|Home | About | Journals | Submit | Contact Us | Français|
Acute pancreatitis (AP) can result in pancreatic necrosis and inflammation, with subsequent multi-organ failure. AP is associated with increased neutrophil recruitment and rise in pro-inflammatory cytokines such as TNFα. Pretreatment with hemin, results in recruitment of hemeoxygenase-1 (HO-1)+ macrophages and protects from experimental pancreatitis. It is not clear whether modulation of HO-1 after onset of disease has a protective role. In this study, we tested the utility of Panhematin, a water-soluble hemin formulation, in activating and inducing pancreatic HO-1, and as a therapeutic agent in treating mouse acute pancreatitis.
We defined the distribution of radiolabeled hemin, then used in-vivo HO-1-luciferase bioluminescence imaging and CO2-release-assay to test Panhematin-induced upregulation of HO-1 transcription and activity, respectively. Using two well-defined AP murine models, we tested the therapeutic benefit of Panhematin, and quantified cytokine release using a luminex assay.
Intravenously-administered Panhematin induces rapid recruitment of HO-1+ cells to the pancreas within 2h and de novo splenic HO-1 transcription by 12h. Despite high baseline spleen HO-1 activity, the pancreas is particularly responsive to Panhematin-mediated HO-1 induction. Panhematin-treated mice, at various time points after AP induction had significant reduction in mortality, pancreatic injury, together with up-regulation of HO-1 and down-regulation of pro-inflammatory cytokines and CXCL1, a potent neutrophil chemoattractant.
Despite AP-associated mortality and morbidity, no effective treatment other than supportive care exists. We demonstrate that Panhematin leads to: i) rapid induction and activation of pancreatic HO-1 with recruitment of HO-1+ cells to the pancreas, ii) amelioration of AP even when given late during the course of disease, and iii) a decrease in leukocyte infiltration and pro-inflammatory cytokines including CXCL1. The utility of Panhematin at modest doses as a therapeutic in experimental pancreatitis, coupled with its current use and safety in humans, raises the potential of its applicability to human pancreatitis.
Acute pancreatitis (AP) accounts for over 220,000 hospital admission every year in the United States alone, with approximately 20% of these patients suffering a severe disease course and can be associated with significant morbidity and mortality. [1, 2] AP is thought to develop as a result of an injury to the pancreatic acini, leakage and activation of pancreatic enzymes within the tissue, and the subsequent initiation of autodigestion and pancreatic injury. [3–5] The most feared complication of AP and the local inflammatory response are systemic and multiple organ failure, which are associated with high mortality.[2, 6] Despite the burden of disease, aside from supportive management, no effective therapy exists for treating AP.  Thus, a search for new treatment modalities for AP remains an important goal in the field.
Hemeoxygenase-1 (HO-1) is a stress inducible enzyme, while its related isoform HO-2 is constitutively expressed.  HOs are crucial rate-limiting enzymes that cleave heme into its iron, carbon monoxide (CO), and biliverdin. [9, 10] We previously showed that pre-treatment with hemin, a hemoglobin prosthetic moiety that upregulates HO-1, protects mice from acute experimentally induced-pancreatitis using two independent pancreatic injury models.  The protective role of hemin is mediated via HO-1+ F4/80+ cells that are recruited to the pancreas.  Given the utility of hemin as a prophylactic agent for pancreatitis development, we hypothesized that hemin may offer a therapeutic benefit in experimental pancreatitis. Hemin is soluble in alkaline pH, and therefore would have to be given intraperitoneally (i.p.) in mice already experiencing significant peritoneal inflammation. Given the limitation of hemin administration via the i.p. route, the use of Panhematin (PH), a water soluble intravenous (i.v.) formulation of hemin, is a clinically attractive possibility, although protocols for reconstitution of hematin for i.v. administration are available.  Notably, PH is an FDA approved drug for the treatment of acute intermittent porphyria. [13, 14]
In the study herein, we show that PH is an effective therapeutic in two experimental models of acute pancreatitis. There is early upregulation of HO-1 and recruitment of HO-1+ cells to the pancreas following PH administration. Although hemin administration does not result in its preferential accumulation in the pancreas, there is an increase in HO-1 protein expression and significant upregulation in HO-1 activity in the pancreas after PH therapy. Consistent with the anti-inflammatory effects of PH, there is a decrease in pancreatic and serum TNFα, IL-12p40, and IL-6. Neutrophils and their recruitment to the inflamed pancreas have been shown to play a central role in the pathogenesis of AP.  We show here that PH treatment leads to a decrease in pancreatic leukocyte infiltration and CXCL1, a potent neutrophil chemoattractant, thereby providing a potential therapeutic mechanism for PH against pancreatitis.
FVB/n were purchased from Taconic. Young female mice (15–19 g) were fasted for 12 h and then fed a choline-deficient diet (CDD; Harlan Teklad) supplemented with 0.5% DL-ethionine (Sigma-Aldrich) or normal chow (control group) for 3 days. In addition, an L-arginine model of pancreatitis was carried out using two intraperitoneal doses of L-arginine (given at time 0 and 1 h) in C56BL/6 mice (Taconic) as described previously. Mice received PH (Ovation Pharmaceuticals) dissolved in sterile water (reported as equivalent to hemin dose, as described on the drug nomogram, per mouse weight) i.v. at various times (Supplementary Table 1). PH was given 24 h following the second dose of L-arginine. Six to eight week-old HO-1-luc transgenic mice, whose transgene contains the full-length (15-kb) HO-1 promoter fused to the reporter gene luciferase (luc) were used for in vivo bioluminescence imaging of HO-1 transcriptional activity.  All animal protocols were approved by the Institutional Animal Care and Use Committees.
Mouse pancreata were isolated and frozen immediately in liquid nitrogen. Frozen pancreata were homogenized using a Dounce and 3% SDS/5 mM EDTA/0.187 M Tris-HCl (pH 6.8). Protein concentrations were determined by the Bicinchoninic acid (BCA) Protein Assay. Equal amounts of protein were separated on SDS-PAGE followed by electrophoretic transfer onto Polyvinylidene fluoride (PVDF) membranes. Anti-HO-1 and anti-HO-2 polyclonal antibodies used were purchased from Assay Designs, Inc.
Mouse pancreata were frozen in Tissue-Tek OCT compound and sectioned using a cryostat. Slides were acetone-fixed and incubated with antibodies to HO-1 (Assay Designs) to visualize HO-1+ cells. To assess pancreatic injury, mouse pancreata were fixed in 10% neutral buffered formalin (pH 6.8–7.2). Pancreata were embedded in paraffin, sectioned, and stained with hematoxylin and eosin. The histologic score and grading were based on edema, inflammation, hemorrhage, and parenchymal necrosis as described previously.  To assess leukocyte infiltration, paraffin embedded sections were stained with biotin-labeled anti-mouse CD45 antibody (BD Biosciences) followed by streptavidin-HRP conjugate and DAB chromogen solution (R&D systems).
Blood was collected from mice by cardiac puncture, and serum was isolated from these samples for subsequent lipase and blood urea nitrogen (BUN) level determination using Beckton Dickinson Microtainer Serum Separator Tubes. Pancreatic trypsin activity was determined as described previously. .
Blood, pancreatic and liver tissues were collected under sterile techniques and submitted for bacterial growth analysis to institutional microbiology lab. Pancreas and liver were first homogenized in sterile saline.
The indicated amounts of 55Fe-hemin (RI consultants) were injected i.v. into mice. Mice were sacrificed at various times, and brain, heart, kidney, pancreas, small intestines, stomach, colon, liver, and spleen were isolated, weighed, and homogenized as described for immunoblotting. Scintillation counts of homogenized tissue were obtained using a Beckman Coulter.
Whole mice were embedded in Tissue-Tek OCT compound and sectioned using a Leica CM3600 Macrotome. Sections were placed on Kodak Biomax XAR film and exposed for 1 month in order to visualize 55Fe-hemin tissue distribution.
HO-1-luc transgenic mice with an FVB/n background received 1 dose of PH on day 0. [11, 17] Whole body imaging of luc expression (hence HO-1 promoter activity) of the PH- or VE-treated HO-1-luc mice were performed at 0, 2, 4, 6, 8, 12, 18, 24 and 30 h post-PH treatment. Luciferin (100 μL, 30 mg/mL) per 20 g of mouse weight (final 150 mg/kg body weight) was administered i.p. to each mouse 10 min prior to imaging. Mice were then placed in an In Vivo Imaging System (IVIS) and the photons/sec emitted from the tissues were quantified using LivingImage software v3.2 (Caliper Life Sciences, Alameda, CA).
Tissues were harvested and prepared as follows: 100±20 mg of tissue was placed in a 1.5 mL microfuge tube and 9 volumes of phosphate buffer was added, diced, and immediately sonicated at 0°C using an ultrasonic cell disruptor with an ice-cold 3-mm microprobe yielding a concentration of 2 mg fresh tissue weight per 20 μL of sonicate. Tissue HO activity was determined through measurement of CO as previously described.  Briefly, 20 μL of tissue sonicate was incubated at 37°C for 15 min in CO-free septum-sealed 2-mL vials, containing 20 μL of methemalbumin (150 μmol heme/L:15 μmol BSA/L) and 20 μL of NADPH (4.5 mmol/L). Blank reaction vials contained 20 μL of phosphate buffer in place of NADPH. The reaction was stopped by the addition of 5 μL of 15% sulfosalicylic acid.  CO produced and released into the vial headspace was then quantified by gas chromatography (GC). HO activity was calculated as pmol CO generated/h/mg wet weight.
The Luminex assay was performed as recommended by the manufacturer (Panomics/Affymetrix). Assays were performed in duplicate using the Luminex 200 IS System (Luminex Corporation). Individual cytokines and chemokines were identified and classified by the red laser, and levels were quantified using the green laser. Digital images of the bead array were captured following laser excitation and were processed on a computer workstation.
Standard curves and reports of unknown samples were prepared using BeadView and MiraiBio software.
In situ hybridization of pancreatic sections was performed as described.  In brief, digoxigenin-labeled sense and anti-sense RNA probes were generated by polymerase chain reaction amplification with the T7 promoter incorporated into the primers. The primers used for CXCL1 were: 5′-ATGATCCCAGCCACCCGCTCGCTT -3′ (sense) and 5′-CCGTTACTTGGGGACACCTTTTAGCATC -3 (antisense). In vitro transcription was performed with a digoxigenin RNA-labeling kit and T7 polymerase (Roche Diagnostics, Indianapolis, IN). Pancreatic sections from paraffin blocks were used for hybridizing with either sense or anti-sense as described.  Sections developed with diaminobenzidine (GenPoint Kit, DAKO) and counterstained in hematoxylin are shown.
Student’s t-test was used to assess statistical significance and a P value of less than 0.05 was considered significant. Values are expressed as mean ± SEM. Unless indicated, results are from at least 3 independent experiments. Survival curve comparison was based on the Log-rank (Mantel-Cox) test.
We previously showed that multiple i.p. injections of hemin during a one-week period upregulate HO-1 significantly by days 4 and 5 of starting the injections.  Here, we examined whether HO-1 expression can be upregulated with a single i.p. or i.v. dose of PH to ascertain whether PH can be utilized for AP therapy. Interestingly, we found that HO-1 expression was upregulated as early as 2 h following a single-dose administration of PH (Figure 1A). Similarly, immunofluorescence staining revealed increasing number of HO-1+ cells at 4 h post-injection, thereby suggesting rapid recruitment of HO-1+ cells into the pancreas regardless of the route of PH administration (Figure 1B).
Previous studies using i.v. 55Fe-hematin in rhesus monkeys showed a half-life in plasma ranging from 2.2 to 4.8 h.  The majority of the injected radioactivity was observed in the liver (48.2%) 21 h after 55Fe-hematin injection, but hematin distribution to the pancreas was not examined.  Thus, based on the beneficial effects of hemin in preventing pancreatitis and the early upregulation of HO-1 protein expression (Figure 1A) following a single dose of PH, we investigated the distribution of 55Fe-hemin in homogenized pancreata and other tissues 4 and 24 h after i.v. injection.  Consistent with previous studies, liver, spleen, and colon (in decreasing amounts) had the highest accumulation of radioactivity relative to other tissues at 24 h (Figure 2A). 55Fe-hemin accumulation in the pancreas was limited as compared with the three major organs of radiolabeled hemin deposition. Similarly, whole mouse section autoradiography demonstrated high activities in the liver and spleen (Figure 2B).
Pre-treatment with hemin protects against mild to moderate (caerulein model) and severe (CDD model) forms of experimental pancreatitis.  In order to evaluate whether PH can treat early or late pancreatitis, we used the CDD-induced pancreatitis model, where mice are fed CDD for 3 days thereby resulting in severe hemorrhagic pancreatitis with peak injury on day 3. Lipase levels in CDD-induced pancreatitis correlate well with pancreatic injury as assessed by histologic scores.  We determined the expression of HO-1 and extent of pancreatic injury during CDD feeding over time. After 8 and 24 h of CDD feeding alone, there was no notable HO-1 up-regulation (Figure 3A; lanes 1,2,7,8). A significant increase in serum lipase was observed at 24 h, but not after 8 h (Figure 3B).
We then assessed pancreatic HO-1 expression after 72 h of CDD feeding in mice treated with either vehicle (VE) or PH at 8 or 24 h after initiating the CDD. Notably, a single dose of PH given either at 8 or 24 h following initiation of CDD induced and maintained HO-1 upregulation at 72 h (Figure 3A; lanes 5,6,11,12). This led us to test whether this dosing regimen after feeding with CDD offered a therapeutic benefit. We found that a single dose of PH, equivalent to a hemin dose of 25 μg/g mouse weight, which previously conferred protection against AP,  dramatically suppressed serum lipase elevation (Figure 3C) and pancreatic injury as determined after scoring of the damage using defined histological criteria (Figure 3D & E) and trypsin activation (Figure 3F) at 72 h. In addition, mice receiving PH at 8 and 24 h had better survival at 72 h (Figure 3G). Importantly, PH was also similarly effective in treating L-arginine induced pancreatitis (Figure 4). Blood, pancreas and liver homogenate cultures did not yield any bacterial growth in either 24 h VE or PH treated mice euthanized at 72 h following initiation of CDD (data not shown).
Given the therapeutic effects of administering PH 8 and 24 h after CDD feeding, we investigated whether PH can reverse or treat late AP. As expected, there was further rise in serum lipase at 36 and 48 h following CDD initiation (Figure 5A). However, PH at the dose of hemin (equivalent to 25 μg/g weight) given at 36 and 48 h did not protect against AP (Figure 5B). The difference in the protective effect of PH at 36 and 48 h versus 8 and 24 h prompted us to examine whether the dose of PH administered caused potential toxic effects similar to that of high dose hemin,  particularly in severely ill mice during late phase pancreatitis. Consistent with this notion, PH treated mice have higher serum blood urea nitrogen (BUN) levels (Supplementary Figure 1). Thus, we examined upregulation of pancreatic HO-1 at lower doses of PH to determine if we could mitigate potential PH toxicity. Dose response studies showed that administration of PH at 6.2 μg/g of mouse weight afforded significant HO-1 overexpression in the pancreas (Figure 6A). Based on this finding and the recommended upper limit dosage in humans [equivalent to 6 mg/kg body weight over 24 h; ], we used 6 μg/g weight dosing to test PH effectiveness in late AP. Notably, this lower dose of PH was effective in treating AP when given at 48 h, but not earlier at the 24 h time point (Figure 6B). For comparison, serum lipases from controls (no CDD) and mice fed CDD only for the first 24 or 48 h are shown (Figure 6C).
In order to address the mechanism of PH action, we studied whether there is de novo transcription of HO-1 and whether the induction of HO-1 by PH correlates with HO activity by measuring HO-1 promoter activity using HO-1-luc transgenic mice (Luc-HO-1+/+) and via release of CO from homogenized tissues, respectively. Mice were imaged 0, 2, 4, 6, 8, 12, 18, 24, and 30 h following a single dose of either VE or PH. Significant bioluminescence was observed at 12 h on the left dorsal and lateral areas in mice treated with PH (Figure 7A), likely corresponding to the spleen. Based on this finding, we then treated normal FVB/n mice (similar to those used in induction of AP) with either VE or PH, and assayed HO activity in lung, liver, spleen, and pancreas sonicates after 12 and 24 h. There was significant increase in HO activity in the liver, spleen, and pancreas, but not lung of PH treated mice at both 12 (not shown) and 24 h (Figure 7B). Interestingly, a higher fold increase in HO activity was observed in the pancreas as compared to the spleen following PH treatment (grey bars; Figure 7B).
Activated leukocytes during the early phase of pancreatitis release cytokines, which in turn mediate and enhance the inflammatory cascade observed in AP. Particularly, TNFα and IL-1 are among the most prominent and well-studied in AP. [6, 22, 23] We performed Luminex assays to determine if PH treatment affects cytokine levels in the pancreas. Although no difference in IL-1 pancreatic levels was noted between 24 h VE or PH treated mice fed with CDD for 60 h, there was a significant decrease in TNFα in PH treated mice (Figure 8A). Nevertheless, we could detect significant decrease in IL-1β in serum of PH treated mice (Supplementary Figure 2). In addition, there were significant decreases in pancreatic IL-12p40 and IL-6 in the PH treated mice.
We previously showed that single dose of hemin given i.p., in non-CDD fed mice, increased pancreatic CCL2 (MCP-1) and CCL3 (MIP-1α) but not CCL5 (RANTES) mRNA 24 h later.  Herein, we did not detect significant differences in pancreatic protein levels of CCL3 or CCL5 after 60 h CDD feeding in mice treated with either VE or PH at 24 h during the feeding (Figure 8B). We selected to harvest the pancreas at 60 h and not 72 h, since all the mice would be alive at 60 h. Based on the fact that CXCL1 is a potent neutrophil attractant and neutrophils play an important role in pancreatic injury during AP,  we tested whether PH can modulate pancreatic CXCL1. Consistent with the findings that PH can treat experimental pancreatitis, CXCL1 was significantly decreased in mice treated with PH (Figure 8B).
In order to determine the cellular source for the increased CXCL1, we performed in situ hybridization with CXCL1 sense (Figure 8C, d) and anti-sense (Figure 8C, a–c) probes. Relative to VE-treated mice, pancreata from PH-treated mice had significantly reduced CXCL1 mRNA levels (Figure 8C, b versus c). In VE-treated mice, the CXCL1 mRNA was found primarily proximal to vascular areas (Figure 8, b). Consistent with the reduction in pancreatic CXCL1, pancreata from PH-treated mice had decreased leukocyte infiltration as demonstrated by a decrease in CD45+ cells (Figure 8D; c and d versus a and b).
AP remains a challenging clinical problem, particularly in patients with severe disease. Despite a handful of clinical trials with pharmacologic agents, no effective treatment exists for AP. We previously demonstrated that pre-treatment, as a prophylactic, with hemin or hemin-activated cells protects against experimentally-induced pancreatic injury.  In the study herein, the beneficial effect of hemin prophylaxis has been extended to therapeutic intervention using the hemin formulation PH in mice undergoing two different models of pancreatitis. Importantly, even at a later time during AP, PH was effective in minimizing the observed pancreatic injury. This finding is likely related to the observation that HO-1 in the pancreas is quickly upregulated and maintained following a single PH dose. Clinically such potency may be critical since patients with AP are a heterogeneous group, presenting to the hospital at different time during the course of the disease.
The limited accumulation of 55Fe-hemin in the pancreas suggests that the pancreas is not a major site for hemin uptake. This is consistent with our previous finding that HO-1+ F4/80+ cells are recruited to the pancreas from elsewhere and that these recruited cells provide the source for pancreatic HO-1 and is responsible for the observed protection. Potentially Panhematin may also affect other organs such as the liver and spleen monocyte/macrophages and suppress systemic inflammatory response. Furthermore, PH not only upregulates HO-1 at the protein level but also functionally as we found significant increases in HO activity in the pancreas. The beneficial effects of HO-1 induction may be mediated via its byproducts, CO and biliverdin/bilirubin, since both have been shown to have anti-inflammatory and anti-oxidant properties. [25–28] In support of our findings, the CO-releasing molecule-2 (CORM-2) demonstrated a protective effect in a rat model of pancreatitis.  Potential therapeutic limitations of such compounds lie in their instability and insolubility in aqueous solutions. In addition, our study demonstrated the efficacy of PH against AP at various time points whereas CORM-2 efficacy was demonstrated only at one time point immediately after AP induction. The relative ease of PH administration, its potency in inducing HO-1 expression in the pancreas, and its efficacy using two independent models of AP at different stages of the disease renders it an attractive potential candidate therapeutic agent for AP.
PH treatment resulted in significant reduction in TNFα, IL-6, and IL-12p40 levels in pancreatic tissue. Serum levels of TNFα are increased in patients with AP and mouse models of the disease, and increased levels are correlated with disease severity.[29–31] Blockade of TNFα in experimental mouse models ameliorates AP, indicating the importance of this cytokine in disease pathogenesis. The reduction of TNFα levels in the pancreas may be one of the mechanisms by which PH improves AP. IL-12 in conjunction with IL-18 has been shown to induce necrotizing AP in obese ob/ob mice. In humans, IL-12p40 is consistently increased during the course of AP; whereas, IL-12p70 was only increased on day 1, but significantly decreased on the second to fourth day of illness.  Our findings are supportive of these results and show that, 60 h post-CDD-feeding, IL-12p40 are much higher as compared to IL-12p70. In addition, PH treatment leads to a decrease in IL-12p40 and an increase in IL-12p70, although the latter was not statistically significant. Increased serum IL-6 levels are also observed in AP patients and in experimental mouse models.  Although IL-6 does not appear to play an important role in initiating or mediating the inflammatory cascade in AP, it is considered a useful predictor of disease severity.  In support of PH’s therapeutic benefit in experimental AP, we find that IL-6 is significantly lower in PH-treated animals as opposed to their VE-treated controls.
IL-1 blockade, via its receptor or via IL-1 converting enzyme (ICE), [36–40] attenuates experimental pancreatitis. PH here did not alter pancreatic IL-1 levels 60 h after the initiation of CDD feeding, suggesting that an IL-1 independent mechanism may account for PH’s therapeutic action. However, our results do not rule out a role for IL-1 down-modulation in PH-mediated protection against AP since the effect of PH on IL-1 levels may occur early during the course of pancreatitis. This latter hypothesis is supported by the fact that at 60 h we could detect significant IL-1β reduction in the serum (Supplementary Figure 2) and IL-β levels in the pancreas are much higher early during CDD feeding (e.g. at 24 h as opposed to 48 h or 60 h, data not shown). For example, CORM-2 suppresses IL-1β levels 12 h after induction of AP.  Similarly and in the same study, levels of the anti-inflammatory cytokine, IL-10, increased 12 h after induction of AP. We did not observe significant changes in pancreatic IL-10 levels 60 h after CDD diet feeding in response to PH treatment. Again, this suggests that PH-mediated protection does not depend on IL-10 solely or, alternatively, that the effect of PH on IL-10 levels occurs at an earlier time point.
In addition to the above-mentioned cytokines, our findings show that the neutrophil chemoattractant, CXCL1, is reduced in PH-treated mice. CXCL1, a potent neutrophil chemoattractant similar to IL-8 in humans, is expressed in fibroblasts, macrophages and endothelial cells particularly in response to inflammatory stimuli. [41, 42] Furthermore, marked reduction in severity of pancreatitis is observed with depletion of neutrophils. [24, 43] The fact that PH suppresses local pancreatic CXCL1 expression supports the notion that PH ameliorates AP by mitigating neutrophil recruitment into the pancreas through reduction in CXCL1 levels. This conclusion is further supported by the reduced leukocyte (CD45+ cells) infiltration that was observed following PH treatment.
One important aspect of our findings is the dosing of PH. For example, higher and lower doses of PH were effective if given earlier and later during the course of AP, respectively. Over time during the CDD feeding, mice become more dehydrated, as shown by the increase in blood urea nitrogen (BUN) levels (Supplementary Figure 1). Treatment with higher doses of PH likely compromises renal function further, since larger doses of PH are known to cause reversible renal failure.  In patients with porphyria, PH is given daily for 3 to 14 days based on the resolution of clinical symptoms and appears to be well tolerated. However, if our findings are translated to patients with AP, care will need to be taken with the dosing. Collectively, our findings show that the use of PH offers a novel potential therapeutic strategy for AP.
Despite significant morbidity and mortality associated with AP, no effective treatments exist aside from supportive care such as fluid resuscitation. The therapeutic benefit of Panhematin in experimental pancreatitis raises the hypothesis that pharmacologic administration of Panhematin may ameliorate disease severity in patients with AP.
We thank Jean Chen, Hui Ye Zheng and Evelyn Resurreccion for their technical assistance. We are also grateful to Timothy C Doyle for guidance with whole animal sectioning, Kelli Montgomery and Rui Li for assistance with in situ hybridization, and Yael Rosenberg-Hasson for assistance with the Luminex assay. This work was supported by NIH grants DK073909, DK47918 and by Ovation Pharmaceuticals (M.B.O.), and NIH Digestive Disease Center grants DK56339 to Stanford University and DK34933 to the University of Michigan.
Declared potential conflict of interest: The supplier of Panhematin (formerly called Ovation Pharmaceuticals) provided support for this study and has a pending patent application for the possible use of Panhematin in human pancreatitis.