Stem Cells and Regeneration in the Digestive System: Keystone Meeting

The gastrointestinal (GI), pancreatic, and hepatic systems work in concert to break down and absorb nutrients, maintain homeostatic glucose levels, and purify the entire bodily organism. Unfortunately, these systems are plagued by many debilitating and life-threatening diseases. Stem cell-based tissue regeneration holds promise for treating diseases of the digestive system, thereby improving patient health and quality of life. To this end, it is of great importance to understand the developmental mechanisms of organ generation and regeneration, what characterizes stem cells in vivo, and how the microenvironments of the pancreas, liver, and/or intestines regulate self-renewal, proliferation, and differentiation of adult stem cells.

The Keystone Symposium X6 “Stem Cells and Regeneration in the Digestive Organs,” held jointly with the Keystone Symposium X5 “Islet Biology: From Cell Birth to Death” (March 13–17, 2016), was the first to bring together a broad representation of scientists (>700) who study stem cells from the interrelated fields of the GI, pancreatic, and hepatobiliary organs. The first point of focus was on the progress made in developing robust, in vitro stem cell isolation and culture protocols. Second, attention was paid to the role of different subsets of stem cells and their exploitation for treating disease, including cancer. Third, recent advances were presented regarding our understanding of digestive system development and the developmental origin of certain GI diseases.

Session Summaries

Deriving GI Lineages from Pluripotent Stem Cells

Hiromitsu Nakauchi, in his opening Keynote address, gave a perspective on the application of interspecific blastocyst complementation for investigation of therapy for diabetes and other diseases with organ defects. A current challenge is deriving human patient-specific islets of Langerhans in pigs, as the evolutionary distance forms a xenobarrier that defies interspecies blastocyst complementation to form human patient-derived tissue. However, the successful generation of a rat pancreas in mouse or vice versa is proof of principle that perhaps one day we will be able to generate functional, human induced pluripotent stem cells (iPSCs)-derived organs in animal models, which can then be used for disease models, therapeutic models, and ultimately for transplantation therapies for many types of GI diseases.¹

The first breakout session addressed a key question: “How can GI lineages be derived from pluripotent stem cells?” Gordon M. Keller reviewed methodologic approaches and outcomes in patterning human iPSC-derived endoderm. Endoderm subpopulations have different hepatic and pancreatic potential, and isolation of these subpopulations requires quantitative assays. Paul Gadue described an in vitro human model to study the most common form of pancreas agenesis: heterozygous mutation of GATA6. The use of this human stem cell-derived GI model is highly relevant, as phenotypes of mouse models do not necessarily replicate human disease. James M. Wells addressed how manipulation of fibroblast growth factor, Wnt, and bone morphogenetic protein signaling patterns iPSC-derived endodermal tissues in human organoids, establishing them as useful models for studying etiology of intestinal diseases. Notably, the Wells group has added complexity to these organoid models, as they incorporated a functional enteric nervous system into human intestinal organoids, generating intestinal tissue capable of peristaltic-like motility. Jason R. Spence also addressed the potency of organoid models, stressing that these systems are multicellular, similar to in vivo models, but have the flexibility of in vitro systems.

Jan Jensen publicly shared for the first time results of a recently developed computerized experimental approach that provides a rapid, high-throughput, systems biology analysis of cell differentiation. Trailhead Biosystems is a new descendant company from these research activities at Case Western Reserve University and the Cleveland Clinic.

Disease Modeling

Disease modeling is a hot topic in liver research. To date, orthotopic liver transplantation is the only plausible therapeutic approach for patients with rescue end-stage liver failure. Unfortunately, the number of available donor livers is limited, and cell transplantation therapy offers an attractive alternative. One potential source of cell based therapy are hepatocyte-like cells derived from iPSCs (iHeps). Patient-derived iPSCs are also relevant for disease modeling of various monogenic disorders. Stephen Duncan’s laboratory is establishing high-throughput screening assays with familial hypercholesterolemia iHeps to identify existing drugs that could be repurposed to correct the pathophysiology of the disease.² Holger Willenbring’s work addresses another aspect of liver disease modeling: in vivo hepatic reprogramming as a therapeutic strategy for liver fibrosis. With adeno-associated virus carrying hepatic transcription factors, myofibroblasts could be reprogrammed into functional iHeps in the mouse liver.³ Surprisingly, only a small portion of myofibroblasts were reprogrammed, perhaps owing to low efficiency of delivery of the transcription factors with individual adeno-associated virus vectors. However, the results suggest that myofibroblasts might be good targets for reprogramming. Shinichiro Ogawa’s group has exploited Notch signaling to transform iPSC-derived hepatoblasts into cholangiocytes.⁴ They could induce cystic fibrosis (CF) patient iPSCs to differentiate into cholangiocytes that do not express CF transmembrane conductance regulator (CFTR). Pharmacologic treatment with VX-809, Corr-4a, and VX770 to treat patients with CF iPSC-derived cysts could rescue the CFTR deficit in vitro, thus demonstrating that this model could be applied to finding novel CF therapeutics.

An example of disease modeling in the intestine is the work of Frederic J. De Sauvage, whose group develops therapeutic approaches that target intestinal stem cells in cancer. They showed that treating colorectal cancer with anti–R-Spondin 3 antibodies leads to loss of a stem cell signature, induces differentiation, and reduces subsequent tumor propagation. William Proctor discussed the role of PI3Kalpha and mammalian target of rapamycin inhibition in endoderm differentiation as a first step toward personalized toxicology, and Martín G. Martín provided an outlook on the application of autologous intestinal stem cell therapies in various intestinal syndromes.

Identity and Function of GI Stem and Progenitor Cells

Tissue regeneration can depend on different mechanisms: stem cell-driven cell proliferation and differentiation, or dedifferentiation of already differentiated cells into stem-cell–like cells, or so-called cell plasticity. However, through which mechanism regeneration is primarily achieved is a hot research topic. In addition, a controversial issue is to what extent cancer stem cells (CSCs) contribute to tumorigenesis and drug resistance.

Differentiation of hepatocytes depends on gene regulatory networks. Core liver-enriched transcription factors are organized as self-enforcing cross and autoregulatory loops. The increased expression of the core factors and self-enforcing stability of the network during hepatocyte differentiation is poorly understood. New insights were presented by Frédéric Lemaigre, whose work illustrated that hepatic differentiation depends on microRNA-dependent regulatory motifs. Hepatocyte fate could be inhibited by mir337-3p, concomitantly inducing duct cells. Moreover, the hepatocyte gene HNF6/Oc1 is targeted by mir122.⁵ Stuart Forbes’ work on the role of hepatic progenitor cells in liver regeneration showed that endogenous beta1 integrin inhibited hepatocyte proliferation, arguably in a stiffness-modulating manner. Thus, Forbes’ laboratory provides the first model for nonhepatocyte factors in liver regeneration.

Timothy C. Wang drew an insightful parallel from the intestinal stem cells and their niches to gastric stem cells and their niches. As in the intestine, the stomach contains rapid and slow cycling progenitor cells; however, the identity and the regulation of gastric corpus stem cells remain elusive. Wang’s laboratory showed that Mist1 expression marks a rare, distinct population of quiescent stem cells in the gastric corpus isthmus, the major site of epithelial proliferation.⁶ Their ongoing work shows that differentiation depends on Notch signaling and Cxcl12⁺-endothelial cells and Cxcr4⁺-gastric innate lymphoid niche cells.

It is still controversial whether cancers within the GI tract are maintained by tumor-initiating cells or CSCs, which can be particularly resistant to conventional cancer treatments. Targeting CSCs may eliminate or reduce the incidence of chemo resistance and relapse. Linheng Li’s laboratory showed how stem cell niche signaling, propagated by chemotherapy-induced inflammatory cells, such as macrophages, contributes to chemoresistance by influencing Bmi⁺ CSC behavior. In his talk, Li updated research conducted by Dr Xi He to develop a combined chemotherapy and targeted therapy treatment (using a cyclo-oxygenase-2 inhibitor that suppresses prostaglandin E2-induced PI3K-Akt and Wnt-β-catenin signaling) to inhibit the oncogenic-driven expansion of CSCs.

Programmed or Reprogrammed Regeneration

Because the autoimmune response destroys insulin-producing pancreatic beta islet cells, treatment of type 1 diabetes relies on replacing these cells. Although patients benefit from pancreatic tissue transplantation, this material is very scarce. Thus, a promising alternate treatment strategy is to derive a new pool of insulin-expressing beta islet-like cells from human iPSCs. Along this line, Louise Laurent’s laboratory is working on reprogramming a patients’ own skin cells into beta islet-like cells. Regulatory factors have been identified by transcriptome analysis of human iPSC cells undergoing directed differentiation to beta islet-like cells. In addition, the Qiao Zhou laboratory has successfully reprogrammed endocrine cells of the stomach into functional insulin producing cells via expression of 3 key regulatory genes: Neurog3, Pdx1, and Mafa.⁷ Antral endocrine cells were shown to reprogram with greater efficiency than those in the intestine, and owing to their regenerative nature, could provide a renewable source of beta cells upon injury. For proof of concept, human gastric organoids were generated with the ability to express these same reprogramming factors. The future hope is to transplant reprogrammed iPSC-derived gastric tissue into humans that would give rise to renewable insulin+ cells outside of the pancreas.

The Lijian Hui laboratory has succeeded in regenerating pig and even human liver cells with iHep cells. However, when doing iHep reprogramming, they observed extensive cell death, senescence, and proliferation arrest in wild-type fibroblasts. A model was proposed wherein the p19^Arf/p53 pathway guards cell identity preservation to maintain the physiologic function of a given tissue. Indeed, it is interesting that the p19^Arf/p53 pathway is found to suppress both tumorigenesis and cell fate conversion; Hui’s data suggest that this model relies on an epigenetic barrier that inhibits lineage conversion, BAF60, ATM, and p53 dependently, but DNA damage independently. The idea of epigenetic barrier limit lineage conversion is supported by a recent report.⁸

Atsushi Suzuki discussed stem cell systems in the developmental liver, and is the first to demonstrate a role for Lin28b and microRNAs. His work showed that Lin28b inhibited microRNA let7b and miR-125a/b, prompting hepatoblast proliferation and lineage differentiation. Inhibition of Lin28b induced cholangiocyte but not hepatocyte differentiation.⁹

Markus Grompe addressed whether liver regeneration occurs via existing cell plasticity or stem cells. Although it is generally believed that the biliary tree harbors liver stem cells, Grompe showed that biliary cells are heterogeneous: only certain cells are genetically clonal and form Lgr5⁺ organoid cells in vitro. Upon injury, hepatocytes but not biliary cells can act as facultative stem cells that differentiate into both hepatocytes and duct cells. Yet, the latter does not resemble a “true” duct morphology, transcriptome profile, or organoid-forming capacity. Instead, this population shows great plasticity and might revert to hepatocytes during recovery.¹⁰

Adult Pancreatic Stem Cells

Understanding the plasticity of adult organs has been of interest to many laboratories for the purpose of reprogramming and regeneration of damaged tissue, and was the focus point of this joint session. Matthias Hebrok’s laboratory has identified hypoxia as a critical stressor for cellular plasticity in the adult pancreas. His laboratory introduced differentiation conditions to generate functional insulin-producing cells from human embryonic stem cells. This system can be used to test human beta cell plasticity and de-differentiation, which could provide information regarding human regenerative capabilities of islets. Lori Sussel discussed the role of Blinc1, a noncoding RNA adjacent to Pax6, and how it regulates lineage allocation in the pancreas.¹¹ Maike Sander discussed the role of metabolic products, such as NAD, in regulating transcription factor activity in beta cells as determinant for signal responsiveness in developmental and adult endocrine cells. These presentations provided additional information regarding mechanisms regulating beta cell identity and how those ideas may be used for future programming or reprogramming of tissue. Olivia G. Kelly presented results of the phase I clinical trial for a stem cell-derived therapy based on macroencapsulated islet progenitor implants for type 1 diabetes. After implantation, encapsulated human embryonic progenitor cells de-differentiate into glucose-responsive, insulin-secreting cells that are protected from immune attack.

Stem and Progenitor Cells in Homeostasis and Diseases

Alternatively, endogenous beta cell mass expansion could facilitate regeneration in patients with diabetes. Maureen Gannon’s laboratory has reported that the secreted protein CTGF promotes adult beta cell replication and mass regeneration after injury.¹² In this session, Gannon provided new insights into beta-cell compensation mechanisms and specifically the role of CTGF together with inflammation signals in promoting expansion of functional pancreatic beta-cell mass. Tsutomu Chiba shared data on doublecortin-like kinase 1 as a marker for pancreatic intraepithelial neoplasia-initiating cells in acinar-to-ductal metaplasia.

For the intestine, Calvin Kuo discussed the interplay of R-spondins and Wnt signaling and their role in balancing self-renewal with differentiation and expansion in LGR5⁺ stem cells. His group showed that Wnt signaling makes cells ‘competent’ to respond to R-spondin signaling, and that the presence of R-spondin induces purely symmetric division of crypt-based columnar cells, whereas the inhibition of R-spondin results in differentiation. Thus, their work highlights the critical role of R-spondins in the process of self-renewal and crypt maintenance.

For the liver, Ben Stanger addressed cellular plasticity. Similar to the Grompe laboratory findings, the Stanger laboratory found that, in an injury model, hepatocytes rather than biliary cells functioned as “stem cells” that de-differentiated into liver ductal cells, while duct cells could not revert into hepatocytes.¹³ Bruce Wang challenged the hypothesis of liver stem cell homeostasis by stating that pericentral hepatocytes expressing Axin, a negative regulator in the Wnt pathway, regenerate liver in wild-type mice. Different from other hepatocytes, these cells are highly proliferative and diploid, giving rise to polyploidy daughter cells. Paracrine Wnt signals from the adjacent endothelium regulate the Axin2⁺ hepatocytes involved in liver regeneration.

Second Keynote Address: Ongoing Clinical Applications of Lgr5 Stem Cell Organoids in CF and Drug Response Screening in Cancer Treatment

In the second Keynote address, Hans Clevers discussed the interconversion of progenitor and stem cells in the stomach and intestine. He showed that, after ablation of LGR5⁺ crypt-based columnar cells, transit-amplifying cells can give rise to new crypt-based columnar cells; similarly, his group found that completely differentiated chief cells in the stomach are capable of creating progeny that result in monoclonal gastric pits. He described how human disease development is being studied using intestinal organoids from CF patients.^14,15 By monitoring the swelling of these organoids in response to various compounds, researchers are able to hasten drug screening and development, and thus bring personalized approaches for this disease closer to fruition. He also demonstrated that the 3-dimensional in vitro assays could be used to screen colonic tumors for rapid sensitivity to various chemotherapeutic agents, and that this approach may someday be used to develop personalized therapeutic approaches for patients with colon cancer. Last, he showed that, through controlling key pathways, liver organoids can be maintained in culture to study disease pathogenesis, toxicologic screening, and liver regeneration.

Conclusion

The goal of the Keystone meeting on Stem Cells and Regeneration in the Digestive System was to promote crosstalk and exchange ideas among 3 distinct but interrelated fields that represent intimately connected biological systems. We conclude that this meeting was very successful in joining expertise and promoting collaborations, which will encourage mutual instruction in addressing key questions common to each of these systems and fields. Although breakthroughs have been made, research efforts need to continue focusing on complete understanding of (1) stemness under homeostasis and stressed conditions, (2) failure of stem cells in vitro to extend to in vivo conditions, (3) stem cell subpopulations, (4) reversion of progenitor cells to stem cells, and (5) effects of the microenvironment on stem cells and progenitor cells.