PFKFB3 controls acinar IP3R-mediated Ca2+ overload to regulate acute pancreatitis severity

Acute pancreatitis (AP) is among the most common hospital gastrointestinal diagnoses; understanding the mechanisms underlying the severity of AP is critical for development of new treatment options for this disease. Here, we evaluate the biological function of phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 (PFKFB3) in AP pathogenesis in 2 independent genetically engineered mouse models of AP. PFKFB3 was elevated in AP and severe AP (SAP), and KO of Pfkfb3 abrogated the severity of alcoholic SAP (FAEE-SAP). Using a combination of genetic, pharmacological, and molecular studies, we defined the interaction of PFKFB3 with inositol 1,4,5-trisphosphate receptor (IP3R) as a key event mediating this phenomenon. Further analysis demonstrated that the interaction between PFKFB3 and IP3R promotes FAEE-SAP severity by altering intracellular calcium homeostasis in acinar cells. Together, our results support a PFKFB3-driven mechanism controlling AP pathobiology and define this enzyme as a therapeutic target to ameliorate the severity of this condition.


Abstract
Acute Pancreatitis (AP) is among the most common hospital gastrointestinal diagnosis; understanding the mechanisms underlying the severity of AP are critical for development of new treatment options for this disease.Here, we evaluate the biological function of phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 (PFKFB3) in AP pathogenesis in two independent genetically engineered mouse models of AP.PFKFB3 is elevated in AP and severe AP (SAP) and knockout of Pfkfb3 abrogates the severity of alcoholic SAP (FAEE-SAP).Using a combination of genetic, pharmacological, and molecular studies we define the interaction of PFKFB3 with inositol 1,4,5-trisphosphate receptor (IP3R) as a key event mediating this phenomenon.Further analysis demonstrated that the interaction between PFKFB3 and IP3R promotes FAEE-SAP severity by altering intracellular calcium homeostasis in acinar cells.Together our results support a PFKFB3-driven mechanism controlling AP pathobiology and define this enzyme as a therapeutic target to ameliorate the severity of this dismal condition.

Introduction
Acute pancreatitis (AP), represents the most common gastrointestinal inflammatory process requiring hospitalization and costs the U.S. healthcare system nearly 3 billion annually (1).Globally, the incidence of AP is about 34 per 100,000 a year, affecting females and males in equal proportion (2).Clinically, AP presents as severe abdominal pain and elevated serum amylase and/or lipase levels (3).In general, patients with mild presentations of AP recover fully in days to weeks, however, approximately 20-30% of patients develop severe or moderately severe AP, often resulting in multiple, persistent organ failure and local complications (4,5).
Thus, understanding the mechanisms contributing to AP severity could facilitate development of new treatment approaches to improve the outcomes of patients with this disease.
Alcohol-related AP is a common type of AP arising as a result of heavy alcohol consumption.In alcohol-related AP, fatty acid ethyl esters (FAEE) from nonoxidative ethanol metabolism cause pancreatic damage and inflammation (6)(7)(8)(9)(10).In this study, we provide evidence for the role of phosphofructo-2-kinase/fructose-2,6biphosphatase 3 (PFKFB3) as a regulator of severity in alcohol-related AP.PFKFB3, encoded by the PFKFB3 gene, plays a key role in maintaining glycolytic rates (11) by controlling the production of fructose-2,6-bisphosphate (F-2,6-BP) , an allosteric activator of phosphofructokinase-1 (PFK-1), which is the rate limiting enzyme of glycolysis (12,13).Here, we show that PFKFB3 expression is increased in AP and SAP.In mice, knockout of Pfkfb3 reduced serum amylase, histologic indicators of pancreatic injury and inflammation in a FAEE-SAP model induced by injection of palmitoleic acid (POA) and ethanol.Interestingly, this effect of PFKFB3 loss was not mediated by dysregulation of glycolysis but through IP3R-mediated Ca 2+ overload, which has been shown to be a central event in AP pathogenesis (14).Further analysis demonstrates the interaction between PFKFB3 and IP3R in pancreatic acinar cells promotes pathologic intracellular calcium signaling.In conclusion, our data indicate that PFKFB3 controls the severity AP model by affecting calcium levels in an IP3R dependent manner and define PFKFB3 as a promising therapeutic target to improve outcomes in alcohol-related AP patients.

Elevated PFKFB3 expression during AP and SAP progression
To define the genes involved in the regulation of AP progression, we analyzed the transcriptome of pancreata from vehicle control and caerulein-induced AP (CAE-AP) mice using bulk RNA-seq (Fig. S1A).Pfkfb3, a known regulator of glycolysis, was among the most upregulated genes in AP mice (Fig. 1A, Fig. S1B).Because of the central role of glycolysis as one of the main sources of ATP in acinar cells under physiological and disease conditions (15)(16)(17), we focused our studies on PFKFB3.
Induction of PFKFB3 transcripts and protein levels were confirmed by RT-qPCR and Western blotting (WB), respectively (Fig. 1B, Fig. S1C).To confirm the biological meaning of these findings, we examined the expression of PFKFB3 in an in vitro system and in two additional AP models, a FAEE-SAP (Fig. S1D) and sodium taurocholate biliary pancreatitis (NaTc-SAP) (Fig. S1E).A substantial induction of Pfkfb3 mRNA and protein was observed during FAEE-SAP (Fig. 1C) and NaTc-SAP (Fig. 1D).Immunohistochemistry (IHC) and immunofluorescence (IF) demonstrated induction of PFKFB3 in the acinar compartment of all three models (Fig. 1E-J, Fig. S1F-G, Fig. S2A-C), as well as in AP patients (Fig. 1K-L, Fig. S2D).These assays also showed that PFKFB3 can localize to both the cytoplasm and nucleus of exocrine cells during AP development.In vitro, TNFA, a pro-inflammatory factor implicated in AP (18), and CoCl2, an inducer of hypoxia -a known component of AP pathogenesis - (19,20), induced PFKFB3 in acinar cell lines (Fig. 1M, Fig. S2E).In summary, our finding identified and confirmed Pfkfb3 as one of the most elevated genes in AP models and suggest a role for this molecule in AP pathogenesis.

Loss PFKFB3 alleviates FAEE-SAP severity
To interrogate the role of Pfkfb3 in AP pathobiology we used CAAG-Cre-ER TM (Cre-Estrogen Receptor, Tamoxifen)-driven inducible Pfkfb3 knockout (iKO) mice in the FAEE-SAP model.No differences in pancreatic/body weight ratio or genders were found in healthy iKO and WT animals (Fig. S3A-G).However, after SAP induction, serum amylase and MCP1 levels were lower in iKO group compared to WT mice (Fig. 2A-B).RT-qPCR showed decreased expression of Il6, Mcp1 and Cxcl2 in the iKO relative to WT mice (Fig. 2C).iKO pancreata had decreased proportional pancreatic necrosis in comparison to WT counterparts in the FAEE-SAP model.Further, iKO showed reduced pancreatic neutrophil infiltration and apoptosis as determined by myeloperoxidase (MPO) and TUNEL staining, respectively (Fig. 2D-G, Fig. S4A-D).Next, we evaluate the effect of Pfkfb3 knockout in a necrotizing pancreatitis model using retrograde injection of sodium taurocholate (NaTc).There were no marked differences in the serum amylase and MCP1 levels or expression of inflammatory cytokines in SAP progression between WT and iKO mice (Fig. S5A-C).H&E and MPO stains showed the same degree of injury and acute inflammation among iKO and WT mice, (Fig. S5D-F) and survival rates were not different between the two groups (Fig. S5G).
To confirm the results of Pfkfb3 inactivation from the iKO mice, we developed a pancreatic specific knockout of Pfkfb3 driven by Cre expressed under the pdx1 promoter (cKO).In the absence of AP, cKO mice have no macroscopic or microscopic differences from the WT animals (Fig. S3H-K).Similar to the iKO animals, cKO mice had lower levels of serum amylase and MCP1 in the FAEE-SAP model of AP (Fig. 3A-B).Il6 and Cxcl2 expression were also lower in the pancreata of the cKO group (Fig. 3C).The percentage of necrotic area, MPO-positive cells and TUNEL positive area were reduced in cKO mice compare to WT mice (Fig. 3D-G, Fig. S6A-D).All together, these findings support a protective role for Pfkfb3 knockout in the regulation of the pathogenesis of pancreatitis.Interestingly, this effect was observed in the FAEE-SAP model thus defining a new etiology-specific pathway driving AP.

PFKFB3 modulates SAP progression independent of glycolysis
Next, we sought to determine if the effect of PFKFB3 in FAEE-SAP severity is mediated by dysregulation of the glycolytic pathway.To this end, we administered fructose-1,6-bisphosphate (FBP), the glycolytic intermediate immediately downstream of PFKFB3-mediated activation of PFK-1, to WT mice with FAEE-SAP to determine if enhanced glycolytic flux affects AP severity (Fig. S7A).Supplementation of FBP protected acinar cells from necrosis and reduced the infiltration of neutrophils and the number of TUNEL-positive cells (Fig. 4A-D, Fig. S7B-D, Fig. S8A).Interestingly, FBP supplement also decreased the serum amylase and MCP1 in serum (Fig. S8B-C), and pancreatic Il6 expression (Fig. S8D).
To further test if the effects of Pfkfb3 knockout on AP severity in the FAEE-SAP model are mediated by alteration of glycolysis, mice were treated with intraperitoneal injection of 3PO (21).3PO inhibits glycolysis by competing with F-6-P instead of binding with PFKFB3 (22-24) (Fig. S7A).Serum amylase was not different between experimental groups; however, the group treated with 3PO had higher serum lipase compared to DMSO-treated mice (Fig. S9A-B).In mice pretreated with 3PO followed by POA and ethanol, MCP1 serum in protein and pancreatic tissue in mRNA were elevated in compared to the DMSO pretreatment group (Fig. S9C-D).Also, 3PO pretreated animals had increased acinar cell necrosis, neutrophil infiltration, and TUNEL-positive cells (Fig. 4E-H, Fig. S9E-G, Fig. S10A).
Overall, exogenous supplement of FBP abrogated rather than aggravated the severity of AP, and glycolysis inhibition by 3PO did not phenocopy Pfkfb3 genetic inactivation.Thus, allosteric modulation of glycolysis by PFKFB3 does not account for the reduced severity of AP observed in Pfkfb3 knockout mice.

PFKFB3 facilitates Ca 2+ signaling activation in FAEE-SAP
To define the mechanisms underlying the role Pfkfb3 in AP pathogenesis, we performed RNA-seq in WT and cKO mice treated with POA and ethanol.Table S1 shows the differentially expressed genes (DEGs) in cKO mice.KEGG analysis revealed the presence of several pathways related with pancreatic function and inflammatory response, such as TNF signaling pathway and cytokine-cytokine receptor interaction (Fig. 5A).Similarly, GO analysis identified a number of biological processes including inflammatory response, acute-phase response, and cell chemotaxis (Fig. 5B).Heatmaps also show genes participating in the pancreatic secretion, protein digestion and absorption and inflammatory pathways (Fig. S11A).
A subset of the DEGs were validated by RT-qPCR (Fig. S11B).
All the aforementioned pathways are associated with calcium signaling, and this particularly important as the dysregulation of intracellular calcium homeostasis is early intra-acinar event in AP (10,25,26).Further in silico analysis revealed that cKO mice have a transcriptional signature suggestive of reduce calcium signaling activity compared with WT mice (Fig. 5C, D).Furthermore, ultra-high speed twophoton microscopy demonstrated that mice pretreated with KAN0438757, a PFKFB3 kinase inhibitor (24) (Fig. S11C-D), had reduced intracellular free calcium following POA or POAEE treatment; cKO mice had an even larger decrease in calcium content (Fig. 5E, F).Together, these results support the dysregulation of calcium homeostasis as a central event mediating the effect of PFKFB3 in alcohol-related AP (Fig. S11E).

PFKFB3 elevates intracellular Ca 2+ level by interacting and activating IP3R
Next, we sought to determine the mechanism by which PFKFB3 is involved in regulating cytosolic calcium flux.Gene set enrichment analysis of the RNA-seq data derived from cKO and WT animals did not demonstrate enrichment of genes involved in calcium signaling pathways including Mitochondrial Calcium Ion Homeostasis, Calcium Ion Export, Voltage Gated Calcium Channel, and Ryanodine Sensitive Calcium Release Channel (Fig. S12A-E).Moreover, POA treatment did not augment PFKFB3 physically interact with RYR (Fig. S12F-G).In contrast to RYR, expression of IP3R and p-IP3R were decreased in cKO mice with FAEE-SAP compared to WT (Fig. 6A, Fig. S13A-E), while there was no difference in transcription levels (Fig. S14A).In vitro, POA-stimulated AR42J cells, inhibition of PFKFB3 reduced apoptosis, the expression and phosphorylation of IP3R (Fig. 6B), and the global free calcium level (Fig. 6C).Conversely, PFKFB3 overexpression in 266-6 cells, even at low doses, increased endogenous and exogenous expression of IP3R and its phosphorylation without altering IP3R transcription (Fig. S14B-D).Next, we evaluated the interaction between PFKFB3 and IP3R by Co-IP, PLA and IF assays.
PFKFB3 has two functional domains, an N-terminal kinase domain and a Cterminal phosphatase domain (27).The N-terminal domain catalyzes the synthesis of F-2,6-BP using F-6-P and ATP as substrates, and the C-terminal domain is mainly responsible for dephosphorylation of F-2,6-BP (28) (Fig. S14H).Further study of the interaction of PFKFB3 with IP3R indicated that is the kinase domain but not phosphatase domain binds to IP3R (Fig. 6H).Functionally, overexpression of PFKFB3 kinase domain, even at low doses, was sufficient to increase the level of IP3R and p-IP3R in 266-6 cells, and elevate the global calcium level in stimulated AR42J cells, while the C-terminal domain did not affect IP3R expression, phosphorylation, or global calcium level (Fig. 6I-J, Fig. S14I).
In summary, our data showed that knockout of Pfkfb3 abrogates tissue injury in alcohol-related AP by attenuating Ca 2+ overload in pancreatic acinar cells.

Discussion
AP is an inflammatory disease of gastrointestinal tract with high morbidity and mortality.Due to the complexity of pathogenesis, the identification of therapeutic agents for AP has been challenging.Our data uncovered a new, glycolysisindependent, role of PFKFB3 in controlling tissue injury in alcohol-related SAP.
Further, we defined an interaction between PFKFB3 and IP3R, a key early regulator of AP development and progression, and proposed that PFKFB3 promotes of severity of alcohol-related AP by upregulating intracellular calcium levels in an IP3Rdependent manner.
Under physiologic conditions, cytosolic Ca 2+ concentration ([Ca 2+ ]c), is maintained at low levels due to plasma membrane Ca 2+ transport ATPase (PMCA), sarco/endoplasmic reticulum (SR/ER) Ca 2+ -ATPase (SERCA) and Na + /Ca 2+ exchanger.In pancreatic acini, transient spikes in cytosolic calcium initiate zymogen exocytosis and maintain mitochondrial function (31).Sustained elevations of calcium under pathological stimuli, such as cholecystokinin hyperstimulation (32), oxidative (33) and non-oxidative (8) metabolites of ethanol and bile acids (26), can trigger AP.This pathological elevation of intracellular calcium in pancreatic acinar cells leads to opening of mitochondrial permeability transition pore (PTP) and subsequent loss of the mitochondrial membrane potential, resulting in ATP depletion (34).Mitigation of AP progression requires ATP-driven apoptosis to eliminate damaged cells (35).In the setting of AP-induced mitochondrial dysfunction, glycolysis is a protective factor in the progression of AP (15)(16)(17).In our study, the protective effect of FBP and deleterious effect of 3PO re-confirmed the protective role of glycolysis in AP progression.
In alcohol-related AP, FAEE leads to the superabundant release of calcium from the ER lumen through IP3R channels.Subsequently, loss of ER calcium concentration stimulates influx of extracellular calcium via calcium release-activated calcium channel protein 1 (ORAI1) (9,10,36,37).Here we demonstrate that the calcium source in FAEE-SAP is mediated by PFKFB3 in an IP3R-dependent manner.In gallstone pancreatitis, or in the NaTc-SAP model, calcium overload is the combined result of extracellular calcium entry and efflux from ER (38)(39)(40), caused by activation of piezo type mechanosensitive ion channel component 1 (PIEZO1) mediated by transient elevations of pancreatic duct pressures (38,41) and bile acid (42) concurrently.The relatively complex sources of calcium in NaTc-SAP model may dilute the role of PFKFB3 in regulation of cytoplasmic calcium such that loss of PFKFB3-regulated calcium is insufficient to abrogate tissue injury in iKO mice.
Our data showed that the kinase domain of PFKFB3 binds to the CD domain of IP3R.Overexpression of PFKFB3 in acinar cells augments IP3R protein expression without influencing its transcription, and promotes the phosphorylation of IP3R.
Consistently, in cKO mice, knockout of PFKFB3 diminished IP3R expression and phosphorylation.Moreover, the PFKFB3 and IP3R interaction may explain why homozygous knockout of Pfkfb3 is embryonic lethal.Knockout of Pfkfb3 and its interaction with IP3R has the potential to disrupt the spatio-temporal Ca 2+ signals responsible for multitude processes including embryonic pattern formation, cell differentiation and proliferation, transcription factor activation (43).This is a supplement to the previous theory, but it needs to be further tested.
Technically, we used ultra-fast two-photon in vivo imaging to perform ultrafast deep synchronous fluorescence imaging to achieve in-situ dynamic calcium measurements in acinar cells.In situ monitoring avoids the over-activation of primary acinar cells during the separation process and the unstable signal acquisition resultant of unstable adhesions of isolated acinar cells.Critically, intraductal injection of Fluo-4 for calcium ion staining did not activate acinar cells and downstream calcium ion signaling pathways, such as PIEZO1 and Trypsinogen channel (Fig. S15A-B), proving the feasibility and value of this method.
In summary, we demonstrated the upregulation of PFKFB3 in AP progression and that knockout of Pfkfb3 reduced the severity of FAEE-SAP.Dysregulated calcium signaling is nexus of event for AP progression, and disruption of the PFKFB3-IP3R interaction resulted in a reduction in calcium signaling, which explains the protective effect of Pfkfb3 knockout in alcohol-related AP.Finally, our study highlights PFKFB3 as a promising therapeutic target in alcohol-related AP.

Sex as a biological variable
In our disease models, sex is not considered a biological variable.

Plasmids and transfection
Full length and truncated Pfkfb3 (Gene ID: 170768) and Ip3r (Gene ID: 16438) were amplified by PCR and cloned into pcDNA3.1 vector.Flag or HA tag was added into the N-terminal of the target genes.The indicated plasmids were transiently transfected into 266-6, HEK293T or AR42J cells at 60-70% confluence using Lipofectamine™ 3000 Transfection Reagent (ThermoFisher Scientific).Master mix of Lipofectamine 3000 reagent and plasmids were prepared in Opti-MEM.Mix was incubated for 15 minutes at room temperature, before adding to cell cultures.

AP and SAP mouse models
Before induction of experimental AP or SAP, mice were fasted for 12 hours.Experimental CAE-AP mice were administered caerulein (MCE, 100 g/kg) i.p. hourly for 10 hours as described previously (44,45).Mice injected with saline were used as control groups.Experimental FAEE-SAP mice, the mixture of 150 mg/kg POA (Sigma) and 1.35 g/kg ethanol were injected i.p. hourly for two hours as described previously (8).200 L saline was injected i.p. immediately prior to POA and ethanol injections to avoid potential local damage by ethanol to peritoneal organs at the injection site, and 0.1 mg/kg buprenorphine hydrochloride was simultaneously given with the first injection of POA and ethanol for animal care.Experimental NaTc-SAP mice were induced by pancreatic ductal retrograde infusion of sodium taurocholate (NaTc) as previously described (40).Briefly, mice were anaesthetized with pentobarbital, laparotomy was performed, and the duodenum was isolated and exposed such that the biliopancreatic duct could be clearly observed.A needle was passed through the duodenal wall directly opposite the papilla and placed within the duct.Once in position, the needle was fixed in place with a tied ligature.Retrograde infusion of 3% NaTc (2 µL/g, Solarbio) was performed at a flow rate of 10-12 µL/min.After infusion, the laparotomy was closed in two layers and the mice were returned to the cage.Sham-operated animals underwent laparotomy and duodenopancreatic manipulation without infusion.All experimental mice were sacrificed 24 hours post AP induction and the pancreatic tissue and blood specimen collected.The pancreas was divided into two parts, one was preserved in 4% paraformaldehyde for pathological sections, and the other was snap-frozen in liquid nitrogen for RNA and protein extraction.Blood kept at room temperature for 2 hours before being centrifuged (7,000 rpm) for 15 minutes at 4℃ to separate the serum.Serum was aliquoted to avoid repeated freeze-thaw cycles and stored in -20℃.

Measurement of amylase, lipase, and MCP1 in serum
All serum samples were properly diluted to ensure that the test values were within the detection range of the kit.Serum amylase was detected with α-Amylase Assay Kit (C016-1) purchased form Nanjing Jiancheng Bioengineering Institute (Nanjing, China).Diluted mouse serum samples were mixed with preheated substrate buffer, reacted at 37℃ for 7.5 min, and absorbance was measured at 660 nm after mixing with iodine solution.Serum lipase was detected with Lipase Activity Kit (A054-2) form Nanjing Jiancheng Bioengineering Institute (Nanjing, China).Diluted mouse serum samples were mixed with reagent 1, reacted at 37 ° C for 4 minutes, add reagent 2 and reacted for another 2 minutes.The absorbance was continuously measured for 2 minutes at 580 nm.The serum MCP1 was measured with a commercial ELISA kit from Biolegend (San Diego, CA).The capture antibody coated 96-well plates were blocked with assay diluent, the diluted serum samples were added and incubated at room temperature for 2 hours.Add detection antibody to capture antigen, and then avidin-HRP solution, TMB substrate solution and stop solution were added successively for chromogenic reaction, measured absorbance at 450nm.Absorbance values are converted according to standard curves or manufacture's recommendations.

RNA extraction and real-time quantitative PCR (RT-qPCR)
For RNA extraction, snap-frozen pancreatic tissues were ground to a powder, then transferred to tubes preloaded with TRIzol (ThermoFisher Scientific).The RNA of cell lines was extracted using the Total RNA Rapid Extraction Kit (Fastagen).
RNA was reverse transcribed to cDNA using PrimeScript TM RT Reagent Kit (Takara).
RT-qPCR was performed using TB Green Premix (TaKaRa).The expression levels of genes were normalized using the 2 -Ct cycle threshold method (46), using gene Rpl32 or B-Actin as the internal control.All primers used in this study were designed by the PrimerBank (47), and all the primer sequences are available upon request.

Protein extraction, immunoprecipitation, and immunoblot
Snap-frozen pancreas tissues were ground to a powder and lysed with RIPA lysis buffer (Beyotime, China) containing complete protease inhibitor (Roche), PhosSTOP (Roche) and PMSF (Beyotime).Protein concentrations were measured by BCA assay (Beyotime).The co-IP lysis buffer (50mM Tris-HCL, pH7.5, 150mM NaCl, 1mM EDTA, 1%(v/v) Triton X-100) supplemented with complete protease inhibitor was prepared for cell lysis.Lysates of transfected 266-6 cell or AR42J cell were subjected to immunoprecipitation with Anti-Flag Sepharose Beads (Millipore), Anti-HA Sepharose Beads (Millipore) or Anti-IgG Sepharose Bead (CST) for 12 hours, lysate of cells transfected with empty vector were prepared as controls.For immunoblot, proteins were separated by 10% SDS-PAGE and transferred to PVDF membranes (Merck &Millipore).After incubating with the indicated primary and secondary antibodies, antigen-antibody complexes were visualized by chemiluminescence (ECL, Millipore) on Bio-Rad ChemiDoc XRS+ system or directly scanned by Odyssey CLx Imaging System (LI-COR).It should be noted that phosphorylated protein antibody was incubated on the same membrane after elution of total protein antibody.In addition, samples in Fig. 6B were run at different times, and Fig. 7C was set up in parallel and run contemporaneously.

Histological examination, immunohistochemistry, and immunofluorescence
Pancreatic tissues fixed in 4% paraformaldehyde for >= 36 hours, were dehydrated and embedded in paraffin.Histologic sections were cut at 5 μm thickness.
For H&E staining, after dewaxing and hydration, the histologic sections were stained with hematoxylin and eosin (Sangon, China), and finally dehydrated, sealed for observation.More than 5 fields views of each H&E-stained section were microscopically evaluated to quantify the degree of necrosis expressed as the percentage of analyzed pancreatic parenchymal area composed of necrosis.For IHC staining, after dewaxing and hydration, the histologic sections were treated with 0.01M sodium citrate buffer for antigen repair.After blocking, MPO antibody (ZSGB-BIO, PV-9001) or PFKFB3 antibody (Abcam, ab181861) and specific secondary antibody were incubated successively, followed by DAB staining and nuclear staining.Finally, the histologic sections were dehydrated and sealed for observation.The whit-point correction was done according to the blank background.
To assess the degree of neutrophil infiltration, the number of MPO-positive cells were counted using ImageJ in several field views of MPO-stained sections.For TUNEL staining, after dewaxed and hydrated, the protein in histologic sections were dissolved in protease K solution, and the TUNEL test solution (Beyotime) was incubated for 1 hour under dark conditions.Finally mounted using ProLong Diamond Antifade Mountant with DAPI (Invitrogen) and wait for observation.In the sections stained with TUNEL, the percentage of positive area relative to total field view area was calculated.All necrotic areas, MPO-positive cells and TUNEL-positive areas were quantified by a reviewer agnostic to the treatment group.
For IF staining, pancreata were embedded in OCT compound (SAKURA); sections were cut at 8 μm thickness by cryotome.Frozen sections were blocked with 5% BSA and incubated overnight with the indicated antibodies.Stained sections were mounted with ProLong Diamond Antifade Mountant with DAPI (Invitrogen).

Proximity ligation assay (PLA)
For Proximity Ligation Assay (PLA), frozen tissue sections or 266-6 cells in confocal dishes were fixed with 4% paraformaldehyde for 1h, 0.2% Triton X-100 was used permeabilize the cell membrane.Cells and tissues were incubated with blocking buffer (Sigma) to prevent non-specific antibody binding.Then the primary antibody was added and incubated at 4 ℃ overnight, and then the secondary antibody was incubated.After that, the ligation reaction and PCR reaction were carried out, and finally mounted using ProLong Diamond Antifade Mountant with DAPI (Invitrogen).
Confocal microscopy was used to detect PLA signal.

Ultra-fast two-photon in vivo imaging
After WT mice were fasted for 12 hours, calcium ion probe Fluo-4 (30 μM, 2 μL/g, ThermoFisher Scientific) was retrogradely infused into the pancreatic main duct 20 minutes before AP models induction.POA (10 μL, 300 mM) or POAEE (10 μL, 50mM, Cayman) were locally injected for 20 seconds for stimulation.Dynamic detection of calcium ions in pancreatic tissue was performed by Two-Photon Fluorescence Microscopy (OLYMPUS, FVMPE-RS) with a water dipping lens (OLYMPUS, N.A 1.05, W.D 2 mm).Femtosecond laser (SpectraPhysics) and barrier filter (BA495-540 nm) were used for excitation and emission filtering, and the green fluorescence was collected using NDD detector.Images (512 x 512 pixels) were acquired at 2 s/frames using MicroManager (HIH).The mean fluorescence intensity (F) of 7 regions of interest (ROI) at each time point were calculated by Image J software, and all fluorescence measurements were expressed as changes from basal fluorescence (F/F0 ratio), where F0 represents initial fluorescence at the beginning of each experiment.Video legends Video 1. WT mice were intraperitoneally injected with 5% DMSO as control, changes in calcium ion levels (green fluorescence signal) in acinar cells was observed when POA stimulates pancreatic tissue.Video 2. WT mice were intraperitoneally injected with KAN0438757 to inhibit PFKFB3, changes in calcium ion levels (green fluorescence signal) in acinar cells was observed when POA stimulates pancreatic tissue.Video 3. CKO mice were intraperitoneally injected with 5% DMSO, changes in calcium ion levels (green fluorescence signal) in acinar cells was observed when POA stimulates pancreatic tissue.Video 4. WT mice were intraperitoneally injected with 5% DMSO as control, changes in calcium ion levels (green fluorescence signal) in acinar cells was observed when POAEE stimulates pancreatic tissue.Video 5. WT mice were intraperitoneally injected with KAN0438757 to inhibit PFKFB3, changes in calcium ion levels (green fluorescence signal) in acinar cells was observed when POAEE stimulates pancreatic tissue.Video 6. CKO mice were intraperitoneally injected with 5% DMSO, changes in calcium ion levels (green fluorescence signal) in acinar cells was observed when POAEE stimulates pancreatic tissue.
AR42J cells were pretreated with DMSO or KAN0438757 (4 μM) for 12 hours, and then incubated with calcium ion probe Fluo-4 (4 μM, ThermoFisher Scientific) for 45 mins.POA (300 mM) was used as a stimulant.Dynamic detection of calcium ions in AR42J cell line was performed by ImageXpress Micro Confocal (MOLECULAR DEVICES).Images were acquired at 5 s/frames using High sensitivity high resolution sCMOS camera.The mean fluorescence intensity (F) of cells (n=40-50) were calculated by Image J software, and all fluorescence measurements were expressed as changes from basal fluorescence (F/F0 ratio), where F0 represents initial fluorescence at the beginning of each experiment.RNA-sequencing and analysisTotal pancreatic RNA from experimental mice were extracted and reverse transcribed to cDNA libraries for sequencing with NEBNext Ultra II Directional RNA Library Prep Kit (NEB).Libraries were sequenced on NovaSeq 6000 platform (Illumina).The pair-end sequencing data was treated by fastp software to filter low quality reads , and then mapped to Ensembl GRCm38.p6reference genome by CLCStatisticsExperiments were performed in at least triplicates.Error bars represent the mean with SD.All statistical analyses were performed with unpaired Student's t-test, logrank (Mantel-Cox) test or one-way ANOVA test (GraphPad Prism 8 software).p<0.05 was considered as a statistically significant difference, *p<0.05,**p<0.01,***p<0.001,N.S., no significance.Study approvalClinical specimens' acquisition was approved by Medical Ethics Committee ofThe First Affiliated Hospital of Wenzhou Medical University (KY2022-R188).All animal experiments were conducted according to the US National Institutes of Health Guide approved by the Animal Service Center of ISM (ISMIACUC-0009-R).

Figure 5 .
Figure 5. PFKFB3 is involved in the regulation of pathological calcium overload in acinar cells.(A-B) KEGG pathway enrichment (A) and GO analysis (B) in pancreas from cKO and WT mice from the FAEE-SAP model.(C) The enrichment plot comparing the transcriptome of cKO mice and WT mice in FAEE-SAP from GSEA analysis.Normalized Enrichment Score (NES), -2.133;Normal p-value, 0.001.(D) Heatmap shows the core enrichment genes in the pathway named response to calcium ion.(E-F) Representative traces of POA (E) or POAEE (F) induced Ca 2+ elevations.The representative images are the first and last frames of pancreas prestained with Fluo-4, scale bar, 40μm.KAN0438757 (25mg/kg) was administered by intraperitoneal injection 1h prior to detection, DMSO in saline solution was used as a control.P-values calculated using unpaired Student's t-test, *p<0.05,**p<0.01,***p<0.001.Error bars depict mean ± SD.Each experiment performed at least in triplicate.

Figure 6 .
Figure 6.PFKFB3 cooperate with IP3R to regulate calcium homeostasis in pancreatic acinar cells.(A) Expression of IP3R, p-IP3R, PFKFB3 and GAPDH by western blotting in mice pancreata from the FAEE-SAP model (B) AR42J cells were pretreated with KAN0438757 (4 μM) or DMSO for 12h, 0.5h or 6h after POA (300 μM) stimulation cells were harvested for Western blotting analysis.(C) Representative traces of POA induced Ca 2+ elevations in AR42J cells were recorded by High Content Analysis Imaging System.KAN0438757 (4 μM) was administered 12h prior to detection, DMSO was used as a control.(D) Frozen section of pancreas damaged by FAEE (collected 2h after induction) evaluated by confocal microscopy, scale bar, 40μm.(E) PLA analyzed in frozen section of pancreas damaged by FAEE, scale bar, 42μm.(F) Whole-cell lysates of 266-6 cells transfected with Flag-PFKFB3 construct subjected to IP analysis, and analyzed by Western blotting.(G) PLA was detected in 266-6 cells co-transfected with Flag-PFKFB3 construct and HA-IP3R construct, scale bar, 50μm.(H) 266-6 cells were transfected with full length Flag-PFKFB3, Flag-PFKFB3-K (kinase domain only) and Flag-PFKFB3-P (phosphatase domain only) constructs for co-IP analysis.(I) 266-6 cells transfected with Flag-PFKFB3, Flag-PFKFB3-K and Flag-PFKFB3-P constructs were analyzed by Western blotting analysis.(J) AR42J cells transfected with Flag-PFKFB3-K construct or Flag-PFKFB3-P construct, representative traces of POA induced Ca 2+ elevations were recorded by High Content Analysis Imaging System.Each experiment performed at least in triplicate.

Figure 7 .
Figure 7.The kinase domain of PFKFB3 binds with the CD domain of IP3R and stabilizes IP3R expression.(A) Whole cell lysate of 266-6 cells transfected with Flag-PFKFB3 and HA-IP3R-C immunoprecipitated and analyzed by Western blotting.(B) PLA signal detected in 266-6 cells co-transfected with Flag-PFKFB3 construct and HA-IP3R-C construct, scale bar, 50μm.(C) Western blot in 266-6 cells were co-transfected with Flag-PFKFB3 and HA-IP3R-C constructs.(D) Whole cell lysates from 266-6 cells transfected with Flag-PFKFB3 and HA-IP3R-CD were used for co-immunoprecipitation (co-IP).(E) Whole cell lysates were prepared from 266-6 cells were co-transfected with Flag-PFKFB3 and HA-IP3R-HD3 constructs were used for co-immunoprecipitation (co-IP).(F) The Flag-PFKFB3-K and HA-IP3R-CD constructs were transfected in 266-6 cells for the co-IP analysis.(G) The schematic containing structural information of IP3R and binding sites of PFKFB3 and IP3R interaction.(H-I) Western blot of IP3R (G) and p-IP3R (H) in 266-6 cells cotransfected with Flag-PFKFB3 construct or GFP construct in the presence of Cycloheximide (CHX, 100μg/ml, 0, 2, 4, 6h).(J) Purified Flag-PFKFB3 protein, GFP protein, HA-IP3R-C, ATP and 1x kinase reaction buffer were used for the in vitro kinase assay.(K) The working model of PFKFB3 in FAEE-SAP progression.Each experiment performed at least in triplicate.