Extracellular pH and NHE1 Regulate Ductal Branching Morphogenesis in Organotypic Cultures of Human Pancreatic Duct Epithelial Cells

TL;DR

This study shows that changes in pH and a transporter called NHE1 influence how pancreatic duct cells form branching structures in a lab setting.

Contribution

The novel finding is that extracellular pH and NHE1 activity regulate ductal branching in human pancreatic duct epithelial cells in a 3D organotypic model.

Findings

Ductal morphogenesis is influenced by acidic extracellular pH (pHe 6.7), leading to hyper-branching.

NHE1 inhibition further increases branching, indicating its regulatory role.

ECM composition affects ductal branching, with Matrigel-rich ECM promoting branching and Collagen I-rich ECM inhibiting it.

Abstract

Branching morphogenesis is a key process for constructing the tree‐like architecture of multiple organs. The mechanisms regulating pancreatic ductal morphogenesis are still poorly understood, especially in the context of the particular pH dynamics of this organ. Indeed, ductal cells periodically release an alkaline juice to balance stomach acidity during digestion. This leads to a drop in extracellular pH (pHe) in the extracellular matrix (ECM) to maintain intracellular pH (pHi) homeostasis. Among the transporters involved in pH regulation, NHE1 also regulates epithelial branching morphogenesis in various tissues/organs. However, neither the effect of the changing pHe nor the role of NHE1 in branching morphogenesis has been investigated in a physiomimetic model in the human pancreas. Here, using 3D organotypic cultures of human pancreatic ductal cells (HPDE), we found that cells seeded on a Matrigel rich‐ECM resembling normal ECM formed branched duct‐like structures, which did not form on a more fibrotic Collagen I‐rich ECM. Further, these cells overexpressed NHE1 mainly at the basolateral membrane. Ductal morphogenesis was affected by acidic pHe (pHe 6.7), which determined a hyper‐branched network, and this was further increased by the inhibition of NHE1. We conclude that ECM composition and extracellular acidosis modulate branching morphogenesis in pancreatic ductal HPDE cells via NHE1 activity.

Figures

ECM composition affects pancreatic ductal epithelial cell growth rate and morphology. Growth rate with Resazurin (A) and morphology (B) of HPDE cells cultured on 2D and 3D organotypic cultures composed of 90%Matrigel:10% Collagen I and 70% Matrigel: 30% Collagen I. Data are mean ± SEM, n = 5. Significance: *p < 0.001 compared to day 1 in each ECM. Scale bar = 500μm. (C) Laser scanning confocal microscopy (LSCM) of representative 5‐day, organotypic cultures of HPDE cells labeled with BCECF–AM. Cross‐sections of the z‐stack (upper) and reconstructed z‐stack (lower) by ImageJ software show acini and tubes with lumens in cells on 90%M: 10%C and a thin monolayer in cells on 70%M: 30%C. Scale bar = 50μm.

Extracellular pH (pHe) affects branching morphology. (A) HPDE cells grown on 90% Matrigel‐10% Collagen I in their growth medium at different pHe values of 7.4, 7.0 and 6.7 organize into branched tubules that connect into nodes and form closed lacunae. Mean number of lacunae (B), branches (C) and (D) nodes in cells cultured at pHe 7.4, pHe 7.0 or pHe 6.7 on 90 M:10 C for the indicated times. *p < 0.05, p < 0.01 in comparison to cells at pHe 7.4 on the same day. Scale bar = 500μm.

NHE1 basolateral membrane expression is higher in cells growing on 90% M:10% C, where it plays a role in branching dynamics. NHE1 mRNA (A) and protein localization (B, red) in confocal images in HPDE cells grown for 7 days as organotypic cultures on 90 M:10 C and 70 M:30 C. In (B), the apical marker CFTR is in green and orthogonal views of both NHE1 and CFTR are reported. Effect of the specific NHE1 inhibitor, Cariporide, on lacunae (C), branch (D), and node (E) formation over time. Cells were grown on 90 M:10 C and 1 μM Cariporide was added 1, 4, and 24 h after cell seeding. *p < 0.05, p < 0.01, *p < 0.001 when compared to control cells of the same day. Scale bar: 50μm.

NHE1 inhibition and extracellular acidosis together create a more stable and hyper‐branched ductal network. Mean number of lacunae (A), branches (B), and nodes (C) in cells cultured at pHe 7.4, pHe 7.0, or pHe 6.7 together with 1 μM Cariporide (applied 24 h after seeding) on 90%M:10%C for the indicated times. In (D), the polygonal features of the lacunae are shown. *p < 0.05, p < 0.01, *p < 0.001, p < 0.0001 and #p < 0.05, ##p < 0.01, ####p < 0.001 and ####p < 0.0001 in comparison to cells treated with Cariporide at pHe 7.4 and with Cariporide at pHe 6.7 within the same day, respectively.

Full text

Introduction

1

Branching morphogenesis is a conservative developmental process in which epithelial buds branch into the surrounding mesenchyme to form the complex epithelial tree‐like cellular network typical of multiple organs and tissues, including the nervous and the respiratory systems, lung, kidney, vasculature, mammary, prostate, and pancreatic glands (Goodwin and Nelson 2020). These networks are critical to organ function by maximizing organ surface area, necessary for absorptive and secretory functions (Shih et al. 2016; Hannezo and Simons 2019; Ingthorsson et al. 2025). Therefore, it is essential to understand how these cellular networks arise and are dynamically modified throughout development and eventually during pathological conditions such as fibrosis and cancer.

Regulation of the branch patterning is a complex process and requires both reciprocal interactions between branching epithelia and their surrounding extracellular matrix (ECM) and the coordinated integration of molecular signal transduction with bio‐physical signals from the surrounding microenvironment (Cozzitorto and Spagnoli 2019; Goodwin and Nelson 2020; Wu et al. 2023). Although these molecular and signaling requirements for branching have been better studied in other organs, the role of some environmental factors in the regulation of the development and spatial patterning of the pancreatic branches is missing, including changes in extracellular pH (pHe). Indeed, a physiological characteristic that sets the pancreas apart from the other branched organs is the cyclic pHi/pHe changes occurring multiple times each day, where the ductal epithelial cells secrete large amounts (2−3 L/day) of bicarbonate‐rich fluid (HCO_3_ ^−^, ≈150 mM) into the ductal lumen. Alkaline secretions maintain inactivated the acinar cells' proteases before they reach the duodenum while neutralizing gastric acid in the duodenum and setting optimal pH for pancreatic enzyme function (Novak et al. 2013; Lee et al. 2020). However, this alkaline secretion leads to a parallel acidification of the pancreatic interstitium such that the ductal cells are exposed to variable gradients of pHe that are translated into specular variations of pHi (Pedersen et al. 2017).

On the basolateral membrane of epithelial cells, including pancreatic ductal epithelium cells, the Na^+^/H^+^ Exchanger isoform 1 (NHE1) plays a fundamental role in the regulation of pHi homeostasis (Reshkin et al. 2012; Jenkins et al. 2012) and indirectly in the mechanism of HCO_3_ ^‐^‐rich fluid secretion during the digestive phase. It is activated by intracellular acidification to restore pHi via H^+^ extrusion especially when HCO_3_ ^‐^ secretion is reduced and this process is dependent on the ECM composition (Alpern et al. 1993; Di Molfetta et al. 2023). We have, indeed, demonstrated that in normal pancreatic cells the NHE1 ability to regulate pHi is higher when the cells grow on a physiological, laminin‐rich ECM and decreases on a more fibrotic, Collagen I‐rich ECM (Di Molfetta et al. 2023). Importantly, NHE1 is also involved in the regulation of branching morphogenesis in various tissues (Rampino et al. 2007; Yu et al. 2008; Jenkins et al. 2012; Jenkins et al. 2014; Sin et al. 2020), but its role in pancreatic branching morphogenesis is unknown. While these considerations suggest that the ability of pancreatic ductal cells to form branching morphogenesis could depend on ECM composition, pHe, and NHE1, their potential involvement/participation in branching morphogenesis has not been investigated in a physiomimetic model of the human pancreas.

Here, we have established a 3D organotypic model of HPDE (Human Pancreatic Duct Epithelial) cells to replicate the normal human pancreatic ductal architecture (Di Molfetta et al. 2023) and investigate the effects of ECM composition, acidic pHe and NHE1 activity on branching morphogenesis. We found that on a Matrigel‐rich ECM resembling normal ECM, ductal cells formed branched duct‐like structures, which did not form on a more fibrotic Collagen I rich‐ECM, resembling a pathological condition such as chronic pancreatitis and pancreatic cancer (PDAC) (Huang et al. 2021; Usman et al. 2023; Li et al. 2025). Moreover, in cells lining the ducts on a Matrigel‐rich ECM, NHE1 was overexpressed on the basolateral membrane. Lastly, both NHE1 inhibition with its specific inhibitor, Cariporide, and cell exposition to acidic pHe (pHe 6.7) resulted in hyperbranched morphogenesis and a greater stability of the tubular reticular structure. Importantly, cell exposition to acidosis during NHE1 inhibition resulted in an additional increase of the branching complexity/stability, indicating that NHE1 is involved both in the basal and acidic pHe‐stimulated branching morphogenesis.

Results

2

Pancreatic Cells Growing on a Physio‐Mimetic ECM Form Branched Duct‐Like Structures

2.1

Here, we studied the pancreatic ductal network using a 3D organotypic culture of the human pancreatic ductal HPDE cells on ECM compositions with increasing Collagen I concentrations in‐order‐to recapitulate Collagen I enrichment in the ECM during the transition from a normal to a fibrotic pancreas (Huang et al. 2021). For this, HPDE cells were seeded in 2D conditions and in an organotypic setup consisting of two ECM mixes: 90% Matrigel:10% Collagen I (90 M:10 C) which is a laminin‐rich ECM biomimetic of the normal pancreatic ECM (Ma et al. 2019; Vigier et al. 2017) and 70% Matrigel:30% Collagen I (70 M:30 C), mimicking the increase in normal pancreatic tissues of fibrotic components (Usman et al. 2023; Li et al. 2025). Cell growth was analyzed with the Resazurin assay (Figure 1A) and cellular morphology was monitored (Figure 1B). As shown in Figure 1A, in 2D cells grew as a monolayer and exhibited a faster growth rate than in 3D conditions. Interestingly, cells grew faster on 90 M:10 C while the enrichment of Collagen I into the ECM mix reduced their growth as described for various cell types (Duval et al. 2017). Interestingly, when grown on 90 M:10 C, cells formed branched duct‐like structures, which did not form on 70 M:30 C ECM (Figure 1B). This loss of ductal morphology with increasing Collagen I in the matrix resembles the structural changes detected during the tissue reorganization from a normal pancreas to pancreatitis and during PDAC progression (Orth et al. 2019) and indicates the role of the ECM composition as an actor of tissue morphogenesis.

*ECM composition affects pancreatic ductal epithelial cell growth rate and morphology. Growth rate with Resazurin (A) and morphology (B) of HPDE cells cultured on 2D and 3D organotypic cultures composed of 90%Matrigel:10% Collagen I and 70% Matrigel: 30% Collagen I. Data are mean ± SEM, n = 5. Significance: **p < 0.001 compared to day 1 in each ECM. Scale bar = 500μm. (C) Laser scanning confocal microscopy (LSCM) of representative 5‐day, organotypic cultures of HPDE cells labeled with BCECF–AM. Cross‐sections of the z‐stack (upper) and reconstructed z‐stack (lower) by ImageJ software show acini and tubes with lumens in cells on 90%M: 10%C and a thin monolayer in cells on 70%M: 30%C. Scale bar = 50μm.

Next, we analyzed the morphological 3D organization of HPDE cells on the different ECMswith Laser Scanning Confocal Microscopy (LSCM) after labeling the cells with the intracellular fluorescent dye BCECF–AM (Figure 1C). On a more physiological matrix composition (90 M:10 C), the branched duct‐like structures of HPDE cells observed in Figure 1B were, indeed, a network of hollow ducts, as they are in vivo. With time, these tubes elongated and bifurcated into two or more smaller tubes, which eventually further split into other smaller branches to form a very complex tubular network. When cultured on 70 M:30 C, the cells grew as a tight, cobblestone‐like single layer of epithelial cells. The relative thickness of the tubules, acini, and monolayers shown in the Y‐Z projections of Figure 1C is shown in Supporting Information S1: Figure 1A.

Extracellular Acidosis Promotes the Formation of a More Stable and Hyper‐Branched Ductal Network

2.2

To investigate whether pHe also affects the development of the pancreatic tubular structures and/or their dynamic remodeling, HPDE cells were seeded on 90 M:10 C at the pHe values of 6.7, 7.0, and 7.4. The morphogenic response of the branched network to the different pHes was quantified by counting the number of (i) lacunae formed/enclosed by the tubular network, (ii) nodes, and (iii) branches (Figure 2A). As shown in Figure 2B−D and in Supporting Information S1: Figure 2, while the number of lacunae, branches and nodes slightly decreased over time at pHe 7.4 as a consequence of the dynamic remodeling of the ductal network, culturing the cells in ever more acidic mediums stepwise increased the number of all three characteristics compared to cells at pHe 7.4, increasing network complexity and stabilizing the branching tubular structure over time.

*Extracellular pH (pHe) affects branching morphology. (A) HPDE cells grown on 90% Matrigel‐10% Collagen I in their growth medium at different pHe values of 7.4, 7.0 and 6.7 organize into branched tubules that connect into nodes and form closed lacunae. Mean number of lacunae (B), branches (C) and (D) nodes in cells cultured at pHe 7.4, pHe 7.0 or pHe 6.7 on 90 M:10 C for the indicated times. *p < 0.05, *p < 0.01 in comparison to cells at pHe 7.4 on the same day. Scale bar = 500μm.

NHE1 Expression Depends on the ECM Composition and its Inhibition Increases Pancreatic Branching Complexity and Stability

2.3

As NHE1 plays a role in tissue branching morphogenesis, we investigated whether it also contributes to pancreatic ductal branching. We first determined if its expression is regulated by changes in ECM composition. HPDE cells were grown on 90 M:10 C and 70 M:30 C for 7 days, and NHE1 mRNA (Figure 3A) expression was quantified by RT‐qPCR. We found that mRNA levels of NHE1 were higher in cells grown on 90 M:10 C compared to 70 M:30 C. Confocal immunofluorescence was used to determine the localized expression of NHE1 and the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) in cells grown on the different ECMs. In Figure 3B, a typical staining for both CFTR (green) and NHE1 (red) in HPDE cells grown on both 90 M:10 C and 70 M:30 C showed an apical distribution of CFTR and basolateral localization of NHE1 on the tight epithelium lining the ducts consistent with their reported distribution in the pancreatic duct epithelium (Kong et al. 2014) and confirmed the higher NHE1 protein expression when grown on 90 M:10 C. Supporting Information S1: Figure 1B showing the staining for both the tight junction marker ZO‐1 (green) and NHE1 (red) in HPDE cells grown on both 90 M:10 C and 70 M:30 C demonstrated identical localization for NHE1. To analyze whether NHE1 also affects the development of the pancreatic tubular structures and their dynamic remodeling, HPDE cells were seeded on 90 M:10 C and NHE1 function was blocked with its specific inhibitor, Cariporide (1 µM) at 1, 4, and 24 h after cell seeding We have previously observed that this Cariporide concentration inhibited NHE1 activity in HPDE cells in the presence of NaHCO3 by approximately 74% (Di Molfetta et al. 2023). Pancreatic ductal cell organization in a branched, tubular epithelial tree‐like network and their morphogenic response to NHE1 inhibition by Cariporide were quantified by analyzing the number of lacunae (Figure 3C), branches (Figure 3D), and nodes (Figure 3E) as described above for pHe. As for the exposure of the cells to an acidic pHe, the inhibition of NHE1 with Cariporide increased network complexity, as indicated by the higher number of lacunae, nodes, and branches compared to the control conditions. Importantly, in experiments lasting 15 days, Cariporide (1 µM) did not affect cell viability compared to control cells (Supporting Information S1: Figure 3A), suggesting that NHE1 inhibition could increase ductal network complexity via a stabilization of the ductal meshwork and/or a reduction of its remodelling, rather than through an increase in cell viability/proliferation.

*NHE1 basolateral membrane expression is higher in cells growing on 90% M:10% C, where it plays a role in branching dynamics. NHE1 mRNA (A) and protein localization (B, red) in confocal images in HPDE cells grown for 7 days as organotypic cultures on 90 M:10 C and 70 M:30 C. In (B), the apical marker CFTR is in green and orthogonal views of both NHE1 and CFTR are reported. Effect of the specific NHE1 inhibitor, Cariporide, on lacunae (C), branch (D), and node (E) formation over time. Cells were grown on 90 M:10 C and 1 μM Cariporide was added 1, 4, and 24 h after cell seeding. *p < 0.05, **p < 0.01, **p < 0.001 when compared to control cells of the same day. Scale bar: 50μm.

Extracellular Acidosis Increases the Effect of NHE1 Inhibition on Branched Ductal Network Development

2.4

As NHE1 activity regulates both pHi and pHe, we next investigated whether the Cariporide‐dependent inhibition of NHE1 and the simultaneous exposure of cells to extracellular acidosis could modify pancreatic ductal morphogenetic branching. As shown in Figure 4A−C, cells cultured for up to 7 days at either pHe 7 or pHe 6.7 in the presence of 1 µM Cariporide further stimulated the branching parameters already increased by the presence of Cariporide at pHe 7.4, indicating that extracellular acidosis facilitates the increased branching morphogenesis caused by NHE1 inhibition. We expanded this characterization of the network complexity by evaluating other network properties as per Dahl‐Jensen et al. 2018, where they used in silico models to map/reconstruct the development of the ductal network in the in vivo fetal pancreas by quantifying the particular type of polygonal shape assumed by the lacunae (i.e, triangles, squares, pentagons, and hexagons) during the in vivo development of the ductal system. As can be seen in Figure 4D, in our organotypic system, these polygonal properties of the network increase over time, particularly under acidic conditions and especially when Cariporide inhibits NHE1 at acidic pHe, supporting the hypothesis that both the acidic microenvironment alone and especially together with NHE1 inhibition increase the formation of a highly interlinked network, typical of the early‐intermediate stages of pancreatic ductal morphogenesis

*NHE1 inhibition and extracellular acidosis together create a more stable and hyper‐branched ductal network. Mean number of lacunae (A), branches (B), and nodes (C) in cells cultured at pHe 7.4, pHe 7.0, or pHe 6.7 together with 1 μM Cariporide (applied 24 h after seeding) on 90%M:10%C for the indicated times. In (D), the polygonal features of the lacunae are shown. *p < 0.05, **p < 0.01, ***p < 0.001, ***p < 0.0001 and #p < 0.05, ##p < 0.01, ####p < 0.001 and ####p < 0.0001 in comparison to cells treated with Cariporide at pHe 7.4 and with Cariporide at pHe 6.7 within the same day, respectively.

Altogether, these data indicate that normal pancreatic HPDE cells create arborized epithelial ductal networks only when grown on hydrogels mimicking the laminin‐rich, natural pancreatic ECM (90% Matrigel:10% Collagen I) and that the regulation of pancreatic morphogenesis is affected not only by the stromal compartment (ECM composition) but also by the metabolic microenvironment (extracellular acidosis) via the involvement of NHE1.

Discussion

3

The function of the mammalian pancreas is intimately linked to its highly branched and interconnected architecture (Shih et al. 2013; Henley and Gannon 2014; Dahl‐Jensen et al. 2018). While a number of key growth factors and their signaling cascades are essential for the induction, regulation, and development of the pancreatic branched epithelial ducts (Shih et al. 2016), the processes governing pancreatic branching morphogenesis remain incompletely understood. At the cellular level, cross‐regulation of integrin‐mediated cell‐ECM interactions and E‐cadherin‐mediated cell‐cell adhesions (Serrill et al. 2018) and laminin deposition (Heymans et al. 2019) has been shown. In this study, we characterized the morphogenesis of pancreatic branching in an in vitro organotypic 3D model by identifying a novel mechanism of branch formation, which involves both extracellular acidosis and one of the main pHi/pHe regulators, the NHE1.

We first evaluated the role of different ECM compositions on the growth and morphology of human pancreatic ductal cells, the HPDE cells. For these purposes, HPDE cells were seeded in 3D ECM mixes composed of 90% Matrigel:10% Collagen I (90 M:10 C) or 70% Matrigel:30% Collagen I (70 M:30 C). Growth analyzed by the Resazurin assay showed that cells grew faster on a laminin‐rich ECM (Matrigel) resembling the normal ECM, while Collagen I enrichment in the ECM reduced their growth rate. Moreover, confocal microscopy analysis revealed that cells growing on a Matrigel‐rich ECM formed well‐shaped, interconnected tubes with a central, empty lumen, connected by large acinar‐like structures. The tubules and acini were hollow and around 35−45 µM and 90 µM in height, respectively, while the monolayers on 70 M:30 C were approximately 10 µM thick (Supporting Information S1: Figure 1A). We conclude that HPDE cells grown on hydrogels mimicking the laminin‐rich, natural pancreatic ECM (90 M:10 C) form arborized epithelial ductal networks, whereas on collagen I‐enriched ECM (70 M:30 C), they lose their ability to form ducts/acini. This mimics the in vivo situation where the epithelial ducts are embedded in a well‐defined stroma, which plays an important role in the construction and maintenance of these highly ordered branched structures (Wu et al. 2023) but, when altered, can equally contribute to cancer development and progression. Indeed, in pancreatic ductal adenocarcinoma (PDAC), Collagen I accumulation in the ECM and the consequent increase in stromal stiffness stimulate malignant cell properties, such as disorganized proliferation, invasion, immune tolerance, and chemoresistance (Henke et al. 2020).

The human pancreas is distinct from other branched organs due to the cyclic pHi/pHe changes occurring multiple times each day with an alkaline secretion that leads to a parallel acidification of the pancreatic interstitium, such that the ductal cells are exposed to periodic acidic pHe. For this reason, we tested the effect on branching morphogenesis of the incubation at the acidic pHes of 7.0 and 6.7 of the cells growing on 90M−10C. Indeed, culturing these branching structures in a more acidic medium resulted in a stepwise increase in complexity and stability of the tubular network characterized by a higher number of lacunae, nodes, and branches compared to the control condition (Figure 2).

We next analyzed the expression and role of NHE1, since it has been shown to play a role in branching morphogenesis in various organs (Jenkins et al. 2012; Jenkins et al. 2014; Rampino et al. 2007; Yu et al. 2008; Sin et al. 2020). We found that NHE1 is overexpressed in cells forming ductal tubes on 90 M−10 C compared to the monolayer growing on 70 M:30 C and is localized at the basolateral membrane, as shown by the orthogonal view of HPDE Z‐stacks (Figure 3A,B, and Supporting Information S1: Figure 1B). Maintenance of this architecture enables apico‐basal cellular polarity, which is essential for proper luminal secretion and the pancreas's primary function. Importantly, as shown in Figure 3, NHE1 has a role in HPDE branching, as its inhibition with its specific inhibitor, Cariporide (1 µM), resulted in hyperbranching; that is, a greater number of lacunae, ducts, and branches to increase stability of the tubular reticular structure, which maintained a high degree of complexity (Figure 3C−E). Interestingly, simultaneously inhibiting NHE1 together with acidic pHe of 6.7 increased all the measured network parameters compared to the control, but did not result in an additional increase in any of them compared to pHe of 7.0, suggesting that extracellular acidity governs three‐dimensional branching morphogenesis through the involvement of NHE1 (Figure 4A−C). Since it has been suggested that increased proliferation is necessary to drive new duct formation (Lee et al. 2024), we determined if this could account, at least in part, for the increased network complexity observed, especially when NHE1 was inhibited in acidic conditions. To this end, we measured cell viability in cells grown at different acidic pH values, both in the absence and in the presence of Cariporide, for up to 5 days. As shown in Supporting Information S1: Figure 3B,C, extracellular acidosis increases cell viability compared to control cells, with an even greater effect when Cariporide inhibits NHE1. In line with Lee et al. (2024), this increased cellular viability could partially explain the expansion in the number of ducts, nodes, and lacunae, as well as the general increase in the complexity of morphogenetic branching induced by acidic pH in the presence of Cariporide.

Altogether, these data indicate that normal pancreatic HPDE cells form arborized epithelial ductal networks only when grown on hydrogels that mimic the laminin‐rich natural ECM (90 M:10 C) and that the regulation of pancreatic morphogenesis in HPDE cells is influenced by ECM composition and extracellular acidosis, which together modulate branching behaviour through NHE1 activity.

Methods

4

Cell Lines

4.1

Human Pancreatic Duct Epithelial cells (HPDE‐H6c7) are immortalized epithelial cells derived from the normal human pancreatic duct epithelial HPDE cells and are used as a model of the normal pancreatic ductal epithelium. HPDE were grown in a mixture of 50% RPMI 1640 (Gibco, Life Technologies), supplemented with 10% FBS, 100 U/mL penicillin G, 0.1 mg/mL streptomycin, and 1% non‐essential amino acids 100X solution (Gibco, Life Technologies) and 50% keratinocyte medium SFM (Gibco, Life Technologies) supplemented with 0.025% bovine pituitary extract, 2.5 mg/L epidermal growth factor (Gibco, Life Technologies). The cell lines were maintained at 37°C in 5% CO_2_. Cariporide was purchased from MedChemExpress (DBA, Italy).

Acidic Low pH Medium

4.2

To adjust the pH of the pHe HPDE culture medium to 7 and 6.7, the medium was supplemented with NaHCO_3_ according to the Henderson‐Hasselbalch equation. The osmolarity of the medium was balanced using NaCl. The initial and final pHs were measured by using the WTW InoLab Benchtop pHmeter.

3D Organotypic Cultures

4.3

The matrix compositions were based on data from the healthy pancreas (Ma et al. 2019; Vigier et al. 2017). Two different mixtures composed of Matrigel Basement Membrane Matrix (Corning) and Collagen I (bovine; Gibco, Life Technologies) were prepared as previously described (Cannone et al. 2022; Forciniti et al. 2021; Biondani et al. 2018). Matrigel and Collagen I were first diluted to a concentration of 7 mg/mL in serum‐free media for Matrigel and 3 mg/mL in distilled sterile water, 10X PBS (Sigma Aldrich) and 0.015 N NaOH for Collagen I and then mixed at 90% Matrigel‐10% Collagen I (90 M−10 C) and 70% Matrigel‐30% Collagen I (70 M−30 C) as previously described (Biondani et al. 2018). A thick (250 μL) layer of the mixes was then used to coat 12 mm glass coverslips in 24‐well cell culture plates, which were then incubated at 37° C with 5% CO_2_ for 1 h in order to allow the mixtures to gel. For the branching morphogenesis experiments, 9 × 10^^4^ cells/well were seeded onto the matrix composed of 90%M:10%C and incubated at 37°C in 5% CO_2_. Preliminary experiments at three different cell seeding concentrations (9 × 10^4^ cells, 4.5 × 10^4^ cells and 1.5 × 10^4^ cells) and five time points starting at 16 h until 5 days, demonstrated that the starting cell density does not affect the formation of the tubes (which form over time regardless of the initial number of cells seeded), but rather the dynamics of their development. Indeed, higher initial cell densities lead to faster cell alignment into tube‐like structures, which interconnect more rapidly and generate an increasingly complex branched network over time (Supporting Information S1: Figure 4). The effect of the different pHe's on branching parameters, either in the absence or in the presence of Cariporide, was examined by imaging each well every day for 7 days and analyzing the images with the ImageJ software.

RNA Extraction and qPCR

4.4

Cells were recovered from each ECM as described above, and total RNA was extracted using the RNeasy Mini Kit (Qiagen, Cat# 74104). RNA was then reverse‐transcribed into cDNA with the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Cat#K1621), following the manufacturer's protocols, as described (Nisco et al. 2025). Real time quantitative PCR was performed using SsoAdvanced Universal SYBR Green Supermix and a custom Primer PCR Assay Plate (cat. n. 10025217, Bio‐Rad), which included validated primer pairs, lyophilized in the well, for SLC9A1/NHE1 (Unique Assay ID: qHsaCID0012512), ACTB (Unique Assay ID: qHsaCED0036269), and HPRT1 (Unique Assay ID: qHsaCID0016375). All samples were measured in triplicate, with at least three biological replicates. ACTB and HPRT1 were used as housekeeping genes. SYBR Green gene expression assays were performed with the CFX96 Touch Real‐Time PCR Detection System (Bio‐Rad Laboratories, Inc., Hercules, CA, USA), using the following cycling protocol: denaturing at 95°C for 3 min, followed by 40 cycles of denaturing at 95°C for 10 s and annealing/extension at 60°C for 40 s. A melt curve was performed at 65°C to 95°C with 5 s/step (0.5°C increments), following the cycling protocol. Gene expression levels were normalized to the housekeeping genes, ACTB and HPRT1, and relative gene expression was calculated using the 2‐ΔΔCT (comparative threshold) method with CFX Maestro software.

Morphological Analysis by LCSM

4.5

A 24‐well cell culture plate was prepared by placing a 15 mm diameter sterile glass in each well. 90 M−10 C and 70 M−30 C mixes were prepared as above, and, respectively, 2 × 10^5^ and 7.5 × 10^4^ cells/well were seeded onto the two ECMs. Five days after seeding, cells were incubated with 10 μM BCECF–AM (Thermo Fisher, Waltham, MA, USA) at 37°C in the dark for 30 min, then washed twice with PBS to remove excess dye. Image acquisition was performed using a Leica TCS SP5 confocal microscope, with the following settings in the LAS AF software: laser Ar 40%, λexc = 488 nm, and emission between 495 nm and 550 nm; XYZ. Images were captured using an X25 wet objective with a 512 × 512 pixel resolution, 0.59 μm Z‐stack distance. For 3D image reconstructions, the VolumeViewer and Isosurface plugins of the ImageJ software (Fiji image processing package) were used.

Immunofluorescence Staining

4.6

A 24–well cell culture plate was prepared by placing a 12 mm mm‐diameter glass in each well. Matrigel and collagen I were prepared as reported above. After matrix polymerization, 1.5 × 10^^4^ cells/well were seeded. On the 5th day, an IF assay was performed as described by Cardone et al. (2007). Primary antibodies, monoclonal anti‐CFTR (R&D Systems clone 24‐1, 1:50 in 1% BSA) and polyclonal anti‐NHE1 (Santa Cruz sc‐28758, 1:25 in 1% BSA) were incubated overnight at 4°C. The next day, cells are washed three times for 5 min with 1% BSA and incubated for 45 min at room temperature with Goat Anti‐mouse 488 (1:1000 in 1% BSA) and Goat Anti‐rabbit 585 (1:1000 in 1% BSA), both from Invitrogen.

Resazurin Viability Assay

4.7

To assess cellular metabolic activity in the 3D cultures, the Resazurin assay was performed according to the manufacturer's instructions. Briefly, a 10% Resazurin (Immunological Sciences) solution was prepared in complete cell culture medium and added to the 3D cultures, followed by 2–3 h of incubation at 37°C. At the end of the incubation period, a change in Resazurin fluorescence was measured using FLUOstar Omega, a microplate reader (BMG LABTECH, Germany) at 530 nm excitation and 590 nm emission. Resazurin (Immunological Sciences) reduction assay was performed at several time points.

Statistical Analysis

4.8

Data are shown as individual representative experiments or as means of at least three independent experiments, with standard error of mean (SEM) error bars. Statistical significance was assessed using a two‐way ANOVA followed by Tukey's multiple comparisons (Biondani et al. 2018).

Author Contributions

Daria Di Molfetta: methodology, investigation, writing original draft preparation, writing review and editing. Marilena Ardone: methodology, investigation. Francesca Fracasso: methodology, writing review and editing. Maria Raffaella Greco: methodology, investigation. Grazia Tamma: formal analysis. Mariangela Centrone: formal analysis, methodology. Maria Barile: formal analysis, investigation. Maria Tolomeo: methodology. Alessia Nisco: methodology. Stephan Joel Reshkin: conceptualization, writing review and editing, supervision, project administration. Rosa Angela Cardone: Conceptualization, writing review and editing, supervision, project administration, funding acquisition.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Suppl Figures updated 2_2_26.