Objectives Previous studies have suggested an important role of the transcription factor Gata6 in endocrine pancreas, while GATA6 haploinsufficient inactivating mutations cause pancreatic agenesis in humans. We aimed to analyse the effects of Gata6 inactivation on pancreas development and function.
Design We deleted Gata6 in all epithelial cells in the murine pancreas at the onset of its development. Acinar proliferation, apoptosis, differentiation and exocrine functions were assessed using reverse transcriptase quantitative PCR (RT-qPCR), chromatin immunoprecipitation, immunohistochemistry and enzyme assays. Adipocyte transdifferentiation was assessed using electron microscopy and genetic lineage tracing.
Results Gata6 is expressed in all epithelial cells in the adult mouse pancreas but it is only essential for exocrine pancreas homeostasis: while dispensable for pancreatic development after e10.5, it is required for complete acinar differentiation, for establishment of polarity and for the maintenance of acinar cells in the adult. Gata6 regulates directly the promoter of genes coding for digestive enzymes and the transcription factors Rbpjl and Mist1. Upon pancreas-selective Gata6 inactivation, massive loss of acinar cells and fat replacement take place. This is accompanied by increased acinar apoptosis and proliferation, acinar-to-ductal metaplasia and adipocyte transdifferentiation. By contrast, the endocrine pancreas is spared.
Conclusions Our data show that Gata6 is required for the complete differentiation of acinar cells through multiple transcriptional regulatory mechanisms. In addition, it is required for the maintenance of the adult acinar cell compartment. Our studies suggest that GATA6 alterations may contribute to diseases of the human adult exocrine pancreas.
- Exocrine Pancreatic Function
- Cell Biology
- Pancreatic Damage
Statistics from Altmetric.com
Significance of this study
What is already known on this subject?
Gata6 is a transcription factor thought to be mainly involved in endocrine pancreas development.
GATA6 haploinsufficient de novo mutations are associated with pancreatic agenesis.
What are the new findings?
We provide evidence that, in mice, Gata6 is not required for the formation of the exocrine or endocrine pancreas but it is essential for normal maturation of acinar cells (full expression of digestive enzymes and normal cell polarity). The effects of Gata6 involve multiple, cooperative, transcriptional regulatory mechanisms.
Gata6 is also required for the maintenance of the acinar compartment in the adult pancreas in the absence of damage.
Gata6 inactivation favours acinar-to-ductal metaplasia and fat replacement of the pancreas, including epithelial-adipocyte transdifferentiation. However, no evidence of exocrine insufficiency is observed.
How might it impact on clinical practice in the foreseeable future?
As previously shown, our findings indicate that massive loss of the acinar compartment can occur in the absence of exocrine insufficiency.
Incomplete acinar differentiation may favour plasticity, which in turn may contribute to aggravate pancreatic function.
In humans, GATA6 genetic variation may contribute to the development of diseases of the exocrine pancreas.
GATA1-6 transcription factors share high homology and bind to a conserved consensus DNA motif: WGATAR.1 Functional specificity is provided by a divergent spatial and temporal expression: GATA1-3 are predominantly expressed in haematopoietic cells while GATA4-6 are expressed in mesoderm- and endoderm-derived tissues. Gata6 is important during development of heart, lung, liver and the gastrointestinal tract.2–4 It is essential for mouse embryo development and its complete ablation is lethal between e5.5–e7.5 due to endodermal defects.5 Gata6 mRNA is detected throughout development in both exocrine and endocrine pancreas. At e9.5, when pancreas formation begins, Gata6 mRNA is detected broadly in epithelial cells and by e16.5 it becomes restricted to the ducts and the endocrine compartment.6 There are conflicting reports on the expression pattern of Gata6 in the adult pancreas but a prominent role in the endocrine compartment was suggested.6–9 Transient transgenic mice expressing a dominant repressor form of Gata6 uncovered a role in early pancreas specification6 but experimental limitations hampered a full understanding of the role of Gata6 in pancreas development. Recently, de novo haploinsufficient inactivating GATA6 mutations have been reported in patients with pancreatic agenesis, arguing for a broader essential function.10
Here, we describe the effects of pancreatic Gata6 deletion at e10.5 using Ptf1a-driven Cre recombinase. Gata6 is dispensable for pancreas development while it is essential for complete differentiation of acinar cells and for adult exocrine homeostasis. Gata6 inactivation caused massive acinar loss and replacement by fat, including adipocyte transdifferentiation. Gata6 null pancreata also showed persistent acinar-to-ductal metaplasia (ADM). By contrast, we have not found any evidence for a requirement of Gata6 in normal endocrine pancreas function.
Materials and methods
Mouse strains were: Gata6loxP/loxP,11 Ptf1a-Cre+/KI,12 Rosa26LSL-EYFP.13 Ptf1a-Cre+/KI mice were used as controls, to rule out effects of loss of one Ptf1a allele. Mice were bred and maintained under sterile and pathogen-free conditions. Experiments were approved by the Animal Ethical Committee of Instituto de Salud Carlos III and performed following guidelines for Ethical Conduct in the Care and Use of Animals as stated in The International Guiding Principles for Biomedical Research involving Animals, developed by the Council for International Organizations of Medical Sciences.
Histopathology and immunohistochemistry
For Oil-Red-O staining, tissues were fixed in 4% paraformaldehyde, passed through a sucrose gradient, embedded in OCT and snap-frozen. For histopathological analysis, pancreata were fixed in 10% formalin, embedded in paraffin and serially sectioned (3 μm); every 10th section was stained using H&E.
Immunostaining followed standard protocols. Primary antibodies were: Gata6 (R&D AF1700, 1 : 1000), Gata4 (C-20 Santa Cruz sc1237, 1 : 2000), α-amylase (SIGMA-Aldrich A8273, 1 : 200), Krt19 (TROMA III Developmental Studies Hybridoma Bank, 1 : 200), Muc1 (CT-1,14 1 : 500), E-Cadherin (BD Laboratories 610182, 1 : 1000), GFP (Roche clones 7.1 and 13.1, 1 : 500), activated-Caspase3 (R&D AF835, 1 : 300), Ki67 (Leica Novocastra, 1 : 1000), perilipin A/B (SIGMA-Aldrich P1873, 1 : 1000). HPRT-conjugated secondary antibodies were from DAKO. For multiple immunofluorescence, species-specific Alexa-conjugated antibodies were used (Invitrogen). Nuclei were counterstained with DAPI. Images were pseudo-coloured using LEICA Application Suite.
Tissue fragments of 1–2 mm3 were fixed with 4% glutaraldehyde in Na-cacodylate buffer, postfixed with osmium tetroxide, dehydrated and embedded in Durcupam (Araldit). Toluidine blue-stained sections were used to select areas of interest. Thin sections mounted on grids were stained with uranyl acetate and lead citrate and examined using a Jeol 10.10 electron microscope.
Gene expression analysis
Total RNA was extracted from pancreas using guanidine-thyocianate buffer, followed by acid phenol-chloroform. RNA was treated with DNase I (Ambion DNA-free kit, Invitrogen) and converted to cDNA using TaqMan reverse transcription reagents (Applied Biosystems). Quantitative PCR was performed using SYBR-green mastermix in a Prism 7900HT instrument (Applied Biosystems). Primers (online supplementary table S1) were designed using Primer3Plus and reactions were done in triplicate. Quantifications were normalised to endogenous Hprt, using the ΔΔCt method.
Pancreatic protein extracts were prepared in Laemmli buffer and sonicated. For SDS-PAGE-western blotting, primary antibodies were: Pparγ (Cell Signalling 2435, 1 : 1000), perilipin A/B (SIGMA-Aldrich P1873, 1 : 800) and β-actin (SIGMA-Aldrich A2228, 1 : 2000).
Sections were stained with the In Situ Cell Death Detection Kit (Roche cat 11 684 795 910). Positive nuclei were counted in 20× microphotographs; at least 10 different fields/animal were scored.
Glycaemia was determined from tail blood using an automatic glucose monitor. For glucose tolerance tests, 2 g/kg of glucose were injected intraperitoneally and glycaemia was measured at the indicated timepoints. Amylase was measured in serum using Reflotron Pancreatic Amylase tests (Roche).
Data are provided as mean±SEM. Statistical analysis was done using two tailed Student's t-test and p<0.05 was considered significant. All statistical analyses were performed using VassarStat.net.
To assess the role of Gata6 in pancreatic development and homeostasis, we generated Gata6loxP/loxP;Ptf1a-Cre+/KI mice (Gata6P−/−) (figure 1A) in which Cre-mediated reporter excision is detected at e10-e10.5.12 Gata6P−/− mice were born at the expected Mendelian frequency and appeared healthy. Gata6 was detected in normal adult acinar, ductal and endocrine cells (figure 1B) and it was efficiently and selectively abrogated in all epithelial cells in Gata6P−/− pancreata (figure 1B and online supplementary figure S1A). Gata6 mRNA levels were 12% of Ptf1a-Cre+/KI control mice (Ctrl) (p=0.003) (figure 1B).
Gata6P−/− pancreata were histologically indistinguishable from controls at birth (not shown), indicating that—after e10.5—Gata6 is dispensable for pancreatic development. At 5 weeks, mild oedema, occasional adipocytes and focal ADM were present. Gata6P−/− pancreata became progressively atrophic and by 30 weeks they were almost completely replaced by fat (figure 1C,D). At 30 weeks, massive atrophy, extensive fat replacement, ADM and mucinous metaplasia occurred (figure 1D). A few acinar remnants persisted and islets of Langerhans appeared unaffected (figure 1D). Gata4 expression was preserved in Gata6P−/− pancreata, indicating that Gata4 cannot replace Gata6 in the adult pancreas (online supplementary figure S1B).
Despite the dramatic histological changes, Gata6P−/− mice appeared healthy, did not present diarrhoea or steatorrhoea and gained weight as controls, even at old age (>60 weeks) (figure 2A). Basal and fasting glycaemia and response to glucose overload were normal (figure 2B,C), indicating preservation of endocrine function. Effective recombination in endocrine cells was demonstrated using immunohistochemistry with antibodies detecting Gata6 in Gata6P−/− pancreas (figure 1B) and with antibodies detecting EYFP in Rosa26LSL-EYFP;Gata6P−/− (EYFP;Gata6P−/−) pancreas (online supplementary figure S1A).
To assess the exocrine function in more detail, we used RT-qPCR. Ptf1a and Rbpjl encode the tissue-specific components of Pancreas Transcription Factor 1 (PTF1), a master regulator of acinar differentiation.15 A significant reduction of their transcripts was observed in 8-week-old Gata6P−/− mice (38% and 22% of controls, respectively; p<0.01 and p<0.001) (figure 3A). mRNA levels of PTF1 target genes coding for digestive enzymes (Amy2a, CtrB, CpA, Pnlip) were also significantly reduced (24%, 35%, 15%, 18% of controls, respectively; p<0.01, all comparisons) (figure 3A); such reduction cannot be explained by the mild atrophy observed at this age (figure 1D). ChIP experiments showed Gata6 binding to Rbpjl and Pnlip promoters, indicating that Gata6 is a novel regulator of the acinar programme (figure 3B). The few remaining acinar cells in 30-week-old Gata6P−/− pancreata showed stronger immunostaining of amylase, CpA, and CtrB1 than controls and contained large granules loaded with digestive enzymes (figure 3C and supplementary figure S2A). Serum amylase levels were significantly higher in young (<20 weeks) Gata6P−/− mice than in controls (2.6-fold increase, p<0.001); these differences disappeared in older mice (figure 3D) despite the dramatic loss of exocrine tissue, suggesting abnormal secretion. To assess polarity we analysed expression of apical Muc1 and of basolateral E-cadherin. These proteins showed a mutually exclusive expression pattern in control mice whereas they showed partial colocalisation and intracellular accumulation in young Gata6P−/− mice, indicating altered apico-basal polarity (online supplementary figure S2B) which contribute to a higher leakage of digestive enzymes into plasma. Mist1 is necessary for full maturation and polarity of acinar cells16: Mist1 mRNA and protein levels were significantly reduced by 8 weeks in Gata6P−/− mice (38% of Ctrl, p=0.003; figure 3A and online supplementary figure S2C) and ChIP confirmed that Gata6 binds the Mist1 promoter (figure 3B). These findings indicate that Gata6 is required for normal acinar differentiation and enzyme secretion and for the maintenance of the acinar compartment in the adult. Multiple mechanisms contribute to this phenotype including regulation of PTF1 complex components and Mist1, and direct regulation of digestive enzyme gene promoters.
To examine the mechanisms leading to the massive acinar loss, we assessed proliferation, apoptosis, autophagy and differentiation. Gata6P−/− pancreata showed significantly more Ki67+ acinar cells than controls both at 5 (2.0-fold, p=0.026) and 30 weeks (16.1-fold, p=0.004) (figure 4A); comparable results were obtained for BrdU uptake (not shown). Apoptosis was measured using TUNEL: while low, the number of apoptotic nuclei increased significantly both at 5 (2.75-fold, p=0.006) and at 30 weeks (2.09-fold, p=0.002) in Gata6P−/− mice (figure 4B). Similar results were obtained by immunostaining for activated Caspase3 (not shown). There were no differences in LC3+ cells, indicating that autophagy does not play a major role in acinar atrophy (not shown).
The most striking histological feature of Gata6P−/− pancreata was a massive accumulation of fat. Ultrastructural analysis at 8 weeks showed the presence of lipid droplets in acinar cells (figure 5A). Lipid droplet-containing E-cadherin+ acinar cells and amylase-expressing adipocytes were also identified at 15 weeks (figure 5B). To genetically trace adipocytes, we used EYFP;Gata6P−/− mice: at 8 weeks, cytoplasmic and nuclear EYFP was detected in all epithelial cells (figure 5C). At 30 weeks, adipocytes co-expressing EYFP and the lipid droplet marker perilipin were observed (figure 5C), indicating that some adipocytes have an epithelial origin. By contrast, EYFP was never detected in the mesenteric white adipose tissue of EYFP;Gata6P−/− mice (figure 5C) or in the fat surrounding Rosa26LSL-EYFP;Ptf1a-Cre+/KI control pancreata (not shown). Transdifferentiation of acinar cells into adipocytes has been described in MycP−/− mice.17 Myc mRNA levels were reduced in Gata6P−/− pancreata at 8 weeks (34% of controls, p=0.002; figure 5D) and adipocyte marker expression was detectable prior to the reduction of Myc. In addition to Myc downregulation, other mechanisms may contribute to the adipocyte transdifferentiation observed in Gata6P−/− pancreata. Upregulation of the master adipogenesis regulator Pparγ (2.1-fold, p=0.021) and its targets Plin1 and Plin4 (19.6-fold, p<0.001; 10.4-fold, p<0.001, respectively) was detected by 8 weeks (figure 5E). Increased expression of Pparγ and perilipin proteins was confirmed (figure 5F).
ADM, a process whereby acinar cells dedifferentiate and acquire ductal features,18 was also prominent in Gata6P−/− mice (figure 1D). Accordingly, in 8-week-old mice reduced levels of digestive enzyme transcripts (figure 3A) were accompanied by increased Krt7 (ductal-specific) mRNA levels (2.8-fold, p=0.001) (figure 6A). Lineage tracing could not be used to assess the acinar origin of ductal-like cells in Gata6P−/− pancreata because the Ptf1a-driven Cre recombinase is active in common precursors. Intermediate cells co-expressing amylase (acinar-specific) and Krt19 (ductal-specific) were found in Gata6P−/− pancreata after 5 weeks and were frequent at 30 weeks (figure 6B). We therefore conclude that apoptosis, adipocyte transdifferentiation, and ADM contribute to acinar cell loss. Increased acinar proliferation suggests a compensating response.
We demonstrate here that Gata6 is dispensable for pancreas development after e10–e10.5 but is critical for complete acinar differentiation and for the maintenance of acinar cells in the adult pancreas. Unlike proposed,6 Gata6 is dispensable for endocrine differentiation and function after e10.5. The fact that Gata6 is dispensable for exocrine pancreas formation but is required for its maintenance indicates the existence of active mechanisms promoting cell survival. The integrity of such mechanisms is necessary for tissue homeostasis in the adult, where cell proliferation is very low, and during aging. Furthermore, the hyperproliferation observed cannot sustain tissue integrity.
The finding that Gata6 is dispensable for pancreas development in this mouse model is not at odds with the pancreatic agenesis observed in patients with GATA6 haploinsufficient mutations.10 This difference suggests that type of mutation, timing and tissue-specificity of Gata6 ablation—among others—account for this phenotype: Cre recombinase driven from the Ptf1a locus is expressed selectively in the pancreas shortly after locus activation, whereas GATA6 mutations in patients can affect endoderm development at earlier time points.
Full differentiation of pancreatic acinar cells implies abundant expression of digestive enzymes, the establishment of apico-basal polarity, and tightly regulated exocytosis.19 We reveal novel roles of Gata6 in all these processes through direct activation of PTF1 components, binding to digestive enzyme gene promoters, and regulation of cell polarity and exocytosis at least in part through Mist1. Gata6 therefore adds to the list of transcription factors that are required for the fine-tuning of acinar differentiation while not being essential for acinar cell formation. Among these are Mist1,16 Xbp1,20 Nr5a2,21 and Hnf1a.22 A striking feature of the phenotype of Gata6P−/− mice is the lack of overt exocrine insufficiency when fed a normal diet. This observation further supports the notion that the exocrine pancreas has a wide functional reserve and that major changes in acinar cell function may not be clinically evident. This finding has broad implications regarding the recognition of both chronic pancreatitis and exocrine insufficiency in patients when assessing genotype-phenotype correlations.
Acinar cells are plastic and can be reprogrammed into other differentiated cell types, both pancreatic (ductal cells,23 β-cells24) and non-pancreatic (hepatocytes,25 adipocytes,17 fibroblasts26) and here we show that Gata6 is a gatekeeper of acinar cell differentiation whose loss favours acinar cell plasticity. Both Gata6P−/− mice and c-MycP−/− mice show epithelial-adipocyte transdifferentiation; interestingly, moderate expression levels of c-Myc mRNA were retained in Gata6P−/− pancreata. MycP+/− have not been reported to have an obvious phenotype, suggesting that other mechanisms—in addition to c-Myc reduction—might be acting in the Gata6P−/− pancreata. Gata2 and Gata3 are known to repress adipogenesis,27 ,28 partly through direct binding to the Pparγ promoter, suggesting that Gata6 might similarly repress this programme in normal pancreas. Only a proportion of adipocytes in the pancreas of Gata6P−/− mice had an epithelial origin as determined by lineage tracing, either because of insufficient sensitivity of the assays or because other mechanisms might also contribute to the fat replacement phenotype. These observations call for further studies in this area taking into account that fat is commonly found in the pancreas of old invididuals.29 The pathophysiological relevance of this process is demonstrated by the fact that adipocytes can promote local inflammation and thereby contribute to the severity of acute pancreatitis.30
ADM, which was also present in the pancreas of Gata6P−/− mice, can be activated in vitro and in vivo upon stress18 and it is an initial step of mutant KRas-driven tumorigenesis.31 ,32 Persistent ADM is associated with increased proliferation in mutant KRas knock-in mice31 and may play a role in the hyperproliferation observed in Gata6P−/− mice. GATA6 has been proposed to participate in the development of ductal adenocarcinoma (PDAC) in humans.33–35 Low GATA6 expression identifies a subset of PDAC characterised by quasi-mesenchymal features and resistance to KRAS knockdown.36 In addition, GATA6 participates in Wnt pathway regulation in this tumour.37 The mouse models reported here may provide clues about the mechanisms involved therein. It will also be important to determine whether genetic variation in GATA6 is associated with diseases of the exocrine pancreas in humans.
The authors thank E. Batlle, M. Barbacid, M.C. Guerra, P. Herrera and M. Serrano for valuable contributions, and Y. Cecilia A. Cortés for technical help, CNIO core facilities for support, and investigators who provided reagents or mice.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
Contributors PM: designed and performed most of the experiments, analysed data, wrote the paper. MC: performed experiments and contributed to data analysis. NdP: provided technical support and performed experiments. FM: performed experiments and analysed data. AZ: performed experiments and analysed data. FXR: directed the research, contributed to experimental design, data analysis, wrote the paper and obtained financial support.
Funding This work was supported, in part, by grants SAF2007-60860, SAF2011-29530 and ONCOBIO Consolider from Ministerio de Ciencia e Innovación (Madrid, Spain) and grant 256974 from European Union Seventh Framework Programme to FXR. PM is recipient of a Juan de la Cierva grant from Spanish Ministry of Science and Innovation and FM is recipient of a La Caixa PhD Fellowship.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement The authors would be glad to share the detailed experimental results included here as well as data not shown reported in the text with other scientists upon their written request to the corresponding author.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.