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

Histological, immunohistochemical and transcriptomic characterization of human tracheoesophageal fistulas

Esophageal atresia (EA) and tracheoesophageal fistula (TEF) are relatively frequently occurring foregut malformations. EA/TEF is thought to have a strong genetic component. Not much is known regarding the biological processes disturbed or which cell type is affected in patients. This hampers the detection of the responsible culprits (genetic or environmental) for the origin of these congenital anatomical malformations. Therefore, we examined gene expression patterns in the TEF and compared them to the patterns in esophageal, tracheal and lung control samples. We studied tissue organization and key proteins using immunohistochemistry. There were clear differences between TEF and control samples. Based on the number of differentially expressed genes as well as histological characteristics, TEFs were most similar to normal esophagus. The BMP-signaling pathway, actin cytoskeleton and extracellular matrix pathways are downregulated in TEF. Genes involved in smooth muscle contraction are overexpressed in TEF compared to esophagus as well as trachea. These enriched pathways indicate myofibroblast activated fibrosis. TEF represents a specific tissue type with large contributions of intestinal smooth muscle cells and neurons. All major cell types present in esophagus are present-albeit often structurally disorganized-in TEF, indicating that its etiology should not be sought in cell fate specification.

TTF-1 and napsin A do not differentiate metastatic lung adenocarcinomas from primary esophageal adenocarcinomas: proposal of a novel staining panel

CONTEXT

When adenocarcinomas arise within the esophagus, particularly when located away from the gastroesophageal junction, it may be important in some patients to differentiate between a primary esophageal adenocarcinoma and metastasis from another site. Lung adenocarcinoma is one tumor that has been reported to frequently metastasize to the esophagus.

OBJECTIVES

To create a panel of immunohistochemical markers that can reliably distinguish between an esophageal and pulmonary primary; within the gastrointestinal pathology literature, including published articles and textbooks, common lung immunohistochemical markers, such as TTF-1, are assumed to be negative in esophageal adenocarcinoma, yet, to our knowledge, no study has yet investigated the veracity of that presumption.

DESIGN

In this study, 24 cases each of pulmonary and esophageal adenocarcinomas were stained with TTF-1, napsin A, CDX2, 34βE12, N-cadherin, and IMP3 in an attempt to define an optimal panel for differentiation. Esophageal adenocarcinomas occurring at the gastroesophageal junction were excluded in this study because a gastric primary tumor cannot be excluded in those cases.

RESULTS

Surprisingly, TTF-1 and napsin A were positive in similar proportions of tumors from both sites. Those markers that differentiated statistically between esophageal and pulmonary adenocarcinoma were IMP3, CDX2, and N-cadherin.

CONCLUSIONS

When differentiating the origin of a tumor as either esophageal or pulmonary, an immunohistochemical panel consisting of IMP3, CDX2, and N-cadherin is superior to either TTF-1 or napsin A.

Adriamycin-Induced Models of VACTERL Association

Animal models are of great importance for medical research. They have enabled analysis of the aetiology and pathogenesis of complex congenital malformations and have also led to major advances in the surgical and therapeutic management of these conditions. Animal models allow us to comprehend the morphological and molecular basis of disease and consequently to discover novel approaches for both surgical and medical therapy. The anthracycline antibiotic adriamycin was incidentally found to have teratogenic effects on rats, producing a range of defects remarkably similar to the VACTERL association of congenital anomalies in humans, providing a reproducible animal model of this condition. VACTERL association is a spectrum of birth defects which includes vertebral, anal, cardiovascular, tracheo-oesophageal, renal and limb anomalies. In recent years, adriamycin rodent models of VACTERL have provided valuable insights into the pathogenesis of this complex association, particularly in relation to tracheo-oesophageal malformations. The adriamycin rat model and adriamycin mouse model are now well established in the investigation of the morphology of faulty organogenesis and the regulation of gene expression in tracheo-oesophageal anomalies.

The lung primordium an outpouching from the foregut! Evidence-based dogma or myth?

BACKGROUND/PURPOSE

This study examines and challenges the "evidence-based legitimacy" of the theory, "the lung primordium is an outpouching from the foregut."

METHOD

A literature review was undertaken using computer database, journals, and relevant anatomical and embryological texts.

RESULTS

The independent path of development taken by the tracheobronchial system and the oesophagus once identified as separate entities; the lack of morphologic, molecular, biological, and genetic supportive evidence for the "common-origin" theory; the distinct longitudinal line of demarcation between the nonsegmented muscles of the esophagus and the highly segmented cartilaginous structure of the tracheobronchial tree; the absence of a tracheoesophageal septum in the process of separation; the differences in epithelial lining; and the diametrically opposed mucociliary cascade of the upper airway vs the mucociliary escalator of the tracheobronchial tree all seriously challenge the authenticity of a common origin to these 2 entities.

CONCLUSION

To the extent that the foregut is seen as consisting of 2 separate semitubular splanchnopleuric entities ventrodorsally juxtaposed, it is true that the lung primordium as an outpouching of, and not from, the foregut. This must never be confused with the notion that the esophagus and tracheobronchial tree have a common origin. In fact, they develop from 2 completely separate segments of the trilaminar germ disk, but because of head fold development are brought together to create a common tracheoesophageal chamber that is later separated, facilitated by the prochordal membrane diverticulum.

Intrinsic esophageal innervation in esophageal atresia without fistula

Esophageal atresia and tracheo-esophageal fistula (EA + TEF) are often associated with malformations of neural crest origin. Esophageal innervation is also derived from the neural crest and it is abnormal in EA + TEF in which there is motor dysfunction. Our aim was to examine the intrinsic esophageal innervation in children with isolated EA in which different embryogenic mechanisms might be involved. Specimens from the proximal and distal esophageal segments of 6/35 patients who had esophageal replacement for isolated EA between 1965 and 2006 were suitable for the study. They were sectioned and immunostained with anti-neurofilament (NF) and anti-S-100 antibodies. The muscle and neural surfaces on each section were measured with the assistance of image processing software. The surface of the ganglia and the number of neurons per ganglion were determined at high power microscopy. The findings were compared with those of six autopsy specimens from newborns dead of other diseases by means of standard statistical tests and a significance threshold of P < 0.05. Unmatched age/size of babies in isolated EA and control groups precluded comparison of the relative surfaces occupied by neural elements. Patients with pure EA had denser fibrilar network and larger ganglia than controls. The number of neurons/ganglion were similar in both groups although the cells from EA patients were larger. The findings were similar at both esophageal levels studied. In spite of methodologic biases, it seems that intrinsic esophageal fibrilar network is denser and the intramural ganglia larger with larger cells in patients with pure EA than in controls on both esophageal ends of the organ. These neural anomalies are only in part reminiscent of those described in regular EA/TEF but may as well explain esophageal dysfunction in patients with repaired isolated EA.

Expression of homeotic genes Hoxa3, Hoxb3, Hoxd3 and Hoxc4 is decreased in the lungs but not in the hearts of adriamycin-exposed mice

Exposure of rat and mouse embryos to adriamycin (doxorubicin chlorhydrate) induces esophageal atresia (EA) and VACTERL association. Sonic hedgehog (Shh) and Gli2/Gli3 pathways are involved in these conditions and knockout mice for homeotic Hox genes Hoxa3, Hoxb3, Hoxc3, Hoxc4 and Hoxa5 show phenotypes with some of the associated VACTERL features. This study aims at evaluating the possible influence of Hoxa3, Hoxb3, Hoxd3 and Hoxc4 as upstream regulators of this complex signalling. Pregnant mice were exposed either to 4 mg/kg of adriamycin (EA group) or vehicle (controls) on embryonic days 7.5 and 8.5. Embryos were recovered at four endpoints (E12.5-E15.5) and randomly assigned for immunohistochemical or molecular biology studies. Lungs and hearts were separately harvested and processed for Hoxa3, Hoxb3, Hoxd3 and Hoxc4 quantitative RT-PCR measurements. Antibodies for Hoxa3, Hoxb3 and Hoxd3 proteins were used for immunohistochemical studies. RT-PCR studies showed a drastic and statistically significant decrease of the four genes in the lungs of EA mice when compared to controls, with a slight recovery from E15.5. Hearts of both groups showed a similar expression of all the genes throughout gestation. Control embryos expressed the hox3 paralogous genes in heart, skin, foregut derivatives and their surrounding mesoderm through E12.5-E15.5 whereas adriamycin-exposed embryos showed a severe decrease in expression of these three proteins in the same tissues but not in the heart. Adriamycin drastically reduced the expression of Hoxa3, Hoxb3, Hoxd3 and Hoxc4 in mice embryonic lungs. Their expression in the heart did not seem to be influenced by adriamycin in this experimental setting.

Faulty bone morphogenetic protein signaling in esophageal atresia with tracheoesophageal fistula

BACKGROUND

The organogenesis of esophageal atresia with tracheoesophageal fistula remains unclear. We have previously demonstrated that the fistula tract develops from a trifurcation of the embryonic lung bud and displays pulmonary lineage traits. Unlike the lung, the fistula grows without branching. Bone morphogenetic proteins (BMPs) are known to be important in lung branching. We studied possible BMP signaling defects as a potential cause for the absence of branching in the fistula tract.

METHODS

Adriamycin was administered to pregnant rats on days 6-9 of gestation to induce tracheoesophageal fistula. Microdissection was performed at E13 and E17 isolating the foregut. Tissues were analyzed using immunohistochemistry for BMP ligand (BMP2, BMP4, BMP7) and receptor (BMPRIA, BMPRIB, BMPRII) expression.

RESULTS

Immunohistochemistry revealed the presence of all 3 BMP ligands at E13, localized specifically to the esophageal mucosa but absent in the fistula and lung. At E17, the ligands were again present in the esophageal mucosa, and additionally in the fistula tract mucosa, but remained absent in the lung. At E17, all of the BMP receptors were also localized to the luminal surface of esophagus and fistula. However, in the lung epithelium, only BMPRII was found, whereas BMPRIA and BMPRIB remained absent.

CONCLUSIONS

The normal expression pattern of BMP4 was increased at the branch tips and low between branches. Among other results, we show here a constant expression level of BMP ligands throughout the entire epithelium of the fistula tract. This diffuse expression suggests defective BMP signaling in the fistula tract and explains its nonbranching phenotype.

Bone morphogenetic protein expression patterns in human esophageal atresia with tracheoesophageal fistula

The organogenesis of esophageal atresia with tracheoesophageal fistula (EA/TEF) remains unknown. The fistula tract appears to develop from a non-branching trifurcation of the embryonic lung bud. The non-branching growth of the fistula differs from the other lung buds and suggests a deficiency in bone morphogenetic protein (BMP) signaling, since BMPs are critical to proper lung development and branching. With IRB approval, portions of newborn human proximal esophageal pouch and distal fistula samples were recovered at the time of surgical repair of EA/TEF. The tissues were processed for immunohistochemistry. Commercially available fetal tissues were used as controls. In control tissues, BMP ligands (BMP 2, 4, and 7) were all present in the esophagus but absent in the trachea. BMPRIA was absent in both tissues. BMPRIB was detected in trachea but not in esophagus and BMPRII was detected in esophagus but not in trachea. In the EA/TEF specimens, all BMP ligands were present in the proximal esophageal pouch but absent in the fistula tract. BMPRIA and BMPRIB were not detected in either tissue. However, BMPRII was found in both fistula tract and proximal pouch. The submucosa of the fistula appears to maintain a mixed (identical neither to lung, esophagus, or trachea) BMP signaling pattern, providing one mechanism which could potentially explain the esophageal dismotility and lack of lung branching seen in the fistula/distal esophagus.

Decrease of parafollicular thyroid C-cells in experimental esophageal atresia: further evidence of a neural crest pathogenic pathway

Adriamycin-induced experimental esophageal atresia (EA) is often associated with malformations of neural crest (NC) origin, such as abnormal pharyngeal pouch derivatives like the thymus and the parathyroids. The aim of the present study was to examine whether NC-derived thyroid C-cells were abnormal in a rat model. Pregnant rats received intraperitoneally either 2 mg/kg Adriamycin (EA) or vehicle (controls) on days 8 and 9 of gestation. Fetuses were recovered on day 21, and blocks including the trachea and thyroid were fixed in formalin, coronally sectioned at 3-mum widths, and stained with standard hematoxylin and eosin until the largest area of thyroid was reached. From this point on, the 1st, 10th, and 20th slices were immunohistochemically stained with anti-calcitonin antibody. Positively-stained cells in each section of the gland were counted using a computer-assisted image analysis method, and the results were averaged. The distribution of the cells within the gland was assessed as well. Comparisons between EA and control rats were made by nonparametric tests with a significance threshold of p<0.05. The number of C-cells was dramatically reduced in EA animals compared with controls (32.4+/-36 vs. 92.3+/-60.5, p<0.001). Histology of the thyroid was similar in both groups, but the distribution of positive C-cells within the gland followed an abnormal pattern in EA rats. Adriamycin causes a pattern of NC-derived malformations, including a severe decrease in thyroid C-cells accompanied by abnormal distribution or migration patterns. These results represent further evidence of the involvement of NC organogenic control dysregulation in the pathogenesis of EA and its associated malformations. The similarities between the rat model and the clinical picture strongly support investigating other subclinical NC-derived anomalies in patients with EA.

Esophageal atresia and other visceral anomalies in a modified Adriamycin rat model and their correlations with amniotic fluid volume variations

The Adriamycin rat model (ARM) has been used to produce visceral malformations in fetuses to explain the mechanisms of foregut division. The models vary in the dosage of Adriamycin (ADR) and in the number of applications. Our study of a modified ARM using 2.2 mg/kg of ADR for 2 days only, intraperitoneally in pregnant rats, is presented. A total of 81 fetuses were obtained with this model from the ADR group, 74 (91%) alive. Uretero-hydronephrosis (UHN) was observed in 70 fetuses (95%), esophageal atresia (EA) in 68 (92%), duodenal atresia (DA) in 68 (92%), bladder hypoplasia (BH) in 67 (90%), plus other malformations. In evaluating amniotic fluid (AF) volume of the fetuses with EA with tracheo-esophageal fistula (TEF) (group I) and EA without TEF (group II), both associated with bilateral UHN when compared with the control group (group III), groups I and II showed higher AF volume in groups I and II than the control group (group III) did ( p=0.0001). In conclusion, ARM was adequate to produce EA and other visceral malformations. The use of ADR in a higher dosage for a shorter period of time produced better results than those presented in previous literature. The increase of AF volume obtained in fetuses presenting EA plus bilateral UHN strongly suggests, despite ureteral dilatation (urinary obstruction), that a malformed communication may exist between the urinary system and the amniotic cavity, permitting the existence of polyhydramnios that is due to digestive obstruction such as EA and DA.

Complete discontinuity of the distal fistula tract from the developing gut: direct histologic evidence for the mechanism of tracheoesophageal fistula formation

The embryogenesis of tracheoesophageal anomalies remains controversial. The purpose of this study was to better define the embryogenesis of developing esophageal atresia with tracheoesophageal fistula (EA/TEF), with specific attention to the controversial issue of whether a discontinuity exists in the foregut during its development of EA/TEF. Pregnant outbred rats were injected with adriamycin (2 mg/kg i.p.) on days 6-9 of gestation (E6-E9). At E12.5 and 13.5, microdissection of the entire foregut was performed. Foreguts were examined by phase microscopy, and serial, precisely transverse sections were created for hematoxylin and eosin (H&E) staining. Gross microdissection of the developing foregut at E12.5 (n = 9) revealed a blind-ending, bulbous fistula tract arising from the middle branch of the tracheal trifurcation (as seen by direct and phase microscopy). No connection with the gut could be appreciated at E12.5, but by E13.5 (n = 10) there was an obvious connection between the fistula and the stomach. Serial H&E transverse sections also demonstrated a blind-ending fistula tract arising from the trachea at E12.5. This fistula tract was clearly discontinuous from the developing stomach, which appeared much further caudal to the end of the fistula tract. These results strongly support a model of experimental TEF wherein the fistula tract arises from a trifurcation of the trachea, and (only during a specific gestational window between days 12.5 and 13.5) there is discontinuity between the fistula tract and the stomach. By day 13.5, the fistula joins with the stomach anlage. These observations in the developing EA/TEF should help to resolve the controversy about the mechanism of EA/TEF formation.

Dorsoventral patterning in oesophageal atresia with tracheo-oesophageal fistula: Evidence from a new mouse model

BACKGROUND/PURPOSE

The well-established Adriamycin rat model of oesophageal atresia (OA) and tracheo-oesophageal fistula (TOF) complements recently described mouse genetic models in which loss of function mutations in foregut patterning genes, such as Nkx2.1 (Ttf 1), lead to OA/TOF. The authors aimed to integrate the 2 systems by adapting the Adriamycin model to the mouse to study molecular aspects of tracheo-oesophageal development.

METHODS

Pregnant CBA/Ca mice were injected intraperitoneally with 4 mg/kg of Adriamycin on embryonic days 7.5 and 8.5. Embryos and fetuses of various gestational ages were subjected to morphologic or histologic examination. Sections were stained with H & E or processed for immunohistochemistry using an antibody specific for Nkx2.1.

RESULTS

Tracheo-oesophageal malformations were observed in 47% of Adriamycin-treated embryos. Early foregut development was similar in Adriamycin-exposed and control embryos but, by E11.5, many treated embryos had an undivided oesophago-trachea, which gave rise to the lung buds and a fistula to the stomach. The fistula originated from the dorsal aspect of the undivided tube and was negative for Nkx2.1, or showed only transient Nkx2.1 expression, compared to the strongly positive bronchi ventrally.

CONCLUSIONS

The Adriamycin model of OA is adaptable to the mouse. In the absence of tracheo-oesophageal separation, the dorsal fistula retains its nonrespiratory commitment suggesting that dorsoventral patterning of foregut development is undisturbed by Adriamycin exposure.