Using Human Induced Pluripotent Stem Cell Derived Organoids to Identify New Pathologies in Patients with PDX1 Mutations

Abstract

Background & Aims:

Two patients with homozygous mutations in PDX1 presented with pancreatic agenesis, chronic diarrhea and poor weight gain, the causes of which were not identified through routine clinical testing. We aim to perform a deep analysis of the stomach and intestine using organoids derived from induced pluripotent stem cells from PDX1^{188delC/188delC} patients.

Methods:

Gastric fundic, antral and duodenal organoids were generated using iPSC lines from a PDX1^{188delC/188delC} patient and an isogenic iPSC line where the PDX1 point mutation was corrected.

Results:

Patient-derived PDX1^{188delC/188delC} antral organoids exhibited an intestinal phenotype, while intestinal organoids underwent gastric metaplasia with significant reduction in enteroendocrine cells. This prompted a re-examination of gastric and intestinal biopsies from both PDX1^{188delC/188delC} patients, which recapitulated the organoid phenotypes. Moreover, antral biopsies also demonstrated increased parietal cells and lacked G-cells suggesting loss of antral identity. All organoid pathologies were reversed upon CRISPR-mediated correction of the mutation.

Conclusion:

These patients will now be monitored for the progression of metaplasia and gastrointestinal complications that might be related to the reduced gastric and intestinal endocrine cells. This study demonstrates the utility of organoids in diagnosing uncovered pathologies.

Graphical Abstract

Lay Summary

We used pluripotent stem cell-derived stomach and intestinal organoids to identify new pathologies and improve care for patients with PDX1 mutations.

Introduction

Pancreatic and duodenal homeobox gene-1 (PDX1) is a parahox transcription factor that is expressed in the pancreas and the gastrointestinal (GI) tract in mice and humans (1–6) and plays a central role in orchestrating pancreatic morphogenesis and is essential for beta cell function. Consistent with this, mice and humans lacking PDX1 develop pancreatic agenesis and early onset diabetes (7,8). PDX1 is also expressed in the gastric antrum and anterior regions of the duodenum and additional phenotypes that have been identified in Pdx1 null mice include a contorted pylorus, ectopic biliary duct epithelium in the duodenum and a reduction in some enteroendocrine cells (EECs) (5,6,9). However, none of these pathologies have been identified in humans, possibly due to limitations of available diagnostic tests. The disruptions in boundary formation of the GI epithelium certainly pose a potential risk for the development of GI cancer in this patient population. Moreover, abnormalities in EEC development and function can negatively affect hunger, satiety, digestion and absorption. Thus, it is essential to identify the exact effects of PDX1 mutations in humans to provide optimal care for these patients.

Over the past decade, we have developed several human organoid models that have allowed for unprecedented modeling of human GI development and function (11–15). Organoids are three-dimensional human tissues that function in a manner similar to native organs and thus allow for unparalleled mechanistic studies of patient pathophysiology in the laboratory. Organoid model systems bioengineered from patient-derived induced pluripotent stem cells (iPSCs) with CRISPR/Cas9 technology for isogenic correction of mutations affords us a unique opportunity to diagnose new pathologies, mechanistically interrogate human diseases, and individualize patient therapies.

We have identified two patients with a homozygous frameshift mutation in PDX1, causing complete loss of functional protein. In addition to having diabetes and exocrine pancreatic insufficiency, these patients had additional GI symptoms including chronic diarrhea and poor weight gain despite being treated with pancreatic enzyme replacement therapy since birth. Initial endoscopies and biopsies failed to identify GI pathologies. We took a novel diagnostic approach and used patient derived iPSCs to generate gastric and duodenal organoids to diagnose how mutations in PDX1 cause a myriad of GI pathologies. This in vitro modeling led us to re-examine the patients’ clinical samples and confirm the in vivo pathophysiology thus impacting their clinical care. For the first time, we demonstrate that lack of PDX1 in humans results in intestinal and gastric metaplasia and increased inflammation in the antrum and intestine respectively. We also show that loss of PDX1 decreases key EECs including somatostatin, PYY, Ghrelin, GIP and CCK and an overall loss of antral identity. Interestingly, the lack of PDX1 in fundic organoids and biopsies failed to reveal metaplasia but did demonstrate an increase in parietal cells. This study represents a paradigm shift in how we use human organoid diagnostics to identify and model complex pathologies in patients and alter patient care.

Materials and Methods

Patient Phenotypes

Patient A: A male infant with a confirmed homozygous mutation in PDX1 (c.188delC (p.P63fs)) was identified via clinical testing. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the Institutional Review Board of Cincinnati Children’s Hospital Medical Center. Written informed consent was obtained from the subject’s legal guardian. He was born small for gestational age with severe intra-uterine growth restriction (birth weight 1125 grams (−3.8SD), birth length 36.2 cm (−4.48SD), head circumference 36.2 cm (−3.13 SD)). At birth, he was noted to be hyperglycemic with blood glucose measurements >600 mg/dL. He was diagnosed with neonatal diabetes and initiated on insulin therapy. Additional GI work up demonstrated partial gallbladder agenesis and pancreatic agenesis causing exocrine pancreatic insufficiency. Pancreatic enzyme replacement therapy was started immediately, and a G tube was placed to ensure adequate caloric intake in the setting of feeding difficulties. At two years of age, the G tube was removed. Esophagogastroduedonsocpy (EGD) was performed at 24 and 40 months of age with routine biopsies obtained from the antrum and duodenum. The EGD at 24 months to close his gastrocutaneous fistula found mucosal nodularity in the antrum and pre-pyloric region. The duodenal mucosa was normal. Luminal narrowing was encountered at the D1/D2 transition. At 40 months of age, a repeat EGD found normal appearing mucosa in the stomach and duodenum without duodenal narrowing. He is currently managed on an insulin pump, pancreatic enzymes and proton pump inhibitor therapy.

Patient B: Similarly, another male infant with a confirmed homozygous mutation in PDX1 (c.188delC(p.P63fs)) was born at 37 weeks with severe intra-uterine growth restriction (birth weight 1560 g (−3.35 SD)) with pancreatic agenesis and gallbladder agenesis. This clinical presentation has been previously described (9). He is managed on insulin injections, pancreatic enzymes, proton pump inhibitor, cyproheptadine, and lactobacillus rhamnosus.

Human Pluripotent Stem Cell culture and Directed Differentiation into Pancreatic Cells, Antral, Fundic and Duodenal Organoids

iPSC lines were generated from blood samples taken from Patient A as previously described (10). iPSC clones were maintained on hESC-qualified Matrigel (Corning) in mTeSR1 media (STEMCELL Technologies) and confirmed to be pluripotent, karyotypic normal and mycoplasma negative.

Human fundic, antral organoids and duodenal organoids, and pancreatic endocrine cells were generated from patient lines as previously reported (10–14). Human antral and duodenal organoids were also generated from HUES8 control and HUES8 PDX1 homozygous iCRISPR lines, which had previously been shown to result in a reduction in pancreatic endocrine cells, a phenotype similar to that observed in our patient (10). For intestinal organoid generation, the AggreWell 400 (StemCell Technologies) system was used to generate spheroids. Plates were rinsed with anti-adherence rinsing solution. Cells were treated with Accutase (StemCell Technologies), rinsed in gut media and plated at a density of 3.6 million cells per well. The following day, spheroids were re suspended in Matrigel and were subject to intestinal protocol as described (11,12). Intestinal organoids were transplanted under the kidney capsule of immune deficient NOD.Cg-Prkdc^scidIl2rg^tm1Wjl/SzJ (NSG) mice and allowed to mature for 10 weeks as previously described (12–15). Fundic, antral, intestinal organoids and pancreatic endocrine cells were generated from individual passages a minimum of three times.

CRISPR Correction

CRISPR-Cas9 was used to correct a mutation in the PDX1^{188delC/188delC} iPSC line. The guide RNA (CTCGTACGGGGAGATGTCCG) targeting the mutation site was designed according to the web tool (http://CRISPOR.org) (16). The complementary DNA oligos were cloned as described (17). A phosphorothioated ssDNA donor oligo (GCCCTGGGCGCGCTGGAGCAGGGCAGCCCtCCGGAtATCTCCCCGTACGAGGTGCCtCCaCTCGCCGACGACCCCGCGGTGGCG) was designed to include the nucleotide insertion (underlined) according to published methods (18). iPSCs were reverse transfected with plasmid and donor oligo using TranIT-LT1 (Mirus). GFP-positive cells were isolated by FACS and replated in Matrigel (Corning). Single clones were manually excised for genotyping, expansion, and cryopreservation. Correctly targeted clones were identified by PCR, enzyme digestion and Sanger sequencing.

Generation of Doxycycline Inducible PDX1 Stem Cell Line

For inducible overexpression, we obtained human PDX1 cDNA in pENTR223.1 (from Harvard PlasmID repository, HsCD00082591), which was then cloned into the lentiviral destination vector pInducer20 (a gift from Stephen Elledge, Addgene #44010) using Gateway cloning methods with LR Clonase II (Invitrogen). Successful cloning was confirmed using Sanger sequencing (CCHMC sequencing core). To generate lentiviral particles, the pInducer20-PDX1 plasmid was co-transfected (Lipofectamine 2000, Invitrogen) into 293T cells with packaging plasmids psPAX2 and pMD2.G. Supernatant containing lentivirus was harvested after 48 and 72 hours post-transfection, pooled, and centrifuged at 1000 x g to remove cellular debris. hPSCs were transduced by addition of lentiviral supernatant to the culture medium immediately after passaging, which was then replaced with fresh medium following 6 hours exposure. The cells were selected with G418 (100 ug/ml) for 4 days, beginning on the following day. At this point, the colonies had normal morphology and growth characteristics, while mock-transduced cells were completely dead. The cell line was then maintained under standard culture conditions, with intermittent exposure to G418 to maintain resistant cells. Inducible PDX1 expression was validated by IF staining following 48 hour exposure to doxycycline (0.5 ug/ml) in the undifferentiated state.

qPCR

RNA was extracted using the Nucleospin RNA extraction kit (Macharey-Nagel) and reverse transcribed into cDNA as previously described (19). Primer sequences are listed in Supplemental Table 2. qPCR was performed as previously described (19). Relative expression was determined using the ΔΔCt method and normalizing to PPIA (cyclophilin A). Organoid samples from at least three independent passages were used for quantification.

Immunofluorescence Staining

Immunofluorescence staining on paraffin embedded tissues was performed as previously described (13). Primary and secondary antibodies are listed in Supplemental Table 1. Images were acquired using a Nikon A1 GaAsP LUNV inverted confocal microscope and were analyzed using NIS Elements (Nikon). Six to eight organoids per passage from at least three independent passages were used for quantification. Data were quantified as number of positive cells divided by the entire epithelium labelled by either CLDN18, CDH17, Beta-Catenin or CDH1.

Statistics

Statistical analyses were performed using GraphPad prism version 8. Quantification of data are represented as the mean +/− SEM. Statistical significance was determined using unpaired student’s t-test or one-way ANOVA with Brown-Forsythe test. *p<0.05, **p<0.01, ***p<0.001 was considered significant.

Results

Loss of PDX1 results in gastritis, intestinal and gastric metaplasia in the antrum and duodenum

Given that PDX1^{188delC/188delC} patients had unexplained chronic diarrhea and weight loss despite being on optimized pancreatic enzyme replacement therapy since birth, we developed an organoid-based strategy to perform deep diagnostic analysis of stomach and intestinal markers (Sup Figure 1). As a control we confirmed that iPSC lines from a PDX1^{188delC/188delC} patient was unable to generate pancreatic tissue, consistent with pancreatic agenesis pathology observed in the patient (Sup Figure 2). We generated antral and duodenal organoids using iPSC lines from the PDX1^{188delC/188delC} patient A and an isogenic iPSC line where we corrected the PDX1 point mutation using CRISPR/Cas9, thus restoring expression of PDX1 protein. While no macroscopic differences were observed between control and PDX1^{188delC/188delC} antral and duodenal organoids (data not shown), the PDX1^{188delC/188delC} antral and duodenal organoids demonstrated metaplasia, with stomach tissue appearing in the duodenal organoids and vice versa (Figure 1 and Sup Figure 1, 3A-F). Intestinal markers upregulated in the antral organoids of the PDX1^{188delC/188delC} patient included Cadherin 17 (CDH17), Villin1 (VIL1), Caudal Type Homeobox 2 (CDX2), and Mucin 2 (MUC2) and decreased gastric markers Claudin 18 (CLDN18) and Mucin 5AC (MUC5AC) (Figure 1A,C and Sup Figure 3A-F). Conversely, gastric markers present in in vitro and transplanted duodenal organoids of PDX1^{188delC/188delC} patients included CLDN18 and MUC5AC (Figure 1A,C and Sup Figure 4A,B). Moreover, regions of metaplasia in antral and duodenal organoids demonstrated a complete conversion from one epithelial type to the other (Figure 1A,C and Sup Figure 3A,C, Sup Figure 4A). These phenotypes were rescued in PDX1 CRISPR corrected organoids indicating that the PDX1 mutation is unambiguously causing the metaplastic phenotypes. Moreover, these above findings were also observed in antral and duodenal organoids generated from a human embryonic stem cell line (HUES8) where PDX1 was disrupted using CRISPR technology (Sup Figure 4C,D). Together these findings suggest that loss of PDX1 results in gastric and intestinal metaplasia in humans.

Figure 1. Loss of PDX1 results in intestinal or gastric metaplasia in the stomach and intestine, respectively.

(A-B) Immunofluorescence and (C-D) quantification of (A) control, PDX1^{188delC/188delC}, and PDX1^{188delC/188delC} CRISPR Corrected antral and post-transplantation intestinal organoids and (B) antral and duodenal biopsies from control, PDX1^{188delC/188delC} patient A and PDX1^{188delC/188delC} patient B biopsies stained with gastric epithelial marker CLDN18 and intestinal epithelial markers CDH17 and VIL1. (E) Histological staining of antral and duodenal biopsies with regions of metaplasia encircled and inflammatory foci marked with asterisk. Sydney classification of PDX1^{188delC/188delC} patient A and B antral and intestinal biopsies compared with controls diagnosed with intestinal and gastric metaplasia respectively. All sections counterstained with nuclear DAPI. Scale bars represent 100 μm.

Based on the above organoid findings, antral and duodenal biopsy samples from both PDX1^{188delC/188delC} patients (Patient A & B) were re-examined to look for evidence of metaplasia. Not only did biopsies from both patients demonstrate gastric and intestinal metaplasia, (Figure 1B,D and Sup Figure 3 A-F), biopsies from both PDX1^{188delC/188delC} patients revealed overt inflammation in the antrum (Figure 1E asterix) especially in regions close to intestinal metaplasia. Additionally, the metaplastic areas in the antrum had increased KI67 proliferation when compared to control patients (Sup Figure 3G) suggesting these areas are mitotically active.

Using the Sydney classification, antral and duodenal biopsies from PDX1^{188delC/188delC} patients as well as control patients with and without a diagnosis of intestinal or gastric metaplasia were scored by a pathologist (Figure 1E). The antral biopsies of both PDX1^{188delC/188delC} patients showed mild intestinal metaplasia which was consistent with the scoring given to three out of four controls that had the previous diagnosis of intestinal metaplasia. Similarly, the duodenal biopsy from PDX1^{188delC/188delC} patient A showed severe gastric metaplasia, while PDX1^{188delC/188delC} patient B showed mild gastric metaplasia (Figure 1E) and this scoring was also consistent with that given to three controls with the diagnosis of gastric metaplasia. Finally, both PDX1^{188delC/188delC} patients have had two separate antral biopsies obtained 2–3 years apart that have demonstrated intestinal metaplasia indicating that the metaplasia is stable and more of a concern for the future development of cancer. Given that metaplasia and associated inflammation are risk factors in the development of cancer, the clinical team caring for these patients will now provide additional screening to monitor for progression of inflammation, intestinal and gastric metaplasia.

Loss of PDX1 results in loss of antral identity and expression of fundic markers

PDX1 is normally expressed in the antrum and not in the fundus of the stomach. In the PDX1^{188delC/188delC} antral organoids and patient biopsies, loss of PDX1 expression is linked to loss of antral markers such as gastrin (GAST). Instead, antral organoids and biopsies are more fundic in nature, with an increase in parietal cells (ATP4B), Ghrelin (GHRL) positive cells, and Chief cells expressing fundic specific pepsinogens (PGA3) (Figure 2A-C). The aforementioned markers were expressed at levels similar to control fundus, indicating that loss of PDX1 results in either a failure of the antrum to form or a possible antral to fundic conversion. To determine if antral organoids lacking functional PDX1 were competent to make parietal cells, we stimulated parietal cell differentiation by the transient inhibition of the MEK pathway and activation of BMP signaling, previously shown to stimulate differentiation of functional parietal cells in fundic organoids (13). PDX1^{188delC/188delC} antral organoids were competent to form ATP4B expressing cells while CRISPR corrected antral organoids were not (Figure 2D), suggesting that PDX1 expression is essential for repressing parietal cell fate and maintaining antral identity. Moreover, we observed ATP4B+ parietal cells in the duodenal organoids and biopsies of one of the PDX1^{188delC/188delC} patients (Figure 2F,G) suggesting the gastric metaplastic tissue was fundic in nature. The loss of a functional antrum may have clinical consequences in these patients including delayed gastric emptying and deranged parietal cell function due to loss of gastrin-producing cells. The increase in parietal cells in the antrum and duodenum may also explain the chronic diarrhea and lack of weight gain noted in these patients. All organoid results were discussed in real time with the team of physicians managing these challenging patients. Patients were continued on proton pump inhibitors and are being monitored for the development of new metaplasia. Given the importance of the antrum in gastric emptying and the regulation of GAST on parietal cell function, these patients will be monitored for additional complications in moving forward.

Figure 2. Loss of PDX1 results in loss of antral identity in the stomach.

(A) Immunofluorescence and (B) quantification of control antral, PDX1^{188delC/188delC} A and PDX1^{188delC/188delC} B biopsies stained with parietal cell marker ATP4B, transcription factor PDX1, pepsinogens PGA3 and PGC and hormones GAST and GHRL. (C) Quantitative pCR of control antral, PDX1^{188delC/188delC} antrum and control fundic organoids for markers GAST, GHRL, ATP4B and PGA5. (D) Immunofluorescence and (E) quantification of PDX1^{188delC/188delC} and PDX1^{188delC/188delC} CRISPR corrected fundic organoids, antral organoids, antral organoids with PD03, BMP4, CHIRON with ATP4B. (F) Immunofluorescence and (G) quantification of control, PDX1^{188delC/188delC}, PDX1^{188delC/188delC} CRISPR corrected post-transplantation intestinal organoids and control, PDX1^{188delC/188delC} A and PDX1^{188delC/188delC} B patient duodenal biopsies with gastric marker ATP4B and epithelial marker CDH1. All sections counterstained with nuclear DAPI. Scale bars represent 100 μm.

PDX1 expression is inversely correlated with parietal cell formation

PDX1 expression is limited to the antrum in both mice and humans. Therefore we predict that the intestinal metaplasia caused by loss of PDX1 would be restricted to the antrum. To investigate this we generated fundic organoids from PDX1^{188delC/188delC} patients and surveyed for intestinal metaplasia. We observed no expression of intestinal markers VIL1 and CDH17 (Figure 3A-B) in PDX1^{188delC/188delC} fundic organoids and biopsies. However, we did observe an increase in the number of parietal cells in fundic PDX1^{188delC/188delC} organoids (Figure 2D) and biopsies (Figure 3E-F). This was surprising given that PDX1 is not expressed at high levels in the fundus. This data suggests that even low levels of PDX1 seem to antagonize parietal cell differentiation. To directly test if PDX1 could repress parietal cell differentiation in the fundus, we ectopically expressed PDX1 in fundic organoids using a tetracycline-inducible PDX1 construct. Expression of PDX1 in control fundic organoids between days 20–34 of differentiation resulted in a significant reduction in parietal cells as marked by ATP4B protein and mRNA (Figure 3 F-G). These findings suggest that PDX1 may have an unappreciated role in inhibiting the differentiation of parietal cells.

Figure. 3. PDX1 controls parietal cell number in the fundus.

(A-B) Immunofluorescence and (C-D) quantification of control PDX1^{188delC/188delC}, and PDX1^{188delC/188delC} CRISPR Corrected fundic organoids and control, PDX1^{188delC/188delC} A and PDX1^{188delC/188delC} B biopsies stained with gastric epithelial marker CLDN18 and intestinal epithelial markers CDH17 and VIL1. (E) Immunofluorescence and (F) quantification of control, PDX1^{188delC/188delC} A and PDX1^{188delC/188delC} B fundic biopsies with ATP4B. (G) Immunofluorescence (H) quantification and qPCR of fundic organoids without and with doxycycline treatment on d20–34 with transcription factor PDX1 and parietal cell marker ATP4B.

PDX1 is required for normal enteroendocrine cell development in the duodenum

EECs are nutrient sensing cells and in the GI tract are essential for nutrient absorption, motility, glucose homeostasis and satiety. Given the role of PDX1 in pancreatic endocrine cell development we surveyed EECs in the duodenum of PDX1^{188delC/188delC} organoids and biopsies. We noted a significant reduction in EECs as marked by chromogranin A (CHGA), as well as loss or reduced numbers of cells expressing the hormones somatostatin (SST), peptide YY (PYY), GHRL, gastric inhibitory peptide (GIP) and cholecystokinin (CCK) (Figure 4A-B) in both organoids and biopsies. We did not observe any changes in serotonin (5-HT) or GAST in duodenal organoids or biopsies from PDX1^{188delC/188delC} patients (Figure 4D-G) demonstrating that PDX1 is required for the differentiation of some, but not all EECs subtypes. This reduction could have direct clinical consequences on nutrient homeostasis including digestion and macronutrient absorption and may be contributing to the poor weight gain and persistent diarrhea observed in PDX1^{188delC/188delC} patients.

Figure. 4. Loss of PDX1 decreases enteroendocrine cells in the intestine.

(A) Immunofluorescence and (B) quantification of control, PDX1^{188delC/188delC}, and PDX1^{188delC/188delC} CRISPR corrected post-transplantation intestinal organoids and control, PDX1^{188delC/188delC} A and PDX1^{188delC/188delC} B duodenal patient biopsies stained for secretory protein CHGA and enteroendocrine hormones SST, PYY, GHRL, GIP, CCK, 5-HT and GAST. All sections counterstained with nuclear DAPI. Scale bars represent 100 μm.

Discussion

In this study, we generated patient derived GI organoids to uncover new pathologies in PDX1^{188delC/188delC} patients. We show that lack of PDX1 in humans results in an overall loss of antral identity, gastric and intestinal metaplasia and increased inflammation in the GI tract. We also show that PDX1 is essential for specific EECs including somatostatin, PYY, Ghrelin, GIP and CCK in humans and is important for regulating parietal cell number. Due to our discoveries using an organoid-based diagnostic strategy, these patients will now be monitored for the development of new areas of metaplasia and inflammation, that together increase the risk of development of gastric cancer. This study represents a paradigm shift in how we use human organoids to identify and model complex pathologies in patients and alter patient care.

Genetic ablation of Pdx1 in mice results in loss of G cells in the antrum (3,4), which we also observed in PDX1^{188delC/188delC} gastric organoids and patient biopsies. However, PDX1^{188delC/188delC} patients still express gastrin in the intestine at similar levels as control, suggesting that the role of PDX1 for endocrine cell formation in the stomach is different from the intestine. Mice that lack gastrin have chronic gastritis, intestinal metaplasia of the stomach and eventually develop adenocarcinoma (20). PDX1^{188delC/188delC} patients also exhibit similar findings, suggesting that the absence of gastrin in the antrum may contribute to these pathologies. Given that gastrin secretion follows a meal and controls acid secretion, loss of gastrin and an increase in parietal cells in both the antrum and fundus might cause deranged gastric pH and improper digestion in the stomach and duodenum. Moreover, we note an increase in parietal cell number in both fundic organoids and patient biopsies from PDX1^{188delC/188delC} patients while overexpression of PDX1 caused a reduction in parietal cells in fundic organoids. This is consistent with findings previously shown by our group and others that PDX1 may regulate parietal cell populations in mice and humans (13,21).

The PDX1^{188delC/188delC} patients have difficulty with weight gain and report diarrhea despite optimizing pancreatic enzyme and proton pump inhibitor therapy. We noted the presence of parietal cells in the intestine in one of the PDX1^{188delC/188delC} patients which may explain the symptoms that these patients are experiencing and the gastritis noted in their biopsies. Moreover, we used the Sydney classification and compared gastric and intestinal biopsies from PDX1^{188delC/188delC} patients with those who had a diagnosis of gastric and intestinal metaplasia. The histological scoring of biopsies was consistent between controls and PDX1^{188delC/188delC} patients. In patients who eventually develop gastric cancer, the extent of metaplasia has been shown to be an important risk factor (22). Thus, from our studies, PDX1^{188delC/188delC} patients will be regularly monitored for further progression of metaplasia and inflammation as well as functional changes in the GI tract.

PDX1 is expressed only in the antrum of the stomach and not in the fundus (3,4). Interestingly, the PDX1^{188delC/188delC} patients demonstrate fundic marker expression in the antrum including increased Ghrelin, parietal and fundic pepsinogen cell populations, suggesting that PDX1 is essential for antral formation. Moreover, patients with Menetrier’s disease, have ectopic expression of PDX1 in the fundus, resulting in antralization and eventual expression of gastrin in fundic regions (21). Taken together with the findings of our study, these data suggest that PDX1 is required for antral formation during development. Interestingly, both PDX1^{188delC/188delC} patients and those with Menetrier’s disease have diarrhea and difficulty with weight gain. This indicates that developmental disruptions in stomach formation may have significant downstream consequences including improper digestion and gastric emptying.

In summary, we have established an organoid diagnostic approach to uncover a myriad of GI pathologies in patients with PDX1 mutations. These include inflammation, gastric and intestinal metaplasia, loss of antral identity, excess number of parietal cells and perturbations in EEC formation. This study represents the utility of human organoids in the identification and modelling of complex pathologies which in turn can alter patient diagnoses and clinical care.

What You Need To Know.

Background and Context:

Two patients with homozygous mutations in PDX1 presented with pancreatic agenesis, chronic diarrhea and poor weight gain, the causes of which were not identified through routine clinical testing. We developed an organoid-based diagnostic strategy to perform deep analysis of the stomach and intestine from induced pluripotent stem cells derived from PDX1^{188delC/188delC} patients.

New Findings:

Organoids and patient biopsies demonstrated gastric and intestinal metaplasia, increased parietal cells and a loss of G cells in the stomach and a reduction in key enteroendocrine hormones in the duodenum.

Limitations:

This study includes patients with rare homozygous PDX1 mutations, only of which three patients have been identified in the world.

Impact:

This study represents how human organoid diagnostics can be used to identify and model complex pathologies in patients and alter patient care.