Fibrosis reduces severity of acute-on-chronic pancreatitis in humans

Abstract

Background & aims: Acute pancreatitis (AP) and chronic pancreatitis (CP) share etiologies, but AP can be more severe and is associated with a higher rate of mortality. We investigated features of CP that protect against severe disease. The amount of intrapancreatic fat (IPF) is increased in obese patients and fibrosis is increased in patients with CP, so we studied whether fibrosis or fat regulate severity of AP attacks in patients with CP.

Methods: We reviewed records from the University of Pittsburgh Medical Center/Presbyterian Hospital Autopsy Database (1998-2008) for patients with a diagnosis of AP (n = 23), CP (n = 35), or both (AP-on-CP; n = 15). Pancreatic histology samples from these patients and 50 randomly selected controls (no pancreatic disease) were analyzed, and IPF data were correlated with computed tomography data. An adipocyte and acinar cell Transwell coculture system, with or without collagen type I, was used to study the effects of fibrosis on acinar-adipocyte interactions. We studied the effects of nonesterified fatty acids (NEFAs) and adipokines on acinar cells in culture.

Results: Levels of IPF were significantly higher in nonobese patients with CP than in nonobese controls. In patients with CP or AP-on-CP, areas of IPF were surrounded by significantly more fibrosis than in controls or patients with AP. Fat necrosis-associated peri-fat acinar necrosis (PFAN, indicated by NEFA spillage) contributed to most of the necrosis observed in samples from patients with AP; however, findings of peri-fat acinar necrosis and total necrosis were significantly lower in samples from patients with CP or AP-on-CP. Fibrosis appeared to wall off the fat necrosis and limit peri-fat acinar necrosis, reducing acinar necrosis. In vitro, collagen I limited the lipolytic flux between acinar cells and adipocytes and prevented increases in adipokines in the acinar compartment. This was associated with reduced acinar cell necrosis. However, NEFAs, but not adipokines, caused acinar cell necrosis.

Conclusions: Based on analysis of pancreatic samples from patients with CP, AP, or AP-on-CP and in vitro studies, fibrosis reduces the severity of acute exacerbations of CP by reducing lipolytic flux between adipocytes and acinar cells.

Keywords: AP; ATP; BMI; CP; CT; FN; IPF; Inflammation; Mechanism; NEFA; Necrosis; PFAN; PI; Pancreas; SAP; UFA; acute pancreatitis; adenosine triphosphate; body mass index; chronic pancreatitis; computed tomography; fat necrosis; intrapancreatic fat; nonesterified fatty acid; peri-fat acinar necrosis; propidium iodide; severe acute pancreatitis; unsaturated fatty acid.

Figure 1. %IPF increases with BMI in Controls, AP patients and AP-on-CP patients but not in CP

Scatter plot showing the association between IPF and BMI for (A) Controls (black) and CP patients (red) and (B) AP patients (black) and AP-on-CP patients (red). Lines represent the best-fit line from a linear regression analysis. Correlation analyses were calculated using the Spearman’s rho correlation coefficient. Results show moderate to high correlations in all groups except in CP patients. Figure C shows the estimated marginal means from the two-way ANOVA that included main effects for BMI and group, and their interaction (BMI × group). Rank data for %IPF (RIPF) was used for this analysis. Post-hoc comparisons were performed using the Sidak adjustment method. Adjusted P values are shown. The two-way ANOVA revealed significant main effects for group and BMI, and a significant interaction. Subsequent post-hoc tests showed that, for the obese group, there were no statistically significant differences in IPF between the disease groups; however, for the normal group, the CP group had significantly greater IPF when compared to Controls (P value shown in green). Post-hoc comparisons between normal (BMI < 30) and obese (BMI ≥ 30) groups are also shown (P values in black). Figures D and E show linear regression analysis between %IPF measured on histology [%IPF (histology)] and Hounsfield units on CT scan or by the CT thresholding method, respectively. Spearman’s correlation coefficient was computed.

Figure 2. Histological quantification of FN (A), PFAN (B), acinar necrosis (C) as percentage of total area

Box plots showing the mean (dashed line), the median (solid line), the 25^th and 75^th percentiles (2 boxes), the 10^th and 90^th percentile (whiskers) and the outliers (dots), comparing Controls, CP, AP and AP-on-CP patients for each of the above parameters show these to be significantly reduced in CP and AP-on-CP compared to AP patients. Significant differences were found between the groups (Kruskal-Wallis test). Post-hoc comparisons were performed using the Mann-Whitney test and adjustment for multiple comparisons was performed using the Dunn-Sidak adjustment method. Adjusted P values are shown.

Figure 3. Fibrotic area (A) and percentage of fat surrounded by fibrosis (B) are increased in a background of CP

Box plots comparing Controls, CP, AP and AP-on-CP patients for (A) fibrosis (as % total area) and (B) % fat surrounded by fibrosis. Significant differences were found between the groups (Kruskal-Wallis test). Post-hoc comparisons were performed using the Mann-Whitney test and adjustment for multiple comparisons was performed using the Dunn-Sidak adjustment method. Adjusted P values are shown. C: Serial pancreas histologic sections from Controls, CP, AP and AP-on-CP patients stained with H&E, Masson’s Trichrome (Trichrome for fibrosis-blue strands;) show fat confined by the fibrosis (arrows) in CP and AP-on-CP. However, fat in Controls and FN in AP has no surrounding fibrosis (arrowheads).

Figure 4. Morphologic analysis of pancreatic injury using serial histologic sections shows fibrosis walling off FN and restricting von Kossa staining to FN

Serial histologic sections from the pancreata of Controls (A-A2), CP (B-B2), AP, (C-C2) patients and those with AP-on-CP (D-D2) stained with H&E (A, B, C, D), Masson’s Trichrome (Trichrome for fibrosis-blue strands; A1,B1, C1, D1) and von Kossa (for saponified NEFA stained brown, A2, B2, C2, D2) show PFAN around FN (dotted shape, B), which is von Kossa positive (dotted shape B2) in AP, where like in Controls there is no fibrosis surrounding the fat (Arrow heads). However, in CP (B) and AP-on-CP (D) the fat, FN and von Kossa positive areas (B2, D2) are confined by the fibrosis (arrows, B1, D1) with preservation of the surrounding parenchyma. Controls (A-A2) do not have significant von Kossa positive areas (A2).

Figure 5. Collagen-I reduces flux of macromolecules between acinar cells and adipocytes

Resistin (A) and NEFA concentrations (B) in the upper compartment and Amylase (C) expressed as a % of total pancreatic acinar content, lipase activity (D) in the lower compartment at the end of five hours incubation under different conditions; i.e. acini alone in the upper compartment (Acini), adipocytes alone in the lower compartment (Adipocytes), Acini in the upper, and adipocytes in the lower compartment without collagen-I (T-A+A), and with collagen-I (C-A+A) coating the base of the transwell. P values for comparing two groups were calculated using the Mann-Whitney test.

Figure 6. Collagen-I reduces acinar cell death in co-culture

A: Images of acini (Ac) after 5 hours incubation in the upper compartment alone (Control) or with adipocytes in the lower compartment without collagen-I (T-A+A) or with collagen-I (C-A+A) coating the base of the transwell. The low power images (Left column) stained with trypan blue (TB) were taken at 4×. The images on the right are two 100x images showing TB and PI staining of identical areas representing each condition. Bar graphs showing percentage of acinar cell positive for PI staining (B), TB staining (C) and acinar cell ATP levels as % Control (29.2 ±1.8 pmol/μg protein) (D) after 5 hours incubation. P values for comparing two groups were calculated using the Mann-Whitney test.