Inflammation-Induced Claudin-2 Up-Regulation Limits Pancreatitis Development by Enhancing Pancreatic Ductal Transport.

See SUPPLEMENTARY METHODS for details FFPE pancreatic tissues were obtained under an IRB-approved protocol from The University of Chicago, including 10 normal and 19 chronic pancreatitis samples. Cldn2 KO mice ( Cldn2 tm1Lex ) 31 on a C57BL/6 background were maintained by crossing Cldn2 KO/+ females with Cldn2 KO/0 males. Male Cldn2 +/0 and female Cldn2 KO/+ mice served as controls, while Cldn2 KO/0 males and Cldn2 KO/KO females were used as Cldn2 KOs. This breeding strategy ensured appropriate comparisons across sex and genotype. For caerulein-induced pancreatitis, mice received eight hourly intraperitoneal injections (75 μg/kg) over two consecutive days, which is a well-established model to induce acute pancreatitis. 32 For acute pancreatitis, tissues were collected at early (day 2) and late (day 8) time points following caerulein treatment, whereas for chronic pancreatitis, pancreatic tissues were harvested at the completion of a repeated caerulein/ethanol injury protocol for histologic and morphologic assessment. For cytokine studies, C57BL6/J mice received intraperitoneal injections of either PBS or PBS plus cytokine. Pancreatic tissues were harvested 1 day post-injection and processed for histological and molecular analysis. H&E staining was performed on human tissues using a Tissue-Tek Prisma automated stainer and manually on mouse tissues to visualize tissue architecture. Trichrome staining of human and mouse pancreas tissues was performed at The University of Chicago’s Human Tissue Resources Center. FFPE human, mouse, and organoid pancreatic tissues were stained for CLDN2, OCLN, CLDN1, CFTR, CD45, and Ki-67 by IHC or IF. IHC included standard deparaffinization, antigen retrieval, polymer-based detection, and hematoxylin counterstaining. IF used Alexa Fluor 488/594–conjugated secondary antibodies with Hoechst 33342 nuclear staining. Slides were scanned at 40× magnification using an Aperio AT2 slide scanner. Image analysis was performed using Aperio eSlide Manager and ImageScope software. Staining was quantified using the positive pixel count algorithm (v8.1) for trichrome (blue), CLDN2 (brown), and zymogen granules (red). CD45-positive cells were quantified using the cell membrane algorithm (v9.1), and Ki-67 staining was quantified using the nuclear algorithm. Pancreatic ductal organoids from control and Cldn2 KO mice were generated using modified established protocols. 33 Pancreatic tissue was minced, enzymatically digested, and filtered through a 100 μm strainer. Cells were embedded in Cultrex UltiMatrix and cultured in organoid growth media. Organoids were passaged every 3–5 days using TrypLE. For IHC, organoids were fixed, embedded in HistoGel, and processed as FFPE samples to examine tight junction proteins. Mouse pancreatic ducts were isolated by two sequential 30-minute enzymatic digestions (DMEM containing soybean trypsin inhibitor, bovine serum albumin, hyaluronidase, and collagenase) with an intermediate wash and gentle trituration to release intact ducts. Digestion was quenched, and the duct-enriched fraction was washed and cultured for 48 hours to allow epithelial resealing. Recovered ducts were resuspended in glucose-supplemented HBSS and plated for downstream assays. Pancreatic organoids were differentiated in media without A83–01, [Leu15]-Gastrin I, and prostaglandin E2 to promote epithelial maturation. Organoids were treated with IL-6, IL-13, IL-22, IFNγ, or TNFα to assess inflammatory effects on gene expression. RNA was extracted (RNeasy Mini Kit), reverse-transcribed to cDNA, and analyzed by qPCR (Bio-Rad iQ SYBR Green Supermix) using mCldn2 and mGAPDH primers. Gene expression was calculated via ΔΔCt method. Live imaging of control and Cldn2 KO pancreatic ductal organoids was conducted in Cultrex UltiMatrix with differentiation media. Organoid size changes were measured under Na + -HBSS and NMDG + -HBSS to assess sodium-driven paracellular water flux. In separate sets of experiments, organoids were treated with forskolin (10 μM), CFTR inhibitor-172 (20 μM) or DIDS (400 μM) as indicated in HBSS. Images were captured every 10 minutes for 180 minutes using a Keyence BZ-X810 microscope, and organoid area was quantified using the OrganoID package. 34 Isolated pancreatic ducts from control and Cldn2 KO mice were equilibrated in HBSS for 2 hours prior to imaging. Ducts were then subjected to time-lapse imaging in the presence or absence of forskolin (10 μM). Images were acquired every 5 minutes for 85 minutes using a Keyence BZ-X810 microscope. Luminal area at each time point was quantified using ImageJ, normalized to the initial area, and converted to relative luminal volume by scaling the normalized area to the 1.5 power. 35 Data analysis and visualization were performed in Prism 8.0 (GraphPad) using Student’s t test or Welch’s t test, one-way ANOVA with Bonferroni post hoc correction, and repeated-measures two-way ANOVA, as appropriate. Significance was defined as * p <0.05, ** p <0.01, *** p <0.001, with statistical tests selected based on data distribution and experimental design.

Previous studies on the pancreas demonstrated that CLDN2 is predominantly expressed in pancreatic ductal epithelial cells, with low expression in pancreatic acinar cells. 36 However, how CLDN2 expression may be altered in chronic pancreatitis remains to be determined. To investigate this, we evaluated CLDN2 expression in pancreatectomy specimens from patients with chronic pancreatitis. H&E staining showed acinar cell loss and fibrosis in the chronic pancreatitis samples, which were not present in normal pancreas samples ( Fig. 1A ). CLDN2 expression was significantly higher in the ducts of chronic pancreatitis tissues compared to normal pancreas, regardless of duct size, as determined by immunohistochemical (IHC) staining ( Fig. 1B – C , clinical parameters for patients studied are listed in Supplementary Table 1 ). CLDN2 staining was also increased in islets of Langerhans in patients with chronic pancreatitis, although such increase was less pronounced than that observed in pancreatic ducts ( Supplementary Fig. 1 ). Similar CLDN2 upregulation in pancreatic ducts was also obtained by IHC staining using a different anti-CLDN2 primary antibody, and by immunofluorescence staining of pancreatic tissues ( Supplementary Fig. 2 ). These data suggest CLDN2 may be involved in the pathophysiological processes associated with chronic pancreatitis. To define if CLDN2 upregulation is a specific process or a general alteration of the tight junction, we assessed expression of CLDN1, a barrier forming claudin, 37 , 38 and occludin (OCLN), a non-claudin transmembrane protein that controls large molecule permeability across the tight junction, 39 , 40 by immunofluorescence staining and IHC respectively 41 . Expression of these proteins was similar in control and pancreatitis samples, suggesting that CLDN2 is selectively altered in chronic pancreatitis ( Supplementary Fig. 3 ). We next sought to understand if CLDN2 upregulation can be recapitulated in experimental pancreatitis. Acute experimental pancreatitis was induced by repeated injection of the cholecystokinin homolog caerulein (first injection at day 0). Following eight hourly caerulein injections per day for two days, pancreatic samples were collected 1d after the final injection (day 2), representing the early phase of caerulein-induced acute experimental pancreatitis and 7 d after the final injection (day 8), representing the later phase of acute pancreatitis ( Fig. 1D ). H&E staining showed that tissues obtained during early and late phases of disease exhibited histological evidence of tissue damage, including mild expansion of the interstitial space due to edema and inflammation ( Fig. 1E ). In the later phase of pancreatitis, inflammation and fibrin deposition were more predominant, particularly in periductal areas ( Fig. 1E ). IHC staining showed CLDN2 expression was increased in the ductal epithelial cells in both the early and late phases of disease ( Fig. 1F – G ). CLDN2 was increased at the tight junction, cell membranes, and in cytoplasmic vesicles. To determine whether CLDN2 upregulation is also a feature of chronic pancreatitis, we examined CLDN2 expression in the trypsinogen T7D23A mouse model of chronic pancreatitis. 42 Pancreatic tissues were collected between 2 weeks and 6 months of age, capturing both early and late stages of disease. Histologic assessment by H&E demonstrated a time-dependent progression of pancreatic injury, with increasing inflammatory infiltrates, edema, and loss of acinar cell mass ( Supplementary Fig. 4A ). IHC staining revealed that CLDN2 expression was similar to control mice at early time points in mice with T7D23A mutation but became significantly higher in T7D23A mice beginning at 4 weeks of age, with persistent elevation in older mice, primarily in pancreatic ductal epithelium ( Supplementary Fig. 4A – B ). These findings demonstrate that CLDN2 upregulation is a common feature for both caerulein-induced acute experimental pancreatitis and trypsinogen mutation induced chronic experimental pancreatitis. Together with increased CLDN2 expression in human chronic pancreatitis samples, these findings suggest CLDN2 upregulation is a common feature of pancreatitis, and this change could impact pancreatitis development. Our data showed CLDN2 is upregulated in pancreatic ductal epithelial cells during pancreatitis. We hypothesized that such CLDN2 upregulation is influenced by pancreatitis-associated cytokines, as IL-13, IL-6, IL-22, TNFα, and IFNγ are known to increase CLDN2 expression in other tissues. 23 – 25 , 43 – 45 To evaluate the effect of these cytokines on CLDN2 expression in pancreatic ductal epithelium, we treated pancreatic ductal epithelial organoids derived from C57BL/6 mice with these agents. In contrast to previous pancreatic ductal adenocarcinoma derived 2D culture models, our non-transformed 3D pancreatic ductal epithelial organoid model provides a system that better reflects normal pancreatic ductal physiology. Out of these cytokines, only IFNγ increased Cldn2 mRNA expression ( Fig. 2A ) in a dose-dependent manner ( Fig. 2B ), as determined by RT-qPCR. CLDN2 protein expression was also increased in organoids treated with IFNγ, demonstrated by IHC staining of organoid sections ( Fig. 2C ). To determine if IFNγ can induce CLDN2 expression in vivo , we injected IFNγ intraperitoneally and collected mouse pancreata 24 hours later. RT-qPCR and IHC staining showed that IFNγ induced CLDN2 expression at both mRNA and protein levels ( Fig. 2D – E ). Together, these data show that IFNγ can induce CLDN2 expression both in vitro and in vivo . To further explore this role, we evaluated whether CLDN2 upregulation in the caerulein-induced mouse model of pancreatitis can be impacted by IFNγ knockout ( Ifng KO). While CLDN2 expression still increased in Ifng KO mice at day 2 following initiation of caerulein treatment, this increase was significantly lower than in control mice treated with caerulein ( Fig. 2F ). Given that CLDN2 upregulation was not completely abrogated in Ifng KO mice, we next asked whether additional cytokines could induce CLDN2 expression in vivo . Intraperitoneal administration of IL-6 did not increase CLDN2 expression in pancreatic ducts, whereas a single dose of TNFα increased ductal CLDN2 staining ( Supplementary Fig. 5 ). Collectively, these data support a role for IFNγ and TNFα in driving CLDN2 expression in pancreatic ductal epithelial cells. Although studies above demonstrated CLDN2 is upregulated in pancreatitis, functional consequences for such upregulation remain unknown. To determine the impact of CLDN2 on pancreatitis development, we assessed how the loss of CLDN2 expression affects the progression of caerulein-induced pancreatitis by using Cldn2 KO mice . Cldn2 KO mice exhibit calcium deposits in the kidney after 4.5 months, but do not exhibit features of systemic disease or weight change in an unstressed state and their pancreata appear to be histologically normal. 46 , 47 Following caerulein treatment, both control and Cldn2 KO mice had similar weight change over time ( Supplementary Fig. 6A , B ). In the early phase (1 d after last caerulein injection), H&E staining showed more significant tissue reorganization in Cldn2 KO mice ( Fig. 3A , Supplementary Fig. 7A ), and IHC staining showed that caerulein induced significant CLDN2 staining increase in pancreatic ducts of control mice, while no CLDN2 staining was observed in Cldn2 KO pancreata ( Fig 3B ). In this phase, little tissue fibrosis occurred in either control or Cldn2 KO mice, as indicated by limited blue staining in the trichrome stained tissue ( Fig. 3C , E ). IHC staining of CD45, a pan-leukocyte marker, was performed to assess tissue inflammation. While the percent of total cells staining positive for CD45 in pancreata from both control and Cldn2 KO mice increased following caerulein treatment, the percentage increased significantly more in Cldn2 KO mice ( Fig. 3D , F ). This indicates CLDN2 limits tissue inflammation in the early phase of caerulein-induced experimental pancreatitis. In the late phase (7 d after last caerulein injection), significant tissue damage differences persisted between control and Cldn2 KO mice ( Fig. 3G , Supplementary Fig. 7B ). Pancreata from Cldn2 KO mice had greater cellular vacuolization, edema, inflammation, and acinar cell dedifferentiation. Trichrome staining showed that despite limited fibrosis in control mice, there was significant tissue fibrosis in Cldn2 KO mice ( Fig. 3H , J , Supplementary Fig. 8 ). CD45 staining showed the percent of CD45 positive cells was lower in both groups. While there was no statistical significance between these groups, there was still a small number of Cldn2 KO mice that had high numbers of tissue leukocytes ( Fig. 3I , K ). Taken together, these data show CLDN2 limits tissue damage and fibrosis in both the early and late phase of caerulein-induced acute pancreatitis. Although studies above demonstrated that CLDN2 can limit tissue damage and reorganization in an acute setting, this model does not present some of the major features of human chronic pancreatitis. To this end, we employed a chronic caerulein- and ethanol-induced pancreatitis model. Mice received twice-weekly caerulein injections administered 3–4 days apart with seven hourly injections per treatment day (100 μg/kg), along with ethanol (3 mL/kg) injections once on each intervening day of the week ( Fig. 4A ). 48 Low-power H&E staining revealed more extensive architectural disruption in pancreata from Cldn2 KO mice compared to controls following 7 weeks of treatment ( Fig. 4A ). While control pancreata retained partially preserved lobular organization, Cldn2 KO pancreata had widespread acinar loss and interstitial expansion. Higher-power H&E images further demonstrated pronounced acinar cell injury in Cldn2 KO mice, characterized by reduced zymogen granule content and increased cellular disorganization ( Fig. 4B , C ). To assess fibrotic remodeling, Masson’s trichrome staining was performed. Cldn2 KO pancreata had mild increase in collagen deposition compared to controls ( Fig. 4D , E ). IHC staining of CD45 further showed Cldn2 KO pancreata had more leukocyte content than control pancreata, indicating more severe tissue inflammation ( Fig. 4F , G ). Together, these data demonstrate that loss of CLDN2 exacerbates acinar injury, increases fibrotic remodeling, and elevates tissue inflammation in a chronic experimental pancreatitis mouse model. CLDN2 is primarily expressed in pancreatic ductal epithelial cells, and its expression is upregulated in the setting of pancreatitis. Furthermore, CLDN2 can form a paracellular channel that is selective for small cations and can also mediate water transport, based on studies on intestinal and renal epithelium. 43 , 46 Thus, it is possible that CLDN2 can also regulate small cation and water transport in pancreatic ductal epithelium. To test this possibility, we generated pancreatic ductal epithelial organoids from control and Cldn2 KO mice. Immunofluorescent staining of formalin fixed organoid sections confirmed that CLDN2 is expressed in control organoids, but not in Cldn2 KO organoids ( Fig. 5A ). Proliferation was similar between control and Cldn2 KO organoids, as assessed by Ki67 staining ( Supplementary Fig. 9 ), indicating that loss of CLDN2 does not affect baseline organoid growth. We hypothesized that if a sodium selective paracellular pathway is important for pancreatic ductal epithelial transport, creating an iso-osmotic Na + gradient across the epithelial layer of a pancreatic organoid would create a driving force for unidirectional Na + transport. If Na + flows through such a selectively permeable pathway, water must follow to maintain ionic equilibrium ( Fig. 5B ). To test this, we substituted Na + in extraluminal imaging media with isomolar CLDN2 impermeable large organic ion N-methyl-D-glutamine (NMDG + ), creating a high intraluminal to extraluminal Na + concentration gradient ( Fig. 5B ). While changing regular imaging media with the same Na + containing media had little effect on organoid size, substitution with iso-osmotic NMDG + containing Na + free media caused progressive decrease in organoid size ( Fig. 5C , D row 2 vs. row 1). In contrast, when such ion substitution was performed on Cldn2 KO organoids, such size change was limited ( Fig. 5C , D row four vs. row 3). This shows that a Na + driven fluid transport pathway exists in pancreatic ductal epithelium, which is CLDN2 dependent. Having identified that CLDN2 can regulate Na + -dependent transport in pancreatic ductal epithelium, we next sought to assess contributions of CLDN2 to active pancreatic ductal secretion using a forskolin induced organoid swelling assay. 49 , 50 In this assay, forskolin activates adenylate cyclase and induces CFTR-mediated cellular secretion of Cl − into the organoid lumen ( Fig. 6A ). This generates an electrochemical gradient that drives Na + entry into the lumen. 51 In control pancreatic ductal epithelial organoids, forskolin induced an increase in organoid size. ( Fig. 6B , C ). In contrast, forskolin only caused a limited increase in Cldn2 KO organoid size. These data further show a functional paracellular Na + permeable CLDN2 channel is required for pancreatic ductal epithelial secretion. Impaired forskolin-induced swelling of Cldn2 KO organoids can be caused by defective transcellular secretion or altered paracellular ion permeability. To differentiate these possibilities, we immortalized organoids in 3D and transitioned them to 2D culture for electrophysiological studies. Despite loss of CLDN2, tight junction organization and actin cytoskeletal architecture were preserved, as assessed by OCLN, ZO-1, and F-actin staining in control and Cldn2 KO pancreatic ductal epithelial monolayers ( Supplementary Fig. 10A ). Loss of CLDN2 resulted in increased transepithelial electrical resistance (TER), indicating a substantial reduction in overall paracellular conductance ( Supplementary Fig. 10B ). Furthermore, NaCl dilution potential measurements demonstrated a reduction in cation selectivity ( Supplementary Fig. 10C ), driven by decreased Na+ permeability, but not increased Cl- permeability, demonstrating that Cldn2 KO monolayers have reduced permeability relative to control monolayers. Furthermore, forskolin-stimulated short-circuit current was not reduced in Cldn2 KO monolayers relative to controls ( Supplementary Fig 10D ). Suggesting impaired forskolin induced Cl − secretion is not the cause for diminished organoid swelling ( Fig. 6 ). Taken together, these data showed that Cldn2 KO pancreatic ductal epithelia are less permeable and have intact Cl − secretion. To further determine the contribution of CFTR to forskolin-induced organoid swelling, we employed the CFTR inhibitor, CFTR inh-172. CFTR inhibition reduced forskolin-induced swelling in control pancreatic ductal organoids, demonstrating forskolin-induced organoid swelling depends on CFTR-mediated Cl − secretion ( Fig. 6D , E ). In contrast, forskolin-induced swelling remained limited in Cldn2 KO organoids, which was not affected by CFTR inhibition ( Fig. 6D , E ). As CFTR induced Cl − secretion is coupled to Cl − /HCO3 − exchanger in pancreatic ducts to impact ionic composition of the pancreatic fluid, it is possible that Cl − /HCO3 − exchanger also participates in forskolin-induced fluid secretion. To test this, we treated pancreatic ductal organoids with 4,4′-Diisothiocyano-2,2′-stilbenedisulfonic acid (DIDS), a Cl − /HCO3 − exchanger inhibitor. DIDS treatment reduced forskolin-induced swelling in control pancreatic ductal epithelial organoids ( Fig. 6F , G ), but did not impact limited size change of Cldn2 KO organoids. Taken together, these data underscore a critical role for CLDN2-facilitated paracellular Na + permeability in pancreatic ductal epithelial secretion. These studies further show that CFTR activity and Cl − /HCO3 − exchange are both required for efficient pancreatic ductal fluid secretion, and in the absence of CLDN2, fluid secretion cannot occur despite forskolin activated transcellular transport pathways remain intact. Studies above suggest CLDN2 can regulate pancreatic ductal secretion in established organoid lines and organoid derived 2D culture. However, contributions of CLDN2 in pancreatic duct secretion were not directly investigated. To this end, we isolated intralobular pancreatic ducts from control and Cldn2 KO mice and measured forskolin-induced fluid secretion by ex vivo live imaging. Similar to what was observed in organoids, control intralobular ducts increased their size over time, indicating fluid secretion, whereas Cldn2 KO intralobular ducts have limited size increase ( Fig. 7 ). These data indicate that Cldn2 KO limits pancreatic ductal secretion.

Pancreatitis is an inflammatory disease of the pancreas that originates from a complex interplay of genetic predispositions, environmental triggers, autoimmune diseases, and metabolic disorders. 1 Aberrant activation of digestive enzymes within the pancreas is thought to play a major role in the development of pancreatitis. This is supported by evidence of genetic mutations implicated in chronic pancreatitis, which often affect the trypsin-dependent pathway. Genes such as PRSS1 , SPINK1 , and CTRC are known to contribute to pancreatitis development through these mutations. Genes that regulate ion and fluid balance in the pancreatic ducts, such as CFTR , and TRPV6 also impact pancreatitis development. 52 , 53 Defects in these ductal associated genes may limit pancreatic fluid secretion, increasing digestive enzyme concentration, leading to autodigestion of the pancreas. This suggests that other proteins that regulate pancreatic fluid secretion may also impact pancreatitis development. Supporting this, GWAS analyses identified a gene that contributes to paracellular cation transport and epithelial barrier function, CLDN2 , as a risk gene for pancreatitis. 4 , 6 , 7 , 54 However, the physiological and pathophysiological function of CLDN2 in pancreas remains unknown. CLDN2 is a tetraspanning transmembrane tight junction associated protein. It functions as a cation-selective ion channel and facilitates water transport via solvent drag. In the intestines, it supports crypt secretion. Notably, while Cldn2 KO increases susceptibility to bacterial infection, it appears to protect against experimental inflammatory bowel disease. 44 , 55 Thus, the role of CLDN2 in disease may be context-dependent, with its physiological functions influencing distinct pathologies. In pancreatitis, CLDN2 alterations could impact disease development, but it remains unclear whether increased or decreased function promotes it. Results of GWAS studies do not readily provide an answer, as all CLDN2 associated SNPs exist outside of protein coding regions of CLDN2 . Thus, they do not affect CLDN2 amino acid sequence but are more likely to alter its expression. Although an initial publication connected a pancreatitis associated SNP ( rs12688220 ) within the CLDN2 - MORC4 region to altered CLDN2 protein localization within the pancreatic ductal epithelial cells, subsequent analysis suggested this locus is associated with MORC4 , rather than CLDN2 . Thus, it remains unclear if increased or decreased CLDN2 function may promote pancreatitis development. To investigate CLDN2’s role in pancreatitis, we examined its expression in pancreatis specimens from patients with chronic pancreatitis of various etiologies, revealing upregulation, particularly in ductal cells. Similarly, CLDN2 expression increased during both acute and chronic pancreatitis mouse models, supporting the use of such models for studying its role in disease. Cldn2 deletion in acute and chronic models of pancreatitis worsened disease, marked by increased immune infiltration and fibrosis. These findings suggest that CLDN2 upregulation may act as a compensatory mechanism to limit pancreatitis, highlighting the potential of enhancing CLDN2 function as a novel therapeutic approach. Building on these findings, it is crucial to understand how CLDN2 is upregulated in pancreatitis. CLDN2 upregulation is a common feature of tissue inflammation across various conditions. In the gastrointestinal tract, its expression is increased in a broad range of diseases, including inflammatory bowel disease, 22 , 23 , 25 infectious colitis, 19 and celiac disease. 21 In the lungs, a similar pattern of CLDN2 upregulation has been observed under chronic inflammatory conditions. 56 , 57 These findings suggest that CLDN2 upregulation is a common response to chronic inflammation in various organs. In the gastrointestinal tract, CLDN2 upregulation has been reported to be driven by multiple inflammatory cytokines. 24 – 30 Using a normal pancreatic ductal epithelial organoid model, our study showed that IFNγ is an effective inducer of CLDN2 expression at the transcriptional level, while the other cytokines tested, including IL-13, IL-6, IL-22, and TNFα, do not increase Cldn2 mRNA expression. When injected into wild type mice, IFN γ also induced CLDN2 expression at both the mRNA and protein levels. These findings may be functionally significant, as Ifng KO mice have been reported to have more severe caerulein-induced experimental pancreatitis. 58 In our study, caerulein-induced CLDN2 expression is reduced in Ifng KO mice. Although our data indicate that IFN γ is not the only driver to upregulate CLDN2 expression, they suggest that IFN γ limits pancreatitis development at least partially through inducing CLDN2 expression. To completely understand how CLDN2 is upregulated to limit pancreatitis development, it is necessary to understand how other factors may drive CLDN2. In contrast to studies using organoids, intraperitoneal injection of TNFα into mice increased CLDN2 expression. Although other possibilities exist, this result may suggest TNFα may regulate CLDN2 expression on pancreatic ductal epithelium through an indirect pathway. 24 , 59 , 60 Additional studies are needed to systematically define regulators for CLDN2 expression in the context of pancreatitis development. We next investigated how CLDN2 expression may protect against pancreatitis. Given the role of CLDN2 in regulating cation and water transport through the paracellular pathway, we assessed its impact on fluid transport using pancreatic ductal epithelial organoids derived from control and Cldn2 KO mice. 49 , 61 This 3D organoid model preserves native ductal epithelial architecture, cell polarity, and expression of epithelial transport machinery, without the oncogenic alterations present in pancreatic ductal adenocarcinoma-derived cell lines. Using organoid-based assays, we obtained new insights into pancreatic ductal epithelial fluid transport. Substituting extraluminal Na + with NMDG + caused a progressive decrease in organoid size, indicating net fluid extrusion from the lumen. Since water movement in our system depends on local osmotic gradients, these observations suggest an asymmetry in Na + and NMDG + transport across the epithelium. This indicates a size-selective cation-permeable pathway in the pancreatic ductal epithelium that drives fluid transport. Because this pathway was largely absent in Cldn2 KO organoids, and CLDN2 mediates selective cation transport, our data suggest that a Na + gradient primarily drives fluid transport through a CLDN2-dependent paracellular pathway. During pancreatic secretion, CFTR-mediated Cl − secretion, and the subsequent Cl − /HCO3 − exchange through SLC26 family members drives HCO3 − and fluid secretion in the pancreatic duct. 62 It is well recognized that a paracellular shunt pathway is required for Na + to follow anion secretion in pancreatic ductal epithelium. 63 However, the molecular entity that creates this paracellular passive Na + permeable pathway is unknown. In addition to the result discussed in the preceding paragraph demonstrating the presence of a CLDN2 dependent Na + permeable pathway, our forskolin-induced organoid swelling assay further suggests paracellular Na + transport during pancreatic ductal secretion is also mediated by CLDN2. Furthermore, forskolin-induced swelling assay performed in acutely isolated intralobular pancreatic ductal segments showed Cldn2 KO ducts have less luminal fluid accumulation, further suggests pancreatic duct fluid secretion is controlled by CLDN2. One of the alternative explanations for the findings discussed above is that loss of CLDN2 can increase transcellular permeability, thus the free movement of ion and water across the epithelium can dissipate ionic and osmotic gradients in any direction, which does not allow net movement of water across the epithelium. This view cannot be true, as electrophysiology studies using immortalized, 2D converted organoid culture showed Cldn2 KO increased transepithelial resistance, along with decreased Na + permeability, without drastic change in Cl − permeability. This indicates Cldn2 KO ductal epithelium is tighter, rather than more permeable, to small molecules. Studies using these monolayers also indicate that forskolin induced Cl − transport in Cldn2 KO epithelium, as measured as short circuit current, is no less than those observed in control epithelium, indicating Cldn2 KO epithelium has intact Cl − secretory machinery. Taken together, our data show that CFTR, Cl − /HCO3 − , and CLDN2 are all key regulators for induced pancreatic secretion. 62 , 63 Although CLDN2 has also been reported to function as a water transporter, and our data show loss of CLDN2 function can limit fluid transport in both organoid models used in this study, it is not clear if observed water transport is directly mediated by CLDN2. It is possible that CLDN2 controlled directional paracellular Na + transport only provides an osmotic gradient that drives water transport through other channels, including aquaporins. It is also possible that both passive Na + and water transport are directly mediated by CLDN2. Relative contributions of water transport directly through CLDN2 channels and other water transporters remain to be investigated in the future. Our studies on chronic pancreatitis patient samples and mouse experimental pancreatitis samples revealed upregulated CLDN2 expression in the pancreatic ductal epithelium during inflammation. Based on our transport findings, this upregulation likely facilitates ion and water secretion into the lumen, representing a compensatory mechanism to limit pancreatitis progression. In our experimental pancreatitis models, the absence of this upregulation, such as in Cldn2 KO mice, results in paracellular Na + secretion into the ductal lumen being reduced, along with reduced CLDN2-dependent pancreatic ductal water secretion. This may result in reduced secretory volume with increased digestive enzyme concentration. This would lead to elevated spontaneous pancreatic enzyme activation that could cause pancreatic ductal damage or further sensitize pancreatic tissue for other insults. In addition, in severe cases, thickened pancreatic fluid due to reduced CLDN2 function may cause ductal obstruction, which may contribute to pancreatitis severity. This leads to increased immune cell infiltration and fibrosis. Mechanistically this is analogous to CFTR mutation associated pancreatitis, where the driving force for Na + and water secretion is reduced by lack of stimulated Cl – secretion. 51 , 64 Knowledge gained in this study may be used to predict how reported risk alleles may impact pancreatitis development: these polymorphisms may cause reduced functional CLDN2 channels at the tight junction, causing increased sensitivity to pancreatitis. Because these risk alleles are located outside the CLDN2 coding region, it is possible such reduced CLDN2 function may be caused by decreased CLDN2 expression. It is also possible that additional polymorphisms that are tightly linked with risk allele are responsible for the increased risk of pancreatitis, and such changes could limit CLDN2 expression, channel activity, or channel localization. These possibilities need to be clarified by deep sequencing at the CLDN2 locus and adjacent region, and functional studies testing how these polymorphisms may affect CLDN2 expression, localization, and function. CLDN2 risk alleles may intersect with other pancreatitis associated genes. It is reported that Cftr KO mice have increased CLDN2 expression in the small intestine, which may compensate for the diminished ion transport. 65 This points to a possibility that CLDN2 upregulation is also higher in patients with pancreatitis risk loci. It is also possible a fraction of patients with CLDN2 risk alleles also have other pancreatitis risk alleles, giving them significantly higher risk for pancreatitis and require careful monitoring and more aggressive clinical management. This highlights the importance of genetic testing of CLDN2 risk alleles. Taken together, our findings suggest that CLDN2 plays a pivotal role in modulating pancreatic ductal epithelial transport, especially in the context of inflammation, and that its upregulation serves as an adaptive mechanism to mitigate pancreatic injury. These results point to a potential therapeutic target in CLDN2, where activating CLDN2-mediated pathways may promote pancreatic secretion, limit inflammation, and reduce the severity of pancreatitis. Further investigation into the complex interplay between CLDN2, CFTR, and other ion channels and transporters could pave the way for novel therapeutic strategies targeting pancreatic ductal function in the management of pancreatitis.

Pancreatitis is characterized by inflammation and auto-digestion of the pancreas, which sometimes progresses to permanent scarring of pancreatic tissue. 1 The condition presents with symptoms such as abdominal pain, nausea, vomiting, and steatorrhea. In more severe or recurrent cases, pancreatitis can lead to complications such as malnutrition due to exocrine pancreatic insufficiency, diabetes due to endocrine dysfunction, and an increased risk of pancreatic ductal adenocarcinoma. In the United States the annual incidence of chronic pancreatitis ranges from 5 to 8 per 100,000 adults, and prevalence estimates between 42 to 73 per 100,000 adults. 2 While lifestyle factors such as alcohol and tobacco abuse are well-recognized contributors to pancreatitis, genetic predispositions also play an important role in its development. Mutations in genes such as PRSS1 , SPINK1 , CTRC , and CFTR have been linked to increased susceptibility to the disease. 3 Genome-wide association studies (GWAS) have identified the rs12688220 locus, which is close to the CLDN2 and MORC4 genes, as a risk locus for chronic pancreatitis in North American patients of European ancestry. 4 Although subsequent analysis has assigned rs12688220 to the MORC4 gene, additional analyses have identified multiple SNPs, including rs4409525, rs12008279 , and rs7057398 within the CLDN2 locus that are associated with increased risk of pancreatitis. 5 – 10 Despite such genetic evidence, how these SNPs may impact pancreatitis development remains unclear, as none of these SNPs are located within the coding region of the CLDN2 gene. Furthermore, these studies do not directly address the physiological and pathophysiological roles that CLDN2 may have during pancreatitis development. Claudins are a 27-member family of tight junction proteins that regulate paracellular permeabilities of small molecules, particularly ions and water. While some claudins form barriers to limit small molecule transport, others create size- and charge-selective pores that allow specific molecules to pass through the tight junction, thus controlling paracellular transport. 11 Expression of distinctive combinations of claudin family members is thought to contribute to unique transport properties of different epithelia. It has been shown that CLDN2 monomers polymerize within a cell membrane to form strands. 12 Trans-interactions of these strands between cells allow the formation of cation-selective ion channels. It has been suggested that claudin channels are formed from four monomers and CLDN2 channels control paracellular Na + , Ca 2+ , and water transport. 13 – 17 The increase in CLDN2 expression during inflammation is well established, particularly in diseases of the gastrointestinal tract such as inflammatory bowel disease, infectious colitis, and celiac disease. 18 – 23 Cytokines, including IFN-γ, IL-6, TNFα, IL-13, and IL-22, are known to modulate the expression of CLDN2 in response to tissue inflammation in a number of different cell types, contributing to changes in paracellular transport and barrier function. 24 – 30 Given the diverse roles of claudins in regulating paracellular transport, the role of CLDN2 in pancreatic ductal epithelial cells, particularly under inflammatory conditions such as pancreatitis, warrants further investigation. We demonstrate that CLDN2 expression is significantly upregulated in pancreatic ductal epithelium from patients with chronic pancreatitis, as well as in multiple acute and chronic mouse experimental models. Our results show that pancreatitis-associated cytokines contribute to the upregulation of CLDN2 in pancreatic ductal epithelial cells. Studies using Cldn2 knockout (KO) mice reveal increased susceptibility to caerulein-induced pancreatic injury, and Cldn2 KO diminishes fluid secretion across pancreatic ductal epithelium. Taken together, our data show that increased CLDN2 expression limits pancreatitis development through regulating paracellular fluid transport in pancreatic ducts.