Loss-of-function of ALDH3B2 transdifferentiates human pancreatic duct cells into beta-like cells

Loss-of-function of ALDH3B2 transdifferentiates human pancreatic duct cells into beta-like cells | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Loss-of-function of ALDH3B2 transdifferentiates human pancreatic duct cells into beta-like cells Peng Yi, Jian Li, Yu-chi Lee, Noelle Morrow, Jennifer Hollister-Lock, and 11 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-2222452/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Replenishment of pancreatic beta cells is a key to the cure for diabetes. Beta cells regeneration is achieved predominantly by self-replication especially in rodents, but it was also shown that pancreatic duct cells can transdifferentiate into beta cells. How pancreatic duct cells were transdifferentiated and whether we could manipulate the transdifferentiation to replenish beta cell mass is not well understood. Using a genome-wide CRISPR screen, we discovered that loss-of-function of ALDH3B2 is sufficient to transdifferentiate human pancreatic duct cells into functional beta-like cells. The transdifferentiated cells have significant increase in beta cell marker genes expression, secrete insulin in response to glucose, and reduce blood glucose when transplanted into diabetic mice. Our study identifies a novel gene that we could potentially target in human pancreatic duct cells to replenish beta cell mass for diabetes therapy. Biological sciences/Genetics Biological sciences/Cell biology Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 INTRODUCTION Diabetes, no matter the cause or type, is a disease of pancreatic beta cell deficiency 1 . Current treatments for diabetes do not provide the same degree of exquisite glycemic control as would a sufficient number of functional beta cells, and thus do not prevent the debilitating correlates of long-term diabetes. To cure diabetes, one has to find a way to stop the recurrent autoimmune attack on beta cells (type 1 diabetes) or resolve persistent peripheral insulin resistance (type 2 diabetes), but restoring a sufficient functional beta cell mass is critical to a cure for both types of diabetes. Beta cells can be replenished by transplanting human cadaveric islets or islet-like cells derived from human embryonic stem cells (hESC) or induced pluripotent stem cells (iPSCs) 2 – 5 . In most cases, transplanted islet cells are HLA-mismatched with the recipient, and immunosuppression is required to prevent graft rejection, creating other complications such as lack of immune defense against pathogens or tumor formation. Alternatively, it may be possible to promote endogenous pancreatic beta cell regeneration. Promoting the regeneration of a patient's own beta cells could be a safer strategy for beta cell mass replenishment; for type 1 diabetes autoimmunity would still need to be tamed but there would be no need for immunosuppression for allo-graft rejection. Beta cell regeneration (reviewed in 6 ) can be achieved by self-duplication 7 or transdifferentiation from other pancreatic cell types such as duct cells 8 , alpha cells 9 , 10 or acinar cells 11 . Pancreatic beta cell replication is the dominant mechanism of beta cell regeneration in adult rodents 7 . However in human, the adult beta cell replication rate is extremely low and it was postulated that human beta cells regeneration is achieved mainly by transdifferentiation from pancreatic duct cells 8 . It has been suggested that pancreatic duct cells may serve as a pool of progenitors for both the islet and acinar tissues after birth and into adulthood 12 – 17 . Although human duct-to-beta cell transdifferentiation has been evidenced by the existence of insulin-expressing cells in the pancreatic duct epithelium, it is a very rare phenomenon that many doubt would be relevant for sufficient beta cell regeneration. However, if its underlying mechanism can be understood, one could potentially manipulate duct-to-beta cell transdifferentiation to a high enough efficiency to replenish functional beta cell mass for the treatment of diabetes. Here we employed a genome-wide CRISPR screen to dissect the mechanism of human duct-to-beta cell transdifferentiation and to identify new therapeutic targets for beta cell mass restoration. RESULTS A genome-wide CRISPR screen identifies ALDH3B2 as a regulator of human duct-to-beta cell transdifferentiation Forward genetic screening, the genome-wide CRISPR screening in particular, is a powerful approach to discover novel genes, signaling pathways and the underlying mechanisms of a complex biological phenomenon. Here, we developed a genome-wide CRISPR screening strategy to search for genes that regulate the transdifferentiation of human pancreatic duct cells into beta cells. To ensure efficient gene editing and sufficient cell numbers for the genome-wide screen with sufficient coverage, we chose to use an immortalized pancreatic cell line, PANC-1, a human pancreatic carcinoma cell line of ductal origin that maintains many of the differentiated characteristics of normal mammalian pancreatic ductal epithelium 18 , 19 . Our strategy is to use PANC-1 cells solely as a CRISPR screen and target discovery tool, and all the findings from the PANC-1 genetic screen will later be validated and characterized in primary human pancreatic duct cells. We reasoned that insulin expression or insulin promoter activation would be the most direct and the simplest readout for cell transdifferentiation into pancreatic beta-like cells. Therefore, we introduced a reporter construct, Rat insulin promoter 3.1 ( R IP)- E GFP- P 2A- Blasticidin-S deaminase ( B SD) (referred to as REPB reporter), into the PANC-1 cells via lentiviral transduction to create a REPB-PANC-1 reporter cell line (Fig. 1a). Rat insulin promoter (RIP) is known to be active in human beta cells and RIP 3.1 is a modified rat insulin promoter that is believed to have higher efficiency and beta cell specificity 20 , which could increase the sensitivity of our CRISPR screen. The P2A peptide ensures the co-expression of the EGFP and BSD reporter genes, the EGFP reporter allows to visually and quantitatively monitor insulin promoter activation, and the expression of the BSD reporter gene confers resistance to blasticidin treatment, making it easy to enrich or select insulin promoter activated cells. To validate the REPB reporter construct, we also transduced the NIT-1 mouse beta cell line with the REPB reporter (REPB-NIT-1 cell line). As shown in Extended Data Fig. 1a and 1b, the EGFP expression can only be observed in the REPB-NIT-1 cells, but not at all in REPB-PANC-1 cells. The REPB-PANC-1 cells, but not the REPB-NIT-1 cells, were sensitive to blasticidin treatment (data not shown). To execute the CRISPR screen, as illustrated in Fig. 1a, we transduced the REPB-PANC-1 cells with a human lentiviral genome-wide CRISPR knockout library (GeCKO v2) 21 that comprises approximately 120,000 guide RNAs (gRNAs) targeting a total of 19,050 genes. We used a low multiplicity of infection (MOI) ~ 0.3 to ensure that most of the cells carry only one mutation. Briefly, approximately 10 8 lentiviral library transduced REPB-PANC-1 cells were treated with low dose of blasticidin (10ug/ml) for 7 days, and then the blasticidin resistant cells were subjected to FACS sorting based on their EGFP intensity (Extended Data Fig. 2). Using next generation sequencing (NGS) and bioinformatic analysis, the gRNA profile of the highest EGFP-expressing cells (EGFP high , blasticidin-resistant) was generated and compared to that of the cells without blasticidin selection and FACS sorting (Extended Table 1). We identified candidates whose sgRNAs were cooperatively and significantly positively selected in the EGFP high , blasticidin-resistant REPB-PANC-1 cells vs control cells using MAGeCK 22 , 23 and MAGeCKFlute 24 at the gene level. As shown in Fig. 1b from MAGeCKFlute where genes that were at least two-fold up are in red and those at least two-fold down are in blue, there is substantial positive selection. In Fig. 1c the half-volcano plot of statistics from MAGeCK-mle illustrates multiple highly significant positively selected hits with robust fold changes, such as ALDH3B2 whose fold change > 10 5 (Fig. 1c). We next considered whether the screening hits were enriched for functional classifications involved in the cell fate change. Top pathways from pathway enrichment analysis of KEGG pathways using MAGeCKFlute are shown in Fig. 1d. Glycolysis/gluconeogenesis, inositol phosphate metabolism, beta-alanine metabolism, and starch and sucrose metabolism represent the highest NES (Normalized enrichment scores) value, which indicates that members of those gene sets tend to participate in the cell transdifferentiation. To understand the protein-protein associations including physical and functional interactions between INS, ACLY, and other GFP-positive screening hits from MAGeCKFlute analysis, the STRING pathway analysis was performed (Fig. 1e). Many CRISPR screening hits, such as ITPKA, GPI, HK2, ALDH3B2, and ALDH3A1 showing independent connection with INS and ACLY, which recognized as an essential gene during cell transdifferentiation 25 and metabolic reprogramming 26 . The screening hits were significantly enriched for shared protein networks, providing additional confidence in the sensitivity to identify interrelated complexes (Fig. 1f). Many hits mapped to the functional categories of metabolic pathway and glycolysis/gluconeogenesis, all of them containing both previously unknown regulators of beta cell transdifferentiation and those with known roles in cell transdifferentiation or metabolic reprogramming. Get together, we then picked the most enriched gRNA of top 8 candidate gene from the list and generated individual mutant PANC-1 cell lines using the corresponding gRNAs identified in our screen. We used quantitative PCR (qPCR) to analyze the expression of endocrine marker genes including insulin (INS), glucagon (GCG) and somatostatin (SST), as well as pancreatic duct cell marker gene keratin 19 (KRT19 or CK19). Several mutant PANC-1 cell lines showed differential expression of the examined marker genes (Fig. 1c, 1d and Extended Data Fig. 3a and 3b). In particular, PANC-1 cells transduced with a gRNA targeting ALDH3B2 showed the highest INS expression and the lowest KRT19 levels compared to non-targeting control (NTC) gRNA transduced PANC-1 cells. ALDH3B2, also known as ALDH8, is one of 19 members of the human aldehyde dehydrogenase (ALDH) superfamily that converts various types of aldehydes to carboxylic acids 27 . It is well documented that ALDH genes are important regulators of stem cells and cell fate determination 28 . Another close member in the ALDH family, ALDH1A3, was recently shown to contribute to pancreatic beta cell failure and de-differentiation in type 2 diabetes 29 . We reasoned that ALDH3B2, as an enzyme, could potentially be an easier therapeutic target for small molecules targeting. Therefore, we prioritized ALDH3B2 for further in-depth validation and characterization. Loss-of-function of ALDH3B2 trans-differentiates PANC-1 cells into pancreatic beta-like cells We generated an ALDH3B2 mutant PANC-1 cell line by lentiviral transduction of SpCas9 and ALDH3B2 gRNA into PANC-1 cells. Genomic sequencing of the targeted region in ALDH3B2 mut PANC-1 cells revealed that more than 75% of the sequenced ALDH3B2 alleles carried indel mutations (Extended Data Fig. 4). Western blot confirmed that ALDH3B2 protein level in the ALDH3B2 mut PANC-1 cells was reduced by ~ 40% compared to the non-targeting-control (NTC) gRNA lentivirus transduced PANC-1 cells (Extended Data Fig. 5a and 5b). To ensure that loss-of-function of ALDH3B2 induces bona fide cell transdifferentiation and did not only just activate the insulin promoter, we conducted a series of qPCR experiments to examine additional genes characteristic for pancreatic beta cells. We found that the expression of pancreatic endocrine hormones, insulin (INS) and somatostatin (SST), but not glucagon (GCG) or pancreatic polypeptide (PP), were significantly increased in the ALDH3B2 mutant PANC-1 cells (Fig. 2a). In addition, the expression of several of key beta cell transcription factors, including PDX1, MAFA, NGN3 and PAX6 , were also significantly upregulated in the ALDH3B2 mutant PANC-1 cells (Fig. 2b). We examined additional genes that are critical for beta cell function and found that GLUT1 (SLC2A1), GLUT2 (SLC2A2), GLUCOKINASE (GCK) , subunits of the ATP-sensitive potassium (K-ATP) (KCNJ11 and ABCC8 ), CPE and IA2 were all significantly upregulated in the ALDH3B2 mutant PANC-1 cells (Fig. 2c). ALDH3B2 mutant PANC-1 cells had slightly decreased expression of pancreatic duct markers KRT19 and CA2 but not of HNF1B or Sox9 (Fig. 2d). Immunofluorescence imaging showed that clusters of ALDH3B2 mutant PANC-1 cells expressed human Insulin (INS) and C-peptide (CPEP), whereas neither insulin and C-peptide were detected in control PANC-1 cells (Fig. 2e). The insulin content of the ALDH3B2 mutant PANC-1 was significantly increased compared to control cells (Fig. 2f). In addition, using electron microscopy (EM), we found that many of the ALDH3B2 mutant PANC-1 cells had insulin granules (vesicles with halo, arrows in Fig. 2g), while no insulin granules were detectable in control -PANC-1 cells (Fig. 2g). Intriguingly, the ALDH3B2 mutant PANC-1 cells had intensive endoplasmic reticulum (ER) network (Fig. 2g), a characteristic found in pancreatic beta cells but not pancreatic duct cells. Collectively, these studies indicate that loss-of-function mutations in ALDH3B2 in PANC-1 cells not only trigger insulin promoter activation but also precipitate a significant cell fate transformation, shifting from a pancreatic ductal phenotype to a beta-like profile. To evaluate whether these trans-differentiated pancreatic beta-like cells were functional, we performed in vitro g lucose s timulated i nsulin s ecretion (GSIS) assay comparing control NTC-PANC-1 cells and transdifferentiated ALDH3B2 mutant PANC-1 cells, and we found that the ALDH3B2 mutant PANC-1 cells can secrete significantly more insulin at baseline (2.8 mM glucose) compared to the control NTC-PANC-1 cells, and also mildly respond to higher glucose (16.7 mM glucose) (Extended Data Fig. 6a). We then examined whether the ALDH3B2 mutant PANC-1 cells is also functional in vivo in diabetic mouse model. We used streptozotocin (STZ) to induce diabetes in NSG mice, where majority of the beta cells were killed by STZ injection (Extended Data Fig. 7), and then transplanted the ALDH3B2 mutant or control PANC-1 cells (NTC) subcutaneously into these diabetic mice (Fig. 2h, upper panel). Mice transplanted with the ALDH3B2 mutant PANC-1 cells showed significantly decreased daily random blood glucose (Fig. 2h, lower panel) and improved glucose tolerance (Fig. 2i) compared to mice transplanted with NTC-PANC-1 cells. Notably, when the ALDH3B2 mutant PANC-1 graft was removed at the end of the study, blood glucose increased to the same level as in the control mice, confirming that the blood-glucose-lowering was indeed caused by the transplanted ALDH3B2 mutant PANC-1 cells (Fig. 2h, lower panel). Human insulin serum levels were also significantly higher in mice transplanted with ALDH3B2 mutant PANC-1 cells (Fig. 2j). Immunofluorescent imaging showed that transplanted ALDH3B2 mutant PANC-1 cells co-expressed PDX1, Insulin, NKX6.1 and C-peptide (Fig. 2k, 2l, Extended Data Fig. 8a and 8b). We observed that a few cells co-express somatostatin (SST) and Insulin, but no cell expresses glucagon (GCG), or the exocrine cell marker gene amylase (AMY) (Extended Data Fig. 8c and 8d). Of all the transplanted ALDH3B2 mutant PANC-1 cells, ~ 20% were PDX1 + (Fig. 2m) and ~ 8% were INS + or NKX6.1 + (Fig. 2n and 2o). Almost all the INS + cells were also NKX6.1 + , suggesting that trans-differentiated beta-like cells adopted a true beta cell phenotype. It should also be noted that only ~ 45% of the PDX1 + cells co-expressed insulin (Fig. 2p), and we speculate that the PDX1 + /INS - cells may represent pancreatic progenitor-like cells that have yet committed to beta cell fate. We employed an inducible shRNA system to ensure that the transdifferentiation of PANC-1 cells into beta-like cells by ALDH3B2 CRISPR knockout was indeed due to the loss-of-function of ALDH3B2 and not caused by off-target effects of the ALDH3B2 gRNA. We generated PANC-1 cell lines carrying a Tet-On inducible ALDH3B2 shRNA or a scrambled control shRNA (Fig. 3a, left panel). The ALDH3B2 shRNA PANC-1 cells with Doxycycline treatment showed significantly reduced ALDH3B2 mRNA expression (Fig. 3a right panel) and protein level (Extended Data Fig. 5c and 5d) after doxycycline (dox) treatment. Similar to the ALDH3B2 CRISPR mutant PANC-1 cells, knock-down of ALDH3B2 by shRNA also trans-differentiated PANC-1 cells into beta-like cells. A series of qPCR experiments showed that the expression of key beta cell transcription factors including PDX1 , MAFA, NGN3, NEUROD and PAX6 (Fig. 3b), endocrine hormone insulin ( INS ) and somatostatin ( SST ) (Fig. 3C), and beta cell function related genes ( SLC2A2, GCK, KCNJ11 and ABCC8 ) were significantly increased (Fig. 3d), whereas the expression of pancreatic duct cell marker genes ( KRT19, CA2 and SOX9 ) were reduced (Fig. 3e). Human insulin could also be detected in shALDH3B2 PANC-1 cells (+ Dox) but not in shControl PANC-1 cells or in shALDH3B2 PANC-1 cells (+ Dox) by immunofluorescence (Fig. 3f). These analyses confirmed that loss-of-function of ALDH3B2 by CRISPR targeting or shRNA silencing allowed PANC-1 cells to transdifferentiate and adopt a beta-like cell fate. Loss-of-function of ALDH3B2 transdifferentiates human primary pancreatic duct cells into beta-like cells. Next, we tested whether loss-of-function of ALDH3B2 was also able to transdifferentiate h uman p rimary p ancreatic d uct (HPPD) cells into beta-like cells. HPPD cells were isolated and affinity-purified from human donor islet-depleted pancreatic acinar tissue from Integrated Islet Distribution Program (IIDP) 30 . qPCR analyses confirmed lack of insulin expression (Fig. 4a) and high expression of the pancreatic duct markers KRT19 (Fig. 4b) in the purified HPPD cells compared to primary human islets. Interestingly, we observed that ALDH3B2 expression levels are markedly lower in human islets compared to pancreatic duct cells (Fig. 4c). This differential expression pattern aligns with our results where the mutation of ALDH3B2 in human pancreatic duct cells promotes their transdifferentiation into beta-like cells. These findings suggest that the reduction of ALDH3B2 could be a critical step in the cellular reprogramming process leading to a beta-cell phenotype. Purified HPPD cells were transduced with lentiviruses carrying SpCas9 and either ALDH3B2 gRNA (ALDH3B2 mut -HPPD) or a non-targeting control gRNA (NTC-HPPD). qPCR analysis demonstrated that ALDH3B2 mut -HPPD cells had significantly higher expression of key beta cell transcription factors ( PDX1 and MAFA ) (Fig. 4d), endocrine hormone insulin ( INS ) and somatostatin ( SST ) (Fig. 4e) and beta cell function-related genes ( GCK, SLC2A1, SLC2A2, KCNJ11 and CPE ) (Fig. 4f). The expression of several pancreatic duct cell marker genes was either unchanged ( KRT19 and HNF1B ) or slightly reduced ( SOX9 ) (Fig. 4g). Using immunofluorescent imaging, we found that a fraction of HPPD-ALDH3B2 mut cells expressed Insulin while still retaining CK19 expression, a possible signature of newly transdifferentiated beta cells from pancreatic duct cells, whereas no insulin expression could be detected in HPPD-NTC cells (Fig. 4h). Furthermore, we also examined whether the ALDH3B2 mut -HPPD cells have insulin granules using electron microscopy (EM). Although not as many insulin granules as in primary human beta cells, the ALDH3B2 mut -HPPD cells do have significant amount of mature insulin granules, while no insulin granules were detectable in the control NTC-HPPD cells (Fig. 4i). We performed in vitro GSIS assay to evaluate the function of the ALDH3B2 mut -HPPD cells, and found that compared to control NTC-HPPD cells, the ALDH3B2 mut -HPPD cells can secrete significantly more insulin and mildly respond to high glucose (Extended Data Fig. 6b). Although the level of insulin secretion from the ALDH3B2 mut -HPPD cells is still much lower than human islets (Extended Data Fig. 6b), we suspected that it is due to relatively low efficacy of the transdifferentiation and immaturity of the trandifferentiated beta-like cells, especially in the in vitro experiment setting. We then transplanted ALDH3B2 mut -HPPD or NTC-HPPD cells under the kidney capsule of STZ induced diabetic NSG mice, and monitored their blood glucose over time (Fig. 5a, upper panel). Mice transplanted with HPPD-ALDH3B2 mut cells had significantly lower blood glucose compared to NTC-HPPD cells transplanted mice (Fig. 5a, lower panel). When the ALDH3B2 mutant HPPD grafts were removed at 56 days post-transplantation, blood glucose increased to similar level as in the control HPPD transplanted mice, suggesting that the blood-glucose-lowering effect was indeed conferred by the transplanted ALDH3B2 mutant HPPD cells (Fig. 5a, lower panel). Importantly, transplanted ALDH3B2 mut -HPPD cells secreted human insulin in response to glucose challenge. We detected a significantly increase in human serum insulin 5 minutes after glucose injection (both at 1 week and 3 weeks post-transplantation). No such response was observed in mice transplanted with NTC-HPPD cells or in non-transplanted control NSG mice (Fig. 5b and 5c). Immunofluorescent analysis revealed that Insulin + cells, C-Peptide + cells, PDX1 + cells, NKX6.1 + cells were only observed in mice transplanted with ALDH3B2 mut -HPPD cells but not with NTC-HPPD cells (Fig. 5d, 5e, Extended Data Fig. 9a-d). Almost all of the transplanted ALDH3B2 mut -HPPD cells expressed pancreatic duct marker genes CK19 and SOX9 but not glucagon (GCG), Somatostatin (SST) or Amylase (AMY) (Extended Data Fig. 9e and 9f). Approximately 40% of the transplanted cells were successfully transduced with the NTC or ALDH3B2 gRNA lentivirus (shown by quantification of the percentage of Cas9 (Flag-tagged) + /CK19 + cells, Fig. 5d and 5f), and among all the gRNA lentivirus infected pancreatic duct cells, ~ 15% of ALDH3B2 mut -HPPD cells expressed insulin (Fig. 5g). Interestingly, we found that the majority of the pancreatic duct cells infected with ALDH3B2 gRNA lentivirus co-expressed PDX1 and CK19 (Fig. 5e and 5h), and approximately 12% of the PDX1 + cells co-expressed insulin (Fig. 5i). Co-expression of PDX1 and CK19 is a signature of pancreatic progenitor cells 31 , and we postulate that ALDH3B2 loss-of-function may cause the de-differentiation of mature duct cells into pancreatic progenitor-like cells, a portion of which then subsequently differentiate into beta-like cells. Loss of ALDH3B2 function in pancreatic duct cells causes epigenetic changes We found that ALDH3B2 loss-of-function allowed pancreatic duct cells to adopt a beta-like cell profile, and we next asked if transdifferentiation was associated with epigenetic changes. To this end, we analyzed DNA methylation in the human insulin gene region by bisulfite conversion assay. DNA methylation was significantly reduced in ALDH3B2 mutant PANC-1 cells compared to control NTC-PANC-1 cells at the + 63, +127 and + 139 positions of the human insulin locus, which are three well-characterized DNA methylation sites in the insulin locus 32 (Fig. 6a and 6b). For DNA methylation analysis of primary human pancreatic duct cells, we included primary human islets for comparison. Again, ALDH3B2 mutation significantly reduced DNA methylation at the same three sites in the insulin gene locus (Fig. 6c and 6d). ALDH3B2 mutation did not reduce the DNA methylation to the level observed in primary islets. This might be due to the fact that only a fraction (8–15%) of pancreatic duct cells were transdifferentiated into beta-like cells with ALDH3B2 mutation. Overall, DNA methylation analyses suggest that loss-of-function of ALDH3B2 caused epigenetic changes in the pancreatic duct cells to induce a stable cell fate change into pancreatic beta-like cells. ALDH3B2 loss-of-function in human pancreatic duct cells induces heterogeneous beta-like cell populations with overlapping endocrine and duct cell identity To investigate the characteristics of transdifferentiated beta-like cells in more detail, we performed 3’ gene expression single cell RNA sequencing of control or ALDH3B2 mutant HPPD cells and identified 13 unique cell cluster (Fig. 7a). Whereas the control NTC-HPPD condition only shows a few insulin positive cells in cluster 10, which could represent rare spontaneous transdifferentiated beta-like cells from duct cells, ALDH3B2 mutant HPPD cells develop insulin-expressing cells in various cell clusters and most prominent in cluster 5, 10, 11 and 12 (Fig. 7b-d). The relative proportion of beta-like-cell-containing cluster 5, 10, 11 and 12 is largely increased in the ALDH3B2 mutant HPPD cells compared to control NTD-HPPD cells (Extended Data Fig. 10a). The percentage of insulin-expressing cells in ALDH3B2 mutant HPPD cells is about 18.1% but only 0.6% in NTC HPPD cells, and about 93% of all cells in both conditions still retain the expression of duct cell marker KRT19 (Extended Data Fig. 10b). The majority of insulin-positive beta-like cells also show significantly higher expression of other beta cell marker genes such as CHGA and TTR but also duct cell identity marker genes KRT17, 19, and 23 (Fig. 7e). Differential gene expression analysis of insulin high-expressing cells (relative intensity > 1) compared to insulin low-expressing cells (relative intensity < 1) within the ALDH3B2 mutant HPPD cell condition showed significant upregulation of key beta cell marker genes such as CHGA, IAPP, SCGN, SCG3 and SCG5 (Fig. 7f). Of note, g ene s et e nrichment a nalysis (GSEA) confirmed the upregulation of key beta cell-specific gene sets related to peptide hormone metabolism, regulation of insulin secretion, and insulin processing (Fig. 7g). Re-analysis of insulin high-expressing cells in the ALDH3B2 mutant HPPD cells also identified heterogeneous cell populations with high beta cell identity (high INS/CHGA/IAPP co-expression), and polyhormonal cells (co-expression of INS/GCG/PPY), and endocrine progenitor-like cells (co-expression of INS and PAX6) (Extended Data Fig. 11). To further characterize the cell identity of insulin-expressing cells we compared the average gene expression of several marker genes specific for beta cells and other endocrine cells, endocrine progenitor cells, and duct cells 33 within all 13 cell clusters (Fig. 7h and 7i). In cluster 10 of the control condition, we detected insulin expression in about 40% of all cells, and the cells in this cluster show a low ductal cell-specific expression profile but high expression of endocrine progenitor cell marker such as PAX6 and INSM1 in 60–80% of all cells, indicating that the majority of cells in cluster 10 may represent endocrine progenitor-like cells. Interestingly, ALDH3B2 loss-of-function in cluster 10 shifts cell identity towards beta-like cells, shown by downregulation of almost all endocrine progenitor marker genes and sustained expression of INS and CHGA. Cluster 11 and 12 show altered duct cell identity compared to cluster 0 to 9 even in the control condition and seem to have a higher potential of beta-like cell transdifferentiation following loss-of-function of ALDH3B2. In comparison, cells in cluster 5 demonstrates strong duct cell identity but admit high potential for beta-like cell transdifferentiation as well. To better understand why cells in cluster 5 and 12 have a higher chance to transdifferentiate into beta-like cells we performed trajectory inference analysis to identify the potential starting point of transdifferentiation (from low to high insulin expression, Extended Data Fig. 10c). Beta-like cells in cluster 5 may originated from cluster 2 and cluster 5 itself. Re-analysis of cluster 5 revealed several insulin low-expressing cell cluster as possible starting points. However, insulin positive cells in cluster 5 show significant upregulation of genes involved in energetic processes such as oxidative phosphorylation, aerobic respiration, and ATP synthesis may representing important prerequisites for beta-like cell transdifferentiation (Extended Data Fig. 10d and 10e). Cluster 3 may represent the originating cluster for beta-like cells in cluster 12. Differential gene expression analysis comparing cluster 2 and 3 (high potential for beta-like cell transdifferentiation) to cluster 4 (low potential of transdifferentiation) reveals that upregulation of genes important for translation and oxidative phosphorylation (cluster 2), enrichment of small GTPases RAC1/RHO (cluster 3), and elevated glycolysis (cluster 2 and 3) may favor beta-like cell transdifferentiation mediated by loss of function of ALDH3B2 (Extended Data Fig. 10f-k). In summary, ALDH3B2 loss-of-function may drive transdifferentation of pancreatic duct cells partially through duct cell-derived endocrine progenitor-like stage, and then into beta-like cells that still keep partial duct cell identity. Elevation of energy metabolism such as oxidative phosphorylation and glycolysis may be a key step for pancreatic duct cells to transdifferentiate into beta-like cells. Dynamic change in ALDH3B2 mutant human primary pancreatic duct cells transdifferentiation is revealed by RNA Velocity and PAGA Trajectory Analysis. We next analyzed RNA velocity of our single cell data using scVelo 34 to investigate a possible transition among duct- and β-cell subclusters. scVelo identified 10 unique cell clusters, which are shown in a UMAP plot with streamlined velocities in Fig. 8a. The plot demonstrated a branching pattern emerging from cluster 2 towards other subclusters, although the specific paths were challenging to enumerate. So we next applied PAGA graph abstraction, which has been benchmarked as a top performing method for trajectory inference (Fig. 8b). It provides a graph-like map of the data topology with weighted edges corresponding to the connectivity between two clusters. We could see that the INS-expressing cells in cluster 5 were developed from cells in cluster 2, and then further developed into INS-expressing cells in cluster 9. The trajectory of INS-negative cells in the cluster also exhibited clear directionality, such as clusters (2_4), (2_1_6), and (2_8_0, 7, or 3). The transcriptional dynamic model enabled the recovery of latent time associated with cellular processes, representing an internal clock for cells undergoing differentiation based on transcriptional dynamics. The velocity latent time highlighted a consistent developmental order, with all clusters progressing from cluster 2 to subsequent subclusters (Fig. 8c). Using velocity length to characterize the speed of transition or differentiation, we observed that clusters 2 and 1 exhibited significantly higher lengths, indicative of robust splicing activity and overall velocity confidence across ductal and β-cell subpopulations. (Fig. 8d). Furthermore, we evaluated the velocity of various representative genes associated with β- and duct-cells, revealing that the expression levels and velocity did not consistently align. Notably, INS expression peaked in the transitional clusters (2_5_9), while RNA velocity was predominantly higher across most clusters, excluding clusters 6, 7, and parts of 9. Selectively elevated INS expression was observed in β-cells, whereas a modest increase in RNA velocity was noted in the ductal subclusters. Subsequent analysis of the β-cell gene TRPM3 demonstrated a concordance between its expression pattern and RNA velocity in cluster 9. Similarly, examination of the duct cell-specific gene KRT19 revealed heightened expression in cluster 1 along with high RNA velocity, contrasting with the observation that the expression pattern of the duct cell marker CFTR did not align with RNA velocity in cluster 4 (Fig. 8e). Collectively, these findings lend support to the hypothesis that cells within cluster 5 may undergo a transition towards a more β-cell-like phenotype in response to ALDH3B2 loss-of-function. Examining the top 40 driver genes within the relevant lineages (clusters 2_5_9), we identified clusters of genes exhibiting specific temporal abundance in distinct cell types. Heatmaps illustrated the primary occurrences of spliced counts for the top-ranked dynamic genes in clusters 2, 5, and 9, organized by the latent time of cells (Fig. 8f). Notably, genes such as ITGB1 35 , NEDD9 36 , HIF1A 37 , and SERPINE1 38 , known for their involvement in transdifferentiation processes linked to TGF-beta signaling, were prominently featured. Furthermore, our analysis revealed a significant upsurge in the expression levels of genes like ATP2A3, CACNA2D1, C2CD4A, and SCGN during the later phase of latent time within a brief duration. This observation suggests a pivotal role for Ca 2+ signaling in the transdifferentiation of duct cells into functional beta cells. Intriguingly, we also noted a dynamic reduction in ALDH1A3 expression during this transdifferentiation process. Additionally, glycolysis-related genes such as LDHA and HIF1A were implicated in the transdifferentiation process, further underscoring the complexity and multifaceted nature of the cellular transformations taking place. DISCUSSION In this study, we have described the first unbiased and genome-wide CRISPR screen in search for genes that regulate the transdifferentiation of human pancreatic duct cells into insulin-producing beta-like cells. We show that loss-of-function of a single gene, ALDH3B2, in human pancreatic duct cells is sufficient to drive them towards a beta-like cell fate. Although the pancreatic duct-to-beta cell transdifferentiation in human has been observed, as evidenced by the existence of INS + /CK19 + cells in pancreatic ductal epithelium, it is still a relatively rare event with the percentage of INS + pancreatic duct cells estimated at ~ 1% 39 . Here we show that disruption of ALDH3B2 can drive transdifferentiation of primary human pancreatic duct cells into beta-like cells with an efficiency of ~ 15%. This significant result carries potential for the development of a therapeutic intervention that could promote pancreatic duct-to-beta cell transdifferentiation for human beta cell mass replenishment. We demonstrated that the trans-differentiated human pancreatic beta-like cells are functional, responsive to glucose challenge in vivo , and able to significantly lower blood glucose in diabetic animal models. Notably, neither ALDH3B2 mut PANC-1 cells nor HPPD cells were able to lower blood glucose to euglycemic levels in our studies, and this could be due to two reasons: (1) The transplanted beta-like cells number was not enough due to experimental limitation, or (2) the glucose sensitivity or set-point of transdifferentiated beta-like cells may be different from that of true pancreatic beta cells. In a separate experiment of transplantation of ALDH3B2 mut -HPPD cells into euglycemic NSG mice, we performed glucose tolerance (GTT) measurements. The peak glucose level in this experiment was approximately 250mg/dL, but transplanted ALDH3B2 mut -HPPD cells were still able to improve the glucose tolerance (Extended Data Fig. 12a) and secrete human insulin in response to glucose challenge (Extended Data Fig. 12b and 7c). Based on these data, it seems unlikely that transdifferentiated beta-like cells had a higher glucose set-point than primary pancreatic beta cells. We may be able to lower the blood glucose further in diabetic mice if a larger number of transdifferentiated cells were transplanted. To this end, we will need to improve lentiviral transduction efficiency and perhaps identify more effective gRNA sequences that target ALDH3B2 gene. At first glance in comparison with primary human islets, the ALDH3B2 mutation transdifferentiated beta-like cells seem to have only modest expression of key beta cell markers (Fig. 4d-g), ability of lowering blood glucose and secrete human insulin in diabetic mice (Fig. 5a-c) and changes in DNA methylation status on insulin promoter (Fig. 6c and 6d). However, all these measurements and characterization were done on a mixed cell population with only less than 15% of the cells being transdifferentiated beta-like cells. In theory, the actual changes in each transdifferentiated beta-like cell should be more dramatic than what our experimental data showed. Our findings of immunofluorescent staining for insulin and key beta cell transcription factors and insulin granules with abundant ER ultrastructurally in the transdifferentiated beta-like cells strongly support the conclusion that loss-of-function of ALDH3B2 does induce bona fide cell transdifferentiation from human pancreatic duct cells into functional beta-like cells. It is interesting that the ALDH family is often considered as a stem cell or progenitor marker. High aldehyde dehydrogenase activity (using Aldefluor assay 40 , 41 ) has been widely used to identify adult stem cells or progenitor cells in various tissues/organs, including hematopoietic stem cells (HSC) 41 , neuronal progenitor cells (NPC) 42 , 43 or potential pancreatic progenitor cells 44 . However, the Aldefluor assay cannot distinguish between the activity of different aldehyde dehydrogenases, so different ALDH members may be marking stem/progenitor cells in different tissues/organs. The function of ALDH3B2 is not yet well understood. One rodent study reported that ALDH3B2 localizes to lipid droplets in cells and catalyzes the conversion of long-chain fatty aldehydes into long-chain fatty acids 45 . How this function would impact pancreatic duct cell transdifferentiation into beta-like cells is unclear and warrants further studies. Notwithstanding, the discovery of ALDH3B2 as a regulator of pancreatic duct-to-beta cell transdifferentiation provides a novel therapeutic target for pancreatic beta cell mass restoration. Endogenous pancreatic duct cells could potentially be targeted by gene-editing to mutate ALDH3B2 and induce transdifferentiation. Alternatively, ALDH3B2 enzymatic activity may be targeted by small molecules inhibitors to achieve the effects similar to those we observed using genetic disruption of the ALDH3B2 gene. The ALDH3B2 enzymatic assay established by Kitamura and colleagues 45 using long chain fatty aldehyde as substrate could be a useful chemical screen platform for potential new drug discovery. In support of this approach, the broad Aldehyde dehydrogenase (ALDH) inhibitors diethylaminobenzaldehyde (DEAB) and Disulfiram (DSF) have been shown to promote beta cell differentiation in zebrafish and in PANC-1 cells 46 , where it is possible that DEAB and DSF elicit their effect through inhibiting ALDH3B2. It will be of great importance to find out which aldehydes are unique substrates for ALDH3B2 and search for ALDH3B2-specific inhibitors, since mutation of the close member ALDH3A1 does not have the same effect (Fig. 1c and 1d) and another close member, ALDH1A3, was shown to be involved in the de-differentiation of pancreatic beta cells in Type 2 diabetes patients 29 . Given the interest in understanding the mechanisms driving beta cell transdifferentiation induced by ALDH3B2 loss-of-function, our primary focus has been on the progression of the INS + cluster. By further examining the subcluster of INS + cells that emerge when incorporating ALDH3B2 mutant datasets, we can visually depict the pathways in a coherent manner (refer to Fig. 8b and 8e). This analysis sheds light on a specific set of connections along the 2_5_9 differentiation axis, highlighting the transdifferentiation from pancreatic duct cells to beta cells. This dynamic model enhances our ability to interpret velocity and identify driver genes more effectively (see Fig. 8f) by exploring correlations between gene kinetics and expression patterns across cells. Notably, we observe that the key potential driver genes are associated with TGF-beta signaling 47 , 48 , glycolysis 25 , and Ca 2+ signaling pathways 49 . These findings are consistent with previous data from other cell types and may serve as valuable insights for future investigations focused on pancreatic duct-to-beta cell transdifferentiation. Our study identifying ALDH3B2 as a regulator of pancreatic duct-to-beta cell transdifferentiation was conducted using a human pancreatic duct cell line and primary human pancreatic duct cells. Given the difficulties of translating rodent studies into human in the beta cell regeneration research field, the findings presented here in human cells are directly relevant and have clear potential for the development of human diabetes therapeutics. MATERIALS AND METHODS Mice NSG (NOD.Cg-Prkdc scid Il2rg tm1Wjl /SzJ) mice were purchased from the Jackson Laboratory (Bar Harbor, ME). Animals were housed in pathogen-free facilities at the Joslin Diabetes Center and all experimental procedures were approved and performed in accordance with institutional guidelines and regulations. REPB reporter construction and REPB PANC-1 cell generation The REPB reporter lentivirus vector was constructed by assembling Rat insulin promoter, RIP3.1 promoter 20 , EGFP and Blasticidin-S deaminase (BSD). The EGFP and BSD genes are fused together with P2A peptide. The REPB reporter lentivirus was used to infect PANC-1 cells (ATCC #CRL-1469), and the infected PANC-1 cells are then single-cell sorted by FACS. PANC-1 clones with confirmed REPB reporter genome incorporation were used in the genome-wide CRISPR screen. CRISPR GeCKO library screen The human GeCKO-v2 (Genome-Scale CRISPR Knock-Out) lentiviral pooled library was obtained from Addgene (Addgene, # 1000000048) and was prepared as previously described 50 . 100 million REPB PANC-1 cells were infected with human GeCKO CRISPR lentiviral library at MOI of 0.3, and then subsequently selected with puromycin (2.5 µg/ml) at day 3 post lentiviral infection. After cells recover from puromycin selection, 20 million cells were collected as baseline control (CON-1 and CON-2), the rest of the cells were further selected with blasticidin (10 µg/ml) for 7 days. The blasticidin-resistant cells were allowed to grow back to full confluence and then subjected to FACS sorting on their EGFP intensity. Cell population with the highest EGFP intensity were collected as experiment group (EXP-1 and EXP-2). Genomic DNA was extracted from the cells (Quick-gDNA midiprep kit, Zymo Research), the NGS (Next Generation Sequencing) libraries were prepared as previously described 51 , and then subjected to NGS sequencing analysis (Novogene, CA). The gRNA sequences from the NGS sequencing data were extracted using standard bioinformatics methods, and the read count of gRNAs were calculated as Count Per Million (CPM). Analysis of pooled CRISPR screen To identify enriched genes based on sgRNAs in this CRISPR screen, we used the MAGeCK algorithm 22 , 23 in R. We used MAGeCK’s ‘MLE’ subcommand for maximum likelihood estimation of gene essentialities using the TMM normalized counts 52 at the gene level. This was followed by applying MAGeCKFlute’s 24 function FluteMLE with the argument to incorporate DepMap genes 53 in normalization, as recommended for human data, which produced pathway enrichment analysis. Among screening hits from MAGeCKFlute, protein connectivity networks based on physical and functional interactions were identified using STRING v12 54 , where only interactions with a medium confidence score of ≥ 0.4 were selected. Cell lines PANC-1 (#CRL-1469) and 293FT (#R7007) cell lines were obtained from ATCC and Thermo Fisher Scientific, respectively. Cells were maintained in DMEM (Gibco, 10313039), supplemented with 10% fetal bovine serum (FBS, Gibco), L-alanyl-L-glutamine (Gibco) and penicillin/streptomycin (Corning), in a 37 o C incubator with 5% CO 2 . To generate non-targeting control (NTC) and ALDH3B2 mut PANC-1 cells, non-targeting (NT) gRNA (5’- GCTTTCACGGAGGTTCGACG-3’) or ALDH3B2 gRNA (5’-GCCCTCCTCACCTGCGGCGA-3’, HGLibA_01571) oligos were cloned into LentiCRISPR-v2 vector. Wild type PANC-1 cells were then transduced by NTC or ALDH3B2 gRNA- containing lentivirus, and subsequently selected by puromycin treatment. Indel mutation in ALDH3B2 mut cells was confirmed by deep sequencing analysis (MGH DNA Core Facility, Cambridge, MA). All plasmid sequences were verified by Sanger sequencing before transduction and transfection. To generate shControl and shALDH3B2 PANC-1 cells, a scrambled or ALDH3B2-targeting shRNA was cloned into FH1t(INSR)UTG-GFP vector (A gift from Dr. Stephan Kissler). shControl and shALDH3B2 lentiviruses were then used to infect PANC-1 cells and the cells were selected by puromycin treatment. PANC-1 cells transplantation studies Experimental diabetes was induced in 8-week-old NSG male mice by intraperitoneal injection of streptozotocin (STZ) (40 mg STZ/kg body weight for five consecutive days). Animals were considered diabetic only if morning-fed blood glucose exceeded 350 mg/dl. Three days after STZ injection, ~10 7 PANC-1 cells (carry NT and ALDH3B2 mut ) were transplanted subcutaneously into each diabetic NSG mouse. Blood glucose was monitored every 3–4 days. Six weeks post cells transplantation, Intraperitoneal glucose tolerance test (IPGTT) was performed. Mice were fasted for 16 hours. The plasma glucose levels of the mice before (baseline) or 15, 30, 60, 90, and 120 minutes after intraperitoneal injection of 2 mg/g body weight glucose were recorded by a Glucose Meter. At day 52 post cells transplantation, grafts were surgically removed from NSG mice, and the blood glucose was measured 4 days later. Human pancreatic ductal cell isolation and purification Human primary pancreatic ductal cells isolation was performed as previously described 30 , 55 . In brief, “Human acinar tissue” from Integrated Islet Distribution Program (IIDP) (Donor information is summarized in Extended Table 2) was washed 2 times with PBS, and then incubated with Trypsin solution (1.5ml 0.25% Trypsin in 20ml PBS) shaking at 37°C for 15 min. Dispersed cells were centrifuged at 1,000 rpm for 5 min, the supernatant was aspirated, and then the pellets were resuspended with mouse antihuman CA19-9 antibody (Invitrogen; clone 116-NS-19-9) in 2 ml PBS solution. After 15 min incubation at 4°C, the cell suspension was mixed with 10ml PBS solution (375mg EDTA and 2.5g BSA in 500ml PBS) gently. Tubes were centrifuged at 1,000 rpm for 5 min, supernatant was aspirated, 250 µl/tube goat anti mouse IgG microbeads (Miltenyi Biotec) in PBS solution were added, and pellets were mixed. After 20 min incubation at 4°C, wash with PBS solution 2 times. Tubes were centrifuged at 1,000 rpm for 5 min, and pellets were resuspended in 20 ml cold PBS solution and passed through 40 m cell strainers to remove newly formed clumps of cells. MACS magnetic LS separation columns (Miltenyi Biotec) were prepared according to the manufacturer’s instructions. Preparation and transplantation of primary human pancreatic duct cells Human primary pancreatic ductal cells isolation was performed as above described. Purified human primary pancreatic ductal cells (HPPD) were immediately cultured in a low-attachment plate in RMPI DMEM/F12 medium (Gibco), supplemented with 10% FBS and penicillin/streptomycin. Lentivirus encoding a NT or ALDH3B2 gRNA together with Cas9 endonuclease was added to the culture media for overnight infection. The next day, HPPD cells were washed with culture media twice and ~ 10 7 cells were transplanted under the left kidney capsule of 8-week-old of STZ induced diabetic male NSG mice. Graft recipients were left to recover from surgery for three weeks. At day 56 post-transplantation, the kidney transplanted with grafts were surgically removed for gene expression analysis by immunofluorescence. The final blood glucose measurement of NSG mice was done 4 days later. Quantitative real-time PCR (qPCR) Cells or grafts were treated with TRIzol (Thermo Fisher Scientific) for RNA extraction following the manufacturer’s protocol. Purified RNA was reverse-transcribed into cDNA using the SuperScript IV first-strand synthesis kit (Invitrogen). INS (Hs00355773_m1), GCG (Hs01031536_m1), SST (Hs00356144_m1) PDX1(Hs00236830_m1), NKX6-1 (Hs00232355_m1), GCK (Hs01564555_m1), SLC2A2 (Hs00165775_m1), SLC2A1 (Hs00892681_m1), CPE (Hs00960598_m1), CA2 (Hs01070108_m1), KRT19 (Hs00761767_s1), SOX9 (Hs00165814_m1), ALDH3B2 (Hs02511514_s1) and ACTB (Hs01060665_g1) probes for TaqMan assays were purchased from Thermo Fisher Scientific. All Gene expression levels were analyzed by SYBR green PowerUp qPCR assays (Applied Biosystems). Primer sequences used is shown in the following table: Gene Forward primer sequence Reverse primer sequence ACTB 5’ CACCATTGGCAATGAGCGGTTC 3’ 5’AGGTCTTTGCGGATGTCCACGT 3’ INS 5’ ACGAGGCTTCTTCTACACACCC 3’ 5’ TCCACAATGCCACGCTTCTGCA 3’ PDX1 5’GAAGTCTACCAAAGCTCACGCG 3’ 5’ GGAACTCCTTCTCCAGCTCTAG 3’ NKX6.1 5’ CCTATTCGTTGGGGATGACAGAG 3’ 5’ TCTGTCTCCGAGTCCTGCTTCT 3’ PAX6 5’ CTGAGGAATCAGAGAAGACAGGC 3’ 5’ ATGGAGCCAGATGTGAAGGAGG 3’ MAFA 5’ GCTTCAGCAAGGAGGAGGTCAT 3’ 5’ TCTGGAGTTGGCACTTCTCGCT 3’ NEUROD 5’ GGTGCCTTGCTATTCTAAGACGC 3’ 5’ GCAAAGCGTCTGAACGAAGGAG 3’ NGN3 5’ CCTAAGAGCGAGTTGGCACTGA 3’ 5’ AGTGCCGAGTTGAGGTTGTGCA 3’ GCG 5’ CGTTCCCTTCAAGACACAGAGG 3’ 5’ ACGCCTGGAGTCCAGATACTTG 3’ SST 5’ CCAGACTCCGTCAGTTTCTGCA 3’ 5’ TTCCAGGGCATCATTCTCCGTC 3’ PP 5’ AGACACAAAGAGGACACGCTGG 3’ 5’ GAGTCGTAGGAGACAGAAGGTG 3’ SLC2A2 5’ ATGTCAGTGGGACTTGTGCTGC 3’ 5’ AACTCAGCCACCATGAACCAGG 3’ GCK 5’ CATCTCCGACTTCCTGGACAAG 3’ 5’ TGGTCCAGTTGAGAAGGATGCC 3’ ABCC8 5’ GACGACAAGAGGACAGTGGTCT 3’ 5’ GCATTCAGACCTCTGGAAGTCC 3’ KCNJ11 5’ TGTGTCACCAGCATCCACTCCT 3’ 5’ GTTCTGCACGATGAGGATCAGG 3’ IA2 5’ TGGAGATCCTGGCTGAGCATGT 3’ 5’ GGTCACATCAGCCAAAGACAGG 3’ KRT19 5’ AGCTAGAGGTGAAGATCCGCGA 3’ 5’ GCAGGACAATCCTGGAGTTCTC 3’ CA2 5’ GTGACCTGGATTGTGCTCAAGG 3’ 5’ GTTGTCCACCATCAGTTCTTCGG 3’ HNF1B 5’ CCCAGCAAATCTTGTACCAGGC 3’ 5’ ACCTCAGTGACCAAGTTGGAGC 3’ SOX9 5’ AGGAAGCTCGCGGACCAGTAC 3’ 5’ GGTGGTCCTTCTTGTGCTGCAC 3’ All qPCR assays were performed using a QuantStudio 6 Flex Real-Time PCR system (Applied Biosystems). Immunofluorescence staining and confocal microscopy The subcutaneously transplanted PANC-1 cells or the kidney with HPPD graft transplantation were surgically removed from the mice, fixed 1 hour in 4% paraformaldehyde at 4°C, and dehydrated using 30% sucrose solution overnight. The tissues were embedded in disposable base molds (Thermo Fisher Scientific) and 10 mm sections were cut. For staining, slides were blocked with PBS + 0.1% Triton X-100 (Thermo Fisher Scientific) + 5% donkey serum (Sigma-Aldrich) for 1 hr at room temperature (RT), incubated with primary antibodies overnight at 4°C, washed, incubated with secondary antibody incubation for 1 hr at RT, incubated with Hoechst 33342 (Invitrogen) for 10 min at RT, and washed. For imaging, samples were mounted in fluorescence mounting medium (Dako), covered with coverslips, and sealed with nail polish. Representative images were taken using a Zeiss LSM 710 confocal microscope. Primary antibody: Insulin (A0564, Dako), Pdx1 (5679S, Cell Signaling Technology), Dykddddk Tag (14793S, Cell Signaling Technology), C-peptide (GN-ID4, Developmental Studies Hybridoma Bank) and Cytokeratin 19 (Abcam, ab7754). Cells were seeded into 4-well culture slide (Falcon) at density of 10 5 cells/well. After another 24 hours, the cells were fixed, stained and subjected to fluorescence microscopic analysis as above described. Serum insulin measurement Mouse blood was collected from the tail tip and allowed to clot. Serum was separated by brief centrifugation according to standard protocol. The serum insulin level was measured using STELLUX® Chemi Human Insulin ELISA kits (Alpco). Electron microscopy NTC-PANC-1 or ALDH3B2 mut -PANC-1 were fixed at RT for 2 hr with a mixture containing 1.25% PFA, 2.5% glutaraldehyde, and 0.03% picric acid in 0.1M sodium cacodylate buffer (pH 7.4). Samples were then sent to the Advanced Microscopy Core of Joslin for further processing and transmission electron microscope imaging. DNA methylation analyses Bisulfite conversion (Zymo Research, EZ DNA Methylation-Direct Kits) of DNA from PANC-1, human primary pancreatic ductal cells carry NT or ALDH3B2 mut and human islets were performed as described previously 56 . Bisulfite-treated DNA was PCR amplified, using primers (human INS promoter forward primer − 5’ AGGATAGGTTGTATTAGAAGAGGTTATTAAG 3’; human INS promoter reverse primer- 5’ CCCCTAAACTCACCCCCACATACTTC 3’) specific for bisulfite treated DNA but independent of methylation status at monitored CpG sites. Reaction conditions for the first round of PCR were 5 cycles of 95°C 1 min, 52°C 3 min, 72°C 3 min followed by 40 cycles of 95°C 30 s, 55°C 45 s, 72°C 45 s followed by 7 min at 72°C. PCR products were gel purified and used for deep sequencing analysis (MGH DNA Core Facility, Cambridge, MA). Single cell RNA sequencing Isolation of human primary pancreatic ductal cells was performed as described above. Purified human primary pancreatic ductal cells (HPPD) were immediately cultured in a low-attachment plate in RMPI DMEM/F12 medium (Gibco), supplemented with 10% FBS and penicillin/streptomycin. Lentivirus encoding a NT or ALDH3B2 gRNA together with a Cas9-mCherry reporter construct was added to the culture media. 6 days later cells were harvested and alive mCherry positive cells were isolated by FACS sorting. 40,000 cells per sample were used for scRNAseq using the Chromium Next GEM Single Cell 3’ GEM, Library & Gel Bead Kit v3.1 (cat # PN-1000213; 10 x Genomics) according to manufactures’ instructions. Illumina NovaSeq 6000 with about 1.3 billion reads total was used for sequencing the purified 3’ gene expression library. The single cell RNA-seq dataset was processed, explored and visualized using Cellenics® community instance ( https://scp.biomage.net/ ) that’s hosted by Biomage ( https://biomage.net/ ) using preset standard filtering thresholds and integration methods. Pre-filtered count matrices were uploaded to Cellenics. Dead or dying cells were removed by filtering droplets with high mitochondrial content (3% cut-off). Outliers in the distribution of number of genes vs number of UMIs were removed by fitting a linear regression model (p-values between 1.97E-04–2.34E-04). Cells with a high probability of being doublets were filtered out using the scDblFinder method (threshold range: 0.50–0.56). Overall filtering rates after processing are in the range of 10.75% − 11.35% of cells. Data normalization, principal component analysis (PCA) and data integration using Harmony were performed on data from high-quality cells. Clusters were identified using the Louvain method, and a Uniform Manifold Approximation and Projection (UMAP) embedding was calculated to visualize the results. Cluster-specific marker genes were identified by comparing cells of each cluster to all other cells using the presto package implementation of the Wilcoxon rank-sum test. Modeling transcriptional dynamics of scRNA-seq database by seVelo To recover the directed dynamic information of newly transcribed, unspliced pre-mRNAs and mature, spliced mRNAs, we used velocyto 57 to count both the unspliced and spliced RNA reads from the counts that were generated by 10X Cellranger 58 . We filtered out genes that had less than 10 counts and were expressed in less than 10 cells (both unspliced and spliced), normalized every cell by its total counts, and logarithmize the normalized counts. We then filtered out cells that had less than 1000 total counts. Finally, we kept the top 3000 highly variable genes. We computed the first and second-order moments (means and uncentered variances) among nearest neighbors in PCA space. First-order is needed for deterministic velocity estimation, while stochastic estimation also requires second-order moments. Transcriptional induction for a particular gene increases newly transcribed precursor unspliced mRNAs while, conversely, repression or absence of transcription decreases unspliced mRNAs. We used scVelo 34 to solve the full transcriptional dynamics of splicing kinetics using a likelihood-based dynamical model. Statistical analyses Statistical analyses were performed by unpaired or paired tests as indicated using the Prism 8 software. All data are presented as mean ± SEM. Significance was defined as *p < 0.05, **p < 0.01 and ***p < 0.001. No samples were excluded from the analysis. Data analysis was not blinded. All data are representative of two or more similar experiments. Declarations ACKNOWLEDGEMENTS This research was supported by funds from the Beatson Foundation to P.Y and S.B-W and by postdoctoral fellowship to J. L. from the Mary K. Iacocca Foundation and JDRF. We acknowledge the support from core facilities funded both by the NIDDK Diabetes Research Center ( P30DK036836) and the Joslin Diabetes Center, especially our CRISPR screen core facility and bioinformatic core (Dr. Hui Pan and Dr. Jonathan Dreyfuss). We also acknowledge the generous comments and editing from Dr. Gordon Weir and Dr. Stephan Kissler of Joslin Diabetes Center. References Weir, G.C., Bonner-Weir, S. & Leahy, J.L. Islet mass and function in diabetes and transplantation. Diabetes 39, 401–5 (1990). Nair, G. & Hebrok, M. Islet formation in mice and men: lessons for the generation of functional insulin-producing beta-cells from human pluripotent stem cells. Curr Opin Genet Dev 32, 171–80 (2015). Nostro, M.C. et al. Efficient generation of NKX6-1 + pancreatic progenitors from multiple human pluripotent stem cell lines. Stem Cell Reports 4, 591–604 (2015). Pagliuca, F.W. et al. Generation of functional human pancreatic beta cells in vitro. Cell 159, 428 – 39 (2014). Shapiro, A.M. et al. Islet transplantation in seven patients with type 1 diabetes mellitus using a glucocorticoid-free immunosuppressive regimen. N Engl J Med 343, 230–8 (2000). Aguayo-Mazzucato, C. & Bonner-Weir, S. Pancreatic beta Cell Regeneration as a Possible Therapy for Diabetes. Cell Metab 27, 57–67 (2018). Dor, Y., Brown, J., Martinez, O.I. & Melton, D.A. Adult pancreatic beta-cells are formed by self-duplication rather than stem-cell differentiation. Nature 429, 41 – 6 (2004). Butler, A.E. et al. Adaptive changes in pancreatic beta cell fractional area and beta cell turnover in human pregnancy. Diabetologia 53, 2167–76 (2010). Ben-Othman, N. et al. Long-Term GABA Administration Induces Alpha Cell-Mediated Beta-like Cell Neogenesis. Cell 168, 73–85 e11 (2017). Collombat, P. et al. The ectopic expression of Pax4 in the mouse pancreas converts progenitor cells into alpha and subsequently beta cells. Cell 138, 449–62 (2009). Zhou, Q., Brown, J., Kanarek, A., Rajagopal, J. & Melton, D.A. In vivo reprogramming of adult pancreatic exocrine cells to beta-cells. Nature 455, 627 – 32 (2008). Bonner-Weir, S., Aguayo-Mazzucato, C. & Weir, G.C. Dynamic development of the pancreas from birth to adulthood. Ups J Med Sci 121, 155–8 (2016). Bonner-Weir, S. et al. The pancreatic ductal epithelium serves as a potential pool of progenitor cells. Pediatr Diabetes 5 Suppl 2, 16–22 (2004). Inada, A. et al. Carbonic anhydrase II-positive pancreatic cells are progenitors for both endocrine and exocrine pancreas after birth. Proc Natl Acad Sci U S A 105, 19915–9 (2008). Qadir, M.M.F. et al. P2RY1/ALK3-Expressing Cells within the Adult Human Exocrine Pancreas Are BMP-7 Expandable and Exhibit Progenitor-like Characteristics. Cell Rep 22, 2408–2420 (2018). Qadir, M.M.F. et al. Single-cell resolution analysis of the human pancreatic ductal progenitor cell niche. Proc Natl Acad Sci U S A 117, 10876–10887 (2020). Wang, D. et al. Long-Term Expansion of Pancreatic Islet Organoids from Resident Procr(+) Progenitors. Cell 180, 1198–1211 e19 (2020). Lieber, M., Mazzetta, J., Nelson-Rees, W., Kaplan, M. & Todaro, G. Establishment of a continuous tumor-cell line (panc-1) from a human carcinoma of the exocrine pancreas. Int J Cancer 15, 741–7 (1975). Madden, M.E. & Sarras, M.P., Jr. Morphological and biochemical characterization of a human pancreatic ductal cell line (PANC-1). Pancreas 3, 512 – 28 (1988). Chai, R., Chen, S., Ding, J. & Grayburn, P.A. Efficient, glucose responsive and islet-specific transgene expression by a modified rat insulin promoter. Gene Ther 16, 1202–9 (2009). Sanjana, N.E., Shalem, O. & Zhang, F. Improved vectors and genome-wide libraries for CRISPR screening. Nat Methods 11, 783–784 (2014). Li, W. et al. MAGeCK enables robust identification of essential genes from genome-scale CRISPR/Cas9 knockout screens. Genome Biol 15, 554 (2014). Li, W. et al. Quality control, modeling, and visualization of CRISPR screens with MAGeCK-VISPR. Genome Biology 16, 281 (2015). Wang, B. et al. Integrative analysis of pooled CRISPR genetic screens using MAGeCKFlute. Nat Protoc 14, 756–780 (2019). Lai, L., Reineke, E., Hamilton, D.J. & Cooke, J.P. Glycolytic Switch Is Required for Transdifferentiation to Endothelial Lineage. Circulation 139, 119–133 (2019). Vaughn, N. & Haviland, D.L. Acly promotes metabolic reprogramming and induction of IRF4 during early CD8(+) T cell activation. Cytometry A 99, 825–831 (2021). Marchitti, S.A., Brocker, C., Stagos, D. & Vasiliou, V. Non-P450 aldehyde oxidizing enzymes: the aldehyde dehydrogenase superfamily. Expert Opin Drug Metab Toxicol 4, 697–720 (2008). Vassalli, G. Aldehyde Dehydrogenases: Not Just Markers, but Functional Regulators of Stem Cells. Stem Cells Int 2019, 3904645 (2019). Kim-Muller, J.Y. et al. Aldehyde dehydrogenase 1a3 defines a subset of failing pancreatic beta cells in diabetic mice. Nat Commun 7, 12631 (2016). Yatoh, S. et al. Differentiation of affinity-purified human pancreatic duct cells to beta-cells. Diabetes 56, 1802–9 (2007). Wescott, M.P. et al. Pancreatic ductal morphogenesis and the Pdx1 homeodomain transcription factor. Mol Biol Cell 20, 4838–44 (2009). Neiman, D. et al. Islet cells share promoter hypomethylation independently of expression, but exhibit cell-type-specific methylation in enhancers. Proc Natl Acad Sci U S A 114, 13525–13530 (2017). Ma, Z. et al. Deciphering early human pancreas development at the single-cell level. Nat Commun 14, 5354 (2023). Bergen, V., Lange, M., Peidli, S., Wolf, F.A. & Theis, F.J. Generalizing RNA velocity to transient cell states through dynamical modeling. Nat Biotechnol 38, 1408–1414 (2020). Sarnobat, D. et al. Taurine rescues pancreatic beta-cell stress by stimulating alpha-cell transdifferentiation. Biofactors 49, 646–662 (2023). Kong, C. et al. NEDD9 is a positive regulator of epithelial-mesenchymal transition and promotes invasion in aggressive breast cancer. PLoS One 6, e22666 (2011). Liu, C. et al. HIF1alpha Regulates Early Metabolic Changes due to Activation of Innate Immunity in Nuclear Reprogramming. Stem Cell Reports 14, 192–200 (2020). Kong, H.J. et al. Crosstalk between YAP and TGFbeta regulates SERPINE1 expression in mesenchymal lung cancer cells. Int J Oncol 58, 111–121 (2021). Mezza, T. et al. Insulin resistance alters islet morphology in nondiabetic humans. Diabetes 63, 994–1007 (2014). Jones, R.J. et al. Assessment of aldehyde dehydrogenase in viable cells. Blood 85, 2742–6 (1995). Storms, R.W. et al. Isolation of primitive human hematopoietic progenitors on the basis of aldehyde dehydrogenase activity. Proc Natl Acad Sci U S A 96, 9118–23 (1999). Cai, J. et al. Membrane properties of rat embryonic multipotent neural stem cells. J Neurochem 88, 212 – 26 (2004). Obermair, F.J. et al. A novel classification of quiescent and transit amplifying adult neural stem cells by surface and metabolic markers permits a defined simultaneous isolation. Stem Cell Res 5, 131–43 (2010). Rovira, M. et al. Isolation and characterization of centroacinar/terminal ductal progenitor cells in adult mouse pancreas. Proc Natl Acad Sci U S A 107, 75–80 (2010). Kitamura, T., Takagi, S., Naganuma, T. & Kihara, A. Mouse aldehyde dehydrogenase ALDH3B2 is localized to lipid droplets via two C-terminal tryptophan residues and lipid modification. Biochem J 465, 79–87 (2015). Rovira, M. et al. Chemical screen identifies FDA-approved drugs and target pathways that induce precocious pancreatic endocrine differentiation. Proc Natl Acad Sci U S A 108, 19264–9 (2011). Ninomiya, K., Takahashi, A., Fujioka, Y., Ishikawa, Y. & Yokoyama, M. Transforming growth factor-beta signaling enhances transdifferentiation of macrophages into smooth muscle-like cells. Hypertens Res 29, 269–76 (2006). Miettinen, P.J., Ebner, R., Lopez, A.R. & Derynck, R. TGF-beta induced transdifferentiation of mammary epithelial cells to mesenchymal cells: involvement of type I receptors. J Cell Biol 127, 2021–36 (1994). Sabatini, P.V., Speckmann, T. & Lynn, F.C. Friend and foe: beta-cell Ca(2+) signaling and the development of diabetes. Mol Metab 21, 1–12 (2019). Joung, J. et al. Genome-scale CRISPR-Cas9 knockout and transcriptional activation screening. Nat Protoc 12, 828–863 (2017). Shalem, O. et al. Genome-scale CRISPR-Cas9 knockout screening in human cells. Science 343, 84–87 (2014). Robinson, M. & Oshlack, A. A scaling normalization method for differential expression analysis of RNA-seq data. Genome Biology 11, 1–9 (2010). Tsherniak, A. et al. Defining a Cancer Dependency Map. Cell 170, 564–576 e16 (2017). Szklarczyk, D. et al. STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res 47, D607-D613 (2019). Corritore, E. et al. beta-Cell differentiation of human pancreatic duct-derived cells after in vitro expansion. Cell Reprogram 16, 456–66 (2014). Millar, D.S., Warnecke, P.M., Melki, J.R. & Clark, S.J. Methylation sequencing from limiting DNA: embryonic, fixed, and microdissected cells. Methods 27, 108 – 13 (2002). La Manno, G. et al. RNA velocity of single cells. Nature 560, 494–498 (2018). Zheng, G.X. et al. Massively parallel digital transcriptional profiling of single cells. Nat Commun 8, 14049 (2017). Additional Declarations There is NO Competing Interest. Supplementary Files SupplementalTable1.xlsx Complete Read Count file for the genome-wide CRISPR screen SupplementalTable2.pdf Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-2222452","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":363238467,"identity":"692f7beb-5918-4e72-9616-d1144327b92e","order_by":0,"name":"Peng Yi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABFklEQVRIiWNgGAWjYPACCTnGBgbGAwkMDAkQATYgPoBfizFQCwNJWhgSG6BqCGsxZ+89/IKxzSK9ub39wYGHOxjy+GefMfxcUcYgx3cjAasWy55zaRaMbRK5jT1nDA4knmEoljiXYyx55hyDsSQOLQY3cswMwFpm5DAcSGwDuvAMW4JkI5CxAZeW+2/AWtIZZ6Q/AGuZf4Yt+SdQSz1OLTd4jB8AtSQwzkgwAGvZcIb5GMiWBAOcfskxY0g4J2EI8UubROJGoBbLBqDIzDMPcITYGeMPH8rq5A3b2x8+/NlmkzjvDGPzzYYyG3m+4zgcBowCiURgLBg2gPkSMAkJrMqhWpg/MPxhYJDHqWQUjIJRMApGPAAAbrVnZh4wFm8AAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0001-7830-4694","institution":"Joslin Diabetes Center","correspondingAuthor":true,"prefix":"","firstName":"Peng","middleName":"","lastName":"Yi","suffix":""},{"id":363238468,"identity":"23f74905-b7a0-40ad-9d41-120a2b48edbc","order_by":1,"name":"Jian Li","email":"","orcid":"","institution":"Joslin Diabetes Center","correspondingAuthor":false,"prefix":"","firstName":"Jian","middleName":"","lastName":"Li","suffix":""},{"id":363238469,"identity":"e6521b72-dca2-4d50-9a55-a2a41de22535","order_by":2,"name":"Yu-chi Lee","email":"","orcid":"","institution":"Joslin Diabetes Center","correspondingAuthor":false,"prefix":"","firstName":"Yu-chi","middleName":"","lastName":"Lee","suffix":""},{"id":363238470,"identity":"48658d1a-d14a-49ca-9e7f-9028cbe54309","order_by":3,"name":"Noelle Morrow","email":"","orcid":"https://orcid.org/0000-0001-7015-4273","institution":"UCLA","correspondingAuthor":false,"prefix":"","firstName":"Noelle","middleName":"","lastName":"Morrow","suffix":""},{"id":363238471,"identity":"cf44cbe5-e692-41fb-8ae2-3042f88add4c","order_by":4,"name":"Jennifer Hollister-Lock","email":"","orcid":"","institution":"Joslin Diabetes Center","correspondingAuthor":false,"prefix":"","firstName":"Jennifer","middleName":"","lastName":"Hollister-Lock","suffix":""},{"id":363238472,"identity":"291a966c-4c7d-4600-8f35-083712e55bf0","order_by":5,"name":"Brooke Sullivan","email":"","orcid":"","institution":"Joslin Diabetes Center","correspondingAuthor":false,"prefix":"","firstName":"Brooke","middleName":"","lastName":"Sullivan","suffix":""},{"id":363238473,"identity":"cb620c95-145c-4c3a-bc5d-62bfd5989af0","order_by":6,"name":"Taylor Stewart","email":"","orcid":"https://orcid.org/0000-0002-0744-2309","institution":"Joslin Diabetes Center, Harvard Medical School","correspondingAuthor":false,"prefix":"","firstName":"Taylor","middleName":"","lastName":"Stewart","suffix":""},{"id":363238474,"identity":"7ed80fa3-c1ad-4aa3-bc0d-5728825b4963","order_by":7,"name":"Siying Wei","email":"","orcid":"","institution":"Joslin Diabetes Center","correspondingAuthor":false,"prefix":"","firstName":"Siying","middleName":"","lastName":"Wei","suffix":""},{"id":363238475,"identity":"dced7823-10c4-4878-bad4-c4e3e9a73a47","order_by":8,"name":"Jessica Pereira","email":"","orcid":"https://orcid.org/0000-0002-8503-1732","institution":"Joslin Diabetes Center","correspondingAuthor":false,"prefix":"","firstName":"Jessica","middleName":"","lastName":"Pereira","suffix":""},{"id":363238476,"identity":"23a1abce-e2f8-4517-bdff-34306216207e","order_by":9,"name":"Alexander Lee-Papastavros","email":"","orcid":"","institution":"Joslin Diabetes Center","correspondingAuthor":false,"prefix":"","firstName":"Alexander","middleName":"","lastName":"Lee-Papastavros","suffix":""},{"id":363238477,"identity":"f27a86f1-6c5d-46b8-b1ae-23243c547f79","order_by":10,"name":"Susan Bonner-Weir","email":"","orcid":"https://orcid.org/0000-0003-4682-0656","institution":"Joslin Diabetes Center","correspondingAuthor":false,"prefix":"","firstName":"Susan","middleName":"","lastName":"Bonner-Weir","suffix":""},{"id":363238478,"identity":"007b342b-2953-420c-951d-309632c82f42","order_by":11,"name":"Kevin Bode","email":"","orcid":"","institution":"Joslin Diabetes Center","correspondingAuthor":false,"prefix":"","firstName":"Kevin","middleName":"","lastName":"Bode","suffix":""},{"id":363238479,"identity":"d8772f3c-fc0d-481c-9950-e556ad314787","order_by":12,"name":"Andy Ma","email":"","orcid":"","institution":"Joslin Diabetes Center","correspondingAuthor":false,"prefix":"","firstName":"Andy","middleName":"","lastName":"Ma","suffix":""},{"id":363238480,"identity":"057c0108-93cc-4665-a67f-fe784d239036","order_by":13,"name":"Mulin He","email":"","orcid":"https://orcid.org/0000-0002-9055-8109","institution":"Joslin Diabetes Center","correspondingAuthor":false,"prefix":"","firstName":"Mulin","middleName":"","lastName":"He","suffix":""},{"id":363238481,"identity":"2c77f3c5-bca1-40d1-8052-e418d523c020","order_by":14,"name":"Hui Pan","email":"","orcid":"https://orcid.org/0000-0001-5748-7217","institution":"Joslin Diabetes Center","correspondingAuthor":false,"prefix":"","firstName":"Hui","middleName":"","lastName":"Pan","suffix":""},{"id":363238482,"identity":"bf838ee9-b373-4105-95c5-6a5261fd2a58","order_by":15,"name":"Jonathan Dreyfuss","email":"","orcid":"https://orcid.org/0000-0001-7242-3991","institution":"Joslin Diabetes Center","correspondingAuthor":false,"prefix":"","firstName":"Jonathan","middleName":"","lastName":"Dreyfuss","suffix":""}],"badges":[],"createdAt":"2022-10-31 15:01:06","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-2222452/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-2222452/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":66153867,"identity":"271ea1bd-168e-44a8-90b2-8cd463993189","added_by":"auto","created_at":"2024-10-08 08:18:37","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":579675,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eGenome-wide CRISPR screen identifies ALDH3B2 as a regulator of pancreatic duct-to-beta cell transdifferentiation.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea, \u003c/strong\u003eThe structure of REPB reporter and an illustration of genome-wide CRISPR screen workflow. The REPB reporter contains Rat insulin promoter (RIP3.1) driving the expression of EGFP and Blasticidin-S deaminase (BSD), the EGFP and BSD genes are fused together with P2A peptide. \u003cstrong\u003eb, \u003c/strong\u003eRank all the genes based on their scores and label genes in the rank plot.\u003cstrong\u003e \u003c/strong\u003eY-axis represents difference in scores between EGFPhigh, blasticidin-resistant REPB-PANC-1 cells vs control cells from MAGeCKFlute at the gene level, where positive scores indicate positive enrichment and negative scores indicate negative enrichment; x-axis represents the gene rank. Large magnitude Score are colored, where the most extreme genes are labeled. \u003cstrong\u003ec, \u003c/strong\u003eHalf-volcano plot at the gene level showing positive enrichment where x-axis is log\u003csub\u003e2\u003c/sub\u003e fold change of EGFPhigh, blasticidin-resistant REPB-PANC-1 cells vs control cells from MAGeCK-mle and y-axis is significance that was winsorized to 10\u003csup\u003e-5\u003c/sup\u003e. The strongest hits are colored in red and labeled. The horizontal dashed line represents a p-value of 0.05. \u003cstrong\u003ed,\u003c/strong\u003e KEGG pathways enriched for positively selected genes from MAGeCKFlute where dot size indicates Normalized Enrichment Score (NES).\u0026nbsp;\u003cstrong\u003ee and f,\u003c/strong\u003e STRING pathway analysis showing protein-protein associations including physical and functional interactions between INS, ACLY, and other GFP-positive screening hits from MAGeCKFlute \u003cstrong\u003e(e)\u003c/strong\u003e and for GFP-positive screening hits from MAGeCKFlute clustered into functionally associated groups \u003cstrong\u003e(f).\u003c/strong\u003e \u003cstrong\u003eg and\u003c/strong\u003e \u003cstrong\u003eh: \u003c/strong\u003eqPCR analysis of INS \u003cstrong\u003e(g)\u003c/strong\u003e and KRT19 \u003cstrong\u003e(h) \u003c/strong\u003eexpression level for top 8 candidate gene mutant PANC-1 cells.\u003c/p\u003e","description":"","filename":"fig1.png","url":"https://assets-eu.researchsquare.com/files/rs-2222452/v1/732d1bc607d9f550d4a2d091.png"},{"id":66153873,"identity":"8172a922-b3ba-4481-acb1-25f0e6c0b521","added_by":"auto","created_at":"2024-10-08 08:18:37","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":855976,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLoss-of-function of ALDH3B2 trans-differentiates PANC-1 cells into insulin-producing beta-like cells.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea-d:\u003c/strong\u003e qPCR analysis of human beta cell and duct signature genes. Data show mean ± SEM of n=4 biological replicates per condition and are representative of 2–3 independent experiments. *p \u0026lt; 0.05, **p \u0026lt; 0.01, calculated by two-way ANOVA with Sidak’s multiple comparisons test. \u003cstrong\u003ea, \u003c/strong\u003eGenes related to pancreatic endocrine hormones. \u003cstrong\u003eb,\u003c/strong\u003e Genes related to beta cell key transcription factors. \u003cstrong\u003ec,\u003c/strong\u003e Genes related to beta cell function genes. \u003cstrong\u003ed,\u003c/strong\u003e Genes related to pancreatic duct cell markers.\u0026nbsp; \u003cstrong\u003ee, \u003c/strong\u003eInsulin and C-peptide immunofluorescence in NTC cells and ALDH3B2\u003csup\u003emut\u003c/sup\u003e-PANC-1 cells. \u003cstrong\u003ef:\u003c/strong\u003e Insulin content in NTC cells and ALDH3B2\u003csup\u003emut\u003c/sup\u003e-PANC-1 cells were measured by insulin ELISA, normalized by cells genomic DNA content. n=7 biological replicates per condition and genotype. Data show mean ± SEM, **p \u0026lt; 0.01. \u003cstrong\u003eg\u003c/strong\u003e, Electron microscopy (EM) analysis of NTC-PANC-1 cells and ALDH3B2\u003csup\u003emut\u003c/sup\u003e-PANC-1 cells. Insulin secretion vesicles are indicated by arrow heads. \u003cstrong\u003eh\u003c/strong\u003e, An illustration of the in vivo function evaluation of trans-differentiated PANC-1 cells, and random\u003cstrong\u003e \u003c/strong\u003eblood glucose levels in the NTC and ALDH3B2\u003csup\u003emut\u003c/sup\u003e-PANC-1 cells transplanted diabetic NSG mice over 8 weeks of monitoring. Data show mean ± SEM of n=5 mice (NTC-PANC-1) and n=6 mice (ALDH3B2\u003csup\u003emut\u003c/sup\u003e-PANC-1) per group and are representative of three experiments, calculated by two-way ANOVA with Tukey’s multiple comparisons test.\u003cstrong\u003e i, \u003c/strong\u003eAn\u003cstrong\u003e \u003c/strong\u003eIntraperitoneal glucose tolerance test (IPGTT) in NTC or ALDH3B2\u003csup\u003emut\u003c/sup\u003e-PANC-1 cells transplanted NSG mice 6 weeks post-transplantation. \u003cstrong\u003ej, \u003c/strong\u003eSerum human\u003cstrong\u003e \u003c/strong\u003einsulin levels measurement in the NTC and ALDH3B2\u003csup\u003emut\u003c/sup\u003e-PANC-1 cells transplanted diabetic NSG mice. \u003cstrong\u003ek,\u003c/strong\u003e Pdx1 and Insulin immunofluorescence in transplanted NTC cells and ALDH3B2\u003csup\u003emut\u003c/sup\u003e-PANC-1 cells. \u003cstrong\u003el,\u003c/strong\u003e Nkx6.1 and Insulin immunofluorescence in transplanted NTC cells and ALDH3B2\u003csup\u003emut\u003c/sup\u003e-PANC-1 cells. \u003cstrong\u003em-o,\u003c/strong\u003e Percentage of Pdx1\u003csup\u003e+\u003c/sup\u003e \u003cstrong\u003e(m)\u003c/strong\u003e, Insulin\u003csup\u003e+\u003c/sup\u003e \u003cstrong\u003e(n)\u003c/strong\u003e, Nkx6.1\u003csup\u003e+\u003c/sup\u003e \u003cstrong\u003e(o)\u003c/strong\u003e cells in all CK19\u003csup\u003e+\u003c/sup\u003e cells. \u003cstrong\u003ep,\u003c/strong\u003e Percentage of Insulin\u003csup\u003e+\u003c/sup\u003e cells in all Pdx1\u003csup\u003e+\u003c/sup\u003e cells.\u003c/p\u003e","description":"","filename":"fig2.png","url":"https://assets-eu.researchsquare.com/files/rs-2222452/v1/bbf15f40226ab839213bc3ca.png"},{"id":66153876,"identity":"6fb09c97-704d-4489-b309-2bef0c59ccb5","added_by":"auto","created_at":"2024-10-08 08:18:37","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1691138,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eKnock-down of ALDH3B2 by shRNA trans-differentiates PANC-1 cells into beta-like cells.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea,\u003c/strong\u003e Generation of inducible non-targeting (shControl) cells and shALDH3B2-PANC-1\u003csup\u003e \u003c/sup\u003ecells, and ALDH3B2 expression is significantly reduced in shALDH3B2 PANC-1\u003csup\u003e \u003c/sup\u003ecells with Doxycycline treatment shown by qPCR analysis. \u003cstrong\u003eb-e:\u003c/strong\u003e qPCR analysis of human beta cell and duct signature genes. Data show mean ± SEM of n=4 biological replicates per condition and are representative of 2–3 independent experiments. *p \u0026lt; 0.05, **p \u0026lt; 0.01, calculated by two-way ANOVA with Sidak’s multiple comparisons test. \u003cstrong\u003eb, \u003c/strong\u003eGenes related to beta cell key transcription factors. \u003cstrong\u003ec,\u003c/strong\u003e Genes related to pancreatic endocrine hormones. \u003cstrong\u003ed,\u003c/strong\u003e Genes related to beta cell function genes. \u003cstrong\u003ee,\u003c/strong\u003e Genes related to pancreatic duct cell markers. \u003cstrong\u003ef, \u003c/strong\u003eInsulin and C-peptide immunofluorescence imaging in shControl and shALDH3B2 PANC-1 cells with or without Doxycycline treatment.\u003c/p\u003e","description":"","filename":"fig3.png","url":"https://assets-eu.researchsquare.com/files/rs-2222452/v1/b3a8cdb56074cbcb84a07ef1.png"},{"id":66153871,"identity":"4483ce0d-ef77-4498-9a8f-ba2596c5d37f","added_by":"auto","created_at":"2024-10-08 08:18:37","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":820242,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLoss-of-function of ALDH3B2 trans-differentiates human primary pancreatic ductal cells into beta-like cells.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea-c: \u003c/strong\u003eqPCR analysis of human Insulin \u003cstrong\u003e(a)\u003c/strong\u003e, KRT19 \u003cstrong\u003e(b)\u003c/strong\u003e and ALDH3B2 \u003cstrong\u003e(c)\u003c/strong\u003e expression in purified human primary pancreatic duct cells (HPPD) and human primary islets. Data show mean ± SEM of n=3. ***p \u0026lt; 0.005. \u003cstrong\u003ed-g\u003c/strong\u003e, qPCR analysis of human beta cell and duct signature genes. Data show mean ± SEM of n=3 biological replicates per condition and are representative of 2 independent experiments. *P \u0026lt; 0.05, **P \u0026lt; 0.01, calculated by two-way ANOVA with Sidak’s multiple comparisons test. \u003cstrong\u003ed, \u003c/strong\u003eGenes related to beta cell key transcription factors. \u003cstrong\u003ee,\u003c/strong\u003e Genes related to pancreatic endocrine hormones. \u003cstrong\u003ef,\u003c/strong\u003e Genes related to beta cell function genes. \u003cstrong\u003eg,\u003c/strong\u003e Genes related to pancreatic duct cell markers. \u003cstrong\u003eh, \u003c/strong\u003eInsulin and CK19 immunofluorescence in NTC-HPPD and ALDH3B2\u003csup\u003emut\u003c/sup\u003e-HPPD cells. \u003cstrong\u003ei\u003c/strong\u003e, Electron microscopy (EM) analysis of NTC-HPPD cells, ALDH3B2\u003csup\u003emut\u003c/sup\u003e-HPPD cells and human primary islet cells. Arrow heads points to insulin secretion vesicles.\u003c/p\u003e","description":"","filename":"fig4.png","url":"https://assets-eu.researchsquare.com/files/rs-2222452/v1/5b218922b4c7546fbfb7d2a1.png"},{"id":66153870,"identity":"7caff729-0798-4e5b-b441-4cd2333df078","added_by":"auto","created_at":"2024-10-08 08:18:37","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1104810,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eALDH3B2 mutation trans-differentiated human primary pancreatic duct cells are functional \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003ein vivo\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea, \u003c/strong\u003eAn illustration of the in vivo function evaluation of NTC-HPPD cells and ALDH3B2\u003csup\u003emut\u003c/sup\u003e-HPPD cells, and random\u003cstrong\u003e \u003c/strong\u003eblood glucose levels in the NTC-HPPD cells and ALDH3B2\u003csup\u003emut\u003c/sup\u003e-HPPD cells transplanted diabetic NSG mice for 60 days of monitoring. Data show mean ± SEM of n=5 mice per group, calculated by two-way ANOVA with Tukey’s multiple comparisons test. \u003cstrong\u003eb and c, \u003c/strong\u003eSerum human\u003cstrong\u003e \u003c/strong\u003einsulin levels in non-transplanted NSG mice (Normal NSG), NTC-HPPD cells and ALDH3B2\u003csup\u003emut\u003c/sup\u003e-HPPD cells transplanted NSG mice 5 minutes after intraperitoneal (IP) glucose injection at 1 week post-transplantation \u003cstrong\u003e(b) \u003c/strong\u003eand 3 weeks post-transplantation \u003cstrong\u003e(c).\u003c/strong\u003e Calculated by two-way ANOVA with Sidak’s multiple comparisons test. * NSG vs. NTC-HPPD; # NSG vs. ALDH3B2\u003csup\u003emut\u003c/sup\u003e-HPPD. \u003cstrong\u003ed\u003c/strong\u003e, Insulin, CK19 and Cas9 (Flag) immunofluorescence of transplanted NTC-HPPD and ALDH3B2\u003csup\u003emut\u003c/sup\u003e-HPPD cells. \u003cstrong\u003ee\u003c/strong\u003e, Pdx1, CK19 and Cas9 (Flag) immunofluorescence of transplanted NTC-HPPD and ALDH3B2\u003csup\u003emut\u003c/sup\u003e-HPPD cells. \u003cstrong\u003ef-i\u003c/strong\u003e, percentage of Cas9 lentivirus transduced cells in all transplanted duct cells \u003cstrong\u003e(f)\u003c/strong\u003e, Insulin\u003csup\u003e+\u003c/sup\u003e cells in all Cas9 lentivirus transduced cells \u003cstrong\u003e(g)\u003c/strong\u003e, Pdx1\u003csup\u003e+\u003c/sup\u003e cells in all Cas9 lentivirus transduced cells \u003cstrong\u003e(h)\u003c/strong\u003e and Insulin\u003csup\u003e+\u003c/sup\u003e cells in all Pdx1\u003csup\u003e+\u003c/sup\u003e cells \u003cstrong\u003e(i)\u003c/strong\u003e.\u003c/p\u003e","description":"","filename":"fig5.png","url":"https://assets-eu.researchsquare.com/files/rs-2222452/v1/ce7655b5eb67c5cbdf697b51.png"},{"id":66153869,"identity":"c6ced8f4-6433-4fe9-a0c7-532ce9cb33df","added_by":"auto","created_at":"2024-10-08 08:18:37","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":431718,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eALDH3B2 loss-of-function in pancreatic duct cells leads to epigenetic changes.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea,\u003c/strong\u003e Pattern and percentage of the DNA methylation at the position +63, +127 and +139 of the human insulin gene locus in NTC or ALDH3B2\u003csup\u003emut\u003c/sup\u003e-PANC-1 cells. \u003cstrong\u003eb\u003c/strong\u003e, quantification of the DNA methylation percentage in \u003cstrong\u003ea\u003c/strong\u003e. \u003cstrong\u003ec\u003c/strong\u003e, Pattern and percentage of the DNA methylation at the position +63, +127 and +139 of the human insulin gene locus in HPPD-NTC, HPPD-ALDH3B2\u003csup\u003emut \u003c/sup\u003eand human primary pancreatic islet cells. \u003cstrong\u003ed\u003c/strong\u003e, quantification of the DNA methylation percentage in \u003cstrong\u003ec\u003c/strong\u003e.\u003c/p\u003e","description":"","filename":"fig6.png","url":"https://assets-eu.researchsquare.com/files/rs-2222452/v1/9c04c24ae4c8551cc1e359ae.png"},{"id":66155307,"identity":"ea88ccae-da28-499f-91e9-89da907ef466","added_by":"auto","created_at":"2024-10-08 08:26:37","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":876017,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSingle cell RNA sequencing of ALDH3B2 mutant human primary pancreatic duct cells identified multiple clusters of insulin producing beta-like cells with overlapping duct cell gene expression profiles.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea-c:\u003c/strong\u003e UMAP plots showing all identified cell cluster (\u003cstrong\u003ea\u003c/strong\u003e), insulin (INS) expression level of non-target control (NTC) (\u003cstrong\u003eb\u003c/strong\u003e) and ALDH3B2\u003csup\u003emut\u003c/sup\u003e (\u003cstrong\u003ec\u003c/strong\u003e) human primary pancreatic duct cells (HPPD). \u003cstrong\u003ed\u003c/strong\u003e, violin plot showing INS expression level of each individual cell grouped by cell cluster for NTC (left) and ALDH3B2\u003csup\u003emut\u003c/sup\u003e HPPD cells (right). \u003cstrong\u003ee\u003c/strong\u003e, Heatmap showing indicated endocrine cell and duct cell maker gene expression of control (blue) and ALDH3B2\u003csup\u003emut\u003c/sup\u003e (red) HPPD cells at single cell level grouped by cell cluster. \u003cstrong\u003ef\u003c/strong\u003e, volcano plot showing differential expressed genes in ALDH3B2\u003csup\u003emut\u003c/sup\u003e HPPD cells comparing cells with high INS expression (intensity \u0026gt;1) vs. low INS expression (intensity \u0026lt;1). \u003cstrong\u003eg\u003c/strong\u003e, Gene set enrichment analysis showing the upregulated Reactome gene sets of ALDH3B2\u003csup\u003emut\u003c/sup\u003e HPPD cells using the top 300 most significant upregulated genes of high INS-expressing (intensity \u0026gt;1) vs. low INS-expressing (intensity \u0026lt;1) cells. \u003cstrong\u003eh\u003c/strong\u003e and \u003cstrong\u003ei\u003c/strong\u003e, Dot plots showing average gene expression of indicated endocrine, progenitor, and duct cell marker genes in all identified cell cluster of NTC (\u003cstrong\u003eh\u003c/strong\u003e) or ALDH3B2\u003csup\u003emut\u003c/sup\u003e HPPD cells (\u003cstrong\u003ei\u003c/strong\u003e).\u003c/p\u003e","description":"","filename":"fig7.png","url":"https://assets-eu.researchsquare.com/files/rs-2222452/v1/2028c792998fb8d9e75fe6bf.png"},{"id":66155306,"identity":"6bebdb4f-b68d-47c9-99ed-d90db2c82b54","added_by":"auto","created_at":"2024-10-08 08:26:37","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":837135,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDynamic change in ALDH3B2 mutant human primary pancreatic duct cells transdifferentiation is revealed by RNA Velocity and PAGA Trajectory Analysis.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea.\u003c/strong\u003e Velocity stream from scVelo analysis. The direction and magnitude of velocities are projected as arrows onto the Uniform Manifold Approximation and Projection (UMAP) plot of gene expression values across cells. Arrows depict predicted fate trajectories. \u003cstrong\u003eb.\u003c/strong\u003e Velocity-based PAGA trajectory inference using scVelo dynamic model. The predicted trajectory reflects the developmental relations shown in \u003cstrong\u003ea.\u003c/strong\u003e \u003cstrong\u003ec.\u003c/strong\u003e Latent time plot (scale from black [0] to yellow [1] indicating initial and terminal states, respectively). \u003cstrong\u003ed.\u003c/strong\u003e The velocity-based length indicating rate of transition (left), and velocity confidence (right). \u003cstrong\u003ee.\u003c/strong\u003e Single gene\u003cstrong\u003e \u003c/strong\u003evelocity visualization with spliced/unspliced scatter plot (left), RNA velocity (mid), and gene expression\u003cstrong\u003e \u003c/strong\u003e(right) for INS, TRPM3, KRT19, and CFTR. \u003cstrong\u003ef.\u003c/strong\u003e Heatmap of velocity genes along INS-expressing cell development showing the means of spliced counts of the top-ranked dynamical genes in clusters 2, 5, and 9 for ALDH3B2\u003csup\u003emut\u003c/sup\u003e-HPPD. The columns were sorted by the latent time of cells in those clusters.\u003c/p\u003e","description":"","filename":"fig8.png","url":"https://assets-eu.researchsquare.com/files/rs-2222452/v1/91784e92bafb9d3611dd84f5.png"},{"id":68266981,"identity":"ee621137-61af-4918-83b1-99311ea31baf","added_by":"auto","created_at":"2024-11-05 13:03:59","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":9237822,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-2222452/v1/b2658c5c-bf12-496f-83da-cd60b55686f1.pdf"},{"id":66153877,"identity":"008f3f33-eac1-46b9-92b4-ede64a8516d7","added_by":"auto","created_at":"2024-10-08 08:18:38","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":33622011,"visible":true,"origin":"","legend":"\u003cp\u003eComplete Read Count file for the genome-wide CRISPR screen\u003c/p\u003e","description":"","filename":"SupplementalTable1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-2222452/v1/0caeea4f4c1d7b1561aa6802.xlsx"},{"id":66153874,"identity":"505c4e67-986c-4a3d-90f5-5f7d6dade77c","added_by":"auto","created_at":"2024-10-08 08:18:37","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":83030,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"SupplementalTable2.pdf","url":"https://assets-eu.researchsquare.com/files/rs-2222452/v1/0d1e5f80403d1ec52169be49.pdf"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Loss-of-function of ALDH3B2 transdifferentiates human pancreatic duct cells into beta-like cells","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eDiabetes, no matter the cause or type, is a disease of pancreatic beta cell deficiency \u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. Current treatments for diabetes do not provide the same degree of exquisite glycemic control as would a sufficient number of functional beta cells, and thus do not prevent the debilitating correlates of long-term diabetes. To cure diabetes, one has to find a way to stop the recurrent autoimmune attack on beta cells (type 1 diabetes) or resolve persistent peripheral insulin resistance (type 2 diabetes), but restoring a sufficient functional beta cell mass is critical to a cure for both types of diabetes. Beta cells can be replenished by transplanting human cadaveric islets or islet-like cells derived from human embryonic stem cells (hESC) or induced pluripotent stem cells (iPSCs) \u003csup\u003e\u003cspan additionalcitationids=\"CR3 CR4\" citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. In most cases, transplanted islet cells are HLA-mismatched with the recipient, and immunosuppression is required to prevent graft rejection, creating other complications such as lack of immune defense against pathogens or tumor formation. Alternatively, it may be possible to promote endogenous pancreatic beta cell regeneration. Promoting the regeneration of a patient's own beta cells could be a safer strategy for beta cell mass replenishment; for type 1 diabetes autoimmunity would still need to be tamed but there would be no need for immunosuppression for allo-graft rejection. Beta cell regeneration (reviewed in \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e) can be achieved by self-duplication \u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e or transdifferentiation from other pancreatic cell types such as duct cells \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e, alpha cells \u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e,\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e or acinar cells \u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. Pancreatic beta cell replication is the dominant mechanism of beta cell regeneration in adult rodents \u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. However in human, the adult beta cell replication rate is extremely low and it was postulated that human beta cells regeneration is achieved mainly by transdifferentiation from pancreatic duct cells \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. It has been suggested that pancreatic duct cells may serve as a pool of progenitors for both the islet and acinar tissues after birth and into adulthood \u003csup\u003e\u003cspan additionalcitationids=\"CR13 CR14 CR15 CR16\" citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. Although human duct-to-beta cell transdifferentiation has been evidenced by the existence of insulin-expressing cells in the pancreatic duct epithelium, it is a very rare phenomenon that many doubt would be relevant for sufficient beta cell regeneration. However, if its underlying mechanism can be understood, one could potentially manipulate duct-to-beta cell transdifferentiation to a high enough efficiency to replenish functional beta cell mass for the treatment of diabetes. Here we employed a genome-wide CRISPR screen to dissect the mechanism of human duct-to-beta cell transdifferentiation and to identify new therapeutic targets for beta cell mass restoration.\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eA genome-wide CRISPR screen identifies ALDH3B2 as a regulator of human duct-to-beta cell transdifferentiation\u003c/h2\u003e \u003cp\u003eForward genetic screening, the genome-wide CRISPR screening in particular, is a powerful approach to discover novel genes, signaling pathways and the underlying mechanisms of a complex biological phenomenon. Here, we developed a genome-wide CRISPR screening strategy to search for genes that regulate the transdifferentiation of human pancreatic duct cells into beta cells. To ensure efficient gene editing and sufficient cell numbers for the genome-wide screen with sufficient coverage, we chose to use an immortalized pancreatic cell line, PANC-1, a human pancreatic carcinoma cell line of ductal origin that maintains many of the differentiated characteristics of normal mammalian pancreatic ductal epithelium \u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. Our strategy is to use PANC-1 cells solely as a CRISPR screen and target discovery tool, and all the findings from the PANC-1 genetic screen will later be validated and characterized in primary human pancreatic duct cells. We reasoned that insulin expression or insulin promoter activation would be the most direct and the simplest readout for cell transdifferentiation into pancreatic beta-like cells. Therefore, we introduced a reporter construct, Rat insulin promoter 3.1 (\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eR\u003c/span\u003eIP)-\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eE\u003c/span\u003eGFP-\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eP\u003c/span\u003e2A- Blasticidin-S deaminase (\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eB\u003c/span\u003eSD) (referred to as REPB reporter), into the PANC-1 cells via lentiviral transduction to create a REPB-PANC-1 reporter cell line (Fig.\u0026nbsp;1a). Rat insulin promoter (RIP) is known to be active in human beta cells and RIP 3.1 is a modified rat insulin promoter that is believed to have higher efficiency and beta cell specificity \u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e, which could increase the sensitivity of our CRISPR screen. The P2A peptide ensures the co-expression of the EGFP and BSD reporter genes, the EGFP reporter allows to visually and quantitatively monitor insulin promoter activation, and the expression of the BSD reporter gene confers resistance to blasticidin treatment, making it easy to enrich or select insulin promoter activated cells. To validate the REPB reporter construct, we also transduced the NIT-1 mouse beta cell line with the REPB reporter (REPB-NIT-1 cell line). As shown in Extended Data Fig.\u0026nbsp;1a and 1b, the EGFP expression can only be observed in the REPB-NIT-1 cells, but not at all in REPB-PANC-1 cells. The REPB-PANC-1 cells, but not the REPB-NIT-1 cells, were sensitive to blasticidin treatment (data not shown).\u003c/p\u003e \u003cp\u003eTo execute the CRISPR screen, as illustrated in Fig.\u0026nbsp;1a, we transduced the REPB-PANC-1 cells with a human lentiviral genome-wide CRISPR knockout library (GeCKO v2) \u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e that comprises approximately 120,000 guide RNAs (gRNAs) targeting a total of 19,050 genes. We used a low multiplicity of infection (MOI)\u0026thinsp;~\u0026thinsp;0.3 to ensure that most of the cells carry only one mutation. Briefly, approximately 10\u003csup\u003e8\u003c/sup\u003e lentiviral library transduced REPB-PANC-1 cells were treated with low dose of blasticidin (10ug/ml) for 7 days, and then the blasticidin resistant cells were subjected to FACS sorting based on their EGFP intensity (Extended Data Fig.\u0026nbsp;2). Using next generation sequencing (NGS) and bioinformatic analysis, the gRNA profile of the highest EGFP-expressing cells (EGFP \u003csup\u003ehigh\u003c/sup\u003e, blasticidin-resistant) was generated and compared to that of the cells without blasticidin selection and FACS sorting (Extended Table\u0026nbsp;1).\u003c/p\u003e \u003cp\u003eWe identified candidates whose sgRNAs were cooperatively and significantly positively selected in the EGFP\u003csup\u003ehigh\u003c/sup\u003e, blasticidin-resistant REPB-PANC-1 cells vs control cells using MAGeCK\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e,\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e and MAGeCKFlute\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e at the gene level. As shown in Fig.\u0026nbsp;1b from MAGeCKFlute where genes that were at least two-fold up are in red and those at least two-fold down are in blue, there is substantial positive selection. In Fig.\u0026nbsp;1c the half-volcano plot of statistics from MAGeCK-mle illustrates multiple highly significant positively selected hits with robust fold changes, such as ALDH3B2 whose fold change\u0026thinsp;\u0026gt;\u0026thinsp;10\u003csup\u003e5\u003c/sup\u003e (Fig.\u0026nbsp;1c).\u003c/p\u003e \u003cp\u003eWe next considered whether the screening hits were enriched for functional classifications involved in the cell fate change. Top pathways from pathway enrichment analysis of KEGG pathways using MAGeCKFlute are shown in Fig.\u0026nbsp;1d. Glycolysis/gluconeogenesis, inositol phosphate metabolism, beta-alanine metabolism, and starch and sucrose metabolism represent the highest NES (Normalized enrichment scores) value, which indicates that members of those gene sets tend to participate in the cell transdifferentiation. To understand the protein-protein associations including physical and functional interactions between INS, ACLY, and other GFP-positive screening hits from MAGeCKFlute analysis, the STRING pathway analysis was performed (Fig.\u0026nbsp;1e). Many CRISPR screening hits, such as ITPKA, GPI, HK2, ALDH3B2, and ALDH3A1 showing independent connection with INS and ACLY, which recognized as an essential gene during cell transdifferentiation \u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e and metabolic reprogramming \u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e. The screening hits were significantly enriched for shared protein networks, providing additional confidence in the sensitivity to identify interrelated complexes (Fig.\u0026nbsp;1f). Many hits mapped to the functional categories of metabolic pathway and glycolysis/gluconeogenesis, all of them containing both previously unknown regulators of beta cell transdifferentiation and those with known roles in cell transdifferentiation or metabolic reprogramming.\u003c/p\u003e \u003cp\u003eGet together, we then picked the most enriched gRNA of top 8 candidate gene from the list and generated individual mutant PANC-1 cell lines using the corresponding gRNAs identified in our screen. We used quantitative PCR (qPCR) to analyze the expression of endocrine marker genes including insulin (INS), glucagon (GCG) and somatostatin (SST), as well as pancreatic duct cell marker gene keratin 19 (KRT19 or CK19). Several mutant PANC-1 cell lines showed differential expression of the examined marker genes (Fig.\u0026nbsp;1c, 1d and Extended Data Fig.\u0026nbsp;3a and 3b). In particular, PANC-1 cells transduced with a gRNA targeting ALDH3B2 showed the highest INS expression and the lowest KRT19 levels compared to non-targeting control (NTC) gRNA transduced PANC-1 cells. ALDH3B2, also known as ALDH8, is one of 19 members of the human aldehyde dehydrogenase (ALDH) superfamily that converts various types of aldehydes to carboxylic acids \u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. It is well documented that ALDH genes are important regulators of stem cells and cell fate determination \u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e. Another close member in the ALDH family, ALDH1A3, was recently shown to contribute to pancreatic beta cell failure and de-differentiation in type 2 diabetes \u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e. We reasoned that ALDH3B2, as an enzyme, could potentially be an easier therapeutic target for small molecules targeting. Therefore, we prioritized ALDH3B2 for further in-depth validation and characterization.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eLoss-of-function of ALDH3B2 trans-differentiates PANC-1 cells into pancreatic beta-like cells\u003c/h3\u003e\n\u003cp\u003eWe generated an ALDH3B2 mutant PANC-1 cell line by lentiviral transduction of SpCas9 and ALDH3B2 gRNA into PANC-1 cells. Genomic sequencing of the targeted region in ALDH3B2\u003csup\u003emut\u003c/sup\u003e PANC-1 cells revealed that more than 75% of the sequenced ALDH3B2 alleles carried indel mutations (Extended Data Fig.\u0026nbsp;4). Western blot confirmed that ALDH3B2 protein level in the ALDH3B2\u003csup\u003emut\u003c/sup\u003e PANC-1 cells was reduced by ~\u0026thinsp;40% compared to the non-targeting-control (NTC) gRNA lentivirus transduced PANC-1 cells (Extended Data Fig.\u0026nbsp;5a and 5b). To ensure that loss-of-function of ALDH3B2 induces \u003cem\u003ebona fide\u003c/em\u003e cell transdifferentiation and did not only just activate the insulin promoter, we conducted a series of qPCR experiments to examine additional genes characteristic for pancreatic beta cells. We found that the expression of pancreatic endocrine hormones, insulin (INS) and somatostatin (SST), but not glucagon (GCG) or pancreatic polypeptide (PP), were significantly increased in the ALDH3B2 mutant PANC-1 cells (Fig.\u0026nbsp;2a). In addition, the expression of several of key beta cell transcription factors, including \u003cem\u003ePDX1, MAFA, NGN3\u003c/em\u003e and \u003cem\u003ePAX6\u003c/em\u003e, were also significantly upregulated in the ALDH3B2 mutant PANC-1 cells (Fig.\u0026nbsp;2b). We examined additional genes that are critical for beta cell function and found that \u003cem\u003eGLUT1 (SLC2A1), GLUT2 (SLC2A2), GLUCOKINASE (GCK)\u003c/em\u003e, subunits of the ATP-sensitive potassium (K-ATP) \u003cem\u003e(KCNJ11\u003c/em\u003e and \u003cem\u003eABCC8\u003c/em\u003e), \u003cem\u003eCPE\u003c/em\u003e and \u003cem\u003eIA2\u003c/em\u003e were all significantly upregulated in the ALDH3B2 mutant PANC-1 cells (Fig.\u0026nbsp;2c). ALDH3B2 mutant PANC-1 cells had slightly decreased expression of pancreatic duct markers KRT19 and CA2 but not of HNF1B or Sox9 (Fig.\u0026nbsp;2d). Immunofluorescence imaging showed that clusters of ALDH3B2 mutant PANC-1 cells expressed human Insulin (INS) and C-peptide (CPEP), whereas neither insulin and C-peptide were detected in control PANC-1 cells (Fig.\u0026nbsp;2e). The insulin content of the ALDH3B2 mutant PANC-1 was significantly increased compared to control cells (Fig.\u0026nbsp;2f). In addition, using electron microscopy (EM), we found that many of the ALDH3B2 mutant PANC-1 cells had insulin granules (vesicles with halo, arrows in Fig.\u0026nbsp;2g), while no insulin granules were detectable in control -PANC-1 cells (Fig.\u0026nbsp;2g). Intriguingly, the ALDH3B2 mutant PANC-1 cells had intensive endoplasmic reticulum (ER) network (Fig.\u0026nbsp;2g), a characteristic found in pancreatic beta cells but not pancreatic duct cells. Collectively, these studies indicate that loss-of-function mutations in ALDH3B2 in PANC-1 cells not only trigger insulin promoter activation but also precipitate a significant cell fate transformation, shifting from a pancreatic ductal phenotype to a beta-like profile.\u003c/p\u003e \u003cp\u003eTo evaluate whether these trans-differentiated pancreatic beta-like cells were functional, we performed \u003cem\u003ein vitro\u003c/em\u003e \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eg\u003c/span\u003elucose \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003es\u003c/span\u003etimulated \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003ei\u003c/span\u003ensulin \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003es\u003c/span\u003eecretion (GSIS) assay comparing control NTC-PANC-1 cells and transdifferentiated ALDH3B2 mutant PANC-1 cells, and we found that the ALDH3B2 mutant PANC-1 cells can secrete significantly more insulin at baseline (2.8 mM glucose) compared to the control NTC-PANC-1 cells, and also mildly respond to higher glucose (16.7 mM glucose) (Extended Data Fig.\u0026nbsp;6a). We then examined whether the ALDH3B2 mutant PANC-1 cells is also functional \u003cem\u003ein vivo\u003c/em\u003e in diabetic mouse model. We used streptozotocin (STZ) to induce diabetes in NSG mice, where majority of the beta cells were killed by STZ injection (Extended Data Fig.\u0026nbsp;7), and then transplanted the ALDH3B2 mutant or control PANC-1 cells (NTC) subcutaneously into these diabetic mice (Fig.\u0026nbsp;2h, upper panel). Mice transplanted with the ALDH3B2 mutant PANC-1 cells showed significantly decreased daily random blood glucose (Fig.\u0026nbsp;2h, lower panel) and improved glucose tolerance (Fig.\u0026nbsp;2i) compared to mice transplanted with NTC-PANC-1 cells. Notably, when the ALDH3B2 mutant PANC-1 graft was removed at the end of the study, blood glucose increased to the same level as in the control mice, confirming that the blood-glucose-lowering was indeed caused by the transplanted ALDH3B2 mutant PANC-1 cells (Fig.\u0026nbsp;2h, lower panel). Human insulin serum levels were also significantly higher in mice transplanted with ALDH3B2 mutant PANC-1 cells (Fig.\u0026nbsp;2j). Immunofluorescent imaging showed that transplanted ALDH3B2 mutant PANC-1 cells co-expressed PDX1, Insulin, NKX6.1 and C-peptide (Fig.\u0026nbsp;2k, 2l, Extended Data Fig.\u0026nbsp;8a and 8b). We observed that a few cells co-express somatostatin (SST) and Insulin, but no cell expresses glucagon (GCG), or the exocrine cell marker gene amylase (AMY) (Extended Data Fig.\u0026nbsp;8c and 8d). Of all the transplanted ALDH3B2 mutant PANC-1 cells, ~\u0026thinsp;20% were PDX1\u003csup\u003e+\u003c/sup\u003e (Fig.\u0026nbsp;2m) and ~\u0026thinsp;8% were INS\u003csup\u003e+\u003c/sup\u003e or NKX6.1\u003csup\u003e+\u003c/sup\u003e (Fig.\u0026nbsp;2n and 2o). Almost all the INS\u003csup\u003e+\u003c/sup\u003e cells were also NKX6.1\u003csup\u003e+\u003c/sup\u003e, suggesting that trans-differentiated beta-like cells adopted a true beta cell phenotype. It should also be noted that only\u0026thinsp;~\u0026thinsp;45% of the PDX1\u003csup\u003e+\u003c/sup\u003e cells co-expressed insulin (Fig.\u0026nbsp;2p), and we speculate that the PDX1\u003csup\u003e+\u003c/sup\u003e/INS\u003csup\u003e-\u003c/sup\u003e cells may represent pancreatic progenitor-like cells that have yet committed to beta cell fate.\u003c/p\u003e \u003cp\u003eWe employed an inducible shRNA system to ensure that the transdifferentiation of PANC-1 cells into beta-like cells by ALDH3B2 CRISPR knockout was indeed due to the loss-of-function of ALDH3B2 and not caused by off-target effects of the ALDH3B2 gRNA. We generated PANC-1 cell lines carrying a Tet-On inducible ALDH3B2 shRNA or a scrambled control shRNA (Fig.\u0026nbsp;3a, left panel). The ALDH3B2 shRNA PANC-1 cells with Doxycycline treatment showed significantly reduced ALDH3B2 mRNA expression (Fig.\u0026nbsp;3a right panel) and protein level (Extended Data Fig.\u0026nbsp;5c and 5d) after doxycycline (dox) treatment. Similar to the ALDH3B2 CRISPR mutant PANC-1 cells, knock-down of ALDH3B2 by shRNA also trans-differentiated PANC-1 cells into beta-like cells. A series of qPCR experiments showed that the expression of key beta cell transcription factors including \u003cem\u003ePDX1\u003c/em\u003e, \u003cem\u003eMAFA, NGN3, NEUROD\u003c/em\u003e and \u003cem\u003ePAX6\u003c/em\u003e (Fig.\u0026nbsp;3b), endocrine hormone insulin (\u003cem\u003eINS\u003c/em\u003e) and somatostatin (\u003cem\u003eSST\u003c/em\u003e) (Fig.\u0026nbsp;3C), and beta cell function related genes (\u003cem\u003eSLC2A2, GCK, KCNJ11\u003c/em\u003e and \u003cem\u003eABCC8\u003c/em\u003e) were significantly increased (Fig.\u0026nbsp;3d), whereas the expression of pancreatic duct cell marker genes (\u003cem\u003eKRT19, CA2\u003c/em\u003e and \u003cem\u003eSOX9\u003c/em\u003e) were reduced (Fig.\u0026nbsp;3e). Human insulin could also be detected in shALDH3B2 PANC-1 cells (+\u0026thinsp;Dox) but not in shControl PANC-1 cells or in shALDH3B2 PANC-1 cells (+\u0026thinsp;Dox) by immunofluorescence (Fig.\u0026nbsp;3f). These analyses confirmed that loss-of-function of ALDH3B2 by CRISPR targeting or shRNA silencing allowed PANC-1 cells to transdifferentiate and adopt a beta-like cell fate.\u003c/p\u003e \u003cp\u003e \u003cb\u003eLoss-of-function of ALDH3B2 transdifferentiates human primary pancreatic duct cells into beta-like cells.\u003c/b\u003e \u003c/p\u003e \u003cp\u003eNext, we tested whether loss-of-function of ALDH3B2 was also able to transdifferentiate \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eh\u003c/span\u003euman \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003ep\u003c/span\u003erimary \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003ep\u003c/span\u003eancreatic \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003ed\u003c/span\u003euct (HPPD) cells into beta-like cells. HPPD cells were isolated and affinity-purified from human donor islet-depleted pancreatic acinar tissue from Integrated Islet Distribution Program (IIDP) \u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e. qPCR analyses confirmed lack of insulin expression (Fig.\u0026nbsp;4a) and high expression of the pancreatic duct markers \u003cem\u003eKRT19\u003c/em\u003e (Fig.\u0026nbsp;4b) in the purified HPPD cells compared to primary human islets. Interestingly, we observed that ALDH3B2 expression levels are markedly lower in human islets compared to pancreatic duct cells (Fig.\u0026nbsp;4c). This differential expression pattern aligns with our results where the mutation of ALDH3B2 in human pancreatic duct cells promotes their transdifferentiation into beta-like cells. These findings suggest that the reduction of ALDH3B2 could be a critical step in the cellular reprogramming process leading to a beta-cell phenotype. Purified HPPD cells were transduced with lentiviruses carrying SpCas9 and either ALDH3B2 gRNA (ALDH3B2\u003csup\u003emut\u003c/sup\u003e -HPPD) or a non-targeting control gRNA (NTC-HPPD). qPCR analysis demonstrated that ALDH3B2\u003csup\u003emut\u003c/sup\u003e -HPPD cells had significantly higher expression of key beta cell transcription factors (\u003cem\u003ePDX1\u003c/em\u003e and \u003cem\u003eMAFA\u003c/em\u003e) (Fig.\u0026nbsp;4d), endocrine hormone insulin (\u003cem\u003eINS\u003c/em\u003e) and somatostatin (\u003cem\u003eSST\u003c/em\u003e) (Fig.\u0026nbsp;4e) and beta cell function-related genes (\u003cem\u003eGCK, SLC2A1, SLC2A2, KCNJ11\u003c/em\u003e and \u003cem\u003eCPE\u003c/em\u003e) (Fig.\u0026nbsp;4f). The expression of several pancreatic duct cell marker genes was either unchanged (\u003cem\u003eKRT19\u003c/em\u003e and \u003cem\u003eHNF1B\u003c/em\u003e) or slightly reduced (\u003cem\u003eSOX9\u003c/em\u003e) (Fig.\u0026nbsp;4g). Using immunofluorescent imaging, we found that a fraction of HPPD-ALDH3B2\u003csup\u003emut\u003c/sup\u003e cells expressed Insulin while still retaining CK19 expression, a possible signature of newly transdifferentiated beta cells from pancreatic duct cells, whereas no insulin expression could be detected in HPPD-NTC cells (Fig.\u0026nbsp;4h). Furthermore, we also examined whether the ALDH3B2\u003csup\u003emut\u003c/sup\u003e -HPPD cells have insulin granules using electron microscopy (EM). Although not as many insulin granules as in primary human beta cells, the ALDH3B2\u003csup\u003emut\u003c/sup\u003e -HPPD cells do have significant amount of mature insulin granules, while no insulin granules were detectable in the control NTC-HPPD cells (Fig.\u0026nbsp;4i).\u003c/p\u003e \u003cp\u003eWe performed \u003cem\u003ein vitro\u003c/em\u003e GSIS assay to evaluate the function of the ALDH3B2\u003csup\u003emut\u003c/sup\u003e -HPPD cells, and found that compared to control NTC-HPPD cells, the ALDH3B2\u003csup\u003emut\u003c/sup\u003e -HPPD cells can secrete significantly more insulin and mildly respond to high glucose (Extended Data Fig.\u0026nbsp;6b). Although the level of insulin secretion from the ALDH3B2\u003csup\u003emut\u003c/sup\u003e -HPPD cells is still much lower than human islets (Extended Data Fig.\u0026nbsp;6b), we suspected that it is due to relatively low efficacy of the transdifferentiation and immaturity of the trandifferentiated beta-like cells, especially in the \u003cem\u003ein vitro\u003c/em\u003e experiment setting. We then transplanted ALDH3B2\u003csup\u003emut\u003c/sup\u003e -HPPD or NTC-HPPD cells under the kidney capsule of STZ induced diabetic NSG mice, and monitored their blood glucose over time (Fig.\u0026nbsp;5a, upper panel). Mice transplanted with HPPD-ALDH3B2\u003csup\u003emut\u003c/sup\u003e cells had significantly lower blood glucose compared to NTC-HPPD cells transplanted mice (Fig.\u0026nbsp;5a, lower panel). When the ALDH3B2 mutant HPPD grafts were removed at 56 days post-transplantation, blood glucose increased to similar level as in the control HPPD transplanted mice, suggesting that the blood-glucose-lowering effect was indeed conferred by the transplanted ALDH3B2 mutant HPPD cells (Fig.\u0026nbsp;5a, lower panel). Importantly, transplanted ALDH3B2\u003csup\u003emut\u003c/sup\u003e -HPPD cells secreted human insulin in response to glucose challenge. We detected a significantly increase in human serum insulin 5 minutes after glucose injection (both at 1 week and 3 weeks post-transplantation). No such response was observed in mice transplanted with NTC-HPPD cells or in non-transplanted control NSG mice (Fig.\u0026nbsp;5b and 5c). Immunofluorescent analysis revealed that Insulin\u003csup\u003e+\u003c/sup\u003e cells, C-Peptide\u003csup\u003e+\u003c/sup\u003e cells, PDX1\u003csup\u003e+\u003c/sup\u003e cells, NKX6.1\u003csup\u003e+\u003c/sup\u003e cells were only observed in mice transplanted with ALDH3B2\u003csup\u003emut\u003c/sup\u003e -HPPD cells but not with NTC-HPPD cells (Fig.\u0026nbsp;5d, 5e, Extended Data Fig.\u0026nbsp;9a-d). Almost all of the transplanted ALDH3B2\u003csup\u003emut\u003c/sup\u003e -HPPD cells expressed pancreatic duct marker genes CK19 and SOX9 but not glucagon (GCG), Somatostatin (SST) or Amylase (AMY) (Extended Data Fig.\u0026nbsp;9e and 9f). Approximately 40% of the transplanted cells were successfully transduced with the NTC or ALDH3B2 gRNA lentivirus (shown by quantification of the percentage of Cas9 (Flag-tagged)\u003csup\u003e+\u003c/sup\u003e/CK19\u003csup\u003e+\u003c/sup\u003e cells, Fig.\u0026nbsp;5d and 5f), and among all the gRNA lentivirus infected pancreatic duct cells, ~\u0026thinsp;15% of ALDH3B2\u003csup\u003emut\u003c/sup\u003e -HPPD cells expressed insulin (Fig.\u0026nbsp;5g). Interestingly, we found that the majority of the pancreatic duct cells infected with ALDH3B2 gRNA lentivirus co-expressed PDX1 and CK19 (Fig.\u0026nbsp;5e and 5h), and approximately 12% of the PDX1\u003csup\u003e+\u003c/sup\u003e cells co-expressed insulin (Fig.\u0026nbsp;5i). Co-expression of PDX1 and CK19 is a signature of pancreatic progenitor cells \u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e, and we postulate that ALDH3B2 loss-of-function may cause the de-differentiation of mature duct cells into pancreatic progenitor-like cells, a portion of which then subsequently differentiate into beta-like cells.\u003c/p\u003e\n\u003ch3\u003eLoss of ALDH3B2 function in pancreatic duct cells causes epigenetic changes\u003c/h3\u003e\n\u003cp\u003eWe found that ALDH3B2 loss-of-function allowed pancreatic duct cells to adopt a beta-like cell profile, and we next asked if transdifferentiation was associated with epigenetic changes. To this end, we analyzed DNA methylation in the human insulin gene region by bisulfite conversion assay. DNA methylation was significantly reduced in ALDH3B2 mutant PANC-1 cells compared to control NTC-PANC-1 cells at the +\u0026thinsp;63, +127 and +\u0026thinsp;139 positions of the human insulin locus, which are three well-characterized DNA methylation sites in the insulin locus \u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e (Fig.\u0026nbsp;6a and 6b). For DNA methylation analysis of primary human pancreatic duct cells, we included primary human islets for comparison. Again, ALDH3B2 mutation significantly reduced DNA methylation at the same three sites in the insulin gene locus (Fig.\u0026nbsp;6c and 6d). ALDH3B2 mutation did not reduce the DNA methylation to the level observed in primary islets. This might be due to the fact that only a fraction (8\u0026ndash;15%) of pancreatic duct cells were transdifferentiated into beta-like cells with ALDH3B2 mutation. Overall, DNA methylation analyses suggest that loss-of-function of ALDH3B2 caused epigenetic changes in the pancreatic duct cells to induce a stable cell fate change into pancreatic beta-like cells.\u003c/p\u003e \u003cp\u003e \u003cb\u003eALDH3B2 loss-of-function in human pancreatic duct cells induces heterogeneous beta-like cell populations with overlapping endocrine and duct cell identity\u003c/b\u003e \u003c/p\u003e \u003cp\u003eTo investigate the characteristics of transdifferentiated beta-like cells in more detail, we performed 3\u0026rsquo; gene expression single cell RNA sequencing of control or ALDH3B2 mutant HPPD cells and identified 13 unique cell cluster (Fig.\u0026nbsp;7a). Whereas the control NTC-HPPD condition only shows a few insulin positive cells in cluster 10, which could represent rare spontaneous transdifferentiated beta-like cells from duct cells, ALDH3B2 mutant HPPD cells develop insulin-expressing cells in various cell clusters and most prominent in cluster 5, 10, 11 and 12 (Fig.\u0026nbsp;7b-d). The relative proportion of beta-like-cell-containing cluster 5, 10, 11 and 12 is largely increased in the ALDH3B2 mutant HPPD cells compared to control NTD-HPPD cells (Extended Data Fig.\u0026nbsp;10a). The percentage of insulin-expressing cells in ALDH3B2 mutant HPPD cells is about 18.1% but only 0.6% in NTC HPPD cells, and about 93% of all cells in both conditions still retain the expression of duct cell marker KRT19 (Extended Data Fig.\u0026nbsp;10b). The majority of insulin-positive beta-like cells also show significantly higher expression of other beta cell marker genes such as CHGA and TTR but also duct cell identity marker genes KRT17, 19, and 23 (Fig.\u0026nbsp;7e). Differential gene expression analysis of insulin high-expressing cells (relative intensity\u0026thinsp;\u0026gt;\u0026thinsp;1) compared to insulin low-expressing cells (relative intensity\u0026thinsp;\u0026lt;\u0026thinsp;1) within the ALDH3B2 mutant HPPD cell condition showed significant upregulation of key beta cell marker genes such as CHGA, IAPP, SCGN, SCG3 and SCG5 (Fig.\u0026nbsp;7f). Of note, \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eg\u003c/span\u003eene \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003es\u003c/span\u003eet \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003ee\u003c/span\u003enrichment \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003ea\u003c/span\u003enalysis (GSEA) confirmed the upregulation of key beta cell-specific gene sets related to peptide hormone metabolism, regulation of insulin secretion, and insulin processing (Fig.\u0026nbsp;7g). Re-analysis of insulin high-expressing cells in the ALDH3B2 mutant HPPD cells also identified heterogeneous cell populations with high beta cell identity (high INS/CHGA/IAPP co-expression), and polyhormonal cells (co-expression of INS/GCG/PPY), and endocrine progenitor-like cells (co-expression of INS and PAX6) (Extended Data Fig.\u0026nbsp;11). To further characterize the cell identity of insulin-expressing cells we compared the average gene expression of several marker genes specific for beta cells and other endocrine cells, endocrine progenitor cells, and duct cells \u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e within all 13 cell clusters (Fig.\u0026nbsp;7h and 7i). In cluster 10 of the control condition, we detected insulin expression in about 40% of all cells, and the cells in this cluster show a low ductal cell-specific expression profile but high expression of endocrine progenitor cell marker such as PAX6 and INSM1 in 60\u0026ndash;80% of all cells, indicating that the majority of cells in cluster 10 may represent endocrine progenitor-like cells. Interestingly, ALDH3B2 loss-of-function in cluster 10 shifts cell identity towards beta-like cells, shown by downregulation of almost all endocrine progenitor marker genes and sustained expression of INS and CHGA. Cluster 11 and 12 show altered duct cell identity compared to cluster 0 to 9 even in the control condition and seem to have a higher potential of beta-like cell transdifferentiation following loss-of-function of ALDH3B2. In comparison, cells in cluster 5 demonstrates strong duct cell identity but admit high potential for beta-like cell transdifferentiation as well.\u003c/p\u003e \u003cp\u003eTo better understand why cells in cluster 5 and 12 have a higher chance to transdifferentiate into beta-like cells we performed trajectory inference analysis to identify the potential starting point of transdifferentiation (from low to high insulin expression, Extended Data Fig.\u0026nbsp;10c). Beta-like cells in cluster 5 may originated from cluster 2 and cluster 5 itself. Re-analysis of cluster 5 revealed several insulin low-expressing cell cluster as possible starting points. However, insulin positive cells in cluster 5 show significant upregulation of genes involved in energetic processes such as oxidative phosphorylation, aerobic respiration, and ATP synthesis may representing important prerequisites for beta-like cell transdifferentiation (Extended Data Fig.\u0026nbsp;10d and 10e). Cluster 3 may represent the originating cluster for beta-like cells in cluster 12. Differential gene expression analysis comparing cluster 2 and 3 (high potential for beta-like cell transdifferentiation) to cluster 4 (low potential of transdifferentiation) reveals that upregulation of genes important for translation and oxidative phosphorylation (cluster 2), enrichment of small GTPases RAC1/RHO (cluster 3), and elevated glycolysis (cluster 2 and 3) may favor beta-like cell transdifferentiation mediated by loss of function of ALDH3B2 (Extended Data Fig.\u0026nbsp;10f-k). In summary, ALDH3B2 loss-of-function may drive transdifferentation of pancreatic duct cells partially through duct cell-derived endocrine progenitor-like stage, and then into beta-like cells that still keep partial duct cell identity. Elevation of energy metabolism such as oxidative phosphorylation and glycolysis may be a key step for pancreatic duct cells to transdifferentiate into beta-like cells.\u003c/p\u003e \u003cp\u003e \u003cb\u003eDynamic change in ALDH3B2 mutant human primary pancreatic duct cells transdifferentiation is revealed by RNA Velocity and PAGA Trajectory Analysis.\u003c/b\u003e \u003c/p\u003e \u003cp\u003eWe next analyzed RNA velocity of our single cell data using scVelo \u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e to investigate a possible transition among duct- and β-cell subclusters. scVelo identified 10 unique cell clusters, which are shown in a UMAP plot with streamlined velocities in Fig.\u0026nbsp;8a. The plot demonstrated a branching pattern emerging from cluster 2 towards other subclusters, although the specific paths were challenging to enumerate. So we next applied PAGA graph abstraction, which has been benchmarked as a top performing method for trajectory inference (Fig.\u0026nbsp;8b). It provides a graph-like map of the data topology with weighted edges corresponding to the connectivity between two clusters. We could see that the INS-expressing cells in cluster 5 were developed from cells in cluster 2, and then further developed into INS-expressing cells in cluster 9. The trajectory of INS-negative cells in the cluster also exhibited clear directionality, such as clusters (2_4), (2_1_6), and (2_8_0, 7, or 3).\u003c/p\u003e \u003cp\u003eThe transcriptional dynamic model enabled the recovery of latent time associated with cellular processes, representing an internal clock for cells undergoing differentiation based on transcriptional dynamics. The velocity latent time highlighted a consistent developmental order, with all clusters progressing from cluster 2 to subsequent subclusters (Fig.\u0026nbsp;8c). Using velocity length to characterize the speed of transition or differentiation, we observed that clusters 2 and 1 exhibited significantly higher lengths, indicative of robust splicing activity and overall velocity confidence across ductal and β-cell subpopulations. (Fig.\u0026nbsp;8d). Furthermore, we evaluated the velocity of various representative genes associated with β- and duct-cells, revealing that the expression levels and velocity did not consistently align. Notably, INS expression peaked in the transitional clusters (2_5_9), while RNA velocity was predominantly higher across most clusters, excluding clusters 6, 7, and parts of 9. Selectively elevated INS expression was observed in β-cells, whereas a modest increase in RNA velocity was noted in the ductal subclusters. Subsequent analysis of the β-cell gene TRPM3 demonstrated a concordance between its expression pattern and RNA velocity in cluster 9. Similarly, examination of the duct cell-specific gene KRT19 revealed heightened expression in cluster 1 along with high RNA velocity, contrasting with the observation that the expression pattern of the duct cell marker CFTR did not align with RNA velocity in cluster 4 (Fig.\u0026nbsp;8e). Collectively, these findings lend support to the hypothesis that cells within cluster 5 may undergo a transition towards a more β-cell-like phenotype in response to ALDH3B2 loss-of-function.\u003c/p\u003e \u003cp\u003eExamining the top 40 driver genes within the relevant lineages (clusters 2_5_9), we identified clusters of genes exhibiting specific temporal abundance in distinct cell types. Heatmaps illustrated the primary occurrences of spliced counts for the top-ranked dynamic genes in clusters 2, 5, and 9, organized by the latent time of cells (Fig.\u0026nbsp;8f). Notably, genes such as ITGB1 \u003csup\u003e35\u003c/sup\u003e, NEDD9 \u003csup\u003e36\u003c/sup\u003e, HIF1A \u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e, and SERPINE1 \u003csup\u003e38\u003c/sup\u003e, known for their involvement in transdifferentiation processes linked to TGF-beta signaling, were prominently featured. Furthermore, our analysis revealed a significant upsurge in the expression levels of genes like ATP2A3, CACNA2D1, C2CD4A, and SCGN during the later phase of latent time within a brief duration. This observation suggests a pivotal role for Ca\u003csup\u003e2+\u003c/sup\u003e signaling in the transdifferentiation of duct cells into functional beta cells. Intriguingly, we also noted a dynamic reduction in ALDH1A3 expression during this transdifferentiation process. Additionally, glycolysis-related genes such as LDHA and HIF1A were implicated in the transdifferentiation process, further underscoring the complexity and multifaceted nature of the cellular transformations taking place.\u003c/p\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eIn this study, we have described the first unbiased and genome-wide CRISPR screen in search for genes that regulate the transdifferentiation of human pancreatic duct cells into insulin-producing beta-like cells. We show that loss-of-function of a single gene, ALDH3B2, in human pancreatic duct cells is sufficient to drive them towards a beta-like cell fate. Although the pancreatic duct-to-beta cell transdifferentiation in human has been observed, as evidenced by the existence of INS\u003csup\u003e+\u003c/sup\u003e/CK19\u003csup\u003e+\u003c/sup\u003e cells in pancreatic ductal epithelium, it is still a relatively rare event with the percentage of INS\u003csup\u003e+\u003c/sup\u003e pancreatic duct cells estimated at ~\u0026thinsp;1% \u003csup\u003e39\u003c/sup\u003e. Here we show that disruption of ALDH3B2 can drive transdifferentiation of primary human pancreatic duct cells into beta-like cells with an efficiency of ~\u0026thinsp;15%. This significant result carries potential for the development of a therapeutic intervention that could promote pancreatic duct-to-beta cell transdifferentiation for human beta cell mass replenishment.\u003c/p\u003e \u003cp\u003eWe demonstrated that the trans-differentiated human pancreatic beta-like cells are functional, responsive to glucose challenge \u003cem\u003ein vivo\u003c/em\u003e, and able to significantly lower blood glucose in diabetic animal models. Notably, neither ALDH3B2\u003csup\u003emut\u003c/sup\u003e PANC-1 cells nor HPPD cells were able to lower blood glucose to euglycemic levels in our studies, and this could be due to two reasons: (1) The transplanted beta-like cells number was not enough due to experimental limitation, or (2) the glucose sensitivity or set-point of transdifferentiated beta-like cells may be different from that of true pancreatic beta cells. In a separate experiment of transplantation of ALDH3B2\u003csup\u003emut\u003c/sup\u003e-HPPD cells into euglycemic NSG mice, we performed glucose tolerance (GTT) measurements. The peak glucose level in this experiment was approximately 250mg/dL, but transplanted ALDH3B2\u003csup\u003emut\u003c/sup\u003e-HPPD cells were still able to improve the glucose tolerance (Extended Data Fig.\u0026nbsp;12a) and secrete human insulin in response to glucose challenge (Extended Data Fig.\u0026nbsp;12b and 7c). Based on these data, it seems unlikely that transdifferentiated beta-like cells had a higher glucose set-point than primary pancreatic beta cells. We may be able to lower the blood glucose further in diabetic mice if a larger number of transdifferentiated cells were transplanted. To this end, we will need to improve lentiviral transduction efficiency and perhaps identify more effective gRNA sequences that target ALDH3B2 gene.\u003c/p\u003e \u003cp\u003eAt first glance in comparison with primary human islets, the ALDH3B2 mutation transdifferentiated beta-like cells seem to have only modest expression of key beta cell markers (Fig.\u0026nbsp;4d-g), ability of lowering blood glucose and secrete human insulin in diabetic mice (Fig.\u0026nbsp;5a-c) and changes in DNA methylation status on insulin promoter (Fig.\u0026nbsp;6c and 6d). However, all these measurements and characterization were done on a mixed cell population with only less than 15% of the cells being transdifferentiated beta-like cells. In theory, the actual changes in each transdifferentiated beta-like cell should be more dramatic than what our experimental data showed. Our findings of immunofluorescent staining for insulin and key beta cell transcription factors and insulin granules with abundant ER ultrastructurally in the transdifferentiated beta-like cells strongly support the conclusion that loss-of-function of ALDH3B2 does induce \u003cem\u003ebona fide\u003c/em\u003e cell transdifferentiation from human pancreatic duct cells into functional beta-like cells.\u003c/p\u003e \u003cp\u003eIt is interesting that the ALDH family is often considered as a stem cell or progenitor marker. High aldehyde dehydrogenase activity (using Aldefluor assay \u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e,\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e) has been widely used to identify adult stem cells or progenitor cells in various tissues/organs, including hematopoietic stem cells (HSC) \u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e, neuronal progenitor cells (NPC) \u003csup\u003e\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e,\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e or potential pancreatic progenitor cells \u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e. However, the Aldefluor assay cannot distinguish between the activity of different aldehyde dehydrogenases, so different ALDH members may be marking stem/progenitor cells in different tissues/organs. The function of ALDH3B2 is not yet well understood. One rodent study reported that ALDH3B2 localizes to lipid droplets in cells and catalyzes the conversion of long-chain fatty aldehydes into long-chain fatty acids \u003csup\u003e\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e. How this function would impact pancreatic duct cell transdifferentiation into beta-like cells is unclear and warrants further studies. Notwithstanding, the discovery of ALDH3B2 as a regulator of pancreatic duct-to-beta cell transdifferentiation provides a novel therapeutic target for pancreatic beta cell mass restoration. Endogenous pancreatic duct cells could potentially be targeted by gene-editing to mutate ALDH3B2 and induce transdifferentiation. Alternatively, ALDH3B2 enzymatic activity may be targeted by small molecules inhibitors to achieve the effects similar to those we observed using genetic disruption of the ALDH3B2 gene. The ALDH3B2 enzymatic assay established by Kitamura and colleagues \u003csup\u003e\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e using long chain fatty aldehyde as substrate could be a useful chemical screen platform for potential new drug discovery. In support of this approach, the broad Aldehyde dehydrogenase (ALDH) inhibitors diethylaminobenzaldehyde (DEAB) and Disulfiram (DSF) have been shown to promote beta cell differentiation in zebrafish and in PANC-1 cells \u003csup\u003e\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e, where it is possible that DEAB and DSF elicit their effect through inhibiting ALDH3B2. It will be of great importance to find out which aldehydes are unique substrates for ALDH3B2 and search for ALDH3B2-specific inhibitors, since mutation of the close member ALDH3A1 does not have the same effect (Fig.\u0026nbsp;1c and 1d) and another close member, ALDH1A3, was shown to be involved in the de-differentiation of pancreatic beta cells in Type 2 diabetes patients \u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eGiven the interest in understanding the mechanisms driving beta cell transdifferentiation induced by ALDH3B2 loss-of-function, our primary focus has been on the progression of the INS\u003csup\u003e+\u003c/sup\u003e cluster. By further examining the subcluster of INS\u003csup\u003e+\u003c/sup\u003e cells that emerge when incorporating ALDH3B2 mutant datasets, we can visually depict the pathways in a coherent manner (refer to Fig.\u0026nbsp;8b and 8e). This analysis sheds light on a specific set of connections along the 2_5_9 differentiation axis, highlighting the transdifferentiation from pancreatic duct cells to beta cells. This dynamic model enhances our ability to interpret velocity and identify driver genes more effectively (see Fig.\u0026nbsp;8f) by exploring correlations between gene kinetics and expression patterns across cells. Notably, we observe that the key potential driver genes are associated with TGF-beta signaling \u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e,\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e, glycolysis \u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e, and Ca\u003csup\u003e2+\u003c/sup\u003e signaling pathways \u003csup\u003e\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e. These findings are consistent with previous data from other cell types and may serve as valuable insights for future investigations focused on pancreatic duct-to-beta cell transdifferentiation.\u003c/p\u003e \u003cp\u003eOur study identifying ALDH3B2 as a regulator of pancreatic duct-to-beta cell transdifferentiation was conducted using a human pancreatic duct cell line and primary human pancreatic duct cells. Given the difficulties of translating rodent studies into human in the beta cell regeneration research field, the findings presented here in human cells are directly relevant and have clear potential for the development of human diabetes therapeutics.\u003c/p\u003e "},{"header":"MATERIALS AND METHODS","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eMice\u003c/h2\u003e \u003cp\u003eNSG (NOD.Cg-Prkdc\u003csup\u003escid\u003c/sup\u003e Il2rg\u003csup\u003etm1Wjl\u003c/sup\u003e/SzJ) mice were purchased from the Jackson Laboratory (Bar Harbor, ME). Animals were housed in pathogen-free facilities at the Joslin Diabetes Center and all experimental procedures were approved and performed in accordance with institutional guidelines and regulations.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eREPB reporter construction and REPB PANC-1 cell generation\u003c/h3\u003e\n\u003cp\u003eThe REPB reporter lentivirus vector was constructed by assembling Rat insulin promoter, RIP3.1 promoter \u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e, EGFP and Blasticidin-S deaminase (BSD). The EGFP and BSD genes are fused together with P2A peptide. The REPB reporter lentivirus was used to infect PANC-1 cells (ATCC #CRL-1469), and the infected PANC-1 cells are then single-cell sorted by FACS. PANC-1 clones with confirmed REPB reporter genome incorporation were used in the genome-wide CRISPR screen.\u003c/p\u003e\n\u003ch3\u003eCRISPR GeCKO library screen\u003c/h3\u003e\n\u003cp\u003eThe human GeCKO-v2 (Genome-Scale CRISPR Knock-Out) lentiviral pooled library was obtained from Addgene (Addgene, # 1000000048) and was prepared as previously described \u003csup\u003e\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/sup\u003e. 100\u0026nbsp;million REPB PANC-1 cells were infected with human GeCKO CRISPR lentiviral library at MOI of 0.3, and then subsequently selected with puromycin (2.5 \u0026micro;g/ml) at day 3 post lentiviral infection. After cells recover from puromycin selection, 20\u0026nbsp;million cells were collected as baseline control (CON-1 and CON-2), the rest of the cells were further selected with blasticidin (10 \u0026micro;g/ml) for 7 days. The blasticidin-resistant cells were allowed to grow back to full confluence and then subjected to FACS sorting on their EGFP intensity. Cell population with the highest EGFP intensity were collected as experiment group (EXP-1 and EXP-2). Genomic DNA was extracted from the cells (Quick-gDNA midiprep kit, Zymo Research), the NGS (Next Generation Sequencing) libraries were prepared as previously described \u003csup\u003e\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e, and then subjected to NGS sequencing analysis (Novogene, CA). The gRNA sequences from the NGS sequencing data were extracted using standard bioinformatics methods, and the read count of gRNAs were calculated as Count Per Million (CPM).\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eAnalysis of pooled CRISPR screen\u003c/h2\u003e \u003cp\u003eTo identify enriched genes based on sgRNAs in this CRISPR screen, we used the MAGeCK algorithm \u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e,\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e in R. We used MAGeCK\u0026rsquo;s \u0026lsquo;MLE\u0026rsquo; subcommand for maximum likelihood estimation of gene essentialities using the TMM normalized counts \u003csup\u003e\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e\u003c/sup\u003e at the gene level. This was followed by applying MAGeCKFlute\u0026rsquo;s\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e function FluteMLE with the argument to incorporate DepMap genes \u003csup\u003e\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u003c/sup\u003e in normalization, as recommended for human data, which produced pathway enrichment analysis. Among screening hits from MAGeCKFlute, protein connectivity networks based on physical and functional interactions were identified using STRING v12 \u003csup\u003e54\u003c/sup\u003e, where only interactions with a medium confidence score of \u0026ge;\u0026thinsp;0.4 were selected.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eCell lines\u003c/h2\u003e \u003cp\u003ePANC-1 (#CRL-1469) and 293FT (#R7007) cell lines were obtained from ATCC and Thermo Fisher Scientific, respectively. Cells were maintained in DMEM (Gibco, 10313039), supplemented with 10% fetal bovine serum (FBS, Gibco), L-alanyl-L-glutamine (Gibco) and penicillin/streptomycin (Corning), in a 37\u003csup\u003eo\u003c/sup\u003e C incubator with 5% CO\u003csub\u003e2\u003c/sub\u003e. To generate non-targeting control (NTC) and ALDH3B2\u003csup\u003emut\u003c/sup\u003e PANC-1 cells, non-targeting (NT) gRNA (5\u0026rsquo;- GCTTTCACGGAGGTTCGACG-3\u0026rsquo;) or ALDH3B2 gRNA (5\u0026rsquo;-GCCCTCCTCACCTGCGGCGA-3\u0026rsquo;, HGLibA_01571) oligos were cloned into LentiCRISPR-v2 vector. Wild type PANC-1 cells were then transduced by NTC or ALDH3B2 gRNA- containing lentivirus, and subsequently selected by puromycin treatment. Indel mutation in ALDH3B2\u003csup\u003emut\u003c/sup\u003e cells was confirmed by deep sequencing analysis (MGH DNA Core Facility, Cambridge, MA). All plasmid sequences were verified by Sanger sequencing before transduction and transfection. To generate shControl and shALDH3B2 PANC-1 cells, a scrambled or ALDH3B2-targeting shRNA was cloned into FH1t(INSR)UTG-GFP vector (A gift from Dr. Stephan Kissler). shControl and shALDH3B2 lentiviruses were then used to infect PANC-1 cells and the cells were selected by puromycin treatment.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003ePANC-1 cells transplantation studies\u003c/h2\u003e \u003cp\u003eExperimental diabetes was induced in 8-week-old NSG male mice by intraperitoneal injection of streptozotocin (STZ) (40 mg STZ/kg body weight for five consecutive days). Animals were considered diabetic only if morning-fed blood glucose exceeded 350 mg/dl. Three days after STZ injection, ~10\u003csup\u003e7\u003c/sup\u003e PANC-1 cells (carry NT and ALDH3B2\u003csup\u003emut\u003c/sup\u003e) were transplanted subcutaneously into each diabetic NSG mouse. Blood glucose was monitored every 3\u0026ndash;4 days. Six weeks post cells transplantation, Intraperitoneal glucose tolerance test (IPGTT) was performed. Mice were fasted for 16 hours. The plasma glucose levels of the mice before (baseline) or 15, 30, 60, 90, and 120 minutes after intraperitoneal injection of 2 mg/g body weight glucose were recorded by a Glucose Meter. At day 52 post cells transplantation, grafts were surgically removed from NSG mice, and the blood glucose was measured 4 days later.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eHuman pancreatic ductal cell isolation and purification\u003c/h2\u003e \u003cp\u003eHuman primary pancreatic ductal cells isolation was performed as previously described \u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e,\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e\u003c/sup\u003e. In brief, \u0026ldquo;Human acinar tissue\u0026rdquo; from Integrated Islet Distribution Program (IIDP) (Donor information is summarized in Extended Table\u0026nbsp;2) was washed 2 times with PBS, and then incubated with Trypsin solution (1.5ml 0.25% Trypsin in 20ml PBS) shaking at 37\u0026deg;C for 15 min. Dispersed cells were centrifuged at 1,000 rpm for 5 min, the supernatant was aspirated, and then the pellets were resuspended with mouse antihuman CA19-9 antibody (Invitrogen; clone 116-NS-19-9) in 2 ml PBS solution. After 15 min incubation at 4\u0026deg;C, the cell suspension was mixed with 10ml PBS solution (375mg EDTA and 2.5g BSA in 500ml PBS) gently. Tubes were centrifuged at 1,000 rpm for 5 min, supernatant was aspirated, 250 \u0026micro;l/tube goat anti mouse IgG microbeads (Miltenyi Biotec) in PBS solution were added, and pellets were mixed. After 20 min incubation at 4\u0026deg;C, wash with PBS solution 2 times. Tubes were centrifuged at 1,000 rpm for 5 min, and pellets were resuspended in 20 ml cold PBS solution and passed through 40 m cell strainers to remove newly formed clumps of cells. MACS magnetic LS separation columns (Miltenyi Biotec) were prepared according to the manufacturer\u0026rsquo;s instructions.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003ePreparation and transplantation of primary human pancreatic duct cells\u003c/h2\u003e \u003cp\u003eHuman primary pancreatic ductal cells isolation was performed as above described. Purified human primary pancreatic ductal cells (HPPD) were immediately cultured in a low-attachment plate in RMPI DMEM/F12 medium (Gibco), supplemented with 10% FBS and penicillin/streptomycin. Lentivirus encoding a NT or ALDH3B2 gRNA together with Cas9 endonuclease was added to the culture media for overnight infection. The next day, HPPD cells were washed with culture media twice and ~\u0026thinsp;10\u003csup\u003e7\u003c/sup\u003e cells were transplanted under the left kidney capsule of 8-week-old of STZ induced diabetic male NSG mice. Graft recipients were left to recover from surgery for three weeks. At day 56 post-transplantation, the kidney transplanted with grafts were surgically removed for gene expression analysis by immunofluorescence. The final blood glucose measurement of NSG mice was done 4 days later.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eQuantitative real-time PCR (qPCR)\u003c/h2\u003e \u003cp\u003eCells or grafts were treated with TRIzol (Thermo Fisher Scientific) for RNA extraction following the manufacturer\u0026rsquo;s protocol. Purified RNA was reverse-transcribed into cDNA using the SuperScript IV first-strand synthesis kit (Invitrogen). INS (Hs00355773_m1), GCG (Hs01031536_m1), SST (Hs00356144_m1) PDX1(Hs00236830_m1), NKX6-1 (Hs00232355_m1), GCK (Hs01564555_m1), SLC2A2 (Hs00165775_m1), SLC2A1 (Hs00892681_m1), CPE (Hs00960598_m1), CA2 (Hs01070108_m1),\u003c/p\u003e \u003cp\u003eKRT19 (Hs00761767_s1), SOX9 (Hs00165814_m1), ALDH3B2 (Hs02511514_s1) and ACTB (Hs01060665_g1) probes for TaqMan assays were purchased from Thermo Fisher Scientific. All Gene expression levels were analyzed by SYBR green PowerUp qPCR assays (Applied Biosystems). Primer sequences used is shown in the following table:\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Taba\" border=\"1\"\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGene\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eForward primer sequence\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eReverse primer sequence\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eACTB\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; CACCATTGGCAATGAGCGGTTC 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo;AGGTCTTTGCGGATGTCCACGT 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eINS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; ACGAGGCTTCTTCTACACACCC 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; TCCACAATGCCACGCTTCTGCA 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePDX1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo;GAAGTCTACCAAAGCTCACGCG 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; GGAACTCCTTCTCCAGCTCTAG 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNKX6.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; CCTATTCGTTGGGGATGACAGAG 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; TCTGTCTCCGAGTCCTGCTTCT 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePAX6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; CTGAGGAATCAGAGAAGACAGGC 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; ATGGAGCCAGATGTGAAGGAGG 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMAFA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; GCTTCAGCAAGGAGGAGGTCAT 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; TCTGGAGTTGGCACTTCTCGCT 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNEUROD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; GGTGCCTTGCTATTCTAAGACGC 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; GCAAAGCGTCTGAACGAAGGAG 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNGN3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; CCTAAGAGCGAGTTGGCACTGA 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; AGTGCCGAGTTGAGGTTGTGCA 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGCG\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; CGTTCCCTTCAAGACACAGAGG 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; ACGCCTGGAGTCCAGATACTTG 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSST\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; CCAGACTCCGTCAGTTTCTGCA 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; TTCCAGGGCATCATTCTCCGTC 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; AGACACAAAGAGGACACGCTGG 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; GAGTCGTAGGAGACAGAAGGTG 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSLC2A2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; ATGTCAGTGGGACTTGTGCTGC 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; AACTCAGCCACCATGAACCAGG 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGCK\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; CATCTCCGACTTCCTGGACAAG 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; TGGTCCAGTTGAGAAGGATGCC 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eABCC8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; GACGACAAGAGGACAGTGGTCT 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; GCATTCAGACCTCTGGAAGTCC 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKCNJ11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; TGTGTCACCAGCATCCACTCCT 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; GTTCTGCACGATGAGGATCAGG 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIA2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; TGGAGATCCTGGCTGAGCATGT 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; GGTCACATCAGCCAAAGACAGG 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKRT19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; AGCTAGAGGTGAAGATCCGCGA 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; GCAGGACAATCCTGGAGTTCTC 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCA2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; GTGACCTGGATTGTGCTCAAGG 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; GTTGTCCACCATCAGTTCTTCGG 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHNF1B\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; CCCAGCAAATCTTGTACCAGGC 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; ACCTCAGTGACCAAGTTGGAGC 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSOX9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026rsquo; AGGAAGCTCGCGGACCAGTAC 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u0026rsquo; GGTGGTCCTTCTTGTGCTGCAC 3\u0026rsquo;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eAll qPCR assays were performed using a QuantStudio 6 Flex Real-Time PCR system (Applied Biosystems).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eImmunofluorescence staining and confocal microscopy\u003c/h2\u003e \u003cp\u003eThe subcutaneously transplanted PANC-1 cells or the kidney with HPPD graft transplantation were surgically removed from the mice, fixed 1 hour in 4% paraformaldehyde at 4\u0026deg;C, and dehydrated using 30% sucrose solution overnight. The tissues were embedded in disposable base molds (Thermo Fisher Scientific) and 10 mm sections were cut. For staining, slides were blocked with PBS\u0026thinsp;+\u0026thinsp;0.1% Triton X-100 (Thermo Fisher Scientific)\u0026thinsp;+\u0026thinsp;5% donkey serum (Sigma-Aldrich) for 1 hr at room temperature (RT), incubated with primary antibodies overnight at 4\u0026deg;C, washed, incubated with secondary antibody incubation for 1 hr at RT, incubated with Hoechst 33342 (Invitrogen) for 10 min at RT, and washed. For imaging, samples were mounted in fluorescence mounting medium (Dako), covered with coverslips, and sealed with nail polish. Representative images were taken using a Zeiss LSM 710 confocal microscope. Primary antibody: Insulin (A0564, Dako), Pdx1 (5679S, Cell Signaling Technology),\u003c/p\u003e \u003cp\u003eDykddddk Tag (14793S, Cell Signaling Technology), C-peptide (GN-ID4, Developmental Studies Hybridoma Bank) and Cytokeratin 19 (Abcam, ab7754).\u003c/p\u003e \u003cp\u003eCells were seeded into 4-well culture slide (Falcon) at density of 10\u003csup\u003e5\u003c/sup\u003e cells/well. After another 24 hours, the cells were fixed, stained and subjected to fluorescence microscopic analysis as above described.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eSerum insulin measurement\u003c/h2\u003e \u003cp\u003eMouse blood was collected from the tail tip and allowed to clot. Serum was separated by brief centrifugation according to standard protocol. The serum insulin level was measured using STELLUX\u0026reg; Chemi Human Insulin ELISA kits (Alpco).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eElectron microscopy\u003c/h2\u003e \u003cp\u003eNTC-PANC-1 or ALDH3B2\u003csup\u003emut\u003c/sup\u003e-PANC-1 were fixed at RT for 2 hr with a mixture containing 1.25% PFA, 2.5% glutaraldehyde, and 0.03% picric acid in 0.1M sodium cacodylate buffer (pH 7.4). Samples were then sent to the Advanced Microscopy Core of Joslin for further processing and transmission electron microscope imaging.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eDNA methylation analyses\u003c/h2\u003e \u003cp\u003eBisulfite conversion (Zymo Research, EZ DNA Methylation-Direct Kits) of DNA from PANC-1, human primary pancreatic ductal cells carry NT or ALDH3B2\u003csup\u003emut\u003c/sup\u003e and human islets were performed as described previously \u003csup\u003e\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e\u003c/sup\u003e. Bisulfite-treated DNA was PCR amplified, using primers (human INS promoter forward primer \u0026minus;\u0026thinsp;5\u0026rsquo; AGGATAGGTTGTATTAGAAGAGGTTATTAAG 3\u0026rsquo;; human INS promoter reverse primer- 5\u0026rsquo; CCCCTAAACTCACCCCCACATACTTC 3\u0026rsquo;) specific for bisulfite treated DNA but independent of methylation status at monitored CpG sites. Reaction conditions for the first round of PCR were 5 cycles of 95\u0026deg;C 1 min, 52\u0026deg;C 3 min, 72\u0026deg;C 3 min followed by 40 cycles of 95\u0026deg;C 30 s, 55\u0026deg;C 45 s, 72\u0026deg;C 45 s followed by 7 min at 72\u0026deg;C. PCR products were gel purified and used for deep sequencing analysis (MGH DNA Core Facility, Cambridge, MA).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eSingle cell RNA sequencing\u003c/h2\u003e \u003cp\u003eIsolation of human primary pancreatic ductal cells was performed as described above. Purified human primary pancreatic ductal cells (HPPD) were immediately cultured in a low-attachment plate in RMPI DMEM/F12 medium (Gibco), supplemented with 10% FBS and penicillin/streptomycin. Lentivirus encoding a NT or \u003cem\u003eALDH3B2\u003c/em\u003e gRNA together with a Cas9-mCherry reporter construct was added to the culture media. 6 days later cells were harvested and alive mCherry positive cells were isolated by FACS sorting. 40,000 cells per sample were used for scRNAseq using the Chromium Next GEM Single Cell 3\u0026rsquo; GEM, Library \u0026amp; Gel Bead Kit v3.1 (cat # PN-1000213; 10 x Genomics) according to manufactures\u0026rsquo; instructions. Illumina NovaSeq 6000 with about 1.3\u0026nbsp;billion reads total was used for sequencing the purified 3\u0026rsquo; gene expression library. The single cell RNA-seq dataset was processed, explored and visualized using Cellenics\u0026reg; community instance (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://scp.biomage.net/\u003c/span\u003e\u003cspan address=\"https://scp.biomage.net/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) that\u0026rsquo;s hosted by Biomage (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://biomage.net/\u003c/span\u003e\u003cspan address=\"https://biomage.net/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) using preset standard filtering thresholds and integration methods. Pre-filtered count matrices were uploaded to Cellenics. Dead or dying cells were removed by filtering droplets with high mitochondrial content (3% cut-off). Outliers in the distribution of number of genes vs number of UMIs were removed by fitting a linear regression model (p-values between 1.97E-04\u0026ndash;2.34E-04). Cells with a high probability of being doublets were filtered out using the scDblFinder method (threshold range: 0.50\u0026ndash;0.56). Overall filtering rates after processing are in the range of 10.75% \u0026minus;\u0026thinsp;11.35% of cells. Data normalization, principal component analysis (PCA) and data integration using Harmony were performed on data from high-quality cells. Clusters were identified using the Louvain method, and a Uniform Manifold Approximation and Projection (UMAP) embedding was calculated to visualize the results. Cluster-specific marker genes were identified by comparing cells of each cluster to all other cells using the presto package implementation of the Wilcoxon rank-sum test.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003eModeling transcriptional dynamics of scRNA-seq database by seVelo\u003c/h2\u003e \u003cp\u003eTo recover the directed dynamic information of newly transcribed, unspliced pre-mRNAs and mature, spliced mRNAs, we used velocyto \u003csup\u003e\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e\u003c/sup\u003e to count both the unspliced and spliced RNA reads from the counts that were generated by 10X Cellranger \u003csup\u003e\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e\u003c/sup\u003e. We filtered out genes that had less than 10 counts and were expressed in less than 10 cells (both unspliced and spliced), normalized every cell by its total counts, and logarithmize the normalized counts. We then filtered out cells that had less than 1000 total counts. Finally, we kept the top 3000 highly variable genes. We computed the first and second-order moments (means and uncentered variances) among nearest neighbors in PCA space. First-order is needed for deterministic velocity estimation, while stochastic estimation also requires second-order moments. Transcriptional induction for a particular gene increases newly transcribed precursor unspliced mRNAs while, conversely, repression or absence of transcription decreases unspliced mRNAs. We used scVelo \u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e to solve the full transcriptional dynamics of splicing kinetics using a likelihood-based dynamical model.\u003c/p\u003e \u003cdiv id=\"Sec23\" class=\"Section3\"\u003e \u003ch2\u003eStatistical analyses\u003c/h2\u003e \u003cp\u003eStatistical analyses were performed by unpaired or paired tests as indicated using the Prism 8 software. All data are presented as mean \u0026plusmn; SEM. Significance was defined as *p\u0026thinsp;\u0026lt;\u0026thinsp;0.05, **p\u0026thinsp;\u0026lt;\u0026thinsp;0.01 and ***p\u0026thinsp;\u0026lt;\u0026thinsp;0.001. No samples were excluded from the analysis. Data analysis was not blinded. All data are representative of two or more similar experiments.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003ch2\u003eACKNOWLEDGEMENTS\u003c/h2\u003e \u003cp\u003eThis research was supported by funds from the Beatson Foundation to P.Y and S.B-W and by postdoctoral fellowship to J. L. from the Mary K. Iacocca Foundation and JDRF. We acknowledge the support from core facilities funded both by the NIDDK Diabetes Research Center ( P30DK036836) and the Joslin Diabetes Center, especially our CRISPR screen core facility and bioinformatic core (Dr. Hui Pan and Dr. Jonathan Dreyfuss). We also acknowledge the generous comments and editing from Dr. Gordon Weir and Dr. Stephan Kissler of Joslin Diabetes Center.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eWeir, G.C., Bonner-Weir, S. \u0026amp; Leahy, J.L. Islet mass and function in diabetes and transplantation. Diabetes 39, 401\u0026ndash;5 (1990).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNair, G. \u0026amp; Hebrok, M. Islet formation in mice and men: lessons for the generation of functional insulin-producing beta-cells from human pluripotent stem cells. Curr Opin Genet Dev 32, 171\u0026ndash;80 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNostro, M.C. \u003cem\u003eet al.\u003c/em\u003e Efficient generation of NKX6-1\u0026thinsp;+\u0026thinsp;pancreatic progenitors from multiple human pluripotent stem cell lines. Stem Cell Reports 4, 591\u0026ndash;604 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePagliuca, F.W. \u003cem\u003eet al.\u003c/em\u003e Generation of functional human pancreatic beta cells in vitro. \u003cem\u003eCell\u003c/em\u003e 159, 428\u0026thinsp;\u0026ndash;\u0026thinsp;39 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShapiro, A.M. \u003cem\u003eet al.\u003c/em\u003e Islet transplantation in seven patients with type 1 diabetes mellitus using a glucocorticoid-free immunosuppressive regimen. N Engl J Med 343, 230\u0026ndash;8 (2000).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAguayo-Mazzucato, C. \u0026amp; Bonner-Weir, S. Pancreatic beta Cell Regeneration as a Possible Therapy for Diabetes. Cell Metab 27, 57\u0026ndash;67 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDor, Y., Brown, J., Martinez, O.I. \u0026amp; Melton, D.A. Adult pancreatic beta-cells are formed by self-duplication rather than stem-cell differentiation. \u003cem\u003eNature\u003c/em\u003e 429, 41\u0026thinsp;\u0026ndash;\u0026thinsp;6 (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eButler, A.E. \u003cem\u003eet al.\u003c/em\u003e Adaptive changes in pancreatic beta cell fractional area and beta cell turnover in human pregnancy. Diabetologia 53, 2167\u0026ndash;76 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBen-Othman, N. \u003cem\u003eet al.\u003c/em\u003e Long-Term GABA Administration Induces Alpha Cell-Mediated Beta-like Cell Neogenesis. Cell 168, 73\u0026ndash;85 e11 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCollombat, P. \u003cem\u003eet al.\u003c/em\u003e The ectopic expression of Pax4 in the mouse pancreas converts progenitor cells into alpha and subsequently beta cells. Cell 138, 449\u0026ndash;62 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou, Q., Brown, J., Kanarek, A., Rajagopal, J. \u0026amp; Melton, D.A. In vivo reprogramming of adult pancreatic exocrine cells to beta-cells. \u003cem\u003eNature\u003c/em\u003e 455, 627\u0026thinsp;\u0026ndash;\u0026thinsp;32 (2008).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBonner-Weir, S., Aguayo-Mazzucato, C. \u0026amp; Weir, G.C. Dynamic development of the pancreas from birth to adulthood. Ups J Med Sci 121, 155\u0026ndash;8 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBonner-Weir, S. \u003cem\u003eet al.\u003c/em\u003e The pancreatic ductal epithelium serves as a potential pool of progenitor cells. Pediatr Diabetes 5 Suppl 2, 16\u0026ndash;22 (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eInada, A. \u003cem\u003eet al.\u003c/em\u003e Carbonic anhydrase II-positive pancreatic cells are progenitors for both endocrine and exocrine pancreas after birth. Proc Natl Acad Sci U S A 105, 19915\u0026ndash;9 (2008).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQadir, M.M.F. \u003cem\u003eet al.\u003c/em\u003e P2RY1/ALK3-Expressing Cells within the Adult Human Exocrine Pancreas Are BMP-7 Expandable and Exhibit Progenitor-like Characteristics. Cell Rep 22, 2408\u0026ndash;2420 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQadir, M.M.F. \u003cem\u003eet al.\u003c/em\u003e Single-cell resolution analysis of the human pancreatic ductal progenitor cell niche. Proc Natl Acad Sci U S A 117, 10876\u0026ndash;10887 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, D. \u003cem\u003eet al.\u003c/em\u003e Long-Term Expansion of Pancreatic Islet Organoids from Resident Procr(+) Progenitors. Cell 180, 1198\u0026ndash;1211 e19 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLieber, M., Mazzetta, J., Nelson-Rees, W., Kaplan, M. \u0026amp; Todaro, G. Establishment of a continuous tumor-cell line (panc-1) from a human carcinoma of the exocrine pancreas. Int J Cancer 15, 741\u0026ndash;7 (1975).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMadden, M.E. \u0026amp; Sarras, M.P., Jr. Morphological and biochemical characterization of a human pancreatic ductal cell line (PANC-1). \u003cem\u003ePancreas\u003c/em\u003e 3, 512\u0026thinsp;\u0026ndash;\u0026thinsp;28 (1988).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChai, R., Chen, S., Ding, J. \u0026amp; Grayburn, P.A. Efficient, glucose responsive and islet-specific transgene expression by a modified rat insulin promoter. Gene Ther 16, 1202\u0026ndash;9 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSanjana, N.E., Shalem, O. \u0026amp; Zhang, F. Improved vectors and genome-wide libraries for CRISPR screening. Nat Methods 11, 783\u0026ndash;784 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, W. \u003cem\u003eet al.\u003c/em\u003e MAGeCK enables robust identification of essential genes from genome-scale CRISPR/Cas9 knockout screens. Genome Biol 15, 554 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, W. \u003cem\u003eet al.\u003c/em\u003e Quality control, modeling, and visualization of CRISPR screens with MAGeCK-VISPR. Genome Biology 16, 281 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, B. \u003cem\u003eet al.\u003c/em\u003e Integrative analysis of pooled CRISPR genetic screens using MAGeCKFlute. Nat Protoc 14, 756\u0026ndash;780 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLai, L., Reineke, E., Hamilton, D.J. \u0026amp; Cooke, J.P. Glycolytic Switch Is Required for Transdifferentiation to Endothelial Lineage. Circulation 139, 119\u0026ndash;133 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVaughn, N. \u0026amp; Haviland, D.L. Acly promotes metabolic reprogramming and induction of IRF4 during early CD8(+) T cell activation. Cytometry A 99, 825\u0026ndash;831 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMarchitti, S.A., Brocker, C., Stagos, D. \u0026amp; Vasiliou, V. Non-P450 aldehyde oxidizing enzymes: the aldehyde dehydrogenase superfamily. Expert Opin Drug Metab Toxicol 4, 697\u0026ndash;720 (2008).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVassalli, G. Aldehyde Dehydrogenases: Not Just Markers, but Functional Regulators of Stem Cells. \u003cem\u003eStem Cells Int\u003c/em\u003e 2019, 3904645 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim-Muller, J.Y. \u003cem\u003eet al.\u003c/em\u003e Aldehyde dehydrogenase 1a3 defines a subset of failing pancreatic beta cells in diabetic mice. Nat Commun 7, 12631 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYatoh, S. \u003cem\u003eet al.\u003c/em\u003e Differentiation of affinity-purified human pancreatic duct cells to beta-cells. Diabetes 56, 1802\u0026ndash;9 (2007).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWescott, M.P. \u003cem\u003eet al.\u003c/em\u003e Pancreatic ductal morphogenesis and the Pdx1 homeodomain transcription factor. Mol Biol Cell 20, 4838\u0026ndash;44 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNeiman, D. \u003cem\u003eet al.\u003c/em\u003e Islet cells share promoter hypomethylation independently of expression, but exhibit cell-type-specific methylation in enhancers. Proc Natl Acad Sci U S A 114, 13525\u0026ndash;13530 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMa, Z. \u003cem\u003eet al.\u003c/em\u003e Deciphering early human pancreas development at the single-cell level. Nat Commun 14, 5354 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBergen, V., Lange, M., Peidli, S., Wolf, F.A. \u0026amp; Theis, F.J. Generalizing RNA velocity to transient cell states through dynamical modeling. Nat Biotechnol 38, 1408\u0026ndash;1414 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSarnobat, D. \u003cem\u003eet al.\u003c/em\u003e Taurine rescues pancreatic beta-cell stress by stimulating alpha-cell transdifferentiation. Biofactors 49, 646\u0026ndash;662 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKong, C. \u003cem\u003eet al.\u003c/em\u003e NEDD9 is a positive regulator of epithelial-mesenchymal transition and promotes invasion in aggressive breast cancer. PLoS One 6, e22666 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, C. \u003cem\u003eet al.\u003c/em\u003e HIF1alpha Regulates Early Metabolic Changes due to Activation of Innate Immunity in Nuclear Reprogramming. Stem Cell Reports 14, 192\u0026ndash;200 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKong, H.J. \u003cem\u003eet al.\u003c/em\u003e Crosstalk between YAP and TGFbeta regulates SERPINE1 expression in mesenchymal lung cancer cells. Int J Oncol 58, 111\u0026ndash;121 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMezza, T. \u003cem\u003eet al.\u003c/em\u003e Insulin resistance alters islet morphology in nondiabetic humans. Diabetes 63, 994\u0026ndash;1007 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJones, R.J. \u003cem\u003eet al.\u003c/em\u003e Assessment of aldehyde dehydrogenase in viable cells. Blood 85, 2742\u0026ndash;6 (1995).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStorms, R.W. \u003cem\u003eet al.\u003c/em\u003e Isolation of primitive human hematopoietic progenitors on the basis of aldehyde dehydrogenase activity. Proc Natl Acad Sci U S A 96, 9118\u0026ndash;23 (1999).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCai, J. \u003cem\u003eet al.\u003c/em\u003e Membrane properties of rat embryonic multipotent neural stem cells. \u003cem\u003eJ Neurochem\u003c/em\u003e 88, 212\u0026thinsp;\u0026ndash;\u0026thinsp;26 (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eObermair, F.J. \u003cem\u003eet al.\u003c/em\u003e A novel classification of quiescent and transit amplifying adult neural stem cells by surface and metabolic markers permits a defined simultaneous isolation. Stem Cell Res 5, 131\u0026ndash;43 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRovira, M. \u003cem\u003eet al.\u003c/em\u003e Isolation and characterization of centroacinar/terminal ductal progenitor cells in adult mouse pancreas. Proc Natl Acad Sci U S A 107, 75\u0026ndash;80 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKitamura, T., Takagi, S., Naganuma, T. \u0026amp; Kihara, A. Mouse aldehyde dehydrogenase ALDH3B2 is localized to lipid droplets via two C-terminal tryptophan residues and lipid modification. Biochem J 465, 79\u0026ndash;87 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRovira, M. \u003cem\u003eet al.\u003c/em\u003e Chemical screen identifies FDA-approved drugs and target pathways that induce precocious pancreatic endocrine differentiation. Proc Natl Acad Sci U S A 108, 19264\u0026ndash;9 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNinomiya, K., Takahashi, A., Fujioka, Y., Ishikawa, Y. \u0026amp; Yokoyama, M. Transforming growth factor-beta signaling enhances transdifferentiation of macrophages into smooth muscle-like cells. Hypertens Res 29, 269\u0026ndash;76 (2006).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMiettinen, P.J., Ebner, R., Lopez, A.R. \u0026amp; Derynck, R. TGF-beta induced transdifferentiation of mammary epithelial cells to mesenchymal cells: involvement of type I receptors. J Cell Biol 127, 2021\u0026ndash;36 (1994).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSabatini, P.V., Speckmann, T. \u0026amp; Lynn, F.C. Friend and foe: beta-cell Ca(2+) signaling and the development of diabetes. Mol Metab 21, 1\u0026ndash;12 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJoung, J. \u003cem\u003eet al.\u003c/em\u003e Genome-scale CRISPR-Cas9 knockout and transcriptional activation screening. Nat Protoc 12, 828\u0026ndash;863 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShalem, O. \u003cem\u003eet al.\u003c/em\u003e Genome-scale CRISPR-Cas9 knockout screening in human cells. Science 343, 84\u0026ndash;87 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRobinson, M. \u0026amp; Oshlack, A. A scaling normalization method for differential expression analysis of RNA-seq data. Genome Biology 11, 1\u0026ndash;9 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTsherniak, A. \u003cem\u003eet al.\u003c/em\u003e Defining a Cancer Dependency Map. Cell 170, 564\u0026ndash;576 e16 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSzklarczyk, D. \u003cem\u003eet al.\u003c/em\u003e STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res 47, D607-D613 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCorritore, E. \u003cem\u003eet al.\u003c/em\u003e beta-Cell differentiation of human pancreatic duct-derived cells after in vitro expansion. Cell Reprogram 16, 456\u0026ndash;66 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMillar, D.S., Warnecke, P.M., Melki, J.R. \u0026amp; Clark, S.J. Methylation sequencing from limiting DNA: embryonic, fixed, and microdissected cells. \u003cem\u003eMethods\u003c/em\u003e 27, 108\u0026thinsp;\u0026ndash;\u0026thinsp;13 (2002).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLa Manno, G. \u003cem\u003eet al.\u003c/em\u003e RNA velocity of single cells. Nature 560, 494\u0026ndash;498 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZheng, G.X. \u003cem\u003eet al.\u003c/em\u003e Massively parallel digital transcriptional profiling of single cells. Nat Commun 8, 14049 (2017).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-2222452/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-2222452/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eReplenishment of pancreatic beta cells is a key to the cure for diabetes. Beta cells regeneration is achieved predominantly by self-replication especially in rodents, but it was also shown that pancreatic duct cells can transdifferentiate into beta cells. How pancreatic duct cells were transdifferentiated and whether we could manipulate the transdifferentiation to replenish beta cell mass is not well understood. Using a genome-wide CRISPR screen, we discovered that loss-of-function of ALDH3B2 is sufficient to transdifferentiate human pancreatic duct cells into functional beta-like cells. The transdifferentiated cells have significant increase in beta cell marker genes expression, secrete insulin in response to glucose, and reduce blood glucose when transplanted into diabetic mice. Our study identifies a novel gene that we could potentially target in human pancreatic duct cells to replenish beta cell mass for diabetes therapy.\u003c/p\u003e","manuscriptTitle":"Loss-of-function of ALDH3B2 transdifferentiates human pancreatic duct cells into beta-like cells","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-10-08 08:18:32","doi":"10.21203/rs.3.rs-2222452/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"c3ba0f14-367c-4963-8113-30c47b4d6210","owner":[],"postedDate":"October 8th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":38638077,"name":"Biological sciences/Genetics"},{"id":38638078,"name":"Biological sciences/Cell biology"}],"tags":[],"updatedAt":"2024-11-05T12:55:48+00:00","versionOfRecord":[],"versionCreatedAt":"2024-10-08 08:18:32","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-2222452","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-2222452","identity":"rs-2222452","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}