Hedgehog signaling drives glial cell plasticity and oncogenic reprogramming in gastroenteropancreatic neuroendocrine neoplasms

Suzann Duan; AnneLeigh B Twer; Ateh Zinkeng; Ikrame Naciri; Juliette W Moore; Rebeca G Gentry; Ricky A Sontz; Reed I Ayabe; Edgar Tapia; Travis W Sawyer; Juanita L Merchant

doi:10.1186/s12943-026-02611-y

. 2026 Feb 21;25:85. doi: 10.1186/s12943-026-02611-y

Hedgehog signaling drives glial cell plasticity and oncogenic reprogramming in gastroenteropancreatic neuroendocrine neoplasms

Suzann Duan ^1,², AnneLeigh B Twer ², Ateh Zinkeng ², Ikrame Naciri ¹, Juliette W Moore ¹, Rebeca G Gentry ¹, Ricky A Sontz ², Reed I Ayabe ³, Edgar Tapia ⁴, Travis W Sawyer ⁵, Juanita L Merchant ^2,^6,^✉

PMCID: PMC13032639 PMID: 41723447

Abstract

Background

Disruption of the Men1 locus in epithelial and endocrine tissues fails to generate the full spectrum of gastroenteropancreatic neuroendocrine tumors (GEP-NETs), raising the possibility of a potential stromal source for these cancers. Neural crest-derived glial cells were previously implicated in neuroendocrine tumors arising in the pituitary and pancreas, yet these studies lacked a clear mechanism for these events. Here, we investigated the hypothesis that Men1-driven Hedgehog (HH) signaling redirects the glial cell fate to give rise to neuroendocrine tumors in the gastrointestinal tract.

Methods

Hyperactivation of the HH signaling pathway in human GEP-NETs was evaluated using immunofluorescent staining and clinicogenomic databases. Men1 was deleted in the glial lineage by expressing Cre recombinase downstream of the human GFAP and Sox10 promoters. Overexpression of HH signaling proteins in mouse GEP-NETs was confirmed by immunofluorescent staining and immunoblot analysis. We generated human and mouse GEP-NET tumoroids and exposed them to agonists and inhibitors of HH signaling. HH activation of Men1-deficient glial cells was blocked by deleting the gene encoding primary ciliary protein KIF3A required for transducing SHH signaling.

Results

We demonstrated that human GEP-NETs overexpress HH signaling pathway components, including SHH and its cognate receptor PTCH1. We showed that patient-derived GEP-NET tumoroids proliferate in response to SHH pathway agonists. In contrast, pharmacologic inhibition of GLI1/2, but not inhibition of SMO alone, attenuated tumoroid growth. Genetic deletion of Men1 in GFAP⁺ and SOX10⁺ glial cells caused the development of pancreatic and intestinal NETs that overexpress HH proteins. Further use of tdTomato⁺ mice demonstrated the involvement of GFAP⁺ and SOX10⁺ glial cells in these tumors. Tumoroid cultures of mouse pancreatic, duodenal, and jejunal NETs recapitulated the drug response shown by patient-derived tumoroids. Lastly, Men1-deficient enteric glial cultures showed a glial-to-neuroendocrine transition that was alleviated upon HH inhibition, and these events were reproduced in genetic mice harboring GFAP⁺ cells with impaired primary cilia.

Conclusions

Our study implicates the HH signaling pathway in GEP-NET development and underscores a glial cell of origin for these tumors.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12943-026-02611-y.

Keywords: Gastroenteropancreatic neuroendocrine tumors, Hedgehog signaling pathway, Enteric neural crest, Menin, Tumor organoids

Background

Gastroenteropancreatic neuroendocrine tumors (GEP-NETs) comprise the most common NET subtype and are rapidly increasing in prevalence, indicating a strong need for developing new approaches to detect and manage these cancers in a growing patient population [1, 2]. GEP-NETs encompass remarkably diverse neoplasms that vary in clinical presentation, mutational status, and response to therapy. Recent efforts to clarify their etiology revealed intra- and inter-tumor cell heterogeneity that was suggested to arise from reprogrammed resident endocrine cell populations [3–5]. A deeper understanding of these reprogramming events and how they contribute to the etiology of these cancers is precluded by a dearth of preclinical models from multi-organ NETs that include pancreas and small intestine tissue sources. Modeling extra-pancreatic gastrointestinal NETs poses a distinct challenge, likely due to the relative low abundance of driver mutations in these neoplasms and reports of tumors containing a mixture of cell populations [5–10]. De novo mutations in the Multiple Endocrine Neoplasia I (MEN1) gene encoding the tumor suppressor protein Menin account for up to 40% of GEP-NETs, with the highest frequency observed in duodenopancreatic NETs (dpNETs) and gastrin-expressing tumors (gastrinomas) that occur in the submucosal Brunner’s glands [11–13]. Whereas, disruption of the Men1 locus in endocrine tissues in mice stimulated well differentiated tumors in the pancreas and pituitary, neither duodenal tumors nor gastrinomas have been reported in these models [14–19]. Consistent with these findings, epithelial-directed Men1 deletion using the Villin Cre transgene failed to produce duodenal NETs (DNETs), yet these mice developed hypergastrinemia, enterochromaffin-like cell hyperplasia, and gastric NETs. Intriguingly, these events followed the non-cell autonomous loss of nuclear Menin in gastrin-expressing enteric glial cells [20]. This evidence raised the possibility that enteric neuroendocrine cells in DNETs and gastrinomas might differentiate from a neuroectoderm lineage rather than solely from endoderm-restricted progenitors [21]. The premise that dpNETs can arise from neural cell progenitors challenges the current dogma and underscores the significance of the current study.

We and others have reported that neural crest derived glial cells, and their progenitors can acquire an endocrine phenotype leading to functional hormone secretion [22–24]. In support of a stromal source for NETs, glial-restricted deletion of Men1 in mice using GFAP and Sox10 Cre transgenes (GFAP-Cre; Men1^FL/FL or GFAP^ΔMen1 and Sox10-Cre; Men1^FL/FL or Sox10^ΔMen1) stimulated pituitary and pancreatic NETs that coincided with loss of the glial-restricted lineage, suggesting that glial cells might undergo reprogramming upon Men1 inactivation [24]. Menin is a ubiquitously expressed nuclear scaffold protein that complexes with multiple transcription factors and epigenetic modifiers to regulate the expression of genes involved in cell growth and endocrine cell-specification [25]. Among these, Menin epigenetically antagonizes components of the Sonic hedgehog (SHH) signaling pathway known to direct neural cell fate patterning and the neuroendocrine phenotype [26, 27]. Consistent with this knowledge, impairment of the SHH signaling pathway in GFAP⁺ glial cells was found to reverse Men1-driven neuroendocrine cell hyperplasia in the GFAP^ΔMen1 mice [24]. The underlying role of SHH activation in these tumors and the potential for pharmacologic intervention in preclinical models of GEP-NETs have not been thoroughly evaluated.

In the current study, we investigated the premise that dpNETs might originate from enteric glial cells that are instructed by SHH signaling. Using a combination of human and mouse derived tumoroids, primary enteric glial cultures, and transgenic mouse models, we demonstrate that pancreatic and duodenal NETs may develop from glial cells suggesting a neural crest origin for these gut tumors. We found that Menin inactivation in enteric glial cells promotes their transition from a glial-restricted lineage to the NET phenotype. Loss of nuclear Menin derepressed a SHH-GLI signaling axis and induced neural progenitor and neuroendocrine cell transcriptional programs. Pharmacologic and genetic inhibition of the GLI1/2 transcriptional effectors dampened Men1-induced glial cell plasticity and blocked GEP-NET development. This study underscores the potential value of targeting non-canonical SHH signaling in cancers with neuroectoderm origins.

Methods

Animals

Mice were housed in individually ventilated cages with food and water ad lib. For serum hormone analysis, the mice were fasted for 16 h with water ad lib prior to blood collection. Transgenic Cre and CreER^T2 expressing mice were bred onto a Men1^FL/FL and Kif3a^FL/FL background to conditionally delete Men1 in GFAP- or SOX10-expressing cells, as reported on previously [24]. All strains were analyzed at 18 to 22 months of age. All histological characterization and downstream comparisons were made using littermate controls that genotyped negative for Cre recombinase expression and included mice with and without the relevant floxed alleles. Further details on the mouse strains are described in the Supplementary Materials.

The cancer genome atlas (TCGA) patient cohort analysis

TCGA data were accessed through the National Cancer Institute’s Genome Data Browser. Patient cohorts were designed using the GDC Cohort Builder and filtered based on the primary cancer diagnosis. Copy number variations were analyzed by filtering for the molecular biomarkers indicated in each table or heatmap. Tables and heatmaps were generated in the GDC data portal and edited for clarity using the Inkscape open access graphic and design editor.

Caris Clinico-genomic database analysis

Retrospective review was performed on molecular profiling data from real-world patient tumors submitted from both comprehensive and community cancer centers throughout the U.S. to a College of American Pathologists (CAP)/Clinical Laboratory Improvement Amendments (CLIA)-certified laboratory (Caris Life Sciences; Phoenix, AZ, USA). Neuroendocrine cancers with primary origin in the pancreas, duodenum, or jejunum and ileum (N = 7,071) were selected for cross tissue analysis. Next-generation sequencing (NGS) of DNA was performed using either a custom-designed SureSelect biotinylated baits panel (Agilent Technologies, Santa Clara, CA, USA) to enrich 592 whole-gene targets on a NextSeq 500 platform (Illumina, Inc., San Diego, CA, USA) (RRID: SCR_014983), or whole exome sequencing (WES) with the SureSelect Human All Exon v7 panel (Agilent Technologies, Santa Clara, CA, USA) and a hybrid pull-down panel of baits on a NovaSeq 6000 System (Illumina, Inc., San Diego, CA, USA) (RRID: SCR_016387). For whole transcriptome sequencing (WTS), a SureSelect Human All Exon v7 panel of biotinylated RNA baits (Agilent Technologies) was hybridized to the synthesized and purified cDNA targets, and the bait-target complexes were amplified in a post-capture PCR reaction, with NGS performed on a NovaSeq 6000 System (Illumina, San Diego, CA, USA) (RRID: SCR_016387). Genetic variants identified were interpreted by board-certified molecular geneticists and categorized as ‘pathogenic,’ ‘likely pathogenic,’ ‘variant of unknown significance,’ ‘likely benign,’ or ‘benign,’ according to the American College of Medical Genetics and Genomics (ACMG) standards. All MEN1 variants were considered for reported mutation frequencies. Clinical outcomes were inferred from insurance claims data, with overall survival (OS) calculated from initial diagnosis to date of last contact or death. Patients without contact/claims data for a period of at least 100 days were presumed deceased, while those with a documented clinical activity within 100 days prior to the latest data update were censored. Kaplan-Meier survival curves were generated for comparison of patient cohorts with tumors representing the top 75th and bottom 25th percentiles of SMO, GLI2, or PTCH1 transcript expression in subpopulations with primary origin of pancreas and duodenum.

MEN1 Mutation Analysis

Genomic DNA was extracted from a single five-micron FFPE tumor section using the QIAamp DNA FFPE Advanced Kit (QIAGEN, Hilden, Germany, Cat #56604). Genomic DNA from snap frozen human tumors PanNET2, PanNET4, and IL-NET-met1, and tumoroids PanNET3 and PanNET5 was extracted using the QIAamp Micro DNA Kit (QIAGEN, Cat #56304). The quality and yield of the FFPE-extracted DNA was insufficient to technically proceed with downstream Sanger sequencing, therefore only the frozen tumors and tumoroid DNA was sequenced. The initial PCR reactions were performed using the ThermoFisher Platinum II Taq Hot-Start DNA Polymerase, including the kit’s GC Enhancer and DMSO controls. PCR products were validated on a 1.5% agarose gel. Validated PCR products were purified with Thermo Fisher Purelink PCR Purification Kit, quantified on the Nanodrop, and Sanger reacted with BigDye 3.1. Sanger reactions were performed using 15 µL of the PCR product per manufacturer’s protocol, with 0.5 µL DMSO. Complete coverage of the MEN1 protein-coding exons (2 through 10) was achieved using the following fourteen primer sets:

Exon 2 Forward: TTGTGGGGGACAAAAAGG; Exon 2 Reverse: GGTGAGGTTGATGATTTGG; Exon 3–6 Forward: GAAGGGATGGAGGGATAG; Exon 3–6 Rev: AGGAAGGACAGTAAGCAG; Exon 3–6 Internal Forward: CTTGGGAGAGTAGAATTG; Exon 3–6 Internal Reverse: CAATTCTACTCTCCCAAG; Exon 7–9 Forward: GGAGTGGAGATGGAGAGGA; Exon 7–9 Reverse: AGAGCAAGGTGAGAGCAA; Exon 7–9 Internal Forward: ACCCCTTCAGACCCTACA; Exon 7–9 Internal Reverse: AAGGCACAGGGTAGAAAC; Exon 10 Forward Alt: CACCTTTCTTGTGCAGTCC; Exon 10 Reverse Alt: CCCATCCCCAATTTCCCA; Exon 10 Internal Forward: GTCGTTAGAATATAGGTCTC; Exon 10 Internal Reverse: TGGGAGAAGAGACCTATATT. Sanger data was aligned to reference sequence and variations were cataloged on QIAGEN CLC Main Workbench. SNVs were classified as benign or pathogenic based on existing annotation in the ClinVar Miner Database.

Immunohistochemical/Immunofluorescence (IHC/IF) staining

Five-micron formalin-fixed and paraffin-embedded (FFPE) tissue sections were baked at 65℃ for 60 min and then deparaffinized in xylene. Tissues were rehydrated in ethanol and then rinsed with Tris Buffered Saline (TBS). Slides were subjected to heat-mediated antigen retrieval in Tris-EDTA buffer (pH 9.0, Cat #ab93684, Abcam) prior to washing in TBS. Slides were blocked in 10% donkey serum prepared in 1% bovine serum albumin (BSA), 0.2% Triton X-100, and 0.05% Tween-20 (TBST) for 1 h at 24℃. For mouse tissues that were subject to a primary antibody that was raised in a mouse host, sections were also blocked for 1 h using the Mouse on Mouse (M.O.M) staining kit (Vector Laboratories). Sections were then incubated in primary antibodies overnight at 4℃ (See Supplementary Materials for antibody list). Slides were washed in TBST and next incubated in Alexa Fluor-conjugated secondary antibodies diluted 1:400 in TBST with 1% BSA (Thermo Fisher Scientific, Waltham, MA). Slides were incubated for 10 min in 4’, 6-diamidino-2-phenylindole (DAPI) diluted 1:5000 in TBS, then mounted using Fluormount-G (Thermo Fisher Scientific). Menin staining was performed using an additional tyramide signal amplification (TSA) step per manufacturer’s instructions (Cat #B40944, Thermo Fisher Scientific, Waltham, MA). Further details are described in the Supplementary Materials. Slides were imaged using the Olympus BX53F epifluorescence microscope (Center Valley, PA), the ECHO Revolve inverted microscope (ECHO, San Diego CA), or with the Nikon AX R Laser-Scanning Confocal Microscope (Tokyo, Japan). Image acquisition settings were kept consistent when any direct comparisons were made. Post hoc image analyses were performed using open-source Image J (FIJI) software, were applied globally (i.e., to the entire image file), and were restricted to channel level adjustments.

Immunocytochemistry (ICC/IF) staining

Primary tumoroids embedded in Matrigel and seeded in 24-well black glass bottom plates were washed twice with pre-warmed Dulbecco’s PBS (DPBS). The tumoroids were fixed in pre-warmed 4% paraformaldehyde (PFA) for 40 min at 24℃. Tumoroids were permeabilized in 0.5% Triton X-100 for 20 min at 24℃. The tumoroids were washed twice with TBS and then blocked in 10% donkey serum and 1% BSA for 1 h at 24℃. The tumoroids were incubated in respective primary antibodies for 16 h at 4℃. To validate staining specificity, select antibodies were preabsorbed with recombinant protein prior to incubating with the tumoroids or tissue slides. Recombinant protein was added at five-times the weight by volume and the preabsorbed antibody was incubated for 24 h with agitation at 4℃. The tumoroids were washed four times with TBST and incubated in Alexa Fluor-conjugated secondary antibodies, followed by DAPI, as described above. Tumoroids were stored in TBS prior to and during imaging using the ECHO revolve microscope or the Nikon AX R Laser-Scanning Confocal Microscope (Tokyo, Japan). Z-stack images were acquired at 100X and 200X magnification and combined into a maximum intensity projection (MIP) image using the Nikon NIS-Elements software. Pseudocoloring and adjustments to brightness levels were applied globally using ImageJ (FIJI) software.

Primary EGCs were seeded in 24-well glass bottom plates or in 6-well plates containing a glass cover slip. In both scenarios, poly-L-lysine was used to coat the plate or coverslip. Cells were stained according to the steps described above, with the following changes: cells were incubated in 4% PFA for only 15 min and in 0.5% Triton X-100 for 5 min. For crystal violet staining of the BON-1 and STC-1 cells, the cells were fixed in the same manner and washed in PBS. Next, the cells were incubated in 0.1% crystal violet dye for 20 min at 24℃. Cells were washed with distilled water and allowed to dry prior to imaging on the ECHO Revolve inverted microscope.

Cell culture

BON-1 and STC-1 cell lines were obtained from ATCC (Manassas, VA). BON-1 cells were grown in DMEM/Ham’s F12 media supplemented with 10% FBS and penicillin-streptomycin (Corning Inc., Corning, NY). STC-1 cells were grown in DMEM supplemented with 10% FBS and penicillin-streptomycin (Corning Inc., Corning, NY). All cells were maintained at 37 °C and used at passage numbers 5–10. Cells were routinely tested every six months for mycoplasma contamination using the Universal Mycoplasma Testing kit (Cat #30–1012 K, ATCC). Primary enteric glial cultures (EGCs) were generated from 2 to 3 adult mice aged 6- to 12-months using an enzymatic digestion method adapted from previously published protocols [28, 29] (see Supplementary Materials for additional details). For the generation of a pure SOX10⁺ cell population, EGCs from Sox10-CreER^T2-LSL-tdTomato mice were treated with 2 µM 4-OHT for 48 h and then dissociated and sorted by FACS using the FACS Aria II. Cre negative cells were used to establish the gating parameters for sorting positive cells. Sorted cells were plated on poly-l-lysine coated plates and subcultured to generate an EGC line.

Tumor organoid culture

Human tumoroid lines were generated from surgically resected PanNET tissues and an ileal NET metastasis to the liver. The tissues were provided fresh or frozen in cryopreservation media containing 10% DMSO. Mouse tumoroid cultures were generated from fresh and cryopreserved tumor tissues. Tumor tissues were diced into 2–3 mm segments and incubated in a series of EDTA incubations and digested with collagenase type 3 and DNase I. Tumor tissues were digested directly in collagenase if they were smaller than 2 mm. The resulting tumor cell pellet was resuspended in growth factor-reduced ice-cold Matrigel without phenol red (Corning) and seeded into pre-warmed 24-well glass bottom or plastic tissue-culture plates. Tumoroids were subcultured in complete organoid media as published on previously [6, 30]. Organoids were used for studies between 14 and 21 days post-seeding and between passage numbers 0 and 3. Further details of the enzymatic buffers and growth media are described in the Supplementary Methods.

In vitro drug studies

Drug compounds used in the studies include: human recombinant SHH N-terminal peptide (R&D Systems, Cat# 1845-GMP), SAG (Tocris, Cat# 4366), vismodegib (Tocris, Cat# 7710), sonidegib (Tocris, Cat# 7826), GANT61 (Tocris, Cat# 3191), and itraconazole (Tocris, Cat# 5981). Men1 knockdown was accomplished using ON-TARGETplus Mouse Men1 siRNA (Horizon Discovery, Cat# L-042675-01-0005, Layfayette, CO). A SMARTpool of four targeting and non-targeting small interfering RNAs were used. The cells were transfected with 25 nM siRNAs using Lipofectamine 3000 without the use of the P3000 reagent (Invitrogen) according to the manufacturer’s instructions. Cells were analyzed by qPCR after 72 h post Men1 silencing or treatment with non-targeting siRNAs.

EdU fluorescence assay

The proliferation of tumoroid cells was determined used the Click-iT™ EdU Cell Proliferation Kit for Imaging and performed as described by the manufacturer (Thermo Fisher Scientific, Cat# C10337). For the studies in the PanNET3 tumoroid line, tumoroids were dissociated using a 25 G needle syringe and resuspended in Matrigel without phenol red. The cell suspension was seeded into 96-wells or 24-wells and incubated in growth media for 5 to 7 days prior to initiating treatment. Following treatment for 5 days, 1X EdU was added to the wells and allowed to incubate for 4 h. For the studies in the PanNET5 and IL-NET-met1 tumoroid lines, the tumoroids were treated 3 to 5 days post-passaging in 24-well plates. The tumoroids were exposed to the drug treatments for 7 days and the media was replenished after three days. The remaining steps were performed as described in the kit instructions to visualize EdU⁺ cells. The number of EdU⁺ cells were quantified per image and normalized to the total number of cells as enumerated by Hoescht staining.

Bromodeoxyuridine (BrdU) incorporation assay

BrdU incorporation was measured using a BrdU assay kit according to the manual provided (Cell Signaling Technologies, Cat# 6813). BON-1 and STC-1 cells were seeded at a density of 10,000 cells in 96-well plates. The cells were serum starved in 1% FBS containing media for 24 h prior to initiating treatment with the respective compounds. Treatments were initiated for tumoroids 2 to 3 days post seeding and in the absence of serum starvation. 1X BrdU was added to each well after 72 h and allowed to incorporate overnight. Cells were fixed and denatured for 30 min, incubated with anti-BrdU antibody for 1 h at 24 °C, washed, and incubated in HRP-conjugated secondary antibody for 30 min at 24 °C. Samples were washed, incubated in TMB substrate, and the HRP reaction was stopped and visualized by measuring the optical density (OD) at 450 nm using the Gen5 Microplate Reader and Data Analysis Software (BioTec, Dorset, UK).

TdTomato fluorescence growth assay

TdTomato⁺ tumoroids were seeded at a 1:5 dilution into 24-well plates in Matrigel without phenol red. Two to three days after seeding, treatment was initiated by replacing the growth media with fresh media containing the respective drug compounds. Tumoroids were incubated for 72 h and then assessed for growth. Tumoroids were extracted from Matrigel by gently dissolving the mixture in cold PBS. TdTomato fluorescence was quantified using the CLARIOstar^® Plus plate reader (BMG LabTech). Values were averaged across triplicate wells and normalized to the DMSO vehicle control treatment group to account for seeding variations across independent experiments.

Serum hormone Enzyme-linked immunosorbent assays (ELISA)

Serum hormone levels were evaluated exactly as reported previously [24]. Glucagon was measured with the Mouse Glucagon enzyme-linked immunosorbent assay (ELISA) Kit (Crystal Chem, #81518). Serum insulin levels were measured with the Ultra Sensitive Mouse Insulin ELISA Kit (Crystal Chem, #90080). Serum GLP-1 levels were measured using the Mouse GLP-1 ELISA kit (Crystal Chem, #81508). Serum gastrin was evaluated with the Gastrin Enzyme Immunoassay kit (Sigma, #RAB0200). Serum glucose levels were measured using the Mouse Glucose Assay Kit (Crystel Chem, #81692). Prolactin levels were measuring using the Mouse Serum Prolactin ELISA Kit (Invitrogen, #EMPRL). Insulin, glucagon, and gastrin levels were measured in 48 h conditioned media of normal pancreatic and duodenal organoids and tumoroids using the same kits. Hormone levels were normalized to organoid protein lysate concentration.

Quantitative polymerase chain reaction (qPCR)

RNA was extracted from the tumoroids using the ReliaPrep miRNA Cell and Tissue Miniprep System following the manufacturer’s instructions (Promega, Madison, Wisconsin). Between 0.1 and 1 µg of cDNA was synthesized using SuperScript VILO IV after treatment with ezDNAse to remove genomic DNA. Quantitative PCR was performed using PowerUp SYBR Green Master Mix (Invitrogen) on 5 to 10 ng of cDNA using the QuantStudio 3 Real-Time PCR System (Applied Biosystems, Waltham, MA) with the following cycling conditions: 2 min at 50 °C, 2 min at 95 °C, denaturing step for 1 s at 95 °C, extension and annealing for 1 min at 60 °C, followed by a dissociation melt curve stage to confirm primer specificity. All forward and reverse primers were purchased as validated predesigned PrimeTime qPCR Assay primer sets used for SYBR Green dye (Integrated DNA Technologies, Coralville, IA). qPCR data was expressed as fold-change using the established 2^− ddCt method.

Western blot analysis

Tissues were homogenized at 10,000 rpm using a rotor-stator homogenizer in 250 to 500 µL of ice-cold RIPA buffer (Thermo Fisher Scientific) supplemented with 1X HALT protease and phosphatase inhibitor (Thermo Fisher Scientific). Tissue lysates were centrifuged for 15 min at 15,000 x g at 4℃ and the supernatant was collected as the protein extract. Organoids were dissociated from the Matrigel (Corning, Corning, NY) by gently pipetting in cold PBS and centrifuging for 5 min at 300 x g at 4℃. The organoid pellet was washed once in cold PBS, and then resuspended in 200 µL of ice-cold RIPA buffer (Thermo Fisher Scientific) supplemented with 1X HALT protease and phosphatase inhibitor (Thermo Fisher Scientific). Organoids were lysed by passing through a 20 G syringe ten times and vortexing prior to centrifuging for 15 min at 15,000 x g at 4℃.

Protein extracts (10–15 µg) were prepared in reducing conditions with 1X SDS buffer with 5% β-mercaptoethanol and denatured by boiling at 95 °C for 5 min. Proteins were run on precast gradient gels (4–12% Bis-Tris, Invitrogen) in cold 1X MOPS gel electrophoresis buffer for 10 min at 80 V, then 90 min at 110 V. Proteins were transferred onto PVDF membranes using the iBlot 2 (Invitrogen) and then blocked in 5% BSA in TBS with 0.05% Tween-20 (TBST) buffer for 1 h on a rocker at 24℃. Blots were incubated in primary antibodies that were diluted in 5% BSA-TBST at 4 °C overnight on a rocker. Membranes were washed in TBST, then incubated in HRP-linked anti-mouse or anti-rabbit IgG antibody for 1 h at 24℃ with gentle rocking (1:3000 dilution, Cell Signaling Technology). Membranes were washed in TBST and protein bands were visualized using the Pierce ECL detection system (Cat #32106, Thermo Fisher Scientific). For quantitation, films were scanned in gray scale and quantified using ImageJ (FIJI) software. Representative images of blots were enhanced using global brightness adjustments that were applied to the entire image.

Bulk RNA sequencing

The Sox10-tdTomato EGC line was treated with pooled siRNAs as described in the previous section. In addition to the si-Non-targeting (NT) and si-Men1 conditions, cells were also co-treated with si-Men1 and si-Gli1 and si-Gli2, or si-Men1 and GANT61. Five days after treatment, the cells were collected in Zymo DNA/RNA shield reagent and submitted to Plasmidsaurus for bulk RNA sequencing and DEG analysis (Watterson Park, Kentucky). Gene Set Enrichment Analysis (GSEA) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were completed for each of the following pair-wise experimental comparisons: si-NT vs. si-Men1, si-Men1-siGli1/2 vs. si-Men1, and si-Men1-GANT61 vs. si-Men1. In each comparison, DEGs were filtered to those with an adjusted p-value less than 0.05, and with a log fold change greater than zero (overexpressed in the Men1 group). GSEA was then conducted using the Python 3 package gseapy, and KEGG analysis with the Python 3 package keggtools. GSEA was completed using the “Hallmark” gene set. Statistical analysis to determine enriched pathways was done using the Fisher exact test, with the Benjamini-Hochberg false discovery rate (FDR) correction. An adjusted p-value threshold of 0.05 was used to determine significance and select significant pathways and gene sets were displayed as bar charts with the horizontal axis representing gene enrichment ratio (# of differential expressed genes / total genes in a pathway or gene set), with the color corresponding to the –log10 of the adjusted p-value.

Heatmap visualization of gene expression among experimental groups was conducted using Python 3, with the “seaborn” and “pandas” analysis packages. Prior to visualization, a spreadsheet containing gene expression reads for each sample was loaded as a data frame. Genes displayed in the heatmap were selected based on single cell expression in the Human Protein Atlas and assigned to specific gene categories relevant to the current study. The data frame was transformed so that the expression of each gene was calculated as the z-score relative to the entire dataset. This was completed by mean subtraction and division by the standard deviation for each gene. The average expression z-score for each experimental group was then calculated for each gene of interest. Finally, heatmaps were generated by plotting the z-scores for each group using the built-in heatmap function.

Statistical analysis

Statistical analyses were performed using GraphPad Prism 10 software. Two-tailed Student’s t test and one-way or two-way analysis of variance (ANOVA) with pertinent post-tests were applied to normally distributed datasets. Datasets with a non-normal sample distribution were analyzed using the non-parametric Kruskal-Wallis test. Error bars indicate means ± SEM. All P values are defined as *P < 0.05, ** P < 0.01, *** P < 0.001, and **** P < 0.0001.

Results

Human dpNETs overexpress Hedgehog signaling components

The Hedgehog (HH) gene family encodes three distinct secreted proteins, Desert hedgehog (DHH), Indian hedgehog (IHH), and Sonic hedgehog (SHH) with critical roles in mammalian development. Of these, SHH directs the differentiation of enteric neural crest derived cells toward the neuronal lineage [31–36], and its aberrant activation is reported in multiple malignancies, including in GEP-NETs [37–39]. To confirm the expression of SHH in human dpNETs, we collected tumor samples across three institutional centers and performed immunohistochemical (IHC) staining on sections from these tumors. The SHH expression patterns in the tumors were compared to normal tissues from non-tumor bearing patients and to normal appearing adjacent tissues in patients with respective dpNETs. In normal adjacent human pancreas, we observed the presence of SHH immunoreactive cells mainly in the endocrine islet and not in the surrounding exocrine tissue (Fig. 1A and fig. S1A). Of the 20 PanNETs evaluated, 11 of the tumors showed positive immunoreactivity for SHH and its expression was restricted to the tumor stroma (10% of total cases) or tumor cells (45% of total cases), but not in both compartments of the same tumor (Fig. 1B and fig. S1A). In the adjacent normal duodenum, SHH expression was restricted to the base of the epithelial lumen and staining was not detected in the submucosal Brunner’s glands from which up to 60% of MEN1 DNETs reportedly develop (Fig. 1C and fig. S1B) [40]. Among the 40 DNETs examined, SHH was strongly expressed in 33 of the tumors and its expression was detected in the tumor-adjacent Brunner’s glands showing neuroendocrine cell hyperplasia (Fig. 1D). We next investigated the expression of PTCH1, the cognate receptor for the SHH ligand, in PanNETs and DNETs. Immunofluorescent staining of the tumors detected strong PTCH1 expression in tumor cells and the tumor stroma, with a greater percentage of both tumor types exhibiting stromal immunoreactivity (fig. S1 and S2). The expression of PTCH1 in tumor cells and the tumor stroma appeared independent of SHH status, as PTCH1 was detected in tumors with positive and negative SHH staining.

Fig. 1 — Hyperactivation of the Sonic hedgehog (SHH) signaling pathway in human pancreatic and duodenal NETs (dpNET). A Immunofluorescence (IF) images showing SHH expression (red) in normal adjacent human pancreas islet (nPANC) compared to (B) pancreatic NETs (PanNET). SHH is expressed in the PanNET microenvironment (top panel, tumor cells indicated by synaptophysin labeling in green) and by PanNETs (bottom panel). C IF images of SHH staining (red) in normal duodenum (nDUO, left panel) and Brunner’s glands (nBGs, top right panel). Bottom right panel shows SHH staining (white) without DAPI overlay. Arrowheads indicate to the luminal epithelium (EPI), lamina propria (LP), and crypts. D IF images showing SHH expression (red) in transitioning BGs (tBGs) and duodenal NETs (DNET). Right panel shows the wide-field image of SHH staining (white) without DAPI overlay. E Circle plots depicting the prevalence of *MEN1* mutations in PanNETs, DNETs, and jejunal-ileal NETs (JI-NET) based on Caris next generation sequencing data. F Kaplan-Meier survival curves depicting the top 75th percentile (high) and bottom 25th percentile (low) of *SMO* and *GLI2* expression in PanNETs (top panels) and *SMO* and *PTCH1* expression in DNETs (bottom panels) based on Caris WTS data. Event-free proportion is defined from the date of initial diagnosis to date of last contact or death. G Bar plots depicting the prevalence of *MEN1* mutations (including benign) in dpNETs with high or low SMO expression. H IF images of a representative PanNET exhibiting strong nuclear Menin expression (green). The same tumor was stained for SHH (red, middle) and PTCH1 (red, right). SYP⁺ tumor cells are shown in green. I IF images of a representative PanNET exhibiting reduced nuclear Menin expression (green) with strong SHH and PTCH1 immunoreactivity in the tumor. J IF images of a representative non-MEN1 DNET showing near absent Menin expression (green) and strong expression of SHH and PTCH1 in the tumor. K Summary table of the patient FFPE specimens with relevant clinical features that were evaluated for Menin, SHH, and PTCH1 expression by immunostaining. L, M Circle plots indicating the number and proportion of dpNETs with nuclear, cytoplasmic, or reduced Menin staining in the tumors

To investigate whether increased SHH expression in tumors correlated with genetic alterations in the SHH signaling pathway, we assessed The Cancer Genome Atlas (TCGA) for copy number variants (CNV) and enrichment of HH signaling genes in patients carrying a primary neuroendocrine diagnosis (TCGA-neuroendocrine). Compared to non-GI NETs, PanNETs showed an enrichment in copy number gains in HH pathway genes, e.g., SHH, SMO, PTCH1, GLI1, and GLI3 (fig. S3 and Table 1), and this was associated with increased transcript expression of HH pathway genes in these tumors (fig. S4). PanNETs exhibited copy number gains in several genes associated with SHH-interacting pathways, e.g., BMP, BCL2, EGFR, whereas neuroendocrine cases in the breast, skin, and lung tended to present with copy number gains in other growth signaling genes, e.g., MYC, PI3K, and GSK3 (Fig. 1F). Consistent with prior reports that Menin antagonizes HH signaling [26, 27], loss-of-function mutations in MEN1 were only identified in PanNETs and in one colon NET represented in the analysis. Human dpNETs were reported to express markers of glial cells [20, 30], and certain types of CNS tumors can arise as part of the MEN1 syndrome [41–43]. Thus, we explored whether genetic alterations in the HH signaling pathway are also shared by cancers with glial cell origins, including astrocytoma, glioma, and glioblastoma. Using TCGA, we examined the prevalence of CNVs in HH pathway genes for both neuroendocrine cancers and cancers with glial cell origins and found that approximately half of these cases exhibit copy number gains in SHH, SMO, and GLI3 (table S1). These trends persisted across MEN1-mutated cases in the combined cohort, with nearly 74% of cases exhibiting loss of MEN1 copy number and 47% with gains in the previous HH signaling genes (table S2).

Table 1.

Prevalence of copy number variations (CNV) in Hedgehog pathway genes for neuroendocrine cancers in the TCGA database

Gene ID	Gene Symbol	Gene Name	Cytoband	# of Cases Tested for CNVs	# of Cases with CNV Gain and (%)	# of Cases with CNV Loss and (%)
ENSG00000106571	GLI3	GLI family zinc finger 3	7p14.1	215	35 (16.3%)	4 (1.9%)
ENSG00000128602	SMO	smoothened, frizzled class receptor	7q32.1	215	32 (14.9%)	12 (5.6%)
ENSG00000164690	SHH	sonic hedgehog signaling molecule	7q36.3	215	30 (14.0%)	12 (5.6%)
ENSG00000111087	GLI1	GLI family zinc finger 1	12q13.3	215	22 (10.2%)	5 (2.3%)
ENSG00000185920	PTCH1	patched 1	9q22.32	215	13 (6.1%)	16 (7.4%)
ENSG00000117425	PTCH2	patched 2	1p34.1	215	6 (2.8%)	113 (52.3%)
ENSG00000074047	GLI2	GLI family zinc finger 2	2q14.2	215	3 (1.4%)	13 (6.1%)
ENSG00000133895	MEN1	Menin 1	11q13.1	215	5 (2.3%)	35 (16.3%)

Open in a new tab

Given the low abundance of gastrointestinal NETs represented in the TCGA-neuroendocrine cohort, we next analyzed the Caris clinico-genomic database containing molecular sequencing and survival data from 7,071 neuroendocrine cases. Menin is a known epigenetic repressor of the HH signaling pathway and MEN1 mutations have been reported in up to 40% of PanNETs [11, 12]. Whole exome sequencing (WES) from prior studies indicate that midgut and distal SI-NETs generally lack mutations in MEN1, however the abundance of MEN1 mutations in DNETs varies across studies, with reported ranges between 30 and 60% [7–10]. Thus, we mined the Caris clinico-genomic database and determined the frequency of MEN1 mutations in PanNETs, DNETs and JI-NETs as a possible mechanism for upregulated HH pathway signaling in dpPNETs. All MEN1 mutations identified by NGS of DNA (WES or a targeted 592-gene panel), including those classified as benign, were considered when calculating mutation frequencies. As expected, MEN1 was most frequently mutated in PanNETs (N = 367 out of 1015 cases, or 36%) followed by DNETs (N = 5 out of 66 cases, or 7.5%), whereas only 1% of JI-NETs carried a MEN1 mutation (N = 3 out of 297 cases) (Fig. 1E). Although DNETs presented with a higher MEN1 mutational burden relative to JI-NETs, this didn’t fully account for upregulated HH pathway signaling in these tumors.

To determine if the expression of individual HH pathway genes were associated with survival outcomes in dpNETs, we generated Kaplan-Meier curves for the top 75th percentile (HIGH) and bottom 25th percentile (LOW) of HH pathway transcript expression among a subset of tumors with whole transcriptome sequencing (WTS) performed and primary origin in the pancreas (N = 839) and duodenum (N = 54). High expression of SMO (P = 0.046) and GLI2 (P = 0.003) were associated with lower survival in PanNETs. The comparison was even more striking in DNETs, with high expression of SMO (P = 0.003) and PTCH1 (P = 0.032) correlating with poor survival (Fig. 1F). Surprisingly, further analysis showed that the mutational frequency of MEN1 across dpNETs did not correlate with high HH pathway transcript expression (Fig. 1G).

Given that the MEN1 mutation frequency in dpNETs did not correlate with high expression of SHH signaling pathway transcripts, we posited that altered subcellular localization of Menin independent of MEN1 status might contribute to increased HH pathway activity in these tumors. We evaluated the expression pattern of Menin protein by performing IHC staining on our original cohort of dpNETs. We consistently observed a loss in nuclear Menin expression, even in cases where no MEN1 mutation was detected (Fig. 1, H–K and fig. S5). Approximately 44% of PanNETs exhibited strong nuclear Menin staining comparable to the endocrine islets in the adjacent pancreas. These tumors were negative for SHH and expressed PTCH1 in SYP⁺ tumor cells (Fig. 1H). Menin expression was restricted to the cytoplasm in 22% of tumors and nuclear Menin levels were reduced, lost, or indistinct in 33% of PanNETs (Fig. 1L). These tumors exhibited strong SHH expression in SYP⁺ tumor cells and intense PTCH1 staining in the tumor stroma (Fig. 1I, fig. S1A, and fig. S2A). In contrast, only 20% of DNETs retained distinct nuclear Menin expression and more than one-half of the tumors showed cytoplasmic-restricted expression (Fig. 1, J and M). DNETs exhibiting reduced nuclear Menin expression showed high SHH immunoreactivity and stromal PTCH1 expression (Fig. 1K, fig. S1B and fig. S2B). Since the MEN1 status was unknown for a majority of the cases, we attempted to perform Sanger sequencing on FFPE sections of these archival tumors but failed to generate high quality genomic DNA for subsequent sequencing analysis. Nevertheless, by analyzing tumors with known MEN1 status, we concluded that decreased levels of nuclear Menin expression can occur in the absence of MEN1 mutation, and the loss of nuclear Menin in dpNETs strongly correlated with upregulated SHH pathway signaling in human dpNETs.

SHH regulates PanNET growth via GLI1/2 signaling

Classical activation of HH signaling occurs when the SHH ligand binds its cognate receptor PTCH1 and activates Smoothened (SMO) at the primary cilium, a nutrient sensing organelle expressed on stromal cells including enteric glia and select endocrine cell types [44]. Aberrant SHH activation, either through canonical SMO signal transduction or non-canonical activation of the transcriptional effectors GLI1/2, promotes tumor cell growth in basal cell carcinoma (BCC), medulloblastoma, and GEP-NETs [27, 38, 45–47]. As prior studies in NETs were limited to the use of monoclonal cell lines, we sought to develop an improved preclinical model and test these events in primary tumor organoids (tumoroids) derived from four surgically resected patient PanNETs. Following subculture for 21-days, we evaluated the expression of CHGA, SHH, and PTCH1 in the PanNET line that demonstrated the most robust growth (PanNET3). As expected, the tumoroids strongly expressed the neuroendocrine marker CHGA, in addition to SHH and PTCH1 (Fig. 2A). The specificity of these antibodies was confirmed following preadsorption with the respective recombinant proteins (fig. S6A). We next tested whether PanNET tumoroids respond to treatment with SHH and vismodegib, an FDA-approved inhibitor of SMO. In the presence of SHH, PanNET tumoroids increased their expression of neuroendocrine transcripts (e.g., SYP and CHGA), whereas treatment with vismodegib reduced CHGA mRNA levels. Given that SHH signaling regulates the cell fate of neural progenitors to acquire a neural or glial cell phenotype, we measured the expression of GFAP as a classical marker of the glial-restricted lineage. Consistent with its role in reprogramming the glial lineage, SHH treatment downregulated the expression of GFAP in PanNET tumoroids, whereas treatment with vismodegib increased GFAP expression (Fig. 2B).

Fig. 2 — Patient-derived PanNET tumoroids express HH signaling proteins and respond to HH pathway activation and inhibition. A Combined fluorescence and phase-contrast images of 21-day-old human PanNET tumoroids (PanNET3) stained for chromogranin A (CHGA) and HH proteins (PTCH1, SHH). B Four patient-derived PanNET tumoroid lines (PanNET1–4) were exposed to recombinant human SHH N-terminal peptide (100 ng/mL) or the SMO inhibitor vismodegib (VISMO, 20 µM) for 72 h and changes in mRNA expression were analyzed by RT-qPCR. mRNA changes were normalized to *HPRT1* expression and DMSO vehicle control. (n = 4 unique patient lines). * = p < 0.05, ** = p < 0.01, **** = p < 0.0001 by Two-way ANOVA with Sidak post-test. C EdU labeling showing proliferation of dissociated PanNET3 cells following 5-day exposure to: the HH agonists SHH-N (100 ng/mL) and SAG (10 nM); inhibitors of the canonical HH signaling pathway vismodegib (20 µM) and sonidegib (10 nM); or inhibitors of the GLI1/2 effectors GANT61 (10 µM) and itraconazole (ITZ, 1 µM). D Quantitation of the percentage of EdU-positive PanNET tumor cells following 5-day treatment. (n = 3 replicates from one patient tumoroid line). * = p < 0.05, by One-way ANOVA with Tukey post-test. E SHH or SAG were co-administered with the respective pharmacologic inhibitors and EdU uptake was evaluated after 7 days. F Immunofluorescent images of CHGA, SHH, and PTCH1 expression in a second PanNET tumoroid line (PanNET5). G, H EdU labeling was evaluated in PanNET5 tumoroids following 7-day treatment with SHH-N (200 ng/mL) and SAG (20 nM); inhibitors of the canonical HH signaling pathway vismodegib (20 µM) and sonidegib (10 nM); or inhibitors of the GLI1/2 effectors GANT61 (10 µM) and itraconazole (ITZ, 5 µM). I Crystal violet staining of human BON-1 PanNET cells after 48 h treatment. J BrdU incorporation in BON-1 cells after 48 h treatment with HH agonists SHH-N and SAG, (K) SMO inhibitors vismodegib and sonidegib, and (L) inhibitors of GLI1/2 signaling. M Immunofluorescent images of CHGA, SHH, and PTCH1 expression in tumoroids derived from a metastatic ileal NET (IL-NET-met1). N, O EdU labeling was assayed in the IL-NET-met1 tumoroid line following 7-day treatment with the same drug concentrations used for PanNET5. * = p < 0.05, ** = p < 0.01, *** = p < 0.001, **** = p < 0.0001 by One-way ANOVA with Dunnett post-test

Next, we evaluated how modulating SHH pathway activation regulated proliferation of PanNET tumoroid cells. We observed a two-fold increase in PanNET tumoroid cell proliferation following treatment with SHH and SAG, a potent SMO agonist, as demonstrated by increased 5-ethynyl-2’-deoxyuridine (EdU) incorporation by tumor cells (Fig. 2C). Treatment with the SMO inhibitors vismodegib and sonidegib did not affect proliferation, however inhibition of the GLI1/2 transcriptional effectors using either GANT61 or itraconazole, a pan inhibitor of SMO and GLI1/2, led to a two-fold reduction in EdU labeling (Fig. 2D). Co-administration of SHH and SAG with the respective HH pathway inhibitors led to a decrease in growth induction that was most pronounced with GLI1/2 inhibition (Fig. 2E). These data suggest that both canonical activation of SMO and non-canonical GLI1/2 signaling contribute to SHH and SAG induced tumoroid growth. To introduce further rigor to these studies, we repeated the cell growth assays using three additional PanNET tumoroid lines expressing CHGA, SHH, and PTCH1 (Fig. 2, F–G, fig. S6B and C). Of these, only one of the tumoroid lines demonstrated robust EdU uptake across multiple passage numbers (PanNET5) (Fig. 2, F–H). Exposure to SHH and SAG increased tumoroid cell growth in all three PanNET lines, whereas treatment with GANT61 resulted in the greatest level of growth suppression (Fig. 2H, fig. S6B and C). Given that primary tumoroid cultures can display heterogeneity in drug response, we repeated the cell growth assays in a well-established monoclonal PanNET cell line (human BON-1 cells) treated with these pharmacological modulators (Fig. 2I). Consistent with the response in patient-derived tumoroids, SHH and SAG treatment increased the proliferation of BON-1 cells in a dose-dependent manner (Fig. 2J). BON-1 cells were generally less sensitive to SMO inhibition (Fig. 2K, whereas direct inhibition of GLI1/2 resulted in significant growth suppression (Fig. 2L). Thus, SHH pathway activation stimulated PanNET growth in a humanized preclinical model and pharmacologic inhibition of the non-canonical HH signaling pathway effectively blocked SHH-mediated tumor cell growth.

We next tested the impact of SHH pathway modulation in a primary tumoroid line derived from a metastatic liver tumor originating from a primary distal ileal NET (IL-NET-met1). This tumoroid line demonstrated strong expression of CHGA and SHH but showed negative PTCH1 staining (Fig. 2M). Compared to the PanNET3 and PanNET5 lines, IL-NET-met1 tumoroids were characterized by a slower growth rate but still showed an increase in EdU uptake in the presence of SHH and SMO that was further dampened with pharmacologic inhibition (Fig. 2, N and O). Finally, to address whether the previous response to HH pathway modulation is influenced by MEN1 status, we performed Sanger sequencing on the parent tumors and tumoroids. Sanger sequencing of the MEN1 protein coding exons 2 through 9 identified multiple SNVs in these tumors (Table 2). Notably, only one tumor carried an SNV that was classified as a pathogenic mutation based on the ClinVar Miner database (PanNET4). This line could not be subcultured for quantitative assays, however the response to HH pathway modulation appeared consistent with the other tumoroid lines carrying benign SNVs in the MEN1 locus (fig. S6B).

Table 2.

MEN1 Sanger mutation analysis of patient-derived tumoroid lines

Patient Tumoroid ID	HGVS Nomenclature	Chromosome Position (GRCh38)	Exon	ClinVar Classification
PanNET2	NM_001370259.2(MEN1):c.1254 C > T	11:64805130	9	Benign
	NM_001370259.2(MEN1):c.1299T > C	11:64805085	9	Benign
	NM_001370259.2(MEN1):c.1621 A > G	11:64804546	10	Benign
PanNET3	NM_001370259.2(MEN1):c.1254 C > T	11:64805130	9	Benign
	NM_001370259.2(MEN1):c.1299T > C	11:64805085	9	Benign
	NM_001370259.2(MEN1):c.1621 A > G	11:64804546	10	Benign
PanNET4	NM_001370259.2(MEN1):c.207del	11:64809903	2	Pathogenic
	NM_001370259.2(MEN1):c.1299T > C	11:64805085	9	Benign
	NM_001370259.2(MEN1):c.1621 A > G	11:64804546	10	Benign
	NM_001370259.2(MEN1):c.*307T > G	11:64804027	2	Benign
PanNET5	NM_001370259.2(MEN1):c.1254 C > T	11:64805130	9	Benign
	NM_001370259.2(MEN1):c.1299T > C	11:64805085	9	Benign
	NM_001370259.2(MEN1):c.1621 A > G	11:64804546	10	Benign
PanNET6	NEG
IL-NET-met1	NM_001370259.2(MEN1):c.1299T > C	11:64805085	9	Benign
	NM_001370259.2(MEN1):c.1621 A > G	11:64804546	10	Benign

Open in a new tab

Loss of Men1 in GFAP⁺ and SOX10⁺ glial cells drives the development of HH-expressing GEP-NETs

Given that human dpNETs overexpress HH signaling proteins and neuroglial markers, we posited that these neoplasms might arise from multiple cellular sources including neuroectoderm derived precursors that are responsive to SHH signaling. In support of this premise, prior work demonstrated that selective deletion of Menin in glial cells expressing the human GFAP promoter or SRY-box transcription factor 10 (SOX10) was sufficient to stimulate the development of pancreatic and pituitary NETs. GFAP-Cre; Men1^FL/FL (GFAP^ΔMen1) mice developed NETs by 15 to 17-months of age, whereas tumors developed by 11-months of age in the Sox10-Cre; Men1^FL/FL (Sox10^ΔMen1) mice [24]. We subsequently analyzed 18 to 20-month-old GFAP^ΔMen1 and Sox10^ΔMen1 mice and observed a low incidence of SI tumors arising in the antropyloric Brunner’s glands, duodenum, jejunum, and ileum. Unlike PanNETs that exhibited a well-differentiated NET phenotype, SI tumors appeared to be poorly differentiated (Fig. 3A). IHC analysis of the tumors identified absent or low nuclear Menin expression compared to adjacent and control tissues, supporting the premise that these NETs might develop from a Menin-deficient cell population (Fig. 3, B and C). Tumors from the ΔMen1 mice expressed neuroendocrine markers, including SYP, CHGA, and PAX6, and thus, we identified them as bona fide NETs (Fig. 3, D and E). Mirroring our previous observations in human dpNETs, the mouse tumors showed increased expression of SHH signaling pathway proteins, with PanNETs showing stronger PTCH1 expression and DNET/JI-NET tumors showing stronger SHH immunoreactivity (Fig. 3E). Further analysis of the ΔMen1 mouse tumors confirmed significantly elevated expression of PTCH1, GLI1, and GLI2 proteins in PanNETs compared to age-matched wild type pancreas tissues (Fig. 3, E and F). In comparison, DNETs/JI-NETs in the ΔMen1 mice were significantly enriched in SHH protein expression but exhibited variable expression of GLI2 and PTCH1 proteins (Fig. 3, E and G). We next determined whether enteric glial cells from normal mouse pancreas and duodenum could also express HH pathway proteins. We isolated glial cells from normal wild type mouse tissues and observed the expression of SHH, PTCH1, and GFAP in primary cell cultures (Fig. 3H). Hence, we found it ostensible that GFAP⁺ glial cells might contribute to the development of dpNETs through upregulated HH pathway signaling.

Fig. 3 — GFAP-directed *Men1* deletion stimulates pancreatic and small intestinal NETs that overexpress HH signaling proteins. A Representative hematoxylin and eosin (H&E) stained images of a pancreatic and duodenal NET (PanNET in top panel and DNET in bottom panel) from 18-month-old *GFAP-Cre: Men1*^FL/FL (*GFAP*^ΔMen1) mice. B IF stained images showing Menin (red) expression in a *GFAP*^ΔMen1 PanNET and the adjacent exocrine and endocrine (islet encircled in white dotted line) pancreas (Adj PAN). Menin expression in the endocrine islet and exocrine pancreas of a littermate control is shown for comparison. C IF stained images showing Menin (red) expression in a *GFAP*^ΔMen1 DNET and the adjacent duodenum (Adj DUO) with arrow indicating to the lamina propria (LP). Menin expression in the duodenum of a littermate control is shown for comparison. DAPI (blue). D IF stained images of a *GFAP*^ΔMen1 PanNET and DNET: Left panel shows SHH (yellow), and right panel shows PTCH1 (yellow) co-stained with SYP (magenta) and DAPI (cyan). E Western blot analysis of HH signaling proteins and neuroendocrine markers in littermate wild type pancreas (WT PAN), duodenum (WT DUO), and *GFAP*^ΔMen1 PanNETs and DNETs/JI-NETs. (n = 4). F, G Quantitation of western blot analysis in panel (E). (n = 4). * = p < 0.05, ** = p < 0.01, **** = p < 0.0001 by Two-way ANOVA with Sidak post-test. H Western blot analysis of four day-old primary enteric glial cell cultures generated from the pancreas and proximal duodenum (including Brunner’s glands) of 6-month-old wild type mice. (n = 3 independent cultures generated from 6 mice)

To visualize the fate of GFAP⁺ and SOX10⁺ glial cells, we crossed the previous GFAP^ΔMen1 and Sox10^ΔMen1 mice to express the lox-stop-lox-tdTomato fluorescent reporter (LSL-tdTomato). We confirmed appropriate tissue-restricted expression of the fluorescent reporter and validated the absence of tdTomato transgene expression in endocrine cells (fig. S7), we analyzed tumors from the ΔMen1 mice and observed the presence of tdTomato⁺ signal in three mouse PanNETs and in one DNET arising in the Brunner’s glands (Fig. 4, A and B). Subsequently, we generated primary tumoroids from these tdTomato⁺ ΔMen1 tumors for further in vitro studies. Prior to conducting these in vitro experiments, we confirmed that tumoroids derived from the tdTomato⁺ ΔMen1 tumors expressed the fluorescent reporter. Subcellular tdTomato expression in the tumoroids was consistent with nuclear SOX10 expression and localization of GFAP to the cytoplasm (Fig. 4, C and D). After confirming that the ΔMen1 tumoroids express CHGA, we further evaluated their expression of the mitotic marker Ki-67 and HH pathway proteins. Mirroring the expression of parent mouse tumors, ΔMen1 tumoroids demonstrated strong immunoreactivity for SHH, PTCH1, and SMO proteins (Fig. 4, E and F). Compared to normal mouse pancreas organoids, PanNET tumoroids showed elevated expression of Shh, whereas Gli2 expression was decreased. In contrast, DNET tumoroids showed higher transcript levels of Gli1 and Gli2, suggesting potential tissue-specific differences in the activation of the HH signaling pathway in these tumors (Fig. 4, G and H). Lastly, we confirmed that mouse PanNET and DNET tumoroids produce elevated levels of insulin, glucagon, and gastrin, hence, recapitulating parent tumor phenotypes (Fig. 4, I and J).

Fig. 4 — Pancreatic and small intestinal NET tumoroids from *GFAP*^ΔMen1 and *Sox10*^ΔMen1 mice recapitulate parent tumor phenotypes. A IF image showing tdTomato (red) expression in a PanNET from a 17-month-old *Sox10-Cre: Men1*^FL/FL: *LSL-tdTomato* mouse (*Sox10*^{ΔMen1; tdTomato}). Inset shows the macroscopic image of the associated pancreas with two large PanNETs. Arrowhead indicates to the tdTomato⁺ PanNET. B IF image showing tdTomato (red) expression in a DNET from a 19-month-old *GFAP-Cre: Men1*^FL/FL: *LSL-tdTomato* mouse (*GFAP*^{ΔMen1; tdTomato}). Inset shows the macroscopic image of the associated stomach and proximal duodenum. Arrowhead indicates to the tdTomato⁺ DNET arising in the pyloric Brunner’s glands. C IF image showing tdTomato (magenta) expression in a 14-day-old tumoroid derived from a *Sox10*^{ΔMen1; tdTomato} PanNET. Left panel shows overlay with DAPI (cyan) and the right panel shows the tdTomato channel only. D IF image showing tdTomato (magenta) expression in a 14-day-old tumoroid derived from a *GFAP*^{ΔMen1; tdTomato} DNET. Left panel shows overlay with DAPI (cyan) and the right panel shows the tdTomato channel only. E IF images of *Sox10*^{ΔMen1; tdTomato} PanNET tumoroids stained for CHGA, Ki-67, and HH signaling proteins (yellow) co-localized with tdTomato (magenta) and DAPI (cyan). F IF images of *GFAP*^{ΔMen1; tdTomato} DNET tumoroids stained for CHGA, Ki-67, and HH signaling proteins (yellow) co-localized with tdTomato (magenta) and DAPI (cyan). Panels (*C-F*) represent maximum intensity projections of Z-stack images of the tumoroids. G Normalized expression of HH pathway transcripts in normal pancreas organoids generated from six-month-old wild type mice (n = 4 mouse organoid lines) compared to PanNET tumoroids (n = 4 mouse PanNET tumoroid lines. H Expression of HH pathway transcripts in normal duodenal organoids from 6-month-old wild type mice (n = 5 mouse organoid lines) compared to DNET tumoroids (n = 2 mouse lines across 2–3 biological replicates). * = p < 0.05 by Two-way ANOVA with Sidak post-test. I Levels of secreted insulin, glucagon, and gastrin hormones in 48 h conditioned media collected from wild type pancreas organoids and PanNET tumoroids (n = 5), and (J) in wild type duodenal organoids (n = 5) and DNET tumoroids (n = 2 lines across 2–3 replicates). * = p < 0.05, ** = p < 0.01 by unpaired Student’s T-test with Tukey post-test

Activation of GLI1/2 regulates the growth of ΔMen1 GEP-NET tumoroids

Unlike patient-derived tumoroids that exhibit indolent growth over subsequent passage numbers [6], the ΔMen1 tumoroids were robust, able to be passaged and proliferated in normal organoid growth media. Thus, we leveraged these in vivo models to further investigate the NET cell response to SHH pathway modulation (Fig. 5A). We generated tdTomato⁺ ΔMen1 PanNET tumoroid lines from two independent mouse tumors and treated the tumoroids with HH pathway agonists, pharmacologic inhibitors of SMO, or inhibitors of GLI1/2 signaling (Fig. 5B). Consistent with the response shown by patient-derived PanNET tumoroids, ΔMen1 PanNET tumoroids demonstrated increased growth upon SHH pathway activation (Fig. 5C). Compared to inhibition of the canonical pathway, pan-inhibition of SMO and GLI1/2 inhibited tumoroid growth by 50 to 90% (Fig. 5, D and E). These patterns remained consistent when tumoroids were treated together with respective pharmacological inhibitors and the SHH ligand (Fig. 5F). Given that patient-derived PanNET tumoroids upregulated their expression of neuroendocrine transcripts in response to SHH treatment, we further tested whether ΔMen1 PanNET tumoroids undergo similar transcriptional changes in response to HH pathway modulation. Indeed, adding SHH ligand increased the expression of Chga in ΔMen1 PanNET tumoroids, whereas treatment with itraconazole or vismodegib suppressed expression (Fig. 5G).

Fig. 5 — HH signaling regulates the growth of *GFAP*^ΔMen1 and *Sox10*^ΔMen1 pancreatic NET tumoroids. A Phase contrast and (B) fluorescence images of PanNET tumoroids from *Sox10-Cre; Men1*^FL/FL;*LSL-tdTomato* mice. Tumoroids were imaged after 72 h exposure to: the HH agonists SHH-N (100 ng/mL) and SAG (10 nM); inhibitors of the canonical HH signaling pathway vismodegib (20 µM) and sonidegib (10 nM); or inhibitors of the GLI1/2 effectors GANT61 (10 µM) and itraconazole (ITZ, 1 µM). C TdTomato fluorescence intensity was used to measure PanNET tumoroid growth in the presence of HH pathway agonists or (D) inhibitors of the canonical and (E) non-canonical HH signaling pathways. Fluorescence signal is compared to DMSO vehicle control. (n = 3 replicates in two unique mouse PanNET tumoroid lines). ** = p < 0.01, *** = p < 0.001, **** = p < 0.0001 by One-way ANOVA with Dunnett post-test. F BrdU incorporation was used to measure tumoroid proliferation in *GFAP*^ΔMen1 and *Sox10*^ΔMen1 PanNET tumoroids after 72 h treatment. (n = 4). ** = p < 0.01, **** = p < 0.0001 by One-way ANOVA with Dunnett post-test. G Relative fold-change in *Chga* mRNA levels in PanNET tumoroids following 72 h treatment. (n = 5). H Western blot analysis of HH pathway proteins in PanNET tumoroids after 72 h treatment. I Quantitation of SHH protein expression normalized to beta-actin and DMSO vehicle control from the western blot analysis in panel (H). (n = 3). J Western blot analysis and (K) associated quantitation of phosphorylated and total ERK and AKT growth pathways in PanNET tumoroids after 72 h treatment. (n = 3). * = p < 0.05, ** = p < 0.01, *** = p < 0.001 by Kruskal-Wallis test

To confirm the downstream signaling targets of SHH pathway modulation, we measured the protein levels of SHH, GLI2, PTCH1, and SMO in ΔMen1 PanNET tumoroids following SHH pathway activation and inhibition. As anticipated, SHH treatment induced the expression of HH pathway proteins and inhibition with itraconazole, but not vismodegib, reversed the induction of these proteins (Fig. 5, H and I). Prior reports showed that SHH signaling can activate ERK and AKT through direct crosstalk with these signaling pathways [48]. Therefore, we examined the impact of SHH pathway modulation on the expression and activation of ERK and AKT proteins in ΔMen1 PanNET tumoroids. Consistent with the previous effects on tumoroid growth, SHH treatment induced the expression of phosphorylated and total ERK and AKT, and these effects were attenuated under pan-inhibition of SMO and GLI1/2, but not when SMO was inhibited alone (Fig. 5, J and K). Collectively, these observations validated our findings in patient-derived PanNET tumoroids and implicated the non-canonical SHH signaling pathway in driving tumor cell growth.

Given that human and dpNETs overexpress SHH, we rationalized that these neoplasms might also respond to HH pathway inhibition. We repeated the previous drug studies on a DNET tumoroid line that was established from a tdTomato⁺ ΔMen1 tumor located in the Brunner’s glands (Fig. 6, A–C). Mimicking the response by ΔMen1 PanNET tumoroids, the DNET tumoroids showed increased growth in response to SHH signaling pathway activation and inhibition of GLI1/2, but not inhibition of SMO alone, resulted in significant growth inhibition. We validated these studies using a second DNET tumoroid line and compared the response to a third tumoroid line derived from a jejunal tumor (Fig. 6, D and E). Interestingly, SHH treatment did not lead to a significant increase in HH protein levels or the expression of ERK and AKT beyond baseline levels, suggesting that high basal expression of SHH ligand in DNETs/JI-NETs contributed to autocrine growth signaling (Fig. 6, F–H). Despite exhibiting reduced sensitivity to SHH ligand, pan-inhibition of SMO and GLI1/2 was nonetheless effective in reducing tumoroid growth and downregulated the expression of ERK and AKT proteins (Fig. 6G). Finally, we compared the drug response of the ΔMen1 DNET tumoroids to STC-1 cells derived from a mouse ileal NET (Fig. 6I). SHH and SAG treatment stimulated STC-1 cell proliferation in a dose-dependent manner and inhibition of both canonical and non-canonical HH pathways resulted in the opposite effect (Fig. 6, J–L). These results indicated that STC-1 cells are characterized by increased sensitivity to modulation of the canonical HH signaling pathway, whereas primary DNET tumoroids more closely mimicked the drug response shown by patient-derived PanNET tumoroids.

Fig. 6 — *GFAP*^ΔMen1 DNETs and jejunal NETs are sensitive to HH pathway activation and inhibition. A Phase contrast and (B) fluorescence images of DNET tumoroids from a *GFAP-Cre; Men1*^FL/FL;*LSL-tdTomato* mouse. Tumoroids were imaged after 72 h exposure to: the HH agonists SHH-N (100 ng/mL) and SAG (10 nM); inhibitors of the canonical HH signaling pathway vismodegib (20 µM) and sonidegib (10 nM); or inhibitors of the GLI1/2 effectors GANT61 (10 µM) and itraconazole (ITZ, 1 µM). C TdTomato fluorescence intensity was used to measure DNET tumoroid growth in the presence of HH pathway inhibitors. Fluorescence signal is compared to DMSO vehicle control. (n = 3 replicates using one mouse DNET tumoroid line). ** = p < 0.01, *** = p < 0.001, **** = p < 0.0001 by One-way ANOVA with Dunnett post-test. D Relative BrdU incorporation in a second *GFAP*^ΔMen1 DNET tumoroid line and (E) a jejunal tumoroid line (J-NET) after 72 h treatment. (n = 4). ** = p < 0.01, *** = p < 0.001, **** = p < 0.0001 by One-way ANOVA with Dunnett post-test. F Western blot analysis of SHH, ERK, and AKT growth pathways in J-NET tumoroids after 72 h treatment. G Western blot quantitation of SHH and (H) phosphorylated and total ERK and AKT proteins normalized to GAPDH and DMSO vehicle control. (n = 3). * = p < 0.05 by Kruskal-Wallis test. I Crystal violet staining of mouse STC-1 SI-NET cells after 48 h treatment with agonists and inhibitors of the HH signaling pathway. J BrdU incorporation in STC-1 cells after 48 h treatment with HH agonists SHH-N and SAG, (K) SMO inhibitors vismodegib and sonidegib, and (L) inhibitors of GLI1/2 signaling. * = p < 0.05, ** = p < 0.01, *** = p < 0.001, **** = p < 0.0001 by One-way ANOVA with Dunnett post-test

Reciprocal signaling by Menin and SHH regulate enteric glial cell identity

Although our previous observations in the tdTomato⁺ ΔMen1 mice indicated that dpNETs can develop from GFAP⁺ and SOX10⁺ enteric glial cells (EGCs), the mechanisms leading up to these events remained undefined. To gain further insight into the transcriptional profiles of the GFAP^ΔMen1 tumors, we performed single cell RNA sequencing on three pooled GFAP^ΔMen1 PanNETs. We identified heterogenous cell populations in the tumors that included acinar, beta, alpha, endothelial, immune, and mesenchymal cell populations (fig. S8A–C). We next mapped the expression of glial and endocrine transcriptional programs based on publicly available single-cell RNA-seq datasets of mouse pancreatic tissue. From this analysis, we identified three distinct clusters of cells characterized by high expression of glial genes (Glial^HIGH), high expression of endocrine genes (Endocrine^HIGH), and elevated expression of both programs (Glial^HIGH/Endocrine^HIGH). The Glial^HIGH population overlapped strongly with the mesenchymal cell cluster, whereas the Glial^HIGH/Endocrine^HIGH program showed the greatest overlap with the beta-neuroendocrine cell cluster (fig. S8D). These observations raise the potential of a glial-to-neuroendocrine transition characterized by global transcriptional reprogramming. To test whether Men1 loss in a glial cell can induce a similar glial-to-endocrine conversion in vitro, we generated primary mouse EGC cultures from the proximal duodenum of Sox10-CreER^T2-LSL-tdTomato mice. TdTomato fluorescence was induced by treating the cells with 4-hydroxytamoxifen (4-OHT) and the resulting tdTomato⁺ EGCs were purified by fluorescence activated cell sorting (FACS) to yield a pure SOX10⁺ cell population (Fig. 7, A–C). We next knocked down the expression of Men1 in EGC subcultures using small interfering RNAs and evaluated the expression of HH signaling pathway genes (Fig. 7D). Men1 silencing resulted in significant upregulation of Shh and Gli1 transcripts, and this was consistent with increased protein expression (Fig. 7, E–G). We further evaluated whether Men1 knockdown led to reprogramming of EGCs by altering the expression of glial and neuroendocrine lineage transcripts. Indeed, we observed a downregulation in several glial lineage-restricted genes (e.g., Gfap, S100b, Fabp7, and Lpar1) and upregulated expression of neuroendocrine and neural progenitor transcripts that are also enriched in GEP-NETs (e.g., Chga, Pax6, and Ascl1) (Fig. 7, H and I). To test whether these events are mediated by upregulated SHH signaling in response to loss of Men1, we treated EGCs with GANT61 or vismodegib and measured changes in the expression of HH and neuroendocrine genes. As predicted, inhibition of GLI1/2 with GANT61 reversed the induction of HH transcripts and this coincided with downregulated expression of Chga and Pax6 (Fig. 7, J and K). In comparison, SMO inhibition with vismodegib did not significantly alter the expression of neuroendocrine transcripts in Men1-deleted cells. Given that Menin was reported to antagonize HH signaling [26, 27], we evaluated whether GLI1/2 inhibition resulted in reciprocal feedback on Menin expression in EGCs. Interestingly, GANT61 treatment rescued nuclear Menin expression in EGCs within three days of Men1 silencing, suggesting that Menin and the GLI1/2 effectors participate in reciprocal regulation (Fig. 7L).

Fig. 7 — Loss of *Men1* in enteric glial cells stimulates GLI1/2-dependent transcriptional reprogramming. A Combined fluorescence and phase contrast images of 5-day-old primary enteric glial cell (EGC) cultures from *Sox10-CreER*^T2;*LSL-tdTomato* mice and *CreER*^T2 negative controls. Top panel shows TdTomato+ EGCs after 48 h exposure to 4-hydroxytamoxifen 4-OHT (2 µM). B TdTomato⁺ EGCs were sorted by FACS to enrich for a pure SOX10 ⁺ cell population. C Combined fluorescence and phase contrast images of FACS-enriched SOX10-tdTomato⁺ EGCs. D Fluctuations in HH pathway mRNA levels were evaluated in SOX10-tdTomato⁺ EGCs 72 h following siRNA-mediated *Men1* silencing. siRNA treatment consisted of four pooled siRNAs targeting the *Men1* gene (si-Men1, 25 nM) or non-targeting (si-NT, 25 nM) controls. (n = 5). E Immunofluorescence images of SHH expression in si-NT and *si-Men1* treated EGCs (SHH = red pseudo-color, DAPI = blue). Inset shows higher power image. F Western blot analysis of si-NT and *si-Men1* EGCs after 72 h treatment. SHH-FL = 55 kDa full length peptide; SHH-N = 22 kDa N-terminal peptide. (n = 3). G Quantitation of protein expression in panel (F) normalized to GAPDH loading control. (n = 3). H Relative fold-change in glial lineage transcripts and (I) neuroendocrine and neural progenitor transcripts in si-NT and *si-Men1* treated EGCs. (n = 6). J qPCR analysis of HH pathway genes and (K) neuroendocrine and neural progenitor transcriptsin si-NT and *si-Men1* EGCs after 72 h treatment with GANT61 (10 µM) or vismodegib (VISMO 20 µM). (n = 3). For all plots, * = p < 0.05, ** = p < 0.01, *** = p < 0.001, **** = p < 0.0001 by Two-way ANOVA with Sidak post-test. L Immunofluorescence images of si-NT and *si-Men1* EGCs after 96 h siRNA knockdown and 72 h treatment with vehicle or GANT61. Menin = green, GFAP = magenta, SHH = yellow. M Significant GSEA pathways in enteric glial cells following 5-days of si-*Men1* knockdown compared to non-targeting control. GSEA was performed on *Men1*-depleted cells and cells co-treated with (N) si-*Men1*, si-*Gli1*, and si-*Gli2*, and (O) cells co-treated with si-*Men1* and GANT61 (10 µM). P–R KEGG pathway enrichment analysis comparing the same groups as shown in panels (M–O). S Heatmap showing significant DEGs mapped to the cell cycle, HH signaling, epigenetic regulation, neural stem cell (NSC) reprogramming, neuronal and neuroendocrine differentiation

To determine whether mEGCs undergo transcriptional reprogramming toward a neuroendocrine lineage, we performed unbiased bulk RNA sequencing on mEGCs following 5-days of siRNA treatment. Consistent with the prior studies, Gene Set Enrichment Analysis (GSEA) identified upregulated HH signaling targets after Men1 silencing (Fig. 7M). Additionally, Men1 depletion in mEGCs led to upregulated Myc, E2F, and G2-M checkpoint targets indicating a shift toward a pro-proliferative program in these cells. To evaluate whether genetic knockdown of Gli1 and Gli2 is sufficient to reverse these transcriptional events, we profiled mEGCs co-treated with siRNAs targeting Men1, Gli1, and Gli2. Complementary to this approach, we evaluated Men1-depleted mEGCs that were exposed to the GLI1/2 inhibitor GANT61. Compared to mEGCs in which Gli1/2 was silenced or pharmacologically inhibited, Men1-depleted mEGCs demonstrated increased expression of gene targets mapped to p53 signaling, apoptosis, and decreased UV response (Fig. 7, N and O). Subsequent KEGG pathway analysis confirmed these findings and identified increased DNA replication, cell cycle, nucleocytoplasmic transport, and chromatin remodeling in Men1-depleted mEGCs (Fig. 7P–R). Consistent with the GSEA and KEGG analyses, Men1-silencing increased the expression of HH target genes, pro-proliferative genes, (Mki67, Top2a, Ccnd1, and Ccnd2), and chromatin-modifying enzymes known to interact with Menin (Atrx, Kmt2a, and Kdm5b). Critically, we identified a significant upregulation in several neuronal (Nrxn1, Robo3, and Slit2), neuroendocrine (Igfbp3, Hap1, Syt11, and Syt13), and neural stem cell reprogramming genes (Nes, Sox9, Klf4, and Olfm1) in Men1-depleted cells that were further downregulated with Gli1/2 knockdown or GANT61 treatment (Fig. 7S). In summary, these studies demonstrated that EGCs reprogram from a glial-restricted lineage and acquire a neuroendocrine phenotype in response to Men1-mediated HH pathway activation.

SHH signaling in GFAP⁺ glia is required for GEP-NET development

Our studies using primary EGCs indicated that pharmacologic inhibition of downstream SHH signaling effectors could reverse the glial-to-neuroendocrine transition. The extent to which these events occur in vivo and their impact on GEP-NET development remained unexplored. Thus, we sought to translate our in vitro observations by generating GFAP^ΔMen1 mice that harbored Kif3a-deficient EGCs showing impaired SHH signaling (GFAP^{ΔMen1; ΔKif3a}). Kif3a encodes a structural protein for primary cilia, a required component of SHH signal transduction present on many stromal cell types including those derived from the neural crest [49–52]. We evaluated mice aged 17-to 20-months for the presence of NETs and observed a remarkable reduction in the number of PanNETs and DNETs/JI-NETs in the GFAP^{ΔMen1; ΔKif3a} mice compared to mice with an intact Kif3a locus (Fig. 8, A–D). Consistent with a reduction in tumors, the GFAP^{ΔMen1; ΔKif3a} mice exhibited significantly lower expression of circulating pancreatic and gastrointestinal hormones (Fig. 8, E–I). Intriguingly, the incidence of pituitary NETs and the expression of prolactin, the dominant hormone expressed by these tumors [24], were not affected by Kif3a deletion (Fig. 8, J and K). This suggested that unlike NETs that developed in the GI tract, the development of pituitary NETs did not depend on activated SHH signaling in GFAP⁺ glia.

We further confirmed downregulated expression of HH proteins in tissues from the GFAP^{ΔMen1; ΔKif3a} mice (Fig. 8, L and M). These results were consistent with the presence of shortened primary cilia surrounding the endocrine islets of these mice compared to the cilia expressed on GFAP^ΔMen1 PanNETs (Fig. 8, N and O) [24]. Further IHC analysis of these tissues showed that nuclear Menin expression was lost in GFAP^ΔMen1 PanNETs, whereas the endocrine islets of GFAP^{ΔMen1; ΔKif3a} mice showed positive nuclear Menin expression. We next analyzed the DNETs and the proximal intestine from the respective genetic mice and observed a similar rescuing of nuclear Menin expression in the lamina propria of the intestinal mucosa (Fig. 8, P and Q). Finally, to confirm the specificity of these events, we generated primary EGCs from GFAP-CreER^{T2ΔMen1; ΔKif3a} mice and validated their expression of glial proteins (Fig. 8R). Following induction with 4-OHT, we observed a reduction in Gli1 and Gli2 transcript levels that coincided with decreased Pax6 expression in EGCs from the GFAP-CreER^{T2ΔMen1; ΔKif3a} mice (Fig. 8S). Collectively, these studies demonstrated that activation of HH signaling in GFAP⁺ EGCs is a required step in ΔMen1 GEP-NET formation and underscores a glial cell of origin for these tumors.

Discussion

Augmented SHH signaling, either through canonical SMO signal transduction or non-canonical activation of the transcriptional effectors GLI1/2, promotes cellular growth in pancreatic adenocarcinoma (PDAC), basal cell carcinoma (BCC), medulloblastoma, and gastroenteropancreatic neuroendocrine tumors (GEP-NETs) [27, 38, 45–47]. Targeting HH clinically has proven more nuanced, with HH inhibitors prolonging overall survival in clinical trials of BCC and medulloblastoma and accelerating disease progression in certain cases of PDAC [45, 46, 53–56]. This paradoxical response to HH inhibition is attributed to fundamental differences in how these tissues respond to HH pathway activation and whether these signals function through cell autonomous, non-cell autonomous, canonical, and non-canonical signaling pathways. In cancers with neuroectoderm origin, such as BCC and medulloblastoma, tumor secreted SHH acts in an autocrine fashion to promote cellular transformation, whereas endoderm derived cancers including PDAC, rely on paracrine SHH signals that can stimulate or restrain cell growth [56, 57]. Thus, identifying the cellular sources of SHH signals and how they drive malignant progression in tissues with ectodermal and endodermal origins are essential to designing effective clinical therapies and underscores the importance of this study.

Our findings confirm a pro-tumorigenic role for activated HH signaling in GEP-NET development and point to the potential promise of targeting aberrant SHH activation in the clinic. Current FDA-approved compounds that target this pathway are restricted to SMO inhibitors, which were shown in our study to have limited therapeutic effect compared to direct inhibitors of the GLI1/2 transcription factors. Hence, our findings warrant a reassessment of pharmacologic inhibitors of the non-canonical HH signaling pathway in targeting these tumors. We further elucidate on a mechanism for HH activation in GEP-NETs and establish the role of SHH in stimulating the transformation of EGCs, a stromal cell population that demonstrates transcriptional plasticity and represents an alternative etiology for these cancers. Our results suggest that HH overactivation in GEP-NETs mimics the pattern of SHH signaling in neuroectoderm derived cancers, and inhibition of this pathway poses therapeutic benefit by restricting the reprogramming of glial cells into hormone-producing NETs.

Recent multiome sequencing demonstrated that EGCs maintain lineage plasticity and can give rise to a neurogenic cell lineage, yet the signaling cues that regulate these differentiation events remain unknown. GFAP⁺ glial cells of the central nervous system were previously shown to reprogram into neural progenitor cells and functional neurons in the presence of activated SHH signaling [35, 58]. Further, SHH pathway activation upregulated the expression of neural progenitor cell (NPC) transcription factors (e.g., PAX6, ASCL1, and SOX2) that support glial-to-neural reprogramming and are essential for neuroendocrine cell specification [59–62]. Although these events are well defined in CNS glia, pathological activation of HH signaling in EGCs that promotes a similar NPC program has not been tested since they are not considered to be a source of cancer. Our findings address this gap in knowledge by showing that EGCs can reprogram toward a neuroendocrine lineage in response to Men1-mediated activation of the HH signaling pathway.

Finally, we point to several limitations in the present study that limit a broader interpretation of these findings. First, genetic deletion in the transgenic mice was accomplished using a constitutive Cre driver, whereas the use of a tamoxifen-responsive CreER^T2 system would enable a true lineage trace of tdTomato⁺ EGCs after Men1 deletion. The use of an inducible CreER^T2 system would allow for precise temporal control of when Men1 is deleted and could therefore answer whether differentiated GFAP⁺ and SOX10⁺ cells in juvenile and adult mice can undergo similar transcriptional reprogramming. Additional application of single cell profiling methods would enable a better dissection of the reprogrammed cell populations and support the identification of potential biomarkers. Second, our use of the Kif3a-null mice demonstrated that SHH signal transduction via primary cilia is a required step in Men1-driven tumor formation, however this model did not provide context into the role of canonical and non-canonical HH signaling pathways. This knowledge is relevant given that patient and mouse tumoroids showed disparate responses to pharmacologic inhibitors of SMO and GLI1/2.

Our in vitro studies using mouse EGCs were limited to cultures derived from the proximal SI, since these cells are better characterized and can be isolated using established protocols [28, 29]. EGCs in the pancreas are more enigmatic but are known to be comprised of Schwann cells that both encapsulate and co-localize with the endocrine islet [62, 63]. Last, our work using patient-derived tumoroids were restricted to PanNETs and a single ileal NET since these tumors were more readily available. Our attempt to generate tumoroid cultures from a lone DNET was unsuccessful, mirroring the high failure rate that others have reported for intestinal NETs [6]. Future work must address the lack of preclinical humanized dpNET models to better translate these findings to the clinic. For instance, the application of patient-derived xenograft models for dpNETs may represent a promising approach to propagate slow-growing tumors and test the delivery of GLI1/2 inhibitors in future preclinical studies.

Conclusions

Our observations implicate neural crest-derived glial cells as potential neuroendocrine cell precursors that are susceptible to transformation through increased SHH signaling upon loss of Menin. By establishing that NETs can arise from a glial cell that responds to SHH, our findings prompt a reassessment of HH inhibitors for potential therapeutic intervention in dpNETs that may include targets beyond inhibiting SMO.

Supplementary Information

Supplementary Material 1.^{(1.2MB, pdf)}

Supplementary Material 2.^{(16.6MB, docx)}

Acknowledgements

The authors wish to acknowledge the support of the Chao Family Comprehensive Cancer Center Experimental Tissue Shared Resource, supported by the National Cancer Institute of the National Institutes of Health under award number P30CA062203. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. We thank the Arizona Genetics Core for their contributions to the sequencing analysis and the University of Arizona Tissue Repository for providing biospecimens used in this study. Finally, we are grateful to Dr. Tobias Else, Dr. Michelle Kim, the Biorepository and Pathology Core laboratory team at Icahn School of Medicine at Mount Sinai, and biorepository participants for their contributions to this research.

Abbreviations

4-OHT: 4-hydroxytamoxifen
ΔMen1: Men1-deleted
BCC: Basal cell carcinoma
BrdU: Bromodeoxyuridine
CNV: Copy number variant
CreER^T2: Tamoxifen-inducible Cre recombinase
DNET: Duodenal neuroendocrine tumor
dpNET: Duodenopancreatic neuroendocrine tumors
EdU: 5-ethynyl-2’-deoxyuridine
EGC: Enteric glial cell
FFPE: Formalin-fixed and paraffin-embedded
GEP-NET: Gastroenteropancreatic neuroendocrine tumors
GFAP: Glial fibrillary acidic protein
GLI: Gli-Kruppel family member
HH: Hedgehog signaling pathway
IHC: Immunohistochemical
IF: Immunofluorescent
IL-NET: Ileal NET
J-NET: Jejunal NET
MEN1/Men1: Multiple Endocrine Neoplasia I
NPC: Neural progenitor cell
PDAC: Pancreatic ductal adenocarcinoma
PanNET: Pancreatic neuroendocrine tumor
PTCH1: Patched-1
SAG: SMO agonist
SI-NET: Small intestinal neuroendocrine tumor
SOX10: SRY-box transcription factor 10
SMO: Smoothened
SHH: Sonic hedgehog
TCGA: The Cancer Genome Atlas

Authors’ contributions

SD and JLM performed the original project conceptualization, methodology, investigation, data visualization and analysis, supervision, and writing of the manuscript. ABT performed cell culture, IHC, and assisted with gene expression analysis. AZ assisted with organoid culture, performed western blot analysis, and contributed to the editing of the original manuscript. IN performed bioinformatic analysis on the single cell RNA sequencing dataset, generated the associated figures, and assisted with image analysis. RGG performed validation studies, western blot, and assisted with image analysis. JWM performed validation studies on primary glial cultures, assisted with organoid culture, and contributed to image analysis. RAS performed animal genotyping and assisted with the animal studies. RIA provided human biospecimens and interpreted clinicopathologic findings. ET provided support for the Caris clinico-genomic database analyses. TWS performed bioinformatic analysis of the bulk RNA sequencing dataset and generated the associated figures. All authors read and approved the final manuscript.

Funding

This work was funded by grants from the National Institutes of Health grant 5K01DK136969 (SD) and University of Arizona Cancer Center Support Grant P30 CA023074.

Data availability

The datasets analyzed during the current study are available from the corresponding author on reasonable request. Data associated with The Cancer Genome Atlas (TCGA) analyses are publicly available through the National Cancer Institute’s Genome Data Commons Portal. Data associated with the Caris clinico-genomic database can be accessed by submitting a letter of intent to Caris Life Sciences.

Declarations

Ethics approval and consent to participate

Deidentified formalin-fixed and paraffin-embedded (FFPE) sections of surgically resected human tumors were provided by the University of Arizona Tissue Acquisition Shared Resource (TACMASR) under Human Institutional Review Board (IRB) protocol #1808861863, University of Michigan Endocrine Oncology Repository under IRB protocol #HUM00115310, and the Biorepository and Pathology Core at Icahn School of Medicine at Mount Sinai, under IRB #STUDY-12-00145. Fresh de-identified surgically resected tumor tissues were provided by the Chao Family Comprehensive Cancer Center Experimental Tissue Shared Resource (protocol # STUDY00000833) and the University of Arizona TACMASR and the University of Arizona TARGHETS Biorepository UA under IRB protocol #1909985869. Informed patient consent was provided prior to tissue collection. Mouse experiments were performed in compliance with the University of Arizona Institutional Animal Care and Use Committee (IACUC) guidelines under protocol #18–440. The study complied with the maximal tumor burden outlined by the IACUC, which permits up to 10% of the animal body weight.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1.Abboud Y, Shah A, Sutariya R, Shah VP, Al-Khazraji A, Gaglio PJ, Hajifathalian K. Gastroenteropancreatic neuroendocrine tumor incidence by sex and age in the US. JAMA Oncol. 2025;11:345–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Xu Z, Wang L, Dai S, Chen M, Li F, Sun J, Luo F. Epidemiologic trends of and factors associated with overall survival for patients with gastroenteropancreatic neuroendocrine tumors in the united States. JAMA Netw Open. 2021;4:e2124750. [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Patte C, Pommier RM, Ferrari A, Chung FF, Ouzounova M, Moullé P, Richaud M, Khoueiry R, Hervieu M, Breusa S, Allio M, Rama N, Gérard L, Hervieu V, Poncet V, Fenouil T, Cahais V, Sertier AS, Boland A, Bacq-Daian D, Ducarouge B, Marie JC, Deleuze JF, Viari A, Scoazec JY, Roche C, Mehlen P, Walter T, Gibert B. Comprehensive molecular portrait reveals genetic diversity and distinct molecular subtypes of small intestinal neuroendocrine tumors. Nat Commun. 2025;16:2197. [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Ye Z, Zhou Y, Hu Y, Li Q, Xu Z, Lou X, Zhang W, Zhu D, Xie C, Zhou Q, Gao J, Zhou H, Yang D, Qin Y, Xu X, Chen J, Ji S, Wang MW, Yu X. Single-cell sequencing reveals the heterogeneity of pancreatic neuroendocrine tumors under genomic instability and histological grading. iScience. 2024;27:110836. [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Somech E, Halder D, Spitzer A, Barbolin C, Tyler M, Halperin R, Biton M, Tirosh A, Tirosh I. Subtypes and proliferation patterns of small intestine neuroendocrine tumors revealed by single-cell RNA sequencing. eLife. 2025;13:RP101153. [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Kawasaki K, Toshimitsu K, Matano M, Fujita M, Fujii M, Togasaki K, Ebisudani T, Shimokawa M, Takano A, Takahashi S, Ohta Y, Nanki K, Igarashi R, Ishimaru K, Ishida H, Sukawa Y, Sugimoto S, Saito Y, Maejima K, Sasagawa S, Lee H, Kim HG, Ha K, Hamamoto J, Fukunaga K, Maekawa A, Tanabe M, Ishihara S, Hamamoto Y, Yasuda H, Sekine S, Kudo A, Kitagawa Y, Kanai T, Nakagawa H, Sato T. An organoid biobank of neuroendocrine neoplasms enables Genotype-Phenotype mapping. Cell. 2020;183:1420–e143521. [DOI] [PubMed] [Google Scholar]
7.Mäkinen N, Zhou M, Zhang Z, Kasai Y, Perez E, Kim GE, Thirlwell C, Nakakura E, Meyerson M. Whole genome sequencing reveals the independent clonal origin of multifocal ileal neuroendocrine tumors. Genome Med. 2022;14:82. [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Banck MS, Kanwar R, Kulkarni AA, Boora GK, Metge F, Kipp BR, Zhang L, Thorland EC, Minn KT, Tentu R, Eckloff BW, Wieben ED, Wu Y, Cunningham JM, Nagorney DM, Gilbert JA, Ames MM, Beutler AS. The genomic landscape of small intestine neuroendocrine tumors. J Clin Invest. 2013;123:2502–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Karpathakis A, Dibra H, Pipinikas C, Feber A, Morris T, Francis J, Oukrif D, Mandair D, Pericleous M, Mohmaduvesh M, Serra S, Ogunbiyi O, Novelli M, Luong T, Asa SL, Kulke M, Toumpanakis C, Meyer T, Caplin M, Meyerson M, Beck S, Thirlwell C. Prognostic impact of novel molecular subtypes of small intestinal neuroendocrine tumor. Clin Cancer Res. 2016;22:250–8. [DOI] [PubMed] [Google Scholar]
10.Francis JM, Kiezun A, Ramos AH, Serra S, Pedamallu CS, Qian ZR, Banck MS, Kanwar R, Kulkarni AA, Karpathakis A, Manzo V, Contractor T, Philips J, Nickerson E, Pho N, Hooshmand SM, Brais LK, Lawrence MS, Pugh T, McKenna A, Sivachenko A, Cibulskis K, Carter SL, Ojesina AI, Freeman S, Jones RT, Voet D, Saksena G, Auclair D, Onofrio R, Shefler E, Sougnez C, Grimsby J, Green L, Lennon N, Meyer T, Caplin M, Chung DC, Beutler AS, Ogino S, Thirlwell C, Shivdasani R, Asa SL, Harris CR, Getz G, Kulke M, Meyerson M. Somatic mutation of CDKN1B in small intestine neuroendocrine tumors. Nat Genet. 2013;45:1483–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Jiao Y, Shi C, Edil BH, de Wilde RF, Klimstra DS, Maitra A, et al. DAXX/ATRX, MEN1, and mTOR pathway genes are frequently altered in pancreatic neuroendocrine tumors. Science. 2011;331:1199–203. [DOI] [PMC free article] [PubMed]
12.Scarpa A, Chang DK, Nones K, Corbo V, Patch AM, Bailey P, Lawlor RT, Johns AL, Miller DK, Mafficini A, Rusev B, Scardoni M, Antonello D, Barbi S, Sikora KO, Cingarlini S, Vicentini C, McKay S, Quinn MCJ, Bruxner TJC, Christ AN, Harliwong I, Idrisoglu S, McLean S, Nourse C, Nourbakhsh E, Wilson PJ, Anderson MJ, Fink JL, Newell F, Waddell N, Holmes O, Kazakoff SH, Leonard C, Wood S, Xu Q, Nagaraj SH, Amato E, Dalai I, Bersani S, Cataldo I, Dei Tos AP, Capelli P, Davì MV, Landoni L, Malpaga A, Miotto M, Whitehall VLJ, Leggett BA, Harris JL, Harris J, Jones MD, Humphris J, Chantrill LA, Chin V, Nagrial AM, Pajic M, Scarlett CJ, Pinho A, Rooman I, Toon C, Wu J, Pinese M, Cowley M, Barbour A, Mawson A, Humphrey ES, Colvin EK, Chou A, Lovell JA, Jamieson NB, Duthie F, Gingras MC, Fisher WE, Dagg RA, Lau LMS, Lee M, Pickett HA, Reddel R, Samra JS, Kench JG, Merrett ND, Epari K, Nguyen NQ, Zeps N, Falconi M, Simbolo M, Butturini G, Van Buren G, Partelli S, Fassan FM, Khanna AJ, Gill DA, Wheeler RA, Gibbs EA, Musgrove C, Bassi G, Tortora P, Pederzoli. Nature. 2017;550:548. J. V. Pearson, N. Waddell, A. V. Biankin, S. M. Grimmond, Corrigendum: Whole-genome landscape of pancreatic neuroendocrine tumours. [DOI] [PubMed] [Google Scholar]
13.Agarwal SK. Exploring the tumors of multiple endocrine neoplasia type 1 in mouse models for basic and preclinical studies. Int J Endocr Oncol. 2014;1(2):153–61. 10.2217/ije.14.16. [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Sedlack AJH, Saleh-Anaraki K, Kumar S, Ear PH, Lines KE, Roper N, Pacak K, Bergsland E, Quelle DE, Howe JR, Pommier Y. Del Rivero, preclinical models of neuroendocrine neoplasia. Cancers (Basel). 2022;14:5646. [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Wong C, Tang LH, Davidson C, Vosburgh E, Chen W, Foran DJ, Notterman DA, Levine AJ, Xu EY. Two well-differentiated pancreatic neuroendocrine tumor mouse models. Cell Death Differ. 2020;27:269–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Fuentes ME, Lu X, Flores NM, Hausmann S, Mazur PK. Combined deletion of MEN1, ATRX, and PTEN triggers development of high-grade pancreatic neuroendocrine tumors in mice. Sci Rep. 2024;14:8510. [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Xu EY, Vosburgh E, Wong C, Tang LH, Notterman DA. Genetic analysis of the cooperative tumorigenic effects of targeted deletions of tumor suppressors Rb1, Trp53, Men1, and Pten in neuroendocrine tumors in mice. Oncotarget. 2020;11:2718–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Veniaminova NA, Hayes MM, Varney JM, Merchant JL. Conditional deletion of Menin results in antral G cell hyperplasia and hypergastrinemia. Am J Physiol Gastrointest Liver Physiol. 2012;303:G752–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Sundaresan S, Kang AJ, Hayes MM, Choi EK, Merchant JL. Deletion of Men1 and somatostatin induces hypergastrinemia and gastric carcinoids. Gut. 2017;66:1012–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Sundaresan S, Meininger CA, Kang AJ, Photenhauer AL, Hayes MM, Sahoo N, Grembecka J, Cierpicki T, Ding L, Giordano TJ, Else T, Madrigal DJ, Low MJ, Campbell F, Baker AM, Xu H, Wright NA, Merchant JL. Gastrin induces nuclear export and proteasome degradation of Menin in enteric glial cells. Gastroenterology. 2017;153(6):1555–e156715. [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Marinoni I, Avanthay S, Alcala N. Novel concepts of cell-of origin in neuroendocrine neoplasms. Virchows Archiv: Int J Pathol. 2025. 10.1007/s00428-025-04311-2. Advance online publication. [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Furlan A, Dyachuk V, Kastriti ME, Calvo-Enrique L, Abdo H, Hadjab S, Chontorotzea T, Akkuratova N, Usoskin D, Kamenev D, Petersen J, Sunadome K, Memic F, Marklund U, Fried K, Topilko P, Lallemend F, Kharchenko PV, Ernfors P, Adameyko I. Multipotent peripheral glial cells generate neuroendocrine cells of the adrenal medulla. Science. 2017;357:eaal3753. [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Santambrogio A, Kemkem Y, Willis TL, Berger I, Kastriti ME, Faure L, Russell JP, Lodge EJ, Yianni V, Kövér B, Oakey RJ, Altieri B, Bornstein SR, Steenblock C, Adameyko I. Andoniadou, SOX2 + sustentacular cells are stem cells of the postnatal adrenal medulla. Nat Commun. 2025;16:16. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Duan S, Sawyer TW, Sontz RA, Wieland BA, Diaz AF, Merchant JL. GFAP-directed inactivation of Men1 exploits glial cell plasticity in favor of neuroendocrine reprogramming. Cell Mol Gastroenterol Hepatol. 2022;14:1025–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Majer AD, Hua X, Katona BW. Menin in cancer. Genes. 2024;15:1231. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Gurung B, Feng Z, Iwamoto DV, Thiel A, Jin G, Fan CM, Ng JM, Curran T, Hua X. Menin epigenetically represses Hedgehog signaling in MEN1 tumor syndrome. Cancer Res. 2013;73:2650–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Gurung B, Feng Z, Hua X. Menin directly represses Gli1 expression independent of canonical Hedgehog signaling. Mol Cancer Res. 2013;11:1215–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Caldwell ML, Cook CA, Mariant CL, Touvron M, Odle J, Blikslager AT, Ziegler AL, Van Landeghem L. Protocol to culture enteric glial cells from the submucosal and myenteric plexi of neonatal and juvenile pig colons. STAR Protoc. 2024;5:103057. [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Wang Z, Ocadiz-Ruiz R, Sundaresan S, Ding L, Hayes M, Sahoo N, Xu H, Merchant JL. Isolation of enteric glial cells from the submucosa and lamina propria of the adult mouse. J Vis Exp. 2018;138:57629. [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Duan S, Sawyer TW, Witten BL, Song H, Else T, Merchant JL. Spatial profiling reveals tissue-specific neuro-immune interactions in gastroenteropancreatic neuroendocrine tumors. J Pathol. 2024;262:362–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Dutton R, Yamada T, Turnley A, Bartlett PF, Murphy M. Sonic Hedgehog promotes neuronal differentiation of murine spinal cord precursors and collaborates with neurotrophin 3 to induce Islet-1. J Neurosci. 1999;19:2601–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Fu M, Lui VC, Sham MH, Pachnis V, Tam PK. Sonic Hedgehog regulates the proliferation, differentiation, and migration of enteric neural crest cells in the gut. J Cell Biol. 2004;166:673–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Nishi Y, Zhang X, Jeong J, Peterson KA, Vedenko A, Bulyk ML, Hide WA, McMahon AP. A direct fate exclusion mechanism by Sonic hedgehog-regulated transcriptional repressors. Development. 2015;142:3286–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Kondo T, A Johnson S, C Yoder M, Romand R, Hashino E. Sonic Hedgehog and retinoic acid synergistically promote sensory fate specification from bone marrow-derived pluripotent stem cells. Proc Natl Acad Sci U S A. 2005;102:4789–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Yang H, Liu C, Fan H, Chen B, Huang D, Zhang L, Zhang Q, An J, Zhao J, Wang Y, Hao D. Sonic Hedgehog effectively improves Oct4-mediated reprogramming of astrocytes into neural stem cells. Mol Ther. 2019;27:1467–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Briscoe J. Making a grade: Sonic Hedgehog signaling and the control of neural cell fate. EMBO J. 2009;28:457–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Gurung B, Hua X, Runske M, Bennett B, LiVolsi V, Roses R, Fraker DA, Metz DC. PTCH1 staining of pancreatic neuroendocrine tumor (PanNET) samples from patients with and without multiple endocrine neoplasia (MEN-1) syndrome reveals a potential therapeutic target. Cancer Biol Ther. 2015;16:219–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Shida T, Furuya M, Nikaido T, Hasegawa M, Koda K, Oda K, Miyazaki M, Kishimoto T, Nakatani Y, Ishikura H. Sonic Hedgehog-Gli1 signaling pathway might become an effective therapeutic target in Gastrointestinal neuroendocrine carcinomas. Cancer Biol Ther. 2006;5:1530–8. [DOI] [PubMed] [Google Scholar]
39.Fendrich V, Waldmann J, Esni F, Ramaswamy A, Mullendore M, Buchholz M, Maitra A. Feldmann, snail and Sonic Hedgehog activation in neuroendocrine tumors of the ileum. Endocr Relat Cancer. 2007;14:865–74. [DOI] [PubMed] [Google Scholar]
40.Anlauf M, Perren A, Meyer CL, Schmid S, Saremaslani P, Kruse ML, Weihe E, Komminoth P, Heitz PU, Klöppel G. Precursor lesions in patients with multiple endocrine neoplasia type 1-associated duodenal gastrinomas. Gastroenterology. 2005;128:1187–98. [DOI] [PubMed] [Google Scholar]
41.Graillon T, Romanet P, Camilla C, Gélin C, Appay R, Roche C, Lagarde A, Mougel G, Farah K, Le Bras M, Engelhardt J, Kalamarides M, Peyre M, Amelot A, Emery E, Magro E, Cebula H, Aboukais R, Bauters C, Jouanneau E, Goudet PA. Cohort study of CNS tumors in multiple endocrine neoplasia type 1. Clin Cancer Research: Official J Am Association Cancer Res. 2024;30(13):2835–45. [DOI] [PubMed] [Google Scholar]
42.Cuevas-Ocampo AK, Bollen AW, Goode B, Pajtler KW, Chavez L, Sharma T, Dai SC, McDermott M, Perry A, Korshunov A, Solomon DA. Genetic confirmation that ependymoma can arise as part of multiple endocrine neoplasia type 1 (MEN1) syndrome. Acta Neuropathol. 2017;133(4):661–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Einarsson HB, Frederiksen AL, Pedersen IS, Ettrup MS, Wirenfeldt M, Boldt H, Nguyen N, Andersen MS, Bjarkam CR, Poulsen F. R. PDP type brain tumor in association with multiple endocrine neoplasia type 1. Heliyon. 2024;10(6):e27418. [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Hilgendorf KI, Myers BR, Reiter JF. Emerging mechanistic Understanding of cilia function in cellular signaling. Nat Rev Mol Cell Biol. 2024;25:555–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Rudin CM, Hann CL, Laterra J, Yauch RL, Callahan CA, Fu L, Holcomb T, Stinson J, Gould SE, Coleman B, LoRusso PM, Von Hoff DD, de Sauvage FJ, Low JA. Treatment of Medulloblastoma with Hedgehog pathway inhibitor GDC-0449. N Engl J Med. 2009;361:1173–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Von Hoff DD, LoRusso PM, Rudin CM, Reddy JC, Yauch RL, Tibes R, et al. Low Inhibition of the hedgehog pathway in advanced basal-cell carcinoma. N Engl J Med. 2009;361:1164–72. [DOI] [PubMed]
47.Spetz J, Langen B, Rudqvist N, Parris TZ, Helou K, Nilsson O, Forssell-Aronsson E. Hedgehog inhibitor sonidegib potentiates 177Lu-octreotate therapy of GOT1 human small intestine neuroendocrine tumors in nude mice. BMC Cancer. 2017;17:528. [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Carballo GB, Honorato JR, de Lopes GPF. L. S. E. Spohr, A highlight on Sonic Hedgehog pathway. Cell Commun Signal. 2018;16:11. [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Schimmack S, Kneller S, Dadabaeva N, Bergmann F, Taylor A, Hackert T, Werner J, Strobel O. Epithelial to stromal re-distribution of primary cilia during pancreatic carcinogenesis. PLoS ONE. 2016;11:e0164231. [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Fendrich V, Oh E, Bang S, Karikari C, Ottenhof N, Bisht S, Lauth M, Brossart P, Katsanis N, Maitra A, Feldmann G. Ectopic overexpression of Sonic Hedgehog (Shh) induces stromal expansion and metaplasia in the adult murine pancreas. Neoplasia. 2011;13:923–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Portal C, Rompolas P, Lwigale P, Iomini C. Primary cilia deficiency in neural crest cells models anterior segment dysgenesis in mouse. eLife. 2019;8:e52423. [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Lai FP, Li Z, Zhou T, Leung AOW, Lau ST, Lui KN, Wong WY, Sham PC, Hui CC. Ngan, ciliary protein Kif7 regulates Gli and Ezh2 for initiating the neuronal differentiation of enteric neural crest cells during development. Sci Adv. 2021;7:eabf7472. [DOI] [PMC free article] [PubMed] [Google Scholar]
53.McCleary-Wheeler AL, Carr RM, Palmer SR, Smyrk TC, Allred JB, Almada LL, Tolosa EJ, Lamberti MJ, Marks DL, Borad MJ, Molina JR, Qi Y, Lingle WL, Grothey A, Pitot HC, Jatoi A, Northfelt DW, Bryce AH, McWilliams RR, Okuno SH, Haluska P, Kim GP, Colon-Otero G, Lowe VJ, Callstrom MR, Ma WW, Bekaii-Saab T, Hung MC, Erlichman C. Fernandez-Zapico, phase 1 trial of vismodegib and erlotinib combination in metastatic pancreatic cancer. Pancreatology. 2020;20:101–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Kim EJ, Sahai V, Abel EV, Griffith KA, Greenson JK, Takebe N, Khan GN, Blau JL, Craig R, Balis UG, Zalupski MM, Simeone DM. Pilot clinical trial of Hedgehog pathway inhibitor GDC-0449 (vismodegib) in combination with gemcitabine in patients with metastatic pancreatic adenocarcinoma. Clin Cancer Res. 2014;20:5937–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
55.De Jesus-Acosta A, Sugar EA, O’Dwyer PJ, Ramanathan RK, Von Hoff DD, Rasheed Z, Zheng L, Begum A, Anders R, Maitra A, McAllister F, Rajeshkumar NV, Yabuuchi S, de Wilde RF, Batukbhai B, Sahin I, Laheru DA. Phase 2 study of vismodegib, a Hedgehog inhibitor, combined with gemcitabine and nab-paclitaxel in patients with untreated metastatic pancreatic adenocarcinoma. Br J Cancer. 2020;122:498–505. [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Quatannens D, Verhoeven Y, Van Dam P, Lardon F, Prenen H, Roeyen G, Peeters M, Smits ELJ, Van Audenaerde J. Targeting Hedgehog signaling in pancreatic ductal adenocarcinoma. Pharmacol Ther. 2022;236:108107. [DOI] [PubMed] [Google Scholar]
57.Nolan-Stevaux O, Lau J, Truitt ML, Chu GC, Hebrok M, Fernandez-Zapico ME, et al. Gli1 is regulated through Smoothened-independent mechanisms in neoplastic pancreatic ducts and mediates PDAC cell survival and transformation. Genes Dev. 2009;23:24–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Yang H, Feng GD, Olivera C, Jiao XY, Vitale A, Gong J, You SW. Sonic Hedgehog released from scratch-injured astrocytes is a key signal necessary but not sufficient for the astrocyte de-differentiation. Stem Cell Res. 2012;9:156–66. [DOI] [PubMed] [Google Scholar]
59.Tang J, Chen L, Wang Z, Huang G, Hu X. Sox2 mediates crosstalk between Sonic Hedgehog and the Wnt/β-catenin signaling pathway to promote proliferation of pituitary adenoma cells. Oncol Lett. 2019;18:81–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Farheen S, Ahmed SP, Mariyath PMM, Kausar T, Hoda MF, Arif SH, Nayeem SM, Ali A, Chosdol K, Shahi MH. Differential role of Pax6 and its interaction with Shh-Gli1-IDH2 axis in regulation of glioma growth and chemoresistance. J Biochem Mol Toxicol. 2023;37:e23241. [DOI] [PubMed] [Google Scholar]
61.Ooki A, Dinalankara W, Marchionni L, Tsay JJ, Goparaju C, Maleki Z, Rom WN, Pass HI, Hoque MO. Epigenetically regulated PAX6 drives cancer cells toward a stem-like state via GLI-SOX2 signaling axis in lung adenocarcinoma. Oncogene. 2018;37:5967–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
62.Rangel-Sosa MM, Mann F, Chauvet S. Pancreatic Schwann cell reprogramming supports cancer-associated neuronal remodeling. Glia. 2024;72:1840–61. [DOI] [PubMed] [Google Scholar]
63.Olaniru OE, Kadolsky U, Kannambath S, Vaikkinen H, Fung K, Dhami P, Persaud SJ. Single-cell transcriptomic and Spatial landscapes of the developing human pancreas. Cell Metab. 2023;35:184–e1995. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1.^{(1.2MB, pdf)}

Supplementary Material 2.^{(16.6MB, docx)}

Data Availability Statement

[CR1] 1.Abboud Y, Shah A, Sutariya R, Shah VP, Al-Khazraji A, Gaglio PJ, Hajifathalian K. Gastroenteropancreatic neuroendocrine tumor incidence by sex and age in the US. JAMA Oncol. 2025;11:345–9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR2] 2.Xu Z, Wang L, Dai S, Chen M, Li F, Sun J, Luo F. Epidemiologic trends of and factors associated with overall survival for patients with gastroenteropancreatic neuroendocrine tumors in the united States. JAMA Netw Open. 2021;4:e2124750. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] 3.Patte C, Pommier RM, Ferrari A, Chung FF, Ouzounova M, Moullé P, Richaud M, Khoueiry R, Hervieu M, Breusa S, Allio M, Rama N, Gérard L, Hervieu V, Poncet V, Fenouil T, Cahais V, Sertier AS, Boland A, Bacq-Daian D, Ducarouge B, Marie JC, Deleuze JF, Viari A, Scoazec JY, Roche C, Mehlen P, Walter T, Gibert B. Comprehensive molecular portrait reveals genetic diversity and distinct molecular subtypes of small intestinal neuroendocrine tumors. Nat Commun. 2025;16:2197. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR4] 4.Ye Z, Zhou Y, Hu Y, Li Q, Xu Z, Lou X, Zhang W, Zhu D, Xie C, Zhou Q, Gao J, Zhou H, Yang D, Qin Y, Xu X, Chen J, Ji S, Wang MW, Yu X. Single-cell sequencing reveals the heterogeneity of pancreatic neuroendocrine tumors under genomic instability and histological grading. iScience. 2024;27:110836. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR5] 5.Somech E, Halder D, Spitzer A, Barbolin C, Tyler M, Halperin R, Biton M, Tirosh A, Tirosh I. Subtypes and proliferation patterns of small intestine neuroendocrine tumors revealed by single-cell RNA sequencing. eLife. 2025;13:RP101153. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR6] 6.Kawasaki K, Toshimitsu K, Matano M, Fujita M, Fujii M, Togasaki K, Ebisudani T, Shimokawa M, Takano A, Takahashi S, Ohta Y, Nanki K, Igarashi R, Ishimaru K, Ishida H, Sukawa Y, Sugimoto S, Saito Y, Maejima K, Sasagawa S, Lee H, Kim HG, Ha K, Hamamoto J, Fukunaga K, Maekawa A, Tanabe M, Ishihara S, Hamamoto Y, Yasuda H, Sekine S, Kudo A, Kitagawa Y, Kanai T, Nakagawa H, Sato T. An organoid biobank of neuroendocrine neoplasms enables Genotype-Phenotype mapping. Cell. 2020;183:1420–e143521. [DOI] [PubMed] [Google Scholar]

[CR7] 7.Mäkinen N, Zhou M, Zhang Z, Kasai Y, Perez E, Kim GE, Thirlwell C, Nakakura E, Meyerson M. Whole genome sequencing reveals the independent clonal origin of multifocal ileal neuroendocrine tumors. Genome Med. 2022;14:82. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR8] 8.Banck MS, Kanwar R, Kulkarni AA, Boora GK, Metge F, Kipp BR, Zhang L, Thorland EC, Minn KT, Tentu R, Eckloff BW, Wieben ED, Wu Y, Cunningham JM, Nagorney DM, Gilbert JA, Ames MM, Beutler AS. The genomic landscape of small intestine neuroendocrine tumors. J Clin Invest. 2013;123:2502–8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR9] 9.Karpathakis A, Dibra H, Pipinikas C, Feber A, Morris T, Francis J, Oukrif D, Mandair D, Pericleous M, Mohmaduvesh M, Serra S, Ogunbiyi O, Novelli M, Luong T, Asa SL, Kulke M, Toumpanakis C, Meyer T, Caplin M, Meyerson M, Beck S, Thirlwell C. Prognostic impact of novel molecular subtypes of small intestinal neuroendocrine tumor. Clin Cancer Res. 2016;22:250–8. [DOI] [PubMed] [Google Scholar]

[CR10] 10.Francis JM, Kiezun A, Ramos AH, Serra S, Pedamallu CS, Qian ZR, Banck MS, Kanwar R, Kulkarni AA, Karpathakis A, Manzo V, Contractor T, Philips J, Nickerson E, Pho N, Hooshmand SM, Brais LK, Lawrence MS, Pugh T, McKenna A, Sivachenko A, Cibulskis K, Carter SL, Ojesina AI, Freeman S, Jones RT, Voet D, Saksena G, Auclair D, Onofrio R, Shefler E, Sougnez C, Grimsby J, Green L, Lennon N, Meyer T, Caplin M, Chung DC, Beutler AS, Ogino S, Thirlwell C, Shivdasani R, Asa SL, Harris CR, Getz G, Kulke M, Meyerson M. Somatic mutation of CDKN1B in small intestine neuroendocrine tumors. Nat Genet. 2013;45:1483–6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR11] 11.Jiao Y, Shi C, Edil BH, de Wilde RF, Klimstra DS, Maitra A, et al. DAXX/ATRX, MEN1, and mTOR pathway genes are frequently altered in pancreatic neuroendocrine tumors. Science. 2011;331:1199–203. [DOI] [PMC free article] [PubMed]

[CR12] 12.Scarpa A, Chang DK, Nones K, Corbo V, Patch AM, Bailey P, Lawlor RT, Johns AL, Miller DK, Mafficini A, Rusev B, Scardoni M, Antonello D, Barbi S, Sikora KO, Cingarlini S, Vicentini C, McKay S, Quinn MCJ, Bruxner TJC, Christ AN, Harliwong I, Idrisoglu S, McLean S, Nourse C, Nourbakhsh E, Wilson PJ, Anderson MJ, Fink JL, Newell F, Waddell N, Holmes O, Kazakoff SH, Leonard C, Wood S, Xu Q, Nagaraj SH, Amato E, Dalai I, Bersani S, Cataldo I, Dei Tos AP, Capelli P, Davì MV, Landoni L, Malpaga A, Miotto M, Whitehall VLJ, Leggett BA, Harris JL, Harris J, Jones MD, Humphris J, Chantrill LA, Chin V, Nagrial AM, Pajic M, Scarlett CJ, Pinho A, Rooman I, Toon C, Wu J, Pinese M, Cowley M, Barbour A, Mawson A, Humphrey ES, Colvin EK, Chou A, Lovell JA, Jamieson NB, Duthie F, Gingras MC, Fisher WE, Dagg RA, Lau LMS, Lee M, Pickett HA, Reddel R, Samra JS, Kench JG, Merrett ND, Epari K, Nguyen NQ, Zeps N, Falconi M, Simbolo M, Butturini G, Van Buren G, Partelli S, Fassan FM, Khanna AJ, Gill DA, Wheeler RA, Gibbs EA, Musgrove C, Bassi G, Tortora P, Pederzoli. Nature. 2017;550:548. J. V. Pearson, N. Waddell, A. V. Biankin, S. M. Grimmond, Corrigendum: Whole-genome landscape of pancreatic neuroendocrine tumours. [DOI] [PubMed] [Google Scholar]

[CR13] 13.Agarwal SK. Exploring the tumors of multiple endocrine neoplasia type 1 in mouse models for basic and preclinical studies. Int J Endocr Oncol. 2014;1(2):153–61. 10.2217/ije.14.16. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR14] 14.Sedlack AJH, Saleh-Anaraki K, Kumar S, Ear PH, Lines KE, Roper N, Pacak K, Bergsland E, Quelle DE, Howe JR, Pommier Y. Del Rivero, preclinical models of neuroendocrine neoplasia. Cancers (Basel). 2022;14:5646. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR15] 15.Wong C, Tang LH, Davidson C, Vosburgh E, Chen W, Foran DJ, Notterman DA, Levine AJ, Xu EY. Two well-differentiated pancreatic neuroendocrine tumor mouse models. Cell Death Differ. 2020;27:269–83. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR16] 16.Fuentes ME, Lu X, Flores NM, Hausmann S, Mazur PK. Combined deletion of MEN1, ATRX, and PTEN triggers development of high-grade pancreatic neuroendocrine tumors in mice. Sci Rep. 2024;14:8510. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR17] 17.Xu EY, Vosburgh E, Wong C, Tang LH, Notterman DA. Genetic analysis of the cooperative tumorigenic effects of targeted deletions of tumor suppressors Rb1, Trp53, Men1, and Pten in neuroendocrine tumors in mice. Oncotarget. 2020;11:2718–39. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR18] 18.Veniaminova NA, Hayes MM, Varney JM, Merchant JL. Conditional deletion of Menin results in antral G cell hyperplasia and hypergastrinemia. Am J Physiol Gastrointest Liver Physiol. 2012;303:G752–64. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR19] 19.Sundaresan S, Kang AJ, Hayes MM, Choi EK, Merchant JL. Deletion of Men1 and somatostatin induces hypergastrinemia and gastric carcinoids. Gut. 2017;66:1012–21. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR20] 20.Sundaresan S, Meininger CA, Kang AJ, Photenhauer AL, Hayes MM, Sahoo N, Grembecka J, Cierpicki T, Ding L, Giordano TJ, Else T, Madrigal DJ, Low MJ, Campbell F, Baker AM, Xu H, Wright NA, Merchant JL. Gastrin induces nuclear export and proteasome degradation of Menin in enteric glial cells. Gastroenterology. 2017;153(6):1555–e156715. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR21] 21.Marinoni I, Avanthay S, Alcala N. Novel concepts of cell-of origin in neuroendocrine neoplasms. Virchows Archiv: Int J Pathol. 2025. 10.1007/s00428-025-04311-2. Advance online publication. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR22] 22.Furlan A, Dyachuk V, Kastriti ME, Calvo-Enrique L, Abdo H, Hadjab S, Chontorotzea T, Akkuratova N, Usoskin D, Kamenev D, Petersen J, Sunadome K, Memic F, Marklund U, Fried K, Topilko P, Lallemend F, Kharchenko PV, Ernfors P, Adameyko I. Multipotent peripheral glial cells generate neuroendocrine cells of the adrenal medulla. Science. 2017;357:eaal3753. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR23] 23.Santambrogio A, Kemkem Y, Willis TL, Berger I, Kastriti ME, Faure L, Russell JP, Lodge EJ, Yianni V, Kövér B, Oakey RJ, Altieri B, Bornstein SR, Steenblock C, Adameyko I. Andoniadou, SOX2 + sustentacular cells are stem cells of the postnatal adrenal medulla. Nat Commun. 2025;16:16. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR24] 24.Duan S, Sawyer TW, Sontz RA, Wieland BA, Diaz AF, Merchant JL. GFAP-directed inactivation of Men1 exploits glial cell plasticity in favor of neuroendocrine reprogramming. Cell Mol Gastroenterol Hepatol. 2022;14:1025–51. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR25] 25.Majer AD, Hua X, Katona BW. Menin in cancer. Genes. 2024;15:1231. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR26] 26.Gurung B, Feng Z, Iwamoto DV, Thiel A, Jin G, Fan CM, Ng JM, Curran T, Hua X. Menin epigenetically represses Hedgehog signaling in MEN1 tumor syndrome. Cancer Res. 2013;73:2650–8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR27] 27.Gurung B, Feng Z, Hua X. Menin directly represses Gli1 expression independent of canonical Hedgehog signaling. Mol Cancer Res. 2013;11:1215–22. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR28] 28.Caldwell ML, Cook CA, Mariant CL, Touvron M, Odle J, Blikslager AT, Ziegler AL, Van Landeghem L. Protocol to culture enteric glial cells from the submucosal and myenteric plexi of neonatal and juvenile pig colons. STAR Protoc. 2024;5:103057. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR29] 29.Wang Z, Ocadiz-Ruiz R, Sundaresan S, Ding L, Hayes M, Sahoo N, Xu H, Merchant JL. Isolation of enteric glial cells from the submucosa and lamina propria of the adult mouse. J Vis Exp. 2018;138:57629. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR30] 30.Duan S, Sawyer TW, Witten BL, Song H, Else T, Merchant JL. Spatial profiling reveals tissue-specific neuro-immune interactions in gastroenteropancreatic neuroendocrine tumors. J Pathol. 2024;262:362–76. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR31] 31.Dutton R, Yamada T, Turnley A, Bartlett PF, Murphy M. Sonic Hedgehog promotes neuronal differentiation of murine spinal cord precursors and collaborates with neurotrophin 3 to induce Islet-1. J Neurosci. 1999;19:2601–8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR32] 32.Fu M, Lui VC, Sham MH, Pachnis V, Tam PK. Sonic Hedgehog regulates the proliferation, differentiation, and migration of enteric neural crest cells in the gut. J Cell Biol. 2004;166:673–84. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR33] 33.Nishi Y, Zhang X, Jeong J, Peterson KA, Vedenko A, Bulyk ML, Hide WA, McMahon AP. A direct fate exclusion mechanism by Sonic hedgehog-regulated transcriptional repressors. Development. 2015;142:3286–93. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR34] 34.Kondo T, A Johnson S, C Yoder M, Romand R, Hashino E. Sonic Hedgehog and retinoic acid synergistically promote sensory fate specification from bone marrow-derived pluripotent stem cells. Proc Natl Acad Sci U S A. 2005;102:4789–94. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR35] 35.Yang H, Liu C, Fan H, Chen B, Huang D, Zhang L, Zhang Q, An J, Zhao J, Wang Y, Hao D. Sonic Hedgehog effectively improves Oct4-mediated reprogramming of astrocytes into neural stem cells. Mol Ther. 2019;27:1467–82. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR36] 36.Briscoe J. Making a grade: Sonic Hedgehog signaling and the control of neural cell fate. EMBO J. 2009;28:457–65. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR37] 37.Gurung B, Hua X, Runske M, Bennett B, LiVolsi V, Roses R, Fraker DA, Metz DC. PTCH1 staining of pancreatic neuroendocrine tumor (PanNET) samples from patients with and without multiple endocrine neoplasia (MEN-1) syndrome reveals a potential therapeutic target. Cancer Biol Ther. 2015;16:219–24. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR38] 38.Shida T, Furuya M, Nikaido T, Hasegawa M, Koda K, Oda K, Miyazaki M, Kishimoto T, Nakatani Y, Ishikura H. Sonic Hedgehog-Gli1 signaling pathway might become an effective therapeutic target in Gastrointestinal neuroendocrine carcinomas. Cancer Biol Ther. 2006;5:1530–8. [DOI] [PubMed] [Google Scholar]

[CR39] 39.Fendrich V, Waldmann J, Esni F, Ramaswamy A, Mullendore M, Buchholz M, Maitra A. Feldmann, snail and Sonic Hedgehog activation in neuroendocrine tumors of the ileum. Endocr Relat Cancer. 2007;14:865–74. [DOI] [PubMed] [Google Scholar]

[CR40] 40.Anlauf M, Perren A, Meyer CL, Schmid S, Saremaslani P, Kruse ML, Weihe E, Komminoth P, Heitz PU, Klöppel G. Precursor lesions in patients with multiple endocrine neoplasia type 1-associated duodenal gastrinomas. Gastroenterology. 2005;128:1187–98. [DOI] [PubMed] [Google Scholar]

[CR41] 41.Graillon T, Romanet P, Camilla C, Gélin C, Appay R, Roche C, Lagarde A, Mougel G, Farah K, Le Bras M, Engelhardt J, Kalamarides M, Peyre M, Amelot A, Emery E, Magro E, Cebula H, Aboukais R, Bauters C, Jouanneau E, Goudet PA. Cohort study of CNS tumors in multiple endocrine neoplasia type 1. Clin Cancer Research: Official J Am Association Cancer Res. 2024;30(13):2835–45. [DOI] [PubMed] [Google Scholar]

[CR42] 42.Cuevas-Ocampo AK, Bollen AW, Goode B, Pajtler KW, Chavez L, Sharma T, Dai SC, McDermott M, Perry A, Korshunov A, Solomon DA. Genetic confirmation that ependymoma can arise as part of multiple endocrine neoplasia type 1 (MEN1) syndrome. Acta Neuropathol. 2017;133(4):661–3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Einarsson HB, Frederiksen AL, Pedersen IS, Ettrup MS, Wirenfeldt M, Boldt H, Nguyen N, Andersen MS, Bjarkam CR, Poulsen F. R. PDP type brain tumor in association with multiple endocrine neoplasia type 1. Heliyon. 2024;10(6):e27418. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR44] 44.Hilgendorf KI, Myers BR, Reiter JF. Emerging mechanistic Understanding of cilia function in cellular signaling. Nat Rev Mol Cell Biol. 2024;25:555–73. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR45] 45.Rudin CM, Hann CL, Laterra J, Yauch RL, Callahan CA, Fu L, Holcomb T, Stinson J, Gould SE, Coleman B, LoRusso PM, Von Hoff DD, de Sauvage FJ, Low JA. Treatment of Medulloblastoma with Hedgehog pathway inhibitor GDC-0449. N Engl J Med. 2009;361:1173–8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR46] 46.Von Hoff DD, LoRusso PM, Rudin CM, Reddy JC, Yauch RL, Tibes R, et al. Low Inhibition of the hedgehog pathway in advanced basal-cell carcinoma. N Engl J Med. 2009;361:1164–72. [DOI] [PubMed]

[CR47] 47.Spetz J, Langen B, Rudqvist N, Parris TZ, Helou K, Nilsson O, Forssell-Aronsson E. Hedgehog inhibitor sonidegib potentiates 177Lu-octreotate therapy of GOT1 human small intestine neuroendocrine tumors in nude mice. BMC Cancer. 2017;17:528. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR48] 48.Carballo GB, Honorato JR, de Lopes GPF. L. S. E. Spohr, A highlight on Sonic Hedgehog pathway. Cell Commun Signal. 2018;16:11. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR49] 49.Schimmack S, Kneller S, Dadabaeva N, Bergmann F, Taylor A, Hackert T, Werner J, Strobel O. Epithelial to stromal re-distribution of primary cilia during pancreatic carcinogenesis. PLoS ONE. 2016;11:e0164231. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR50] 50.Fendrich V, Oh E, Bang S, Karikari C, Ottenhof N, Bisht S, Lauth M, Brossart P, Katsanis N, Maitra A, Feldmann G. Ectopic overexpression of Sonic Hedgehog (Shh) induces stromal expansion and metaplasia in the adult murine pancreas. Neoplasia. 2011;13:923–30. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR51] 51.Portal C, Rompolas P, Lwigale P, Iomini C. Primary cilia deficiency in neural crest cells models anterior segment dysgenesis in mouse. eLife. 2019;8:e52423. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR52] 52.Lai FP, Li Z, Zhou T, Leung AOW, Lau ST, Lui KN, Wong WY, Sham PC, Hui CC. Ngan, ciliary protein Kif7 regulates Gli and Ezh2 for initiating the neuronal differentiation of enteric neural crest cells during development. Sci Adv. 2021;7:eabf7472. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR53] 53.McCleary-Wheeler AL, Carr RM, Palmer SR, Smyrk TC, Allred JB, Almada LL, Tolosa EJ, Lamberti MJ, Marks DL, Borad MJ, Molina JR, Qi Y, Lingle WL, Grothey A, Pitot HC, Jatoi A, Northfelt DW, Bryce AH, McWilliams RR, Okuno SH, Haluska P, Kim GP, Colon-Otero G, Lowe VJ, Callstrom MR, Ma WW, Bekaii-Saab T, Hung MC, Erlichman C. Fernandez-Zapico, phase 1 trial of vismodegib and erlotinib combination in metastatic pancreatic cancer. Pancreatology. 2020;20:101–9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR54] 54.Kim EJ, Sahai V, Abel EV, Griffith KA, Greenson JK, Takebe N, Khan GN, Blau JL, Craig R, Balis UG, Zalupski MM, Simeone DM. Pilot clinical trial of Hedgehog pathway inhibitor GDC-0449 (vismodegib) in combination with gemcitabine in patients with metastatic pancreatic adenocarcinoma. Clin Cancer Res. 2014;20:5937–45. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR55] 55.De Jesus-Acosta A, Sugar EA, O’Dwyer PJ, Ramanathan RK, Von Hoff DD, Rasheed Z, Zheng L, Begum A, Anders R, Maitra A, McAllister F, Rajeshkumar NV, Yabuuchi S, de Wilde RF, Batukbhai B, Sahin I, Laheru DA. Phase 2 study of vismodegib, a Hedgehog inhibitor, combined with gemcitabine and nab-paclitaxel in patients with untreated metastatic pancreatic adenocarcinoma. Br J Cancer. 2020;122:498–505. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR56] 56.Quatannens D, Verhoeven Y, Van Dam P, Lardon F, Prenen H, Roeyen G, Peeters M, Smits ELJ, Van Audenaerde J. Targeting Hedgehog signaling in pancreatic ductal adenocarcinoma. Pharmacol Ther. 2022;236:108107. [DOI] [PubMed] [Google Scholar]

[CR57] 57.Nolan-Stevaux O, Lau J, Truitt ML, Chu GC, Hebrok M, Fernandez-Zapico ME, et al. Gli1 is regulated through Smoothened-independent mechanisms in neoplastic pancreatic ducts and mediates PDAC cell survival and transformation. Genes Dev. 2009;23:24–36. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR58] 58.Yang H, Feng GD, Olivera C, Jiao XY, Vitale A, Gong J, You SW. Sonic Hedgehog released from scratch-injured astrocytes is a key signal necessary but not sufficient for the astrocyte de-differentiation. Stem Cell Res. 2012;9:156–66. [DOI] [PubMed] [Google Scholar]

[CR59] 59.Tang J, Chen L, Wang Z, Huang G, Hu X. Sox2 mediates crosstalk between Sonic Hedgehog and the Wnt/β-catenin signaling pathway to promote proliferation of pituitary adenoma cells. Oncol Lett. 2019;18:81–6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR60] 60.Farheen S, Ahmed SP, Mariyath PMM, Kausar T, Hoda MF, Arif SH, Nayeem SM, Ali A, Chosdol K, Shahi MH. Differential role of Pax6 and its interaction with Shh-Gli1-IDH2 axis in regulation of glioma growth and chemoresistance. J Biochem Mol Toxicol. 2023;37:e23241. [DOI] [PubMed] [Google Scholar]

[CR61] 61.Ooki A, Dinalankara W, Marchionni L, Tsay JJ, Goparaju C, Maleki Z, Rom WN, Pass HI, Hoque MO. Epigenetically regulated PAX6 drives cancer cells toward a stem-like state via GLI-SOX2 signaling axis in lung adenocarcinoma. Oncogene. 2018;37:5967–81. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR62] 62.Rangel-Sosa MM, Mann F, Chauvet S. Pancreatic Schwann cell reprogramming supports cancer-associated neuronal remodeling. Glia. 2024;72:1840–61. [DOI] [PubMed] [Google Scholar]

[CR63] 63.Olaniru OE, Kadolsky U, Kannambath S, Vaikkinen H, Fung K, Dhami P, Persaud SJ. Single-cell transcriptomic and Spatial landscapes of the developing human pancreas. Cell Metab. 2023;35:184–e1995. [DOI] [PubMed] [Google Scholar]

PERMALINK

Hedgehog signaling drives glial cell plasticity and oncogenic reprogramming in gastroenteropancreatic neuroendocrine neoplasms

Suzann Duan

AnneLeigh B Twer

Ateh Zinkeng

Ikrame Naciri

Juliette W Moore

Rebeca G Gentry

Ricky A Sontz

Reed I Ayabe

Edgar Tapia

Travis W Sawyer

Juanita L Merchant

Abstract

Background

Methods

Results

Conclusions

Supplementary Information

Background

Methods

Animals

The cancer genome atlas (TCGA) patient cohort analysis

Caris Clinico-genomic database analysis

MEN1 Mutation Analysis

Immunohistochemical/Immunofluorescence (IHC/IF) staining

Immunocytochemistry (ICC/IF) staining

Cell culture

Tumor organoid culture

In vitro drug studies

EdU fluorescence assay

Bromodeoxyuridine (BrdU) incorporation assay

TdTomato fluorescence growth assay

Serum hormone Enzyme-linked immunosorbent assays (ELISA)

Quantitative polymerase chain reaction (qPCR)

Western blot analysis

Bulk RNA sequencing

Statistical analysis

Results

Human dpNETs overexpress Hedgehog signaling components

Fig. 1.

Table 1.

SHH regulates PanNET growth via GLI1/2 signaling

Fig. 2.

Table 2.

Loss of Men1 in GFAP+ and SOX10+ glial cells drives the development of HH-expressing GEP-NETs

Fig. 3.

Fig. 4.

Activation of GLI1/2 regulates the growth of ΔMen1 GEP-NET tumoroids

Fig. 5.

Fig. 6.

Reciprocal signaling by Menin and SHH regulate enteric glial cell identity

Fig. 7.

SHH signaling in GFAP+ glia is required for GEP-NET development

Fig. 8.

Discussion

Conclusions

Supplementary Information

Acknowledgements

Abbreviations

Authors’ contributions

Funding

Data availability

Declarations

Ethics approval and consent to participate

Consent for publication

Competing interests

Footnotes

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases

Loss of Men1 in GFAP⁺ and SOX10⁺ glial cells drives the development of HH-expressing GEP-NETs

SHH signaling in GFAP⁺ glia is required for GEP-NET development