Patterning of the "distal esophagus" in esophageal atresia with tracheo-esophageal fistula: is thyroid transcription factor 1 a player?

BACKGROUND

We have recently proposed that the "distal esophagus" in esophageal atresia with tracheo-esophageal fistula (EA/TEF) is actually embryologically derived from the middle branch of a trifurcation of the embryonic lung bud, which subsequently grows caudally in the foregut to connect with the developing stomach. We hypothesized that differential mRNA expression of the lung-specific patterning transcription factor, thyroid transcription factor 1 (TTF-1), in the developing fistula tract in TEF relative to the bronchi (the other branches of the lung bud trifurcation) might explain the unique nonbranching pattern of growth of the fistula tract.

MATERIALS AND METHODS

EA/TEF was induced in Sprague-Dawley rat embryos via intraperitoneal injection of 2.2 mg/kg adriamycin into pregnant dams on Days 6-9 of gestation. The foregut from embryos developing EA/TEF and from control embryos (no adriamycin) were isolated on Gestational Days 13.5, 15.5, and 17.5 (term = 21 days). Some were processed for whole-mount in situ hybridization for TTF-1, while others were embedded and sectioned for histologic analysis via in situ hybridization for TTF-1.

RESULTS

The expression of the respiratory-specific transcription factor TTF-1 is conserved in the epithelium of the developing fistula tract in TEF. The pattern of expression of TTF-1 in the fistula tract mirrors the expression in the large airways of the developing lungs, despite the fact that the fistula tract does not form secondary branches to give rise to a lung.

CONCLUSIONS

The fistula tract in TEF is a respiratory-derived structure that expresses the lung-specific transcription factor TTF-1 throughout its development in the foregut. Contrary to the patterning role that it normally plays in the developing lung, TTF-1 does not induce branching morphogenesis in the fistula tract. Thus, the nonbranching pattern of growth of the fistula tract may be attributable to local mesenchymal-epithelial interactions that override TTF-1 patterning activity.

Basement membrane exposure defines a critical window of competence for pancreatic duct differentiation from undifferentiated pancreatic precursor cells

We previously showed that the undifferentiated pancreatic epithelium can differentiate into islets, ducts, or acini depending on its milieu and that laminin is necessary for pancreatic duct formation. Therefore we wanted to study the plasticity of laminin-induced duct differentiation the better to understand mechanisms of pancreatic duct lineage selection induced by basement membrane. Mouse embryonic pancreases were dissected at gestational day 11 (E11.5), and epithelium was isolated from its surrounding mesenchyme. Some epithelia were cultured in a collagen gel devoid of laminin. These epithelia were "rescued" at days 1-7 of culture by transferring them to a laminin-rich matrix (Matrigel) for 7 additional days. Other epithelia were instead first cultured in Matrigel, and then placed into collagen. Immunohistochemistry was performed for insulin, amylase, and carbonic anhydrase II. Pancreatic epithelia rescued from collagen into laminin during days 1-4 after harvest were still able to form ducts, whereas epithelia deprived of laminin for longer than this 4-day window were not. Pancreatic epithelia exposed to laminin for as little as 1 day, and then placed into collagen, still retained the ability to make ducts. Thus there is a clear cut-off in the development of the pancreatic epithelium at E11.5, after which laminin appears necessary to induce duct formation. We believe that such "windows of competence" in embryonic development imply that developmental programs in the embryo allow some flexibility.

Expression and role of laminin-1 in mouse pancreatic organogenesis

Previous studies have suggested that basement membrane alone may induce ductal differentiation and morphogenesis in the undifferentiated embryonic pancreas. The mechanism by which this induction occurs has not been investigated. Studies of other organ systems such as the lungs and mammary glands, where differentiation has been shown to be induced by basement membrane, have suggested a major role for laminin as a mediator of ductal or tubular morphogenesis and differentiation. We first defined the ontogeny of laminin-1 in the developing mouse pancreas. To determine the specific role of basement membrane laminin in pancreatic ductal morphogenesis and differentiation, we microdissected 11-day mouse embryonic pancreatic epithelium free from its surrounding mesenchyme and then suspended the explants in a 3-dimensional organ culture to allow us to assay cell differentiation and morphogenesis. When the pancreatic epithelium buds off the foregut endoderm, the pancreatic mesenchyme diffusely expresses laminin-1. This laminin subsequently organizes to the interface between the epithelium and the mesenchyme by E12.5. As gestation progresses, epithelial cells in direct contact with laminin-1 seem to differentiate into ducts and acini, whereas those spared intimate contact with laminin-1 appeared to organize into islets. Although basement membrane gel could induce pancreatic ductal morphogenesis of embryonic pancreatic epithelium, this induction was blocked when we added neutralizing antibodies against any of the following: 1) laminin (specifically laminin-1), 2) the "cross-region" of laminin-1, and 3) the alpha6 moiety of the integrin receptor, which is known to bind laminins. Immunohistochemistry, however, showed that pancreatic duct cell-specific differentiation (carbonic anhydrase II) without ductal morphogenesis was still present, despite the blockage of duct morphogenesis by the anti-laminin-1 neutralizing antibodies. Interestingly, there appeared to be a decrease in carbonic anhydrase II expression over time when the epithelia were grown in a collagen gel, rather than in a basement membrane gel. The pattern of laminin-1 expression in the embryonic pancreas supports the conclusion that laminin-1 is important in the induction of exocrine (ducts and acini) differentiation in the pancreas. Furthermore, our data demonstrate that 1) pancreatic ductal morphogenesis appears to require basement membrane laminin-1 and an alpha6-containing integrin receptor; 2) the cross-region of basement membrane laminin is a biologically active locus of the laminin molecule necessary for pancreatic ductal morphogenesis; 3) duct-specific cytodifferentiation, in the form of carbonic anhydrase II expression, is not necessarily coupled to duct morphogenesis; and 4) the basement membrane gel may contain components (e.g., growth factors) other than laminin-1 that can sustain both carbonic anhydrase II expression and, possibly, the capacity to form ducts, despite the absence of duct structures.

In vitro validation of duct differentiation in developing embryonic mouse pancreas

BACKGROUND

Early embryonic pancreatic epithelia have the capacity for either endocrine or exocrine lineage commitment. Recent studies demonstrated the pluripotential nature of these undifferentiated cells. Isolated pancreatic epithelia grown under the renal capsule formed primarily islets. However, when these same epithelia were grown in a basement-membrane-rich gel (Matrigel) they formed mostly ducts. Currently, there is no model for in vitro pancreatic duct formation and therefore, the mechanism of duct morphogenesis has never been described. The purpose of this study was to provide such a model by characterizing the expression of two duct markers, carbonic anhydrase II (CAII) and the cystic fibrosis transmembrane conductance regulator (CFTR), in isolated undifferentiated pancreatic epithelia grown in vitro.

MATERIALS AND METHODS

We microdissected embryonic pancreases at Embryonic Days (E)9.5-11.5 and performed RT-PCR for CAII and CFTR on E9.5 whole pancreases, E10. 5 and E11.5 epithelia, as well as E11.5 epithelia grown for 7 days in Matrigel. Next we performed in situ hybridization for CAII and CFTR and immunohistochemistry for CAII on E11.5 epithelia grown for 7 days in Matrigel.

RESULTS

Early, undifferentiated embryonic pancreatic epithelium does not express CAII and CFTR by RT-PCR. When E11.5 epithelia were grown for 7 days in Matrigel, however, gene expression for both markers is upregulated as ducts form. Furthermore, CAII was seen by IHC and both CAII and CFTR were seen by in situ hybridization in the ducts after 7 days in Matrigel.

CONCLUSIONS

These data validate our in vitro system as a model for studying the mechanism of normal pancreatic duct differentiation and may potentially help us to understand the faulty mechanism involved in pancreatic ductal carcinogenesis.

Ontogeny of activin B and follistatin in developing embryonic mouse pancreas: implications for lineage selection

Activin, a member of the transforming growth factor-beta superfamily, has been shown to be a critical regulator in exocrine and endocrine pancreas formation. The purpose of our study was to describe the ontogeny of activin B and its inhibitor, follistatin, in developing pancreas and to elucidate potential mechanisms for exocrine and endocrine lineage selection. Mouse embryonic pancreata were dissected at various ages (day 10 [E10.5] to birth [E18.5]), sectioned, and immunostained for activin B (one of two existing isomers, A and B), follistatin, insulin, and glucagon. In addition, reverse transcriptase-polymerase chain reaction was employed to determine the messenger RNA expression of follistatin in isolated pancreatic epithelia and mesenchyme of various ages. Activin B was first detected at E12.5 in epithelial cells coexpressing glucagon. At E16.5 these coexpressors appeared as clusters in close proximity to early ducts. By E18.5 activin B was localized to forming islets where cells coexpressed glucagon and were arranged in the mantle formation characteristic of mature alpha cells. Follistatin was found to be ubiquitous in pancreatic mesenchyme at early ages by immunohistochemical analysis, disappearing sometime after E12.5. Follistatin reappeared in E18.5 islets and remains expressed in adult islets. Follistatin messenger RNA was first detected in epithelium at E11.5, preceding its protein expression in islets later in gestation. We propose that mesenchyme-derived follistatin inhibits epithelium-derived activin at early embryonic ages allowing for unopposed exocrine differentiation and relative suppression of endocrine differentiation. At later ages the decrease in the amount of mesenchyme relative to epithelium and the subsequent drop in follistatin levels liberates epithelial activin to allow differentiation of endocrine cells to form mature islets by the time of birth.

Transforming growth factor-beta 1 in the developing mouse pancreas: a potential regulator of exocrine differentiation

Transforming growth factor-beta 1 (TGF-beta 1) is known to regulate cell growth, differentiation, and function in developing mammalian systems. Altering TGF-beta 1 expression in the developing pancreas has been shown to affect both exocrine and endocrine development, suggesting that it is an important regulator of pancreatic organogenesis. We proposed to examine the ontogeny of TGF-beta 1 mRNA expression in the developing pancreas, as well as characterize the patterns of relative TGF-beta 1 gene expression and activity. We performed in situ hybridization for TGF-beta 1 on pancreas specimens obtained from CD-1 mice on gestational days 12.5 (E12.5), 15.5 (E15.5), and 18.5 (18.5). We also isolated mRNA from the pancreas on each of these days and performed a semi-quantitative reverse transcriptase polymerase chain reaction (RT-PCR) to assess relative TGF-beta 1 expression as a function of gestational age. Finally, we performed a TGF-beta 1 ELISA with media conditioned by embryonic pancreas from gestational days 15.5 and 18.5. By in situ hybridization, TGF-beta 1 mRNA is expressed exclusively in the E12.5 pancreatic epithelium, sparing the surrounding mesenchyme. As pancreatic organogenesis progresses, TGF-beta 1 mRNA expression localizes predominantly to the developing acini. TGF-beta 1 gene expression appears modest through E15.5 but is upregulated near the end of gestation, at E18.5. TGF-beta 1 activity, by ELISA, is also upregulated at E18.5. TGF-beta 1 may thus be a modulator of pancreatic organogenesis. Modest TGF-beta 1 expression through E15.5 may be permissive for exocrine lineage selection. TGF-beta 1 expression may then become critical for terminal acinar differentiation. Upregulated TGF-beta 1 expression at the end of gestation may be important for islet formation, and it may be necessary to inhibit continued proliferation and differentiation of pluripotent cells within the pancreatic ductal epithelium.

Defective epithelial-mesenchymal interactions dictate the organogenesis of tracheoesophageal fistula

We have previously suggested that the fistula tract in esophageal atresia with tracheoesophageal fistula (EA/TEF) arises from a trifurcation of the embryonic lung bud. Thus, it appears to be a respiratory-derived structure, and expresses the lung-specific transcription factor TTF-1 in its epithelium. The fistula tract does not give rise to lungs like the other branches from the bud. It grows caudally until it fistulizes with the stomach. We hypothesized that epithelial-mesenchymal interactions (EMI) dictate the differential pattern of growth of the respiratory-derived fistula tract in EA/TEF. EA/TEF was induced in rat embryos via prenatal exposure to adriamycin. Microdissection was performed on E13.5 embryos to isolate developing lung bud, fistula tract, or esophagus from adriamycin-treated or control animals, respectively. The mesenchyme and epithelium from each of these foregut structures were separated. The individual epithelia were recombined with each of the various mesenchymes and grown in culture. They were assayed for relative degrees of branching. Isolated lung-bud epithelia (LBE) or fistula epithelium were also cultured in Matrigel with exogenous fibroblast growth factors (FGF) and subsequently assayed for branching. The fistula-tract mesenchyme relatively inhibited branching of lung epithelium. The epithelium of the fistula tract could be induced to branch by non-fistula (lung or esophageal) mesenchyme. The fistula-tract and adriamycin-treated LBE both branched in response to FGF1. In contrast, neither responded to FGF7 or FGF10. EMI are defective in the developing EA/TEF. The inability to respond to FGF7 and FGF10 suggests an epithelial defect involving the receptor FGF2R-IIIb, to which these mesenchymal factors obligately bind. Thus, the mesenchyme around the developing fistula tract may lack an FGF branching morphogen(s), such as FGF1. Hence, this mesenchyme is unable to induce branching of respiratory epithelia and allows the middle branch of the embryonic tracheal trifurcation to grow caudally as an unbranched tube until it fistulizes into the stomach.