Pancreatitis severity in mice with impaired CFTR function but pancreatic sufficiency is mediated via ductal and inflammatory cells‐Not acinar cells

Abstract Mutations in the cystic fibrosis transmembrane conductance regulator gene (CFTR) are an established risk factor for cystic fibrosis (CF) and chronic pancreatitis. Whereas patients with CF usually develop complete exocrine pancreatic insufficiency, pancreatitis patients with CFTR mutations have mostly preserved exocrine pancreatic function. We therefore used a strain of transgenic mice with significant residual CFTR function (CFTRtm1HGU) to induce pancreatitis experimentally by serial caerulein injections. Protease activation and necrosis were investigated in isolated acini, disease severity over 24h, pancreatic function by MRI, isolated duct stimulation and faecal chymotrypsin, and leucocyte function by ex vivo lipopolysaccharide (LPS) stimulation. Pancreatic and lung injury were more severe in CFTRtm1HGU but intrapancreatic trypsin and serum enzyme activities higher than in wild‐type controls only at 8h, a time interval previously attributed to leucocyte infiltration. CCK‐induced trypsin activation and necrosis in acini from CFTRtm1HGU did not differ from controls. Fluid and bicarbonate secretion were greatly impaired, whereas faecal chymotrypsin remained unchanged. LPS stimulation of splenocytes from CFTRtm1HGU resulted in increased INF‐γ and IL‐6, but decreased IL‐10 secretion. CFTR mutations that preserve residual pancreatic function significantly increase the severity of experimental pancreatitis—mostly via impairing duct cell function and a shift towards a pro‐inflammatory phenotype, not by rendering acinar cells more susceptible to pathological stimuli.


Abstract
Mutations in the cystic fibrosis transmembrane conductance regulator gene (CFTR) are an established risk factor for cystic fibrosis (CF) and chronic pancreatitis. Whereas patients with CF usually develop complete exocrine pancreatic insufficiency, pancreatitis patients with CFTR mutations have mostly preserved exocrine pancreatic function. We therefore used a strain of transgenic mice with significant residual CFTR function (CFTR tm1HGU ) to induce pancreatitis experimentally by serial caerulein injections. Protease activation and necrosis were investigated in isolated acini, disease severity over 24h, pancreatic function by MRI, isolated duct stimulation and faecal chymotrypsin, and leucocyte function by ex vivo lipopolysaccharide (LPS) stimulation. Pancreatic and lung injury were more severe in CFTR tm1HGU but intrapancreatic trypsin and serum enzyme activities higher than in wild-type controls only at 8h, a time interval previously attributed to leucocyte infiltration. CCK-induced trypsin activation and necrosis in acini from CFTR tm1HGU did not differ from controls. Fluid and bicarbonate secretion were greatly impaired, whereas faecal chymotrypsin remained unchanged. LPS stimulation of splenocytes from CFTR tm1HGU resulted in increased INFγ and IL-6, but decreased IL-10 secretion. CFTR mutations that preserve residual pancreatic function significantly increase the severity of experimental pancreatitismostly via impairing duct cell function and a shift towards a pro-inflammatory phenotype, not by rendering acinar cells more susceptible to pathological stimuli.

| INTRODUC TI ON
Pancreatitis is a common disease, often leading to hospital admission 1 and its underlying genetic susceptibilities are increasingly recognized, 2 particularly in patients with chronic or relapsing forms of pancreatitis. Although most pathophysiological studies have focused on acinar cells or acinar cell-specific events such as a pathological secretagogue response, 3 the balance between digestive protease activation and degradation via secretory 4 or lysosomal enzymes, [5][6][7][8] and the role of acinar cells as the initial site of injury, 9 the role of other cell types such as duct cells 10,11 and immune cells [12][13][14] has only recently become apparent. The cystic fibrosis transmembrane conductance regulator (CFTR) is an ABC transporter-class ion channel protein that conducts chloride ions across epithelial cell membranes.
Loss-of-function mutations in the CFTR gene were originally discovered to be associated with cystic fibrosis (CF) in affected patients 15 and later found to play a role in male vas deference infertility and chronic pancreatitis. The latter association was initially detected using the original CF-mutation panels, 16,17 subsequently characterized by full exome sequencing, and found to increase the susceptibility for pancreatitis around 2.5-fold to threefold. 18,19 It later became obvious that, unlike in CF, mainly patients with exocrine pancreatic sufficiency and carriers of CFTR mutations with maintained residual CFTR function develop pancreatitis. 20 The question whether and why only some carriers of the more than 1700 known CFTR mutations (around 15% of the population in Western countries carry CFTR mutations) develop pancreatitis, and which cell types known to contribute to the disease pathogenesis are predominantly affected by CFTR impairment, is still under debate. 21 It is further complicated by the fact that not only the type of mutation, 22 but also a number of pleotropic and modifier effects 23 appear to determine the manner and severity by which the pancreas is involved. Elegant experimental studies that were undertaken to elucidate the underlying mechanism of CFTR mutation-associated pancreatic damage have either used mice with a deleted cftr-gene (null mice) and thus no residual function 24 or with transgenically expressed cftr-deletion mutations also causing absence of cftr activity. 25,26 To experimentally recreate a situation closer to the human geno-and phenotype, we have studied experimental pancreatitis in a mouse strain (CFTR tm1HGU ) with significant residual CFTR function and no spontaneous pancreatic phenotype before adulthood. 27 Pancreatic enzyme output in these mice was found to be unimpaired, but ductal bicarbonate and fluid secretion were strongly reduced. The severity of the disease was significantly increased, whereas the response and injury of isolated acinar cell remained unaffected. In vitro inflammatory cells from CFTR tm1HGU mice were found to respond to lipopolysaccharide (LPS) with an increased release of pro-inflammatory, and a decreased release of anti-inflammatory cytokines. These data indicate that "mild" CFTR mutations with significant residual exocrine pancreatic function increase the susceptibility towards pancreatitis, as well as the disease severity. This occurs independently of acinar cell events, but is due to an impairment of CFTR function in pancreatic duct cells and the immune system.

| Induction of acute pancreatitis in wildtype and CFTR transgenic mice
All animal experiments were carried out after prior approval by the institutional animal care committee. CFTR tm1HGU mice were originally generated by Dorin et al 28 and kindly provided by the colleagues B. Tümmler and U. Seidler from the Medical University of Hannover (MHH). By using an insertional vector that encompasses a part of intron 9 and extends into exon 10 ( Figure 1A), the CFTR gene is disrupted and offspring of heterozygous mice were crossed leading to a phenotype with an impaired chloride ion transport but significant residual CFTR function of over 20% because the gene disruption by insertion does not result in the loss of any genomic sequence and the mutation is 'leaky' allowing for aberrant splicing of significant wild-type cftr mRNA. 27 These mice have no spontaneous pancreatic phenotype and good long-term survival. 29 Wild-type littermates were used as controls. All animals were maintained according to institutional guidelines and protocols of the local animal facility and approved by the institutional animal care and use committee. For this study, 14-week-old male animals were used. Mice were starved overnight with access to water ad libitum. Acute pancreatitis was induced by 8

| Preparation of serum and tissue samples
After kill, serum was aliquoted, snap-frozen and stored at −20°C for further analysis. Pancreatic tissue was partitioned and either frozen in liquid nitrogen and stored at −80°C for later enzymatic measurements, fixed in 5% paraformaldehyde for paraffin embedding, or put into Cryomold ® embedding folds (Tissue Tek, Sakura Finetek) and snap-frozen. Parts of tissue used later for enzymatic assays were homogenized on ice in a buffer consisting of 100 mmol/L Tris, 5 mmol/L CaCl 2 and pH 8.0.

| Isolation of pancreatic ducts
Pancreatic ducts were isolated using a protocol as described previously. 33 Intra-and interlobular ducts were isolated by enzymatic digestion and microdissection. Isolated ducts were incubated overnight. During the overnight incubation, the ends of the ducts seal and begin to swell due to fluid secretion of the ductal cells.

| Isolation of splenocytes
Leucocytes were isolated from spleens of CFTR tm1HGU and wild-type mice using a 70µm nylon cell filter (BD Falcon) to remove aggregates and cell debris. 34 Erythrocytes were lysed by a solution containing 155 mmol/L NH 4 Cl, 10 mmol/L KHCO 3 and 0.1 mmol/L EDTA.
Neutrophil granulocytes were isolated from total leucocytes using a MACS® cell separation kit according to the manufacturer's instruction (Miltenyi Biotech, Bergisch-Gladbach, Germany). 35,36 For separation of neutrophils, Ly6g served as a specific antibody. Both splenocytes and neutrophil granulocytes were counted in a Neubauer chamber and transferred to sterile PBS.

| Biochemical assays
Activity of serum lipase and amylase was measured by photometric absorbance assays (Roche Hitachi, Grenzach-Whylen, Germany) as kinetics over 30 min with an absorbance at 570 nm at 37°C. Purified enzymes (Sigma) were used for standardization. Trypsin activity in pancreatic homogenates was measured fluorometrically at 37°C for 1 hour using the substrate R110-CBZ-Ile-Pro-Arg (Invitrogen, Karlsruhe, Germany). Protein amount was quantified using the Bradford assay. To determine exocrine enzyme secretion, we measured chymotrypsin activity in stool as previously reported. 37 Briefly, stool was collected, stored at −20°C and later samples of For measurement of cytokine secretion, splenocytes were stimulated with 1 μg/mL LPS as reported 12,32,36 and activity in the supernatants was quantified with the CBA mouse inflammation kit following the manufacturer's instructions (Becton Dickinson) and fluorescence activated cell sorting (FACS).

| Histology, immunohistochemistry and immunofluorescence staining of pancreatic tissues
Paraffin embedded tissues were sliced and stained with haematoxylin and eosin. For each animal, ten randomly chosen microscopic fields were investigated under a light microscope. For quantification of oedema, necrosis and leucocyte infiltration, a modified histopathologic score adapted from Kyogoku et al 39 was used, that scored 0 for absent, 1 for less than 20%, 2 for 20%-50% and 3 for more than 50% as previously described. 40 Immunohistochemistry was performed from paraffin embedded tissue samples as previously reported. The following primary antibodies were used in a dilution of 1 to 200: anti-CFTR (Clone: A-3, Santa Cruz Biotechnology, sc-376683), p67-phox (Santa Cruz Biotechnology, sc-374510) and anti-CD11b (abcam, ab133357). For immunofluorescence staining of CFTR snap-frozen tissue embedded with Tissue-Tek® OCT TM compound (Sakura Finetek, Alphen aan den Rijn) was cut into 1-to 2µm-thick slices, fixed with 20% acetone and washed with PBS.
The anti-CFTR antibody (Clone: M3A7, DLN-06996, Dianova) was used in a 1:100 dilution in 20% FCS, and incubation was performed over night at 4°C. An anti-rat IgG FITC-labelled antibody (Jackson ImmunoResearch) served as a secondary antibody. Nuclei were stained by DAPI and slides were mounted with DACO mounting medium (Agilent Technologies Inc) for immunofluorescence.

| Measurement of intracellular pH
Intracellular pH (pH i ) was estimated using techniques as described before. 41 Briefly, isolated ducts were bathed in standard HEPES solution at 37°C and were loaded with the membrane-permeable acetoxymethyl derivative of BCECF (2 μmol/L) for 20-30 minutes. After loading, the ducts were continuously perfused with solutions at a rate of 5-6 mL/min. pH i was measured using a CellR imaging system.  and possibly CFTR. 42 The initial rate of intracellular acidification (dpH/dt) over the first 60 seconds starting from the administration of inhibitors was calculated by linear regression analysis using 60 data points (one pH i measurement per second).

| Measurement of in vivo fluid secretion
Small animal magnetic resonance imaging (MRI) was used to determine stimulated pancreatic fluid output as previously reported 44 and wild-type controls compared to CFTR tm1HGU animals.
Animals were allowed free access to pineapple juice 12 hours be-

| Statistical analysis
The collected data are shown as means ± SEM from at least five animals per time point and group. Statistical analysis was performed by SigmaPlot (Systat Software GmbH, Erkrath, Germany) using the Students t-test for independent samples. The data from the pancreatic duct experiments were analysed by one-or two-way ANOVA tests.

| Characterization of the CFTR tm1HGU mouse model
The insertion mutation (CFTR tm1HGU ) between exons 9 and 10 ( Figure 1A Figure 1C). This indicates that CFTR Rtm1HGU mice have no exocrine pancreatic enzyme insufficiency.

| Increased severity of acute pancreatitis in CFTR tm1HGU mice
To investigate whether a modified CFTR expression alters disease severity, we induced acute pancreatitis in CFTR tm1HGU mice and corresponding wild-type animals by repeated intraperitoneal caerulein injections. During the first 24 hours, the CFTR transgenic mice clearly developed a more severe disease course. In all animals, serum amylase, lipase as well as intrapancreatic trypsin activities increased after induction of pancreatitis. As previously shown, 46 these enzyme activities showed a biphasic curve with peaks at one and 8 hours of which the first peak is thought to be induced by direct secretagogue effects on acinar cells, whereas the second is thought to be mediated by activated inflammatory cells. Interestingly, only the later, second peaks were greater in CFTR tm1HGU animals when compared to their wild-type controls with pancreatitis, suggesting the direct response to supramaximal CCK stimulation of acinar cells did not differ from controls in the pancreas of CFTR-disrupted animals (Figure 2A-C).
MPO activity, a surrogate marker for leucocyte infiltration, was higher in CFTR tm1HGU animals than in controls, and it peaked at 24h in the pancreas and at 8h in the lungs ( Figure 2D&E).
On histology, the pancreas of CFTR tm1HGU mice clearly showed no signs of damage in the absence of caerulein stimulation, whereas all animals developed cellular vacuolization, interstitial oedema, tissue necrosis and inflammatory infiltrates in response to disease induction ( Figure 3A). Tissue injury was clearly more prominent in CFTR tm1HGU animals and when we scored and quantitated the results for histology score, which was most pronounced 8h after disease onset of pancreatitis ( Figure 3B). A detailed examination of individual parameters of the histology score showed a significant increase of inflammatory infiltrate and a trend towards increased oedema and necrosis at 8h in the CFTR tm1HGU mice ( Figure 3C).
Immunohistochemical stainings of macrophages by anti-CD11b ( Figure 3D) and neutrophils by anti-p67-phox ( Figure 3E) in pancreatic tissue samples indicated in both wild-type and CFTR tm1HGU animals that CD11b + macrophages represent the dominant infiltrating immune cell type, whereas p67-phox positive neutrophils 34 could only be detected sporadically. In CFTR tm1HGU animals there was a markedly stronger infiltration of CD11b + macrophages, which could explain the increased pancreatic damage.

| Premature intracellular zymogen activation and cell injury in CFTR tm1HGU mice
To test whether the absence of a greater disease severity after one hour of pancreatitis was due to an acinar cell-specific effect in CFTR tm1HGU mice, we isolated acini from the pancreas of CFTRdisrupted and wild-type animals and incubated them in parallel with CCK for up to 60 minutes.

| Duct cell function in CFTR tm1HGU mice
Previous studies have shown that a lack of CFTR in pancreatic duct cells leads to decreased bicarbonate secretion and impairs their physiological function. 47,48 We therefore investigated the degree of duct cell impairment in CFTR tm1HGU mice by quantitating HCO 3 (B) Summarized data for the dpH/dt changes that were calculated by linear regression analysis of pH i measurements made over the first 60 s after exposure of the transport inhibitors. (C) Wild-type animals clearly responded to forskolin stimulation by increased fluid secretion, but the CFTR tm1HGU animals did not. (D) The maximum of fluid secretion was significantly lower in CFTR tm1HGU compared to wild-type mice. Pancreatic ducts were prepared for at least from five mice per group and measurements were performed in triplicates. Asterisks indicate significant differences with P <.05 of dpH/dt changes after addition of H 2 DIDS and amiloride in the different groups is displayed in Figure 5B. space. The secretory rate, calculated from the rate of change in luminal volume, was 211 ± 58 pl/min/mm 2 after stimulation. In contrast, ducts from TG mice did not show increased secretion in response to forskolin ( Figure 5C). The secretory rate was 31 ± 74 pL/min/mm 2 , which represents a reduction of 85% compared with the WT controls. The summary of fluid secretion data is shown in Figure 5D. The maximal changes of relative luminal volume are significantly reduced in mutant mice compared to wild-type controls.
To test whether this impairment translates to the in vivo situation, we used Magnetic Resonance Imaging ( Figure 6A) of mice stimulated with secretin and found a prominent decrease in fluid secretion in the CFTR tm1HGU mice, which was reduced to one fifth of that in wild-type controls ( Figure 6B). These data demonstrate a strong impairment of pancreatic duct cell function in CFTR tm1HGU mice, which differed greatly from the effects on acinar cells.

| Impairment of leucocyte function in CFTR tm1HGU mice
The differences between CFTR tm1HGU and wild-type mice in terms of serum pancreatic enzyme and intrapancreatic trypsin activities in the later (rather than the early) disease course of pancreatitis, as well as the differences in MPO activity in the lungs and pancreas of the animals, suggested that the effects of the CFTR transgene are ultimately connected to infiltrating inflammatory cells. These findings prompted us to analyse the response of immune cells from CFTR tm1HGU animals directly. When we stimulated isolated splenocytes with lipopolysaccharide (LPS) and measured the cytokine release, we detected a significantly increased secretion of the pro-inflammatory cytokines TNFα, IFNγ and IL-6 in splenocytes of the transgenic mice ( Figure 7A-C). On the other hand, the anti-inflammatory cytokine IL-10 was found to be reduced in the transgenic animals ( Figure 7D). This combination of an increased pro-inflammatory cytokine secretion along with a decreased antiinflammatory chemokine release could represent an additional explanation for greater disease severity in CFTR tm1HGU mice.
F I G U R E 6 Fluid secretion after secretin stimulation in magnetic resonance cholangiopancreatography (MRCP). Fluid volume was markedly reduced in CFTR tm1HGU compared to wild-type mice after secretin stimulation (A + B) A B

| D ISCUSS I ON
Mutations in the CFTR gene lead to a defective CFTR channel protein, a major regulator of the transcellular transport of chloride and bicarbonate. CFTR mutations were originally identified in patients with cystic fibrosis (CF), but later they were also found to be associated with male vas deference infertility and chronic or recurrent pancreatitis. CF is perhaps the best investigated genetic disorder and most frequently based on a loss of the amino acid phenylalanine in codon 508 (delF508), either on both alleles or in combination with other variants of the cftr-gene. 15 CF patients often show a very variable clinical course including pulmonary complications such as bronchiectasis, increased mucus production and chronic endobronchial inflammation but also an impairment of other organs that include the reproductive system leading to infertility, the liver with abnormal liver function and cirrhosis, and the pancreas with exocrine and endocrine insufficiency. 50 The variable extent of pancreatitis that develops in patients with CFTR mutations [16][17][18]   In conclusion, our results confirm that impairment of CFTR function increases the susceptibility of the pancreas towards exogenous injury and clearly renders the disease course of acute pancreatitis more severe. In an experimental model in which sufficient residual CFTR function is maintained and animals have not developed exocrine pancreatic insufficiency, the underlying mechanisms appear to involve duct cells and the immune system, whereas acinar cells are not the source but a target of the CFTR pancreas phenotype.

ACK N OWLED G M ENTS
The authors thank the following team members for technical sup-

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.