Weill Cornell autologous gastric insulin-secreting organoids (GINS)

Overview

Developed by Dr. Qiao ("Joe") Zhou and colleagues at the Hartman Institute for Therapeutic Organ Regeneration at Weill Cornell Medicine (PMID: 37106062, PMID: 41202817), Gastric Insulin-Secreting (GINS) organoids offer a fundamentally different cellular starting point for beta-cell replacement than embryonic stem cells (ESCs, used in Vertex's VX-880 / zimislecel) or induced pluripotent stem cells (iPSCs, used in RGB-5088 and SC451).

Because the embryonic stomach and pancreas arise from adjacent domains of the developing foregut endoderm, adult human gastric stem cells (hGSCs) — which naturally regenerate the stomach lining every few days including gut hormone-secreting enteroendocrine cells — share a closely related epigenetic landscape with pancreatic islet progenitors (PMID: 26908146).

How GINS Organoid Reprogramming Works

1. Routine Endoscopic Biopsy & Exponential Stem-Cell Expansion: A standard 15-minute outpatient upper gastrointestinal endoscopy collects a few millimeters of stomach mucosa (antrum or corpus). From a single biopsy, adult human gastric stem cells (hGSCs) can be expanded exponentially in culture into billions of cells while maintaining genomic stability and never passing through a pluripotent (teratoma-capable) state.
2. Sequential 3-Factor (NPM) Induction in 10 Days (PMID: 37106062):
   • Step 1 (NGN3 transient pulse, 2 days): Transient activation of Neurogenin-3 (NGN3 / NEUROG3) commits proliferating gastric stem cells to an endocrine fate via a SOX4-high / Galanin-high precursor state.
   • Step 2 (PDX1 + MAFA maturation, ~8 days): Expression of the pancreatic beta-cell transcription factors PDX1 and MAFA drives spontaneous 3D aggregation into islet-like GINS organoids containing approximately 70% insulin-producing beta-like cells alongside smaller populations of glucagon-producing (α-like), somatostatin-producing (δ-like), and ghrelin-producing (ε-like) endocrine cells.
   • Within 10 days of induction (compared with 30–45 days for 6- to 7-stage pluripotent stem-cell protocols), GINS organoids exhibit glucose-stimulated insulin secretion (GSIS) and respond to GLP-1 receptor agonists.

Preclinical Evidence (2023–2026)

• Reversal of Diabetes in Mice (Nature Cell Biology 2023, PMID: 37106062): When human GINS organoids were transplanted under the kidney capsule of streptozotocin (STZ)-diabetic immunocompromised (NSG) mice, the grafts rapidly vascularized, secreted human C-peptide in response to glucose challenges, and restored normoglycemia (70–180 mg/dL [3.9–10.0 mmol/L]) within days, maintaining stable blood glucose control for over 100 days until surgical graft removal (nephrectomy) caused immediate relapse to severe hyperglycemia. No teratomas or uncontrolled proliferation occurred.
• In Vivo Induction in Human Stomach Organoids (Stem Cell Reports Dec 2025, PMID: 41202817): Lu, Kim, Zhu, and Zhou demonstrated that human gastric stem cells already engrafted in vivo can also be induced to generate functional insulin-secreting cells in situ.
• Hypoimmunogenic GINS Organoids (bioRxiv July 2026, PMID: 42523425): Because unmodified autologous beta-like cells still express HLA class I molecules and remain vulnerable to recurrent Type 1 Diabetes autoimmune T-cell attack, Dattoli, Brown, Zhou, and colleagues reported the generation of hypoimmunogenic human GINS organoids engineered to evade both autoimmune and allogeneic immune destruction.

Limitations & Watch Items

1. Preclinical Stage (Zero Human Recipients): As of October 2026, GINS organoids have not yet entered a human clinical trial or reported large-animal (non-human primate) transplantation outcomes.
2. Autoimmune Protection & Non-Integrating cGMP Manufacturing: Moving from academic proof-of-concept into a Phase 1 human trial requires locking down a cGMP-grade, non-integrating factor-delivery workflow and deciding whether the first clinical candidate will be autologous (potentially paired with encapsulation or targeted immune modulation) or an off-the-shelf hypoimmune-edited GINS line (PMID: 42523425).

Summary of Strengths & Limitations

Strengths

• Starts from abundant adult human gastric stem cells obtained via a routine, minimally invasive outpatient endoscopic stomach biopsy — avoiding pluripotent stem-cell reprogramming (iPSCs/ESCs) and eliminating residual teratoma risk
• Rapid 10-day differentiation protocol (transient NGN3 activation followed by PDX1 and MAFA) yields 3D islet-like organoids (~70% beta-like GINS cells alongside alpha-, delta-, and epsilon-like cells) that restored normoglycemia for >100 days in diabetic mice (Huang et al., Nature Cell Biology 2023, PMID 37106062)
• Autologous derivation eliminates allogeneic HLA rejection, and a July 2026 preclinical study (Dattoli et al., PMID 42523425) demonstrated generation of hypoimmunogenic GINS organoids to address autoimmune attack and off-the-shelf scaling

Limitations

• Zero human clinical trials or large-animal (non-human primate) studies have been published as of October 2026 — all in vivo efficacy data are in immunocompromised diabetic mice
• Unmodified autologous beta-like cells still express patient HLA class I molecules presenting beta-cell autoantigens and remain vulnerable to recurrent Type 1 Diabetes autoimmune T-cell destruction unless paired with hypoimmune gene editing or encapsulation
• Clinical translation requires replacing research-grade integrating lentiviral/inducible transcription-factor cassettes with cGMP-compliant non-integrating delivery