Fibroblast Growth Factor Signaling in the Developing TEF-2003

Fibroblast Growth Factor Signaling in the DevelopingTracheoesophageal FistulaBy Troy L.Spilde,Amina M.Bhatia,Julie K.Marosky,Barry Preuett,Hiroyuki Kobayashi,Mark J.Hembree, Krishna Prasadan,Erica Daume,Charles L.Snyder,and George K.GittesKansas City,MissouriBackground/Purpose:The Adriamycin-induced rat model of esophageal atresia and tracheoesophagealfistula(EA/TEF) provides a reliable system for the study of EA/TEF pathogen-esis.The authors previously hypothesized that faulty branch-ing lung morphogenesis pathways were a critical component of its pathogenesis.The authors have found evidence for faultyfibroblast growth factor(FGF)signaling related to epithelial-mesenchymal interactions in thefistula tract.To better define FGF signaling,the differential expression of FGF ligands and their receptors between lung,fistula tract, and esophagus are described.Methods:Time-dated pregnant,Sprague-Dawley rats were injected with Adriamycin(2mg/kg intraperitoneally)on days 6through9of gestation.Tissues were processed for histol-ogy and reverse transcriptase polymerase chain reaction. FGF-1,-7and-10were measured from whole lung,fistula tract,and esophagus of TEF or normal embryos.Expression of FGF2RIIIb and FGF2RIIIc receptors was measured in iso-lated epithelium and mesenchyme of lung andfistula tract of TEF embryos as well as lung and esophagus from normal controls.Results:FGF-1mRNA was present in thefistula tract and normal and Adriamycin-exposed lung but absent from whole esophagus.Interestingly,FGF-7mRNA was present only in normal lung.FGF-10was present in all tissues examined. FGF2RIIIb mRNA was absent infistula mesenchyme but present in all other tissues examined.However,the splice variant FGF2RIIIc mRNA was present in all tissues examined. Conclusions:Thesefindings support defective FGF signaling in the rat model of EA/TEF.Absence of FGF-7mRNA in Adriamycin-exposed tissues suggests the primary effect of Adriamycin may be to inhibit FGF-7expression.Moreover, absence of FGF2RIIIb infistula mesenchyme may be caused by loss of positive feedback from FGF-7,its normal obligate ligand.Understanding these specific defects in FGF signaling may provide insight into faulty mechanisms of EA/TEF.J Pediatr Surg38:474-477.Copyright2003,Elsevier Science (USA).All rights reserved.INDEX WORDS:Esophageal atresia,tracheoesophagealfis-tula,rat,fibroblast growth factor.T HE ADRIAMYCIN-INDUCED rat model of esoph-ageal atresia and tracheoesophagealfistula(EA/ TEF)provides a good model system for the study of EA/TEF pathogenesis.Based on previous results showing a respiratory origin for the developingfistula tract(also called distal esophagus)with expression of the respiratory-specific marker,thyroid transcription factor-1(TTF-1),1we hypothesized that faulty lung branching morphogenesis pathways were a critical component of the pathogenesis of EA/TEF.Dominant-negativefibroblast growth factor receptor2RIIIb (FGF2RIIIb)mice have completely blocked branching of the airways,but instead the lung buds grow straight caudally toward the diaphragm without branching.2 This straight,nonbranching growth is similar to the caudally growing,nonbranching third lung bud that appears at the tracheal trifurcation on embryonic day12.5(E12.5)and then extends to connect with the distal developing foregut by E13.5to form the TEF. We have shown by axial sectioning of E12.5foregut specimens that thefistula tract is not continuous with the developing foregut early in formation of the anomaly.3MATERIALS AND METHODS Adriamycin Induction of EA/TEFAfter obtaining protocol approval from our Institutional Animal Care and Use Committee,time-dated pregnant,Sprague-Dawley rats(day 0.5corresponding with noon of the day of discovery of a vaginal plug) were obtained from Charles River Laboratories(Wilmington,MA).On gestational days6through9,the experimental rats were injected intraperitoneally using2mg/kg of Adriamycin from Bedford Labora-tories(Bedford,OH).Control animals were not injected.Litters of embryos from both groups were harvested on gestational day13.5 (E13.5).Using this protocol,a greater than90%incidence of EA/TEF was seen in the experimental groups at age E13.5.Tissues then wereFrom the Division of Pediatric Surgery,Childen’s Mercy Hospital,Kansas City,MO.Presented at the33rd Annual Meeting of the American PediatricSurgical Association,Phoenix,Arizona,May19-23,2002.Address reprint requests to George K.Gittes,MD,Director,Labo-ratory for Surgical Organogenesis,Children’s Mercy Hospital,2401Gillham Rd,Kansas City,MO64108.Copyright2003,Elsevier Science(USA).All rights reserved.0022-3468/03/3803-0037$35.00/0doi:10.1053/jpsu.2003.50082474Journal of Pediatric Surgery,Vol38,No3(March),2003:pp474-477processed for histology and reverse-transcriptase polymerase chain reaction (RT-PCR).mRNA Isolation and Reverse Transcriptase Polymerase Chain ReactionEmbryonic foregut specimens were preserved on ice in Dulbecco ’s Modi fied Eagle ’s Medium (DMEM;Sigma,St Louis,MO)while preparing for microdissection at up to 400ϫpower.Whole compo-nents,consisting of epithelium and mesenchyme,were obtained for Adriamycin-treated lung,fistula tract,as well as normal lung and esophagus.Epithelia and mesenchyme also were isolated for each of these groups.The tissues then were snap-frozen in liquid nitrogen,and mRNA was isolated using an RNeasy isolation kit from Qiagen (Valencia,CA).RNA samples then were reverse transcribed into cDNA using Sensiscript Reverse Transcriptase (Qiagen).Polymerase Chain Reaction (PCR)then was performed with AmpliTaq Gold (Roche,Branchburg,NJ)and PCR Nucleotide Mix (Promega,Madi-son,WI).FGF-1and -7were measured from whole lung,fistula tract,and esophagus of embryos with either the EA/TEF or normal pheno-type.Expression of the lung morphogenic ligand FGF-10,as well as the 2-splice variants of the FGF receptors FGF2R-IIIb (KGF receptor)and FGF2R-IIIc (bek receptor)were measured in the isolated epithelium and mesenchyme of the lung and fistula tract of EA/TEF rat embryos as well as from the lung and esophagus of embryos with the normal phenotype.Oligonucleotide sequences for primers are as follows:FGF2RIIIb sense CAATGTGACGGAGATGGATG,antisense GAT-GACTGTCACCACCATGC;FGF2RIIIc sense GTGCTTGGCGGG-TAATTCTA,antisense GCTTCTTGGTCGTGGTCTTC;FGF-1sense AGATCACAACCTTTGCAGCC,antisense ATGGCTTTCTGGCCG-TAGTG;FGF-7sense ACTCACACAAGCGCACGCTC,antisense TGCGCATTATAGAGCAGTGC;and FGF-10sense GGATACTGA-CACATTGTGCC,antisense ACATTTGCCTGCCGTTGTGC.RESULTSFGF-2receptor IIIb isoform (KGF receptor)mRNA was speci fically absent in the fistula mesenchyme by RT-PCR,but was present in all other tissues (Fig 1).The FGF-2receptor IIIc (bek receptor)mRNA was present in all tissues by RT-PCR (Fig 1).Interestingly,FGF-7was completely absent from the entire fistula,Adriamycin-exposed lung buds,and esophagus,but was present throughout normal lung (Fig 2).FGF-10mRNA wasabsent in the fistula epithelium and the esophagus epi-thelium but was present in normal lung epithelium and all tissue mesenchymes (Fig 3).Lastly,FGF-1mRNA was present in the fistula tract and both normal and Adriamycin-exposed lung,but was absent from whole esophagus (Fig 3).␤-actin was used as a positive control gene (data not shown).DISCUSSIONThese findings support that there is a defect in FGF signaling mechanisms in the Adriamycin-induced rat model of EA/TEF.Of the several members of the FGF family,we chose to speci fically study FGF-1,-7,-10and their recep-tors,FGF2RIIIb (receptor for FGF-1,-7,and -10)and FGF2RIIIc (receptor for FGF-1),because these factors are important to normal branching morphogenesis of the devel-oping embryonic lung.2,4-11The absence of FGF-7mRNA in the Adriamycin-exposed tissues suggests that the primary effect of the adriamycin treatment may be to inhibit FGF-7,creating secondary defects in FGF-10signaling and thus inducing a more esophaguslike phenotype in the fistula tract.These signaling defects possibly could lead to the nonbranching growth of the fistula tract,similar to that seen with the normal esophagus,even thought the fistula tract retains expression of the respiratory-speci fic morphogen TTF-1.1The presence of receptor FGF2RIIIc throughout all tissues may allow for atypical FGF signaling tooccurFig 1.FGF-2Receptor isoforms IIIb (KGF receptor)and IIIc (bek receptor).FGF2RIIIb is specifically absent in fistula mesenchyme (arrow),whereas FGF2RIIIc is present in all tissues.A dashed line separates the receptors.Fist,fistula;epi,epithelium;mes,mesenchyme.Fig 2.FGF-7is present only in the normal lung,suggesting that a primary effect of the Adriamycin treatment may be to inhibit FGF-7signaling,whose obligate receptor is FGF2RIIIb.475FGF SIGNALING IN TEF(eg,FGF-1)in the developing embryonic foregut,per-mitting the nonbranching growth of the fistula tract.FGF-7and -10must signal through receptor FGF2RIIIb,whereas FGF-1may signal through either splice variant 2RIIIb or 2RIIIc.10,12-14FGF-1is not present in the normally developing esophagus,but is present in the fistula tract,suggesting that FGF-1may be the key regulator of the nonbranching growth of the fistula tract in the rat model of EA/TEF.Epithelial-mesenchymal interactions are important in the development of many organ systems.In the contextof the developing EA/TEF,it has been shown that the fistula mesenchyme does not allow for branching of normal lung bud epithelia,whereas the fistula tract epi-thelia is induced to branch when cultured with normal lung mesenchyme.15Moreover,the absence of KGF receptor in the fistula mesenchyme may be caused by the epithelial loss of both of the normal obligate ligands to this receptor (FGF-7and -10).The use of the distal fistula tract as an esophageal replacement continues to be the mainstay of surgical treatment for this human anomaly.We previously have shown evidence for a respiratory origin of the human neonatal fistula tract,which also expresses TTF-1.16Others also have shown that the fistula tract shows tracheobronchial remnants such as abnormal mucus pro-duction in ducts and glands,cartilage,and a disorganized muscular coat as well as unilateral pulmonary agenesis and sequestrations.17-20This respiratory origin of the fistula tract may account for the poor motility seen after the standard EA/TEF repair.We currently are working to de fine the expression of the different fibroblast growth factors in the human neonatal tracheoesophageal fistula.Hopefully,understanding these speci fic defects in FGF signaling in the rodent model of EA/TEF may provide insight into the faulty mechanisms leading to esophageal atresia and tracheoesophageal fistula.Per-haps quanti fication of the different FGF receptors and ligands will help clarify the issue of aberrant signaling.REFERENCES1.Crisera CA,Connelly PR,Marmureanu AR,et al:TTF-1and HNF-3␤in the developing tracheoesophageal fistula:Further evidence for the respiratory origin of the ‘distal esophagus.’J Pediatr Surg 34:1322-1326,19992.Peters K,Werner S,Liao X,et al:Targeted expression of a dominant negative FGF receptor blocks branching morphogenesis and epithelial differentiation of mouse lung.EMBO J 13:3296-3301,19943.Spilde TL,Bhatia AM,Marosky JK,et al:Complete discontinuity of the distal fistula tract from the developing gut:Direct histologic evidence for the mechanism of formation of tracheoesophageal fistula.Anat Rec 267:220-2244.Crisera CA,Maldonado TS,Longaker MT,et al:Defective fibroblast growth factor signaling allows for nonbranching growth of the respiratory-derived fistula tract in esophageal atresia with tracheo-esophageal fistula.J Pediatr Surg 35:1421-1425,20005.Bellusci S,Grindley J,Emoto H,et al:Fibroblast growth factor 10(FGF10)and branching morphogenesis in the embryonic mouse lung.Development 124:4867-4878,19976.Nogawa H,Ito T:Branching morphogenesis of embryonic mouse lung epithelium in mesenchyme-free culture.Development 121:1015-1022,19957.Post M,Souza P,Liu J,et al:Keratinocyte growth factor and its receptors are involved in regulating early lung branching.Development 122:3107-3115,19968.Cardoso W,Itoh A,Nogawa H,et al:FGF-1and FGF-7induce distinct patterns of growth and differentiation in embryonic lung epithelium.Dev Dyn 208:398-405,19979.Shiratori M,Oshika E,Ung L,et al:Keratinocyte growth factor and embryonic rat lung morphogenesis.Am J Respir Cell Mol Biol 15:328-338,199610.Mason I:The ins and outs of fibroblast growth factors.Cell 78:547-552,199411.Mason I,Fuller-Pace F,Smith R,et al:FGF-7(keratinocyte growth factor)expression during mouse development suggests roles in myogenesis,forebrain regionalization,and epithelial-mesenchymal in-teractions.Mech Dev 45:15-30,199412.Celli G,LaRochelle W,Mackem S,et al:Soluble dominant-negative receptor uncovers essential roles for fibroblast growth factors in multi-organ induction and patterning.EMBO J 17:1642-1655,199813.Jaye M,Schlessinger J,Dionne C:Fibroblast growth factor receptor tyrosine kinases:Molecular analysis and signal transduction.Biochem Biophys Acta 1135:185-199,199214.Orr-Urtreger A,Redford M,Burakova T,et al:Developmental localization of the splicing alternatives of fibroblast growth factor receptor-2(FGFR2).Dev Biol 158:475-48615.Crisera CA,Grau JB,Maldonado TS,et al:Defective epithelial-mesenchymal interactions dictate the organogenesis of tracheoesopha-geal fistula.Pediatr Surg Int 16:256-261,200016.Spilde TL,Bhatia AM,Miller KA,et al:Thyroid transcription factor 1expression in the human neonatal tracheoesophageal fistula.J Pediatr Surg 37:1065-1067,200217.Hokama A,Myers NA,Kent M,et al:Esophageal atresia with tracheo-esophageal fistula:A histopathological study.Pediatr Surg Int 1:117-121,1986Fig 3.FGF-1(upper photograph)expression is present in the fistula tract,as well as treated and normal lungs,but is speci fically absent from the esophagus (arrow),suggesting it may be the key regulator of nonbranching fistula growth.FGF-10(lower photograph)is present in all tissues.476SPILDE ET AL18.Dutta HK,Mathur M,Bhatnagar V:A histopathological study of esophageal atresia and tracheoesophagealfistula.J Pediatr Surg35: 438-441,200019.Steadland KM,Langham MR,Greene MA,et al:Unilateral pulmonary agenesis,esophageal atresia,and distal tracheoesophageal fistula.Ann Thorac Surg59:511-513,199520.Lewis JE,Murray RE:Pulmonary sequestration with broncho-esophagealfistula.J Pediatr Surg3:575-579,1968DiscussionM.Schwartz(Wilmington,DE):This is a very nice study,and it points to perhaps an additional mechanism by which these anomalies occur.I am wondering if you have studied or are aware of anyone else who has studied human tissue.Certainly,we can get a piece of thefistula during repair,and your methods and techniques applied to human tissue.T.Spilde(response):Actually,I have not seen any literature supporting PCR results of human tissue;how-ever,there are several historical cases that have shown bronchial remnants to be present in the distalfistula tract and other suggestions of respiratory origin.We actually, in our current human studies,are looking at the FGF pathways.D.Cass(Houston,TX):I have a question that pertains to this study and the previous study that you presented today.With regard to differences in the expression of sonic hedgehog protein and FGF proteins,you have shown differences between the proximal esophageal pouch and the distalfistulous tract.As a control have you had an opportunity to compare the expression of these proteins in the proximal and distal esophagus in those fetuses that were treated with Adriamycin but in whom the TEF malformation did not occur?T.Spilde(response):That is a great question,and actually over90%of the embryos that we have examined have the anomaly and those that do not are usually involuted and unable to dissect easily so the tissue usually is destroyed by the time we are able tofinish the dissection.477FGF SIGNALING IN TEF。