The role of phosphate in alcohol-induced experimental pancreatitis

Abstract

Background and Aims:

Heavy alcohol consumption is a common cause of acute pancreatitis, however, alcohol abuse does not always result in clinical pancreatitis. As a consequence, the factors responsible for alcohol-induced pancreatitis are not well understood. In experimental animals it has been difficult to produce pancreatitis with alcohol. Clinically, alcohol use predisposes to hypophosphatemia and hypophosphatemia has been observed in some patients with acute pancreatitis. Due to abundant protein synthesis, the pancreas has high metabolic demands, and reduced mitochondrial function leads to organelle dysfunction and pancreatitis. We proposed, therefore, that phosphate deficiency might limit ATP synthesis and, thereby, contribute to alcohol-induced pancreatitis.

Methods:

Mice were fed a low phosphate diet (LPD) prior to orogastric administration of ethanol. Direct effects of phosphate and ethanol were evaluated in vitro in isolated mouse pancreatic acini.

Results:

LPD reduced serum phosphate levels. Intragastric administration of ethanol to animals maintained on a LPD caused severe pancreatitis that was ameliorated by phosphate repletion. In pancreatic acinar cells, low phosphate conditions increased susceptibility to ethanol-induced cellular dysfunction through decreased bioenergetic stores, specifically affecting total cellular ATP and mitochondrial function. Phosphate supplementation prevented ethanol-associated cellular injury.

Conclusion:

Phosphate status plays a critical role in predisposition to and protection from alcohol-induced acinar cell dysfunction and the development of acute alcohol-induced pancreatitis. This finding may explain why pancreatitis develops in only some individuals with heavy alcohol use and suggests a potential novel therapeutic approach to pancreatitis. Finally, low phosphate diet + ethanol provides a new model for studying alcohol-associated pancreatic injury.

Short summary

Phosphate status is critical for normal pancreatic acinar cell function and underlies susceptibility to alcohol-associated pancreatic injury. Low phosphate diet + ethanol is a new model for studying alcohol-induced pancreatitis.

Introduction

Acute pancreatitis affects over 280,000 individuals per year in the United States and is consistently identified as a leading cause of hospitalization for gastrointestinal disorders ¹. One of the most common causes of acute pancreatitis is alcohol abuse ². Although the prevalence of pancreatitis is increased 4-fold in people with a history of alcoholism compared to those without ³, the prevalence of pancreatitis among heavy drinkers is only 3% ^{4, 5}. Thus, it appears that factors other than alcohol affect susceptibility to acute alcoholic pancreatitis. Identification of these risk factors is important for understanding the pathogenesis of acute alcohol-associated pancreatitis and for developing specific therapeutics for clinical use.

Developing experimental models for alcoholic pancreatitis has proven difficult. Similar to what has been observed in humans, laboratory animals do not reliably develop overt pancreatitis with exposure to alcohol alone. In current models, an additional “trigger” is required to produce pancreatic injury ⁶. For example, injection of lipopolysaccharide (LPS) has been used to induce pancreatitis in ethanol-fed animals ⁷. Similarly, co-administering ethanol and fatty acids ⁸ causes acute pancreatic injury. Although a binge alcohol model—initially developed to study alcoholic liver disease ⁹—may induce pancreatitis via ethanol exposure alone ¹⁰, this model relies on chronic ethanol feeding followed by a high-dose binge to trigger pancreatitis, a pattern of alcohol ingestion that does not reliably produce acute pancreatitis in humans. Nevertheless, all of these models indicate that alcohol is detrimental to the pancreas in the proper experimental setting and some have provided insight into the mechanisms of acinar cell damage. However, they have not answered the question of why heavy drinking induces pancreatitis in some individuals but not others.

Based on existing clinical data linking abnormal phosphate levels and critical illness ¹¹, we hypothesized that serum phosphate may modify susceptibility to pancreatic disease. Serum phosphate is critical for energy metabolism, nucleic acid synthesis, and cell signaling ¹². Strict regulation of serum phosphate levels is maintained by the kidneys, the gut, parathyroid hormone, fibroblast growth factor 23, and vitamin D, making hypophosphatemia relatively uncommon ¹³. However, certain populations are at high-risk for the development of hypophosphatemia, including patients with alcohol use disorder ¹⁴. In the setting of alcoholism, hypophosphatemia may develop due to nutritional deficiency, inappropriate phosphaturia, respiratory alkalosis, or gastrointestinal losses ¹⁵. Low serum phosphate levels have been associated with respiratory muscle dysfunction, decreased cardiac contractility, hemolysis, platelet dysfunction, insulin resistance, and myopathy ¹⁴. Furthermore, hypophosphatemia has been observed in clinical cases of alcohol- and gallstone-induced pancreatitis and has been correlated with a more severe clinical course ^16–18. However, the potential causal relationship between hypophosphatemia, alcohol use, and pancreatic dysfunction has not been investigated.

Disordered cellular energetics have been implicated in alcohol-associated acute pancreatitis. Alterations in intracellular calcium concentrations ([Ca²⁺]_i) cause pancreatic acinar cell dysfunction, intracellular enzyme activation, and produce acute pancreatitis in vivo ^{19, 20}. [Ca²⁺]_i homeostasis is maintained through mitochondrial generation of intracellular ATP. Diminished ATP stores leads to dysregulated [Ca²⁺]_i and acute pancreatitis ^21–23. In pancreatic acinar cells, high [Ca²⁺]_i itself may cause mitochondrial dysfunction and lower intracellular ATP²². However, the effects of phosphate depletion on cellular ATP levels and mitochondrial function have not been explored. The susceptibility of patients with heavy alcohol use to hypophosphatemia and the effects of reduced phosphate levels on pancreatic function led us to postulate that low serum phosphate levels increase susceptibility to and severity of acute alcoholic pancreatitis.

In the current study, we describe a model of acute pancreatitis that is reliably produced by acute alcohol exposure in the setting of diet-induced hypophosphatemia. We further demonstrate that low phosphate diet-associated acute alcohol-associated pancreatitis can be ameliorated by phosphate repletion. In vitro effects of phosphate on pancreatic acini support the concept that low phosphate levels increase susceptibility to ethanol-induced pancreatic dysfunction at the cellular level.

Methods

Animals.

Four-six week old C57BL/6J male mice (Jackson Laboratory) were used for in vivo experiments and 6-8-week-old mice were used for the preparation of pancreatic acini. At four weeks of age animals were maintained on a normal mouse chow diet [Purina 5053 (4 Kcal/g, phosphate 0.33%)] or a low phosphate diet [Envigo Teklad Custom Diet TD.140659 (3.8 Kcal/g, phosphate 0.02%)] for 2 weeks prior to experimental use and were kept on the low phosphate diet for the duration of the in vivo studies.

Mice were housed in a 12:12 hour light-dark cycle and given water and food ad libitum. Studies were approved by the Institutional Animal Care and Use Committee of Duke University.

In vivo experiments.

Orogastric gavage was performed using standard technique ⁹ with the following specifications. A weight-controlled volume (10 μL/g) of experimental solution was administered to each mouse using a 1.5-inch, 20-gauge animal feeding needle (Thermo Fisher, catalog number 14-153-200). Animals were not anesthetized during the procedure and were observed for 15 minutes post-gavage to ensure that there were no immediate complications.

Mice maintained on normal (ND) and low phosphate diets (LPD) were subjected to daily orogastric gavage for 5 days. Control animals received deionized water. Experimental animals received ethanol at a dose of 1.43 or 2.86 g/kg or a combined solution of ethanol (2.86 g/kg) and Na₂HPO₄ (0.3 mmol/kg). A 2.86 g/kg dose of ethanol reliably induces intoxication and models moderate binge drinking ²⁴ which was confirmed by measurement of blood alcohol levels 1 hour after ethanol gavage. Oral gavage was performed at 24-hour intervals and experimental solutions were prepared fresh each day to prevent concentration changes from evaporation. A total of 5 doses was given to each animal.

The in vivo experiments concluded 24 hours after the final orogastric gavage. Blood from experimental animals was collected via decapitation for serum amylase and lipase measurements. The entire pancreas was removed and weighed and pancreatic edema was assessed by changes in the ratio of pancreas weight (mg) to total body weight (g). The gastric and duodenal lobes were isolated for histological and biochemical assessments as previously described ²⁵. Lung sections were also isolated. Tissues for histological assessment were preserved in 10% formalin. Tissues used for biochemical assessments were stored at −80°C.

Assays.

Serum phosphate was measured using the Phosphate Colorimetric Assay Kit (Sigma, #MAK030) and tissue phosphate was measured using the Phosphate Assay Kit (Abcam, #ab65622) according to the manufacturer’s instructions. Serum alcohol and serum lipase were measured using colorimetric assay kits (Abcam, #ab65343 and ab 102524, respectively). Serum amylase was measured as previously described ²⁵. Pancreatic myeloperoxidase (MPO) was measured according to the previously described protocol ²⁶.

Histologic grading.

Pancreatic sections were embedded in paraffin, sectioned at 5 μm, stained with hematoxylin and eosin (H&E), and coded for examination by an investigator blinded to the experimental design. The severity of pancreatitis was graded using previously described scoring criteria and included measures for pancreatic edema, inflammatory cell infiltration, and necrosis ²⁷. Formalin fixed lung tissues were embedded in paraffin and 5 μm tissue sections were stained with H&E. The severity of lung damage was graded by considering four pathological categories (intra-alveolar edema, intra-alveolar hemorrhage, capillary congestion, and neutrophil infiltration) and each category was scored on a scale of 0 (no injury) to 4 (maximum injury). The total score was the composite of the four categories on a scale of 0 to 16 ²⁸.

In vitro studies on isolated pancreatic acini.

Pancreatic acini were isolated from 1-3 mice maintained on ND or LPD as previously described ^{25, 29}. Acini were incubated in variable concentrations of ethanol (0-50 mM) in HEPES buffer (composition [in mM]: 140 NaCl, 4 KCl, 1 MgCl₂, 2 CaCl₂, 10 glucose and 10 HEPES, pH 7.4) containing 0 mM, 2 mM or 5 mM Na₂HPO₄ for one hour prior to assessment of biochemical markers of cellular dysfunction.

Acini were plated on Matrigel-coated 24-well plates and incubated with ethanol (0-50 mM) and 0-5 mM Na₂HPO₄ for 1 hour at 37°C. Lactate dehydrogenase (LDH) release was measured by the CytoTox96 Non-Radioactive Cytotoxicity Assay Kit (Promega, #G1780), trypsin activity was measured using the Trypsin Assay Kit (Sigma, # MAK290), and ATP content was measured uisng the ATP Assay Kit (Sigma, # MAK190) according to the manufacturers’ instructions.

Mitochondrial activity was measured by a fluorometric method with the TMRE-Mitochondrial Membrane Potential Assay Kit (Abcam, # ab113852) ³⁰. Acini were incubated in Eppendorf tubes with ethanol (0-50 mM) and 0-5 mM Na₂HPO₄ for 1 hour at 37°C. Acini were incubated with 200 nM TMRE for 30 minutes at 37°C and 20 μM FCCP was used as positive control per the manufacturer’s instructions for microplate assay-based experiments. Acini were washed in HEPES buffer and plated on a black 96-well plate. Fluorescence intensity was assessed at Ex/Em 535 nm/590 nm.

Ca²⁺ imaging.

Live-cell calcium imaging was performed on pancreatic acini from ND or LPD mice as previously described with the following modifications ²⁵. HEPES buffer (with 0 or 2 mM Ca²⁺) containing 1 mM or 5 mM Na₂HPO₄ was used during imaging. A Zeiss Axio observer Z1 with a 20x objective was used to capture images at 400 ms intervals. The captured images were analyzed with MetaMorph software (Molecular Devices).

Statistical analysis.

Data were analyzed using GraphPad Prism 8.4.1. Results were reported as mean ± SEM. For two-group comparisons, mean differences were analyzed by 2-tailed Student’s t test. For multi-group comparisons, mean differences were analyzed by one-way analysis of variance with the Tukey’s multiple comparison post-test. For the comparison of survival curves, survival analysis was performed using the Log-rank (Mantel-Cox) test. P values < 0.05 were considered significant. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001.

Results

Low phosphate diet-induced hypophosphatemia increased susceptibility to alcohol-induced pancreatitis.

To induce hypophosphatemia, mice were placed on a LPD for 2 weeks ³¹ and continued on the LPD throughout the in vivo experiments. Serum phosphate levels measured at the end of the study were consistent with previously reported measurements in murine models ^{32, 33} (Fig. 1A). LPD and ND mice had similar weight and no significant weight loss was observed in either group over the 5 days of ethanol feeding. Animals maintained on the LPD had significantly lower serum phosphate levels compared to animals maintained on the normal diet (ND).

Figure 1. Low phosphate diet (LPD) increased susceptibility to alcohol-induced pancreatitis.

(A) LPD produced hypophosphatemia in mice. (B) Blood alcohol levels in normal diet (ND) and LPD mice 1 hour after ethanol (2.86 g/kg) administration. Mice maintained on ND or LPD were subjected to daily orogastric gavage with water (control) or ethanol for 5 days and (C) pancreatic edema, (D) serum amylase, (E) serum lipase (F) pancreatic MPO, and (G) histological scoring of pancreatic tissue samples were assessed 24h after the final gavage. Statistical analysis was performed using Student’s t test (serum phosphate) and ANOVA test with Tukey’s post-test. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001. (n = 5-8). (H) Representative images of pancreatic tissue stained with H&E. Scale bar = 200 μm.

To determine if hypophosphatemia sensitizes the pancreas to alcohol, mice fed the LPD were gavaged with ethanol (2.86 g/kg) or deionized water (control) for 5 days. Previously described murine models of binge drinking have utilized ethanol doses as high as 5-7 g/kg ^{9, 24}, but much lower doses have been shown to induce behavioral changes suggestive of alcohol intoxication ³⁴. Serum alcohol levels were similar (~150 mg/dL) in both ND and LPD mice 1 hour after ethanol gavage (Fig. 1B). This level of serum alcohol is reported in humans with moderate intoxication ³⁵. Therefore, this dose was chosen to examine the association between phosphate status and susceptibility to alcohol-induced pancreatitis. Pancreatic tissue was harvested 24 hours after the final orogastric gavage for biochemical and histologic evaluation.

Ethanol exposure in LPD-fed mice caused significant elevations in pancreatic edema, serum amylase and lipase, and pancreatic myeloperoxidase (MPO), an indicator of neutrophil infiltration (Fig. 1C–F). Significant histological changes were observed in pancreatic tissue samples from the LPD-fed, ethanol-treated group (Fig. 1G). Notably, none of these parameters of pancreatitis was elevated in the ND-fed ethanol-exposed group. However, pancreatic MPO was slightly increased in LPD-fed mice that were not exposed to ethanol. While the other markers of pancreatitis were not elevated, this finding suggests that pancreatic tissue may be more sensitive to stress in general in the setting of hypophosphatemia. Considered together, these data indicate that hypophosphatemia increased susceptibility to acute alcohol-induced pancreatitis. To determine the effects of lower doses of alcohol in LPD mice, we treated ND- and LPD- fed mice with ethanol (1.43 g/kg) or deionized water (control) administered by orogastric gavage for 5 days. This alcohol dose did not cause significant pancreatic injury (Sup Fig. 1).

Phosphate supplementation restored euphosphatemia and reduced the severity of LPD-associated alcoholic pancreatitis.

Phosphate supplementation in LPD-fed mice was performed via orogastric gavage. Animals in the LPD + ethanol + Na₂HPO₄ group received a solution of Na₂HPO₄ dissolved in ethanol instead of separate orogastric gavages in order to minimize procedural stress. Serum phosphate levels were measured at the end of the study. Intragastric phosphate supplementation restored euphosphatemia in LPD-fed mice (Fig. 2A).

Figure 2. Phosphate supplementation ameliorated LPD-associated alcohol-induced pancreatitis.

(A) Oral phosphate supplementation restored euphosphatemia in mice maintained on LPD. Mice maintained on LPD were subjected to daily orogastric gavage with ethanol (2.86 g/kg) or ethanol and Na₂HPO₄ (0.3 mmol/kg) for 5 days and (B) serum amylase, (C) pancreatic edema, (D) pancreatic MPO, and (E) histological scoring of pancreatic tissue samples were assessed 24 hours after the final gavage. Statistical analysis was performed using Student’s t test. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001. (n = 7-9). (F) Representative images of pancreatic tissue stained with H&E. Scale bar = 200 μm. (G) Survival curve over 5-day orogastric gavage experiment (n = 6 for LPD, n = 15 for LPD + EtOH, n 14 for LPD + EtOH + Na₂HPO₄). Statistical analysis was performed using the logrank test.

To examine the importance of serum phosphate level on predisposition to and severity of acute alcohol-associated pancreatitis, we determined if LPD-associated alcohol-induced pancreatitis could be ameliorated by phosphate supplementation. Mice on the LPD received daily orogastric gavage with ethanol (2.86 g/kg) or a combined solution of ethanol and Na₂HPO₄ (0.3 mmol/kg) for 5 days. Phosphate supplementation in LPD-fed mice restored serum phosphate to levels similar to mice fed a ND. Pancreatic edema was unchanged by phosphate supplementation, however, the other parameters of pancreatitis, including serum amylase, pancreatic MPO and histological measures of pancreatitis were significantly decreased in LPD-fed, ethanol-exposed mice that received phosphate supplementation (Fig. 2B–F). These results demonstrate that the restoration of euphosphatemia reduced the severity of LPD-associated alcohol-induced pancreatitis.

Survival was also significantly decreased in mice that developed LPD-associated alcohol-induced pancreatitis (Fig. 2G). Mice were observed in the post-gavage period for any complications to ensure that deaths were not attributable to procedural trauma. While LPD-fed mice that did not receive ethanol survived the experiment, LPD-fed, ethanol-exposed mice were highly susceptible to the development of pancreatitis and experienced a high experimental death rate. Pancreatitis-associated mortality was significantly reduced by correction of hypophosphatemia.

Ethanol-induced acinar cell dysfunction was modified by phosphate levels.

In order to determine the effect of ethanol exposure and phosphate levels on acinar cell function, we conducted studies in vitro. We harvested pancreata from ND-fed and LPD-fed mice and isolated acinar cells using phosphate-free buffer solutions. We assessed acinar cell function at various concentrations of ethanol correlating with physiologically relevant doses in humans, with 20-50 mM ethanol representing midrange concentrations commonly found in binge drinking ^{36, 37}. LDH release—a measure of cytotoxicity—was significantly increased in acini from mice fed the LPD and exposed to 50 mM ethanol (Fig. 3A). This group also demonstrated significantly increased levels of trypsin activity (Fig. 3B). Notably, the same effects were not seen in pancreatic acini harvested from mice maintained on the ND that were exposed to the same concentration of ethanol. These data demonstrate that low phosphate levels exacerbated ethanol-induced acinar cell dysfunction.

Figure 3. Ethanol-induced acinar cell dysfunction was modified by phosphate levels.

Acini from mice maintained on a ND or LPD were incubated with 0-50 mM ethanol. (A) LDH release and (B) trypsin activity were assessed. Acini from LPD-fed mice were then prepared in phosphate-free buffer and treated with 50 mM ethanol and 0-5 mM Na₂HPO₄ (C-D) Fold change represents the results of test groups compared to the control condition (no addition of ethanol or phosphate). Statistical comparisons were made by ANOVA with Tukey’s post-test. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001. (n = 3).

We next investigated whether phosphate supplementation could preserve pancreatic acinar cell function. Pancreatic acini from LPD-fed mice were isolated in phosphate-free buffer and treated with 50 mM ethanol and 0-5 mM Na₂HPO₄. The addition of 5 mM Na₂HPO₄ significantly decreased LDH and trypsin activity (Fig. 3C–D) in pancreatic acini exposed to ethanol. These data indicate that the deleterious effects of ethanol are effectively prevented by restoration of normal phosphate levels.

Phosphate regulated basal intracellular calcium in pancreatic acini isolated from ND-fed mice.

Dysregulated intracellular calcium ([Ca²⁺]_i) is detrimental to the acinar cell and is an early step in the development of pancreatitis ¹⁹. To determine the effect of phosphate on acinar cell calcium homeostasis, we conducted live-cell calcium imaging of pancreatic acini isolated from ND-fed mice. We first assessed basal intracellular calcium stores by treating acini with 5 μM ionomycin which facilitates the transport of calcium ions across the plasma membrane and the release of calcium from intracellular stores ³⁸. Cells were imaged in Ca²⁺-free buffer to ensure that intracellular calcium recordings represented the release of intracellular calcium stores without a confounding influx of extracellular calcium. ND acini isolated in phosphate-free buffer had high basal levels of ionomycin-stimulated cytoplasmic calcium while acini isolated in 5 mM Na₂HPO₄ buffer had significantly lower basal ionomycin-stimulated intracellular calcium levels (Fig. 4A–B). Ionomycin stimulated an increase in Ca²⁺ fluorescence in acini from both groups, but a more sustained elevation in [Ca²⁺]_i was observed in acinar cells in the low phosphate condition (Fig. 4A). These data indicate that phosphate plays an important role in regulating intracellular calcium levels, a critical aspect of cellular homeostasis.

Figure 4. Phosphate regulated [Ca²⁺]_i in pancreatic acini from ND-fed mice.

Live-cell calcium imaging of pancreatic acini from ND-fed mice loaded with calcium 6-QF. Acini were prepared in buffer containing 0 mM or 5 mM phosphate (A) The relative fluorescence intensity (ΔF/F₀) of pancreatic acini in response to ionomycin (5 μM) in the absence of external Ca²⁺ with imaging buffer containing 0 mM or 5 mM phosphate. (B) Statistical analysis of ionomycin mediated peak calcium elevation (F_max/F₀) from 45 to 51 cells. F₀ and F_max are the baseline and peak intensity, respectively. (C) Representative traces for relative fluorescence intensity (ΔF/F₀) of calcium 6-QF–loaded cells are shown in response to ethanol (50 mM) in the presence of 2 mM external Ca²⁺ with imaging buffer containing 0 mM or 5 mM phosphate (D) The average peak [Ca²⁺]_i intensity of F_max/F₀ is shown from 34 cells. Statistical analyses were performed using 2-tailed Student’s t test. Data are shown as the mean ± SEM. ****p ≤ 0.0001.

To assess the effect of alcohol on pancreatic acinar cell calcium dynamics, acini from ND-fed mice were isolated in 0 mM or 5 mM Na₂HPO₄ buffer and exposed to 50 mM ethanol. The addition of ethanol to the buffer did not induce a significant [Ca²⁺]_i elevation. There was no significant difference observed between the two phosphate conditions (Fig. 4C–D). These data are similar to those previously reported in which small elevations in [Ca²⁺]_i were observed in response to continuous perfusion of 50-500 mM ethanol⁸. These results are consistent with our other in vitro studies, demonstrating that acini from mice fed a normal diet are not particularly susceptible to ethanol-induced acinar cell dysfunction.

Phosphate supplementation reduced cytosolic calcium levels in pancreatic acini isolated from LPD-fed mice.

Having discovered a potential link between phosphate and calcium homeostasis, we postulated that more significant dysregulation of calcium homeostasis may be observed in acini from LPD-fed mice. LPD acini were prepared in 1 mM or 5 mM Na₂HPO₄ buffer and treated with 5 μM ionomycin in the absence of external Ca²⁺. LPD acini in low phosphate buffer had significantly higher ionomycin-stimulated intracellular calcium levels than LPD acini in 5 mM Na₂HPO₄ buffer (Fig. 5A). Interestingly, phosphate supplementation had an effect during multiple stages of ionomycin-provoked intracellular Ca²⁺ dynamics (Fig. 5A–C). LPD acini isolated in 5 mM Na2HP04 buffer had significantly lower peak calcium elevation (Fig. 5B) and significantly lower sustained calcium elevation 10 minutes after the addition of ionomycin (Fig. 5C). These findings support the concept that chronic low phosphate conditions impact calcium dynamics by increasing cytoplasmic [Ca²⁺]_i and reducing the ability of the acinar cells to remove Ca²⁺ from the cytoplasm. Furthermore, phosphate supplementation effectively reversed these effects.

Figure 5. Phosphate supplementation reduced basal [Ca²⁺]_i in pancreatic acini isolated from LPD-fed mice.

Live-cell calcium imaging of pancreatic acini isolated from mice maintained on a LPD. (A) The traces represent the relative fluorescence intensity (ΔF/F₀) of pancreatic acini in response to ionomycin (5 μM) from three experiments in the absence of external Ca²⁺with imaging buffer containing 1 mM or 5 mM phosphate. (B and C) Statistical graphs represent the ionomycin-mediated initial peak calcium elevation (F_a/F₀) calculated from elapsed time between 2 to 4 minutes and sustained peak calcium elevation (F_b/F₀) calculated at 10 min. (D and E) Effects of ethanol (50 mM) on pancreatic acini at 1 mM and 5 mM phosphate. (D) The relative fluorescence intensity (ΔF/F₀) of calcium dye over time and (E) the average peak [Ca²⁺]_i intensity of pancreatic acini from 32 cells. Statistical analyses were performed using 2-tailed Student’s t test. Data are shown as the mean ± SEM. ****p ≤ 0.0001.

To determine if ethanol affected [Ca²⁺]_i in low phosphate conditions, we exposed LPD acini isolated in 1 mM or 5 mM Na₂HPO₄ buffer to 50 mM ethanol. As shown in Fig. 5D–E, ethanol did not evoke a significant intracellular Ca²⁺ rise in either of the experimental groups. Interestingly, fluorescence intensity tracings of LPD-low phosphate buffer acini demonstrated a gradual increase in [Ca²⁺]_i after ethanol exposure.

Phosphate level influenced acinar cell susceptibility to ethanol-induced injury by altering cellular mitochondrial function and ATP.

To further explore the mechanism underlying the effect of phosphate on ethanol-induced acinar cell dysfunction, we investigated cellular ATP content and mitochondrial function.

In order to determine the effects of phosphate and ethanol on mitochondrial function, we assessed mitochondrial membrane potential using the specific fluorescent dye, TMRE. TMRE fluorescence intensity was significantly decreased in LPD acini treated with 50 mM ethanol (Fig. 6C) indicating a reduction in mitochondrial membrane potential. 20 μM FCCP, a potent mitochondrial uncoupler, was used in each group to confirm the health of acinar cells used in the experiments and the background intensity after FCCP treatment was adjusted during analysis. These data point to mitochondrial dysfunction and altered cellular energetics as potential drivers of increased cell death and increased zymogen activation resulting in LPD-associated acute alcohol-induced pancreatitis. Ethanol-induced decreases in mitochondrial function were not observed in ND-derived acini (Fig. 6C). These findings suggest that chronic phosphate depletion may significantly affect mitochondrial function and thereby alter cellular energetics in pancreatic acinar cells. To address this possibility, we examined the effects of ethanol on ATP levels in acini under ND and LPD conditions (Fig. 6E) and observed that mitochondrial dysfunction was associated with lower intracellular ATP.

Figure 6. Phosphate level influenced acinar cell susceptibility to ethanol-induced injury by altering mitochondrial function and cellular ATP.

(A) Serum and (B) pancreas tissue phosphate levels in mice fed ND or LPD and treated with ethanol. (C and D) Mitochondrial membrane potential (TMRE fluorescence). Fold change represents the results of test groups compared to the control condition (no addition of test agents). Acini from mice maintained on a ND or a LPD were prepared in phosphate-free buffer and treated with 0-50 mM ethanol. (E and F) Total cellular ATP content. Acini from LPD-fed mice were isolated in phosphate-free buffer and treated with 50 mM ethanol and 0-5 mM supplemental Na₂HPO₄. Statistical analysis was performed using ANOVA test with Tukey’s post-test. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001. (n = 3).

This concept is further reinforced by the effects of phosphate supplementation on LPD-derived, ethanol-exposed acinar cell function. Ethanol administration did not affect serum phosphate levels in mice fed a ND or the LPD (Fig. 6A). However, ethanol slightly reduced pancreatic tissue phosphate levels in mice fed the LPD (Fig. 6B). Ethanol dose-dependently caused mitochondrial depolarization (Fig. 6C) that was reversed by phosphate supplementation (Fig. 6D). Thus, normal acinar cell mitochondrial activity was restored by phosphate. This was manifested by the reversal of ethanol-induced lowering of acinar cell ATP levels by phosphate supplementation in mice fed the LPD (Fig. 6E and 6F). These results indicate that phosphate has a significant protective effect on pancreatic metabolism at the cellular level.

Discussion

Alcohol abuse is one of the leading causes of acute pancreatitis, accounting for 20-30% of cases ³⁹. However, the factors influencing individual susceptibility to acute alcohol-induced pancreatitis are not well understood. Up to 30% of alcoholic patients admitted to the hospital have hypophosphatemia (serum phosphate <2.4 mg/dL) ⁴⁰. The hypophosphatemia is commonly caused by malnutrition, however, alcohol also impairs phosphate absorption in the intestine ⁴¹. Following hospitalization, refeeding may exacerbate hypophosphatemia ⁴² and alcohol withdrawal may create respiratory alkalosis which can also exacerbate hypophosphatemia. Therefore, several factors render alcoholic patients susceptible to hypophosphatemia. In the current study, we asked if serum phosphate levels could play a role in predisposition to and protection from acute alcohol-induced pancreatitis.

A fundamental pathophysiologic change that initiates pancreatitis is a sustained elevation in [Ca²⁺] ^{19, 43}. Much experimental evidence supports the concept that circulating ethanol and fatty acids are taken up by pancreatic acinar cells and metabolized to fatty acid ethyl esters by a nonoxidative pathway. Fatty acid ethyl esters then accumulate in mitochondria where they are converted to fatty acids that affect cellular calcium homeostasis and induce cytoplasmic calcium overload ⁴⁴. This cytoplasmic calcium overload leads to mitochondrial failure and progressive ATP depletion ^{22, 45–47}. Insufficient cellular ATP production leads to a cycle in which sarcoendoplasmic reticulum Ca²⁺-ATPase (SERCA) and plasma-membrane Ca²⁺-ATPase (PMCA) are unable to remove excess Ca²⁺ ions from the cytoplasm ^{48, 49} thus reinforcing calcium overload. As this pathophysiologic cycle accelerates, intracellular enzyme activation, vacuolization and cellular necrosis occur, producing pancreatitis ²⁰. Significantly, ATP repletion has been shown to reduce pancreatic acinar dysfunction ^{45, 50}.

Our findings suggest that phosphate status has a significant effect on the susceptibility of pancreatic acinar cells to ethanol-induced cellular dysfunction, likely through exacerbation of the pathophysiologic events described above. These results also indicate that disordered cellular energetics are a key abnormality in low phosphate-associated sensitization to ethanol-induced cellular injury, leading to gradual increases in cytosolic Ca²⁺ and elevations in markers of cell death and dysfunction after ethanol exposure.

Compared to acini from euphosphatemic mice, acini with chronic exposure to a low phosphate environment demonstrated higher rates of cell death and increased trypsin activity when treated with ethanol. It may be that combining hypophosphatemia with ethanol administration reduces the availability of phosphate in the acinar cell mitochondria below the concentration needed to synthesize sufficient ATP to reduce cytoplasmic calcium overload via SERCA and PMCA ¹⁹.

Phosphate is both a substrate for ATP generation and a regulator of mitochondrial oxidative phosphorylation ^{51, 52}. We observed decreases in both total cellular ATP content and mitochondrial function in ethanol-exposed acini maintained in low phosphate conditions. Our results suggest that disordered cellular energetics are the key abnormality in low phosphate-associated sensitization to ethanol-induced cellular injury. The finding that phosphate supplementation protected cells against these changes demonstrates that phosphate status is an important factor in modifying susceptibility of pancreatic acinar cells to alcohol-associated injury. The concept that acute pancreatitis is accompanied by reduced cellular energy production is supported by the recent observation that circulating leukocytes from human patients with acute pancreatitis exhibit reduced mitochondrial function and diminished bioenergetics ⁵³. Serum phosphate levels were not measured in these patients.

The hypothesis that serum phosphate levels modify the risk of developing acute alcohol-induced pancreatitis was further substantiated by our in vivo studies. In this new model of alcoholic pancreatitis, LPD-associated pancreatitis was reliably produced by acute ethanol treatment without the need for an additional trigger or chronic alcohol exposure. A hypophosphatemic state predisposed animals to ethanol-induced pancreatitis. Notably, ND-fed mice (euphosphatemic state) and LPD-fed mice that received phosphate supplementation (thus restoring the euphosphatemic state) did not develop acute pancreatitis when treated with ethanol. These findings suggest that the coexistence of acute ethanol exposure and low serum phosphate is a high-risk combination for the development of acute alcoholic pancreatitis.

We observed mortality in hypophosphatemic, ethanol-treated mice but not LPD-fed mice suggesting that the physiological effects of hypophosphatemia alone were not responsible for increased mortality. Alcohol-treated mice on the LPD had more extensive lung involvement that accompanied more severe pancreatitis (Sup-Fig. 2). Although it was not possible to determine the exact cause of death in these animals, it is possible that respiratory compromise may have been a complicating factor. In humans, the presence of hypophosphatemia during critical medical illness is an indicator of a poor prognosis ⁵⁴, particularly if this imbalance reflects a persistent deficiency rather than an acute alteration ⁵⁵. Thus, the lower survival rate of LPD-fed, ethanol-treated mice can be reasonably attributed to severe pancreatic disease. Nevertheless, both the elevation in markers of pancreatitis and the increased mortality rate observed in LPD-fed, ethanol-treated mice support the conclusion that hypophosphatemia is an important risk factor for pancreatitis.

Although cytoplasmic calcium overload is an essential feature of experimental alcoholic pancreatitis, ethanol alone does not cause calcium overload in mouse pancreatic acinar cells ^{8, 44}. Thus, a reasonable conclusion is that fatty acids must circulate in the blood at sufficient concentrations to cause pancreatitis in human alcoholic pancreatitis by reaching a threshold concentration in the acinar cell when combined with ethanol by a nonoxidative pathway producing fatty acid ethyl esters ⁸. This is supported by the fact that hyperlipidemia, high fat diet, and fasting all increase fatty acids as metabolic substrates ^{56, 57} and that long-chain fatty acid infusion causes acute pancreatitis and significant ATP depletion ⁵⁸. In addition, administration of ethanol together with a high fat diet causes experimental pancreatitis ^{59, 60}. Furthermore, various combinations of ethanol and a high fat diet worsen both caerulein- and bile acid-induced experimental acute pancreatitis ^61–63. Finally, the excessive ethanol consumption that often occurs in human alcoholic pancreatitis may be accompanied by decreased nutritious food intake leading to malnourishment and an increase in fatty acid metabolism.

Malnutrition in alcoholic patients may contribute to other metabolic disturbances including magnesium and potassium deficiency ^{64, 65}. Magnesium deficiency is particularly interesting because it has the potential to exacerbate experimental pancreatitis and magnesium administration in vitro reduces cytosolic calcium and intracellular protease activation ⁶⁶. Thus, several metabolic derangements may increase the risk of alcohol-induced pancreatitis.

The present findings suggest that in the presence of euphosphatemia, circulating ethanol does not cause enough calcium overload in acinar cells to suppress mitochondrial ATP production below the levels needed to prevent cellular damage. However, these same levels of circulating ethanol in the presence of hypophosphatemia, in which the phosphate substrate for ATP synthesis is reduced below the level needed for normal ATP production, are sufficient to cause acute pancreatitis. There is substantial evidence that dysregulation of energy production in the pancreas is a key component of acute pancreatitis. For example, since ATP supplementation ameliorates Ca²⁺ clearance from the cytosol and prevents damage in acinar cells ⁴⁵, it may be possible to treat acute pancreatitis by boosting cellular energy provision. It has been demonstrated that enteral nutrition is beneficial compared to a nil per os diet not only in severe, but also in mild and moderate acute pancreatitis ⁶⁷. In addition, a recent study showed that enhancing cellular energy production by administering galactose ameliorated acute pancreatitis ⁶⁸. It will be important to determine if the same mechanisms are involved in human alcoholic pancreatitis.

What you need to know.

Background and context

Even though alcoholic pancreatitis is common in humans, it has been difficult to produce pancreatitis in animals with alcohol.

New findings

The current study demonstrates that hypophosphatemia predisposes mice to alcohol-induced pancreatitis and that phosphate status is critical for maintaining pancreatic acinar cell energy stores.

Limitations

It remains unknown if hypophosphatemia underlies human alcohol-associated pancreatitis and if phosphate administration reverses other causes of acute pancreatitis.

Impact

Low phosphate diet + ethanol provides a new model for studying alcohol-associated pancreatitis. These findings suggest that phosphate may ameliorate the adverse effects of ethanol on the pancreas.

Abbreviations:

ATP

adenosine triphosphate

[Ca²⁺]_i

intracellular calcium concentration

EtOH

ethanol

FCCP

carbonyl cyanide-4-(trifluoromethoxy) phenylhydrazone

H&E

hematoxylin and eosin

LDH

lactate dehydrogenase

LPD

low phosphate diet

MPO

myeloperoxidase

ND

normal diet

PMCA

plasma-membrane Ca²⁺-ATPase

SERCA

sarcoendoplasmic reticulum Ca²⁺-ATPase

TMRE

tetramethylrhodamine, ethyl ester