Defective efferocytosis and dysregulated sphingomyelin metabolism are involved in the fibrotic response in CP.
To investigate whether defective efferocytosis exist during CP progression, PRSS1Tg mice were administered different treatment schedules of caerulein to induce CP as reported previously. Pancreatic acinar cell apoptosis was increased in the early stage of CP development, peaking in week 1, and then decreased over the remaining three weeks of the experiment (Fig. 1A). Furthermore, substantial collagen deposition and morphological changes in the pancreas were observed in CP tissue compared with 0-week from PRSS1Tg mice (Fig. S1A).
The major cells responsible for apoptotic cell recognition and clearance in vivo are macrophages (“professional phagocytes”) including M1 and M2a subtypes, which prompted us to investigate temporo-spatial changes in macrophage infiltrate. The expression level of F4/80, a key marker of macrophages, markedly increased at 1 week, but decreased during CP progression (Fig. 1A, S1E). It indicated that macrophage polarization changed concurrently. Flow cytometry analysis indicated that the proportion of pancreatic macrophages were achieved to the highest level at 1 week but decreased in the following stages. In addition, a significant increase in alpha-smooth muscle actin (α-SMA) expression over time (Fig. S1D) indicated that pancreatic parenchymal fibrosis increased after CP induction. M1 (CD86 and TLR2) and M2a (CD206 and ARG1) macrophage markers expression in CP tissues from PRSS1Tg mice were further assessed. Specifically, M1 macrophage marker expression was increased in CP tissues compared with 0-week from PRSS1Tg mice, but this expression did not significantly differ among week 1, week 2 and week 4. In contrast, M2a macrophage marker expression peaked at week 1 and then decreased gradually after model induction (Fig. 1C), the flow cytometric analysis indicated that the proportion of M2a macrophages reached the highest level at 1 week, but gradually reduced along with the aggravation of fibrosis (Fig. 1B). Additionally, the levels of the inflammatory factors interleukin (IL)-1 beta, IL-6, and tumour necrosis factor alpha (TNF-α) were significantly elevated in 1, 2, 4-week groups compared with 0-week group (Fig. S1B).
Because modified membrane lipids, especially sphingomyelin pathway, accounts for the most important apoptosis-specific “find-me” signals of efferocytosis[17, 18], we next examined 11 sphingomyelin-related metabolites and 7 selected phospholipid metabolites. In caerulein-treated PRSS1Tg mice at 0, 1, 2, 4-week, the levels of 4 sphingomyelin-related metabolites (n-acetyl-sphingosylphosphorylethanolamine, dihydrosphingosylphosphorylcholine, ceramide, and LPC) were significantly altered (Fig. 1D). The level of an important “find-me” signal-lysophosphatidylcholine (LPC), a phospholipid present at different locations in the plasma membrane, peaked in week 1 and then decreased over time after CP induction (Fig. 1D).
Collectively, above results showed that the quantity of macrophages, especially M2a macrophages which play an important role in engulfment of dying cells, decreased from week 1 to week 4 in PRSS1Tg CP model. Meanwhile, abundance of LPC, an important “find me” signal from apoptotic acinar cells, also decreased over time. These data indicate that defective efferocytosis and dysregulated sphingomyelin metabolism exist in CP and that changes in the LPC concentration produced via the sphingomyelin pathway are probably associated with the pathogenesis of CP.
Atp8b1 is downregulated in pancreatic acinar cells during CP development.
To clarify potential genes responsible for sphingomyelin metabolism in CP, epigenetic assays and screening are performed on pancreatic tissues of PRSS1Tg mice that were treated with caerulein for 4 weeks (Group B) and those treated with caerulein for 1 week (Group A). RNA-seq and H3K27me3 ChIP-seq results were used to compare Group A and Group B and 528 genes were found to be differentially expressed in both datasets (Fig. 2A), among which 32 differentially expressed genes were identified as downregulated by RNA-seq analysis (Fig. 2B). We then searched protein databases (UniProt and STRING) for gene products specifically related to sphingomyelin metabolism and identified Atp8b1 as a potential regulator that contributes to the dysregulation of sphingomyelin metabolism in CP pathogenesis (Fig. 2C). Atp8b1 was identified by histone 3 lysine 27 trimethylation (H3K27me3) ChIP-seq, RNA pol II ChIP-seq and RNA-seq analyses of pancreatic acinar cells from PRSS1Tg mice (Fig. 2D) qPCR (Fig. 2E) and Western blotting (Fig. 2F) analysis revealed that the increase in mRNA and protein expression of Atp8b1 in Group A was attenuated in Group B.
To confirm the association of Atp8b1 downregulation with CP, Atp8b1 expression in the pancreas was complemented using an Atp8b1-overexpressing adeno-associated virus (AAV; adAtp8b1) (Fig. S3A); qRT-PCR and Western blotting for Atp8b1 expression confirmed a high transduction efficiency (Fig. S3B, C). Notably, the caerulein-induced fibrosis and histologic changes in the pancreas in PRSS1Tg mice were alleviated upon restoration of Atp8b1 expression (Fig. 3A). Moreover, the levels of IL-1β, IL-6, and TNF-α were significantly were inhibited after Atp8b1 expression was restored (Fig. 3B). Furthermore, adAtp8b1-injected PRSS1Tg mice exhibited a higher F4/80 level (Fig. 3C) and LPC concentration (Fig. 3F) than negative control AAV-injected mice in pancreatic tissues, as well as increased expression of M2a macrophage markers (CD206 and ARG1) (Fig. 3D, E). These results suggest that Atp8b1 expression is downregulated in CP and that restoration of Atp8b1 expression can significantly improve CP outcomes.
Atp8b1 increases acinar LPC concentration, promoting M2a macrophage efferocytosis of apoptotic acinar cells in CP.
To further investigate the mechanism by which Atp8b1 is involved in CP pathogenesis, PRSS1Tg mice were treated with CLOs to deplete macrophages after injection of adAtp8b1 and then treated with caerulein to induce CP; inflammation severity was then assessed by histological examination. Substantial fibrosis, histologic changes, increased inflammatory factor expression and decreased M2a macrophage marker expression were observed in CLO-treated PRSS1Tg + adAtp8b1 mice compared with non-CLO-treated PRSS1Tg + adAtp8b1 mice (Fig. 4A, B). Additionally, LPC concentration was significantly elevated in CLO-treated PRSS1Tg + adAtp8b1 mice compared to PRSS1Tg + NC AAV mice but not significantly different between the CLO-treated PRSS1Tg + adAtp8b1 mice and non-CLO-treated PRSS1Tg + adAtp8b1 mice (Fig. 4C). These results reveal that overexpression of Atp8b1 increased the LPC concentration and reversed the impairment of efferocytosis and dysregulation of acinar sphingomyelin metabolism in PRSS1Tg CP mice, confirming Atp8b1 as protective for CP pathogenesis. In addition, macrophage-depleted PRSS1Tg mice failed to clear apoptotic acinar cells and developed fibrotic lesions histologically, regardless LPC abundance was elevated or not, which further reveal that macrophage-related efferocytosis is involved in CP pathogenesis.
To ascertain the role of LPC in CP development, PRSS1Tg mice were treated with the G2A receptor agonist commendamide, followed by assessment of H&E staining and Masson’s trichrome staining. The increases in substantial pancreatic impairment and collagen expression observed in PRSS1Tg CP mice were significantly ameliorated upon treated with commendamide (Fig. 4E). CD206 and ARG1 expression levels were increased in PRSS1Tg mice CP model in the commendamide-treated group compared with those in the untreated group (Fig. 4D, E), which indicate that activation of LPC receptor G2A relieves fibrotic response by triggering macrophage-related efferocytosis.
Collectively, these results indicate that Atp8b1 promotes efferocytosis through the binding of LPC to its receptor G2A.
Bhlha15 is a regulatory binding transcription factor of Atp8b1.
To identify upstream transcription factors binding to Atp8b1, ATAC-seq analysis of pancreatic acinar cells from PRSS1Tg CP mice was performed. H3K27me3 modification was increased at the Atp8b1 transcription start site (TSS) and Atp8b1 expression was downregulated in Group B compared to Group A (Fig. 5A). By predicting the binding of transcription factors to the open regions around the TSS using Jaspar, we identified 11 transcription factors (Fig. 5B). We then analysed the transcription factors with the UniProt database to identify those with functions strongly correlated with sphingomyelin metabolism, and thereby screened out 3 transcription factors (Fig. S4B); we ultimately identified Bhlha15 as a potential upstream transcription factor of Atp8b1. The mRNA and protein expression of Bhlha15 in the pancreatic tissue of PRSS1Tg mice was significantly downregulated in Group B compared with Group A (Fig. 5C, D).
To investigate the role of Bhlha15 in CP development, we used a Bhlha15-overexpressing AAV for pancreatic transduction in caerulein-treated PRSS1Tg mice and then assessed pathological changes by H&E staining and Masson’s trichrome staining. Inflammatory cell infiltration, collagen deposition and inflammatory cytokine levels were decreased in PRSS1Tg CP mice upon Bhlha15 overexpression (Fig. 6A, B), indicating disease amelioration. Furthermore, the mRNA and protein levels of Bhlha15 and Atp8b1 were significantly increased in Bhlha15-overexpressing PRSS1Tg CP mice compared with PRSS1Tg + NC AAV CP mice (Fig. 6C, D). The LPC concentration was also examined in these groups, revealing a significant increase in the Bhlha15 overexpression group (Fig. 6E). Moreover, the expression of M2a macrophage markers in PRSS1Tg CP mice was enhanced upon Bhlha15 overexpression (Fig. 6F).
Next, ChIP with quantitative PCR (ChIP-qPCR) (Fig. 6G) was conducted to verify the Jaspar prediction that two potential binding sites for Bhlha15 are located in the Atp8b1 promoter. The ChIP assay showed substantial increases in the binding of Bhlha15 to chromatin in the S2 and S3 regions of the Atp8b1 promoter (Atp8b1-S2 and Atp8b1-S3) in 293T cells but no changes in binding to the Atp8b1-S1 and Atp8b1-S4 regions (Fig. 6G). This result indicated that Bhlha15 can bind to the Atp8b1 promoter in CP, but only to the predicted sites in Atp8b1-S2 and Atp8b1-S3, not the predicted sites in Atp8b1-S1 and Atp8b1-S4, and the binding degree of Atp8b1-S2 is the highest. Then, to confirm the predicted binding site in Atp8b1-S2, we performed a luciferase assay, which indicated that Bhlha15 binds to the Atp8b1-S2 region and identified two binding sites in this region (Fig. 6H). In summary, we concluded that Bhlha15 regulates the transcriptional activity of Atp8b1 by binding to the Atp8b1 promoter.