Dynamic Imaging of Pancreatic Nuclear Factor κB (NF-κB) Activation in Live Mice Using Adeno-associated Virus (AAV) Infusion and Bioluminescence*

Background: NF-κB is an important signaling molecule in the development of acute pancreatitis. Results: Adeno-associated virus-NF-κB-luciferase-infused mice showed a 77- and 140-fold increase in pancreas-specific NF-κB bioluminescence following caerulein and caerulein + LPS pancreatitis, respectively. Conclusion: NF-κB activation can be examined in a live, dynamic fashion during pancreatic inflammation. Significance: This technique offers a valuable tool to study real-time activation of NF-κB in vivo. Nuclear factor κB (NF-κB) is an important signaling molecule that plays a critical role in the development of acute pancreatitis. Current methods for examining NF-κB activation involve infection of an adenoviral NF-κB-luciferase reporter into cell lines or electrophoretic mobility shift assay of lysate. The use of adeno-associated viruses (AAVs) has proven to be an effective method of transfecting whole organs in live animals. We examined whether intrapancreatic duct infusion of AAV containing an NF-κB-luciferase reporter (AAV-NF-κB-luciferase) can reliably measure pancreatic NF-κB activation. We confirmed the infectivity of the AAV-NF-κB-luciferase reporter in HEK293 cells using a traditional luciferase readout. Mice were infused with AAV-NF-κB-luciferase 5 weeks before induction of pancreatitis (caerulein, 50 μg/kg). Unlike transgenic mice that globally express NF-κB-luciferase, AAV-infused mice showed a 15-fold increase in pancreas-specific NF-κB bioluminescence following 12 h of caerulein compared with baseline luminescence (p < 0.05). The specificity of the NF-κB-luciferase signal to the pancreas was confirmed by isolating the pancreas and adjacent organs and observing a predominant bioluminescent signal in the pancreas compared with liver, spleen, and stomach. A complementary mouse model of post-ERCP-pancreatitis also induced pancreatic NF-κB signals. Taken together these data provide the first demonstration that NF-κB activation can be examined in a live, dynamic fashion during pancreatic inflammation. We believe this technique offers a valuable tool to study real-time activation of NF-κB in vivo.

pancreatic inflammation, cancer, and regeneration. NF-B is a heterodimeric complex composed of members of the Rel protein family (1)(2)(3). The most common dimeric form is the p50-p65 heterodimer. Prior to activation, NF-B proteins are predominantly restricted to the cytosol by associating with members of the inhibitor of B family (IB).
There are two distinct pathways that result in NF-B activation (4). The canonical pathway for NF-B activation is induced by most physiological NF-B stimuli. Phosphorylation of IB by IB kinases (IKK) triggers the degradation of IB and allows translocation of NF-B to the nucleus. Upon translocation, NF-B mediates the expression of numerous inflammatory response genes including pro-inflammatory cytokines, chemokines, adhesion molecules, and inducible enzymes such as cyclooxygenase-2 and inducible nitric-oxide synthase (5)(6)(7).
The non-canonical pathway for NF-B activation is induced by certain tumor necrosis factor (TNF) family cytokines, such as CD40L, BAFF, and lymphotoxin-␤ (4,8). Phosphorylation of p100 by IKK␣ leads to partial processing of p100 and the generation of transcriptionally active p52-RelB complexes. IKK␣ activation and phosphorylation of p100 depends on NIK, which is subject to complex regulation by TRAF3, TRAF2, and additional ubiquitin ligases (9 -11).
NF-B activity in pancreatic acinar cells plays an important, initiating role in the inflammatory response during acute pancreatitis (12) and, in this context, induces the expression of interleukin-1 (IL-1), IL-6, IL-8, tumor necrosis factor-␣ (TNF-␣), and platelet-activating factor. For this reason, it is of keen interest to monitor pancreatic NF-B activation in a dynamic fashion.
Traditional methods for examining NF-B activity in vitro include protein determination of NF-B pathway markers (i.e. phosphorylated IB; p65 nuclear translocation; IKK up-regulation), electromobility shift assay (EMSA), and immunohistochemistry for phosphorylated p65 (13,14). Newer techniques include the transfection (or infection via viruses) of NF-Bluciferase reporters. With these techniques, binding of NF-B subunits to a nuclear response element drives transcription of the luminescent protein luciferase. The commonly used luciferase reporters are firefly (15) and Renilla luciferases (16). The development of secreted luciferases such as Gaussia (Gluc), secreted alkaline phosphatase, and Cypridina allows for serial determination from the media of NF-B activity from the same population of cells (17)(18)(19)(20).
Bioluminescence (i.e. luciferase-based) imaging in vivo has emerged as a powerful tool in biomedical research for monitoring transgene expression, viral vector infection, tumor growth, and metastasis, as well as inflammation and gene therapy (21). Bioluminescence imaging is highly sensitive, cost-effective, noninvasive, and it facilitates real-time analysis in vivo. For this reason, luciferase-based reporters provide a major advantage over less sensitive reporters such as LacZ (22,23) and eGFP (24).
The first globally expressed transgenic NF-B reporter mouse line used a firefly luciferase, and the stimuli TNF-␣, IL-1␣, or LPS increased luminescence in most tissue, with the strongest activity observed in skin, lungs, spleen, Peyer patches, and the wall of the small intestine (25). In the current study, we first demonstrated that this mouse line poorly expressed NF-B-luciferase in the pancreas and failed to induce robust pancreatic luminescence signals above background noise, even with a pancreatitis stimulus that is known to cause pancreatic NF-B activation. Instead, novel gene delivery of NF-B-luciferase through intrapancreatic duct infusion of an adeno-associated virus (AAV6-NF-B-luciferase) led to robust pancreas-specific NF-B signals. We propose that this improved method can be used to measure pancreatic NF-B activity in a dynamic fashion and can be adapted to examine NF-B in other organ systems.

EXPERIMENTAL PROCEDURES
Reagents and Animals-All reagents were purchased from Sigma unless otherwise stated. NF-B-luciferase transgenic reporter mice were from Taconic Farms (strain NF-B-RE-luc; Taconic Farms, Hudson, NY). These mice were generated by microinjecting a transgene containing 6 NF-B response elements, a CMV␣ promoter, and a basal SV40 promoter, which are upstream to a modified firefly luciferase called Luciferase 2P (Promega, pGL3, Fig. 1A). The transgene was microinjected into BALB/cJ zygotes. The resultant mice were bred with a BALB/cJ strain.
Female Swiss Webster mice weighing 20 -25 g (Charles River, Wilmington, MA) were used for the AAV6-NF-B-luciferase infusion. They were fed standard laboratory chow, given free access to water, and randomly assigned to control or experimental groups.
Design and Purification of the AAV6-NF-B-luciferase Reporter Construct-The AAV6-NF-B-luciferase plasmid was generated by cloning a pGL4.32[luc2P/NF-B-RE/Hygro] vector (Promega number E8491) into a pAAV-MCS plasmid (Cell Biolab number VPK-410; Fig. 1B). The pGL4.32[luc2P/NF-B-RE/Hygro] vector contains five copies of an NF-B response element that drives transcription of the luciferase reporter gene luc2P (Photinus pyrlais). As indicated on the Promega website, Luc2P is a synthetically derived luciferase sequence with humanized codon optimization that is designed for high expression and reduced anomalous transcription. Once cloned, the pAAV6-NF-B-luciferase plasmid was cotransfected into HEK293 cells along with two helper plasmids: 1) pAAV-RC, which is used as a packaging plasmid carrying the serotype 6 rep and cap genes and 2) pHelper, a helper plasmid carrying the adenovirus helper functions (Fig. 1C). Cells were collected after 72 h and suspended in lysis buffer.
To purify the virus, cells were freeze/thawed three times to release the AAV6-NF-B-luciferase. Cell lysates were treated with benzonase (0.05 units) at 37°C for 30 min followed by 10% sodium deoxycholate at 37°C for 30 min. To clean cell debris, lysates were spun at 2500 ϫ g for 10 min. AAV6-NF-B-luciferase was precipitated using a 1:4 mixture of 40% polyethylene glycol (PEG-800) and 2.5 M sodium chloride for 2 h at 0°C. The solution was spun at 2500 ϫ g for 30 min to collect the PEG precipitate. The PEG pellet was resuspended in HEPES buffer (50 mM), treated with an equal volume of chloroform (100%), spun at 2500 ϫ g, and air dried for 30 min. Two phase partitioning was performed using 50% ammonium sulfate and 40% PEG, and the solution was spun at 2500 ϫ g for 15 min. The ammonium sulfate phase was collected and dialyzed using a Slide-A-Lyser Dialysis Cassette (10K MWC0; Thermo Scientific, Rockford, IL) for 4 h. Dialysis was repeated a second time for 16 h. The AAV6-NF-B-luciferase was concentrated using a concentrator filter tube (Millipore number UFC905024) and stored at Ϫ80°C. Viral concentrations were quantified using the Quick-Titer AAV Quantitation Kit (Cell Biolabs, San Diego, CA). To verify that the construct was functional, HEK293 cells were infected with the AAV6-NF-B-luciferase along with a plasmid expressing the constitutively active p65 NF-B subunit (Fig.  1D).
Intrapancreatic Duct Infusion of AAV6 -The procedure for retrograde infusion into the common bile duct (CBD) and pancreatic duct has been recently described (26 -28). Briefly, Swiss Webster mice were anesthetized with isoflurane. A midline incision was made to reveal the abdominal cavity. The duodenum was flipped to reveal its distal side and held in place by ligatures. A 30-gauge needle was inserted through the antimesenteric aspect of the duodenum to cannulate the CBD. A small bulldog clamp was applied to the distal CBD (near the duodenum) to prevent backflow of the infusate into the duodenal lumen and to hold the cannula in place. A larger bulldog clamp was applied to the proximal CBD (near the liver) to prevent infusion into the liver and thus to direct flow to the pancreatic duct. One hundred microliters of AAV6-NF-B-luciferase (titer 2.31 ϫ 10 12 pfu) were infused at 10 l per min for 10 min using a P33 perfusion pump (Harvard Apparatus, Holliston, MA). Upon completion of the infusion, the bulldog clamps were released. The exterior abdominal wound was closed using 7-mm wound clips, and a single injection of buprenorphine (0.075 mg/kg) was given immediately after the surgery. Mice recovered on a heating pad for 30 min after the procedure. They were given free access to food and water after the surgery.
Bioluminescence Imaging and Quantification-Bioluminescence imaging was performed using an IVIS imaging system (Xenogen, Alameda, CA). Mice were anesthetized with 1-3% isoflurane prior to a subcutaneous injection with D-luciferin (150 mg/kg). After 7 min, the mice were imaged. The regions of interest from displayed images were quantified using the LivingImage software 4.2 (Xenogen, Alameda, CA), represented as average radiance (photons/s/cm 2 /steradian). In vivo abdominal signals were normalized to a baseline value for each individual mouse. Ex vivo organ signal measurements were not normalized. In these particular experiments, mice had to have similar baseline luminescence values (i.e. before administration of stimuli).
Experimental Pancreatitis Models-Pancreatitis was induced in mice by administering hourly subcutaneous injections of caerulein (50 g/kg body weight) for up to 12 h (29). In additional experiments, a more severe model of pancreatitis, and one that augmented non-pancreatic sources of NF-B, was induced by administering 6 hourly caerulein injections followed by a subcutaneous injection of lipopolysaccharide (LPS; 10 mg/kg) as modified from Ding et al. (30). A second distinct and clinically relevant model of pancreatitis, mimicking post-ERCP (endoscopic retrograde cholangiopancreatography) pancreatitis, was induced by intraductal infusion of the radiocontrast iohexol (Omnipaque-300; 100 l total per mouse; GE Healthcare) at 20 l/min for 5 min. Animals receiving intraductal infusion of normal saline served as sham controls.
Statistical Analysis-Data were expressed as mean Ϯ S.E. unless otherwise stated. Statistical analysis was performed using a Student's t test. Statistical significance was defined as a p value Յ 0.05. NF-B-luciferase was measured as average radiance (photons/s/cm 2 /steradian). D, to validate that the AAV6 construct was functional, HEK293 cells were co-infected with AAV6-NF-B-luciferase and a constitutively active form of the NF-B subunit p65 (pCI-p65) or an empty vector (pCI alone). E, schematic of the surgical procedure in which the distal common bile duct (arrows) was cannulated via a trans-duodenal puncture, and methylene blue was added to the infusate containing the AAV6-NF-B-luciferase construct.

RESULTS
Globally Expressing NF-B-luciferase Transgenic Mice-The NF-B-luciferase reporter mouse is a frequently employed tool to study activation of the transcription factor NF-B in vivo (21,25,31,32). In this transgenic mouse strain, the luciferase gene is placed downstream of the NF-B response elements and a minimal CMV and SV40 promoter (Fig. 1A). Previous studies in isolated pancreatic acinar cells, as well as from in vivo pancreatic tissue, demonstrate that NF-B activation occurs early in the course of pancreatic injury (5, 33, 34). To specifically exam-   /kg). E, the pancreas, liver, spleen, kidney, and heart/lung were removed and imaged for NF-B bioluminescence (left). Quantification of organ bioluminescence is shown on the right. MAY 1, 2015 • VOLUME 290 • NUMBER 18

JOURNAL OF BIOLOGICAL CHEMISTRY 11313
ine pancreatic NF-B activation in the whole mouse, we defined 4 regions of interest that outlined the (1) neck, (2) thorax, (3) upper abdomen, and (4) lower abdomen ( Fig. 2A). The luminescent intensity values for the neck and thorax were 4and 2.5-fold higher, respectively, than the collective average values of the upper and lower abdomen. We also observed some level of diurnal variation at baseline (Fig. 2C) and found that there was an artificial signal at the site of intraperitoneal injections of luciferin (Fig. 2D). For this reason, we subsequently only gave subcutaneous injections.

NF-B-luciferase Transgenic Mice Demonstrate Mild Elevations in Upper Abdominal NF-B Signal following Caerulein
Pancreatitis-We next determined whether the transgenic NF-B-luciferase mice could elicit a pancreatic signal in response to pancreatitis. A highly reproducible model of experimental pancreatitis in rodents is induced by giving hourly injections of high doses of the cholecystokinin analog caerulein (50 g/kg; subcutaneously) for 12 h (Fig. 3A) (35,36). Quantification of the NF-B luminescent signal from the upper abdomen revealed a 2.2-and 1.8-fold increase with caerulein pancreatitis above normal saline-injected controls at 9 and 12 h, respectively (Fig. 3, B and C). Ex vivo imaging of the organs revealed that caerulein induced a higher signal in the pancreas compared with the other organs of the abdomen and thorax (Fig. 3E). However, the signal intensities in the pancreas were focally enhanced and were reminiscent of intrapancreatic lymph nodes. To test a more potent inducer of NF-B activation, we administered the endotoxin lipopolysaccharide (10 mg/kg) 1 h after the sixth caerulein injection. LPS caused a global increase in NF-B activation with large elevations in the heart, lungs, and spleen and only mild elevations in the pan-creas (Fig. 3, D and E). Taken together, these results indicate that NF-B transgenic mice poorly express NF-B-luciferase in the pancreas and fail to demonstrate a pancreatic signal above background noise with a pancreatitis stimulus that is known to cause NF-B activation.
Mice Infected with Intraductal AAV6-NF-B-luciferase-To address the problem of specificity of signal to the pancreas we designed an adeno-associated virus serotype 6 (AAV6) carrying an NF-B-luciferase reporter as described under "Experimental Procedures" (Fig. 1, B and C). Once purified, we infused the AAV6-NF-B-luciferase (titer 10 12 pfu) into the pancreatic duct of wild type mice and tracked the emergence of bioluminescence as the mice healed from the surgery (Fig. 4). The signals from the upper abdomen peaked 7-10 days after the infusion and reached a new baseline after 3 weeks. What was particularly noticeable was that the bioluminescence was restricted primarily to the upper abdomen (Fig. 5A), and it took on the shape of the pancreas. In contrast to the globally expressing NF-B-luciferase transgenic mice, the AAV6-NF-B-luciferase intraductally infused mice had less diurnal variation (Fig.  5B). However, there was still considerable inter-animal variability at baseline.
Pancreatitis Models Induce a Pancreas-specific NF-B Signal in Vivo Using Mice That Underwent Intraductal Infusion of AAV6-NF-B-luciferase-We next determined using caerulein hyperstimulation whether intraductally infused AAV6-NF-Bluciferase mice could manifest a pancreas-specific signal in response to pancreatitis (Fig. 6, A and B). Quantification of the NF-B luminescent signal from the upper abdomen revealed that caerulein-treated mice had a 15-fold peak in NF-B luciferase 7 h after the first caerulein injection compared with nor- mal saline-treated controls (p Ͻ 0.05; Fig. 6C). Comparing this time course with pancreatic histologic severity, the results confirm previously published findings (33,34) that pancreatic NF-B is an early signal for pancreatic inflammation. In a separate batch of experiments with caerulein, the intra-abdominal organs were removed 8 h after the first hourly caerulein injection. The signal in the pancreas of the caerulein-induced mouse was 77.5-fold higher than in the pancreas of a normal salineinjected control mouse (Fig. 6, D and E). The spleen and stomach had no increased intensity, but there was a 12-fold increase in signal in the liver. A combination of 6 hourly caerulein injections followed by 1 injection of LPS, to provoke extra-pancreatic NF-B, led to a more intense upper abdominal signal (Fig.  6C) and ex vivo the pancreas-specific signal was 9-fold higher than with caerulein alone. However, in this case, the liver contributed the greatest intensity, which attests to the finding that there was likely some leakage of AAV6 into the hepatic ducts during the intraductal infusion. Nonetheless, the results indicate that with a pancreatitis-specific stimulus (i.e. caerulein alone), AAV gene delivery of NF-B-luciferase through intrapancreatic duct infusion leads to robust pancreatic NF-B bioluminescent signals in vivo.
Isolated acinar cells from AAV6-NF-B-luciferase-infused mice had high baseline levels of NF-B activity (data not shown). This is most likely the result of cellular stress caused by the isolation and dispersion of acini. For this reason, it was not possible to examine NF-B activity in vitro using the in vivoinfused luciferase.
We next examined a novel model of pancreatic injury that mimics the clinical scenario of post-endoscopic retrograde cholangiopancreatography pancreatitis (ERCP). An ERCP is a common gastrointestinal procedure in which the bilio-pancreatic ducts are cannulated (through a side port on an endoscope) and a small amount of radiocontrast is instilled to radiograph- ically visualize the ducts (37,38). The most common complication of ERCP is pancreatitis, occurring in 4 -7% of patients (39,40). To recapitulate this scenario in mice, we performed an intraductal infusion of the radiocontrast iohexol (Omnipaque-300) in mice that had received intraductal AAV6-NF-B-lucif-erase 5 weeks prior (Fig. 7A). Radiocontrast infusion led to a 13-fold increase in pancreatic NF-B luciferase signals 4 h after surgery above the signals observed in a sham-operated mouse that received intraductal normal saline (p Ͻ 0.05; Fig. 7, B and C). These data provide complementary evidence that AAV6-   MAY 1, 2015 • VOLUME 290 • NUMBER 18

JOURNAL OF BIOLOGICAL CHEMISTRY 11317
NF-B-luciferase infusion is a sensitive and specific method for detecting pancreatic NF-B activity.

DISCUSSION
In this study we provide the first demonstration of live, dynamic NF-B signaling within the pancreas and successful pancreatic gene delivery of a bioluminescent reporter through AAVs. We validated the technique using two complementary pancreatic injury models. We believe that this tool will be particularly useful to examine a time course of NF-B in the same mouse, thus minimizing the number of mice necessary for each experiment.
A growing body of evidence suggests that the NF-B-dependent pathway is critical in the development and persistence of pancreatic inflammation and injury (36,(41)(42)(43)(44). Thus, NF-B is an attractive target for pharmacological intervention and being able to monitor the NF-B-dependent pathway under various pathological states would facilitate the validation of novel therapeutic agents and approaches.
Previous efforts to examine live, dynamic pancreatic NF-B have been challenging. Using globally expressing NF-B-luciferase mice, Gray et al. (45) found that mice fed a choline-deficient, ethionine-supplemented diet to induce pancreatitis had a prominent bioluminescent signal in the thorax and a broad nonspecific signal in the upper abdominal region. Specific organs emitting these signals could not, however, be well delineated. Subsequent studies in these NF-B-luciferase reporter mice showed that administration of TNF-␣, IL-1␣, or LPS increased luminescence in a tissue-specific manner, with the strongest activity observed in skin, lungs, spleen, Peyer patches, and the wall of the small intestine (25). We found, however, that this transgenic mouse poorly expressed luciferase in the pancreas and failed to demonstrate a pancreatic signal above background noise with caerulein hyperstimulation pancreatitis.
Instead, we were able to detect specific pancreatic NF-B activation dynamically in vivo by gene delivery of the NF-Bluciferase reporter gene through intraductal infusion of AAV6. AAV6 and AAV8 appear to have the highest infection efficiency in the pancreas compared with other AAV serotypes and transduce pancreatic acinar cells, islets, and ducts (27,46). In previous work, intrapancreatic ductal infusion of AAVs successfully yielded pancreas-specific infectivity (27,46,47). This route of delivery is preferred over hydrodynamic injection into the systemic circulation, because it is targeted to the pancreas. Unlike adenoviral vectors, AAVs offer the major advantage of evading an immunogenic response (48).
Bioluminescence imaging in vivo is a highly sensitive method for monitoring gene expression in luciferase reporter transgenic mice (49,50). The noninvasive nature of this technology also enables convenient longitudinal studies and the ability to perform a paired analysis within the same mouse (51)(52)(53). Transient transfection of a plasmid with a luciferase reporter has been used to monitor NF-B activation in the liver (54). However, reporter expression lasted for only a few days and was thus unsuitable for monitoring NF-B activity during chronic pathological conditions. AAVs, on the other hand, integrate within the genome and should provide a long duration of effect.
In the current study, we demonstrate stable expression of an AAV6-NF-B-luciferase reporter. A drawback of the technique, however, is that there is a smaller extra-pancreatic signal in the liver, even with caerulein alone, a stimulus that primarily evokes pancreatic inflammation. The hepatic signal was further pronounced when NF-B was globally induced using LPS. It is also important to note that the intraductal infusion of AAVs causes incorporation of the reporter only within the native pancreatic cells. However, the signal does not account for the contribution of NF-B, for example, from infiltrating immune cells during inflammation.
We used two relevant pancreatic inflammation models to induce pancreatic NF-B activation. In the first model, caerulein-treated mice manifested an increase in signal after the 4-h time point, which is similar to what has been previously published (33,34).
In this report, we also provide the first demonstration that pancreatic NF-B is activated in an experimental model of post-ERCP pancreatitis. The results support the claim that NF-B is crucial to a range of pancreatitis etiologies. In summary, we report here a novel technique of intraductal infusion of an AAV6-NF-B-luciferase to examine pancreatic NF-B activation in a live, dynamic fashion.