mg) in controls vs. 313265 ± 104180 AU/mg (p<0.01) in the GTL group. There was an insignificant increase in LDH leakage in the GTL group at this this point (4.6 ±1.7 vs. 1.8 ±0.4 in controls, p = 0.12), followed by a large increase in LDH leakage over 4 hours (Fig 5C). It is notable that the increase in NIR 2-DG uptake was associated with a trend of decreasing pancreatic ATP levels in the GTL group which became significant only in the group that was moribund prior to sacrifice (Fig 6E).

Discussion
In this study we find that severe biliary AP is associated with an early and sustained retention of the near-infrared 2-DG probe (IRDye 1 800CW 2-DG; NIR 2-DG; Li-Cor) in the pancreas in rats. This retention is associated with worse pancreatic necrosis and mortality. The study is unique for the following reasons: 1) It uses a rapidly evolving inflammatory disease model which is unlike the much longer duration cancer [12,13,35] or chronic inflammation [5,36] models commonly employing 2-DG probes. 2) it shows the feasibility of detecting NIR signals in large (250-350 g) rats in contrast to mice which are commonly used for such imaging studies [12,13,37]. 3) It shows that the increased retention of NIR 2-DG between 1 and 3 hours of AP induction has a stronger relation with severe outcomes than commonly used blood parameters used to determine AP severity in animal models.
Preliminary studies helped us decide the positioning of the rat and the timing of the imaging. While previous studies have shown a rapid first pass effect in mice with fluorescence overlying the kidneys to increase within 5 seconds of IRDye 1 800CW 2-DG administration [35], we did not note such an increase during the time fluorescence over the pancreas became intense (Fig 6A). We did note a rapid increase over the bladder to begin after the first minute and become intense by 10 minutes on imaging in the supine position. However we could not visualize the increase over the pancreas in this position, perhaps due to the stomach which overlies it. Placing the rat in a right lateral position allowed imaging the fluorescence over the pancreas (Fig 1A, dashed outline). The pancreatic signal was more cephalic and posterior than the bladder in the right lateral position. The right lateral position also resulted in the signal from the bladder to be largely masked by the pelvis and thigh, thus preventing it from interfering with the signal from the pancreas. In some cases we did note a distinct small signal from the bladder (seen in some images of Fig 1A-1I) which is anterior and inferior to the dashed outline corresponding to the signal from the pancreas. AP induction took 10-15 minutes and was executed within the 30 minutes that the signal peaked in controls (Figs 1K and 2A). This observation and the similarity in baseline glucose and other parameters between the 3 groups prior to AP induction prompted us to design the experiments to study 2 end points: 1) whether Severe AP is associated with a progressive increase in necrosis and mortality. A: time course of percentage of pancreatic parenchyma what was necrotic in the GTL (red) and DL (green) groups measured when electively sacrificed at 1, 3 hours and at euthanasia ( 8 hours). Values at 1 hour, 3 hours and at the time of euthanasia for each group were compared to each other and the controls by ANOVA. "*" depicts a significant (p<0.05) increase vs. control and "#" depicts significant differences in GTL vs. DL. Representative microscopic (5x, B) and gross (C) images are shown. D: Kaplan-Meyer curve showing % survival in the different study groups. There was 100% mortality in the GTL group by 8 hours, but none in the duct ligation group. There were 8-12 animals per group. Scale bars are 1mm in length. AP induction affects the increase in signal from NIR 2-DG over the pancreas. 2) Whether the pattern of NIR 2-DG signal over time can be used to predict the severity of AP.
As seen in Figs 1 and 2A, we note that AP induction does not interfere with the increase in ARE overlying the pancreas, which is similar in the controls and AP groups immediately after induction. After 30 minutes the trend changes with controls depicting a clear decline, whereas the GTL group shows a sustained increase over the course of the 3 hours these animals were imaged. The generalized increase over the posterior abdomen corresponding to where the kidney, pancreas, duodenum are located is at least partly contributed by the duodenum irrespective of AP as shown in the white outlines (Fig 6B and 6C). While bowel fluorescence could be  Parameters potentially affecting the NIR 2-DG signal measured in our system. A: Images of a rat administered NIR 2-DG collected on the IVIS imaging system in the supine position over the first 10 minutes (first five images) and in the right lateral position (last image). Note the increase over the bladder but the lack of signal over the pancreas in the supine position. This pattern is reversed in the right lateral position. Fluorescent (B) and black and white(C) images of the duodenum, pancreas and spleen removed en block after 2 hours of NIR 2-DG administration in controls or pancreatitis in the DL and GTL groups. Please note an increased uptake in the duodenum (white outlines) especially in controls. D: Data from groups of 8-12 animals showing the time course of the increase in pancreatic edema (measured as % water content or pancreatic ATP concentrations (E) in the DL and GTL groups. The "*" depict p<0.05 vs. corresponding controls "#" depicts significant differences in GTL vs. DL. doi:10.1371/journal.pone.0149073.g006 Near Infrared 2-Deoxyglucose Imaging in Acute Pancreatitis easily excluded ex-vivo after removing the pancreas (Fig 2C and 2D) this could not be done in vivo since the imaging was planar. This uptake by the bowel likely contributed to the in vivo background signal in controls, making the signal from the GTL group comparatively less impressive. It remains to be seen whether such interference can be reduced by cross sectional imaging.
Lipolysis of GTL by pancreatic lipases results in the release of the unsaturated fatty acid linoleic acid. Human pancreatic necrosis collections have high concentrations of linoleic acid [14,31,38], which can result in necrotic cell death [39,40] via inhibition of mitochondrial complexes I and V [22]. This would also be consistent with the decrease in ATP levels we note in the damaged pancreas (Fig 6E). The amounts of GTL used in this study are less than the adipocyte mass in the obese human pancreas [14,22,23,24,25,26]. Further proof validating the GTL model as being representative of human severe biliary AP is discussed in the manuscript in which this was originally described [14]. Additionally, GTL was preferred over the bile salt infusion model since the concentrations of bile salts used (20-100 mM) are >10 times their critical micellar concentrations and thus have a detergent effect [41,42] on cell membranes. Moreover these high bile acid concentrations are in stark contrast to those in human pancreatic necrosis collections which are < 200 μM (unpublished data, not shown).
There are several factors that could be associated with increased NIR 2-DG retention during the progression of SAP. These include increased vascular permeability [43]measured as an increase in percentage water content of the pancreas during pancreatitis (Fig 6D), hemorrhage into the pancreas resulting in persistent NIR 2-DG accumulation, reduced clearance of NIR 2-DG in the GTL group, and hypoxia resulting in increased glucose requirement [5]. We have previously shown that lipotoxicity from unsaturated fatty acids such as linoleic acid causes a reduction in ATP levels associated with acinar necrosis [22,23], which would be consistent with the large increase in necrosis (Fig 4A-4C) and drop in ATP levels we note in GTL group (Fig 6E). These in combination with hypoxia could increase glycolysis dependence of the GTL treated pancreas and thus result in the increased NIR 2-DG uptake we note. The contribution of vascular permeability to the increased in vivo signal noted in the GTL group while possible is unlikely, since the fluorescence increase during pancreatitis in vivo (Fig 2) is similar to that in acini exposed to GTL in vitro (Fig 5A and 5B). Similarly, the increase in NIR 2-DG signal noted ex-vivo in the GTL group did not correlate with the amount or location of hemorrhages (Figs 2C, 6B and 6C), and while hemorrhages were sometimes noted within an hour of GTL injection (Fig 4C), these did not result in an increased of NIR 2-DG signal at 1 hour (Fig 1J and  1K). These observations along with the in vitro findings in acini mentioned above make the contribution of hemorrhage to the in vivo NIR 2-DG signal also unlikely. The possibility of variable clearance in affecting the NIR 2-DG signal is also unlikely since all animals were imaged at 5-10 minute intervals after its administration, and the signal increase was similar in all groups up to 90 minutes after administration. A selective reduction in clearance of NIR 2-DG the GTL group should have resulted in a generalized increase in NIR signal, but the increase noted ex-vivo was only over the pancreas and not the duodenum or spleen (Figs 2C, 6B and 6C). We have not looked at the contribution of increased expression of glucose transporters, hexokinase or the direct effect of inflammatory mediators to the increased NIR 2-DG retention noted in the GTL group. However the in vitro data in isolated acinar cells showing increased NIR 2-DG retention within an hour of GTL exposure, in the absence of exogenous inflammatory mediators (Fig 5A and 5B) makes the contribution of these to the in vivo signal unlikely.
These studies make a case that measuring the increase in metabolic demand of the pancreas helps predict the subsequent severity of local injury. This is important since most clinical cases of AP are mild, and currently there are no reliable early predictors to distinguish these mild ones from the ones that subsequently develop severe necrosis. In the clinical setting necrosis typically develops after the first several days of AP, and an early predictor of necrosis may allow early changes in management and help focus therapeutic approaches to patients with predicted severe AP.
In summary, the current studies show increased and persistent accumulation of the nearinfrared 2-DG probe, IRDye 1 800CW 2-DG, over the pancreas during the initial hours of AP to be a more reliable predictor of severity than conventional serum markers of AP. This binding of the 2-DG probe precedes the drop in ATP levels and necrosis that eventually results in severe AP. Based on the utility demonstrated in this rapidly evolving inflammatory disease model in rats, this study opens the scope of imaging 2-DG analogs as early predictors of inflammatory disease severity in humans.