| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Published online before print November 9, 2006
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Pathology and Microbiology, University of Nebraska Medical Center, Omaha, Nebraska, USA
1 Correspondence: Department of Pathology and Microbiology, University of Nebraska Medical Center, 986495 Nebraska Medical Center, Omaha, NE 68198-6495, USA. E-mail: tjerrell{at}unmc.edu
ABSTRACT
Acute and chronic pancreatitis is associated with alcohol abuse, but symptomatic pancreatitis develops in only a small proportion of persons (10–20%) who abuse alcohol. This apparent paradox has led to the notion that additional cofactors are involved in the development of alcoholic pancreatitis. Potential cofactors, such as diet and smoking, have been suggested, but there are no compelling epidemiologic data to support this idea. A number of viruses and some bacteria have been shown to infect the pancreas and produce pancreatitis. One important mediator of pancreatitis in persons with a compromised immune system is a viral infection. The increased susceptibility of immunocompromised persons to viral pancreatitis led to the hypothesis, described in this paper, that the well-known immunosuppression associated with alcohol abuse would result in a more severe viral pancreatitis in mice, which are provided ethanol, than in control animals. To test this hypothesis, C57BL/6 mice were infected with a virulent strain of coxsackievirus B3, which preferentially induces pancreatitis, or with a strain that is naturally avirulent. The study findings presented in this paper show that ethanol consumption alone does not produce pancreas damage but results in a more severe and prolonged pancreatitis after infection with a virulent virus and interestingly, after infection with the avirulent strain of virus. This was associated with an increased number of viruses in the pancreas and spleen, which correlated with decreased humoral immune responses to the virus.
Key Words: inflammation • immunosuppression • coxsackievirus
INTRODUCTION
A number of viruses have been shown to infect the pancreas and induce acute and chronic pancreatitis [1 ]. Viral infections, which are relevant to the studies done to date to evaluate the role of viral infection in the development of alcoholic pancreatitis, include cytomegalovirus infection, which has been shown to be important in the development of pancreatitis in immunodeficient persons, and coxsackieviruses, especially the B3 type (CVB3) of this viral group [2 , 3 ]. The mechanisms of pathogenesis of viral infections of the exocrine pancreas have been described clearly with the use of animal models of pancreatitis and coxsackievirus infections [4 , 5 ]. An important finding resulted from the studies conducted by Tracy and colleagues [4 ], who showed that pathogenesis of CVB3 in mouse models can faithfully reproduce the infection of human beings mediated by each individual strain. In addition, strains have been characterized that only induce pancreatitis, although other organs are infected, and strains differ in virulence to include an avirulent strain [4 ]. It is important to note that the viruses used in the studies described by Tracy et al. [4 ] and in this paper infect and damage the exocrine pancreas (i.e., acinar cells) and do not affect the endocrine pancreas (i.e., islets and ducts) [3 , 4 , 6 ].
Pancreatitis, acute and chronic, is closely associated with alcohol abuse, but symptomatic pancreatitis develops in only 10–20% of persons who abuse alcohol for long periods [1 , 7 ]. This apparent paradox has led to several investigations to define potential cofactors, such as smoking and consumption of a high-fat diet, which would interact with alcohol abuse for the development of pancreatitis [8 9 10 ]; however, no clear association of these factors with alcoholic pancreatitis has been established. We proposed the hypothesis that viral infections are important cofactors for the development of alcoholic pancreatitis. A publication from this laboratory [6 ] and data presented in this paper have provided preliminary data that show that CVB3 infection of mice, provided ethanol with a liquid diet protocol [11 ], was associated with the development of severe pancreatitis, even with the avirulent strain (CVB3-GA) of this virus. It is important to note that findings of studies to date with the use of this relatively short-term feeding protocol [6 ] and a chronic ethanol consumption protocol shown in this paper have shown that ethanol consumption by mice does not in itself result in pathologic changes in the pancreas. Results of these studies have led to the hypothesis that the well-known suppression of innate and adaptive immune responses to infectious agents associated with alcohol consumption leads to a decreased ability to control the viral infection, which results in the development of a more severe acute pancreatitis and ultimately, the development of chronic pancreatitis and fibrosis. The study results presented in this paper provide data to support this hypothesis.
MATERIALS AND METHODS
Mice and ethanol-feeding protocols
Two protocols were used in the studies reported in this paper. The first protocol is a well-described model of feeding ethanol in a defined liquid diet, which is a modification of the Lieber-DeCarli liquid diet [11
]. In brief, female C57BL/6 mice, 6–8 months of age (Charles River, Portage, MI, through a contract with the National Cancer Institute, Bethesda, MD), are acclimated gradually to the ethanol-containing diet (usually over 10 days) until the mice are consuming diet that provides 36% calories, such as ethanol. The modification of the diet for these studies is to use a relatively low-fat diet. When the ethanol group begins consumption of the diet containing 36% ethanol-derived calories, a group of mice is provided a control liquid diet, which is isocaloric to the ethanol diet by the inclusion of dextrin and maltose with a pair-feeding protocol in which each mouse in this group receives a volume of diet equal to the average of the amount of diet consumed by the mice in the ethanol-fed group. The second protocol provides ethanol in the drinking water as a 20% (v/v) solution, essentially as described by Song et al. [12
]. In brief, mice are acclimated over 7–10 days to reach the goal of 20% ethanol in drinking water. In most experiments, mice were provided 20% ethanol for 6–8 weeks before infection, and the feeding protocol was continued throughout the infection period.
Virus and infection
In the studies reported in this paper, mice were infected with CVB3-CO, which is a pancreas-specific virulent strain of CVB3, or CVB3-GA [4
]. It is important to note that these viruses have not been passed in mice, which would likely increase virulence for mice, and have been propagated in HeLa cells as described before [4
, 13
]. Viral concentrations in stock cultures and tissue homogenates were determined with the use of a plaque assay performed with HeLa cells [4
, 13
]. Mice were infected i.p. with indicated doses of virus (100 µl) diluted in cell culture medium (DMEM, Mediatech, Inc., Herndon, VA), supplemented with 2% FBS.
Parameters of pathogenic effects of viral infection
To evaluate pancreas damage, serum concentrations of lipase and amylase were determined with the use of assay kits [Lipase Color Assay, Thermo Electron Corp., Pittsburgh, PA (formerly Data Medical Associates, Inc., Arlington, TX), and Infinity Amylase EPS single liquid stable reagent, Thermo Electron Corp. (formerly ThermoDMA, Louisville, CO)]. The extent of pancreas damage was also evaluated by examining tissue sections of formalin-fixed pancreas stained with H&E. A qualitative evaluation of pancreatic fibrosis was done with fixed tissue sections stained with a Sirrius red stain, essentially as described by Haber et al. [14
].
Immunoblot for 
-smooth muscle actin
A further evaluation of the induction of fibrosis was done with immunoblots of tissue homogenates of pancreata, which were developed with antismooth muscle actin antibody, which identifies activated stellate cells in the pancreas [15
, 16
].
Immunoblot assays were done essentially as described before [17
] to determine the extent of pancreatic stellate cell activation by a semiquantative assay for -smooth muscle actin. Briefly, pancreata were isolated at various times after infection and homogenized as a 10% w/v suspension in radioimmunoprecipitation assay buffer containing a mixture of protease inhibitors (Complete Mini, Roche Applied Science, Indianapolis, IN), and after electrophoresis, proteins were transferred to nitrocellulose membranes (Immobilon-P transfer membranes, Millipore, Bedford, MA). Monoclonal anti--smooth muscle actin (Clone 1A4) was obtained from Sigma Chemical Co. (St. Louis, MO). Membranes were probed with anti--smooth muscle actin or anti-ß-actin (Sigma Chemical Co.) to ensure uniform loading, and bands were visualized by ECL (Amersham Pharmacia Biotech, Piscataway, NJ).
Isolation of mononuclear cells and flow cytometric analysis of phenotype
Intact pancreata were removed from the animals and placed immediately in cold calcium- and magnesium-free HBSS containing a mixture of protease inhibitors (Complete Mini Protease Inhibitor Cocktail Tablets, Roche Diagnostics Corp., Roche Applied Science). Pancreata were disrupted by passage through a metal screen, and an equal volume of RPMI 1640 supplemented with 10% FBS was added immediately to each cell preparation. The entire cell fraction was centrifuged (1200 RPM for 10 min at 4°C), and the pellet was resuspended in RPMI 1640 supplemented with 5% FBS. Mononuclear cells were enriched by gradient centrifugation (Lympholyte-M, Cedarlane Laboratories Ltd., Hornby, Ontario, Canada), viable cell numbers were determined by Trypan blue exclusion, and cells were adjusted to 106/mL in PBS containing 2% FBS and 0.1% azide (FACS buffer). Mononuclear cells were stained with mAb to CD3 (145-C11), CD4 (RM4-5), CD8 (53-6.7), F4/80 for macrophages, NK1.1 (PK136), and B220 (RA3-6B2). Fluorescence was evaluated with the use of FACSCalibur (Becton Dickinson, San Jose, CA) in the University of Nebraska Medical Center Cell Analysis Core (Omaha). A minimum of 104 events was obtained, and the percentage of each phenotype was determined. Total numbers of each phenotype were determined by multiplying the percentage and the total number of mononuclear cells obtained at each experimental time.
To provide a visual evaluation of mononuclear cell infiltration, isolated pancreata were fixed with IHC zinc fixative solution (BD Biosciences PharMingen, San Diego, CA) and stained with anti-CD8 mAb by standard immunohistochemistry protocols. Briefly, tissue was deparanninized, and endogenous peroxidase was neutralized with a 0.3% solution of H2O2 and blocked with blocking serum provided in the Vectastain peroxidase mouse IgG ABC kit (Vector Laboratories, Inc., Burlingame, CA). Biotin-labeled antibody binding was visualized with the use of peroxidase goat antimouse IgG and 3'-diaminobenzidine tetrahydrochloride substrate with a metal (nickel)-enhanced protocol, according to the manufacturer’s instructions (Vectastain peroxidase mouse IgG ABC kit, Vector Laboratories, Inc.).
Serum cytokine concentrations were determined with the use of a flow cytometric method [cytometric bead array mouse inflammation kit, BD Biosciences PharMingen], according to the manufacturer’s instructions.
Serum concentrations of neutralizing antibodies were determined with an in vitro assay to measure inhibition of viral cytopathic effects with the use of HeLa cells as the target cell. In brief, an initial 1:100 dilution and serial twofold dilutions of heat-inactivated (56°C for 30 min) sera were prepared in DMEM-2%, and each dilution was incubated with virus at 37°C for 30 min. The concentration of virus used was the lowest concentration of virus that produced a 100% cytopathic effect of HeLa cell monolayers in 18–24 h. HeLa cell monolayers were established in 96-well plates, and 100 µl each serum/virus dilution was added to triplicate wells. Controls included virus and diluted normal mouse sera, serum samples obtained from the infected mice, and virus only. After 18–24 h of incubation, the medium was removed from each well. Cells were fixed with 10%-buffered formalin for 60 min and stained with crystal violet. To evaluate protection, the OD of each well was determined with the use of an ELISA plate reader (Awareness Technology, Palm City, FL) after extraction of the crystal violet with methanol at 570 nm.
Statistical analyses
Where appropriate, a one-way ANOVA in association with the Student-Newman-Keuls multiple comparison test was used to determine statistical significance of data that contained three or more groups. Percentage data were evaluated with the Kruskal-Wallis statistic. Data obtained from experiments from two groups, which were normally distributed, were analyzed with a two-tailed t-test. Differences at a P value less than 0.05 were considered statistically significant. Calculations were done with StatView software (SAS Institute Inc., Cary, NC).
RESULTS
To evaluate the effects of short-term ethanol administration on the pathogenesis of CVB3 infection, mice were provided the modified Lieber-DeCarli diet, infected (2.5x105 PFU) with CVB3-GA, and evaluated for pancreas damage on the basis of serum concentrations of amylase and lipase. The resultant data are presented in Figure 1
. As previously reported [14
], pancreas damage is greatest 4 days after infection of mice, and there is more damage in the ethanol-fed mice on the basis of amylase concentrations in serum samples. Amylase concentrations are significantly (P
0.05) higher in serum samples obtained from ethanol-fed mice, 4 and 6 days after infection (Fig. 1a)
. It is apparent from these data that infection of control mice (pair-fed and chow-fed) with this strain of virus yields little to no pancreas damage at the dose used, as the serum levels of amylase and lipase were essentially normal values (see below). Serum concentrations of lipase were also higher in the serum samples obtained from ethanol-fed animals, in comparison with findings for control animals, which were essentially normal values, at similar times after infection (Fig. 1b)
.
|
There were no significant differences in these control levels of amylase or lipase in sera of mice provided ethanol in the liquid diet or the alcohol in water protocols.
To assess the effects of chronic ethanol consumption on coxsackievirus-mediated pancreatitis, mice that were provided ethanol in drinking water for 6 weeks or control mice were infected with CVB3-CO, and the severity of the pancreatitis was evaluated as above. Serum concentrations of lipase and amylase after infection of mice, provided ethanol in drinking water, and control mice, maintained with drinking water without ethanol for the same time, are shown in Figure 1c
and 1d
, respectively. The control group shows the expected rise in serum concentrations of both enzymes 72 h after infection with this virulent strain of CVB3. These data show significantly (P0.05) higher serum concentrations of both enzymes at 96 h after infection and amylase at 72 h after infection, which provides evidence that consumption of ethanol for relatively long periods in this chronic model results in increased pancreas damage associated with ethanol consumption. Therefore, the remainder of the studies presented in this paper was done with mice provided ethanol by the more relevant chronic model, essentially as described by Song et al. [12
].
To assess the possibility that chronic ethanol consumption results in the development of chronic pancreatitis, the chronic model described by Song et al. [12 ] was used to provide ethanol to mice. After 8 weeks of ethanol consumption, mice were infected with 5 x 105 PFU CVB3-CO, and infected mice were evaluated at indicated times after infection for up to 16 days. The group of mice provided ethanol in the drinking water was provided ethanol for the total duration of the experiment.
To evaluate chronic pancreatitis histologically, sections of pancreata obtained from mice provided ethanol for 6 weeks and control mice infected with CVB3-CO and evaluated up to 16 days after infection were stained with H&E. Figure 2a
2b
2c
2d
2e
, shows representative H&E-stained tissue sections from ethanol-fed mice, control mice, 10 and 12 days after infection, and an uninfected mouse provided ethanol for 8 weeks in the same experiment. It is clear from the histologic data that there are no obvious effects of ethanol consumption alone, as the histologic aspects of the pancreata obtained from ethanol-consuming mice were indistinguishable from those of the pancreata obtained from uninfected control mice (data not shown). There are clearly more inflammatory cells in the pancreas obtained from the ethanol-fed mouse in comparison with findings for the control animals. These data are in agreement with the increased numbers of mononuclear cells recovered from the pancreata of ethanol-fed mice when compared with the cell numbers isolated from the control mice (Fig. 3a
3b
3c
) at a similar time after infection as the histologic evaluation described above. Data in Figure 3a
show that similar numbers of mononuclear cells (lymphocytes and macrophages) were found in pancreata obtained from mice provided ethanol and control mice through 8 days of infection. Evaluation of mononuclear cell numbers 12 and 16 days after infection revealed that the numbers in the control group had essentially returned to baseline, but higher numbers were noted in the pancreata obtained from the mice provided ethanol. Although there was considerable variability in the numbers of each mononuclear cell phenotype in the pancreata from each group, it appears that there are significant (P0.05) increases in the number of NK1.1 cells and CD4+ T cells 8 days after infection in the group provided alcohol (Fig. 3b)
, the number of CD4+ T cells remained elevated 12 days after infection, and numbers of CD8+ T cells were higher in the group provided alcohol 12 and 16 days after infection as compared with the number of each cell type in the control group (Fig. 3c)
. Figure 3
, d and e, shows representative immunohistochemistry staining for CD8+ T cells in pancreata from a representative animal from each group. (Fig. 3d
is a control mouse, and Fig. 3e
is a mouse from the group provided ethanol in water for 12 weeks and 12 days after infection.) It appears from these stained sections that there are more CD8+ T cells in the tissue of the mice provided ethanol as compared with the control group. The majority of lymphocytes is in the inflammatory lesions, as indicated by the presence of ductals in both groups. Similar data and distribution of positive cells were seen in sections stained for CD4+ T cells (data not shown).
|
|
|
As can be seen in Figure 5a and 5b , there are extensive areas of Sirrius red-stained material in the representative sections of pancreata obtained from ethanol-fed mice 16 days after infection with CVB3-CO. Figure 5b shows a higher power view of the Sirrius red staining to show an area of staining around a ductal, where the stellate cells would be expected to produce collagen. In comparison, staining in the pancreata obtained from the infected control mice 16 days after infection (Fig. 5c) shows less-intense staining. Uninfected mice provided ethanol in the same experiment (Fig. 5d) show minimal staining, mostly in areas around the ducts and vessels.
|
-smooth muscle actin (data not shown) support the suggestion that stellate cells are activated in this model system. The data obtained in these assays show that -smooth muscle actin is detectable in the pancreas of mice consuming alcohol for at least 12 days after infection. In these experiments, we found that the inflammation and fibrosis continued, although there was no detectable virus in the pancreas after 12 days of infection (Table 1 ). Also, there was ten- to 100-fold more infectious virus in the pancreas of ethanol-fed animals at 4–8 days of infection in comparison with the viral numbers in the pancreas of the control mice.
|
and MCP-1 peaked 3 days after infection (Fig. 6a
and 6b
), which was similar to the times serum concentrations of amylase and lipase peaked. These were the only cytokines found to be elevated significantly in serum samples in these experiments.
|
|
The data presented in this paper show that relatively short-term and chronic (i.e., 6–8 weeks) ethanol consumption by mice results in a more severe pancreatitis mediated by a pancreas-specific virus (i.e., CVB3-CO) in comparison with the pancreatitis in the appropriate control groups. There is also a clear association of chronic ethanol consumption with the continuation of pancreatitis and fibrosis for relatively long times, as the virus has been cleared, which is not seen in the infected control animals. Of interest is the association of proinflammatory cytokine production (notably, TNF) with early pancreas damage, as determined by serum concentrations of amylase and lipase, which supports the suggestion that the exaggerated production of this cytokine is a critical factor in the pathogenesis of infection of mice provided alcohol. We also propose that the higher serum concentrations of MCP-1, a chemokine for macrophages and NK cells in the ethanol-fed mice, are a possible mechanism for the exaggerated, early pancreatic inflammation noted in these studies, and studies are in progress to address this point. It is interesting that the numbers of other inflammatory cells (notably, CD4+ and CD8+ T cells) are higher in the ethanol-fed mice than in the control mice at the later times of infection, which is predicted to be the result of a failure to control the replication of the virus early in the infection associated with ethanol consumption and a resultant, continued stimulation of the immune system. The study results reported in this paper show that ethanol consumption results in a continued inflammatory response that persists after the infectious virus is no longer demonstrable in the pancreas. The chronic pancreatitis is clearly associated with the development of fibrosis in the pancreas, which is one of the key pathologic features of chronic pancreatitis [14
, 16
, 18
, 19
]. In this model system, we have been able to demonstrate stellate cell activation on the basis of -smooth muscle actin concentrations in the pancreas, as well as -smooth muscle-positive cells in the pancreas detected by immunohistochemistry (not shown), which supports this finding. Whether the fibrosis noted in this model is reversible remains to be seen, but it is anticipated that chronic ethanol consumption will facilitate the continued fibrosis.
One mechanism that we can propose for the prolonged pancreatitis is the inability of the ethanol-consuming animal to control the early viral infection in the pancreas. In the current study, in comparison with the number of virus in control animals, there was from ten- to 100-fold more infectious virus in pancreatic tissue obtained from the ethanol-fed groups early in the infection. The increased and perhaps prolonged infection would likely increase the inflammation produced by the infection and subsequent pancreas damage. It is also possible that the incomplete immune response to the virus, resulting in a persistent infection of the pancreas, which has been described by others [20 21 22 ], would also perpetuate the inflammation. The hypothesized effect of ethanol consumption on the immune response to CVB3 is supported by our finding that concentrations of neutralizing antibody are lower in serum samples obtained from ethanol-consuming mice in comparison with the concentrations in control animals. Although there are no apparent differences in the numbers of CD8+ T cells at the early times of infection that were evaluated, it is possible that there is a delay in the response of antigen-specific CD8+ T cells, and studies are in the process to evaluate the specific CD8+ T cell response to the virus, as well as the effect of ethanol on this response.
Another possible effect of ethanol consumption on the increased pathogenesis of the virulent and avirulent strains of CVB3 virus is a "sensitization" of the pancreas to the "second hit" provided by the virus and subsequent production of proinflammatory cytokines, especially TNF [18 , 23 ]. This concept has been accepted in terms of cytokine-mediated damage to the liver, and this has been proposed to be primarily because of depletion of antioxidants mediated by the metabolism of ethanol in the liver [24 , 25 ]. The pancreas has the ability to metabolize ethanol [26 ], and this is a possible cofactor for the increased pathogenesis of viral infection of the pancreas modeled with CVB3, again because of oxidative stress. Perhaps lessons learned from the study results defining the mechanisms of alcoholic liver disease may apply to the mechanisms of alcoholic pancreatitis.
It has been proposed by Beck and colleagues [27 , 28 ] that immunosuppression and oxidative stress result in the development of a population of coxsackievirus that has mutated and as a result, become more virulent. It has been shown that ethanol consumption by mice and alcohol abuse by human beings result in demonstrable suppression of the immune response, which is associated with an increased susceptibility to pathogenic organisms [11 , 29 30 31 ]. We propose that a similar situation described by Beck and Levander [27 ] may occur as a result of the suppression of specific viral immunity, along with oxidative stress associated with ethanol metabolism in the pancreas, in the ethanol-consuming mice, which results in a change in virulence of the CVB3-GA strain, which is normally avirulent. Studies with plaque isolates recovered from the infected pancreata of control and ethanol-consuming mice are in progress.
In summary, we have shown that ethanol consumption is associated with a more severe and prolonged pancreatitis mediated by pancreas infection with viruses that specifically infect the pancreas. Of importance is that ethanol consumption is associated with a pathologic infection with an otherwise avirulent strain of the CVB3 virus. The animals chronically provided ethanol showed evidence of a chronic pancreatitis and fibrosis. In a paper by Clemens and Jerrells [32 ], data were presented that showed that the exocrine pancreas regenerates quickly after the extensive damage induced by viral infection, and this is associated with large numbers of acinar cells that have proliferated. It is tempting to speculate that the chronic inflammation, fibrosis, and repeated proliferation of acinar cells provide a situation that would favor the production of tumor cells in the pancreas, as suggested by others [10 , 33 ]. A recent paper published by Ostrowski et al. [34 ] reveals evidence that a number of genes associated with cell growth, angiogenesis, and inhibition of apoptosis are expressed in one of their coxsackievirus model systems, and these data would further support the notion that extensive and repeated damage to the pancreas may be a cofactor in the ultimate production of pancreatic cancer. This is obvious speculation but warrants further study.
ACKNOWLEDGEMENTS
This research was supported by National Institutes of Health Grant AA013841 (to T. R. J.). The assistance with some of the assays by Dr. Dahn Clemens, Omaha VA Medical Center, is acknowledged. Corrina Gibbons was involved in the development of the antibody neutralization assay, and her help is also acknowledged. The excellent conceptual and technical assistance and providing of stocks of virus by Drs. Steve Tracy and Nora Chapman with these studies are gratefully acknowledged.
Received October 29, 2004; revised August 29, 2006; accepted September 15, 2006.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
