TBX3 is dynamically expressed in pancreatic organogenesis and fine-tunes regeneration
BMC Biology volume 21, Article number: 55 (2023)
The reactivation of genetic programs from early development is a common mechanism for injury-induced organ regeneration. T-box 3 (TBX3) is a member of the T-box family of transcription factors previously shown to regulate pluripotency and subsequent lineage commitment in a number of tissues, including limb and lung. TBX3 is also involved in lung and heart organogenesis. Here, we provide a comprehensive and thorough characterization of TBX3 and its role during pancreatic organogenesis and regeneration.
We interrogated the level and cell specificity of TBX3 in the developing and adult pancreas at mRNA and protein levels at multiple developmental stages in mouse and human pancreas. We employed conditional mutagenesis to determine its role in murine pancreatic development and in regeneration after the induction of acute pancreatitis. We found that Tbx3 is dynamically expressed in the pancreatic mesenchyme and epithelium. While Tbx3 is expressed in the developing pancreas, its absence is likely compensated by other factors after ablation from either the mesenchymal or epithelial compartments. In an adult model of acute pancreatitis, we found that a lack of Tbx3 resulted in increased proliferation and fibrosis as well as an enhanced inflammatory gene programs, indicating that Tbx3 has a role in tissue homeostasis and regeneration.
TBX3 demonstrates dynamic expression patterns in the pancreas. Although TBX3 is dispensable for proper pancreatic development, its absence leads to altered organ regeneration after induction of acute pancreatitis.
TBX3 shows switching expression patterns during embryonic development in the pancreas and leads to fine-tuning of regeneration from acute pancreatitis via limiting proliferation and fibrosis during regeneration.
The family of T-box transcription factors includes five subfamilies and comprises 17 genes in both humans and mice. They all share highly conserved “T” DNA-binding domains and are essential during embryonic development and tissue homeostasis [1,2,3,4].
TBX3, along with TBX2, TBX4, and TBX5, belongs to the TBX2-family [1, 2]. Heterozygous mutations in TBX3 lead to the malformation of multiple structures resulting in ulnar-mammary syndrome in humans . Accordingly, homozygotes for hypomorphic mutations in Tbx3 in murine embryos and adults resulted in lethal maldevelopment of the heart and cardiac arrhythmias [4,5,6].
TBX3 has multiple molecular functions  and participates in transcription factor networks that guide organ patterning in multiple tissues. TBX3 and TBX2 are closely entangled in other regulatory programs, such as Wnt and Hedgehog signaling, and both are essential for proper limb , lung , and ureter development . Interestingly, redundancy between Tbx3 and Tbx2 to compensate for each other has been reported in several organs [9, 10].
Finally, TBX3 is not only critically involved in pluripotency maintenance but also in lineage-specific exit from the pluripotency circuitry in mouse embryonic stem cells [11,12,13,14,15,16,17]. TBX3 is also a putative inhibitor of pancreatic lineage entry in differentiating human pluripotent stem cells . Interestingly, TBX3 was associated with higher aggressiveness in pancreatic cancer via its ability to (i) promote angiogenesis and (ii) induce cancer stem cell properties . Such stem cell-related features may also become relevant during tissue regeneration after organ injury.
Common mechanisms and similar genetic programs between stem cell induction, organ homeostasis, and regeneration from injury have been identified for the pancreas [20, 21]. Among others, augmented Hedgehog, Notch, and Wnt/b-catenin signaling were delineated as mandatory for regeneration in mice after experimental induction of pancreatitis [21,22,23,24]. Interestingly, Tbx3 interacts with certain parts of the Wnt and Hedgehog signaling pathways [8,9,10].
To investigate the functional roles of Tbx3 during pancreatic development, adult pancreatic homeostasis, and pancreatic regeneration after injury, we here re-analyzed publicly available single-cell transcriptomic data sets and employed a Tbx3Venus reporter system as well as a compartment-specific Tbx3 knockout mouse model and a human knockdown induced pluripotent stem cell (iPSC) differentiation model. While TBX3 is expressed in both mesenchyme and epithelium during embryonic development, its deletion does not significantly alter organogenesis. However, loss of Tbx3 leads to overshoot proliferation of acinar cells, accumulation of fibrosis, and enhanced inflammatory stimuli, Il6-Jak-Stat3, and acinar cell-specific NF-κB signaling during pancreatitis.
TBX3 is dynamically expressed during pancreatic organogenesis and in adult pancreatic stellate cells
We first performed an RNA-based analysis to examine the differential expression patterns of Tbx3 in mice by re-analyzing published datasets of mouse development (GSE101099)  as well as of adult murine pancreata (GSE109774)  revealing dynamic expression patterns over distinct stages (Fig. 1a–c). At E12.5 (embryonic day 12.5) and E14.5, Tbx3 was expressed in various mesenchymal cells and neural crest cells, as indicated by cluster assignment using Col3a1 or Tlx2 [27, 28] (Fig. 1a,b, Additional file 1: Fig. S1a,b). Interestingly, while Tbx3 expression levels were retained in a few cells of the mesenchymal compartment, only a limited number of endothelial cells expressed Tbx3 at E17.5 as indicated by Pecam1 positivity (Fig. 1c, Additional file 1: Fig. S1c). Of note, a few acinar precursor cells marked by Cpa1 demonstrated expression of Tbx3. In the adult pancreas, Tbx3 was detected only in few stellate (Col3a1) and endothelial (Pecam1) cells but not exocrine (Cpa1, Krt19) or endocrine (Ins1, Gcg, Sst) cells (Fig. 1d, Additional file 1: Fig. S1d).
To substantiate our analysis across another species, we investigated the expression of human TBX3. First, we chose to re-analyze our recently published bulk transcriptomic datasets (GSE131817) [29, 30] to assess T-box gene expression during in vitro differentiation of human PSCs toward pancreatic progenitor cells (Additional file 1: Fig. S2a,b). While other T-box genes (TBX6, TBX19, EOMES) were highly expressed in human pluripotent stem cells and in definitive endoderm, TBX3 was among the most prominent transcripts at the pancreatic endoderm and pancreatic progenitor stage (Additional file 1: Fig. S2b). As mouse pancreatic progenitors start to arise and specify between E9.5 and E12.5 [31,32,33,34,35], this human analysis enabled us to zoom into earlier time points than those from the murine embryonic pancreata (Fig. 1a–c). To investigate TBX3 expression in the more mature pancreatic lineages [36, 37], we investigated our recently published single-cell transcriptome dataset (GSE162547) monitoring, particularly pancreatic ductal differentiation . Here, we noticed that TBX3 expression was heterogeneously, still robustly expressed throughout the arising lineages, including ductal, endothelial, and endocrine cells, indicative of a widely distributed embryonic expression of TBX3 during human development (Additional file 1: Fig. S2c). Next, we analyzed the time-resolved expression patterns of a set of TBX3 interaction partners [38,39,40] in our published bulk transcriptomic dataset [29, 30] (Additional file 1: Fig. S2d). Here, we observed a decrease in well-known embryonic factors (e.g., NANOG), while several genes demonstrated higher expression levels over time at the pancreatic progenitor stage (e.g., LEF1).
To further elucidate the cell type-specific expression profiles of these T-box genes in adult pancreata, we performed a single-cell transcriptomic in silico analysis from four published human pancreatic islet donors (GSE84133)  (Additional file 1: Fig. S2e,f). In addition to endocrine cells (GCG, INS; SST, PPY), also exocrine cell types (marked by KRT19 and CPA1), stellate (COL3A1), and endothelial (PECAM1) cells were identified (Additional file 1: Fig. S2f). As anticipated from the murine adult pancreata, the expression of TBX3 was restricted to the stellate cells (Additional file 1: Fig. S2e).
Thus, TBX3 is dynamically expressed during the differentiation of pancreatic cells at several developmental stages (PSC-derived pancreatic progenitors, duct-like cells, endothelial-like cells, and endocrine-like cells), and its expression shifts from epithelial cells to stromal cells in the adult human pancreas. Thus, expression in the adult human and murine organisms appears comparable in both species.
TBX3 in site analysis confirms transcriptionally derived expression patterns
Dynamic Tbx3 expression during pancreatic development prompted us to resolve stage-specific protein expression in the pancreatic anlage during embryonic and postnatal development. To visualize the activity of the murine Tbx3 promoter at the protein level, we used a validated Tbx3-Venus (Tbx3tm1(Venus)Vmc) [11, 42,43,44,45,46,47] reporter system (Fig. 2a). In this model, the production of Venus protein reflects Tbx3 promoter activity as detected with anti-GFP antibodies [42, 45]. By using a global knockout of Tbx3 through the replacement of the first three exons by a Cre-knockin , we confirmed TBX3 antibody specificity (Additional file 1: Fig. S3a). Similarly, the anti-GFP antibody was licensed to capture the Venus protein in Tbx3-positive cells of the seminal vesicle serving as a positive control (Additional file 1: Fig. S3b). We chose multiple developmental stages (E12.5, E15.5, E18.5, adult) to visualize Venus expression. In addition, we investigated the pancreata of P7 (postpartum day 7 after birth) newborn mice. These experiments revealed a good correlation between protein and the scRNA-seq data (Fig. 1). Venus was significantly expressed in the pancreatic mesenchyme (Fig. 2b), a structure known to mediate inductive cues to pattern the pancreatic epithelium during organogenesis . Specifically, at E12.5 Venus signal was detected around epithelial buds as indicated by co-staining for transcription factors NKX6-1, PDX1, SOX9, FOXA2, and basal membrane proteins such as laminin (LAM) (Fig. 2c). At E15.5, the expression pattern of Venus was virtually the same as at E12.5, demonstrated by colocalization with mesenchymal markers vimentin and collagen IV (COL IV), but not the epithelial markers NKX6-1, PDX1, CDH1, and CD49f (Fig. 2d). Of note, Venus protein was detected both in the mesenchymal and epithelial compartment of the pancreas at E18.5 and P7 whereas RNA expression of Tbx3 was more abundant in mesenchymal and endothelial cells at E17.5 (Figs. 1d and 2e-g). However, concordant to the few acinar precursor cells expressing Tbx3 at E17.5 (Fig. 1c), the epithelial expression was mostly in acinar cells as proven by colocalization of Venus with amylase (AMY) and not in endocrine cells labeled by insulin (INS) and glucagon (GCG). Furthermore, Venus was detected in mesenchymal cells (collagen IV +), and also detected in PECAM-1 expressing endothelial cells. To further validate the expression of Venus in acinar (precursor) cells at E18.5, we performed staining of Venus and TBX3 on consecutive sections of E18.5 pancreata (Fig. 2f) confirming the expression of both markers in acinar structures (Zoom-ins in Fig. 2f). Immunohistochemistry analysis further licensed the acinar expression pattern (Additional file 1: Fig. S3c). In adult mice, concordant with the RNA data, Venus was not detected in pancreas-specific epithelial (CDH-1, AMY, CK-19, INS, GCG) cell types (Figs. 1e and 2h). Thus, TBX3 protein expression, as assayed with the Venus reporter, presents specific and developmentally dynamic expression patterns in both the pancreatic epithelium and the mesenchyme. These patterns correlate mostly with the RNA findings, except the weak but robust acinar expression around E18.5 which appears underestimated in scRNA sequencing data potentially attributed to technical limitations for weakly expressed genes [50,51,52,53,54].
Compartment-specific deletion of Tbx3 leads to virtually unaltered pancreatic development
To study the functional role of TBX3 in the context of pancreatic organogenesis, two different conditional knockout models targeting Tbx3 in either the pancreatic epithelium or mesenchyme were employed. Embryonic deletion of Tbx3 in the epithelial compartment of the pancreas was achieved by crossing Tbx3flox/+ with Ptf1aCre/+ mice, a model known to target all epithelial lineages of the pancreas, including acinar, ductal, and most endocrine cells starting from E10.5 (Fig. 3a) . Specific recombination in epithelial cells of the pancreas was confirmed by immunohistochemical stainings for tdRFP in Ptf1aCre/+  x LSL-tdRFPKI/KI [55, 56] mice (Additional file 1: Fig. S3d) . Histologic morphology was similar between homozygous knockout mice (Tbx3-KO (epi, epithelial)) and control mice (no knockout of Tbx3) (Fig. 3b). Furthermore, immunofluorescence analysis labeling exocrine pancreatic markers (CK-19, AMY2A; Fig. 3c) revealed virtually no difference between Tbx3-KO (epi) and control mice. This was also valid for the endocrine compartment, where only subtle differences with a trend toward smaller islets in Tbx3-KO (epi) mice was documented while maintaining similar percentages of INS+ beta-cells and GCG+ alpha-cells (Fig. 3d–g).
Given the essential role of the pancreatic mesenchyme for proper lineage formation [49, 57], Tbx3 was also deleted in this compartment. The Nkx3-2Cre/+ mouse strain was used (Fig. 3h) as it has been employed to demonstrate the relevance of the pancreatic mesenchyme for proper pancreatic development . As with the epithelial compartment, homozygous deletion of Tbx3 (Tbx3-KO (mes, mesenchymal)) in the mesenchyme did not alter pancreatic morphology (assayed by histomorphologic investigation of H&E staining) (Fig. 3i). Additional analysis of exocrine markers (CK-19 and AMY2A) showed no difference between Tbx3-KO (mes) and control mice (Fig. 3j). Interestingly, also the investigation of the endocrine markers insulin and glucagon revealed no significant difference (Fig. 3k–n). The size of islets (Fig. 3l) and the percentages of insulin- and glucagon-expressing cells (Fig. 3m,n) were not significantly different between Tbx3-KO (mes) and control mice. Thus, although Tbx3 displayed specific and dynamic regulation patterns on mRNA and protein levels in both mesenchyme and epithelium, its targeted ablation in either compartment does not relevantly alter proper pancreatic organ formation.
To check whether the well-described homolog TBX2 might compensate for TBX3 loss [9, 10], we performed immunohistochemistry for TBX2 in pancreata of both knockout models (Additional file 1: Fig. S4a,b). However, the overall staining intensity remained similar in control (epi) and Tbx3-KO (epi), as well as control (mes) and Tbx3-KO (mes) pancreata, albeit nuclear and cytoplasmatic protein distribution varied across various animals without a relevant genotype-specific trend.
Finally, we wanted to challenge the relevance of TBX3 in human pancreatic development. Thus, we employed an inducible knockdown of TBX3 by an shRNA in a previously reported iPSC line  during in vitro pancreatic progenitor differentiation [58,59,60,61]. We assessed trilineage potential (generation of acinar, ductal, and endocrine cells) in a recently described porcine urinary bladder (PUB) organ culture model for the maturation of pancreatic progenitor cells [62, 63] (Additional file 1: Fig. S5a). Of note, we did observe the generation of morphologically similar (Additional file 1: Fig. S5b), ductal (Additional file 1: Fig. S5c), acinar (Additional file 1: Fig. S5d), and endocrine (Additional file 1: Fig. S5e) structures on the PUB scaffold. Thus, the trilineage potential was not significantly altered by TBX3-knockdown during pancreatic progenitor differentiation. We conclude that TBX3 is dispensable for proper pancreatic organ formation in mice and men.
Loss of Tbx3 results in increased fibrosis and excessive proliferation during recovery from acute pancreatitis
Since pancreatic development shares common regulatory patterns with pancreatic repair, we challenged Tbx3-KO (epi) mice and control mice with caerulein-induced acute pancreatitis to investigate the potential relevance of Tbx3 under exogenous stress conditions (Fig. 4a). Notably, no significant differences between control and Tbx3-KO mice were detected based on the amount of acinar-to-ductal metaplasia (ADM) as well as edema and inflammatory infiltration scores (Fig. 4b–e). As expected, a substantial increase of ADMs at 72h preceded by increased edema and infiltration of inflammatory cells (24h) was documented after induction of caerulein-driven acute pancreatitis in both groups of animals. The damage was largely repaired 7 days after the first injection. However, when assessing the proliferation rates of acinar cells with normal morphological configuration by KI-67 quantification, we noticed a significantly higher amount of KI-67+ cells per field in the Tbx3-KO (epi) mice compared to the control counterpart at 72h (4.3 cells/per field vs 1.7 cells/field) and 168h (3.0 cells/field vs 1.2 cells/field) (Fig. 4f–h). Interestingly, increased acinar proliferation was accompanied by significantly more fibrosis as quantified upon Sirius Red stainings at 72h in Tbx3-KO (epi) mice (Fig. 4i,j). The amount of fibrosis was reflected by the presence of ACTA2-positive cells around ADM structures (Fig. 4k). Interestingly, gene expression analysis of Acta2 at 72h after the onset of pancreatitis confirmed the significant difference between Tbx3-KO (epi) and control mice (Fig. 4l). Finally, we checked if apoptosis at 72h after induction of pancreatitis might play a role in the altered regeneration (Fig. 4m,n). However, no significant differences were observed concerning the rate of apoptotic cells. Thus, despite the fact that regeneration was achieved in Tbx3-KO (epi) mice, prolonged proliferation in acinar cells and accumulation of fibrosis suggest that Tbx3 is involved in proper regeneration from organ injury, but its absence is well compensated by other genes.
Loss of Tbx3 results in increased susceptibility of acinar cells to fibroinflammatory stimuli
To further specify the subtle differences across the two genotypes during regeneration (Tbx3-KO (epi) vs. control (epi) mice), we performed RNA sequencing for the two most interesting time points (72 and 168h). Interestingly, we noticed a clear difference in the principal component analysis (PCA) of whole transcriptomes from Tbx3-KO (epi) mice compared to their respective controls at 72h, while the differences at 168h were virtually absent (Fig. 5a). We could identify 1152 differentially expressed genes (DEG) at 72h (762 up, 390 down in Tbx3-KO (epi)), while only 1 DEG was detected at 168h (Fig. 5b). Subsequently, we performed gene set enrichment analysis (GSEA) for hallmark (HM) gene sets and one acinar-specific NF-κB response gene set (Additional file 2: Supplementary Table 1) adapted from  (Fig. 5c). Interestingly, the significantly enriched gene sets included the Il6-Jak-Stat3-, Il2-Stat5-signaling, Inflammatory-Response, E2F-Targets, G2M-Checkpoint, and the acinar-specific NF-κB response. Notably, most gene sets are related to the immune system. We also checked other well-known gene sets relevant to pancreatic organ regeneration and detected no significant enrichment for the HM-Hedgehog, Wnt, Tgf-β, and Notch signaling. Concerning depleted gene sets, we noticed a non-significant depletion of Pancreatic-Beta-Cell-signature and Protein-Secretion-Signature, the latter being potentially indicative of normal acinar function.
Fitting with the enrichment of the Il6-Jak-Stat3 gene signature, a variety of immune cell-related genes (e.g., Cxcl9, Il2rg) were significantly overexpressed at 72h post pancreatitis (Fig. 5d), which prompted us to investigate specific immune cell signatures (Additional file 2: Supplementary Table 2 ; (Fig. 5e). Interestingly, signatures for B cells and neutrophils were significantly upregulated. However, validation on protein level could only confirm an increased abundance of B cells in Tbx3-KO (epi) pancreata as compared to control (epi) counterparts at 72h after induction of pancreatitis (Fig. 5f,h). Neutrophil granulocyte quantification revealed only a trend (Fig. 5g,i). Infiltrating T cells and macrophages remained similar across the two genotypes (Additional file 1: Fig. S6a-d). To identify potential compensators for Tbx3 loss in the epithelial compartment, we analyzed the expression of all expressed Tbx genes at 72h (Fig. 5j). Interestingly, most Tbx genes remained unaltered, with Tbx18 being significantly downregulated and Tbx21 overexpressed in the Tbx3-KO (epi) mice. Next, a set of Tbx3 target genes were investigated  (Fig. 5k). Amongst these, the sole strongly and significantly upregulated gene was Lef1. Accordingly, we calculated the putative binding score and significance using the palindromic TBX3 binding site (18 bp) for the Lef1 promoter consensus sequence. Indeed, the relative similarity score was highly significant (score = 0.82, p = 0.0009), indicating direct TBX3 binding to the Lef1 promoter, most likely triggering its transcriptional repression (Fig. 5l).
In this study, we (i) determined the expression of TBX3 during pancreatic organogenesis, (ii) demonstrated expression switching between mesenchymal and epithelial cells during pancreatic development, and (iii) revealed expression in the stromal compartment of the adult pancreas. Surprisingly, epithelial and mesenchymal knockout of Tbx3 did not lead to statistically significant phenotypic alterations in the murine pancreas, and knockdown of TBX3 in a human iPSC-based system for pancreatic differentiation did not alter ex vivo pancreatic development. It is likely that TBX3 function may be compensated by other genes in these compartments during development. Driven by the fact that developmental programs can become re-activated during regeneration from injury, we investigated the relevance of Tbx3 for regeneration after caerulein-induced acute pancreatitis. Our findings indicate that epithelial deletion of Tbx3 does not significantly impact overall organ regeneration but does prolong proliferation and increase fibrosis. Again, this lack of a major effect in the face of Tbx3 ablation is likely attributable to other factors, while Tbx3 in itself is required for the fine-tuning of the regenerative process. Whole transcriptome analysis finally revealed putative mechanisms that alter the regeneration via enhanced immune cell-acinar interactions, fibroinflammatory stimuli, and dysregulated proliferation (Fig. 5m). Altogether, while Tbx3 is not a master regulator of pancreatic development and organ regeneration, our data show a relevant but subtle contribution of this factor to adult pancreatic organ homeostasis and development.
Interestingly, Tbx3 RNA expression levels demonstrated a virtually lower percentage of Tbx3+ cells than assayed in immunofluorescence analysis. This discrepancy was higher for acinar cells than for mesenchymal or stromal cells, which may be related to either technical limitations of single-cell RNA-seq [51,52,53,54] with a bias to filter out genes with low expression  and eventually boosted by the fact that the protease- and ribonuclease-rich acinar cells raise difficulties to isolate intact RNA [66,67,68]. Another potential explanation would be that Tbx3 expression levels in mesenchymal/stromal cells are, in general, higher (also evident in the violin plots of Fig. 1), which eases the detection.
Nevertheless by employing a knockout-validated antibody and a Venus reporter system [11, 42,43,44,45,46,47], the expression of TBX3, or Venus, respectively, was proven on the protein level. In addition, the evidence of human TBX3 expression in an in vitro iPSC differentiation in the epithelial compartment highlights that TBX3 is indeed expressed during embryonic pancreatic development.
Indeed, important roles for TBX3 have been shown in lung branching morphogenesis and ureter organogenesis [9, 10]. In both cases, TBX3 was located in the mesenchyme, and depletion of Tbx3 (together with Tbx2, which otherwise would have compensated Tbx3) resulted in defective organ patterning [9, 10]. Based on such recent publications, we chose to not only perform a knockout of Tbx3 in the epithelial compartment by Ptf1a-Cre but also in the mesenchymal compartment by Nkx3-2-Cre. Other than our anticipation from the expression level profiles and the recent publications [9, 10], neither of the knockout models led to a strong phenotype. TBX2 and 3 are members of the same T-box subfamily . Their partially redundant roles in lung or ureter development [5, 9, 10] raise the question if this leads to redundant function in pancreatic evolution. However, our solely on gene and protein quantification relying study was underpowered to detect evidence for TBX2 compensating during pancreatic development as TBX2 remained unchanged in both the two knockout models and the pancreatic injury model. To ultimately clarify this hypothesis, a double-knockout of Tbx2 and Tbx3 in both pancreatic compartments would be necessary. However, this was clearly beyond the scope of the current study.
Even though we did not observe a gross impact of Tbx3 on recovery from acute pancreatitis, we observed (i) higher amounts of fibrosis, (ii) prolonged proliferation in acinar cells in its absence, and (iii) higher variance in some investigated parameters of regeneration of Tbx3-KO mice. Of note, we assessed the putative compensation by other Tbx genes during pancreatitis. Here we observed an upregulation of Tbx21. However, as Tbx21 (T-bet) and also Eomes (increased expression, though not significant) also play relevant roles in immune cell signaling, we attributed the enhanced expression to the increased inflammatory signature. Considering the downregulation of Tbx18, we pinpoint a recent publication of cardiac pacemaker differentiation where Tbx3 and Tbx18 are co-regulated, implying that a similar program could occur during pancreatic regeneration .
Based on the fact that Tbx3 is involved in the regulation of Wnt signaling in embryonic development  and Wnt signaling controls the proliferation of acinar cells during organ regeneration , Tbx3 may have a regulatory role on Wnt signaling during acinar cell regeneration. By demonstrating that Lef1, a central mediator of the Wnt signaling axis [24, 70,71,72,73], is not only significantly upregulated in Tbx3-KO (epi) mice but also possesses a putative DNA binding site for TBX3 in its promoter region, we postulate a previously unknown potential suppressive function of TBX3 on Lef1 building the following hypothetical model: The absence of Tbx3 in the pancreatic epithelium could trigger excessive proliferation eventually involving Lef1, a known member of the Wnt signaling [74, 75], and acinar-specific upregulated NF-KB-pathway members leading to abnormal proliferation of acinar cells . Such proliferation pulse is accompanied by an increased fibroinflammatory response resulting in transient accumulation of B cell-infiltrated fibrotic tissue (Fig. 5m). This suggests an ongoing homeostatic process during pancreatic regeneration from an injury involving only Tbx3 for fine-tuning without the generation of substantial alterations upon genetic loss.
In summary, Tbx3 shows switching expression patterns in the developing pancreas. Although Tbx3 seems dispensable for proper pancreatic development and lineage entry, its absence results in altered organ regeneration after induction of acute pancreatitis marked by enhanced fibrosis and inflammation.
Single-cell RNA sequencing re-analysis
Expression patterns of Tbx3 on a single-cell level were derived from two different murine datasets: GSE101099  for embryonic pancreata at E12.5, E14.5, and E17.5, as well as GSE109774 for adult pancreata . Expression of TBX3 in a recently published human pancreatic duct-like cell differentiation approach from pluripotent stem cells was performed as recently described in detail (GSE162547) . The expression of TBX3 in adult pancreata of four human donors was interrogated in a published dataset: GSE84133 . Before expression analysis, preprocessing steps were conducted in RStudio with the version 4.0.4 and the “dplyr”, “Seurat” version 4.0.6 [76,77,78,79], and “patchwork” packages. For the murine datasets, cells with less than 1000 expressed genes, more than 4000 genes for E12.5, E14.5, and more than 5000 genes for E17.5 and genes expressed in fewer than 3 cells were filtered out before analysis. For GSE109774, cells expressing more than 5% ERCC RNA spike-ins were additionally removed. For human datasets, cells expressing either less than 300 genes or more than 4000 genes and genes expressed in fewer than 10 cells were filtered out. Data were merged for E12.5 for GSM2699156_E12_B2 and GSM3140915_E12_v2, for E14.5 for GSM2699154_E14_B1, GSM2699155_E14_B2, and GSM3140916_E14_v2, and for E17.5 for GSM2699157_E17_B2, GSM3140917_E17_1_v2, and GSM3140918_E17_2_v2 of the respective dataset. Human datasets were also merged. Batch effects were corrected with the FindIntegrationAnchor() and the IntegrateData() function in the reciprocal PCA approach of the Seurat workflow. Datasets were Log-normalized, and the top 2000 highly variables genes were identified by the “vst” method for the murine datasets. For the human datasets, similar to our recent work , the top 4000 highly variable genes were identified. Data were scaled. A standard Seurat workflow was performed. Single-cell neighborhood was calculated with the first 30 principal components and clustering was performed with the Louvain algorithm with a resolution of 0.015 for E12.5, 0.5 for E14.5, 0.5 for E17.5, 0.2 for the adult murine pancreas, and 0.5 for human datasets. Cluster identity was annotated by specific marker genes.
Bulk RNA expression re-analysis from publicly available RNA sequencing data sets
RNA expression levels of different T-Box transcription factors in embryonic stem cells, definitive endoderm, pancreatic endoderm, and pancreatic progenitors were derived as FPKM (fragments per kilobase of transcript per million mapped reads) from a previously published dataset (GSE131817, wildtype samples) [29, 30]. Data were analyzed in RStudio with the R version 4.0.4. The heatmap was generated with the “pheatmap” package (version 1.0.12), scaling was set to columns, and clustering followed the “ward.D2” method. A list of described TBX3 interaction partners was derived from . A heatmap was generated as indicated above from the same dataset (GSE131817) [29, 30]. The scale was set to the row factor.
All animal care and procedure were conducted in compliance with the German legal regulations and were previously approved by the local governmental review board of the state of Baden-Württemberg (Permission no. 1477, O.195-4, O.195-6, O.195-10, and O.195-12) or conducted in compliance with the local government of Bavaria. All mouse work aspects were performed according to acknowledged guidelines of the Society of Laboratory Animals (GV-SOLAS) and of the Federation of Laboratory Animal Science Associations (FELASA).
Ptf1aCre (Ptf1atm1(cre)Cvw) mouse strain  was previously described [80, 81]. Nkx3-2Cre (Nkx3-2tm1(cre)Wez) was a kind gift of Warren E. Zimmer (Texas A&M University) . Tbx3flox (Tbx3tm3.1Moon) (kind gift of Anne M. Moon) , and Tbx3Venus (Tbx3tm1(Venus)Vmc) mouse (kind gift of Vincent M. Christoffels)  strain were maintained on a complex C57BL/6×129/Sv genetic background. Mice were housed and bred in a conventional health status-controlled animal facility. All animal care and procedures followed German legal regulations and if applicable were previously approved by the governmental review board of the state of Baden-Württemberg. All the aspects of the mouse work were carried out following strict guidelines to insure careful, consistent, and ethical handling of mice.
Analysis of Ptf1aCre-specific recombination in adult pancreata
Formalin-fixed paraffin-embedded (FFPE) tissue sections of pancreata of Ptf1aCre  x R26-LSL-tdRFPKI  mice which were previously described  were a kind gift of Patrick Hermann (Ulm University Hospital, Department of Internal Medicine I).
Knockout validation of TBX3-antibodies
TBX3 antibody was analyzed by immunofluorescence in archived Tbx3Cre/Cre at E13.5 or Tbx3wt/wt. .
Mouse embryos and adult pancreata preparation
Embryos were collected at 12.5, 15.5, and 18.5 days post-coitum. Newborn mice were euthanized at post natal day 7 (P7). Pancreata were fixed in 4% PFA for 2 h at room temperature or for 16 h at 4°C, cryopreserved, and embedded in optimal cutting temperature compound (OCT compound). Adult mice were at least 8 weeks old. Pancreata were fixed in cold 4% PFA for 16 h at 4°C and embedded in paraffin for histological analysis.
Model of Caerulein-induced acute pancreatitis
Acute pancreatitis in Tbx3-KO (epi) and control mice was induced in adult animal (≥ 8 weeks) by hourly injections (8 times) of 50 μg caerulein /kg bodyweight (Sigma-Aldrich) dissolved in PBS (vehicle). Mice were sacrificed 24, 72, and 168 h after the first caerulein injection. Corresponding volumes of vehicle were injected in Tbx3-KO (epi) and control mice, and pancreata were isolated 24 h after the first injection.
Differentiation of pancreatic progenitor cells and organ culture model
Maintenance culture and differentiation to pancreatic progenitors of a well-described iPSC line with a doxycycline-inducible shRNA against TBX3  was performed as described recently in a step-by-step protocol [58, 59] with slight modifications. For shRNA expression, 3 μg/mL doxycycline was added to the cells. The medium was changed on a daily basis. After the differentiation of pancreatic progenitor cells, organ culture of porcine urinary bladders was performed as described recently [62, 63]. Briefly, porcine urinary bladders (PUB) were cleaned, de-epithelialized, and sterilized with 0.1% peroxy-acetic acid. In total, 500,000 cells per ring were seeded in 30 μL of 50% Matrigel and 50% basal medium as in [58, 59], including 5% FCS and 10 μM Y-27632. The medium of PUBs was changed once a week, and doxycycline was added twice a week freshly. After 2 weeks, PUBs were fixed in 3.7% formaldehyde and processed for histology as described recently [62, 63].
All histological experiments on FFPE tissue were performed as previously described [20, 59, 62]) following standard procedures. Four-micrometer-thin sections were rehydrated. Heat-mediated antigen retrieval was either performed with citrate-based buffer (pH=6) or TRIS-based buffer (pH=9). Blocking was performed in 5% normal donkey serum in 0.1% Triton X-100 in PBS for 30 min at RT. Primary and secondary antibodies and respective dilution factors are listed in Additional file 2: Supplementary Table 3. Bright-field images were acquired using a Leica DM5500B microscope (Leica) equipped with a Leica DMC5400 camera and Leica Application Suite software (Leica) or by using a Zeiss Axioscope2 microscope (Carl Zeiss) ZEN3.1 imaging software (Carl Zeiss). Immunofluorescence images were obtained with a Zeiss Axioscope2 microscope equipped with an Axiocam 702 (Carl Zeiss). Acquired pictures were subsequently analyzed using ImageJ software (National Institutes of Health).
Immunofluorescence stainings of cryosections were performed following standard protocols. Briefly, 10-μm-thick cryo-sectioned pancreata were rehydrated in PBS for 30 min, washed twice with PBS containing 0.1 % Tween 20 (PBS-T), and permeabilized using 0.1 M glycine (Merck) and 0.1% Triton X-100 (for NKX6.1: 0.5% Triton X-100; Merck) in MilliQ water for 15 min at RT and blocked using 0.1% Tween - 20, 10% heat inactivated fetal calf serum (FCS), 0.1% BSA, and 3% donkey serum in PBS for 1 h at RT. Incubation with primary antibodies diluted in blocking solution occurred overnight at 4°C or 1 h at RT. The slides were washed 3× with PBS-T for 10 min each, and subsequently, secondary antibody solution was added for 2 h at RT. DAPI/PBS solution was added for 20 min before washing of slides 3× with PBS for 10 min each. Finally, the slides were mounted with Elvanol and kept 24 h at RT to dry.
The total number of cells per Langerhans islets, as well as the number of insulin and glucagon-expressing cells, was counted manually from at least 3 different islets per pancreas on one section. Acinar-to-ductal metaplasias (ADMs) were quantified by counting at least ten 200× fields. Edema and immune cell infiltration were scored from 0 to 3 (0, no evident edema to 3, maximal degree of edema; 0, no immune infiltration to 3 maximal degree of immune infiltration) as previously described . Proliferation was quantified by counting KI-67-positive cells in immunohistochemistry from ten 400× fields of each pancreas. KI67-positive cells within only clearly identifiable acinar structures were considered. Immune cells were excluded from the quantification. For picrosirius red-positive area/hematoxylin-based whole pancreas surface ratio quantification, pictures from at least five 100× fields were loaded into ImageJ software (National Institutes of Health) to perform color deconvolution. Areas covered by picrosirus red (red) and hematoxylin (yellow) were quantified automatically. Immune cell populations were quantified by counting the absolute number of positive cells per field in at least ten 200× fields. Caspase-3 (CASP3)-positive cells per field of view were quantified manually by counting the number of positive cells in each pancreas in at least six 100× fields. Monocytes, neutrophils, and T cells were manually quantified in at least five different 200× fields. B cells were quantified in at least 8 different 200× fields by quantifying the B220 positive area per field in ImageJ.
RNA extraction, cDNA synthesis, qRT-PCR
RNA extraction was performed as described previously  with the RNeasy Plus Mini Kit (Qiagen) according to the manufacturer’s instructions. RNA concentration was determined with a NanoDrop. cDNA synthesis was performed with the iScript™ cDNA Synthesis Kit (Bio-Rad) following the manufacturer’s instruction. Briefly, 1 μg of RNA was transcribed. cDNA was diluted 15-fold. PCR was performed at the Rotor-Gene-Q (Qiagen) using 4 μL of diluted cDNA with 5 μL of Green Master Mix (Genaxxon) and 0.4 μM forward and reverse primer (each 0.5μL). Gene expression was normalized to ribosomal protein S18. Primer (Biomers) sequences are as follows: Acta2: fwd 5′-GTTCAGTGGTGCCTCTGTCA-3′, rev 5′-ACTGGGACGACATGGAAAAG-3′, S18: fwd 5′-GTAACCCGTTGAACCCCATT-3′, rev 5′- CCATCCAATCGGTAGTAGCG-3′.
The amount of total RNA was quantified using the Qubit 2.0 Fluorometric Quantitation system (Thermo Fisher Scientific, Waltham, MA, USA), and the RNA integrity number (RIN) was determined using the Experion Automated Electrophoresis System (Bio-Rad, Hercules, CA, USA). RNA-seq libraries were prepared with the NEBNext® Ultra™ II Directional RNA sample preparation kit (New England Biolabs, Inc., Ipswich, MA, USA). Library concentrations were quantified with the Qubit 2.0 Fluorometric Quantitation system (Life Technologies, Carlsbad, CA, USA), and the size distribution was assessed using the Experion Automated Electrophoresis System (Bio-Rad, Hercules, CA, USA). For sequencing, samples were diluted and pooled into NGS libraries in equimolar amounts.
Next-generation sequencing and raw data acquisition
Expression profiling libraries were sequenced on a HiSeq 3000 instrument (Illumina, San Diego, CA, USA) following a 50-base-pair, single-end recipe. Raw data acquisition (HiSeq Control Software, HCS, HD 184.108.40.206) and base calling (Real-Time Analysis Software, RTA, 2.7.7) was performed on-instrument, while the subsequent raw data processing off the instruments involved two custom programs (https://github.com/epigen/picard/) based on Picard tools (2.19.2) (https://broadinstitute.github.io/picard/). In a first step, base calls were converted into lane-specific, multiplexed, unaligned binary alignment map (BAM) files suitable for long-term archival (IlluminaBasecallsToMultiplexSam, 2.19.2-CeMM). In a second step, archive BAM files were demultiplexed into sample-specific, unaligned BAM files (IlluminaSamDemux, 2.19.2-CeMM).
Next-generation sequencing (NGS) reads were mapped to the Genome Reference Consortium GRCm38 assembly via “Spliced Transcripts Alignment to a Reference” (STAR, 2.7.9a)  utilising the “basic” Ensembl transcript annotation from version e100 (April 2020) as reference transcriptome. Since the mm10 assembly flavour of the University of California, Santa Cruz (UCSC) Genome Browser was preferred for downstream data processing with Bioconductor packages for entirely technical reasons, Ensembl transcript annotation had to be adjusted to UCSC Genome Browser sequence region names. STAR was run with options recommended by the ENCODE project. NGS read alignments overlapping Ensembl transcript features were counted with the Bioconductor (3.14) GenomicAlignments (1.30.0) package via the summarizeOverlaps function in Union mode, ignoring secondary alignments and alignments not passing vendor quality filtering. Since the NEBNext® Ultra™ II Directional RNA protocol leads to sequencing of the first strand, all alignments needed inverting before strand-specific counting in feature (i.e., gene, transcript, and exon) orientation. Transcript-level counts were aggregated to gene-level counts, and the Bioconductor DESeq2 (1.34.0)  package was used to test for differential expression based on a model using the negative binomial distribution.
An initial exploratory analysis included principal component analysis (PCA), multidimensional scaling (MDS), sample distance, and expression heatmap plots, all annotated with variables used in the expression modelling (ggplot2 , 3.3.6, and Bioconductor ComplexHeatmap , 2.10.0), as well as volcano plots (Bioconductor EnhancedVolcano  , 1.12.0). Biologically meaningful results were extracted from the model, and log2-fold values were shrunk with the CRAN ashr  (2.2.-54) package, while two-tailed p-values obtained from Wald testing were adjusted with the Bioconductor Independent Hypothesis Weighting  (IHW, 1.22.0) package.
Gene set enrichment analysis
GSEA was performed using the hallmark data sets from the Molecular signatures database v7.4 (MSigDB, Broad Institute; http://software.broadinstitute.org/gsea/msigdb) and the GSEA software version 2.4.3 [90, 91]. Additional GSEA was performed for an acinar-specific NF-κB gene set. Genes upregulated with a significant fold change of least 2 from  were chosen to generate the reference gene list (Additional file 2: Supplementary Table 1). Immune cell gene sets were derived from  (Additional file 2: Supplementary Table 2). Significant enrichments were defined false discovery rate <0.25.
Putative TBX3-DNA binding analysis
All the analysis was performed using R and Bioconductor (R Core Team, 2022). Lef1 promoter sequence was retrieved from M. musculus UCSC genome version mm10 known Gene database. The position weight matrix (PWM) was generated using the 8 entries of Lef1 promoter regions and was subsequently used for the sequence logo. Mouse TBX3 palindromic binding site was obtained from Uniprot (https://www.uniprot.org/uniprotkb/P70324/entry). Sequence logos, relative similarity score, correlation, Euclidean distance, and p-value were calculated using Transcription Factors binding sites tools in R.
Statistical analysis was performed with GraphPad Prism 9.3.1 software. Data are expressed as individual data points and mean ± SEM (indicated in figure legends). Significance levels for islet size, insulin+, and glucagon+ cells were calculated with a Mann-Whitney test. Significance was evaluated with a two-way ANOVA and a Sidak post-test for the investigation of number of ADM per field, edema score, infiltration, score, proliferation, and fibrosis in the acute pancreatitis model. Significance levels for apoptosis, Acta2 gene expression, and immune cell infiltrations at 72h after onset of the pancreatitis were assessed by Mann-Whitney test. Significance levels were defined as the following: p < 0.05 = *, p < 0.01 = **.
Availability of data and materials
All data generated or analysed during this study are included in this published article, its supplementary information files, and publicly available repositories. Quantitative data generated in this study are presented in Additional file 3 and Additional file 4. The newly generated raw data from RNA-seq can be accessed from the Gene Expression Omnibus repository via the accession code GSE216889 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE216889). GSE101099, GSE109774, GSE162547, and GSE109774 were accessed for data analysis through the Gene Expression Omnibus. Other data will be made available upon reasonable request.
Binary alignment map
Differentially expressed genes
Embryonic day post conception
Fetal calf serum
Fragments per kilobase of transcript per million mapped reads
Gene set enrichment analysis
Induced pluripotent stem cells
Postpartum day after birth
Principal component analysis
Porcine urinary bladder
Position weight matrix
Small hairpin RNA
Spliced Transcripts Alignment to a Reference
University of California, Santa Cruz
Uniform manifold approximation and projection
Papaioannou VE. The t-box gene family: emerging roles in development, Stem cells and cancer. Development. 2014;141(20):3819–33.
Khan SF, Damerell V, Omar R, Du Toit M, Khan M, Maranyane HM, et al. The roles and regulation of TBX3 in development and disease. Gene. 2020;726:144223.
Bertolessi M, Linta L, Seufferlein T, Kleger A, Liebau S. A fresh look on T-Box factor action in early embryogenesis (T-Box factors in early development). Stem Cells Dev. 2015;24:1833–51.
Frank DU, Carter KL, Thomas KR, Burr RM, Bakker ML, Coetzee WA, et al. Lethal arrhythmias in Tbx3-deficient mice reveal extreme dosage sensitivity of cardiac conduction system function and homeostasis. Proc Natl Acad Sci U S A. 2012;109:E154–63.
Singh R, Hoogaars WM, Barnett P, Grieskamp T, Sameer Rana M, Buermans H, et al. Tbx2 and Tbx3 induce atrioventricular myocardial development and endocardial cushion formation. Cell Mol Life Sci. 2012;69(8):1377–89.
Mesbah K, Harrelson Z, Théveniau-Ruissy M, Papaioannou VE, Kelly RG. Tbx3 is required for outflow tract development. Circ Res. 2008;103:743–50.
Kumar P, Franklin S, Emechebe U, Hu H, Moore B, Lehman C, et al. TBX3 regulates splicing in vivo: a novel molecular mechanism for ulnar-mammary syndrome. PLoS Genet. 2014;10(3):e1004247.
Emechebe U, Pavan Kumar P, Rozenberg JM, Moore B, Firment A, Mirshahi T, et al. T-box3 is a ciliary protein and regulates stability of the Gli3 transcription factor to control digit number. Elife. 2016;5:e07897.
Lüdtke TH, Rudat C, Wojahn I, Weiss AC, Kleppa MJ, Kurz J, et al. Tbx2 and Tbx3 act downstream of Shh to maintain canonical Wnt signaling during branching morphogenesis of the murine lung. Dev Cell. 2016;39(2):239–53.
Aydoğdu N, Rudat C, Trowe MO, Kaiser M, Lüdtke TH, Taketo MM, et al. TBX2 and TBX3 act downstream of canonical WNT signaling in patterning and differentiation of the mouse ureteric mesenchyme. Development. 2018;145(23):dev171827.
Russell R, Ilg M, Lin Q, Wu G, Lechel A, Bergmann W, et al. A dynamic role of TBX3 in the pluripotency circuitry. Stem Cell Rep. 2015;5:1155–70.
Klingenstein M, Raab S, Achberger K, Kleger A, Liebau S, Linta L. TBX3 knockdown decreases reprogramming efficiency of human cells. Stem Cells Int. 2016;2016:6759343.
Weidgang CE, Russell R, Tata PR, Kühl SJ, Illing A, Müller M, et al. TBX3 directs cell-fate decision toward mesendoderm. Stem Cell Reports. 2013;1:248–65.
Quarta C, Fisette A, Xu Y, Colldén G, Legutko B, Tseng YT, et al. Functional identity of hypothalamic melanocortin neurons depends on Tbx3. Nat Metab. 2019;1(2):222–35.
Cioffi M, Vallespinos-Serrano M, Trabulo SM, Fernandez-Marcos PJ, Firment AN, Vazquez BN, et al. MiR-93 controls adiposity via inhibition of Sirt7 and Tbx3. Cell Rep. 2015;12(10):1594–605.
Kartikasari AER, Zhou JX, Kanji MS, Chan DN, Sinha A, Grapin-Botton A, et al. The histone demethylase Jmjd3 sequentially associates with the transcription factors Tbx3 and Eomes to drive endoderm differentiation. EMBO J. 2013;32(10):1393–408.
Lu R, Yang A, Jin Y. Dual functions of T-box 3 (Tbx3) in the control of self-renewal and extraembryonic endoderm differentiation in mouse embryonic stem cells. J Biol Chem. 2011;286:8425–36.
Mukherjee S, French DL, Gadue P. Loss of TBX3 enhances pancreatic progenitor generation from human pluripotent stem cells. Stem Cell Reports. 2021;16(11):2617–27.
Perkhofer L, Walter K, Costa IG, Carrasco MCR, Eiseler T, Hafner S, et al. Tbx3 fosters pancreatic cancer growth by increased angiogenesis and activin/nodal-dependent induction of stemness. Stem Cell Res. 2016;17:367–78.
Arnold F, Mahaddalkar PU, Kraus JM, Zhong X, Bergmann W, Srinivasan D, et al. Functional genomic screening during somatic cell reprogramming identifies DKK3 as a roadblock of organ regeneration. Adv Sci. 2021;8(14):2100626.
Murtaugh LC, Keefe MD. Regeneration and repair of the exocrine pancreas. Annu Rev Physiol. 2015;77:229–49.
Fendrich V, Esni F, Garay MVR, Feldmann G, Habbe N, Jensen JN, et al. Hedgehog signaling is required for effective regeneration of exocrine pancreas. Gastroenterology. 2008;135(2):621–31.
Siveke JT, Lubeseder-Martellato C, Lee M, Mazur PK, Nakhai H, Radtke F, et al. Notch signaling is required for exocrine regeneration after acute pancreatitis. Gastroenterology. 2008;134(2):544–55.
Keefe MD, Wang H, De La OJP, Khan A, Firpo MA, Murtaugh LC. β-catenin is selectively required for the expansion and regeneration of mature pancreatic acinar cells in mice. DMM. Dis Model Mech. 2012;5(4):503–14.
Byrnes LE, Wong DM, Subramaniam M, Meyer NP, Gilchrist CL, Knox SM, et al. Lineage dynamics of murine pancreatic development at single-cell resolution. Nat Commun. 2018;9(1):3922.
Tabula Muris Consortium; Overall coordination; Logistical coordination; Organ collection and processing; Library preparation and sequencing; Computational data analysis; Cell type annotation; Writing group; Supplemental text writing group; Principal investigators. Single-cell transcriptomics of 20 mouse organs creates a Tabula Muris. Nature. 2018;562(7727):367–72. https://doi.org/10.1038/s41586-018-0590-4.
Hatano M, Iitsuka Y, Yamamoto H, Dezawa M, Yusa S, Kohno Y, et al. Ncx, a Hox11 related gene, is expressed in a variety of tissues derived from neural crest cells. Anat Embryol (Berl). 1997;195(5):419–25.
Shirasawa S, Yunker AMR, Roth KA, Brown GA, Horning S, Korsmeyer SJ. Enx (HOX11L1)-deficient mice develop myenteric neuronal hyperplasia and megacolon. Nat Med. 1997;3(6):646–50.
Philippi A, Heller S, Costa IG, Senée V, Breunig M, Li Z, et al. Mutations and variants of ONECUT1 in diabetes. Nat Med. 2021. https://doi.org/10.1038/s41591-021-01502-7.
Heller S, Li Z, Lin Q, Geusz R, Breunig M, Hohwieler M, et al. Transcriptional changes and the role of ONECUT1 in hPSC pancreatic differentiation. Commun Biol. 2021;4(1):1298.
Burlison JS, Long Q, Fujitani Y, Wright CVE, Magnuson MA. Pdx-1 and Ptf1a concurrently determine fate specification of pancreatic multipotent progenitor cells. Dev Biol. 2008;316(1):74–86.
Haumaitre C, Barbacci E, Jenny M, Ott MO, Gradwohl G, Cereghini S. Lack of TCF2/vHNF1 in mice leads to pancreas agenesis. Proc Natl Acad Sci U S A. 2005;102(5):1490–5.
Gu G, Dubauskaite J, Melton DA. Direct evidence for the pancreatic lineage: NGN3+ cells are islet progenitors and are distinct from duct progenitors. Development. 2002;129(10):2447–57.
Henseleit KD, Nelson SB, Kuhlbrodt K, Hennings JC, Ericson J, Sander M. NKX6 transcription factor activity is required for α- and β-cell development in the pancreas. Development. 2005;132(13):3139–49.
Kawaguchi Y, Cooper B, Gannon M, Ray M, MacDonald RJ, Wright CVE. The role of the transcriptional regulator Ptf1a in converting intestinal to pancreatic progenitors. Nat Genet. 2002;32(1):128–34.
Gonçalves CA, Larsen M, Jung S, Stratmann J, Nakamura A, Leuschner M, et al. A 3D system to model human pancreas development and its reference single-cell transcriptome atlas identify signaling pathways required for progenitor expansion. Nat Commun. 2021;12(1):3144. Available from: http://www.nature.com/articles/s41467-021-23295-6.
Wiedenmann S, Breunig M, Merkle J, von Toerne C, Georgiev T, Moussus M, et al. Single-cell-resolved differentiation of human induced pluripotent stem cells into pancreatic duct-like organoids on a microwell chip. Nat Biomed Eng. 2021;5(8):897–913. https://doi.org/10.1038/s41551-021-00757-2.
Washkowitz AJ, Gavrilov S, Begum S, Papaioannou VE. Diverse functional networks of Tbx3 in development and disease. Wiley Interdiscip Rev Syst Biol Med. 2012;4:273–83.
Eblaghie MC, Song SJ, Kim JY, Akita K, Tickle C, Jung HS. Interactions between FGF and Wnt signals and Tbx3 gene expression in mammary gland initiation in mouse embryos. J Anat. 2004;205(1):1–13.
Davenport TG, Jerome-Majewska LA, Papaioannou VE. Mammary gland, limb and yolk sac defects in mice lacking Tbx3, the gene mutated in human ulnar mammary syndrome. Development. 2003;130:2263–73.
Baron M, Veres A, Wolock SL, Faust AL, Gaujoux R, Vetere A, et al. A single-cell transcriptomic map of the human and mouse pancreas reveals inter- and intra-cell population structure. Cell Syst. 2016;3(4):346–60 e4.
Kunasegaran K, Ho V, Chang THT, De Silva D, Bakker ML, Christoffels VM, et al. Transcriptional repressor Tbx3 is required for the hormone-sensing cell lineage in mammary epithelium. PLoS One. 2014;9:e110191.
van Eif VWW, Stefanovic S, van Duijvenboden K, Bakker M, Wakker V, de Gier-de Vries C, et al. Transcriptome analysis of mouse and human sinoatrial node cells reveals a conserved genetic program. Development. 2019;146:dev173161.
Mohan RA, Bosada FM, Weerd JHV, van Duijvenboden K, Wang J, Mommersteeg MTM, et al. T-box transcription factor 3 governs a transcriptional program for the function of the mouse atrioventricular conduction system. Proc Natl Acad Sci U S A. 2020;117(31):18617–26.
Mohan RA, Mommersteeg MTM, Domínguez JN, Choquet C, Wakker V, de Gier-de Vries C, et al. Embryonic Tbx3+ cardiomyocytes form the mature cardiac conduction system by progressive fate restriction. Development. 2018;145:dev167361.
van Weerd JH, Mohan RA, van Duijvenboden K, Hooijkaas IB, Wakker V, Boukens BJ, et al. Trait-associated noncoding variant regions affect tbx3 regulation and cardiac conduction. Elife. 2020;9:e56697.
van Eif VWW, Stefanovic S, Mohan RA, Christoffels VM. Gradual differentiation and confinement of the cardiac conduction system as indicated by marker gene expression. Biochim Biophys Acta Mol Cell Res. 2020;1867(3):118509.
Hoogaars WMH, Engel A, Brons JF, Verkerk AO, De Lange FJ, Wong LYE, et al. Tbx3 controls the sinoatrial node gene program and imposes pacemaker function on the atria. Genes Dev. 2007;21(9):1098–112.
Landsman L, Nijagal A, Whitchurch TJ, VanderLaan RL, Zimmer WE, MacKenzie TC, et al. Pancreatic mesenchyme regulates epithelial organogenesis throughout development. PLoS Biol. 2011;9(9):e1001143.
Svensson V, Natarajan KN, Ly LH, Miragaia RJ, Labalette C, Macaulay IC, et al. Power analysis of single-cell rnA-sequencing experiments. Nat Methods. 2017;14(4):381–7.
Hicks SC, Townes FW, Teng M, Irizarry RA. Missing data and technical variability in single-cell RNA-sequencing experiments. Biostatistics. 2018;19(4):562–78.
Zheng Y, Zhong Y, Hu J, Shang X. SCC: an accurate imputation method for scRNA-seq dropouts based on a mixture model. BMC Bioinformatics. 2021;22(1):5.
Islam S, Zeisel A, Joost S, La Manno G, Zajac P, Kasper M, et al. Quantitative single-cell RNA-seq with unique molecular identifiers. Nat Methods. 2014;11(2):163–6.
Papalexi E, Satija R. Single-cell RNA sequencing to explore immune cell heterogeneity. Nat Rev Immunol. 2018;18:35–45.
Luche H, Weber O, Rao TN, Blum C, Fehling HJ. Faithful activation of an extra-bright red fluorescent protein in “knock-in” Cre-reporter mice ideally suited for lineage tracing studies. Eur J Immunol. 2007;37(1):43–53.
Walter K, Tiwary K, Trajkovic-Arsic M, Hidalgo-Sastre A, Dierichs L, Liffers ST, et al. MEK inhibition targets cancer stem cells and impedes migration of pancreatic cancer cells in vitro and in vivo. Stem Cells Int. 2019;2019:8475389.
Cozzitorto C, Mueller L, Ruzittu S, Mah N, Willnow D, Darrigrand JF, et al. A specialized niche in the pancreatic microenvironment promotes endocrine differentiation. Dev Cell. 2020;55(2):150–62 e6.
Breunig M, Merkle J, Melzer MK, Heller S, Seufferlein T, Meier M, et al. Differentiation of human pluripotent stem cells into pancreatic duct-like organoids. STAR Protoc. 2021;2(4):100913 Available from: https://linkinghub.elsevier.com/retrieve/pii/S2666166721006195 [cited 12 Dec 2021].
Breunig M, Merkle J, Wagner M, Melzer MK, Barth TFE, Engleitner T, et al. Modeling plasticity and dysplasia of pancreatic ductal organoids derived from human pluripotent stem cells. Cell Stem Cell. 2021;28(6):1105–24 e19.
Merkle J, Breunig M, Schmid M, Allgöwer C, Krüger J, Melzer MK, et al. CDKN2A-mutated pancreatic ductal organoids from induced pluripotent stem cells to model a cancer predisposition syndrome. Cancers (Basel). 2021;13(20):5139 Available from: https://www.mdpi.com/2072-6694/13/20/5139.
Krüger J, Breunig M, Pasquini LP, Morawe M, Groß A, Arnold F, et al. Functional genomic screening in human pluripotent stem cells reveals new roadblocks in early pancreatic endoderm formation. Cells. 2022;11(3):582.
Melzer MK, Breunig M, Arnold F, Wezel F, Azoitei A, Roger E, et al. Organoids at the PUB: the porcine urinary bladder serves as a pancreatic niche for advanced cancer modeling. Adv Healthc Mater. 2022;11:e2102345 Available from: https://onlinelibrary.wiley.com/doi/full/10.1002/adhm.202102345 [cited 2 Mar 2022].
Melzer MK, Breunig M, Lopatta P, Hohwieler M, Merz S, Azoitei A, et al. Protocol to use de-epithelialized porcine urinary bladder as a tissue scaffold for propagation of pancreatic cells. STAR Protoc. 2022;3(4):101869.
Liou GY, Döppler H, Necela B, Krishna M, Crawford HC, Raimondo M, et al. Macrophage-secreted cytokines drive pancreatic acinar-to-ductal metaplasia through NF-ΚB and MMPs. J Cell Biol. 2013;202:563–77.
Nirmal AJ, Regan T, Shih BB, Hume DA, Sims AH, Freeman TC. Immune cell gene signatures for profiling the microenvironment of solid tumors. Cancer Immunol Res. 2018;6(11):1388–400.
Assi M, Dauguet N, Jacquemin P. DIE-RNA: A reproducible strategy for the digestion of normal and injured pancreas, isolation of pancreatic cells from genetically engineered mouse models and extraction of high quality RNA. Front Physiol. 2018;9:129.
Griffin M, Abu-El-Haija M, Abu-El-Haija M, Rokhlina T, Uc A. Simplified and versatile method for isolation of high-quality RNA from pancreas. Biotechniques. 2012;52(5):332–4.
Li D, Ren W, Wang X, Wang F, Gao Y, Ning Q, et al. A modified method using TRIzol® reagent and liquid nitrogen produces high-quality RNA from rat pancreas. Appl Biochem Biotechnol. 2009;158(2):253–61.
Zhao H, Wang F, Zhang W, Yang M, Tang Y, Wang X, et al. Overexpression of TBX3 in human induced pluripotent stem cells (hiPSCs) increases their differentiation into cardiac pacemaker-like cells. Biomed Pharmacother. 2020;130:110612.
Morris JP IV, Cano DA, Sekine S, Wang SC, Hebrok M. β-catenin blocks Kras-dependent reprogramming of acini into pancreatic cancer precursor lesions in mice. J Clin Invest. 2010;120(2):508–20.
Zhang Y, Yu J, Shi C, Huang Y, Wang Y, Yang T, et al. Lef1 contributes to the differentiation of bulge stem cells by nuclear translocation and cross-talk with the notch signaling pathway. Int J Med Sci. 2013;10(6):738–46.
Zhou P, Byrne C, Jacobs J, Fuchs E. Lymphoid enhancer factor 1 directs hair follicle patterning and epithelial cell fate. Genes Dev. 1995;9(6):700–13.
Singhi AD, Lilo M, Hruban RH, Cressman KL, Fuhrer K, Seethala RR. Overexpression of lymphoid enhancer-binding factor 1 (LEF1) in solid-pseudopapillary neoplasms of the pancreas. Mod Pathol. 2014;27(10):1355–63.
Santiago L, Daniels G, Wang D, Deng FM, Lee P. Wnt signaling pathway protein LEF1 in cancer, as a biomarker for prognosis and a target for treatment. Am J Cancer Res. 2017;7:1389–406.
Clevers H. Wnt/beta-catenin signaling in development and disease. Cell. 2006;127:469–80.
Satija R, Farrell JA, Gennert D, Schier AF, Regev A. Spatial reconstruction of single-cell gene expression data. Nat Biotechnol. 2015;33(5):495–502.
Butler A, Hoffman P, Smibert P, Papalexi E, Satija R. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat Biotechnol. 2018;36(5):411–20.
Stuart T, Butler A, Hoffman P, Hafemeister C, Papalexi E, Mauck WM, et al. Comprehensive integration of single-cell data. Cell. 2019;177(7):1888–902 e21.
Hao Y, Hao S, Andersen-Nissen E, Mauck WM, Zheng S, Butler A, et al. Integrated analysis of multimodal single-cell data. Cell. 2021;184(13):3573–87 e29.
Russell R, Perkhofer L, Liebau S, Lin Q, Lechel A, Feld FM, et al. Loss of ATM accelerates pancreatic cancer formation and epithelial-mesenchymal transition. Nat Commun. 2015;6:7677.
Perkhofer L, Schmitt A, Romero Carrasco MC, Ihle M, Hampp S, Ruess DA, et al. ATM deficiency generating genomic instability sensitizes pancreatic ductal adenocarcinoma cells to therapy-induced DNA damage. Cancer Res. 2017;77:5576–90.
Verzi MP, Stanfel MN, Moses KA, Kim BM, Zhang Y, Schwartz RJ, et al. Role of the homeodomain transcription factor Bapx1 in mouse distal stomach development. Gastroenterology. 2009;136(5):1701–10.
Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2013;29(1):15–21.
Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550.
Wickham H. Package ‘ggplot2’: elegant graphics for data analysis. New York: Springer-Verlag; 2016.
Gu Z, Eils R, Schlesner M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics. 2016;32(18):2847–9.
Blighe K, Rana S, Lewis M. EnhancedVolcano: Publication-ready volcano plots with enhanced colouring and labelin. R-Package; 2019.
Stephens M. False discovery rates: a new deal. Biostatistics. 2017;18(2):275–94.
Ignatiadis N, Klaus B, Zaugg JB, Huber W. Data-driven hypothesis weighting increases detection power in genome-scale multiple testing. Nat Methods. 2016;13(7):577–80.
Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102(43):15545–50.
Mootha VK, Lindgren CM, Eriksson KF, Subramanian A, Sihag S, Lehar J, et al. PGC-1α-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet. 2003;34(3):267–73.
Brulet P, Babinet C, Kemler R, Jacob F. Monoclonal antibodies against trophectoderm-specific markers during mouse blastocyst formation. Proc Natl Acad Sci U S A. 1980;77(7 II):4113–7.
We are deeply grateful to Warren E. Zimmer from Texas A&M University who gave his permission to use the Nkx3.2tm1(cre)Wez mouse strain for generating the mesenchymally depleted Tbx3-deficient mice. We further thank Limor Landsman from Sackler Faculty of Medicine for sending the Nkx3.2tm1(cre)Wez mouse strain to our facility. We further profoundly acknowledge Kuhn Elektro-Technik GmbH for supporting our research to fight pancreatic cancer. The authors thank Katrin Köhn, Aref Saed, Ralf Köhntop, Ulrike Mayr-Beyrle, Rashmi Bijegatte, Franziska Ott, Claudia Längle, Meike Hohwieler, Markus Breunig, Elodie Roger, Eleni Zimmer, Vibha Kumaraswamy Bharadwaj, Karolin Walter, Patrick Hermann, André Lechel, and Lisa Appel for excellent technical support and for helpful discussions. We thank the Biomedical Sequencing Facility at CeMM – Research Center for Molecular Medicine of the Austrian Academy of Sciences for supporting our RNA-seq analysis. The CK-19 antibody (TROMA-III) was deposited to the DSHB by Kemler, R. (DSHB Hybridoma Product TROMA-III) and kindly provided by DSHB for our studies . Illustrative figures were partially modified from Servier Medical Art, licensed under a Creative Common Attribution 3.0 Generic License https://smart.servier.com/, or created in BioRender.com.
Open Access funding enabled and organized by Projekt DEAL. The main funding was acquired by AK from the Deutsche Forschungsgemeinschaft (DFG) “Sachbeihilfe” KL2544/7-1 and “Heisenberg-Programm” KL2544/6-1. Further funding was acquired by AK from the German Cancer Aid (AK70114761) and by LP from the DFG (PE 3337/1-1). A.K. is speaker of an Else Kröner Research School for Physicians. M.K.M is a clinician scientist within the “Clinician-Scientist-Programm” of Ulm University and the Else Kröner Research School for Physicians. The work was further supported by funds from the Helmholtz Association (or in specific case of Helmholtz Munich) and the German Center for Diabetes Research (DZD).
Ethics approval and consent to participate
All animal care and procedure were conducted in compliance with the German legal regulations and were previously approved by the local governmental review board of the state of Baden-Württemberg (Permission no. 1477, O.195-4, O.195-6, O.195-10, and O.195-12) or conducted in compliance with the local government of Bavaria. No new human sequencing data were generated.
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Expression patterns of specific marker genes for cluster assignment in murine pancreata. Fig. S2. TBX3 is expressed during human pancreatic differentiation and in stellate cells of the adult pancreas. Fig. S3. Antibody and co-expression validation of TBX3 and Venus markers and validation of pancreatic recombination. Fig. S4. TBX2 expression in Ptf1a-Cre and Nkx3-2-Cre driven Tbx3-KO mice pancreata. Fig. S5. TBX3-knockdown does not impair the formation of human pancreatic tissue. Fig. S6. Tbx3 depletion does not alter T cell and macrophage infiltration during tissue regeneration after acute pancreatitis.
Upregulated acinar-specific NF-KB response geneset. Supplementary Table 2. Immune cell gene sets. Supplementary Table 3. Antibodies for histology.
About this article
Cite this article
Melzer, M.K., Schirge, S., Gout, J. et al. TBX3 is dynamically expressed in pancreatic organogenesis and fine-tunes regeneration. BMC Biol 21, 55 (2023). https://doi.org/10.1186/s12915-023-01553-x
- Pancreatic development
- Acute pancreatitis
- Organ regeneration
- Embryonic development