Activin A Modulates Inflammation in Acute Pancreatitis and Strongly Predicts Severe Disease Independent of Body Mass Index

Abstract

Introduction: Acute pancreatitis (AP) is a healthcare challenge with considerable mortality. Treatment is limited to supportive care, highlighting the need to investigate disease drivers and prognostic markers. Activin A is an established mediator of inflammatory responses, and its serum levels correlate with AP severity. We hypothesized that activin A is independent of body mass index (BMI) and is a targetable promoter of the AP inflammatory response.

Methods: We assessed whether BMI and serum activin A levels are independent markers to determine disease severity in a cohort of patients with AP. To evaluate activin A inhibition as a therapeutic, we used a cerulein-induced murine model of AP and treated mice with activin A-specific neutralizing antibody or immunoglobulin G control, both before and during the development of AP. We measured the production and release of activin A by pancreas and macrophage cell lines and observed the activation of macrophages after activin A treatment.

Results: BMI and activin A independently predicted severe AP in patients. Inhibiting activin A in AP mice reduced disease severity and local immune cell infiltration. Inflammatory stimulation led to activin A production and release by pancreas cells but not by macrophages. Macrophages were activated by activin A, suggesting activin A might promote inflammation in the pancreas in response to injury.

Discussion: Activin A provides a promising therapeutic target to interrupt the cycle of inflammation and tissue damage in AP progression. Moreover, assessing activin A and BMI in patients on hospital admission could provide important predictive measures for screening patients likely to develop severe disease.

Figure 1.

Serum activin A level of patient is independent of BMI as a prognostic tool for AP disease severity. (a) ELISA of serum activin A levels (y axis) plotted against the patient BMI (x axis) distributed by disease severity group (control = inverted triangle; mild AP = square; moderate AP = triangle; severe AP = circle) n = 10 for each group with activin A estimated at admission and 24 hr and 48–72 hr after admission for each patient. (b) Serum activin A levels in those subjects with a BMI ≤30. GEE multivariate analysis confirmed that serum activin A level is predictive of disease severity for patients with low BMI (P < 0.0001). AP, acute pancreatitis; BMI, body mass index; GEE, generalized estimating equation.

Figure 2.

Inhibition of activin A in cerulein-induced AP murine model reduces disease severity and inflammation. (a) Schematic of treatment and cerulein injection schedule. Mice were pretreated with activin A-specific neutralizing antibody (Anti-Act) or IgG control 30 minutes before the start of cerulein injections. Mice were given 12 hourly injections of 50 μg/kg cerulein or equal volume PBS. Blood and tissue were collected 24 hr after the start of cerulein injections. (b) Stained sections of pancreatic tissue from mice with cerulein-induced AP. Top row, H&E staining; middle row, IHC staining of CD68; and bottom row, IHC staining of TGFβ. Representative images were taken at ×40 total magnification. (c) Quantification of pancreatic tissue staining. n = 5–10 mice per group. *P < 0.05; **P < 0.01; ***P < 0.001. AP, acute pancreatitis; IHC, immunohistochemical; PBS, phosphate-buffered saline.

Figure 3.

Inhibition of activin A treatment of cerulein-induced AP decreases disease severity but does not reduce pancreas regeneration. (a) Schematic of treatment and cerulein injection schedule. Mice (n = 3–5) were given 12 hourly injections of 50 μg/kg cerulein or equal volume PBS. At 7 hr of cerulein injection, mice were treated with activin A-specific neutralizing antibody (Anti-Act) or IgG control. Blood and tissue were collected 24 hr after the start of cerulein injections and at 7 d. (b) Hematoxylin/eosin stained pancreatic sections at ×10 magnification. (c) Edema score at 24 hr and 7 d of cerulein injection with and without Anti-Act. (d) Percentage of pancreatic necrosis, (e) pancreas:body weight ratio, (f) neutrophil infiltration score, and (g) AP severity score. **P < 0.01, ***P < 0.001. AP, acute pancreatitis; PBS, phosphate-buffered saline.

Figure 4.

Activin A is produced and released by pancreatic cell lines but not by macrophage cell lines. (a) qRT-PCR results of INHBA mRNA levels in hPSC and HPDE cells after 24-hr treatment with 10 ng/mL TNF. (b) Representative immunoblots of the βA subunit proform of activin A in hPSC and HPDE cells after 24-hr treatment with 10 ng/mL TNF. Loading control is α-tubulin for hPSC cells and β-actin for HPDE cells. (c) Densitometry quantification of immunoblots of at least 3 independent experiments. (d) ELISA of activin A in cell culture media of hPSC and HPDE cells after treatment with increasing concentrations of TNF or LPS for 24 hr. (f) ELISA of activin A in cell culture media of RAW264.7 cells treated with 10 ng/mL TNF or 10 μg/mL LPS for 24 hr. Activin A in the media was undetectable. (e) INHBA mRNA level in RAW264.7 cells after activation with 10 ng/mL TNF. *P < 0.05; **P < 0.01. HPDE, human ductal epithelial cells; hPSC, human pancreas stellate cells; qRT-PCR, quantitative real-time polymerase chain reaction.

Figure 5.

Activin A activates RAW264.7 macrophages. qRT-PCR results of TNF and IL1B mRNA expression in RAW264.7 macrophages after 24-hour treatment with 25 ng/mL activin A. *P < 0.05. qRT-PCR, quantitative real-time polymerase chain reaction.