Key role of down-regulated in adenoma (SLC26A3) chloride/bicarbonate exchanger in linaclotide-stimulated intestinal bicarbonate secretion upon loss of CFTR function

Duodenal bicarbonate secretion is critical to epithelial protection, nutrient digestion/absorption and is impaired in cystic fibrosis (CF). We examined if linaclotide, typically used to treat constipation, may also stimulate duodenal bicarbonate secretion. Bicarbonate secretion was measured in vivo and in vitro using mouse and human duodenum (biopsies and enteroids). Ion transporter localization was identified with confocal microscopy and de novo analysis of human duodenal single cell RNA sequencing (sc-RNAseq) datasets was performed. Linaclotide increased bicarbonate secretion in mouse and human duodenum in the absence of CFTR expression (Cftr knockout mice) or function (CFTRinh-172). NHE3 inhibition contributed to a portion of this response. Linaclotide-stimulated bicarbonate secretion was eliminated by down-regulated in adenoma (DRA, SLC26A3) inhibition during loss of CFTR activity. Sc-RNAseq identified that 70% of villus cells expressed SLC26A3, but not CFTR, mRNA. Loss of CFTR activity and linaclotide increased apical brush border expression of DRA in non-CF and CF differentiated enteroids. These data provide further insights into the action of linaclotide and how DRA may compensate for loss of CFTR in regulating luminal pH. Linaclotide may be a useful therapy for CF individuals with impaired bicarbonate secretion.


INTRODUCTION
Proximal duodenal bicarbonate secretion, together with pancreatic bicarbonate secretion, is integral to preventing intestinal mucosal injury from gastric secretions and establishing a non-acidic intraluminal environment for activation of digestive enzymes.This importance is highlighted by cystic fibrosis (CF) intestinal disease and malnutrition, where decreased bicarbonate secretion is central to disease pathogenesis (1,2).Duodenal bicarbonate secretion is highly regulated, driven by luminal acid and neurohormonal signaling, ultimately resulting in coordinated transepithelial bicarbonate secretion by the cystic fibrosis transmembrane conductance regulator (CFTR), chloride/bicarbonate exchange (e.g., downregulated in adenoma, DRA; putative anion transporter-1, PAT-1), and/or decreased proton transport through sodium/hydrogen exchange (e.g.NHE3) (3).Despite its physiologic importance, there are limited therapies targeting duodenal bicarbonate secretion.
Linaclotide is a Food and Drug Administration (FDA)-approved medication for the treatment of constipation-type irritable bowel syndrome and chronic idiopathic constipation whose effect relies on its structural similarity to the heat-stable enterotoxin of Escherichia coli (STa).Both STa and linaclotide bind guanylyl cyclase C (GC-C) receptors on the apical surface of enterocytes and cause increases in CFTR-dependent chloride secretion and inhibition of NHE3-mediated sodium absorption (4)(5)(6).Previously, we identified that STa stimulates duodenal bicarbonate secretion through CFTR and CFTR-independent means (7,8).Linaclotide has been investigated as a potential therapy for CF-related constipation (6,9,10), however, it is unclear if linaclotide stimulates duodenal bicarbonate secretion and if linaclotide could improve impaired intraluminal pH in CF patients by stimulating CFTR-independent bicarbonate secretion.
We undertook a series of experiments utilizing in vivo and in vitro bicarbonate secretion measurements in mice and humans to examine the effect of linaclotide on duodenal bicarbonate secretion.We coupled these experiments with scRNA-seq analysis of acid-based transporters in the human duodenum and confocal microscopy to characterize the brush border expression .CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is made The copyright holder for this preprint (which this version posted April 22, 2024.; https://doi.org/10.1101/2023.05.05.539132 doi: bioRxiv preprint of the DRA chloride/bicarbonate exchanger in human duodenum in the presence or absence of CFTR activity and/or expression.
Bicarbonate secretion and I sc were accompanied by small, but significant (P<0.05)changes in transepithelial resistance (Figure 1H).
We next examined the potential for linaclotide to stimulate bicarbonate secretion in human duodenum using endoscopically obtained duodenal biopsies from subjects without known acid-base disturbances or histologic evidence of duodenal disease.Biopsies were mounted in Ussing chambers and combined bicarbonate secretion and I sc measurements were performed, similar to Figure 1D-H.As seen in Figure 1I, 10 -7 M and 10 -5 M linaclotide stimulated significant increases in human duodenal bicarbonate secretion (P=0.006,n=11 and P=0.004, n=10, respectively), at similar magnitudes previously reported for STa in human duodenal biopsies (11).In contrast to mice, linaclotide did not stimulate a significant change in I sc in human biopsies (10 -7 M: P=0.071, n=11; 10 -5 M: P=0.186, n=10) (Figure 1J) or transepithelial resistance (Figure 1J, K).Thus, linaclotide stimulates electrogenic bicarbonate secretion in mouse duodenum and electroneutral bicarbonate secretion in human duodenum.
DRA expression in human duodenum.DRA and PAT-1 are the primary chloride/bicarbonate exchangers expressed in mouse intestine, with PAT-1 showing dominant RNA expression in the small intestine and DRA RNA being more abundant in the colon.(17) There is limited data regarding DRA expression in human duodenum.(18,19) We capitalized on two previously published human duodenum scRNA-seq datasets (20,21) (20) contained 702 crypt cells and 923 villus cells from 2 adult duodenal samples, with an average of 10,381 reads per cell.Data from one crypt sample ("Crypt 1") was excluded due to very low sequence reads (2,088 per cell).We verified accurate classification into crypt and villus cells by examining LGR5 expression and proliferation markers MKI67 and PCNA, which are generally restricted to the crypt region.These were all more abundant in the crypt group, with LGR5 being exclusively expressed in the crypt, but not villus, group (Supplementary Figure 2A-C).Using these two datasets, we examined expression of the key acid-base transporters in the duodenum, DRA (SLC26A3), PAT-1 (SLC26A6), CFTR, and NHE3 (SLC9A3).Our analysis showed that SLC26A3 mRNA is predominantly expressed in enterocytes (Figure . 4A), in both the crypt and villus groups (Figure 4E), with higher expression and expressed in more cells than SLC26A6, CFTR, or SLC9A3 (Figures 4A-D and     E-H).Of note, Crypt 2 had lower SLC26A3 expression than Crypt 3, which correlated with higher CFTR, MKI67, PCNA expression, more LGR5+ cells, and less SLC9A3+ cells, suggesting Crypt 2 contained more cells deeper within the crypts than Crypt 3 ( Figures 4E, G, H and Supplementary Figure 2).We confirmed DRA protein expression in human duodenal biopsies (n=3) using immunofluorescence imaging.The majority of DRA is localized to the intracellular apical region, with a smaller fraction present at the apical brush border (Figures 4I-K).Given DRA's functional independence from CFTR in mediating linaclotide-stimulated duodenal bicarbonate secretion, we quantified the proportion of SLC26A3-expressing cells that coexpressed CFTR.In the Elmentaite et al. dataset, 88.8% of SLC26A3-expressing enterocytes co-expressed SLC26A3 and CFTR (Figure 5a).Using the Busslinger et al. dataset to examine crypt vs. villus localization, 63.0% of SLC26A3 positive crypt cells co-expressed SLC26A3 and CFTR, whereas in the villus cells, only 29.6% did so (Figures 5B and C).These findings were not ubiquitous for all chloride/bicarbonate exchangers.SLC26A6 and CFTR co-expression was relatively low in enterocytes (16.7%) (Supplementary Figure 3A), but examination of the crypt and villus cell types showed a higher percentage of CFTR co-expression with SLC26A6 than SLC26A3, with 81.8% of crypt cells and 42.9% of villus cells co-expressing SLC26A6 and CFTR (Supplementary Figures 3B and C).In the Elmentaite et al. dataset, SLC9A3 and CFTR were co-expressed in 14.3% of enterocytes, similar to SLC26A6 (Supplementary Figure 3D).The Linaclotide alters the apical localization of DRA.Ion channel activity can be regulated by altering membrane expression of channels.We examined if linaclotide alters DRA activity by changing its apical brush border expression.For these experiments we utilized human derived three-dimensional duodenal enteroids.Given that the linaclotide receptor GC-C is expressed at the apical surface, we generated reverse polarity apical-out duodenal enteroids, which allow facile access to the apical membrane via the enteroid exterior (Supplementary Figures 4A and     B).( 22) Using qPCR, we verified that apical-out enteroids express CFTR, SLC26A3, SLC26A6, SLC9A3, and GUCY2C in undifferentiated and 3-day differentiated enteroids (Supplementary Figures 4C and D. 4).As expected, LGR5 was decreased in differentiated apical-out enteroids.
Using this model of human duodenum, we incubated apical-out differentiated enteroids with linaclotide (10 -7 M, 40 minutes) or vehicle control (water, 40 minutes) and performed confocal immunofluorescence to localize DRA.As seen in Figures 6A-D, linaclotide significantly increased apical brush border expression of DRA (n=12-16).In contrast, when we performed the same experiment with NHE3 localization, we found no change in the apical brush border expression of NHE3 upon linaclotide stimulation (Supplementary Figures 5A-D).We verified that myosin VI (MYO6), a key protein in NHE3 trafficking, (23) mRNA was present in apical-out enteroids.MYO6 mRNA was expressed at similar levels in this model as many of the other transporters we examined (Supplementary Figures 4C and D).To determine if linaclotide also stimulates DRA localization upon lack of functional CFTR, we repeated the experiments in the presence of CFTR inh -172 (2 x 10 -5 M, 40 minutes).Compared to vehicle control (DMSO, 40 minutes), CFTR inhibition resulted in an increase in the apical expression of DRA (P=0.032,n=26-31)(Figures 6E-H), but not NHE3 (P=0.633,n=16-30) (Supplementary Figures 5E-H).Of note, DMSO alone increased DRA brush border expression compared to water (MFI means ± SEM: 10.94 ± 1.32 vs. 3.43 ± 1.03, n=11-15, P<0.001).Linaclotide stimulation in the presence of CFTR inh -172 did not alter DRA or NHE3 brush border expression (P= 0.730, n=31; P=0.535, n=29-30, respectively) (Figures 6E-G and I, Supplementary Figures 5E-G and I).
To determine how our CFTR inh -172 studies compare to chronic loss of CFTR expression and function, we examined DRA brush border expression in human duodenal enteroids isolated from a CF subject with F508del homozygous CFTR variants.Comparison of baseline (water vehicle control, 40 minutes) DRA brush border expression between non-CF and CF enteroids showed CF enteroids had increased DRA expression at the apical brush border (Figures 6J-M  Inhibition of CFTR transiently increases intracellular pH.To investigate why CFTR inhibition increases DRA brush border expression, we examined the effect of CFTR inh -172 on duodenal enteroid intracellular pH (pH i ).We hypothesized that loss of CFTR activity would lead to intracellular bicarbonate trapping (resulting in increased pH i ) and DRA may traffic to the apical brush border as part of the cell's effort to restore bicarbonate transport and normalize the altered pH i .Undifferentiated and 3-day differentiated enteroids grown on coverslips and CFTR inh -172-induced (or DMSO) changes in pH i were monitored with live cell fluorescence imaging.CFTR inh -172 (2 x 10 -5 M) resulted in significant increases in pH i in undifferentiated and differentiated enteroids, although the slopes and magnitudes of these increases were different (Figure 7, n=57-89, P<0.001), with undifferentiated enteroids showing more rapid and greater pH i changes (P<0.001).DMSO (1:1000) had no effect on pH i (data not shown).Following CFTR inh -172, all enteroids underwent pH i recovery.Pre-treatment with DRA inh -A250 (10 -5 M, 5 min) did not alter the CFTR inh -172-induced increase in pH i but did alter the pH i recovery.In undifferentiated enteroids, DRA inh -A250 decreased the slope and magnitude of pH i recovery (Figures 7A-C, n=31, P<0.001).In differentiated enteroids, the magnitude of pH i recovery was not different with DRA inh -A250 (Figure 7F, n=37, P=0.213), however, the recovery slope was altered, albeit in a different way than undifferentiated enteroids (Figure 7D and E).Thus, inhibition of CFTR function causes an increase in pH i , the recovery of which is at least in part influenced by DRA activity.

DISCUSSION
Numerous studies in animal models and humans have shown the critical importance of duodenal bicarbonate secretion in maintaining epithelial integrity and creating the appropriate intestinal intraluminal pH for digestion.In CF, with genetic loss of CFTR, or Helicobacter pylori infection, with down-regulation of CFTR and PAT-1,(24) decreased duodenal mucosal bicarbonate secretion contributes to malabsorption and duodenal ulcers, respectively.While eradication of Helicobacter pylori with antibiotics may restore bicarbonate function, the most common way to improve small intestinal pH in CF is through chronic gastric acid suppression.
Ivacaftor, a CFTR potentiator, was shown to improve proximal intestinal pH in CF patients with the G551D CFTR gating mutation (accounts for 4-5% of CFTR disease-causing variants).(25) While additional CFTR modulators are FDA-approved, it remains unknown whether these positively impact small intestinal pH.This gap, coupled with lack of potential bicarbonatecorrecting therapy for CF patients ineligible for CFTR modulator therapies (due to genotype or medical contraindication), has driven us to seek ways to improve intestinal bicarbonate secretion independent of CFTR.Our current data are in line with prior mouse and human duodenum studies with STa (7,11) and a recent study by Tan et el.showing that linaclotide can stimulate jejunal bicarbonate secretion in mice.(9) Uniquely, we have shown that linaclotide stimulates both mouse and human bicarbonate secretion in the duodenum, where bicarbonate secretion is physiologically important.Furthermore, we have provided evidence for the roles (or lack of) membrane trafficking of DRA and NHE3 in this process and how decreases in CFTR function may impact cellular pH.
Linaclotide represents an attractive target for restoration of deficient duodenal bicarbonate secretion in CF.Our data in both mice and humans provide greater confidence in the translatability of our results to clinical care.While we did not study other segments of the small intestine, linaclotide stimulates jejunal bicarbonate secretion in F508del CF mice.( 9) Braga Emidio et al. identified that linaclotide's half-life in small intestinal fluid was 48 minutes, (26) suggesting that its effect is likely to be most influential in the proximal small intestine.Comparing linaclotide-stimulated and acid-stimulated duodenal bicarbonate secretion using the same methodology, (27,28) the linaclotide effect is about 75-100% of the stimulatory effect of acid, indicating linaclotide could produce a clinically meaningful impact for patients.
With existing FDA approval, and the use of clinically relevant dosing,(9) our results may provide immediate application to patients.
In the absence of CFTR, the logical targets for intestinal pH modulation are either 1) increasing chloride/bicarbonate exchanger activity, and/or 2) decreasing sodium/hydrogen exchange.Linaclotide increases intraluminal small intestinal fluid in mice by inhibiting NHE3dependent sodium absorption.(6,9) Another byproduct of NHE3 inhibition is a decrease in proton secretion, which can raise intraluminal pH.Consistent with this, NHE3 inhibition in our studies "increased bicarbonate secretion," likely by producing less luminal protons to neutralize secreted bicarbonate anions.In our NHE3 trafficking studies, we did not see a significant decrease in NHE3 membrane localization with linaclotide treatment.NHE3 activity may be impacted by shifting along the microvilli,(23) however, our imaging resolution was unable to determine these potential nanometer shifts.Nonetheless, our in vivo data supports that NHE3 inhibition alone does not account for linaclotide's ability to increase duodenal pH.Our studies  30) Thus, we hypothesized that upon CFTR blockade (or genetic lack of CFTR), the rise in pH i may trigger an increase in DRA expression and protein localization to at least partially restore bicarbonate secretion and limit alterations in pH i .Our current pH i data with CFTR inh -172, supports this hypothesis.Bijvelds et al. reported a nearly 2,000-fold increase in DRA mRNA in a CF ileal biopsy compared to non-CF ileum (n=1 each).(31) The lack of additional membrane trafficking of DRA by linaclotide in the presence of CFTR inhibition may indicate that DRA trafficking requires CFTR function or that a ceiling affect for DRA membrane expression has already been achieved.Based on our current data we favor the latter interpretation since linaclotide increased DRA trafficking in F508del homozygous CF enteroids and a comparison of MFIs across conditions showed that while the relative magnitude of DRA trafficking may differ between vehicle controls/conditions, the absolute magnitudes are similar.
While CFTR inhibition caused similar pH i changes in undifferentiated (crypt-like) and differentiated (villus-like) enteroids, how DRA inhibition impacted pH i recovery was different.
This may be due to increased expression of a larger variety of acid-base transporters in differentiated enteroids that may also contribute to pH i normalization.(19) Further study on how the expression and activities of DRA and other acid-base transporters are altered in CF intestine, especially in human samples, may identify new potential therapies for CF intestinal disease.
Prior to our study, our knowledge on the intestinal expression profile of DRA and other acid-base transporters was largely based on mouse studies.Wang et al. reported high expression of Slc26a3 (DRA) RNA in the colon, but negligible RNA expression in the small intestine (duodenum, jejunum, ileum).Conversely, Slc26a6 (PAT-1) RNA expression was high in mouse duodenum, jejunum, and ileum, with minimal in the colon.(17) Yet, Slc26a3 KO mouse studies suggest a prominent role for DRA in basal and stimulated bicarbonate transport.(32,33) Human studies are sparse.Yin et al. used human duodenal enteroids to examine SLC26A3, SLC26A6, and SLC9A3 (NHE3) mRNA expression in undifferentiated and terminally differentiated human duodenal enteroids.(19) Taking advantage of published scRNA-seq datasets from the human duodenum, we have provided a comparison on the expression of SLC26A3, SLC26A6, and SLC9A3, from human duodenum tissue.SLC26A3 mRNA was generally expressed ~3-4x higher than SLC26A6, CFTR, and SLC9A3.In the CFTR high expressing BEST4+ cells, CFTR expression levels approached that of SLC26A3.Of note, SLC26A3 was also highly expressed in BEST4+ cells, although enterocytes comprise the majority of SLC26A3-expressing cells.These findings, put in context of prior studies, suggest a more prominent role for DRA in the human duodenum than may have generally been appreciated from mouse studies alone.Future work examining not only the mRNA expression, but also protein expression, and especially pertinent to membrane ion transporters, membrane localization of DRA, PAT-1, and NHE3 across different segments of human intestine may advance our knowledge regarding the regional roles of these acid-base transporters in health and disease.
A potential limitation of our study is the general reliance on pharmacologic inhibitors throughout our experiments.In Xenopus levis oocytes, CFTR inh -172 (2 x 10 -5 M) inhibited ~70% of human CFTR current in patch-clamp experiments (34).In our experiments, CFTR inh -172 (2 x 10 -5 M) significantly inhibited CFTR ion transport in mouse tissue and human enteroids, with observed results similar to that by Yin et al. (19) Complementary results in Cftr KO mice and CF enteroids provide further confidence in our findings with CFTR inh -172.For DRA, we did not perform experiments in Slc26a3 KO animals or cells, however, we did use 3 different DRA inhibitors, from 2 different structural categories and modes of inhibition.While confidence in pharmacological inhibitors can wane with time and increased use, we believe the consistency of results across these different inhibitors provides confidence in the interpretation of the data.
In summary, we have identified that linaclotide may have more pleiotropic ion transport effects beyond CFTR-and NHE3-mediated chloride and sodium transport, respectively.We have shown, using mice, duodenal biopsies, and human duodenal enteroids, that linaclotide stimulates bicarbonate secretion, and does so at least in part, through DRA recruitment to the apical membrane.Our findings that this remains possible independent of CFTR activity and/or expression has potentially important implications for CF patients, who may benefit from linaclotide's bicarbonate stimulatory effect, in addition to its inhibitory effect on NHE3, even if they can't leverage its pro-chloride secretory properties.(6) Furthermore, our identification of high DRA expression in the human duodenum may spur interest in identifying DRA-activating drugs targeting proximal small intestinal disease.
Animals. Adult (8-16 week-old) C57BL/6J mice from Jackson Laboratories (Bar Harbor, ME) were used as wildtype control mice.CFTR tm1Unc knockout (KO) mice were obtained from Case Western University CF Mouse Models Core Facility and maintained on 50% PEG-3350 with electrolytes solution in their drinking water with standard chow.Similar number of male and female mice were used in experiments.
In Vivo Measurement of Bicarbonate Secretion.In vivo measurement of duodenal bicarbonate secretion was performed using a well-validated technique.(36)Animals were maintained with free access to food and water for up to 1 hour before the experiment, then fasted from chow.
Mice were anesthetized with oxygen-delivered isoflurane (1-3%) at 1 L/minute via a vaporizer (Braintree Scientific, Inc, Braintree, Mass).Mouse temperature was monitored by rectal probe and maintained at 37°C through automated warming using a controlled warming pad (ATCC 2000, World Precision Instruments, Sarasota, FL).Anesthetic plane was assessed by respiratory rate and toe pinch reflex.Respiratory and heart rates were determined every 15 minutes.Animals could be sustained for more than 3 hours under these experimental conditions.
To isolate the proximal duodenum, the abdomen was opened by a central vertical incision, and the proximal 5-10 mm of duodenum, from the pylorus to just proximal to the entry of the common bile duct, was isolated in situ without compromising vascular supply.A small polyethylene tube (PE-50) with a distal flange was advanced to the duodenal bulb via the stomach, and a ligature was secured around the pylorus.A distal intestinal incision was made, and PE-50 flanged tubing was advanced to just proximal to the entry of the pancreaticobiliary duct to prevent entry of the pancreatic or biliary secretions and allow for collection of effluent from isolated segment.The isolated duodenal segment was gently flushed and then continuously perfused (Harvard Infusion Pump, Harvard Apparatus, South Natick, MA, USA) at a rate of 0.21 mL/min with 154 mM NaCl (37°C) ± stimulatory/inhibitor drugs.Effluents from the isolated segment were visually free of bile and blood throughout all experiments.After an initial 15-minute washout period, basal bicarbonate secretion (with luminal saline perfusion) was measured for 30 minutes.Subsequently, linaclotide (10 -9 M, 10 -7 M, 10 -5 M) ± DRA inh -A250 inhibitor (10 -5 M), CFTR inh -172 (2 x 10 -5 M), or S3226 (10 -5 M) was perfused intraluminally for 30 minutes.When multiple doses of drugs were studied, they were perfused sequentially in the same animals.These doses were selected based on previously published data and similarities with clinical dosing.Forskolin (10 -4 M) ± CFTR inh -172 (2 x 10 -5 M) was perfused intraluminally for 45 minutes.After each experiment, the length of the duodenal test segment was measured in situ to the nearest 0.5 mm.Sample volumes were measured by weight to the nearest 0.01 mg.The amount of bicarbonate in the effluents was quantitated by a validated micro back-titration method.(36) Briefly, 100 μL of 5 x 10 -2 M HCl was added to 2 mL sample with 2 mL of double-distilled water.
Samples were then gassed with N 2 , prewashed in Ba(OH) 2 to remove all CO 2 , and back-titrated with 2.5 x 10 -2 M NaOH to an endpoint of pH 7.0 using an automated pH meter/titration unit (TIM 856 Radiometer, Copenhagen, Denmark).Bicarbonate secretion was determined in 15-minute periods and expressed as micromoles per centimeter per hour and presented as bicarbonate output over time or net peak bicarbonate output (peak output minus average basal).
In Vitro Measurement of Ion Transport and Bicarbonate Secretion.For murine measurements, mice were anesthetized as above, and the proximal 1 cm of duodenum was removed and placed in ice-cold 3 x 10 -1 M mannitol solution with indomethacin (10 -5 M) to prevent endogenous prostaglandin formation.The mesentery was dissected off and the duodenum was opened along the mesenteric edge.The muscular layer was scraped from the mucosa using a glass slide while being bathed in the indomethacin-containing solution.Stripped duodenal mucosae were mounted between two 0.1 cm 2 aperture sliders placed in temperature controlled Ussing chambers maintained at 37°C (P2300, Physiologic Instruments, San Diego, CA).For human endoscopic biopsy measurements, biopsies were collected in the same iso-osmolar mannitol solution with indomethacin and then mounted directly between two 0.031 cm 2  CaCl 2 1.2, Glucose 10, gassed with 95% O 2 /5% CO 2 .Ion transport was measured by shortcircuit current (I sc ), where tissues were voltage clamped at 0 mV using a voltage-clamp apparatus (VCC600, Physiologic Instruments, San Diego, CA).To monitor transepithelial resistance, the voltage clamp was released, and the open circuit voltage was recorded every 10 minutes.Resistance was calculated using Ohm's Law.To measure bicarbonate secretion, we utilized the automatic pH-stat method, as previously described.(7,37) In brief, a pH microelectrode was placed in the apical bath and set at the pH of the unbuffered apical solution (typically ~6.8).The titrator (Titrando 902, MetrOhm, Switzerland) was set to titrate 5 x 10 -3 M HCl in 0.2 μL aliquots as the pH of the apical bath rose above the pH set point.Tiamo software (MetrOhm, Switzerland) was used to control the rate of titration and continuously measure the amount of HCl titrated into the apical bath and the apical bath pH.The set-up as described allows for simultaneous measurement of bicarbonate secretion and I sc without electrical interference.Bicarbonate secretory rates (μmol/cm 2 •h) were calculated in 5-minute intervals by noting the amount titrated, the concentration of titrant, and the surface area of the slider aperture.In a subset of experiments, forskolin (10 -4 M, serosal) was added at the end of the experiment to assess tissue viability.
Single cell sequencing analysis.The R toolkit Seurat was employed to conduct an analysis and quality control of single-cell RNA sequencing (scRNA-seq) data from existing literature.(38) We clustered the cells according to the Seurat default settings.In brief, we embedded the cells through a K-nearest neighbor graph, reduced the data with principal component analysis, and applied the Louvain algorithm for clustering.We visualized these clusters with UMAP (uniform manifold approximation and projection).To annotate major cell types and compare gene expression in crypts versus villi, we used the metadata annotations provided by Elmentaite et al. (21) and Busslinger et al.,(20) respectively.We utilized violin and feature plots to display gene expression profiles, and FeatureScatter to compare the co-expression of two genes by cell type.
Human duodenal organoids.Duodenal organoids were established and cultured as previously described.(39) In short, crypt cells were isolated from duodenal endoscopic biopsies using a series of PBS washes followed by incubation with cold chelation buffer (distilled water with 5.6 mM Na 2 HPO 4 , 8.0 mM KH 2 PO 4 , 96.2 mM NaCl, 1.6 mM KCl, 43.4 mM sucrose, 54.9 mM dsorbitol, 1 mM dl-dithiothreitol) with 2 μM EDTA for 30 mins on ice.After incubation, EDTA was removed, and the biopsies were vigorously resuspended in cold chelation buffer (without EDTA).The resuspension was allowed to sit on ice for 1 minute to allow biopsies to settle by gravity.The supernatant was collected.This process was repeated 3-5 times until no more crypts were present in the supernatant.The crypt containing supernatants were pooled and centrifuged at 600g for to examine DRA mRNA expression in human duodenum.The Elmentaite et al. dataset(21) contained 5,944 cells with an average of 4,164 reads per cell from 5 healthy adult duodenal samples, allowing for high confidence in celltype separation.The Busslinger et al. dataset Busslinger et al. dataset possessed insufficient numbers of SLC9A3-expressing cells to analyze SLC9A3 and CFTR co-expression in crypts and villi.
, n=8-11, P=0.006).Linaclotide stimulation (10 -7 M, 40 minutes) increased DRA brush border expression in F508del homozygous CF enteroids, compared to CF enteroids treated with vehicle control (water, 40 minutes)(Figures 6J-L and N, n=8-10, P=0.010).To examine the discrepancies between CFTR inh -172 + linaclotide in non-CF enteroids and linaclotide in the CF enteroids, we examined the absolute mean fluorescent intensities (MFI) of each condition.We found that CFTR inh -172 alone caused similar increases in absolute DRA brush border expression as linaclotide in non-CF enteroids (with or without CFTR inh -172) or CF enteroids (Figure 6O, n=10-31, P=0.676), suggesting that lack of additional DRA membrane expression in non-CF enteroids treated with CFTR inh -172 + linaclotide may be due to maximal DRA brush border expression already being achieved upon CFTR inh -172 treatment.Taken together with our bicarbonate secretory measurements, these data show that linaclotide can alter DRA activity and/or brush border expression, in the presence or absence of functional CFTR.In contrast, linaclotide appears to modulate NHE3 activity, rather than brush border expression, regardless of CFTR function.
using specific DRA inhibitors and DRA protein trafficking studies in apical-out human enteroids, indicate that DRA-mediated chloride/bicarbonate exchange is critical to linaclotide stimulated duodenal bicarbonate secretion upon loss of CFTR activity.This helps explain the lack of linaclotide-stimulated short-circuit current in Cftr KO mice, human biopsies with CFTR inh-172, and prior studies by McHugh et al. in F508del mice.(6)Interestingly, we found that CFTR inhibition alone caused an increase in DRA membrane trafficking.There was also increased apical membrane DRA expression in F508del homozygous duodenal enteroids even without stimulation.In mouse intestinal crypts, Strubberg et al. identified that pH i was raised in cftr KO mice or wildtype mice treated with CFTR inh -172.(29)In fact, ~20 years ago, Kaunitz et al. proposed the "CF Paradox" whereby diminished bicarbonate transport may increase intracellular buffering in CF and prevent duodenal ulcer formation despite diminished bicarbonate secretion.( aperture sliders.Tissue was bathed in apical solutions containing (in mM): NaCl 115, KGluconate 5.2, NaGluconate 25, MgCl 2 1.2, CaCl 2 1.2, Mannitol 10, gassed with 100% O 2 and basolateral solutions containing (in mM): NaCl 115, K 2 HPO 4 2.4, KH 2 PO 4 0.4, NaHCO 3 25, MgCl 2 1.2,

Figure 2 .
Figure 2. Linaclotide stimulates duodenal bicarbonate secretion independent of CFTR. A. In vivo measurement of duodenal bicarbonate secretion in mice, similar to Figure 1A, with the exception that each perfusate also contained CFTR inh -172 (2 x 10 -5 M) (n=12).Dotted line is the mean response without CFTR inh -172 (from Figure 1A).*, P<0.05 vs. baseline by ANOVA.B and C. In vitro duodenal mucosal bicarbonate secretion (B) and I sc (C) in wildtype and Cftr KO mice.Data are expressed as fold increase in linaclotide (10 -7 M) stimulated responses over baseline bicarbonate secretion or I sc in the same mouse.Each point (n=7) is a separate piece of duodenum from 5 mice.**, P<0.01 by unpaired Student's t-test.D. Transepithelial resistance measurements in wildtype (WT) and Cftr KO mice.E-G.In vitro duodenal mucosal bicarbonate secretion (E), I sc (F), and transepithelial resistance (G) in human endoscopic biopsies with or without CFTR inh -172 pre-treatment (2 x 10 -5 M, 40-60 minutes).Each point (11-14) represents a different biopsy.All data are means ± SEM.

Figure 4 .
Figure 4. Cellular and membrane expression of SLC26A3 (DRA) in human duodenal enterocytes.A-H.Cellular mRNA expression of SLC26A3 (DRA), SLC26A6 (PAT-1), CFTR, and SLC9A3 (NHE3) based on re-analysis of Elmentaite et al.(21) (A-D) and Busslinger et al.(20) (E-H).Violin plots represent expression relative to all cells of that type, with each point representing the expression of individual cells within each type.I-K.Representative confocal immunofluorescence imaging of DRA (I), villin (marker of apical brush border, J), and nucleus (K) in human duodenum from endoscopic biopsy (n=3).Scale bar = 20 μm.

Figure 5 .
Figure 5. SLC26A3 and CFTR co-expression in human duodenum by single cell RNA sequencing.A-C.Coexpression of SLC26A3 (DRA) and CFTR mRNA using FeatureScatter based on Elmentaite et al.(21) enterocytes (A) and Busslinger et al.(20) crypt and villi (B and C) datasets that were analyzed in Figure 4. Numbers on top of graphs represent the percentage of SLC26A3-expressing cells that express SLC26A3 only (left) or SLC26A3 and CFTR (right).

Figure 6 .
Figure 6.Linaclotide increases membrane expression of DRA in apical-out human duodenal enteroids.A. Representative confocal microscopy images of DRA (A), villin (B), and DRA and villin and Hoescht (C) during control conditions (water, 40 minutes, top) or linaclotide (10 -7 M, 40 minutes, bottom).D-F.Quantification of DRA present at the apical brush border using villin to define this cell region.Apical brush border DRA mean fluorescence intensity following linaclotide (10 -7 M, 40 minutes, n=16 ) treatment was normalized to vehicle controls for comparison.CFTR inh -172 (2 x 10 -5 M, 40 minutes) and/or linaclotide (10 -7 M, 40 minutes) treatment was normalized to vehicle controls for comparison.Vehicle control for linaclotide was water and for CFTR inh -172 was DMSO.Columns with whiskers are mean ± SEM with each dot representing a different enteroid.Enteroids from three different patients were used for each condition.Significance determined by unpaired Student's t-test.For enteroids treated with both CFTR inh -172 and linaclotide, enteroids were pretreated with CFTR inh -172 for 40 minutes prior to linaclotide treatment.G and H. DRA membrane expression in non-CF and CF (F508del homozygous) enteroids during vehicle control conditions (G and H, water) or during linaclotide stimulation (H, 10 -7 M, 40 minutes).N=8-11 enteroids from 1 patient across 2 different passages.Data were normalized to non-CF enteroids (G) or vehicle control in CF enteroids (H).Significance determined by unpaired Student's t-test.I. To compare DRA membrane expression across non-CF and CF enteroids in different situations, absolute DRA membrane mean fluorescence intensity was plotted.Data are from experiments performed in D-H.Significance determined by ANOVA.Arrows provided to assist with identifying regions of interest.All data are means ± SEM, unless stated otherwise.

Figure 7 .
Figure 7. Inhibition of CFTR activity induces increases in intracellular pH.A and B. Mean timecourse of CFTR inh -172 (2 x 10 -5 M) induced changes in pH i in the absence (A) or presence (B) of DRA inh -A250 (10 -5 M) pretreatment in undifferentiated human duodenal enteroids.C. Quantitative comparison of the peak change (peakbaseline) or recovery (plateau -baseline) in pH i units for enteroids exposed to CFTR inh -172 alone or DRA inh -A250 then CFTR inh -172.Data are expressed as means ± SEM.D-F.Similar experiments and analyses were performed in 3-day differentiated enteroids as A-C.Each circle represents a different cell within an enteroid.Measurements were taken from multiple enteroids across multiple coverslips.***, P<0.001 by paired (within control or DRA inh -A250 groups) or unpaired (across control or DRA inh -A250 groups) Student's t-test.+++, P<0.001 compared to undifferentiated enteroids by unpaired Student's t-test.
by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is made by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.It is made