UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Morphometric studies of normal and abnormal primary palate formation in noncleft and cleft lip strains… Wang, Kang-Yee 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1992_spring_wang_kang_yee.pdf [ 2.69MB ]
Metadata
JSON: 831-1.0086710.json
JSON-LD: 831-1.0086710-ld.json
RDF/XML (Pretty): 831-1.0086710-rdf.xml
RDF/JSON: 831-1.0086710-rdf.json
Turtle: 831-1.0086710-turtle.txt
N-Triples: 831-1.0086710-rdf-ntriples.txt
Original Record: 831-1.0086710-source.json
Full Text
831-1.0086710-fulltext.txt
Citation
831-1.0086710.ris

Full Text

MORPHOMETRIC STUDIES OF NORMAL AND ABNORMAL PRIMARY PALATEFORMATIONIN NONCLEFT AND CLEFT LIP STRAINS OF MICEBYKANG-YEE WANGB.D.S., National Taiwan University, 1979A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENT FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(Department of Oral Biology)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAFebruary 1992© Kang-Yee Wang, 1992In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of Oral BiologyThe University of British ColumbiaVancouver, CanadaDate April 15, 1992DE-6 (2/88)ABSTRACTGrowth and development of the primary palate in the human embryo iscomplicated. Failure of the fusion of the maxillary prominence, with the lateralnasal and the medial nasal prominences results in cleft lip. Although manyqualitative descriptions have been published on both normal and abnormalpalatogenesis, there is little definitive information about mechanisms of primarypalate formation. This study will address mechanisms using the mouse as theexperimental model. The embryogenesis of primary palate of mouse is similar tothat of human and spontaneous cleft lip is associated with genotype in both.Therefore the mouse provides an experimental model for studying primarypal atogen esi s.The purpose of my research was to compare primary palate developmentin different strains of mice. Two of the strains studied, BALB/cByJ and C57BLJ6J,have normal primary palate development, and three of the strains studied, A/J,AIWySn and CL/Fr, have stable frequencies of cleft lip. The first part of this studywas to determine the cleft lip frequency and the resorption rate of the three cleftlip strains of mice. The results confirmed previous observations that NJ has alower frequency of cleft lip and a higher resorption rate than that of A/WySn andCL/Fr.The purpose of second part of this study was to determine the stage ofbody development and the chronological age at which primary palatogenesistakes place in these five strains. The results showed that in each strain, thenumber of tail somites was highly correlated with development of primary palatein that strain. Somite development in the strains with genetic cleft lip liability wasapproximately twelve hours later than in the C57BLJ6J strain. However, in thenoncleft BALB/cByJ strain, the somite was also twelve hours later than inC57BLI6J. Thus, delayed chronological development appears not to be a factorcontributing to cleft lip malformation.In the third part of the study, the phases of primary palatogenesis werecompared in the noncleft and cleft lip strains. The results showed fusion of theepithelia to form the nasal fin and replacement of the epithelial seam bymesenchyme was delayed in cleft lip strains relative to tail somite stages. Thestrains with higher cleft lip frequency, A/WySn and CL/Fr, were more delayedthan the NJ strain with lower cleft lip frequency. The time of replacement of thenasal fin by mesenchyme occurred at about 12-13 tail somites in noncleft strains,14 tail somites in NJ, 15 tail somites in NWySn and 16 tail somites in CL/Fr.Forward growth of the maxillary prominence had the same pattern asmesenchymal replacement in the strains. The position of the maxillaryprominence was highly correlated with the size of the nasal fin and themesenchymal component in all strains. One major gene could explain thisdelayed formation in cleft lip strains and maternal effect could explain the moredelayed formation in NWySn and CL/Fr than in NJ strain.The purpose of the fourth part of this study was to determine the tailsomite stage when the oronasal membrane ruptures to form the primary choanadorsal to the primary palate. A definitive primary palate is established as aconsequence of this rupture. Primary choana formation occurred at 18 tailIIIsomites in CL/Fr and 20 tail somites in C57BLJ6J. This earlier occurrence in CL/Frthan in C57BL/6J appears to provide a more limited developmental interval formesenchymal replacement and enlargement in the cleft lip strain. A multifactorialthreshold model was suggested from this study. The tail somite stage ofmesenchymal replacement is applied as a scale of liability for the cleft lipmalformation. Unfavorable growth may move embryos toward the threshold andresult in an increased incidence of cleft lip. The threshold may be affected by thetiming of primary choana formation.ivTABLE OF CONTENTSPAGEABSTRACT iiTABLE OF CONTENTS VLIST OF FIGURES viiiLIST OF TABLES xACKNOWLEDGEMENT xiiINTRODUC11ON 11. Review of the literature 2A. Normal development of the primary palate 3I. The origin of facial mesenchyme 3II. Induction of the nasal organ 4III. Epithelial fusion 6IV. Mesenchymal penetration 8V. Primary choana formation 12B. Abnormal development of the primary palate and cleft lipmalformation 13I. Morphogenesis of cleft lip malformation 13II. Morphometric study of postnatal human cleft lip 14Ill. Morphogenesis of cleft lip in laboratory animals 15C. The genetic causes of cleft lip 16I. One major gene in mouse cleft lip 16II. Maternal effects 17III. Reciprocal relationship between cleft lip andresorption rate 19IV. Uterine site effects 20V. General concept of multifactorial threshold model 20VI. Multifactorial threshold and one major gene inhuman cleft lip 23D. Environmental effects on cleft lip malformation 26I. Hadacidin 26II. 6-aminonicotinamide 27III. Phenytoin 28IV. Oxygen 302. Rationale 31MATERIALS AND METHODS 34A. Cleft lip frequency and resorption rate 34VI. Embryo collection 34II. Data anayss 35B. Stages of primary palate development 35I. Embryo collection 35II. Measurements 36a. Nasal fin 36b. Mesenchymal component 36c. Position of the maxillary prominence 37Ill. Data analysis 37a. Analysis of covariance 37i. Testing adjusted treatment means 38ii. Homogeneity of regression coefficients 43b. Partial least square analysis 44IV. Error of measurement 50C. Stages of primary choanae formation 54RESULTS 55A. Cleft lip frequency and resorption rate 55I. Strain effect 55II. Uterine site effect 55B. The development of chronological age and tail somites of five strainsduring primary palate formation 58C. Delineation of phases of primary palate development 58I. Early primary palate development 60II. Primary palate development with mesenchymal formation 70a. Analysis of primary palate area formation 79b. Partial least squares and covariance analysis of primarypalate area formation 80c. Analysis of mesenchymal bridge formation 85d. Analysis of growth of the maxillary prominence 92Ill. Late primary palate development and primary choanaeopening 95DISCUSSION 102A. Cleft lip frequency and resorption rate 102I. Strain effect 102II. Uterine site effect 103B. Tail somite stage and chronological age 104C. Delineation of phases of primary palate development 105I. Early primary palate development 105II. Primary palate development with mesenchymal formation 106a. Primary palate area formation 106vib. Primary palate area formation and partial least squaresanalysis 11 0c. Temporal and spatial analysis of mesenchymal bridgeformation 111d. Analysis of the growth of maxillary prominence 113e. Multifactorial threshold for primary palate development... 116Ill. Primary choana formation 119CONCLUSIONS 124BIBLIOGRAPHY 130APPENDIX 1 143APPENDIX 2 144APPENDIX 3 149viiLIST OF FIGURESPAGEFigure 1 Falconer’s concept of threshold 21Figure 2 Relation between two-tailed t and one-tailed F 43Figure 3 Diagram for a two-block latent variable model relating 3indicators of development to 5 parameters from theprimary palate 47Figure 4 Error of measurement for the position of the maxillaryprominence from the superior view 51Figure 5 Error of measurement for the position of the maxillaryprominence from the sagittal view 53Figure 6 Tail somite distribution (8 to 18 tail somite stage) comparedwith primary palate development by strains 59Figure 7 Frontal view of C57BL/6J, day 10 hour 17, 8 tail somitestage 61Figure 8 Frontal section of C57BL/6J at the level of nasal fin,day 10 hour 17, 8 tail somite stage 62Figure 9 Frontal section of C57BL/6J at the end of nasal fin,day 10 hour 17, 8 tail somites 64Figure 10 Frontal view of C57BL/6J, day 11 hour 0, 12 tail somitestage 66Figure 11 Frontal section of C57BL/6J at the level of nasal fin,day 11 hour 0, 12 tail somite stage 67Figure 12 Regression lines of right nasal fin depth of C57BLJ6Jand CL/Fr from 8 to 12 tail so mite stage 69Figure 13 Regression lines of the position of the right maxillaryprominence of C57BL/6J and CL/Fr from 8 to 12 tailsomite stage 72Figure 14 Frontal section of C57BL/6J showing nasal epithelium,mesenchymal bridge and oral ectoderm, day 11 hour 0, 13tail somite stage 75viiiFigure 15 Frontal view (A) and a frontal section (B) of CL/Fr embryo,day 11 hour 13, 15 tail so mite stage 76Figure 16 Regression lines of right primary palate area of the fivestrains from 13 to 16 tail somite stage 81Figure 17 Regression lines of the right primary palate area on thelatent variable of development including the tail somite,body weight and position of the right maxillary prominenceof the five strains 84Figure 18 Regression lines of growth of the mesenchymal area ofthe five strains from 12 to 19 tail so mite stage 88Figure 19 Regression lines of right mesenchymal area on latentvariable of development including the tail somite, bodyweight and position of the right maxillary prominenceof the five strains 90Figure 20 Regression lines of the position of the right maxillaryprominence relative to the end of the nasal fin of thefive strains from 13 to 16 tail somite satge 93Figure 21 Regression lines of the right primary palate area on theposition of the right maxillary prominence relative to theend of the nasal fin of the five strains 96Figure 22 Regression lines of right mesenchymal area on the positionof the right maxillary prominence relative to the end of thenasal fin of the three strains 98Figure 23 Summary of hypotheses of cleft lip gene effects andmaternal effects 115Figure 24 A conventional multifactorial threshold showing thatcleft lip may occur in some strains becausemesenchymal formation is relatively late 11 8Figure 25 A diagram illustrating the multifactorial nature of cleft lip 120Figure 26 Outlines of sagittal sections showing the steps in theformation of the primary palate in human embryo stage16,l7andl8 123ixLIST OF TABLESPAGETable 1 Analysis of covariance for a randomized complete-blockeddesign 39Table 2 Cleft lip and resorption rate in embryos of day 13 hr 12 ofthree cleft lip strains 56Table 3 Effects of implantation site on frequency of cleft lip andresorption 57Table 4 Nasal fin depth and the position of the maxillaryprominence of C57BLJ6J and CL/Fr from 8 to 12 tailsomite stage (TS) 65Table 5 Test for homogeneity of slopes of right nasal fin from8 to 12 tail somite stage (TS) between C57BL’6J and CL/Fr 70Table 6 Analysis of covariance of the position of the right maxUlaryprominence of C57BL/6J (C57) and CL/Fr from 8 to 12 tailsomite stage (TS) 73Table 7 The percentage of embryos which have mesenchymereplace the nasal fin of the right side at each tail so mitestage (TS) of the five strains 74Table 8 The correlation coefficients between the right andleft sides of the maxillary prominence, primary palatal andmesenchymal components in noncleft lip strains from 13 to16 tail so mite stage 77Table 9 The correlation coefficients between the right and left sidesof the maxillary prominence, primary palatal andmesenchymal components in cleft lip strains from 13 to16 tail so mite stage 78Table 10 Analysis of covariance of growth of right primary palatearea from 13 to 16 tail somite stage (TS) of the fivestrains (ST) 82Table 11 Two-block partial least squares analysis of the threedevelopmental scores and the five primary palateparameters blocks of the five strains 83xTable 12 Analysis of covariance of growth of right primary palatearea on latent variable of development (LVDEV) of fivestrains (ST) 86Table 13 Analysis of covariance of growth of right mesenchymal areaon tail somite stage (TS) of five strains (ST) 89Table 14 Analysis of covariance of right mesenchymal area onlatent variable of development (LVDEV) of five strains (ST).. 91Table 15 Analysis of covariance of the position of the right maxillaryprominence relative to the end of the nasal fin on tail somitestage (TS) of the five strains (ST) 94Table 16 Analysis of covariance of growth of right primary palatearea on the position of the right maxillary prominence(RMXP) of five strains (ST) 97Table 17 Analysis of covariance of growth of right mesenchymalarea on the position of the right maxillary prominence(RMXP) of five strains (ST) 99Table 18 Percentage of embryos with primary choana opening inembryos of 17-20 tail somite stage in C57BLI6J and CL/Frstrains 100xiACKNOWLEDGEMENTThis study was guided by Dr. Virginia M. Diewert and supported by theMedical Research Council of Canada grant MT 4543 to Dr. Diewert. I would liketo extend my thanks to Mrs. Barbara Tait for her help in animal care aridhistology; Dr. Scott Lozanoff for his instruction in computer analysis; Dr. D. M.Brunette, Ms. E. Robertson and Dr. Fred Bookstein for their statistical assistancewith morphometric data; Mr.B. McCaughey and Mr. H. Traeger for theirphotographic technical help; and Ms. C. Lo, Mr. M. Wong and Mr. D. Nagy forassistance with proof reading the manuscript. I would thank my thesiscommittee members, Dr. D. Juriloff, Dr. M. Todd and Dr. D. Waterfield for theirvaluable comments during this study. I also want to express my appreciation forthe care and support of my wife and family without whom this work could nothave been possible.xiiINTRODUCTIONThe palate develops from two primordia: the primary palate and thesecondary palate. Palatogenesis begins toward the end of the fifth week afterconception in humans and is not complete until about the twelfth week. Thesecondary palate closes later in development than the primary palate. Geneticand environmental factors that influence its closure are different from those thatinfluence the primary palate. Abnormal development of the primary palate,leading to a cleft lip, may interfere secondarily with secondary palate closure.Thus, on both embryologic and genetic grounds, congenital cleft lip (CL) and cleftlip with cleft palate CLP, appear to be etiologicatly related [in data combining thetwo they may be designed CL(P)]. Isolated clefting of the secondary palate (CP) isan etiologically independent entity (Fraser and Baxter 1954; Trasler and Fraser,1963; Woolf etal, 1963; Fraser, 1970).Cleft lip with or without cleft palate [CL(P)] is one of the most commoncraniofacial congenital malformations affecting primary palate development inhuman embryos. There are a large number of syndromes in which CL(P) may beone of the features. For most of these the cause is unidentified. A few areassociated with recognizable chromosomal aberrations, and about a third arecaused by major mutant genes. Each of these syndromes is rare, and togetherthey may account for perhaps 5% of all cases. Most cases of CL(P) withoutassociated malformations in humans appear to be multifactorially determined(Fraser, 1970).1There are striking differences in the CL(P) frequency between races:North American Indians in British Columbia have very high frequencies(2.75/1 000) (Lowry and Renwick, 1969); Orientals have relatively highfrequencies (1.7/1,000 births) (Kobayashi, 1958; Neel, 1958), Caucasians areintermediate (1/1,000) while Blacks tend to have low frequencies (0.4/1 ,000)(Chung and Myrianthopoulos, 1968; Khoury etal, 1983). Although the etiology ofnonsyndromic CL(P) remains unknown (Fraser, 1989), these differences persistin different geographic regions, suggesting that they do not result fromenvironmental alternations. It is thought that they may be associated withdifferences in face shape (Fraser and Pashayan, 1970). Because themultifactorial threshold model, to be discussed latter has been applied to cleft lip(Carter, 1969; Fraser, 1970, 1980), it would be useful to identify some biologicalattribute of liability such as face shape, that could be an indicator of increasedrisk.1. Review of the literatureTo provide a background for these studies, the results from previousinvestigations are presented comprehensively. This includes a review of normaldevelopment of the primary palate and the cleft lip malformation in human as wellas animal models. The genetic causes of cleft lip in human and mouse arediscussed. As three mouse strains used for these studies (NJ, A/WySn andCL/Fr) are genetically predisposed to CL(P) and susceptible to environmentaleffects, the effects of teratogens have also been addressed.2A. Normal development of the primary palateI. The origin of facial mesenchymeBefore any primary palate formation occurs, neural crest cells migratefrom the neural tube to the craniofacial region to form mesenchyme of the facialprominences. The importance of the role of the neural crest cells in forming themesenchyme of the facial prominences was investigated by Johnston (1964,1966) who conducted two significant experiments in chick embryos on this topic.In the first experiment small segments of neural crest from either mid- or forebrainwere labelled with tritiated thymidine in donor embryos, and were then implantedin the corresponding region of host embryos. The results indicate that these cellsmade a significant contribution to the formation of the facial mesenchyme. In asecond experiment, segments of the midbrain neural crest were removed prior tocell migration, and the embryos were then incubated for an additional period.This resulted in severe facial malformations which included an absence of apronounced fronto-nasal prominence and a mandibular prominence. Removal ofthe forebrain neural crest, which normally makes a lesser contribution to facialmesenchyme, frequently resulted in the clefting of the primary palate only.Using interspecific grafts of the neural primordium between quail andchick embryos, Le Lievre and Le Douarin (1975) have shown that themesenchyme of the maxillary prominence and the branchial arches is composedof mesectodermal cells. Noden (1975, 1983) has demonstrated that the aviancrest cells from the posterior mesencephalon migrate en masse away from theneural tube between the epidermis and the underlying mesoderm. Then this3neural crest population migrates and proliferates throughout both the regionventrolateral to the mesencephalon and the future maxillary prominence. Neuralcrest cells migrating from the anterior mesencephalon and posteriordiencephalon move rostrally, and at later stages they overlap with those derivedfrom the posterior mesencephalon, contributing to the formation of the maxillaryprominence. Nichols (1986) has shown that in mouse embryos, neural crestformation and emigration at midbrain-rostral hindbrain neural folds are completeat late 4 to 5 somites of development. The neural crest leaves behind overlyingsquamous epithelium. At approximately 10 somites of development, this neuralcrest mesenchyme is distributed dorsolateral to the pharynx and displacedventromedially in a narrow, transient subectodermal space functionally similar tothat observed in the chick embryo. If the fate of the crest mesenchyme in themouse is similar to those in birds and amphibians, this mesenchyme will formbone, cartilage and connective tissue of the first branchial arch.II. Induction of the nasal organThe induction of the nasal organ starts primary palate formation. At about33 days (stage 15) after conception in humans, the area which will form the nosebecomes induced and begins to elevate forming the medial and lateral nasalprominences. In the Salamander, nasal organ induction involves a succession ofdifferent inductors. In the gastrula and the neurula stages portions of theendoderm and mesoderm act as inductors. The final inductor of the nasal organis a portion of the central nervous system. (Jacobson, 1 963a, 1 963b).4The mesenchyme underlying the ectoderm of the facial region originatesfrom the neural crest cell population (Johnston, 1964, 1966; Le Lievre and LeDouarin, 1975). Its high rate of proliferation (Minkoff and Kuntz, 1977, 1978) isthought to be maintained by an epithelial-mesenchymal interaction. Movement ofmesenchymal tissue from the medial nasal prominence into the base of the nasalgroove and into the medial area of the lateral nasal prominence was observedwhen H thymidine was implanted with a sable hair probe (Patterson et al, 1984;Patterson and Minkoff, 1985). A series of separation and recombinationexperiments involving a variety of tissue configurations in organ cultures fromchick (Saber et al, 1989) suggested that the influence of the epithelium onmesenchyme viability was stage dependent, and epithelial-mesenchymalinteractions appear to evoke, within the mesenchyme, a zonal growth-sustainingeffect. In an effort to determine whether epithelial-mesenchymal interaction mayalso be relevant to the maintenance of the growth rates in the facialprominences, Bailey et a! (1988) have found that cells located deeper within themesenchyme have lower proliferation rates than those closer to the epithelium inthe chick. During the latter stages of development, however, this trend is notobserved. The reason for this mechanism is still unknown.There are many theories as to the kind of macromolecules involved inepithelial-mesenchymal interaction. The presence of serotonin uptake sites in theepithelia and the serotonin binding protein in the underlying mesenchyme raisesthe possibility that serotonin might be involved in epithelial-mesenchymalinteraction (Lauder et al, 1988). Xu et a! (1990) have analyzed the distribution of5a group of macromolecules associated with the basement membrane in thedeveloping primary palate in chick embryos. The results indicate that the regionaldifferences within the maxillary prominence and between the maxillaryprominence and adjacent regions, such as the roof of the stomodeum, are relatedto developmentally regulated changes. These changes are associated with thepresence and distribution of type IV collagen. Type IV collagen expression wasdecreased in actively growing regions, i.e. regions of maxillary outgrowth on thelateral surface.Ill. Epithetial fusionAt about 37 days after conception in humans (Stage 16), after the facialprominences become induced to form the nasal placode, the lateral wall of thenasal placode is formed caudo-occipitally by the maxillary prominence andcranio-frontally by the lateral nasal prominence, Its medial boundary consists ofthe medial nasal prominence, which contacts the maxillary prominence in thefrontal portion of the nasal groove and the lateral nasal prominence in the caudalportion. The area of contact is called the epithelial plate or nasal fin (Streeter,1948; Vermeij-Keers, 1972) and it forms the continuity between the nasal cavityand the roof of the mouth. This continuity between the nasal sac and the roof ofthe mouth becomes interrupted and replaced by the active proliferation of themesenchyme of maxillary prominence and nasal prominences at stage 17(Streeter, 1948; Warbrick, 1960; Vermeij-Keers, 1972; Diewert and Shiota, 1990;Diewert and Van der Meer, 1991).6In the mouse embryo, the development of the primary palate starts atabout 10 days and 18 hours (Reed, 1933; Trasler, 1968). From the time of theirappearance the lateral nasal and medial nasal prominence are connected by anisthmus (Trasler, 1968). Fusion of lateral nasal and medial nasal prominencescommences from the back portion of the mouth and proceeds ventrally. Trasler(1968) showed that at the “crescent” face stage, the epithelium of the medial andlateral nasal prominences begins to make contact in a posterior to anteriordirection forming a flat plate of double epithelium called the nasal fin. A “zippingup” process follows this fusion of epithelia giving the nasal opening a commashape.The initial contact between the cells of the medial and lateral nasalswellings is made by short projections from one superficial cell to the surface ofan opposing superficial cell in the mouse embryos (Gaare and Langman,1977a).Trasler and Ohannessian (1983) have also shown that cells approaching or incontact with opposing cells form cell projections, intercellular junctions,desmosomes, and microfilaments, demonstrating firm contact between theopposing epithelia. Millicovsky and Johnston (1981) have shown that in themouse embryo, epithelial cells lose their surface microvilli before contact. After abrief period of quiescence, they begin to fill the groove separating the facialprominences by producing a series of surface projections that increase in sizeand complexity as the process of fusion progresses.7Thecarbohydratesurfacecoathasbeensuggestedtobeanessentialfactorinmediatingadhesionbetweenopposingpalatalshelves(GreeneandPratt,1976).Whentheseinvestigatorsinhibitedsurfacecoatproductionbymeansofdiazo-oxo-norleucine(DON)invitro,thepalatalshelvesfailedtofuse.Acellsurfacecoatisalsofoundovertheepithelialliningsofthenasalprominencesintheregionofpresumptivefusionusingrutheniumredandradioactiveprecursors(GaareandLangman,1977a;FigueroaandPratt,1979).Burketal(1979)foundtheconcentrationofsurfacecoatmaterialontheepitheliumofthepresumedfusionareatobehigherthanotherregionsofthenasalfoldsusing3H-ConcanavalinA.Thesefindingssupportthehypothesisthatthecellsurfacecoatisassociatedwiththeabilityofepithelialshelvesorfoldstoadhereandfuse.Infact,complexcarbohydrateshaveforsometimebeenimplicatedincellaggregationandintercellularadhesioninvariousinvitrocellsystems(PessacandDefendi,1972;Oppenheimer,1973;Roseman,1974;GreigandJones,1977).IV.MesenchymalreplacementAtabout41daysafterconceptioninhumans(stage17),amajorportionofthenasalfinisinterruptedbytheactiveproliferationofthemesenchymefromthelateralnasal,medialnasalandmaxillaryprominences.Thismesenchymeconnectsthemedialandlateralwallsofthenasalgrooveandestablishestheprimordiumofthepalate(Streeter,1948;Warbrick,1960).Therearetwoconceptsaboutmechanismsofreplacementofthenasalfinbythemesenchyme.Oneof8these is fusion between the facial prominences which is similar to that information of the secondary palate. As the two palatal processes come together,the covering epithelial layers are brought into contact. An epithelial seam forms,and shortly thereafter the epithelial seam begins to fragment as the cells eitherdegenerate (Greene and Pratt, 1976) or transform into mesenchyme (Fitchett andHay, 1989). In contrast with fusion is a series of events which was namedmerging by Patten (1961). In merging, mesenchymal growth and migrationunderlying the epithelium of the prominences eliminates the interveningepithelium. The epithelium is pushed out from between the elevations instead ofbeing apposed and then disintegrating or transforming as in fusion.Tondury (1950) noticed in the antenor part of the epithelial platedegenerative processes expressed by the occurrence of pyknotic nuclei andnuclear fragments followed by destruction of the basement membrane. Incontrast, Anderson and Matthiessen (1967) interpreted these nuclear fragmentsas peripherally sectioned mitotic figures. In addition, they did not detecthistiocytes, which they postulate to be present wherever embryonic epitheliumdisappears. Therefore, they were inclined to agree with Patten (1961) whoexplains the disappearance of the epithelial plate by a process called merging, inwhich the epithelium between two swellings is squeezed out by pressure exertedby the underlying mesenchyme upon the epithelium of the groove. Vermeij-Keers(1972) found that the basement membrane of the epithelial plate haddisintegrated locally. Between the normal epithelial cells in this plate nuclearfragments were found. The limited number of human embryos of the relevant9stages available for these studies limited definite conclusions.Recently, Diewert and Van der Meer (1991) quantified growth of theepithelial nasal fin and mesenchymal replacement of the nasal fin during normalhuman primary palate formation. Thirty serially-sectioned human embryos ofstage 16 to 19 in the Carnegie Collection (Streeter, 1948; O’Rahilly and Muller,1987) were studied. The results showed that during stage 16 the nasal fin formedbetween the medial nasal and maxillary prominences. During stage 17, amesenchymal bridge formed through the nasal fin, and the size of themesenchymal bridge increased rapidly to occupy up to 50% the total area. Duringstages 18 and 19 total primary palate area increased and the mesenchymalbridge enlarged to constitute 65 to 85% of the total area.Gaare and Langman (1977b) reported that in mouse embryos, shortlybefore the epithelial linings of the opposing nasal prominences make contact, celldegeneration characterized by condensation and fragmentation occurs in theepithelial linings of the prospective fusion areas. After fusion has established thenasal fin, epithelial cells continue to degenerate in the same manner. However,cell degeneration can not account for complete regression of the nasal fin, sincemany morphologically healthy epithelial cells are always present. Theysuggested that these surviving epithelial cells incorporate into the adjacentepithelial linings of the expanding pnmary nasal and oral cavities.The interchange of tissue phenotype, especially epithelial tomesenchymal, is a common phenomenon during early embryogenesis. It ispossible that there exists in most epithelia, a readily triggered mechanism that10turns on the mesenchymal genetic program. Greenburg and Hay (1982, 1986,1988) demonstrated that a variety of adult and embryonic epithelia that normallydo not give rise to mesenchyme do so when the isolated tissue is immersedinside hydrated Type I collagen gels. Confronted with collagen fibrils in closecontact on all sides, these well-established epithelia express the potential fortissue-type conversion in response to an abnormal extracellular matrixenvironment.Fitchett and Hay (1989) showed that palatal medial edge epithelium is anectoderm that retains the ability to transform into mesenchymal cells. They reportthat cell death is not the major mechanism leading to removal of the midlineepithelial seam created by contact of the two palatal shelves. Rather, opposingbasal cells adhere, after sloughing of the periderm, proliferate, and then transforminto mesenchyme. The basal lamina disappears as basal cells extend filopodiaand then pseudopodia into the adjacent connective tissue compartment. Theglycogen rich basal cells have euchromatic, vesicular nuclei and abundant roughendoplasmic reticulum. Before they begin to elongate and move into theextracellular matrix, they acquire a vimentin-rich cytoskeleton and lose keratinexpression. Vimentin filaments are the characteristic intermediate filament type ofmesenchymal cells. The changes in cell shape and cytoskeleton are similar tothose reported in already established epithelial mesenchymal transformations(Hay, 1968: Bernanke and Markwald, 1979, 1982; Nichols, 1981; Franke etal,1982).11V. Primary choana formationAt about 44 days after conception in humans (stage 18), the width ofnasal septum and medial nasal prominence decreased to 0.5 to 0.7 times that ofstage 16 (Diewert et al, 1989; Diewert and Lozanoff, 1989, 1990). At the samestage, the nasal fin dorsal to the zone of mesenchymal penetration persists andcavitates, forming the oronasal membrane which separates the nasal pit from thecavity of stomodeum. The rupture of the oronasal membrane is brought about bythe disintegration of the cells that form it. This results in the opening of arespiratory passage from the nostril through the primary choana to the pharynx(Streeter, 1948; Warbrick, 1960).In the mouse embryo, after the disintegration of nasal fin and penetrationof mesenchyme, a portion of the nasal fin remains at the back of the nasal pitforming the oronasal membrane, with the formation of interstitial gaps occurring atthe 11th day after conception (Trasler, 1968; Tamarin, 1982). The gaps enlargeand coalesce so that a completely patent opening between nasal passage andstomodeum is established by 13 days. The membrane consists of two layers ofsimple squamous epithelium which become separated as involution progresses.The form of the choanal antrum changes from a simple funnel-shaped ellipseearly in the 13th day to a complex slit like opening within the following 24 hours.This coincides with the completion of a definitive primary palate and theenlargement and elevation of secondary palatal shelves.12B. Abnormal development of the primary palate and cleft lip malformationI. Morphogenesis of cleft lip in the human embryoCleft lip is a result of failure of fusion between the medial nasalprominence, lateral nasal prominence and/or maxillary prominence. Onepossible cause of this defect is that the ventral ends of the prominences fail tocome into actual contact with each other for fusion. This can occur if growth isdefective in either or both prominences. TOndury (1964) described an embryowith unilateral complete cleft lip where the primary cause for the faultydevelopment appears to be defective growth of the lateral nasal prominences.Another explanation is that the nasal fin persists throughout the developmentalstages preventing the mesenchyme of the maxillary and fronto-nasal prominencefrom making contact. Subsequently, when the dorsal part of the nasal finundergoes the normal cavitation and cleavage, resulting in the formation of theprimary choana, the ventral part of the nasal fin also undergoes cleavageresulting in the formation of cleft lip (Stark, 1954; Warbrick, 1960).Anderson and Matthiessen (1967) have suggested that complete cleft lipwill also appear if mesenchymal proliferation is retarded in the medial nasal andmaxillary prominences, and incomplete cleft lip will appear in cases with a lessmarked retardation of the mesenchymal proliferation in the above mentionedmesenchymal centers. Furthermore by investigating human embryos with primarypalatal clefting in the Kyoto collection, Diewert and Shiota (1990) showeddeficient mesenchymal bridge growth and a visible deficiency of tissue in the cleftareas. The results also showed regional growth deficiency or developmental13abnormality in the palatal tissues during the critical time of rapid mesenchymalbridge enlargement in cases of partial or incomplete clefting.II. Morphometric study of postnatal human cleft lipFraser and Pashayan (1970) have shown that parents of children withcleft lip tend to differ from the general population in certain dimensions of facialtopography. There was a significant tendency for the anterior surface of themaxilla to be flatter in the experimental group than in the controls. In addition, inthe experimental group the mean dizygomatic and intraocular chinmeasurements were larger, the frequency of rectangular and trapezoid shapeswas higher, and the upper lip was less protuberant relative to the lower lip.Coccaro et al (1972) have also shown that parents who lack facial deformities,but have cleft lip and palate children, have faces that are less convex with atendency toward mandibular prognathism. Vertical and horizontal measurementsof the upper face and the nose length were found to be shorter for parents of cleftlip and palate children.Erickson (1974) analyzed three proposed microforms in the normal sibsof children with cleft lip but with or without cleft palate: (1) facial profile, (2) dentalarch shape, and (3) palatal form. It is concluded that the sibs of these faciallymalformed children are likely to be different from normal children. However, thesedifferences are not sufficient to classify these people as a group having a distinctmalformation with any degree of certainty.14Studies of facial morphology in monozygotic twins discordant for CL(P)suggest that approximately two thirds of cleft lip cases are caused byunderdevelopment of the medial nasal prominences (Johnston and Hunter,1989). The twin studies indicated that the remaining one-third of cleft lip casesresult from underdevelopment of the maxillary prominences. Thus, Johnston andHunter (1989) proposed the existence of two major CL(P) groups; one with smallmedial nasal prominences in the other with small maxillary prominences.Ill. Morphogenesis of cleft lip in laboratory animalsReed (1933) has proposed that harelip in animals is due primarily to thelateral nasal prominence and the medial nasal prominence failing to fuse. Thisfailure is probably due to a retarded growth rate of the maxillary prominence.Trasler (1968) has shown that formation of the normal lip requires the posteriorportions of the medial and lateral nasal prominences to remain continuous witheach other and with the medial portion of the maxillary prominence. In an NJembryo genetically predisposed to clefting, the medial nasal prominences do notdiverge laterally as much as they do in an embryo that is not predisposed. Thisresults in a decrease or failure of epithelial fusion between the medial and lateralnasal prominences, and consequently, a lack of consolidation of the isthmus thenoccurs.The hypothesis that face shape is a causal factor in geneticpredisposition to cleft lip in mice was further tested by Juriloff and Trasler (1976).Their results from measuring photographs of embryos support this hypothesis.15However, it is suggested that the susceptibility to another type of cleft lip in othergenotypes could arise for example, through hypoplasia of one of the facialprocesses. Trasler and Machado (1979) have found that a particular facialcomplex is associated with cleft lip predisposition. Premaxillary length issignificantly shorter in newborns and adults in mouse cleft lip lines CL/Fr, NJ andL than non-cleft lines M and C57BLI6J. Premaxilla width also tended to benarrower in the adults, and gum length, in newborns and adults, tended to beshorter in cleft lip lines.Developmental alterations associated with spontaneous cleft lip andpalate in CL/Fr mice (Millicovsky et al, 1982; Forbes et al, 1989) include: alteredfacial geometry, in which the orientation of the medial nasal prominences isalmost parallel to the mid-sagittal plane, depressed ability of the surfaceepithelium of primary palate primordia to participate in the fusion process, andhypoplasia of the lateral nasal prominences. Ohbayashi and Eto (1986) havesuggested that the medial nasal prominences (MNP) play a critical role in normalfacial development and cleft lip formation based on culture experiments done invitro using rat embryos that have had a part of each facial prominence removed.They found that cleft lip like malformation was observed only in the MNP-excisedgroup.C. The genetic causes of cleft lipI. One major gene in mouse cleft lipThe genetic causes of cleft lip have been controversial. In the 1930’s,16Reed (1936) outcrossed CL(P)-liable stock to test if one single major essentialgene for harelip was present. The results were consistent with the single majorlocus hypothesis. Reed noted, however, that it was also probable that cleft lipresulted from the cumulative effect of a small number of recessive genescontaining the same genetic information. By backcrossing A/J and C57BLI6Jmice, Juriloff (1980) showed that one or two loci were involved in the expressionof CL(P). In a further study, Juriloff (1986) has produced a congenic strain, inwhich the cleft lip gene is transfered into an unrelated AEJ/GnRK strain. Thestable frequency of cleft lip from detected carriers in each backcross generation ismost compatible with the one locus mutation model. Biddle and Fraser (1986)have also proposed that the difference between the A/J mouse embryo and theC57BL/6J strain appears to be determined by a single recessive gene.II. Maternal effectsDavidson et al (1969) demonstrated a maternal effect on the frequency ofspontaneous cleft lip in the mouse. In repeated backcross studies of A/J andC57BL/6J the frequency of CL(P) was higher for genetically equivalent embryosfrom NJ mothers compared with those from hybrid mothers. The authors havesuggested that clefting could be mediated through a genetically determinedmaternal uterine biochemistry or physiology. If this is so, then the geneticdifference between the strains must be multifactorial. Alternatively, it is possiblethat the clefting strain mother did not provide a cytoplasmic factor, found inC57BL/6J, which provides resistance to clefting. Bornstein et a! (1970) tested17these hypotheses and have shown that the maternal effect on cleft lipsusceptibility was present in the CL/Fr strain, but it was not transfered through anycytoplasmic factor.Juriloff and Fraser (1980) have found that the difference in CL(P)frequency between the NJ and CL/Fr strains was not determined by the embryogenotype but by the maternal genotype. Also the data shows a reciprocalrelationship between cleft lip frequency and resorption frequency. This suggeststhat the maternal trait may cause a difference in the survival of cleft lip fetuses.Juriloff (1982) has repeated her study with A/WySn and A/l-leJ and found thesame result.Maternal effects have been sought for human CL(P) data. Maternalgenetic effects did not account for the racial differences in frequency of CL(P) inHawaii (Ching and Chung, 1974). No evidence of maternal effects was foundwhen comparing CL(P) recurrence risk between maternal half and paternal halfsiblings (Bingle and Niswander, 1 977). Juriloff (1 980) has suggested that if themouse populations were as heterogeneous as the human populations, thematernal effects probably would not be revealed. The maternal effects weredetected when the genetic risk of the embryos was sizable and constant. A morerecent paper compared clefting frequencies in Blacks and Whites and reciprocalcrosses (Khoury et al, 1983). The results showed that the difference in thereported rates of CL(P) between reciprocal crosses of Whites and Blacks is due tothe effect of mother’s race. When the mother is Black, offspring of White fathersdid not have a higher rate CL(P) than those of Black fathers. This study18documents the existence of maternal determinants of CL(P) in humans.Ill. Reciprocal relationship between cleft lip and resorption rateIn the study of genetic maternal effects on cleft lip frequency in NJ andCL/Fr mice, Juriloff and Fraser (1980) have shown a reciprocal relationshipbetween cleft lip and resorption frequency of the two strains. In another study(Juriloff, 1982) this reciprocal relationship was observed to be directly related tothe segregation of genetic variation. The existence of similar genetic variation inhumans would partially explain inverse association between clefting andspontaneous abortion (Stein et a!, 1975; Bear, 1978). The concept of“terathanasia” the natural abortion of defective embryos (Warkany, 1978) is alsosupported.It has been thought that thyroxin may reduce the frequency of cleft lip(Woollam and MilIen, 1960). However, by the administration of thyroxin, it hasbeen shown that the reduced frequency of cleft-lip embryos collected at term afterthyroxin treatment is due to their increased mortality rate, and not to prevention ofthe lip defect (Brown et a!, 1974; Juriloff, 1981). A further study by Juriloff andHarris (1985) on the thyroxin-induced differential mortality of cleft lip in mouseembryos has also shown that following the thyroxin treatment cleft lip and normalembryos died, but cleft lip embryos died at higher rate. Therefore thyroxin doesnot affect the events that lead to cleft lip but the presence of cleft lip increases theliability for thyroxi n-induced death.19IV. Utenne site effectsTrasler (1960) has found that within A/J strain, embryos in the uterine sitenearest the ovary develop cleft lip significantly more often than embryos in otherpositions in the uterine horn. Kalter (1975) showed that the frequency of CL(P)was higher at the ovarian and cervical sites than elsewhere. The resorption waslower at the ovarian site than elsewhere, and this mortality trend ran along withfetal weight, ie, as the former increased the latter decreased. Juriloff (1980) hasreinvestigated this uterine site effect and shown increased cleft lip and decreasedresorption at the ovarian site. It is suggested that a relatively privileged area atovarian site in both A/J and CLJFr allows the survival of cleft lip embryos thatwould have died elsewhere (Juriloff, 1980).V. General concept of the multifactorial threshold modelThere are many characteristics of biological interest which appear to varyin a discontinuous manner but are not found to be inherited in a simpleMendetian fashion. These can be classified into two phenotypic classes, affectedor not-affected. Characteristics of this sort appear at first sight to be outside therealm of quantitative genetics; yet when they are subjected to genetic analysisthey are found to be inherited in the same way as the continuously varyingcharacteristics.“The clue to understanding the inheritance of such characteristics lies inthe idea that the characteristics have an underlying continuity with a thresholdwhich imposes a discontinuity on the visible expression, as depicted in Figure 1.200 +1 ÷2Scale of liability (standard deviations from threshold)Fig. 1. Falconer’s concept of threshold. Two populations or groups withdifferent mean liabilities. The liability is normally distributed, with the samevariance in the two groups. The groups are compared by references to a fixedthreshold. The stippled portions are the affected individuals with the incidencesshown (from Falconer, 1965).Incidence = 5%Incidence = 20%MeanThreshold—4—3 —2 —121When the underlying variable is below this threshold level the individual has oneform of phenotypic expression, which is conceived ‘normal’; when it is above thethreshold the individual has the other phenotypic expression, i.e. ‘affected’.”(Falconer, 1965). For such characteristics with a threshold, the underlyingcontinuous variable has been called the liability in the context of humandiseases. The continuous variation of liability is both genetic and environmentalin origin, It might be thought of as the developmental rate of a specific process,and thus, in principle, it could be measured and studied as a metric character inthe ordinary way.The classical threshold analysis is described in a paper written byWright (1934) on polydactyly in guinea pigs. Three closely inbred guinea pigs(strains 2, 1 3, 32) with normal 3-toed hind feet were crossbred with strain D withpolydactyly. The crosses between 2 and D simulate one factor Mendelianheredity to a remarkable extent in the dominance of 3-toe in Fl, and apparentsegregation in F2 in a fairly close approach to a 3:1 ratio and in the backcross tostrain D in a 1:1 ratio. This interpretation breaks down in the tests of the supposedsegregants. These tests indicate that there are at least 3 factors of comparableimportance and more probably 4 by which strains 2 and D differ. There is a closeapproach to blending inheritance in a character which approaches alternativeexpression because of physiological thresholds. The crosses between 32 and Dgave closely similar results to those of 2 and D. On the other hand, the crossesbetween 13 and D gave a very different result which indicates that strain 13 ismuch closer to the threshold for polydactyly.22A study was made by Gruneberg (1951) on the CBA mouse strain. Theseanimals have abnormally small third molars, and some lack these teethcompletely. It appears that the absence of the third molar occurs when the dentallamina, which forms the tooth bud, falls below a certain size. The differencebetween the mean third molar size of the CBA strain and that of C57BL is broughtabout by multiple genes. This result shows that if the Fl males were backcrossedto CBA females, there was a shift in the variable distribution of the offspringtoward that of the CBA. In a later paper, Grüneberg (1952) has described quasi-continuous characters in the sense that the underlying genetic basis is acontinuous variable, with multiple factor inheritance, which is divided by aphysiological threshold into normal and abnormal animals. The peculiargenetical properties of quasi-continuous characters are regarded to be due partlyto the fact that a continuous distribution may shift in relation to a physiologicalthreshold, and partly that they share with ordinary continuous variables, themultiple gene basis and the sensitivity to influence of the environment. Green’sstudy (1971) of presacral vertebrae and Tom et ats study (1991) of exencephalyare further examples of this type of approach.VI. Multifactorial threshold model and one major gene in human cleft lipA multifactonal threshold model which accounts for the liability of thecommon, familial, human disorders was developed by statistical geneticists(Falconer, 1965). Using Falconer’s procedure on cleft lip and palate, an additivepolygenic model has been fitted to the cleft lip data from human populations23(Carter, 1969, 1976; Bixler et a!, 1971, Woolf, 1971, Czeizel and Tusnady, 1972,Chung etal, 1974; Bear, 1976).From the analysis of Danish and Japanese families of probands withCL(P) Chung et a! (1986) found that the Danish data is best explained by themixed model, which combines the major gene and multifactorial inheritancemodels. On the other hand, the Japanese data is best accounted for only by themultifactorial inheritance model. These findings appear to explain the puzzlingobservation that the Japanese population which has a higher general incidenceof CL(P) has a lower recurrence risk of having CL(P) relative to the Caucasianpopulation (Chung et a!, 1986). In other words, factors outside the major cleftinducing gene have a greater effect on the Japanese population. The factorswhich affect one generation may not be present in the next generation, thus therisk for clefting in the second generation does not need to be the same as theirparents. To identify the major gene, Ardinger et a! (1989) have shown that eitherthe transforming growth factor-a (TGF-a) gene itself, or other DNA sequences inan adjacent region, contribute to the development of a proportion of the cases ofcleft lip in humans. Further evidence for an association between genetic variationin transforming growth factor a and cleft lip and palate has been shown byChenevix-Trench et a! (1991). However, in another study of seven families withCL(P) segregating in a dominant manner, the association of the A2B2C2haplotype reported by Ardinger et a! (1989) was not found, with none of theaffected parents having this TGF-a haplotype (Hecht eta!, 1990).24Studies of facial morphology in monozygotic twins discordant for CL(P)suggest that approximately two thirds of the cleft lip cases are caused byunderdevelopment of the medial nasal prominences and the remaining one-thirdof cleft lip cases result from underdevelopment of the maxillary prominences(Johnston and Hunter, 1989). These groups have also been consideredcomparable to Chung et al’s (1986) multifactorial and single major gene groups,respectively. Marazita et al (1986) have studied cleft lip with or without cleft palatein the families of non-syndromic CL(P) probands who were surgically corrected.The data, which come from three populations, provide no support for themultifactorial threshold model but did provide evidence of the presence of a majorgene responsible for cleft lip in at least a portion of cases.Melnick et al (1980) have tested the multifactorial threshold inheritancemodel by studying 1,895 persons born in Denmark between 1941 and 1968. Theindividuals were born with cleft lip with or without cleft palate. The resultsrevealed that neither the multifactorial threshold model nor the single-major locusmodel provided an adequate fit. As an alternative model, the monogenicdependent susceptibility (MDS) to a variety of teratogens was proposed in light ofexperimental mouse and human data. In humans, a study by Bonner et a! (1 978)suggests an association between particular HLA haplotypes and clefting.However, Van Dyke et al (1980) have shown that it is very unlikely thatspontaneous cleft lip with or without cleft palate is closely linked to HLA. Inmouse, it has been shown that the maternal H-2 haplotype significantly affects theincidence of corticosteroid-induced isolated cleft palate (Bonner and Slavkin,251975) and cleft lip with or without cleft palate (Silberman et al, unpublished, citedfrom Melnick et al, 1980). Juriloff (1982) showed that the H-2 gene region did notappear to influence CL(P) frequency on the A/WySn strain background.D. Environmental effects on cleft lip malformationI. HadacidinHadacidin is an antibiotic isolated from broth culture of Penicilliumfrequentants (Chaube and Murphy 1963). It is a potent inhibitor of the enzyme,Adenylsuccinic synthetase in normal rat tissue in vitro, by competing with Laspartate for the active site on the enzyme molecule (Shigeura and Gordon,1962a, 1962b).Lejour-Jeanty (1966) has shown that Hadacidin induces harelips in therat which are comparable with human abnormalities. The effect of the drug isconfined to the lateral nasal prominence, which seems to be responsible for theabsence or incompleteness of fusion of both edges of the olfactory pit. Themaxillary prominence does not take part in the fusion, but its slow anddisorientated growth is thought to contribute to the maintenance of the nasal finwhich persists or may be partially destroyed by the mesenchyme. Another study(Lejour, 1969) of cleft lip induced by Hadacidin in rats indicated that completeclefting is more often the result of a disruption of the contacting edges of the nasalgroove rather than of an absence of fusion, as usual. Clefts with bridges areproduced when the nasal fin is only partly penetrated by mesenchyme in theregion of the maxillary arch.26II. 6-aminonicotinamideSpecific vitamin antimetabolites utilized in experimental mammalianteratology have been incoporated into deficient diets (Nelson, 1957) or injectedinto pregnant animals. (Wilson, 1959, 1964). The niacin antimetabolite, 6-aminonicotinamide (6-AN) has been shown to be teratogenic in various species(Murphy etal, 1957; Pinsky and Fraser, 1959). 6-AN has been shown to be anantimetabolite for nicotinamide (Pinsky and Fraser, 1960). It has beendemonstrated (Dietrich eta!, 1958) that 6-AN inhibits the diphosphopyridinenucleotide (D.P.N.)- dependent reaction by substituting the nicotinamide in theD.P.N. molecule and rendering the inactive analogue incapable of functioning inthe hydrogen and electron transfer reactions essential to the normal metabolismof the cell.Pinsky and Fraser (1960) found that cleft lip and cleft palate woUld resultfrom an injection of 6-AN on day 9 1/2 followed by nicotinamide two hours later.However, no malformations were observed if treatment was given on day 10 1/2.They concluded that the study of the protective effect of vitamins against theteratogenic activity of their antagonists appears to be useful for analyzing thevitamin requirements of various organogenetic processes in the developingembryo. This shows that nicotinamide requirements of the embryo-mother systemappears to vary from one-gestational period to another.Trasler and Leong (1982) have found that treatment with 6-AN on day 9produced 18% median cleft lip and no lateral cleft lip, whereas treatment on day10 produced 22% lateral cleft lip and no median cleft lip in near-term C57BL1627fetuses. A median cleft lip is associated with a critical reduction in growth of themedial nasal areas. Histologically, the nasal ectoderm of the treated group hadfewer cells per unit area in the medial nasal area than of the controls, adjacentmesenchyme had an increase in the number of dense bodies, a possible result ofcell death occurring, and a significant reduction of mitotic index in the nasal area.On the other hand, mechanisms for 6-AN-induced lateral cleft lip may involvefailure of less organized denser nasal ectoderm to fuse, and a growth reductionof both lateral and medial nasal prominences. Histologically, the mesenchyme inthis area also contained a large number of dense bodies, and the mitotic indexwas significantly reduced in both nasal and neural areas suggesting that mitoticinhibition may have caused the observed abnormalities.In a further investigation, Trasler and Ohannessian (1983) madecomparisons of the ultrastructure of cleft lip liable and control embryos treatedwith 6-AN. A few 6-AN-treated embryos showed abnormal contact that appearedmalpositioned and tenuous. The teratogen also produced increased cell deathand a denser epithelium and mesenchyme. The denser appearance of themesenchyme may be associated with a decrease in the intercellular matrix aswas found in another study by Flint and Ede (1978).Ill. PhenytoinThe anticonvulsant phenytoin (PHT) is teratogenic to inbred strains ofmice (Massey, 1966). Treatment of pregnant mice with PHT increases thefrequency of cleft lip and/or palate in surviving fetuses (Johnston et a!, 1978). The28first association of phenytoin with birth defects in humans was reported in 1968(Meadow). A group of abnormalities in growth and performance have beencharacterized as the Fetal Hydantoin Syndrome (Hanson and Smith, 1975) andthey include: craniofacial abnormalities, cardiac and limb defects, generaldeficiencies in growth, and mental retardation.Martz et al (1977) studied the possibility that the teratogenesis of PHT isdue to the arene oxide produced by the molecule. Arene oxide can form covalentbonds to gestational embryonic tissue and cause abnormal development.Scanning and transmission EM analyses of high incidence of cleft lip and palateproduced by maternal intraperitoneal administration of phenytoin on gestationalday lOin A/J mice have been reported by Sulik etal(1979). In the phenytointreated embryos, the mesenchymal cellular processes, which form a densemeshwork that interact with the epithelial basement membrane, areunderdeveloped or are absent. The hypothesized effect is secondary tointerference of the drug with oxidative metabolism and ATP production by thecells. Mackler et a! (1975) proposed that PHT, or the arene oxides from PHT,inactivate oxidative enzymes such as DPNH oxidase, and thus interfere withoxidative metabolism.Genetically determined differences in metabolism of phenytoin mayexplain why Fetal Hydantoin Syndrome only occurs in a small portion of humanfoetuses who are exposed to the drug (Strickler eta!, 1985). Inbred and congenicstrains of mice have been studied for susceptibility to phenytoin induced cleft lip(Goldman et a!, 1983). The role of genes linked to the H-2 complex on29chromosome 17 has been confirmed.In a study, where Hicks et al (1983) collected mouse embryos fromphenytoin treated mothers, it was found that the DNA and protein synthesis werealtered by this drug. DNA synthesis in these tissue was only 26% of the controlgroup value, but the protein synthesis was 2.6 times that of the control primarypalate. These effects of phenytoin are not limited to primary palates for similarchanges in DNA and protein synthesis also occur in embryonic limb buds.IV. OxygenHypoxia-induced cleft lip in CL/Fr mice has been reported by Millicovskyand Johnston (1981). The spontaneous clefting rate of 36% in CL/Fr mice wasshown to increase to approximately 90% when pregnant mothers were exposedto hypoxia, (10% 02), and to decrease to 13% when they were exposed tohyperoxia, (50% 02), during the critical time of primary palate development. In amorphological study (Bronsky et al, 1986), cellular debris was present in hypoxicembryos at stages prior to primary palate fusion and absent in comparable stagenormoxia embryos. It was suggested that this cellular debris was associated withthe retardation of placodal invagination and was primarily responsible for theincreased incidence of CL(P). In addition, it was hypothesized that placodeinvagination involves energy-dependent actin-myosin interactions, and is thussensitive to the ATP-reducing effects of hypoxia.302. RationaleIn the past, many studies have looked at human primary palate and cleftlip formation (Stark, 1954; Warbrick, 1960; Tondury, 1964; Anderson andMatthiessen, 1967; Vermeij-Keers, 1972; Hinrichsen, 1985; Diewert and Shiota,1990). These studies utilized several human embryos at relevant stages, but toofew embryos are available to reach any definite conclusion. Early facialdevelopment and morphology of human and mouse embryos are very similar andof comparable size at the time of lip formation (Trasler, 1968). The mouse modelhas been studied for cleft lip morphologically (Reed, 1933; Trasler, 1968; Juriloffand Trasler,1 976; Trasler and Machado, 1979; Millicovsky et a!, 1982; Cirianiand Diewert, 1985), teratogenetically (Sulik eta!, 1979; Juriloff, 1981; Eto etal,1981; Trasler and Leong, 1982; Trasler and Ohannessian, 1983; Juriloff andHarris, 1985; Bronsky eta!, 1986) and genetically (Reed, 1936; Davidson eta!,1969; Bornstein eta!, 1970; Juriloff, 1980; Juriloff and Fraser, 1980; Biddle andFraser, 1986; Juriloff, 1986). However, there have been no quantitative studieslooking at internal structures of normal and abnormal lip formation in mice withdifferent genetic backgrounds.The purpose of this study was to give insight into the etiology of cleft lipby measuring and analyzing internal structures of the forming primary palate intwo noncleft, and three cleft lip strains of mice. The rationale behind the use oftwo noncleft lip strains was to compare the effect of different genetic backgroundwithout the genetic cleft lip liability on primary palatal structures. The three cleft lipstrains are investigated to determine the effects of different genetic backgrounds,31with different cleft lip and resorption rates, on the internal structures of primarypalate.The first specific aim was to collect embryos of three cleft lip strains at day13 hour 12, and determine the cleft lip frequency and resorption rates of eachstrain. The rates were then compared among three cleft Lip strains using statisticalanalysis. The uterine site effect on the cleft lip and resorption rate was alsocompared statistically.The second specific aim was to determine the stage of bodydevelopment and the chronological age at which primary palatogenesis takesplace in these five strains. The chronological ages of these strains was used onlyto collect embryos at similar tail somite stages. The analysis of primary palateinternal structure in the noncleft lip and cleft lip strains is based on developmentaltail somite number which is more representative of developmental stage than ischronological age.The third specific aim was to delineate and quantitatively compareprogressive phases of primary palate development by studying the internalanatomical structures which were suggested as essential for successful upper lipformation. These phases are epithelial fusion forming nasal fin, interruption of thenasal fin with mesenchyme, mesenchymal enlargement, and cleavage of nasalfin forming the oronasal membrane and primary choana.At an early stage of primary palate formation involving fusion of facialprominence epithelia and forward growth of the maxillary prominence, onenoncleft lip and one cleft lip strain of mice were studied. This was done to32understand the possible factors which may contribute to cleft lip at the stagebefore the mesenchyme connects the medIal and lateral walls of the primarypalate. The next stage of primary palate formation with the mesenchymereplacing the epithelial nasal fin was investigated using two noncleft lip, andthree cleft lip strains. An analysis of covariance comparing area of primary palateand mesenchymal component was used. The partial least squares analysis wasused to determine the best predictor of primary palate formation. The primarychoana dorsal to the primary palate forms at a later stage during primary palateformation when the oronasal membrane retracts, and a respiratory passageopens from the nostril through the primary choana to the pharynx. The timing ofthis event was investigated and compared in one noncleft lip strain and one cleftlip strain.Finally, a multifactorial threshold model was suggested. Within themodel, the genetic cleft lip liability is proposed to be influenced by the biologicaltraits of primary palate formation. At the threshold, a critical amount ofmesenchyme must be present for normal primary palate formation to occur.Unfavorable growth of any biological traits brings the period of mesenchymalformation closer to the threshold resulting in an occurrence of cleft lip. Thisthreshold may be affected by the timing of primary choana formation.33MATERIALS AND METHODSA. Cleft lip frequency and resorption rate.I. Embryo collectionThe cleft lip strains studied were NJ obtained from Jackson Laboratory,Bar Harbor, Maine, A/WySn obtained from Jackson Laboratory, Bar Harbor,Maine, and CL/Fr (developed in the laboratory of Dr. Fraser in McGill University).The strain of CL/Fr was imported to our laboratory in 1985. Mice were maintainedon a diet of Purina mouse chow and filtered water and were housed with a 12 hrlight cycle from 7 AM to 7 PM. Three or four adult females were caged overnightwith a male and were examined in the morning for the presence of a vaginal plug.It was assumed that ovulation took place at midnight, therefore 9 AM of the daythe plug was found was designated as day 0 hour 9 of gestation (Snell et al,1940). Ten litters of NJ, eight litters of A/WySn and ten litters of CL/Fr wericollected at day 13 hour 12. The embryos were dissected and numbered by thethree digit system: the first digit indicated the right or the left site in the uterinehorn; the right side was indicated as one and the left side was indicated as two;the second digit indicated the sequential order of the embryos from sites nearestthe ovary to the cervix, and the third digit indicated the sequential order of theembryos from the sites nearest the cervix to the ovary; both the second and thirddigits were counted by numbers starting with one. No distinction was made as tothe extent and form of the malformation. Thus, whether the clefting was completeor incomplete, unilateral or bilateral, the malformations were called cleft lip.Resorptions were defined as all dead embryos up to day 13 hour 12 of34development. Litters of less than 5 live embryos were discarded.II. Data analysisThe frequency of resorptions among implantations and frequency of cleftlip among embryos in each litter were transformed to their Freeman-Tukey arcsine values (Mosteller and Youtz, 1961). One-way analysis of variance wasapplied to the transformed data. Effects of site of uterine implantation (ovarian ornonovarian) were tested in 2 X 2 Chi square tests for each resorption amongimplants and cleft lip among embryos.B. Stages of primary palate development.I. Embryo collectionFive strains of mice were used in this study: C57BLI6J (Charles RiverCanada Inc., Montreal), BALB/cByJ (Jackson Laboratory, Bar Harbor, Maine),and cleft lip strains, NJ, A/WySn and CL/Fr. Pregnant females were sacrificed atvarious times from day 10 hour 14 onwards. The uteri were removed from theanimals and embryos were dissected and numbered using the three digitnumbering system described above. The embryos were then fixed in Bouin’ssolution for at least 24 hours. Embryos were weighed and tail somite stagenumber, calculated by counting the number of pairs of somites from the hind limbto the last somite pair at the end of the tail, was recorded. Embryonic heads werethen removed from the body with a No. 15 surgical blade and were photographedwhile immersed in 70 percent ethanol under standard conditions of lighting andmagnification in frontal and ventral profile. The embryonic heads were processed35in an autotechnicon, embedded in paraffin, serially sectioned at 7 jim thickness inthe frontal plane. Slides were then put in a 50 °C oven, then stained withhematoxylin and eosin or Periodic acid Schiff.II. Measurementsa. NasalfinThe nasal fin is the fused epithelia between medial nasal prominence,lateral nasal prominence and/or maxillary prominence and is connected to thenasal groove at the upper portion and oral ectoderm at the bottom portion.Calculation of the depth of the nasal fin in the serial anteroposterior sections wasstarted from the first section containing fusion of the three facial prominences andended at a point where the nasal fin disintegrated and preoptic mesenchymeappeared. A microruler of 100 grids (each grid equals 9.8 jim) in the Nikonmicroscope was used to measure the height of the fused epithelia (nasal fin)which spanned from the top of the nasal epithelium to the bottom of the oralectoderm. The total area of the nasal fin was the summation of the areas from theserial sections calculated by multiplying the measured height by the tissuesection thickness.b. Mesenchymal componentThe mesenchymal component is the mesenchymal tissue replacing thenasal fin epithelium. The mesenchyme grows in about 12 hours after epithelialnasal fin formation begins. The same micrometer and magnification as abovewas Used to measure the height of the mesenchymal component, which spannedfrom the bottom of the nasal epithelium to the top of the oral ectoderm. At the36stage where the mesenchymal tissue replaces the nasal fin, the region was nolonger called the nasal fin but the primary palate. The primary palate area is thesummation of the areas from each section calculated by multiplying the height ofthe nasal epithelium at the top, the mesenchymal component in the middle, andthe oral ectoderm at the bottom by the tissue section thickness. A similarcalculation is done for the area of the mesenchymal component using the heightof the mesenchyme only.c. Position of the maxillary prominenceThe maxillary prominence is an anatomical structure lateral to the lateralnasal prominence and superficially distinguished from the nasal prominence bythe naso-maxillary groove. The maxillary prominence and lateral nasalprominence grow in a frontomedial direction and become fused with the medialnasal prominence. The position of the maxillary prominence was determined bycounting the sections from the frontomedial end of the maxillary prominence asdefined by the naso-maxillary groove to the dorsal end of the nasal fin. Thenumber of sections was multiplied by 7 p.m to determine the measurement of theposition of the maxillary prominence.Ill. Data analysisa. Analysis of covarianceThe analysis of covanance makes use of the concepts of both analysis ofvariance and regression. In a one-way classification, the typical analysis ofvariance model for the value Yij of the jth observation in the ith class is= a +37where the x represent the population mean of the classes and the e1 are theresiduals. But suppose that on each unit we have also measured anothervariable Xij that is linearly related to Yij. It is natural to set up the model,Y1=ci+I3(X-X..)+e1where 3 is the regression coefficient of Y on X. This model is typical for theanalysis of covariance. If X and Y are closely related, we may expect this model tofit the Y1 values better than the original analysis of vanance model. That is, theresiduals ej should be in general, smaller than the e1. In this study, a simpleexample of the use of covariance in randomized experiments is demonstrated.With a completely randomized design, the data form a one-way classification withthe strains being the classes.i. Testing adjusted treatment meansTable 1 gives the analysis of covariance for a randomized complete-block design and, at the same time, illustrates the general procedure. Thegeneral procedure requires all three sums of products for treatments and for errorafter adjustment for all other sources of variation included in the model. For acompletely randomized design, there would be no block sums of products. From1— X..x-=nx..=zxi38Sums ofSource df products of df Adjusted y2 MSxx )çy y,yTotal ri - 1 xy Ey2Blocks r-1 Rxx Rxy RyyTreatments t - 1 T, Txy Tyy (Exy )2S’.x(r—1)(t—1)—1 Eyy —Error (r-1)(t-1) E Exy Eyy ExxTreatments (Sxy )2r(t-1) Sxx Sxy Syy r(t-1)-1 s, —+ errorSjt-i (syy -(Y)2)Treatmentsadjustedsxx- (Exy)2ExxTable 1. Analysis of covariance for a randomized complete-blockeddesign (From Steel and Torrie, 1980).39the treatments and error line, a line for treatments plus error is obtained byaddition. The sums of squares E and S are adjusted by subtracting thecontributions due to linear regression. The difference between these adjustedsums of squares is the sum of squares for testing adjusted treatment means. Totest the mean square for adjusted treatments, the appropriate error mean squareis S.x. Notice that the lines for treatments and error are essential from which thetest of adjusted treatment means is constructed.The calculation of sums of products for the randomized-block design isas follows (Steel and Torrie, 1980)2?. X..y2y_ Y.Lv XY13 — X..Y..Sum of products for blocks are:40R X.J H..XX— rtR Y”UU t — rtR-_____t rt 2Sums of products for treatments are:T X..XX— r— rty.r rtT-X1.Y X.Y..r— rt3Sums of products for error are found by subtraction and are:2 X.1 : the dot indicates that all observations for the jth block havebeen added to give this total.3 X. : the dot indicates that all observations for the ith treatment have beenadded to give total.41E <x = — —..‘cij2_DLjy L.f r\—‘U=y—— TxuTo test the hypothesis of no differences among treatment means for Y adjusted forthe regression of Y on X,MS (adjusted treatments means)Fv.x— (Exg)22sv.x = (r — 1 )(t— 1) — 1MS=“The one-tailed Ftest with 1 and n degrees of freedom corresponds to the twotailed ttest with n degrees of freedom. This ttest does not specify the direction ofthe difference between two treatment means for the alternative hypothesis; thus itis like the one-tailed Ftest which specifies which mean square is to be the largert— 142as the result of differences of unspecified direction between treatments. Thesetests can be shown to be algebraically equivalent; in particular t2= F. Therelation is shown graphically in Figure 2. Small numerical values of t, whensquared, become small values of F, positive quantities. Large numerical valuesof t, when squared, become large values of F (Steel and Torrie, 1980).”F(t2Fig. 2. Relation between two-tailed t and one-tailed F (curves are onlyapproximate) (From Steel and Torrie, 1980).ii. Homogeneity of regression coefficientsWhere the experimental design is a completely random one, theregression of V on X can be computed for each treatment. In this case, the usualt0 5%one—tailed F43assumption of homogeneity of the regression coefficients can be posed as a nullhypothesis and tested by an appropriate F test in an analysis of covanance(Sokal and Rohlf,1 969).The procedure that follows is an F test for difference between two regressioncoefficients:F=—2—2where S y. is the weighted average. For two groups we can write its formula as(Sokal and Rohlf,1 969).2 - (xy) 2 -_____‘ç .2 ‘V .2fli+ fl24Since there is a single degree of freedom in the numerator, t = ‘IF.b. Partial Least Squares (PLS) analysisPLS is a hybrid regression analysis and factor analysis which hasrecently been applied to diverse scaling problems in the natural and social44sciences (Bookstein, 1982; 1986; Wold, 1982). It is a method of data reduction.An investigator has collected two “blocks” of indicators and wishes to summarizethe predictive interrelations among the set of these two blocks consideredtogether. Each indicator was intended to tap some aspect of a constructunderlying its entire block. (For example, the construct for the general body itemis “develop”, and for the primary plate parameters is “palate”.) Yet our interest isnot so much in that underlying construct (its factors, its reliability, etc.) as in itscorrelations with other construct or constructs of the full data set, which are alsomeasured indirectly via their own indicators. Regardless of the correlationsamong primary palate parameters, different parameters are sensitive to effects ofbody size to different extents. We therefore wish to scale the items of each blockof indicators to best explain the cross-block relationships (correlations). Thesemutually scaled scores are the latent variables as they are constructed by PLSfor two blocks at the same time.Appendix 3 displays the ordinary correlation coefficients between each ofthree measures of development indices and each of the five primary palateparameters from five mouse strains. This array displays a clear pattern of signs:the correlations of all of the developmental indices with respect to all of theprimary palate parameters are positive. There appears to be a stable positivecorrelation between general body size development as measured in this battery,and primary palate growth as assessed by the parameters of five mouse strains.In combining different estimates of the same quantity that vary inprecision, it is standard to weight the contribution of each in proportion to its45precision, so that the more precise estimates are given more weight in formingthe average. Likewise, in attempting to construct a net score (latent variables) fordevelopment that is to correlate with palate, we should weight the development inproportion to their correlation with the sum of the primary palate parameters. Thisis what a partial least squares does.Such a two-block analysis is typically diagrammed as shown in Figure 3.Observed variables are indicated by squares and latent variables by circles. Thesingle line between the two latent variables indicates our intention to explain thepattern of correlations between observations of different blocks in terms of asingle pair of latent variables. We are not attempting to explain the correlationsamong indicators of the same block; instead we are determining the linearcombinations of the indicators in each block which are predictive of items in theopposite block.The prescription just given for the latent variables of interest can beexpressed simply in algebra as follows (Sampson et a!, 1989). The Developmentlatent variable (LV) scores are written asLVA = + . . . +=where A1,... ,A3 are the three developmental indicators, scaled to have varianceone, and x ,...,cz are three positive weights to be computed. The a s are to beproportional to the correlations of the A’s with a similarly weighted sum46TAIL50 MITEMAXILLARYPROMI NENCEBODYWEIGHTPRIMARYD PALATAL AREAQ MESENCHYMEAREAPRIMARYPALATAL DEPTHQ MAXIMAL PRIMARYPALATAL HEIGHTMAXIMAL MESENCHYME HEIGHTFig. 3. Diagram for a two-block latent variable model relating 3 indicatorsof development to 5 parameters from the primary palate.47LViBi +..+ =of the five primary palate parameters. That is,corr (A1, I31B)coy (Al, 23B)= (1)where is the correlation of development item i and palate item j, the (i, i)element of the matrix RAB of correlations given in Appendix 3. Note that all of the(scaled) palate variables are treated equally in determining the coefficient o ofthe development variable A1. The weights l3 of the palate variables are similarlyrequired to satisfy3cci=1 (2)For convenience, we scale the weights so that = = 1.“Thus each coefficient, aj or f3 is computed as a simple covariance, orregression coefficient (salience), corresponding to an optimal least squaresprediction using part of the data. For this reason we call the linear combinationcqA the” net partial predictor” (NPP) of LV B denoted NPP (LV B I A1,...These NPP’s stand in contrast to multiple regression predictors. Estimates of the48coefficients are typically computed using an iterative algorithm, alternatelyupdating estimates of the cx from equation (1), and then the 3i from equation (2).Such an iterative procedure constitutes a Partial Least Squares algorithm (Wold,1982). The algorithm may be summarized conveniently as follows:0. Initialize LVA cc (A1 + . . . + A3 ); that is, set cxl =. . . = (X3 = 1 I 131. Compute the linear combination of 5 primary palate parameters asLVB = I3B1 = NPP (LV,,, I B1,.. . ,B5), where each 6 is defined as inequation (2) with Z 32 = 1.2. Compute the linear combination of 3 developmental scores asLVA = cxA = NPP (LVB I A1,. . . ,A3), where each is defined as inequation (1) with Z a 2=1.3. Return to 1 and iterate until LVA and LVB fail to change to some presettolerance.” (Sampson et al, 1989).A computer program which implements this algorithm is from Dr. F. Bookstein.For example, the differences in the primary palate parameters, whichinclude the primary palate area and the mesenchymal bridge area, among fivemouse strains are analyzed by analysis of covariance, with the tail somite ascovariable to test the specific strain effects. In addition, the latent variables of tailsomite, body weight and maxillary prominence depth based on partial leastsquare analysis, were also used as the covariable in the study of strain specific49effects. The latent variable of tail somite, body weight, with the entailments ofmaxillary prominence growth were operated by normalizing each indicator byassigning 0 as the mean and 1 as the standard deviation. Normalized indicatorswere then multiplied by each simple covariance generated from partial leastsquare analysis called “salience”. Summation of each multiplied value is thelatent variable of each embryo. Such strain effects on primary palate growth weremore precisely expressed as residuals because the growth of maxillaryprominence is linearly related to the regional growth of primary palateparameters.IV. Error of measurementThe maxillary prominences were tested for strain specific effects amongfive strains by analysis of covariance using tail somite as the covariable. Theerrors of measurement of the position of maxillary prominence were determinedfrom both superior and sagittal view. As the angulation of the cutting may deviatelo right or left side of the head in superior view as shown in Figure 4, the realposition of the maxillary prominence can be expressed by the measured positionof the maxillary prominence and an angulation of 9. The angulation 0 can becalculated from the bimaxillary distance and difference between right and leftdepth of the maxillary prominence as they are cut in different levels. From theangulation and the measured position of the maxillary prominence, the true depthcan be calculated as in the equation of Figure 4. The error from angulation ofsection is then generated by deducting the true depth from the measured depth.50Tip of maxillary prominenceb_______CX=(a-(b-c-(axtanO))xtanO)xcosOa = measured depth of the maxillary prominence.b = half of the measured bimaxillary width.c = measured distance from the end of the nasal fin to midline.Fig. 4. Error of measurement for the depth of the maxillary prominencefrom the superior view. The cutting deviates from the perpendicular line tothe midline with an angle 0. This angle can be calculated from the bimaxillarywidth and the difference between left and right depth of maxillaryprominences. From the angulation and the equation shown in this graph, thetrue depth of maxillary prominence can be generated.Medial nasal prominence MidlineLateral nasal prominenceTip of maxillary prominenceAngulation of cutting___\__ 9/0I End of nasal fin Midline51Then the results of angle error were compared with standard deviation of positionof right maxillary prominence. In BALB/cByJ, the errors from angulation, whichaveraged two degrees, are 5.05 im at 13 tail somites and 10.87 urn at 16 tailsomites, while the respective standard deviations were 22.13 i.tm and 21.12 .tm.For NJ, the same errors of 13 and 16 tail somites are 6.54 urn and 8.70 p.m, andstandard deviations of 13 and 16 tail somites were 21.76 p.m and 21.80 p.m. Asthe errors fall into the standard deviation’s range, it is not necessary to correct thedata for error in angulation.From the sagittal view, as the angulation of the cutting may deviateforward or backward as shown in Figure 5, the position of the maxillaryprominence can be varied by the angulation of tipping as well. From thecalculation by using the mandibular prominence as reference, the angles oftipping can be calculated (Fig. 5). If the position of maxillary prominence, usingthe cutting through both tips of maxillary and mandibular prominence asreference, is considered as the true position, it can be calculated from measureddepth and angulation shown in Figure 5. The error from angulation of section isthen generated by deducting the true depth from the measured depth. The resultsof angle error were compared with the standard deviation of the position of theright maxillary prominence. In BALB/cByJ, the errors from angulation, whichaveraged twenty degrees, are 10.08 p.m at 13 tail somites and 14.58 p.m at 16 tailsomites, while the respective standard deviations were 22.13 p.m and 21.12 p.m.For NJ, the same errors for 13 and 16 tail somites are 8.74 p.m and 13.08 p.m, andstandard deviations for 13 and 16 tail somites were 21.76 p.m and 21.80 p.m. As52Fig. 5. Error of measurement of the depth of the maxillary prominence from thesagittal view. The cuttings may deviate forward or backward. The depth of themaxillary prominence are varied by the angulation 8. This angulation can becalculated from the distance between tip of maxillary and mandibular prominenceand the depth between two cuttings (equation 1). From the angulation and theequation 2, the true depth of maxillary prominence using the cutting through both thetip of maxillary and mandibular prominence as reference can be generated.Mandibular prominencesin 0= a/b/Tip of maxillaryprominence(1)(2)4— End of nasal finX=d xcos053the errors also fall into the standard deviation’s range, correction of the data forerror in angulation is not necessary.C. Stages of primary choana formationEmbryos of C57BLJ6J and CL/Fr were also collected for studying theformation of the primary choana. From the serial frontal sections, the structure ofthe oronasal membrane was identified and the cavitation of the oronasalmembrane to form the primary choana was studied in these two strains. The tailsomites of embryos with primary choanae formation were used as reference forthe comparison. Effects of strains on time of primary choana formation weretested by 2 X 2 Chi square tests.54RESULTSA. Cleft lip frequency and resorption rateI. Strain effectThe frequency of cleft lip from 10 litters of NJ is 4.0% (Table 2). Muchhigher cleft lip frequencies occurred in A/WySn and CL/Fr, 22.5% and 23.9%. Thearc sine transformed value was significantly smaller in A/J but did not differbetween A/WySn and CL/Fr (Table 2). Thus the cleft lip frequency falls into twogroups; NJ (low cleft lip frequency) versus A/WySn and CL/Fr (high cleft lipfrequency). Embryonic resorption rates are also shown in Table 2. Day 13 hr 12resorption rate was higher in A/J (18%) and lower in A/WySn and CL/Fr, 5.3%and 12.2%. The arc sine transformed values differed significantly between NJand NWySn.II. Uterine site effectThe relationship of cleft lip and resorption frequency and uterine positionwere analyzed by dividing each uterine horn into ovarian site and other sites. Themalformation and resorption frequencies in them were compared (Table 3). Thecleft lip frequency was higher in ovarian sites than other sites in all three strains;however, for each strain, the difference analyzed by the Chi square test was notsignificant. Analysis of the segments of three strains pooled for independence by2 X 2 tests showed that the frequency of cleft lip at the ovarian site (13/48 =27.1%) was significantly different (Fisher’s exact test, p = 0.014) from other sites(24/190 = 12.6%). However, the resorption rate at the ovarian site (7/55 = 12.7%)55Mean arc sine(%)resorbed ± SE126.37 ± 2.1 7a15.27± 234b21.19±5oab% cleft lipembryos4.022.523.6Mean arc sine(%)CL ± SE213.78 ± 217a29.80 ±30.59 ± 429bTable 2. Cleft lip and resorption rate in embryos of day 13 hour 12 of three cleft lip strains.1 F 2, 25 = 5.30; p < 0.05.2 F2 25=9;p< 0.001.a,b means sharing the same superscript do not differ (p > 0.05) by Duncan multiple-range test (Cody and Smith, 1987).No. of No. of Mean number ResorbedStrain Litters lmplantations of implantations (%)A/J 10 117 11.7 18.0A/WySn 8 75 9.4 5.3CL/Fr 10 82 8.2 12.2No. ofembryos967172Number of fetuses % cleft lip Number of implants % resorbedStrain Ovarian Other Ovarian Other Ovarian Other Ovarian Othersite sites site sites site sites site sitesNJ 16 80 12.5 2.5 20 97 20.0 17.5AIWySn 15 56 33.3 19.6 16 59 6.3 4.0CL/Fr 17 54 35.3 20.4 19 62 10.5 12.9All pooled 48 190 27.la 12.6 55 218 12•7b 12.8Table 3. Effects of implantation site on frequency of cleft lip and resorption.a % of cleft lip at the ovarian site is significantly different from other sites (Fisher’s exact test, p = 0.014)b % of resorption at the ovarian site is not different from other sites.was not significantly different from other sites (28/218 = 12.8%).B. The development of chronological age and tail somites of five strains duringprimary palate formationPrimary palate development in mice starts when the epithelia of the threefacial prominences fuse together. This occurs at 8 tail somites in C57BLJ6J andCL/Fr. The oronasal membrane breaks down to form the primary choana. Thisalso happens at 18 tail somites in CL/Fr and C57BLJ6J. Since the enlargementand elevation of the shelves of the secondary palate follows primary choanaeformation (Tamarin, 1982), the primary choana is considered as an anatomicaldelineation for successful primary palate development. The five strains of micestudied have a distribution of chronological age and tail somite stage as shownin Figure 6.NJ, NWySn and CL/Fr have the same chronological age (day 11 hour 2to day 11 hour 18) at the same interval of tail somites (8 to 18 tail somites).BALB/cByJ has the same chronological age as the three cleft lip strains at thesame interval of tail somites. In contrast, at the same range of tail somites thechronological age of C57BL/6J is younger (from day 10 hour 17 to day 11 hour11) than the above four strains.C. Delineation of phases of primary palate developmentFrom 8 to 12 tail somites, the early stage of primary palate developmentincluding the fusion of medial and lateral nasal epithelia to form the nasal fin and58Dl1/0Dl1/1Dl1/2D11/3Dl1/4Dl1/8Dli/91212121212121313131313131414141414141515151616141516161717171717171818181818181313141414141415151516161616171717178814151515151516171717121414151616168881010108891213Fig.6.Tailsomitedistribution(8to18tailcomparedwithprimarypalatedevelopmentspecimen(s)ofeachchronologicalageis/arethesectioned.somitestage)bystrains.Theembryo(s)coHectedandC57BL/6J8910119991010101011111212D10/17Dl0/18Di0/19D10120Di0/21Di0/22Dl0/23A/WvSnCL/FrDli/5D11/6Dli/7BALB/cByJ101112121314151511141414141112121314141515151515151516171710101113131313131414141516161415161617201517NJ14151313141515151313141414141515151616161616141516161718181818Dil/lODil/li011/12Dl 1/13Dl 1/14011/15D11/16D11/17Dl 1/18121215151515171718121313131415131414171717171719131416169991111111214151616161616171818812131314141515161616171714161713141516101113131416161714151559the forward growth of the maxillary prominence was studied in noncleft lip(C57BL/6J) and cleft lip (CL/Fr) strains. Depth of the nasal fin and position of themaxillary prominence of the noncleft lip strain was compared with the cleft lipstrains. From 13 to 16 tail somites, replacement of the epithelial seam withmesenchyme ingrowth forming the mesenchymal component is anticipatedduring. Areas of the primary palate and the mesenchymal component, andposition of the maxillary prominence, were compared among two noncleft lip(BALB/cByJ and C57BLJ6J), and three cleft lip (A/J, A/WySn and CL/Fr) strains.Primary choana formation occurs from 18 to 20 tail somites, after the primarypalate has definitely formed. Timing of primary choana formation of a noncleft lipstrain (C57BLJ6J) is compared with that of a cleft lip strain (CL/Fr).I. Early primary palate developmentAt the stage of 8 tail somites, in CL/Fr and C57BL/6J, the lateral wall ofthe nasal pit is formed frontally by the lateral nasal prominence, while morecaudally the maxillary prominence replaces the lateral nasal prominence. Thesetwo prominences are separated at the surface by the naso-maxillary groove (Fig.7). The medial nasal prominence, which forms the medial boundary of the nasalpit, now touches both the lateral nasal prominence frontally and the maxillaryprominence caudally (Fig. 7). The interposition of the epithelial plate or nasal finmaintains an epithelial continuity between the nasal cavity and the roof of themouth (Fig. 8). The definition of the back end of the nasal fin is the separation ofthe nasal epithelium and the oral ectodermal epithelium by the preoptic60Fig. 7. Frontal view of C57BL/6J, day 10 hour 17, 8 tail somitestage. The lateral nasal prominence (LNP) and maxillary prominence (MxP)are separated by the naso-maxillary groove. The medial nasal prominence(MNP) which forms the medial boundary of the nasal pit touches both thelateral nasal prominence frontally and maxillary prominence caudally.61Fig. 8. Frontal section of C57BL/6J at the level of nasal fin, day 10hour 17, 8 tail somite stage. The nasal fin (NF) which is composed ofmedial nasal prominence (MNP) at the medial side and lateral nasal (LNP)and maxillary prominence (MxP) at lateral side maintains the continuitybetween the nasal cavity and the roof of the mouth.62mesenchyme (Fig. 9).The maxillary prominence in C57BLJ6J mice extends frontally past theback end of the nasal fin with a positive position of the maxillary prominencerelative to the end of the nasal fin at 8 tail somites (Table 4). In contrast theposition of the maxillary prominence in CL/Fr is behind the end of the nasal fin at8 tail somites. Thus the position of the maxillary prominence is defined negative(Table 4). Calculations of the number of sections show that the depth of nasal finof C57BLJ6J is not significantly larger than that of CLJFr.From the stage 9 to 12 tail somites, the chronological age of CL/Fr andC57BL/6J is approximately 11 days 8 hours to 11 days 14 hours and 10 days 20hours to 11 days 0 hours (Fig. 6). The location of nasal fin, indicated on thesurface by a groove, becomes deeper and marks the boundary between themedial nasal, lateral nasal and maxillary prominence in both the CL/Fr andC57BLJ6J embryos (Fig. 10). Internally the proliferation of mesenchymal tissueinside the maxillary, lateral nasal and medial nasal prominence is very active inthis stage. The nasal fin is not replaced by the mesenchymal cells in this stage(Fig. 11). By calculation of the nasal fin of CL/Fr frontocaudally the depth issmaller than C57BLJ6J, although only left sides of 10 and 12 tail somites aresignificantly different (Table 4). The right and left maxillary prominences, whichbulge on the surface as a prominent ridge, are still widely separated. In CL/Fr themaxillary prominence extends ventromedially over the end of the nasal fin with apositive position relative to the end of the nasal fin. The position of the maxillaryprominence in CL/Fr is significantly less advanced than in C57BL/6J at 10 and 1263•1.•50 pmFig. 9. Frontal section of C57BL/6J at the end of nasal fin, day 10hour 17, 8 tail somite stage. The end of the nasal fin is shown in thisphotograph. The nasal epithelium (NE) and oral epithelium (CE) areseparated by the preoptic mesenchyme (between the arrow heads). Thus theepithelial continuity does not exist between the nasal cavity and the roof of themouth.LNPMxP—64TS Strain (n) Nasal Fin Depth (l.tm) Position of maxillary, prom. (tim)Right Left Right LeftMean±S.D. Mean±S.D. Mean±S.D. Mean±S. D.8 C57 3 39.7±8.0 46.7±14.6 28.0±18.5 37.3±4.0CL/Fr 6 30.8±6.3 32.2±3.8 -18.2±26.5 -12.6±18.19 C57 4 64.7±22.4 69.3±23.5 54.3±25.2 51 .8±28.3CL/Fr 4 54.3±6.7 57.8±1 4.4 28.0±1 5.1 30.3±4.010 C57 5 146.4±34.9 137.2±28.3 102.2±25.5 91.0±16.4CL/Fr 6 99.2±60.8 80.5±32.2 * 45.5±21.1* 47.8±28.8*11 C57 3 182.0±7.0 175.0±14.0 116.7±22.5 105.0±12.1CL/Fr 5 128.8±45.0 133.0±40.0 86.8±34.9 75.6±31.512 C57 5 221.2±13.7 226.2±21.2 142.8±15.3 148.4±15.2CL/Fr 4 127.8±78.2 89.3±61.9 * 75.3±14.5* 66.5±4.0 *Table 4. Nasal fin depth and the position of the maxillaryprominence of C57BL/6J and CL/Fr from 8 to 12 tail somite stage(TS). The position of the maxillary prominence depth of CL/Fr is negativerelative to the end of the nasal fin at 8 TS because the position of the tip of themaxillary prominence is behind the end of the nasal fin. At 10 and 12 TS, theposition of the maxillary prominence are significantly less advanced in CL/Frthan in C57BL/6J. The nasal fin depth is smaller in CL/Fr as well; however,only left sides of 10 and 12 TS are significantly different.* p <0.05 (between CL/Fr and C57).65Fig. 10. Frontal view of C57BL/6J, day 11 hour 0, 12 tail somitestage. The boundary among the medial nasal prominence, lateral andmaxillary prominences is marked by a deeper groove which indicates theposition of the nasal fin.66Fig. 11. Frontal section of C57BL/6J at the level of nasal fin, day11 hour 0, 12 tail somite stage. The proliferation of mesenchymal tissueinside the maxillary, lateral and medial nasal prominence is very active. Nonasal fin is replaced by mesenchymal tissue (between the arrow heads).LNPMNP-67tail somites (Table 4).Since the correlation between the depth of right and left side of nasal finand the maxillary prominence is significant (R = 0.99 in nasal fin depth and 0.94in maxillary prominence for C57BLI6J; R = 0.79 in nasal fin depth and 0.86 inmaxillary prominence for CL/Fr), the right side was chosen randomly for testingthe strain effect on the growth of the nasal fin and forward movement of themaxillary prominence. The analysis of covariance was used to analyze andcompare the growth patterns of the nasal fin and the maxillary prominence. Itcombines the methods of regression and analysis of variance. The basic problemis to make inferences about group means of a dependent variable, such as depthof nasal fin or position of the maxillary prominence, that is measured in .tm.Another variable, the tail somite number, called a covariable, is measured inwhole units. Analysis of covariance makes use of information about depth ofnasal fin or position of the maxillary prominence compared with tail somitedevelopment. Associated with the analysis of covariance is the study ofdifferences in regression relationships among the groups by the test forheterogeneity of slopes. Covariance analysis test for differences in interceptsassuming a constant regression relationship among groups. The test forheterogeneity is a test for the validity of this assumption, and it tests whether ornot the regression coefficients are constant over groups (Freund and Littell,1981). Analysis of covariance of the nasal fin depth with the tail somite as acovariable showed that CL/Fr nasal fin depth increases slower than C57BL/6J’s(p < 0.05) (Fig. 12, Table 5). However, covariance analysis of the position of right680 C57=IUizU--JCl)zI—=300200100y = - 353.35 + 48.537x RA2 = 0.898• CL/FrTAIL SOMITEFig. 12. Regression lines of the right nasal fin depth of C57BL/6Jand CL/Fr from 8 to 12 tail somite stage. The growth is slower in CL/Frthan in C57BIJ6J (p <0.01) (Table 5). Equations include slopes, intercepts,and squares of correlation coefficients. Linear regressions are not significantin test for lack of fit of C57BL/6J (F = 2.0, p> 0.05) and CL/Fr (F = 0.38, p>0.05).y = - 187.57 ÷ 27.241x R’2 = 0.51500••0••7 8 9 10 11 12 1369GENERAL LINEAR MODELS PROCEDUREDEPENDENT VARIABLE: RIGHT NASAL FIN DEPTHT -value Standard errorParameter Estimate for hypothesis Probablity of estimateof slope parameter=0 for T-value of slopeTS:C57 48.54 7.90 0.0001 6.14TS:CL/Fr 27.24 4.81 0.0001 5.69TS*ST C57 21.30 2.53 0.015 0.37CL/Fr 0.00 .Table 5. Test for homogeneity of slopes of right nasal fin from 8 to12 tail somite stage (TS) between C57BL/6J (C57) and CL/Fr.Regression coefficients of C57BLI6J and CL/Fr are tested for heterogeneity.There is significant difference in the nasal fin depth / tail somite relationship fordifferent strains (ST) by setting to zero effect of the CL/Fr (p = 0.015).70maxillary prominence shows that slopes of these two strains are homogeneousand intercepts of these two strains are significantly different (Fig. 1 3, Table 6).II. Primary palate development with mesenchymal formationFive strains were included for studying the regression of nasal fin andmesenchyme formation. In noncleft lip strains BALB/cByJ and C57BL/6J at 13 tailsomites, most of nasal fin, which maintains the continuity between the nasalcavity and the roof of the oral cavity, becomes interrupted by the activeproliferation of the mesenchyme of the maxillary prominence, lateral nasalprominence and medial nasal prominence, growing across from one side to theother (Table 7) (Fig. 14). This forms the primary palate. In contrast, most of theembryos of the cleft lip strains still maintain the nasal fin at 1 3 tail somites. Theingrowth of mesenchyme across the nasal fin does not take place until 14tailsomites in A/J, 15 tail somites in A/WySn and 16 tail somites in CL/Fr (Table 7). ACL/Fr embryo of 15 tail somites as shown in Figure 15 without any mesenchymeacross the nasal fin may develop into a cleft lip foetus. The internal structuresincluding areas of primary palate and mesenchymal component and the positionof maxillary prominence of the five strains of mice of 13-16 tail somites are listedin Appendix 2. The correlation between right and left sides are high in bothnoncleft lip strains (Table 8) and cleft lip strains except for the mesenchymalcomponent of CL/Fr (0.38) (Table 9). Only 7 of 37 embryos have bilateralmesenchymal formation, 23 embryos have no mesenchymal formation and 7embryos have mesenchymal formation in one side in CL/Fr. This correlation71TAIL SOMITEFig. 13. Regression lines of the position of the right maxillaryprominence of C57BL/6J and CL/Fr from 8 to 12 tail somite stage.The growth rates are not significantly different between C57BL/6J and CL/Fr(p > 0.05) and the group mean of C57BL/6J is significantly larger than CL/Fr (p<0.01) (Table 6). Equations include slopes, intercepts and squares ofcorrelation coefficients. Linear regressions are not significant in testing for lackof fit of C57BL/6J (F = 0.78, p > 0.05) and CL/Fr (F = 1 .82, p > 0.05).y = - 198.88 + 28.835x RA2 = 0.795 El C57y = - 211.60 + 25.554x RA2 = 0.617 • CL/FrElDELiiC.)zLUz00.>-—I-J>cI—z0L.0z0ICl,00a2001000-100.7 8 9 10 11 12 1372GENERAL LINEAR MODELS PROCEDUREDEPENDENT VARIABLE: POSITION OF THE RIGHT MAXILLARYPROMINENCE (RMXP)LEAST SQUARE MEANS (LSMEAN)Probability ofST RMXP (jim) Standard error Probability of hypothesisLSMEAN of LSMEAN hypothesis LSMEAN C57=LSMEAN=0 LSMEAN CL/FrC57 90.97 5.71 0.0001 0.0001CL/Fr 45.23 5.21 0.0001T -value Standard errorParameter Estimate for hypothesis Probablity of estimateof slope parameter=0 for T-value of slopeTS:C57 28.84 6.96 0.0001 4.14TS:CLJFr 25.55 6.66 0.0001 3.84TS*ST C57 3.29 0.58 0.564 5.65CL/Fr 0.00Table 6. Analysis of covariance of the position of the rightmaxillary prominence of C57BL/6J (C57) and CL/Fr from 8 to 12tail somite stage (TS). The difference of group least square meansbetween these two strains (ST) is significant (p <0.01). Regressioncoefficients of C57BL/6J and CL/Fr are tested for heterogeneity of slopes.There is no significant difference in the position of the maxillary prominence /tail somite relationship for different strains by setting to zero effect of the CL/Fr(p > 0.05).73BALB/cBy C57BL/6J NJ NWySn CL/Fr% % % % %No. of mesenchymal No. of mesenchymal No. of mesenchymal No. ot mesenchymal No. of mesenchymalTS embryos formation embryos formation embryos formation embryos formation embryos formation13 4 75 5 100 4 0 5 0 7 1414 7 100 7 100 7 71 6 17 10 1015 7 100 7 100 8 100 6 50 5 016 4 100 7 100 6 83 5 80 15 66Table 7. Th percentage of embryos which have mesenchyme replace the nasal fin of the right side at eachtail somite stage (TS) of the five strains.Fig. 14. Frontal section of C57BL/6J showing nasal epithelium,mesenchymal component and oral ectoderm, day 11 hour 0, 13tail somite stage. The mesenchymal tissue with a blood vessel inside themaxillary, lateral and medial nasal prominence starts to grow across the nasalfin. The nasal fin becomes nasal epithelium (NE) on the top, mesenchymalcomponent in the middle (between the arrow heads) and oral ectoderm (CE)on the bottom. The primary palate area is defined as including these threeparts.75MxP OEFig. 15. Frontal view (A) and a frontal section (B) of CL/Fr embryo,day 11 hour 13, 15 tail somite stage. The medial and lateral nasalprominence stay apart from each other (A). The nasal fin is small in the right ofthe embryo and has barely contact in the left (B). No mesenchymal componentformation was found in this section and in this embryo.76oPEARSON CORRELATION COEFFICIENTSBALB/cByJ N=22LMXP LPP LM LPD LPH LMHRMXP 0.88 0.83 0.87 0.76 0.84RPP 0.84 0.83 0.88 0.81 0.86RM 0.82 0.88 Q. 0.88 0.73 0.85RPD 0.84 0.87 0.83 0.63 0.75RPH 0.77 0.81 0.69 0.69 Q. 0.85RMH 0.85 0.90 0.88 0.83 0.87C57BL/6J N=26LMXP LPP LM LPD LPH LMHRMXP 0.84 0.86 0.64 0.82 0.86RPP 0.87 0.90 0.80 0.79 0.87RM 0.88 0.88 0.71 0.81 0.89RPD 0.52 0.70 0.61 0.35 0.57RPH 0.82 0.76 0.79 0.43 0.85RMH 0.87 0.85 0.89 0.62 0.90Table 8. The correlation coefficients between the right and leftsides of the maxillary prominence, primary palatal andmesenchymal components in noncleft lip strains from 13 to 16 tailsomite stage. The correlation coefficients between the same parameters ofthe right and left sides are underlined. The abbreviations are explained inAppendix 1.77PEARSON CORRELATION COEFACIENTSNJ N=25LMXP LPP LM LPD LPH LMHRMXP 0.90 0.77 0.87 0.72 0.84RPP 0.76 9..Q 0.71 0.89 0.80 0.76RM 0.80 0.75 0.63 0.62 0.82RPD 0.69 0.86 0.59 Q..4 0.70 0.65RPH 0.58 0.83 0.60 0.77 I.4 0.64RMH 0.83 0.83 0.85 0.69 0.71NWySn N=22LMXP LPP LM LPD LPH LMHRMXP 0.91 0.69 0.76 0.86 0.68RPP 0.94 0.64 0.85 0.89 0.65RM 0.66 0.64 0.48 0.65 0.81RPD 0.78 0.78 0.35 0.68 0.36RPH 0.90 0.82 0.65 0.68 0.67RMH 0.72 0.75 0.81 0.61 0.71CL/Fr N=37LMXP LPP LM LPD LPH LMHRMXP 0.77 0.43 0.73 0.76 0.58RPP 0.81 0.34 0.79 0.73 0.44RM 0.59 0.51 *0.38 0.44 0.48 0.54RPD 0.67 0.68 0.21 0.55 0.26RPH 0.84 0.76 0.41 0.66 0.56RMH 0.73 0.67 0.56 0.56 0.64 *0.65Table 9. The correlation coefficients between the right and leftsides of the maxillary prominence, primary palatal andmesenchymal components in cleft lip strains from 13 to 16 tailsomite stage. The correlation coefficients between the same parameters ofthe right and left sides are underlined. The abbreviations are explained inAppendix 1.* Only 7 of 37 embryos have bilateral mesenchymal formation (Appendix 2).78coefficient increases from 0.38 to 0.87 by studying the growth of themesenchymal component of CL/Fr from 13 to 19 tail somites. Thus, analysis ofcovariance in the growth of primary palate, mesenchymal component andposition of the maxillary prominence of right side only are presented in the dataanalysis. The increases of the mesenchymal components of the cleft lip strainswere studied from 13 to 17 tail somites in NJ and from 13 to 19 tail somites inNWySn and CL/Fr.a. Analysis of primary palate area formationThe growth of the primary palate area including both the area replacedand the area not replaced by mesenchyme relative to the tail somite wasanalyzed among the five strains by the analysis of covariance. The purpose ofusing covariance analysis is to analyze and compare the growth pattern oi theprimary palate of the noncleft lip and cleft lip strains. The analysis is to determineif the increases of primary palate area from 13 to 16 tail somites of the five strainshave a constant growth rate and if the group least square means of the primarypalate area are significantly different among these five strains.The results of the analysis of covariance shows that slopes of the fivestrains are homogeneous (p> 0.01) and group mean of the primary palate areaof BALB/cByJ is significantly larger than those of the three cleft lip strains (p <0.01). In the noncleft lip strains, the group mean of primary palate area ofBALB/cByJ is larger than that of C57BLJ6J. In the cleft lip strains, A/J has a largergroup mean of primary palate area than both NWySn and CL/Fr while the mean79of AIWySn is also larger than CL/Fr (Fig. 16, Table 10). Linear regressions werealso examined for validity in each strain by conducting tests of lack of fit of thelinear regression model. All five strains are not significant in these tests ( p >0.01). Thus, cleft lip strains have smaller primary palate area than noncleft lipstrains during the 13 -16 tail somite stage of development. Each strain in noncleftlip and in cleft lip groups has a significantly larger or smaller primary palate thanother strains during the 13-16 tail somite stage of development as well.b. Partial least squares and covariance analysis of the primary palate areaformationThe latent variables of development including tail somite number, bodyweight, and position of the maxillary prominence were first normalized with 0 asthe mean and 1 as standard deviation. Normalized indicators were thenmultiplied by each simple covariances generated from partial least squarescalled ‘salience’ as shown in Table 11. The position of the maxillary prominencehas been shown the best predictor for primary palate formation among the threedevelopmental scores (Table 11). Then the summation of each multiplied valuewas operated to generate the latent variable of development for each embryo.The plot of primary palate area on latent variable of development in each strainshows that this grouping simplifies the linearization of smoothed scatters betweenprimary palate area and latent variable with a higher square of correlationcoefficient in each strain (Fig. 17) compared to linear regression between primarypalate area and tail somite (Fig. 16), especially in the cleft lip strains. Results of80y = - 5.7143e+4 + 7565.4x RA2 = 0.664 BALB/cByJC••4 y = - 5.7781e+4 + 6683.7x R’2 = 0.619 0 C57BL/6JE y = - 8.0757e+4 + 7988.5x RA2 = 0.633 m NJ80000 y = - 9.2416e+4 + 8361.4x RA2 = 0.725 • A/WySnw y 4 21 69e+4 + 4477 6x R’2 0 181 CL/Fr60000UII-J0.. 40000>-20000 :0F— 1 00 o•12 13 14 15 16 17TAIL SOMITEFig. 16. Regression lines of the right primary palate area of thefive strains from 13 to 16 tail somite stage. The keys are in same orderas lines. The slopes of five strains are all not significantly different (p > 0.01)and the group means of five strains are all significantly different (p <0.01)except between C57BLI6J and NJ (Table 10). Linear regressions are notsignificant in test for lack of fit of each strain. Equations include slopes,intercepts and squares of correlation coefficients.81GENERAL LINEAR MODELS PROCEDUREDEPENDENT VARIABLE: RIGHT PRIMARY PALATE AREA (RPP)LEAST SQUARE MEANS (LSMEAN)Standard probabilityST RPP (J.tm2) error for of hypothesis Probability of hypothesis:LSMEAN(I)=LSMEAN(J)LSMEAN LSMEAN LSMEAN=0 1/J 1 2 3 4 5BALB 53250 1705 0.0001 1C57 39842 1567 0.0001 2 0.0001NJ 35891 1631 0.0001 3 0.0001 0.0832NW 29518 1705 0.0001 4 0.0001 0.0001 0.0079CL/Fr 22959 1299 0.0001 5 0.0001 0.0001 0.0001 0.0027Estimate Test for hypothesis Probability Standard error ofPARAMETER of slope Parameter=0 forT-value estimate of slopeTS*ST BALB 3087.86 1.52 0.1321 2036.87C57 2206.09 1.21 0.2276 1819.19NJ 3510.96 1.79 0.0765 1965.32A/W 3883.77 2.02 0.0454 1920.95CL/Fr 0.00Table 10. Analysis of covariance of growth of right primary palatearea from 13 to 16 tail somite stage (TS) of the five strains (ST).The group least square means are all significantly different among five strainsexcept between A/J and C57BLJ6J (p = 0.0832). Slopes are not significantlydifferent by setting to zero effect of the CL/Fr (p 0.0454).82Latent Variables BALB CL/FrDevelopmentTail somite 0.39 0.37 0.38 0.38 0.37Position of the right maxillary 0.39 0.39 0.58 0.40 0.65prominenceBody weight 0.32 0.34 0.20 0.31 0.18Primary palateRight primary palate area 0.21 0.24 0.22 0.27 0.26Right mesenchymal area 0.23 0.24 0.25 0.16 0.16Right primary palate depth 0.21 0.16 0.19 0.25 0.23Right maximal primary palate 0.19 0.22 0.17 0.25 0.28heightRight maximal mesenchymal 0.23 0.24 0.27 0.19 0.21heightTable 11. Two-block partial least squares analysis of the threedevelopmental scores and the five primary palate parametersblocks of the five strains. The values indicate the simple covariance,saliences, of development and primary palate. The salience is the contributionof each indicator in proportion to its precision. The position of the rightmaxillary prominence is the more precise estimate and gives more weight informing the latent variable of development than tail somite and body weightespecially in the cleft lip strains.83y = 5.2556e+4 + 8073.3x Rd%2 = 0.733 • BALB/cByJC’.JE:1LULUI-j0>-0I080000600004000020000LATENT VARIABLE OF DEVELOPMENTo C57BLI6JEl NJ• A/WySn• CL/FrFig. 17. Regression lines of the right primary palate area on thelatent variable of development including the tail somite, bodyweight and position of the right maxillary prominence of the fivestrains. The keys are in same order as lines. The slopes of five strains are allnot significantly different (p> 0.01) and the group means are all significantlydifferent (p < 0.01) (Table 1 2). Regression lines are not significant in test forlack of fit of each strain (p > 0.05). Equations include slopes, intercepts, andsquares of correlation coefficients. The square of correlation coefficient ofCL/Fr increases from 0.181 to 0.613 compared with the regression of rightprimary palate area on tail somite as shown in Fig. 16 because the rightmaxillary prominence provides more weight on the growth of primary palate inCL/Fr.y = 3.9904e+4 + 8237.3x RA2 = 0.783y = 3.5646e+4 + 8199.4x RA2 = 0.678y = 2.8824e+4 + 1.0042e+4x RA2 = 0.859y = 2.3817e+4 + 9763.8x R’2 = 0.613UEl 00000-3 -2 -1 0 1 284the regression of this area on latent variable of tail somite, body weight, andmaxillary prominence also indicate that there is no difference in slope. Also, thereare differences by strains in the intercepts of these regressions with the order ofBALB/cByJ > C57BL/6J > NJ > NWySn > CL/Fr (p < 0.05) (Table 12). Thus, thedelayed formation of the primary palate area can be partly attributed to thedelayed forward growth of the maxillary prominence.c. Analysis of mesenchymal component formationAs shown in Table 7, mesenchymal component formation of BALB/cByJand C57BLJ6J starts at 13 tail somites; 71% NJ at 14 tail somites, 80% NWySn at16 tail somites and 67% CL/Fr at 16 tail somites. To verify the analysis ofcovariance, these five strains have been divided into three groups. The first group(noncleft lip group) is composed of C57BLJ6J and BALB/cByJ as most of theembryos at 13 tail somites had mesenchymal formation in both strains. Thesecond group (low cleft lip frequency group) is NJ which shows mesenchymalcomponent formation at 14 tail somites. The third group (high cleft lip frequencygroup) is composed of A/WySn and CL/Fr as most of the embryos of these twostrains have mesenchymal component formation at 16 tail somites. To analyzethe formation of the mesenchyme, the noncleft lip and NJ strains wereinvestigated from 13 to 17 tail somites while both A/WySn and CL/Fr strains wereinvestigated from 13 to 19 tail somite animals to generate regression forincreases of the mesenchyme. The embryos without mesenchymal componentformation were taken away from the data to avoid bias in covariance analysis.85GENERAL LINEAR MODELS PROCEDUREDEPENDENT VARIABLE: RIGHT PRIMARY PALATE AREA (RPP)LEAST SQUARE MEANSStandard probabilityST RPP (p.m2) error for of hypothesis Probability of hypothesis:LSMEAN(I)=LSMEAN(J)LSMEAN LSMEAN LSMEAN=0 IJJ 1 2 3 4 5BALB 52995 1236 0.0001 1C57 39922 1137 0.0001 2 0.0001NJ 35656 1184 0.0001 3 0.0001 0.0105NW 28842 1236 0.0001 4 0.0001 0.0001 0.0001CL/Fr 23835 940 0.0001 5 0.0001 0.0001 0.0001 0.0016Estimate Test for hypothesis Ptobability Standard error ofPARAMETER of slope Parameter=0 forT-value estimate of slopeLVDEV*STBALB-1809.61 -1.15 0.2517 1571.13C57 -1526.46 -1.01 0.3140 1509.83NJ -1564.45 -1.03 0.3073 1526.15NW 278.11 0.17 0.8620 1596.51CL/Fr 0.00Table 12. Analysis of covariance of growth of right primary palatearea on latent variable of development (LVDEV) of five strains(ST). The group least square means of five strains are all significantlydifferent and regression coefficients are not significantly different by setting tozero effect of the CL/Fr.86Analysis of covariance of two noncleft strains shows homogeneous slopeand same intercepts of mesenchyme between these two normal strains (p>0.05). In the high cleft lip frequency group, the same analysis of covariance ofgrowth trends of A/WySn and CL/Fr shows the same slope and same intercepts(p > 0.05) between these two high cleft lip frequency strains as well. Betweennoncleft lip and cleft lip strains, the slopes are homogeneous and intercepts ofnoncleft lip strains are significantly larger than cleft lip strains (Fig. 1 8, Table 13).Mesenchymal area of NJ has significantly smaller intercept compared with bothBALB/cByJ and C57BLI6J to a common slope and significantly larger interceptthan CL/Fr to a common slope (Fig. 18, Table 13). Tests of lack of fit of five linearregressions show that all five strains are not significant (p> 0.01).The mesenchymal areas of these strains were also subjected to analysisof covariance with latent variable of tail somite, body weight and position of themaxillary prominence as covariable. As mesenchymal area data of AJWySn andCL/Fr has been extended to 19 tail somites while the other three strains havebeen extended to 17 tail somites only, the latent variables are generated for allthe strains pooled. Results show that in the five strains, the slopes arehomogeneous. The squares of correlation coefficients for regression lines arebetter than those for regression lines against tail somites only. Within two noncleftlip strains there is no difference in group mean of mesenchymal area betweenBALB/cByJ and C57BL/6J; in the high cleft lip frequency group, there is nodifference in group mean between A/WySn and CL/Fr as well (Fig. 19, Table 14).There is a difference in group means of mesenchymal area between noncleft lip87y = - 7.0971e+4 + 5828.5x RA2 = 0.833 i BALB/cByJELULU>-IC)zUiCl)LUIIC-,5000040000300002000010000018o C57BL/6JEl NJ• A/WySn• CLJFrFig. 18. Regression lines of growth of the right mesenchymal areaof the five strains from 12 to 19 tail somite stage. The keys are insame order as lines. They are all not significantly different in slopes. Thegroup means fall into three significantly different groups which are BALB/cByJand C57BL/6J (noncleft lip strains), A/J (low cleft lip frequency strain) andA/WySn and CL/Fr (high cleft lip frequency strains) (Table 13). The data of‘zero’ mesenchyme has been taken out from the analysis to reduce the bias ofregression. Linear regressions are not significant in test for lack of fit of eachstrain (p > 0.05) although the fit of A/WySn and CL/Fr to linear regression ispoor. Reasons are that very few embryos in the early tail somite stages whichhave mesenchymal formation in A/WySn and CL/Fr are included in theanalysis.y = - 8.1395e+4 + 641 4.Ox R’2 = 0.798yy = - 8.9109e+4 + 6365.6x RA2 = 0.649y- 9.0673e+4 + 6244.4x RA2 =- 9.9296e+4 + 6671.8x0.4350RA2 = 0.4780QU•010 12 14 16TAIL SOMITE2088GENERAL LINEAR MODELS PROCEDUREDEPENDENT VARIABLE: RIGHT MESENCHYMAL AREA (RM)LEAST SQUARE MEANS (LSMEAN)Standard probabUityST RM (g2) error for of hypothesis Probability of hypothesis:LSMEAN(I)=LSMEAN(J)LSMEAN LSMEAN LSMEAN=0 IIJ 1 2 3 4 5BALB 19822 1366 0.0001 1C57 17962 1208 0.0001 2 0.2916NJ 9540 1449 0.0001 3 0.0001 0.0001NW 6054 1586 0.0002 4 0.0001 0.0001 0.1123CL/Fr 4546 1408 0.0016 5 0.0001 0.0001 0.0170 0.4519Estimate Test for hypothesis Probability Standard error ofPARAMETER of slope Parameter=0 for T-value estimate of slopeTS*ST BALB -843.35 -0.61 0.5440 1385.36C57 -257.85 -0.19 0.8483 1344.56NJ -306.25 -0.17 0.8692 1856.03NW -427.45 -0.26 0.7954 1644.46CLJFr 0.00Table 13. Analysis of covariance of growth of right mesenchymalarea on tail somite stage (TS) of five strains (ST). The group leastsquare means are significantly different between noncleft lip and cleft lipstrains and between NJ and CL/Fr (low and high cleft lip frequency strains).Regression coefficients are not significantly different by setting to zero effect ofthe CL/Fr.89y = 1.7232e+4 + 1.0519e+4x R2 = 0.854 BALBIcByJEzLLULU>-IC.)zLUcrLUII0700006000050000400003000020000100000LATENT VARIABLE OF DEVELOPMENTo C57BLJGJ• A1WySnNJ• CL/FrFig. 19. Regression lines of right mesenchymal area on latentvariable including the tail somite, body weight and position of theright maxillary prominence of the five strains. The keys are in sameorder as lines. The slopes of five strains are homogeneous.The group meansalso fall into three significantly different groups which are BALB/cByJ andC57BLJ6J (noncleft lip strains). NJ (low cleft lip frequency strain) and A/WySnand CL/Fr (high cleft lip frequency strains) (Table 14). The squares ofcorrelation coefficient (R2) for regression lines are better than those forregression against somites. Linear regressions are not significant in test forlack of fit of each strain. Equations include slopes, intercepts and squares ofcorrelation coefficients.y = 1.5560e+4 + 1.1092e+4x R’2 = 0.856y = 7859.1 + 1.2253e+4x RA2 = 0.731y = 1.3892e+4 + 9331.lxy = 6727.7 ÷ 1.0565e+4xRA2= 0.693R’2 0.526.-3 -2 -1 0 1 2 3 490GENERAL LINEAR MODELS PROCEDUREDEPENDENT VARIABLE: RIGHT MESENCHYMAL AREA (RM)LEAST SQUARE MEANS (LSMEAN)PARAM ETEREstimate Test for hypothesisof slope Parameter=0Probabilityfor T-valueStandard error ofestimate of slopeStandard probabilityST RM Qim2) error for of hypothesis Probability of hypothesis:LSMEAN(l)=LSMEAN(J)LSMEAN LSMEAN LSMEAN=0 lJJ 1 2 3 4 5BALB 17248 1147 0.0001 1C57 15437 1031 0.0001 2 0.2403NJ 12198 1292 0.0001 3 0.0035 0.0511NW 8220 1350 0.0001 4 0.0001 0.0001 0.0389CL/Fr 6465 1178 0.0001 5 0.0001 0.0001 0.0019 0.3172LVDEV*ST BALB -45.49- 0.02 0.9819 1998.68C57 527.03 0.28 0.7778 1862.91NJ -539.33 -0.22 0.8249 2431.77NW 1688.47 0.78 0.4383 2170.27CL/Fr 0.00Table 14. Analysis of covariance of growth of right mesenchymalarea on latent variable of development (LVDEV) of five strains(ST). The group least square means fall into three significantly differentgroups which are BALB/cByJ and C57BLI6J (noncleft lip strains), NJ (low cleftlip frequency strain) and NWySn and CL/Fr (high cleft lip frequency strains).Regression coefficients are not significantly different by setting to zero effect ofthe CL/Fr.91and cleft lip strains and between low cleft lip frequency strain, NJ and high cleftlip frequency strains, A/WySn and CL/Fr. In summary, the mesenchymal areas ofthe five strains fall into three significantly different groups, noncleft lip strains, lowcleft lip frequency, and high cleft lip frequency strains.d. Analysis of growth of the maxillary prominenceThe growth of the maxillary prominence is analyzed in two levels. First,the position of the maxillary prominence versus tail somite is compared byanalysis of covariance for cleft and noncleft lip strains. Second, growth of primarypalate area and mesenchymal area was investigated by using the position of themaxillary prominence as covariable in analysis of covariance. Growth of primarypalatal and mesenchymal area at certain positions of the maxillary prominencewas compared within noncleft and cleft lip strains.Analysis of covariance of the position of the maxillary prominence versustail somite number indicated no significant difference in the slopes among the fivestrains. There are differences noticed in intercepts between noncleft lip and cleftlip strains (Fig. 20, Table 15). There is no difference between the noncleft strainsBALB/cByJ and C57BLI6J at P = 0.05 level and between the cleft lip strainsA/WySn and CL/Fr at p = 0.05 level. The amount of jut of maxillary prominencebeyond the end of the nasal fin is greater in A/J than A/WySn and CL/Fr and lessthan in BALB/cByJ and C57BLJ6J at P = 0.05 level (Fig. 20, Table 15).The frequency of cleft lip in NJ is 4.0% while the cleft lip frequency inA/WySn and CL/Fr is 22.5 and 23.6% as shown in the data. The deficiency of the92y = - 138.45 + 23.767x RA2 = 0.722 0 C57BL/6JI -13 14TAIL• BALB/cByJa A/J• A/WySn• CLJFrFig. 20. Regression lines of the position of the right maxillaryprominence relative to the end of the nasal fin of the five strainsfrom 13 to 16 tail somite stage. The keys are in same order of lines. Theslopes of five strains are parallel and the group means fall into threesignificantly different groups, noncleft lip strains (BALB/cByJ and C57BLJ6J),low cleft lip frequency strain (NJ) and high cleft lip frequency strains (AJWySnand CL/Fr) (Table 15) which are similar to the regressions of mesenchymearea compared with tail somites as shown in Fig. 18. Linear regressions arenot significant in test for lack of fit of each strain. Equations include slopes,intercepts and squares of correlation coefficients.0w y = - 153.96 + 25.161x RA2 = 0.708y = - 278.63 + 30.986x RA2 = 0.802wZ 300 y = - 306.69 + 31.706x R’2 = 0.684y = - 167.77 + 21.559x RA2 = 0.301 02001x<I—z!2 100U0z0ECl)00.•00• 80015 16SOMITE1793GENERAL LINEAR MODELS PROCEDUREDEPENDENT VARIABLE: POSITION OF RIGHT MAXILLARY PROMINENCE (RMXP)LEAST SQUARE MEANS (LSMEAN)Standard probabilityST RMXP (urn) error for of hypothesis Probability of hypothesis:LSMEAN(I)=LSMEAN(J)LSMEAN LSMEAN LSMEAN=0 IJJ 1 2 3 4 5BALB 213.41 5.77 0.0001 1C57 213.52 5.30 0.0001 2 0.9880NJ 173.84 5.52 0.0001 3 0.0001 0.0001A/W 155.81 5.77 0.0001 4 0.0001 0.0001 0.0256CLJFr 146.53 4.39 0.0001 5 0.0001 0.0001 0.0002 0.2032Estimate Test for hypothesis Probability Standard error ofPARAMETER of slope Parameter=0 forT-value estimate of slopeTS*ST BALB 4.81 0.69 0.4906 6.96C57 3.60 0.58 0.5635 6.22NJ 9.43 1.40 0.1630 6.72A’W 10.15 1.55 0.1248 6.56CLJFr 0.00Table 15. Analysis of covariance of the position of the rightmaxillary prominence relative to the end of the nasal tin on tailsomite stage (TS) of the five strains (ST). The group least squaremeans are significantly different between noncleft lip and cleft lip strains,between NJ and A/WySn, and between NJ and CL/Fr (low and high cleft lipfrequency strains). Regression coefficients are not significantly different bysetting to zero effect of the CL/Fr.94maxillary prominence in cleft lip strains can be used to compare with the noncleftlip strains. The deficiency in the high cleft lip frequency group compared with thelow cleft lip frequency groups also suggests that the forward growth of themaxillary prominence is significant to the etiology of cleft lip.The linear regressions of primary palate area on maxillary prominencegrowth in noncleft and cleft lip strains were compared with analysis of covariance.Results show that regression lines are parallel among five strains. The groupmean of noncleft lip (BALB/cByJ) strain is significantly larger than the groupmeans of cleft lip strains, and the group mean of low cleft lip frequency strain (NJ)is larger than those of high cleft lip frequency strains (NWySn and CL/Fr) (Fig. 21,Table 16). Mesenchymal area is also subjected to the analysis of covariance onmaxillary prominence and results show the same differences of group means ofmesenchymal areas as those of primary palate areas in five strains (Fig. 22,Table 17).Ill. Late primary palate formation and primary choana openingAt 18 tail somites, the oronasal membrane, which is the posteriorboundary separating the primary palate from secondary palate, starts to open toform the primary choana in 80% embryos in CL/Fr. However, in C57BLJ6J, in only12.5% of embryos at 18 tail somites does the primary choana start to open. Thedifference between C57BL/6J and CL/Fr is significant (Chi square test, p = 0.015)(Table 18). At 20 tail somites, 80% of embryos in CL/Fr and 70% of embryos inC57BLJ6J have the primary choana open. Growth of the nasal fin and subsequent95y = - 8782.5 + 297.50x RA2 = 0.804 D BALB/cByJ6000040000200000y = - 7702.4 + 250.77x RA2 = 0.747 D300POSITION OF RIGHT MAXILLARY PROMINENCE (rim)• A/WySno C57BL/6J• CL/FrFig. 21. Regression lines of the right primary palate area on theposition of the right maxillary prominence relative to the end ofthe nasal fin of the five strains. The keys are in same order of lines. Theslopes of five strains are homogeneous (Table 16). The group mean ofBALB/cByJ is significantly larger than three cleft lip strains and the groupmean of NJ (low cleft lip frequency strain) is significantly larger than CL/Fr(high cleft lip frequency strain) (Table 16). Linear regressions are notsignificant in the test of lack of fit for each strain. Equations include slopes,intercepts and squares of correlation coefficients.80000 yy- 6866.9 + 233.20x RA2 = 0.828y=- 1.2579e÷4 + 245.51x- 1.1865e+4 + 237.96xRA2= 0.747RA2= 0.791C4ELiiLUI-J0>-0IC,00000 100 20096GENERAL LINEAR MODELS PROCEDUREDEPENDENT VARIABLE: RIGHT PRIMARY PALATE AREA (RPP)LEAST SQUARE MEANS (LSMEAN)Standard probabilityST RPP (J.tm2) error for of hypothesis Probability of hypothesis:LSMEAN(I)=LSMEAN(J)LSMEAN LSMEAN LSMEAN=0 IJJ 1 2 3 4 5BALB 44583 1159 0.0001 1C57 31180 1090 0.0001 2 0.0001NJ 36734 1047 0.0001 3 0.0001 0.0004A/W 34659 1129 0.0001 4 0.0001 0.0363 0.1779CL/Fr 30395 890 0.0001 5 0.0001 0.6015 0.0001 0.0024Estimate Test for hypothesis Probability Standard error ofPARAMETER of slope Parameter=0 forT-value estimate of slopeRMXP*ST BALB -1.63 -0.04 0.9664 38.59C57 7.54 0.21 0.8373 36.69NJ 12.81 0.36 0.7222 35.93NW -4.75 -0.15 0.8843 32.63CL/Fr 0.00Table 16. Analysis of covariance of growth of right primary palatearea on the position of the right maxillary prominence (RMXP)relative to the end of the nasal fin of five strains (ST). The groupleast square means are significantly different between noncleft lip and cleft lipstrains, and between NJ and CLJFr (low and high cleft lip frequency strains).Regression coefficients are not significantly different by setting to zero effect ofthe CL/Fr.97y = - 3.5474e+4 + 238.58x R’2 = 0.793 i BALBPOSITION OF RIGHT MAXILLARY PROMINENCE (tim)Fig. 22. Regression lines of right mesenchymal area on theposition of the right maxillary prominence relative to the end ofthe nasal fin of the three strains. The keys are in same order of lines.The group mean of BALB/cByJ is significantly larger than cleft lip strains (A/Jand CL/Fr) and low cleft lip frequency strain (NJ) is larger than high cleft lipfrequency strain (CL/Fr) (Table 17). The slopes are homogeneous amongthree strains (p > 0.01) (Table 17). Linear regressions are not significant in thetest of lack of fit for each strain. Equations include slopes, intercepts andsquares of correlation coefficients.y = - 3.0190e+4 + 198.OOxy = - 2.8476e÷4 + 168.16xRA2= 0.684RA2= 0.8080 NJ• CL/Frc%JEUiUizC.)zUiC,)UiI9000080000700006000050000400003000020000100000100 150 200 250 300 350 400 45098GENERAL LINEAR MODELS PROCEDUREDEPENDENT VARIABLE: RIGHT MESENCHYMAL AREA (RM)LEAST SQUARE MEANS (LSMEAN)Standard probabilityST RPP (i.tm2) error for of hypothesis Probability of hypothesis:LSMEAN(l)=LSMEAN(J)LSMEAN LSMEAN LSMEAN=0 [/J 1 2 3 4 5BALB 15945 1046 0.0001 1C57 13917 949 0.0001 2 01546NJ 12745 1196 0.0001 3 0.0437 0.4458NW 10255 1229 0.0001 4 0.0001 0.0200 0.1532CL/Fr 7687 1066 0.0001 5 0.0001 0.0001 0.0026 0.1142Estimate Test for hypothesis Probability Standard error ofPARAMETER of slope Parameter=0 forT-value estimate of slopeRMXP*ST BALB 71.60 2.02 0.0460 35.47C57 66.87 2.15 0.0338 31.10NJ 31.03 0.77 0.4414 40.16NW 36.34 1.20 0.2309 30.16CL/Fr 0.00Table 17. Analysis of covariance of growth of right mesenchymalarea on the position of the right maxillary prominence (RMXP)relative to the end of the nasal fin of five strains (ST). The leastsquare means are significantly different between noncleft lip and cleft tip -strains and between A/J and CL/Fr (low and high cleft lip frequency strains).Regression coefficients are not significantly different by setting to zero effect ofthe CL/Fr (p> 0.01).99C57BL/6J CL/FrNo of No. of primary % of primary No of No. of primary % of primaryTail Somites embryos choanae opened choanae opened embryos choanae opened choanae opened17 4 0 0 6 1 1718 8 1 *12.5 5 4 8019 8 2 25 2 2 10020 9 7 78 5 4 80Table 18. Percentage of embryos with primary choanae opened in embryos of 17-20 tail somites in C57BL/6J andCLIFr strains.* The percentage of primary choanae opened in CL/Fr is significantly greater than in C57BLJ6J at 18 tail somite stage (Chi square test. p <0.05).HCCmesenchymal replacement and enlargement are anticipated in both noncleft andcleft lip strains from 13 to 16 tail somites. The opening of the primary choanaeindicates a definite primary palate formation for CLJFr at 18 tail somites and forC578L/6J at 20 tail somites. The results showed that there is a longer interval formesenchymal ingrowth and enlargement in C57BL/6J from 13 tail somites to 20tail somites than in CL/Fr from 1 3 tail somites to 1 8 tail somites.101DISCUSSIONA. Cleft lip frequency and resorption rateI. Strain effectThe present study demonstrates that the frequency of cleft lip falls intotwo groups; AIJ is the low cleft lip frequency (4.0%) group; A/WySn and CL/Fr arein the high cleft lip frequency group (22.5% and 23.6%). The frequency ofresorption of the low cleft lip frequency strain (A/J) is higher than high cleft lipfrequency group (18.0% of A/J versus 5.3% of A/WySn and 12.2% of CL/Fr).Thus a reciprocal relationship is expressed among these three strains. A/WySnand CL/Fr both have a higher frequency of cleft lip and a lower resorption ratecompared with the lower frequency of cleft lip and higher rate of resorption of A/J.In previous studies of fetal A/J mice, the frequency of cleft lip was found tobe 8.5% (Trasler, 1960) and 7.9% (Juriloff, 1982). In studies of fetal A/WySn mice,the frequency of cleft lip was 29% (Juriloff, 1982), and 19.5% (Juriloff and Harris,1985). Studies of fetal CL/Fr mice showed frequencies of 21.2% (Juriloff andFraser, 1980), 21% (Juriloff, 1981), and 26% (Staats, 1972). There is goodagreement between previous studies and this one for A/WySn and CL/Fr forpercent CL(P). The frequency of CL(P) in A/J is known to have dropped since1980 (Trasler and Trasler, 1984). This difference in frequency of cleft lip amongthe A strains of mice may be due to a genetic maternal effect (Juriloff, 1982). Inthe study by Juriloff (1982), the cleft lip frequency in NJ was 9.9% and theresorption rate was 24.8%, while in CL/Fr the cleft lip frequency was 21.2% andthe resorption rate was 5.8%. It was suggested that maternal effects account for102the difference and that the maternal trait may be the difference in survival rate ofcleft lip fetuses.Present results show CL(P) frequencies similar to those in previousreports. The results also provide a novel model for comparing the development ofpalatal structures of higher and lower frequency cleft lip strains as well ascomparing normal and cleft lip strains of mice. The hypothesis is that the highfrequency of cleft lip in the A/WySn and CL/Fr results from deficient primarypalatal structure in these two strains compared with A/J which has low cleft lipfrequency. Another hypothesis is that the low cleft lip frequency in A/J results fromthe higher mortality of embryos which may develop cleft lip. Since these cleft lipNJ embryos may die due to the maternal effect, there are less cleft lip embryossurviving in A/J litters than in AJWySn and CL/Fr litters. These surviving cleft lipembryos in A/WySn and CL/Fr may have more severe deficiencies of the palatalstructures than NJ embryos.II. Uterine site effectIt has been found that within the A/J strain, embryos in the uterine sitenearest the ovary develop cleft lip significantly more often than embryos in otherpositions in the uterine horn (Trasler, 1960). Kalter (1 975) found that thefrequency of cleft lip was higher at both the ovarian and cervical sites andresorption was lower at the ovarian site. Juriloff (1980) has reinvestigated thisuterine site effect and shown increased cleft lip and decreased resorption at theovarian site. It is suggested that a relatively privileged area at ovarian site in both103NJ and CL/Fr allows the survival of cleft lip embryos that would have diedelsewhere (Juriloff, 1980). In this study, cleft lip frequency is significantly higher atthe ovarian site than at other sites when three cleft lip strains are analyzedtogether, but the resorption rate is not significantly different between these twosites. This pattern was present in each of the strains but the sample size limitedidentification of statistical significant differences. By analyzing three cleft lipstrains together my findings support the hypothesis that the implantation site hasan effect on frequency of cleft lip but not on resorption rate. The effects ofimplantation site on the size of primary palate and the mesenchyme will beinvestigated in the future. The hypothesis is that the embryos at ovarian sites tendto have less developed primary palates than embryos at other uterine sites.B. Tail somite stage and chronological ageThe two cleft lip strains (A/J and A/WySn) belonging to the highly inbredN- strain and one cleft lip strain (CL/Fr) which is related to the N- strain arestudied. The N-strain is derived from a cross between Cold Spring Harbor albinoand Bagg albino as long as 50 years ago (Staats, 1972; Bailey, 1978). Thepresent investigation demonstrates that at equivalent tail somite stages thechronological ages of these three cleft lip strains are similar to those ofBALB/cByJ originating from albino (white coat with pink eyes) breeding stockwithout spontaneous cleft lip. They are thought to share more alleles with the Astock than C57BLJ6J (Taylor, 1972). C57BL/6J is a very distinctive inbred strainon the basis of the average number of shared aHeles with the other four strains104(Taylor, 1972). In the present study, the chronological age of C57BL/6J isyounger (from Dl 0/17 to Dl 1/8) than those of other four strains (from Dl 1/2 toDl 1/18) at the same tail somite stage (8 to 18 tail somites). The differences inchronological age distribution can be explained on the basis of different geneticbackground.C. Delineation of phases of primary palate developmentI. Early primary palate developmentThe data from 8 to 12 tail somites show that forward growth of themaxillary prominence is retarded in CL/Fr compared with C57BLI6J. At 8 tailsomites, the maxillary prominence of C57BL/6J extends frontally past the end ofthe nasal fin and joins with the lateral nasal prominence to form the lateral wall ofnasal pit which fuses with epithelium of medial nasal prominence. In contrast, themaxillary prominence of CL/Fr is still left behind the end of the nasal fin at thesame stage. From 9 to 12 tail somites, the maxillary prominence of CL/Fr growsfrontally over the end of nasal fin, but the depth is significantly smaller than thedepth in C57BL/6J. These results confirm previous studies suggesting thatregional growth deficiency or developmental abnormality in the maxillaryprominence may be a common feature in primary palatal clefting (Reed, 1933:Johnston and Hunter, 1989, Diewert and Shiota, 1990).The results of nasal fin depth measurements are similar to the results ofReed (1933) who measured the nasal floor (fused portion) of the nasal fossa in asmall number of cleft lip embryos and their normal littermates. His results have105shown a floor length of 90 im in cleft lip mice and of 320 m in normal mice of 11days 2.5 hours. The present study of 11 day old mice has shown the length ofnasal fin, which is the fused portion, ranged from 40 to 220 itm in C57BL/6J andfrom 30 to 120 urn in CL/Fr.Analysis of covariance in the depth of nasal fin in C57BL/6J and CL/Frfrom 8 to 12 tail somites suggests that the growth rate is different (p <0.05)between these two strains and growth of the nasal fin of CL/Fr is slower than thatof C57BL/6J. On the other hand, the analysis of covariance on the position ofmaxillary prominence results in a homogeneous growth rate of two strains withthe forward growth of maxillary prominence being delayed in CL/Fr (p <0.05).The slower growth rate of the nasal fin depth in CL/Fr compared to C57BLJ6J isprobably a result of deficient nasal fin depth in several embryos of CL/Fr whichwill probably end with complete or partial cleft lip as shown in Figure 11. Thegrowth rate of the nasal fin is slower in CL/Fr compared with C57BL/6J and thesize of the maxillary prominence is smaller, however, the growth rate of themaxillary prominence is similar in the two strains (Fig. 12). Thus the delayedforward growth of maxillary prominence in CL/Fr appears to contribute partially tothe deficiency of the nasal fin. Also this result partially supports the conclusionfrom Reed (1933) that the failure of nasal prominences to fuse with each other isattributable to a retarded maxillary prominence growth.II. Primary palate development with mesenchymal component formation.a. Primary palate area formation.106From 13 to 16 tail so mites, mesenchyme forms at the epithelial nasal finin these five strains. It forms earlier in noncleft lip strains (13 tail somites) than incleft lip strains (14 to 16 tail somites). The primary palate area including bothepithelial and mesenchymal areas were analyzed to understand the generalmechanisms involved during primary palate development. Comparing theincrease of right primary palate area to tail somite number in five strains I foundthe growth rate of primary palate area of five strains to be the same (p 0.05).Group means of right primary palate areas of noncleft lip strain (BALB/cByJ) aresignificantly larger than those of the three cleft lip strains. In the three cleft lipstrains, group means of the low cleft lip frequency strain (A/J) are larger thanthose of high cleft lip frequency strains (A/WySn and CL/Fr) at this critical stage ofprimary palatogenesis.Although one major gene has been shown involved in the expression ofCL(P) in mice (Juriloff, 1986; Biddle and Fraser, 1986), there are controversiesabout the effect of the gene. The expression of this major gene could beexpressed as developmental deficiencies in different areas of the facialprominences in the cleft lip strains. Both deficient growth of the maxillaryprominence and the less divergent medial nasal prominence could lead todeficiency in contact between the lateral and medial nasal prominences (Reed,1933; Trasler, 1968). These biological traits are possible factors contributing tothe smaller primary palatal area in the cleft lip strains than noncleft strains.In addition, Millicovsky et a! (1982) and Forbes etal(1989) have shownthat lack of divergence of medial nasal prominences and the depressed activity of107surface epithelium may contribute to the higher cleft lip frequency observed inCL/Fr and A/WySn than in NJ. A/J embryos show no depressed activity of thesurface epithelium. My results showing significantly smaller internal nasal contactarea in AIWySn and CL/Fr than in A/J support the idea that NJ with lower cleft lipfrequency quantitatively has less failure of contact than A/WySn and CL/Fr withhigher cleft lip frequency. This reduced failure of contact in A/J mice compared toAA’VySn and CL/Fr may be associated with an additional depressed ability of thesurface epithelium in NWySn and CL/Fr. The maternal effect on the uterusthrough the serum of the pregnant mice results in more embryos having cleft lip inNWySn and CL/Fr and fewer embryos having cleft lip in NJ. Since the resorptionrate is higher in A/J mice than in NWySn and CL/Fr mice, the higher frequency ofCL(P) in A/WySn and CL/Fr compared with A/J may be explained by thepossibility that the resorbed embryos in A/J mice may have gone on to developcleft lip had they not been resorbed. Depressed activity of the epitheliumobserved in NWySn (Forbes eta!. 1989) and CL/Fr (Millicovsky eta!., 1982)embryos compared with A/J embryos may also contribute to the smaller internalcontact area in these strains.Both BALB/cByJ and C57BL/6J are noncleft lip strains; however, theprimary palate area is significantly smaller in C57BL/6J than in BALB/cByJ. Thegenetic background of C57BL/6J is very distinctive on the basis of the averagenumber of shared alleles with other strains including BALB/cByJ (Taylor, 1972). Inaddition, the chronological age is younger in C57BL/6J than in BALB/cByJ at asimilar tail somite number (Fig. 4). Thus, the smaller primary palate area in108C57BL/6J compared to BALB/cByJ may result from the different geneticbackground between these two inbred strains.A/WySn has a larger primary palate area than CL/Fr, although these twostrains have the same cleft lip frequency. The CLJFr strain is derived from aheterogeneous stock crossed to NJ and inbred by brother-sister mating withselection for high frequency of spontaneous cleft lip (Bornstein et al, 1970).Millicovsky et al (1982) reported that after primary fusion fails to consolidate thearea at the bottom of the nasal pit at 6 tail somites from the genital tubercle(equivalent to 10 tail somites in this study), CL/Fr embryos have a secondopportunity to fuse their primary-palate primordia at 10 tail somites (equivalent to14 tail somites in this study). As the nasal prominences continue to grow in size,in approximately two-thirds of the embryos, the medial and lateral nasalprominences gain close apposition. This contact may be preceded by isolatedbursts of epithelial activity in regions adjacent to the initial fusion area. Theauthors suggested that the process of this “secondary fusion” may facilitate thesuccessful fusion of the primary palate in CL/Fr embryos without cleft lip. Thissecondary fusion may only exist in CL/Fr and not in A/WySn. Although theprimary palate contact area is larger in A/WySn than in CL/Fr before themesenchymal component forms, some CL/Fr embryos will catch up later and forma successful primary palate through secondary fusion. Consequently thefrequency of cleft lip between CL/Fr and A/WySn is similar. Further study ofNWySn and CL/Fr embryos is needed to test this hypothesis.109b. Primary palate area formation and embryonic developmentThe principles of partial least square analysis as they apply tomorphometric and developmental studies (Bookstein, 1991) operate to optimizethe covariance available for effective statistical predictions in complex systems.For instance, developmental index is taken not as a single indicator but as alatent variable (LV) combining the tail somite number, body weight, and positionof the maxillary prominence relative to the end of the nasal fin. Because theseindicators are correlated quite strongly over samples of any range of maturities, acomposite of these scores can be expected to show a stronger pattern ofcovariance with morphometric outcomes, and thereby to underlie more preciseanalyses of primary palate area, than that available by reference to any singlemeasure (Bookstein eta!, 1985).For the relation between the three developmental indicators and the fiveprimary palate parameters, the developmental LV is dominated by maxillaryprominence position variable, especially in the cleft lip strains (with greaterweight placed on the developmental measures) (Table 11). The primary palatelatent variable is more evenly weighted across all the measures of parameters.The regression coefficient of primary palate area on latent variable in each strainis better than the regression coefficient of primary palate area on tail somitenumber, especially in the cleft lip strains (compare Fig. 16 and Fig. 17). Thus, thisgrouping simplifies the linearization of smoothed scatters between LV and theprimary palate areas because forward growth of the maxillary prominence is abetter predictor for primary palate area formation than tail somite and body110weight, especially in cleft lip strains.c. Temporal and spatial analysis of mesenchymal replacement of the epithelialseamThe mechanisms by which the mesenchyme replaces the epithelial seamremain poorly understood. The growing-through of the epithelial plate by activeproliferation of the mesenchyme has been proposed to follow fusion of medialnasal, lateral nasal and maxillary prominences (Warbrick, 1960; Trasler, 1968;Vermeij-Keers, 1972; Gaare and Langman, 1977b). Different mechanismsproposed for loss of the epithelial seam include programmed peridermal celldeath and transformation of basal epithelial cells to mesenchymal cells (Fitchettand Hay, 1989). Anderson and Matthiessen (1967) have suggested that completecleft lip will appear if mesenchymal proliferation is retarded, and cleft lip with amesenchymal bridge will appear in cases where there is less marked retardationof the mesenchymal proliferation. The results of this study of mesenchymalreplacement of the nasal fin relative to developmental age marked by tail somitestages show that mesenchymal component formation is delayed in cleft lip strainscompared with noncleft lip strains. Seventy five percent of embryos of noncleft lipstrain (BALB/cByJ) start mesenchymal replacement at 13 tail somites. Cleft lipstrains start mesenchymal replacement at 14 tail somites in 70% of the embryosof NJ, at 16 tail somites in 80% of the embryos of A/WySn and in 70% of theembryos of CL/Fr. Thus, mesenchymal replacement is delayed in the cleft lipstrains, and the replacement by mesenchyme of low clefting frequency (A/J) strain111was less retarded than the high clefting frequency strains (A/WySn and CL/Fr).This study also provides a quantitative analysis of mesenchymalcomponent formation. The analysis of covariance of the mesenchymal area from13 to 17 tail somites in BALB/cByJ and C57BLJ6J, from 14 to 17 tail somites in A/Jand from 15 to 19 tail somites in A/WySn and CL/Fr includes mesenchymalproliferation and enlargement in the five strains. Results have shown that theslopes of the five strains are homogeneous, suggesting the growth rates of themesenchymal components of noncleft and cleft lip strains are not different. Thesignificantly smaller mesenchymal area in the cleft lip strains compared to thenoncleft lip strains indicates that mesenchymal growth in cleft lip strains isretarded. There is a significantly smaller mesenchymal area in the high cleft lipfrequency strains (CL/Fr) compared with the low cleft lip frequency strain (A/J).This suggests that the formation and enlargement of the mesenchymal area is aspecific indicator for determining the cleft lip malformation in noncleft lip, low cleftlip frequency and high cleft lip frequency strains of mice.The partial least squares analysis has also been applied for comparingmesenchymal areas of noncleft and cleft lip strains. Because the tail somiteinterval is different among the five strains, normalization of indicators for latentvariables were based on all the strains to avoid hiding the delayed mesenchymalcomponent formation relative to the tail somite number. The results indicate thatthe five strains can be divided into three significantly different groups ofmesenchymal component formation, noncleft lip strains (BALB/cByJ andC57BL/6J), low cleft lip frequency strain (A/J) and high cleft lip frequency strains112(A/WySn and CL/Fr). Thus the strain effect on the cleft lip frequency is the sameas the strain effect on mesenchymal component formation. The results also showthat the regression coefficients for mesenchymal area in each strain are betterwhen latent variables, rather than tail somite number, are used as the covariable(compare Figure 18 and Figure 19).d. Analysis of the growth of maxillary prominenceThe relationship of the position of the maxillary prominence relative to thenasal fin, epithelial fusion and mesenchymal replacement is still not wellunderstood. The present results show that the forward growth of the maxillaryprominence had the same pattern as mesenchymal replacement in the strains.The maxillary prominence is in a more advanced position in a noncleft lip strainthan in cleft lip strains. Comparing the low cleft lip frequency strain (A/J) with highcleft lip frequency strains (AJWySn and CL/Fr), the former is significantly moreadvanced than the latter two. The less advanced position of the maxillaryprominence in the high cleft lip frequency strain is clearly associated with ahigher cleft lip frequency. In addition, the maternal effect on the higher survivalrate of cleft lip embryos in A/WySn and CL/Fr compared to NJ results in moreembryos having cleft lip in A/WySn and CL/Fr and fewer embryos having cleft lipin A/J. Since the resorption rate is higher in NJ mice than in CL/Fr mice, thehigher frequency of cleft lip in A/WySn and CL/Fr may be explained by thepossibility that the resorbed embryos in A/J mice may have gone on to developcleft lip had they not been resorbed. Thus, more cleft lip embryos surviving with113less advanced position of the maxillary prominences in high cleft lip frequencystrains (AIWySn and CL/Fr) than in the low cleft lip frequency strain (A/J) mayexplain the less advanced mean position of maxillary prominence in the high cleftlip frequency strains.A further question that has to be answered is whether a retardedmaxillary prominence can explain retarded primary palate formation. The primarypalate and mesenchymal area are both analyzed with the position of themaxillary prominence as a covariable for testing the strain effect. Results showthat at a certain position of maxillary prominence, the primary palate andmesenchymal areas are still significantly larger in the noncleft lip strain(BALB/cByJ) than in the cleft lip strains. However, when tail somites is used as thecovariable in analysis of covariance, the difference between the group means inthe noncleft lip strain and the cleft lip strains is only half as large. In otherwords,at a certain tail so mite number the difference between noncleft lip and cleft lipstrains can be partially explained by the retarded maxillary prominence. Otherpossible pathogeneses of these cleft lip strains include: less divergent medialnasal prominences in cleft lip strains than in noncleft lip strains (Trasler,1968;Juriloff and Trasler, 1976), and depressed ability of surface epithelium in highcleft lip strains (A/WySn and CL/Fr) but not in a low cleft lip frequency strain (NJ)(Millicovsky et al, 1982; Forbes et al, 1989).In summary, a single recessive gene involved in the expression of CL(P)may be expressed in different developmental deficiencies of facial prominencesduring primary palate formation. As shown in Figure 23, this single gene may be114Major Gene(Growth Factor / Receptor)Maxillary Prominence Position Medial Nasal Prominence Divergence4’ Noncleft lip strains 4’ Noncleft lip strains4i Cleft lip strains Cleft lip strains+ +Primary palatal and mesenchymal areas Primary palatal and mesenchymal areas4’ Noncleft lip strains 4’ Noncleft lip strainsCleft lip strains \jf Cleft lip strainsMaternal EffectMaxillary Prominence Position Epithelial Activity4’ NJ 4A/WySn and CL/Fr J1 A’WySn and CL/Fr+ +Primary palatal and mesenchymal areas Nasal FinA’J 4’ A/JNWySn and CL/Fr \IJ A/WySn and CL/FrFig. 23. Summary of hypotheses of cleft lip gene effects and maternaleffects.115related to a growth factor or a receptor of a growth factor. The damage of thenormal function affecting expression of a growth factor, a receptor or other geneproduct may lead to: retarded forward growth of the maxillary prominence, lessdivergent medial nasal prominences or deficient medial nasal prominences. Thedifference of primary palatal and mesenchymal component area between thenoncleft lip and cleft lip strains can be partially explained by the retarded forwardgrowth of the maxillary prominence and partially by the less divergent or deficientmedial nasal prominences. Another genetic factor related to the difference of cleftlip frequency in the cleft lip strains is the maternal effect (Fig. 23). In humans,maternal effects have been shown to increase the frequency of CL(P) in childrenof White mothers of mixed Black/White parentage (Khoury etal, 1983). In mice,the maternal effect through the uterine environment or through the serum of thepregnant mice affects the frequencies of cleft lip. In the high cleft lip frequencystrains, more embryos surviving with the less advanced position of the maxillaryprominence and reduced epithelial activity compared with the low cleft lipfrequency strain may be due to the maternal effect. More embryos surviving withsmaller maxillary prominences and reduced epithelial activity may explain thedeficient primary palatal and mesenchymal component areas in the high cleft lipfrequency strains (A/WySn and CL/Fr) compared with the low cleft lip frequencystrain (A/J).e. Multifactorial threshold model for primary palate developmentMy results may be considered under the threshold model for116mesenchymal formation in the different strains of mice tested (Fig. 24). Possibly,the earlier mesenchymal replacement in noncleft lip strains provides favorablegrowth of the primary palate while delayed mesenchymal replacement in the lowcleft lip strain (NJ) results in less favorable growth of primary palate with about 5% cleft lip frequency. High cleft lip frequency strains have the poorest primarypalate growth. Mesenchymal replacement of high cleft lip frequency strainsoccurs late compared with both noncleft lip and low cleft lip frequency strains,resulting in about 20 to 30% cleft lip frequency. Susceptibility to cleft lip dependson where the embryo lies in relation to the threshold. Since the mesenchymalcomponent forms later in development in cleft lip strains compared with noncleftlip strains, this suggests that the cleft lip strain’s genes put it relatively closer tothe threshold.Cleft lip is considered here as an example of a congenital malformationthat is clearly multifactorial. Development of the primary palate appears todepend on adequate growth of the maxillary prominence that must providecontact of epithelia and then set up mesenchymal replacement at the epithelialseam. The facial geometry, size of lateral nasal prominence, and activity ofsurtace epithelium are factors that must be overcome so that primary palatedevelopment can occur. These factors may act together to prevent facialprominence contact and robust mesenchymal formation. The more these factorsimpinge on the developmental process, the greater the severity of the cleft lipmalformation.A multifactorial/threshold model for cleft lip is modified from Fraser (1976)117TNONCLEFT LIPBALB?c C57>1eftect of CLICL genotypeSIWySnfcCL?FrJCLEFT LIPScale of liability (Timing of mesenchymal component formation)Fig. 24. A conventional multifactorial threshold showing that cleft lipmay occur in some strains because mesenchymal formation is relativelylate. Within strains, variation among individuals would occur due toenvironmental and stochastic effects, and the time of mesenchymal formationin some individuals would lie beyond some biological tolerable limit, athreshold, and in those individuals primary palate formation would fail.118and is illustrated in Figure 25. It postulates that the stage at which themesenchyme replaces the epithelial seam and the replacement is continuouslydistributed. In some embryos mesenchymal formation occurs relatively early andin others relatively late. A discontinuous variable (cleft lip versus normal lip) isdetermined by whether a continuous variable (stage at which mesenchymeforms) puts the embryo on one side or the other of a developmental threshold(latest stage at which mesenchyme forms). Both the distribution of the variableand the threshold can be influenced by genetics and environment.The diagram in Figure 25 illustrates the position of the distribution(relative to the threshold) as being determined primarily by the interaction ofgrowth of the maxillary prominence, facial geometry, size of lateral nasalprominence and the activity of surface epithelium. The growth of the maxillaryprominence can be influenced by other factors such as the migration of neuralcrest cells which may contribute to the forward growth of the maxillaryprominence (Noden, 1975; Le lievre and Le Douarin, 1975). Epithelialmesenchymal interaction is required for the growth of the maxillary prominence(Bailey et a!, 1988; Saber et a!, 1989). It has also been suggested that thepresence of serotonin uptake sites in epithelia and serotonin binding protein inthe underlying mesenchyme indicates that serotonin might be involved inepithelial-mesenchymal interactions (Lauder et al, 1988).III. Primary choana formationDuring vertebrate evolution, the development of primary choanae or119Ea)‘Aaa)EzFig. 25. A diagram illustrating the multifactorial nature of cleft lip. Stagesat which the mesenchymal component forms is represented by the growth ofmaxillary prominence on the one side and influenced by the face shape, sizeof lateral nasal prominence and epithelial activity on the other side. Position ofthe threshold varies with the timing of primary choana opening. See text fordetails (Modified from Fraser, 1976).size of lateralnasal prominenceprimarlj choanaopeningVCleftlipSceie o liability(Stage ot mesench.mai component formation)120internal nares signifies an important landmark in the adoption of an air-breathingexistence (Hyman, 1942; Carter,1967). A definitive primary palate is alsoestablished as a consequence of this connection (Tamarin, 1982). Later, theelevation and fusion of the palatal shelves create the nasopharyngeal canalthereby causing the internal nares to assume a more posterior position as thesecondary choanae (Hyman, 1942). Preliminary observations of primary choanaformation provide a reference for definitive primary palate formation atprogressive tail somite developmental stages. One noncleft (C57BLJ6J) and onecleft lip strain (CL/Fr) were observed. Primary choana formation takes place at 18to 20 tail somites in both strains. The number of embryos with primary choanaeformation at 1 8 tail somites is significantly higher in CL/Fr (80%) than in C57BL/6J(12.5%).Since most of the embryos in the cleft lip strain (CL/Fr) have their primarychoanae formed at 18 tail somites compared with primary choanae formation at20 tail somites in noncleft lip strain (C57BLJ6J), the earlier cavitation andcleavage of the dorsal part of the nasal fin of cleft lip strain embryos may extendto the ventral part of the nasal fin which has delayed mesenchymal replacement.This earlier cleavage of the dorsal part of the nasal fin in CL/Fr than in C57BL/6Jmay be related to the more convergent medial nasal prominences in CL/Fr thanin C57BL/6J (Trasler, 1968; Juriloff and Trasler, 1976). Streeter (1948) hasshown that in normal human embryos at stage 17, the nasal fin becomestransformed from an epithelium to an epithelial-lined passage as a result of thecoalescence of its cleavage spaces, which results in primary choana formation121(Fig. 26). Warbrick (1960) has suggested that the nasal fin may persist throughthe developmental stages and prevent the mesenchyme of maxillary and frontonasal prominences from making contact. Subsequently the dorsal part of thenasal fin undergoes the normal cavitation and cleavage to form the primarychoana. An extension of this cavitation and cleavage into an abnormallypersisting ventral part of the nasal fin would result in cleft lip formation.The cleft lip mouse strain (CL/Fr) in this study shows both earlier primarychoana formation at 18 tail somites and delayed mesenchymal componentformation at 16 tail somites after the fusion of the epithelia compared to noncleftlip strain (C57BLJ6J). Here primary choana formation takes place at 20 tailsomites and mesenchymal penetration takes place at 13 tail somites. These twofactors were addressed in the threshold model of Figure 25 as an explanation ofthe etiology of cleft lip from various genetic and environmental factors. Thenumber of embryos falling beyond the threshold also varies with the position ofthe threshold, which in turn varies with the timing of primary choana formation.Thus the position of the threshold is also depicted as a continuous variable. Insummary, the primary choana forms at a certain stage, so that if mesenchymalcomponent formation is delayed by more than a certain critical amount, themesenchymal component will form too late to accomplish normal primary palateformation, and a cleft lip will result. For example, in CL/Fr, after primary andsecondary fusion, certain embryos have no mesenchymal component formationby 16 tail somites which may fail to allow a primary palate formation at 18 tailsomites when the opening of primary choana starts.122Stage 16 Early Stage 17 Late Stage 17MX ILLAOLFACT. PIT NASAL FINEarly Stage 18 Late stage 18Fig. 26. Outlines of sagittal sections showing the steps in the formationof the primary palate in human embryo stage 16, 17 and 18. The platelikenasal fin is the epithelium which by characteristic splitting phenomenonproduces the primary choanae (Modified From Streeter, 1948).NASAL FINNESENCHVMkL COMPONENT SPLITTiNGNASALWING123CONCLUSIONSThe present study of primary palate formation in mice with genetic cleft lipand mice without cleft lip at specific chronological age and tail somite numberprovides base line data for primary palatogenesis. In noncleft lip strains,BALB/cByJ has similar chronological age to three cleft lip strains and isconsidered an appropriate control for studying the cleft lip malformation. AlsoBALB/cByJ shares many alleles with the cleft lip strains except for the cleft lipgenes. C57BL/6J is another noncleft lip strain and shows younger chronologicalage than both BALB/cByJ and Al- strains at the stage of primary palatedevelopment, which may make this tissue relatively “younger” than in otherstrains. Future work studying normal and abnormal primary palate formationshould use BALBIcByJ as the primary control.In each cleft lip strain, cleft lip frequencies and resorption rates arereciprocally related, and the difference between low cleft lip frequency strain (AIJ)and high cleft lip frequency strains (A/WySn and CL/Fr) are significant. It isbelieved that maternal effects cause the differences in cleft lip frequencies andresorption rates for different cleft lip strains, and this maternal trait may be due toa difference in survival rate of cleft lip fetuses (Davidson et al. 1969; Juriloff andFraser, 1980). It has also been shown that this maternal effect on cleft lip ismediated through uterine environment (Bornstein et al. 1970). We expect that thismaternal effect on the cleft lip frequency and resorption rate affects the primarypalatal structures in the form of general developmental delays and results in moredeficient palatal structures in high cleft lip frequency strains (A/WySn and CL/Fr)1 24than in the low cleft lip frequency strain (NJ).The ovarian site of the uterine horn has a higher cleft lip frequency thanother sites within the uterus, but resorption rate is not influenced by the site withinthe uterus. It has been suggested that cleft lip embryos have higher survival rateat the ovarian site. Consequently, the resorption rate is lower and the cleft lipfrequency is higher at the ovarian site than at other sites of the uterus (Juriloff andFraser, 1980). My results show higher cleft lip frequency at the ovarian sitecompared to other sites but no difference in resorption frequency. Further studieslooking at internal development of the primary palate in embryos taken from theovarian and other sites are required to determine the morphological effects theuterine site has on development of the primary palate.During early stages of primary palate development, forward growth of themaxillary prominence is delayed in CL/Fr compared to C57BLI6J. The possiblemechanism is either delayed neural crest migration after neural tube formation, orreduced neural crest cell proliferation at the stage of induction of the nasalplacode. Formation of the nasal fin represents the early fusion of facialprominences. The slower growth rate of the nasal fin in CL/Fr mice compared toC57BLJ6J may be associated with delayed forward growth of the maxillaryprominence and other factors such as a less divergent medial nasal prominences(Trasler, 1968). These etiologic factors may represent the biological traits of theexpression of the same single gene which causes cleft lip in mice.A further analysis of fused primary palates was undertaken at a stageduring mesenchymal replacement. Noncleft lip strain mice had a larger total1 25primary palate area than cleft lip strains due to their genotype. Within noncleft lipstrains, larger primary palate areas in BALB/cByJ than C57BL/6J are associatedwith different genetic backgrounds and may be affected by the youngerchronological age of C57BL/6J at similar tail somite stages. The strain of low cleftlip frequency (NJ) has a larger primary palate area than strains of high cleft lipfrequency (A/WySn and CL/Fr) which may be associated with depressed ability ofthe surface epithelium (Millicovsky eta!, 1982; Forbes eta!, 1989). The maternaleffect on the uterine environment of the pregnant mice results in more embryoshaving cleft lip in A/WySn and CL/Fr compared to less embryos having cleft lip inNJ. Thus there are many embryos in NWySn and CL/Fr with depressed activity ofepithelium which results in smaller internal contact area in A/WySn and CL/Frcompared with NJ strain.The primary palate area is also larger in A/WySn than in CL/Fr.Nevertheless they still have the same cleft lip frequency. A possible explanationis that a large portion of CL/Fr embryos have secondary fusion as reported byMillicovsky et a! (1982). CL/Fr embryos with smaller fused areas may representprimary fusion with or without secondary fusion, compared to A/WySn which maynot have secondary fusion. Eventually, these CL/Fr embryos with secondaryfusion will have successful primary palate formation, and the cleft lip frequencywill end up the same between NWySn and CL/Fr.Partial least squares analysis was applied to determine the best predictorfor primary palate development. In this study, forward growth of the maxillaryprominence was shown to be a better predictor for primary palate area formation126than tail somite number and body weight, especially in cleft lip strains. Thedelayed formation of primary palate area can be partially attributed to the delayedforward growth of the maxillary prominence.Both the qualitative and quantitative results of mesenchymal componentformation in the five strains can be divided into noncleft lip, low cleft lip and highcleft lip frequency groups as the result of the cleft lip frequency study. Thus, fromeither the time of formation or the size of the mesenchymal component, we canpredict if the embryo belongs to a noncleft lip strain, a low or a high cleft lipfrequency strain. The partial least squares analysis also shows that the position ofthe maxillary prominence is a better predictor than tail somite for mesenchymalcomponent formation.The forward growth of the maxillary prominence in the five strains studiedcan also be divided into noncleft lip, low and high cleft lip frequency groups whichare similar to the results of the primary palatal and mesenchymal componentareas study. Hence, growth of the primary palatal area, mesenchymal componentand position of the maxillary prominence are specific indicators for primary palatedevelopment. Comparison of the primary palatal and mesenchymal areas at thesame position of the maxillary prominence showed that both primary palatal andmesenchymal areas are still significantly different among noncleft, low and highcleft lip frequency groups, although these differences are reduced to half thedifferences observed in analysis of covariance with tail somite number. Delayedforward growth of the maxillary prominence may be partly associated withdelayed primary palatal and mesenchymal formation in cleft lip strains. Other1 27etiologic factors such as less divergent medial nasal prominences may contributeto the delayed primary palatal and mesenchymal formation as well (Trasler, 1968;Juriloff and Trasler, 1976; Millicovsky etal, 1982).The maternal effect on the higher survival rate of cleft lip embryos inA/WySn and CL/Fr compared with A/J may explain the less advanced position ofthe maxillary prominence in the high cleft up frequency strains. This lessadvanced position of the maxillary prominence may partially contribute to thedeficient primary palatal and mesenchymal areas in the AJWySn and CL/Frstrains compared with the NJ strain. Depressed activity of the surface epitheliummay also contribute to the reduced size of the primary palatal and themesenchymal bridge areas in A/WySn and CL/Fr than in A/J strain ( Millicovsky etal, 1982; Forbes etal, 1989).A multifactorial threshold model is suggested from this study. The stageof mesenchymal replacement is applied as a scale of liability for the cleft lipmalformation. Genotypes of noncleft lip or cleft lip determine the continuousdistribution toward the favorable growth of the primary palate, which is away fromthe threshold, or unfavorable growth of primary palate which is closer to thethreshold. Both the distribution of the variable and threshold can be influenced bygenetic and environmental factors. Evaluation of the hypothesis that the positionof the continuous distribution is determined by the growth of the maxillaryprominence will require further study of the relation between the growth of themaxillary prominence, epithelial seam replacement and the enlargement of themesenchymal component. Future studies should focus on whether delayed128growth of the maxillary prominence is caused by delayed formation, migration, orproliferation of neural crest and/or insufficient epithelial-mesenchymalinteractions. Also, it should be confirmed whether this delay in maxillaryprominence growth goes on to cause insufficient mesenchymal replacementleading to various degrees of cleft lip severity.1 29BIBLIOGRAPHYAnderson, H., and Matthiessen, M.E. (1967): Histochemistry of the earlydevelopment of the human central face and nasal cavity with special reference tothe movement and fusion of the palatine processes. Acta. Anat. (Basel), 68:473-508.Ardinger, H.H., Buetow, K.H., Bell, G.l., Bardach, J., VanDemark, D.R., Murray,J.C. (1989): Association of genetic variation of the transforming growth factoralpha gene with cleft lip and palate. Am. J. Hum. Genet. 45:348-353.Bailey, L.J., Minkoff, R., and W.E. Koch. (1988): Relative growth rates of maxillarymesenchyme in the chick embryo. J. Craniofac. Genet. Dev. Biol. 8:167-177.Bailey, D.W. (1978): Sources of subline divergence and their relative importancefor subline of six major inbred strains of mice. In: Morse H.C. lii (ed) “Origins ofInbred Mice.” Academy Press, New York, pp 197-215.Bear, J.C. (1976): A genetic study of facial clefting in Northern England. Clin.Genet. 9:277-284.Bear J.C. (1978): Spontaneous abortion, sex ratio and facial cleft malformations.Clin. Genet. 13:1-7.Bernanke, D.H. and Markwald, R.R. (1979): Effects of hyaluronic acid on cardiaccushion cells in collagen matrix cultures. Tex. Rep. Biol. Med. 39: 271 -285.Bernanke, D.H. and Markwald, R.R. (1982): Migratory behavior of cardiac cushioncells in collagen-lattice culture system. Dev. Biol. 91: 235-245.Biddle, F.G., and Fraser, F.C. (1986): Major gene determination of liability tospontaneous cleft lip in the mouse. J. Craniofac. Genet. Dev. Biol. [suppl] 2:67-88.Bingle, G.J. and Niswander, J.D. (1977): Maternal effects in human cleft lip andpalate. Am. J. Hum. Genet., 29: 605-609.Bixier, D., Fogh-Andersen, P., and Conneally, P.M. (1971): Incidence of cleft lipand palate in the offspring of cleft parents. Clinical Genetics 2:155-1 59.Bonner,J.J. and Slavkin, H.C. (1975): Cleft palate susceptibility linked tohistocompatibility-2 (H-2) in the mouse. Immunogenetics 2: 21 3-21 8.Bonner, J.J., Terasaki, P.1., Thompson, P., Holve, L.M., Wilson, L., Ebbin, A.J. andSlavkin, H.C. (1978): HLA phenotype frequencies in individuals with cleft lip130and/or cleft palate. Tissue Antigens 12: 228-232.Bookstein, F.L. (1982): The geometric meaning of soft modeling, with somegeneralizations. In: Systems under indirect observation: Causality, structure,prediction. H. Wold and K. Joreskog (Eds.) North-Holland, Amsterdam. Vol. II, pp55-74.Bookstein, F.L. (1986): The elements of latent variable models: A cautionarylecture. In: Brown, A.L. and Rogoff, B. (Eds.) “Advances in developmentalpsychology.” Vol. 4, pp 203-230.Bookstein, F.L. (1991): Morphometric Tools for Landmark Data. Book manuscriptaccepted for publication, Cambridge University Press.Bookstein, F.L., Chernoff, B., Elder, R., Humphries, J., Smith, G. and Strauss, R.(1985): Morphometrics in evolutionary biology. (Special publication 15). TheAcademy of Natural Sciences, Philadelphia.Bornstein, S., Trasler, D.G. and Fraser, F.C. (1970): Effect of the uterineenvironment on the frequency of spontaneous cleft lip in CL/FR mice. Teratology3: 295-298.Bronsky, P.T., Johnston, M.C. and Sulik, K.K. (1986): Morphogenesis of hypoxia-induced cleft lip in CL/Fr mice. J. Craniofac. Genet. Dev. Biol. (Suppl.) 2:113-128.Brown, K.S., Murphy, P.F. and Johnston, M.C. (1974): Frequency of CLP in NJfetuses after maternal thyroxin injection. J. Dent. Res. 53:1 90.Burk, D., Sadler, T.W., and Langman, J. (1979): Distribution of surface coatmaterial on nasal folds of mouse embryos as demonstrated by concanavalin Abinding. Anat. Rec. 193:185-196.Carter, C.O. (1969): Genetics of common disorders. Br. Med. Bull. 25: 52-57.Carter, C.O. (1976): Genetics of common single malformations. Br. Med. Bull.32:2 1 -26.Carter, G.S. (1967): Structure and Habit in Vertebrate Evolution. Univ.Washington Press, Seattle, pp 122-166.Chaube, S. and Murphy, M.L. (1963): Teratogenic effect of hadacidin (a newgrowth inhibitory chemicals) on the rat fetus. J. Exp. Zool. 152:67-73.131Chenevix-Trench, G. Jones, K., Green, A. and Martin, N. (1991): Further evidencefor an association between genetic variation in transforming growth factor alphaand cleft lip and palate. Am. J. Hum. Genet. 48:1012-1 01 3.Ching, G.H.S. and Chung, C.S. (1974): A genetic study of cleft lip and palate inHawaii. I. Interracial crosses. Am. J. Hum. Gent. 26:162-176.Chung, C.S. and Myrianthopoulos, N.C. (1968): Racial and prenatal factors inmajor congenital malformations. Am.. J. Hum. Genet. 20:44-60.Chung, C.S., Ching, G.H.S., and Morton, N.E. (1974): A genetic study of cleft lipand palate in Hawaii. II. Complex segregation analysis and genetic risks. Am. J.Hum. Genet. 26:177-188.Chung, C.S., Bixler, D., Watanabe, T., Koguchi, H., and Fogh-Andersen, P.(1986): Segregation analysis of cleft lip with or without cleft palate: a comparisonof Danish and Japanese data. Am. J. Hum. Genet. 39 :603-61 1.Ciriani, D. and Diewert, V.M. (1986): A comparative study of development duringprimary palate formation in A/WySn, C57BLJ6J, and their Fl crosses. J.Craniofac. Genet. Dev. Biol. 6: 369-377.Coccaro, P.J., D’Amico, R., Chavoor, A. (1972): Craniofacial morphology ofparents with or without cleft lip and palate children. Cleft Palate J. 9:23-28.Cody, R.P. and Smith, J.K. (1987): Applied Statistics and the SAS ProgrammingLanguage, 2nd ed. North-Holland, New York, pp 107-1 29.Czeizel, A. and Tusnady G. (1972): A family study on cleft lip with or without cleftpalate and posterior cleft palate in Hungary. Hum. Hered. 22:405-416.Davidson, J.G., Fraser, F.C. and Schiager G. (1969): A maternal effect on thefrequency of spontaneous cleft lip in the A/J mouse. Teratology 2: 371 -376.Dietrich, L.S., Friedland, l.M., and Kaplan, L.A. (1958): Pyridine nucleotidemetabolism: Mechanism of action of the niacin antagonist, 6-aminonicotinamide.J. Biol. Chem. 233:964-968.Diewert, V.M. and Shiota, K. (1990): Morphological observations in normalprimary palate and cleft lip embryos in the Kyoto collection. Teratology 41:663-667.Diewert, V.M. and Lozanoff,S. (1989): A frontal plane morphometric analysis ofhuman embryonic craniofacial growth during primary palate formation. Anat. Rec.132223 :33A.Diewert, V.M. and Lozanoff, S. (1990): Morphogenesis of the human midfaceduring primary palate formation. J. Dent. Res. 69:1 59.Diewert, V.M. and Van der Meer, D. (1991): Growth of the human primary palate.J. Dent. Res. 69:1 56.Diewert, V.M., Wang, K. and Lozanoff, S. (1989): Growth of the frontonasalprocess in human embryos. J. Dent. Res. 68:247.Erickson, J.D. (1974): Facial and oral form in sibs of children with cleft lip with orwithout cleft palate. Ann. Hum. Genet. 38:77-88.Eto, K., Figueroa, A., Tamura, G. and Pratt, R.M. (1981): Induction of cleft lip incultured rat embryos by localized administration of tunicamycin. J. Embryol. Exp.Morph., 64: 1-9.Falconer, D.S. (1965): The inheritance of liability to certain diseases, estimatedfrom the incidence among relatives. Ann. Hum. Genet. 29:51-71.Figueroa, A. and Pratt, R.M. (1979): Autoradiographic study of macromolecularsynthesis in the fusion epithelium of the developing rat primary palate in vitro. J.Embryol. Exp. Morph. 50:1 45-1 54.Fitchett, J.E. and Hay, E.D. (1989): Medial edge epithelium transforms tomesenchyme after embryonic palatal shelves fuse. Dev. Biol. 131: 455-474.Flint, O.P., and Ede, D.A. (1978): Facial development in the mouse; a comparisonbetween normal and mutant (amputated) mouse embryos. J. Embryol. Exp.Morph. 48:249-267.Forbes, D.P., Steffek, A.J. and Klepacki, M. (1989): Reduced epithelial surfaceactivity in related to a higher incidence of facial clefting in A/WySn mice. J.Craniofac. Genet. Dev. Bid. 9:271 -283.Franke, W.W., Grund, C., Kuhn, C., Jackson, B.W., and lilmensee, K. (1982):Formation of cytoskeletal elements during mouse embryogenesis. Ill. Primarymesenchymal cells and the first appearance of vimentin filaments. Differentiation23: 43-59.Fraser, F.C. (1970): The genetics of cleft lip and cleft palate. Am. J. Hum. Genet.22: 336-352.133Fraser, F.C. (1976): The multifactorial/threshold concept - uses and misuses.Teratology 14 267-280.Fraser, F.C. (1980): The genetics of cleft lip and palate: Yet another look. In: PrattR.M., Christiansen R.L. (eds) Current trends in prenatal craniofacial development.North-Holland, New York, pp 357-366.Fraser, F.C. (1989): Invited editorial: Mapping the cleft lip genes: The first fix? Am.J. Hum. Genet. 45: 345-347.Fraser, F.C. and Baxter, H. (1954): The familial distribution of congenital clefts ofthe lip and palate. Am. J. Surg. 87:656-659.Fraser, F.C. and Pashayan, H. (1970): Relation of face shape to susceptibility tocongenital cleft lip: a preliminary report. J. Med. Genet. 7:112-117.Freund, R.L. and Littell R.C. (1981): “SAS for Linear Models: A Guide to theANOVA and GLM Procedures.” SAS Institute Inc. Cary, North Carolina, pp. 187-205.Gaare, J.D., and Langman, J. (1977a): Fusion of nasal swellings in the mouseembryo : Surface coat and initial contact. Am. J. Anat. 150:461 -476.Gaare, J.D., and Langman, J. (1977b): Fusion of nasal swellings in the mouseembryo : Regression of the nasal fin. Am. J. Anat. 150:477-500.Goldman, A.S., Fishman, C.L., and Baker, M.K. (1983): Phenytoin teratogenicity inthe primary and secondary mouse embryonic palate is influenced by the H-2histocompatibility locus (41613). Proc. Soc. Exp. Biol. Med. 173: 82-86.Green, E.L. (1971): The threshold model: Uses and Limitations. In Moyers, R.E.and Krogman, W.M. (Eds): “Craniofacial Growth in Man” Pergamon Press Oxford& New York, pp 143-1 53.Greenburg, G. and Hay, E.D. (1982): Epithelia suspended in collagen gels canlose polarity and express characteristics of migrating mesenchymal cells. J. CellBiol. 95: 333-339.Greenburg, G. and Hay, E.D. (1986): Cytodifferentiation and tissue phenotypechange during transformation of embryonic lens epithelium to mesenchyme-likecells in vitro. Dev. Biol. 115: 363-379.Greenburg, G. and Hay, E.D. (1988): Cytoskeleton and thyroglobulin expressionchange during transformation of thyroid epithelium to mesenchyme-like cells.134Development 102: 605-622.Greene, R.M. and Pratt, R.M. (1976): Developmental aspects of secondary palateformation. J. Embryol. Exp. Morphol. 36: 225-245.Greig, R.G. and Jones, M.N. (1977): Mechanisms of intercellular adhesion.BioSystems 9: 43-55.Gruneberg, H. (1951): The genetics of a tooth defect in the mouse. Proc. Roy.Soc. B, 138:437-451.Gruneberg, H. (1952): Genetics studies on the skeleton of the mouse. IV. Quasi-continuous variations. J. Genet. 51:95-114.Hanson, J.W. and Smith D.W. (1975): The fetal hydantoin syndrome. J. Pediatric.87:285-290.Hay, E.D. (1968): Organization and fine structure of epithelium and mesenchymein the developing chick embryo. In: Treistad, R.L. (ed) “The Role of ExtracellularMatrix Development.” A.R. Liss, New York. pp 1-31.Hecht, J.T., Wang, Y., Blanton, S.H., Daiger, S.P. and Michels, V.V. (1990):Nonsyndromic cleft lip with or without cleft palate: no evidence of linkage totransforming growth factor alpha. Am. J. Hum. Genet. 47:A220.Hicks, H.E., Johnston, M.C., and Banes, A.J. (1983): Maternal phenytoinadministration affects DNA and protein synthesis in embryonic primary palate.Teratology 28: 389-397.Hinrichsen, K. (1985): The early development of morphology and patterns of theface in the human embryo. Adv. Anat. Embryol. cell biol. 98:1 -76.Hyman, L.H. (1942): “Comparative Vertebrate Anatomy” (2nd ed) Univ. ChicagoPress, Chicago, pp 255.Jacobson, A.G. (1963a): The determination and positioning of the nose, lens andear. I. Interactions within the ectoderm, and between the ectoderm andunderlying tissues. J. Exp. Zool. 154:273-283.Jacobson, A.G. (1963b): The determination and positioning of the nose, lens andear. II The role of the endoderm. J. Exp. Zool. 154:283-291.Johnston, M.C. (1964): Facial malformations in chick embryos resulting fromremoval of neural crest. J. Dent. Res. 43:822 (abstract).135Johnston, M.C. (1966): A radiographic study of the migration and fate of thecraniofacial neural crest cells in the chick embryo. Anat. Rec. 156:1 43-1 55.Johnston, M.C. and Hunter, W.S. (1989): Cleft lip and/or palate in twins : evidencefor two major cleft lip groups. Teratology 39:461.Johnston, M.C., Sulik, K.K. and Dudly, K.H. (1978): Diphenylhydantoin (DPH,Dilantin)- induced cleft lip and palate (CLP) in A/J mice. J. Dent. Res. 57A:326.Juriloff, D.M. (1980): The genetics of clefting in the mouse. Prog. Clin. Blot. Res.46: 39-79.Juriloff, D.M. (1981): Thyroxin-induced differential mortality of mouse embryoswith cleft lip. Dev. Pharmacol. Ther. 2: 17-31.Juriloff, D.M. (1982): Differences in frequency of cleft lip among the A strains ofmice. Teratology 25:361 -368.Juriloff, D.M. (1986): Major genes that causes cleft tip in mice: progress in theconstruction of a congenic strain and in linkage mapping. J Craniofac. Genet.Dev. Biol. [suppl] 2:55-66.Juriloff, D.M. and Trasler, D.G. (1976): Test of the hypothesis that embryonic faceshape is a causal factor in genetic predisposition to cleft lip in mice. Teratology14: 35-42.Juriloff, D.M. and Fraser, F.C. (1980): Genetic maternal effects on cleft tipfrequency in A/J and CL/Fr mice. Teratology 21: 167-1 75.Juriloff, D.M. and Harris, M.J. (1985): Thyroxin-induced differential mortality ofcleft lip mouse embryos: Dose- and time-response studies of the A/WySn strain.Teratology 31: 319-329.Kalter, H. (1975): Prenatal epidemiology of spontaneous cleft lip and palate, openeyelid, and embryonic death in A/J mice. Teratology 12:245-258.Khoury, M., Erickson, J.D. and James, L.M. (1983): Maternal factors in cleft lip withor without palate: evidence from interracial crosses in the United States.Teratology 27:35 1 -357.Kobayashi, Y. (1958): A genetic study of harelip and cleft palate. Jap. J. Hum.Ge net. 3:73-107.136Lauder, J.M., Tannir, H., and Sadler, T.H. (1988): Serotonin and morphogenesis.I. Sites of serotonin uptake and binding protein immunoreactivity in the mid-gestation mouse embryo. Development 102: 709-720.Lejour, M., (1969): Cleft lip induced in the rat. Cleft Palate J. 17:1 69-1 86.Lejour-Jeanty, M. (1966): Becs-de-lièvre provoqués chez le rat par un dérivé deIa pénicilline, l’hadacidine. J. Embryol. Exp. Morph. 15:193-211.Le Lievre, C. and Le Douarin, N.M. (1975): Mesenchymal derivatives of theneural crest: Analysis of chimeric quail and chick embryos. J. Embryol. Exp.Morphol. 34: 125-154.Lowry, R.B. and Renwick, D.H.G. (1969): Incidence of cleft lip and cleft palate inBritish Columbia Indians. J. Med. Genet. 6:67-69.Mackler, B., Grace, R., Tippit, D.F., Lemire, R.J., Shepard, T.H., and Kelley, V.C.(1975): Studies of the development of congenital anomalies in rats : Ill. Effects ofinhibition of mitochondrial energy systems on embryonic development.Teratology 12:291-296.Marazita, M.L., Spence, A.M., Melnick, M. (1986): Major gene determination ofliability to cleft lip with or without cleft palate : a multiracial view. J. Craniofac.Gent. Dev. Biol. [Suppl] 2:89-97.Martz, F., Failinger III, C. and Blake, D.A. (1977): Phenytoin teratogenesis:Correlation between embryopathic and covalent binding of putative arene oxidemetabolite in gestational tissue. J. Pharmacol. Exp. Ther. 203: 231 -239.Massey, K.M. (1966): Teratogenic effects of diphenylhydantoin sodium. J. OralTher. Pharmacol. 2:380-385.Meadow, S.R. (1968): Anticoagulant drugs and congenital abnormalities. Lancet2:1296.Melnick, M., Bixler, D., Fogh-Andersen P., Conneally, P.M. (1980): Cleft lip with orwithout cleft palate: An overview of the literature and an analysis of Danish casesborn between 1941 and 1968. Am. J. Med. Genet. 6:83-97.Millicovsky, G., and Johnston, M.C. (1981): Active role of embryonic facialepithelium : New evidence of cellular events in morphogenesis. J. Embryol. Exp.Morph. 63:52-66.137Millicovsky, G., Ambrose, L.J.H., and Johnston, M.C. (1982): Developmentalalterations associated with spontaneous cleft lip and palate in CL/FR mice. Am. J.Anat.164: 29-44.Minkoff, R. and Kuntz, A.J. (1977): Cell proliferation during morphogeneticchange; analysis of frontonasal morphogenesis in the chick embryo employingDNA labeling indices. J. Embryol. Exp. Morph. 40:101-113.Minkoff, R. and Kuntz, A.J. (1978): Cell proliferation and cell density ofmesenchyme in the maxillary process and adjacent regions during facialdevelopment in the chick embryo. J. Embryol. Exp. Morph. 46: 65-74.Mosteller, F. and Youtz, C. (1961): Tables of the Freeman-Tukey transformationsfor the binomial and Poisson distributions. Biometrika 48: 433-440.Murphy, M.L, Dagg, C.P., and Kamofsky, D.A. (1957): Comparisons of teratogenicchemicals in the rat and chick embryos. Pediatrics 19:701-71 4.Neel, J.V. (1958): A study of major congenital defects in Japanese infants. Am. J.Hum. Genet. 10:398-445.Nelson, M.M. (1957): Production of congenital anomalies in mammals bymaternal dietary deficiencies. Pediatrics 19:764-776.Nichols, D.H. (1981): Neural crest formation in the head of the mouse embryousing a new histological technique. J. Embryol. Exp. Morphol. 64: 105-1 20.Nichols,D.H. (1986): Formation and distribution of neural crest mesenchyme tothe first pharyngeal arch region of the mouse embryo. Am. J. Anat. 176: 221-231.Noden, D.M. (1975): An analysis of the migratory behavior of avian cephalicneural crests cells. Dev. Biol. 42:106-130.Noden, D.M. (1983): The role of the neural crest in patterning of avian cranialskeletal, connective and muscle tissue. Dev. Biol. 96:1 44-1 65.Ohbayashi, N. and Eto, K. (1986): Relative contributions of the facial processes tofacial development: a microsurgical assay. J. Carniofac. Genet. Dev. Biol. [Suppi]2:41-44.Oppenheimer, S.B. (1973): Utilization of L-glutamine in intercellular adhesion:Ascites tumor and embryonic cells. Exp. Cell Res. 77: 175-1 82.138O’Rahilly, R. and MUller, F. (1987): Developmental Stages in Human Embryos.Carnegie Inst. Wash. PubI. 637, Carnegie Institute of Washington, pp 1 75-302.Patten, B.M. (1961): The normal development of the facial region. In: Pruzansky,S. (ed) “Congenital anomalies of the face and associated structures.” Springfield,Ill. Thomas. p. 11-45.Patterson, S.B. and Minkoff, R. (1985): Morphometric and autoradiographicanalysis of frontonasal development in the chick embryo. Anat. Rec. 212: 90-99.Patterson, S.B., Johnston, MC. and Minkoff, R. (1984 ): An implant labelingtechnique employing sable hair probes as carriers for 3H-thymidine: Applicationsto the study of facial morphogenesis. Anat. Rec. 210: 525-536.Pessac, B. and Defendi, V. (1972): Cell aggregation: Role of acidmucopolysaccharides. Science 175: 898-900.Pinsky, L., and Fraser, F.C. (1959): Production of skeletal malformations in theoffspring of pregnant mice treated with 6-aminonicotinamide. Biol. Neonat. 1:106-112,Pinsky, L., and Fraser, F.C. (1960): Congenital malformations after a two-hourinactivation of nicotinamide in pregnant mice. Brit. Med. J. 2:195-1 97.Reed, S.C. (1933): An embryological study of harelip in mice. Anat. Rec.56:101-110.Reed, S.C. (1936): Harelip in the house mouse. II. Mendelian units concernedwith harelip and application of the data to the human harelip problem. Genetics21:361-374.Roseman, S. (1974): Complex carbohydrates and intercellular adhesion. In: Lee,E.Y.C. and Smith, E.E. (eds) “Biology and Chemistry of Eukaryotic Cell Surfaces.”Academic Press, New York, pp. 31 7-354.Saber, G.M., Parker, S.B., and Minkoff, R. (1989): Influence of epithelialmesenchymal interaction on the viability of facial mesenchyme in vitro. Anat. Rec.225:56-66.Sampson, P.D., Streissguth, A.P., Barr, H.M. and Bookstein, F.L. (1989):Neurobehavioral effects of prenatal alcohol: Part II. partial least squares analysis.Neurotoxicol. Teratol. 11: 461-476.139Shigeura, H.T. and Gordon, C.N. (1962a): Hadacidin, a new inhibitor of purinebiosynthesis. J. Biol. Chem. 237: 1932-1 936.Shigeura, H.T. and Gordon, C.N. (1962b): The mechanism of action of Hadacidin.J. Biol. Chem. 237: 1937-1 940.Snell, G.D. Fekete, E. Hummel, K.P. and Law, L.W. (1940): The relation of mating,ovulation and the estrus smear in the house mouse to time of day. Anat. Rec.76:39-54.Sokal, R.R., and Rohif, F.J. (1969): Biometry. Freeman, San Francisco. pp 404-493.Staats, J. (1972): Standardized nomenclature for inbred strains of mice: fifthlisting. Cancer Res. 32:1609-1646.Stark, R.B. (1954): The pathogenesis of harelip and cleft palate. Plast.Reconstruct. Surg. 13:20-32.Steel, R.G.D., and Torrie, J.H. (1980): Principles and procedures of statistics.McGraw-Hill New York. pp 401 -437.Stein, Z., Susser, M., Warburton, D. Wittes, J. and Kline, J. (1975): Spontaneousabortion as a screening device. Amer. J. Epidemiol. 102: 275-290.Streeter, G.L. (1948): Developmental horizons in human embryos. Contr.Embryol. Carneg. lnstn. 32: 133-204.Strickler, S.M., Dansky, L.V., Miller, M.A., Seni, M.H., Andermann, E., andSpielberg, S.P. (1985): Genetic predisposition to phenytoin-induced birth defects.Lancet 2:746-749.Sulik, K.K., Johnston, M.C., Ambrose, L.J.H., and Dorgad, D. (1979): Phenytoin(Dilantin)-induced cleft lip and palate in NJ mice: A scanning and transmissionelectron microscopic study. Anat. Rec. 195:243-256.Tamarin, A. (1982): The formation of the primitive choanae at the junction of theprimary and secondary palates in the mouse. Am. J. Anat. 165: 31 9-337.Taylor, B.A. (1972): Genetic relationship between inbred strains of mice. J. ofHeredity 63: 83-86.Tom, C., Juriloff, D.M., and Harris, M.J. (1991): Studies of the effect of retinoicacid on anterior neural tube closure in mice genetically liable to exencephaly.140Teratology 43:27-40.Töndury, G. (1950): Zum Problem der Gesichtsentwicklung und der Genese derHasenscharte. Acta Anat. (Base I) 11: 300-328.Tondury, G. (1964): Embryology of clefts. In: Hotz, R. (ed.) “Early treatment of cleftlip and palate.” Huber, Stuttgart. pp 17-24.Trasler, D.G. (1960): Influence of uterine site on occurrence of spontaneousclefting in mice. Science 132:420-421.Trasler, D.G. (1968): Pathogenesis of cleft lip and its relation to embryonic faceshape in NJ and C57BL mice. Teratology 1:33-50.Trasler, D.G. and Fraser, F.C. (1963): Role of the tongue in producing cleft palatein mice with spontaneous cleft lip. Dev. Biol. 6:45-60Trasler, D.G. and Leong, S. (1982): Mitotic index in mouse embryos with6-aminonicotinamide-induced and inherited cleft lip. Teratology 25: 259-265.Trasler D.G. and Machado, M. (1979): Newborn and adult face shapes related tomouse cleft lip predisposition. Teratology 19: 197-206.Trasler, D.G. and Ohannessian, L. (1983): Ultrastructure of initial nasal processcell fusion in spontaneous and 6-aminonicotinamide-induced mouse embryo cleftlip. Teratology, 28: 91-101.Trasler, D.G. and Trasler, T.A. (1984): Left cleft lip predominance and geneticsimilarities of L line and CL/FR strain mice. Teratology 30:423-427.Van Dyke, D.C., Goldman, A.S., Spielman, R.S., Zmijewski, C.M. and Oka, S.W.(1980): Segregation of HLA in siblings with cleft lip or cleft lip and palate:Evidence against genetic linkage. Cleft Palate J. 17:189-193.Vermeij-Keers, C. (1972): Transformation in the facial region of the humanembryo. Adv. Anat. Embryol. Cell Biol. 46: 7-28.Warbrick, J.C. (1960): The early development of the nasal cavity and upper lip inthe human embryo. J. Anat. (Lond.) 94:351-362.Warkany, J. (1978): Terathanasia. Teratology 17:187-192.Wilson, J.G. (1959): Experimental studies on congenital malformations. J. Chron.Dis. 10:111-130.141Wilson, J.G. (1964): Experimental teratology. Am. J. Qbst. Gyn., 90:1181-1192.Wold, H. (1982): Soft modeling: The basic design and some extensions. In: WoldH.and Joreskog, K. (Eds.) “Systems under indirect observation: Causality,structure, prediction.” North-Holland, Amsterdam. Vol. II, pp. 1-54.Woolf, C.M. (1971): Congenital cleft lip. A genetic study of 496 propositi. J. Med.Genet., 8:65-71.Woolf, C.M., Woolf, R.M. and Broadbent, T.R. (1963): A genetic study of cleft lipand palate in Utah. Am. J. Hum. Genet. 6:244-256.Woollam, D.H.M. and Millen, J.W. (1960): Influence of thyroxin on the incidence ofharelip in the ‘strong A’ line of mice. Br. Med. J. 1:1253-1254.Wright, S. (1934): An analysis of variability in number of digits in an inbred strainof guinea pigs. Genetics 19: 506-536.Xu, Z., Parker S.B., and Minkoff, R. (1990): Distribution of type IV collagen,laminin, and fibronectin during maxillary process formation in the chick embryo.Am. J. Anat. 187:232-246.142APPENDIX 1: The abbreviations applied in the Appendix 2 and 3.TS: Tail somites.HR: Age in hours.WT: Body weight in mg.RMXP:Maxillary prominence depth of right side in jimRPP: Primary palate area of right side in jim2.RM: Mesenchymal area of right side in jim2.RPD: Primary palate depth of right side in jim.RPH: Maximal primary palate height of serial sections of right side in jim.RMH: Maximal mesenchymal height of serial sections of right side in jim.LMXP:Maxillary prominence depth of left side in jimLPP: Primary palate area of left side in jim2.LM: Mesenchymal area of left side in jim2.LPD: Primary palate depth of left side in jim.LPH: Maximal primary palate height of serial sections of left side in jim.LMH: Maximal mesenchymal height of serial sections of left side in jim.143TSHR13277132771327713277142761427614275142741427514275142751527615281152771527415276152761527616276162771627716278WTRMXPLMXPRPPLPPRMLM221401402908630664003118918944384432863292452722161168420514198333613567221821683971940679246941126196189585845241097419192322172104973554811624211792311681754603052067795713102321751754561944864596872713118219645207471961207399473021018957349543311193614187322102105412558310157781317127217217557035803516052138573221019646373456871385715435292312106009356457177671358235224224502155776112279187273722420359819600932490124421332312316016256869190701749333224203602995995620991235293622421059819606421865919894282312246084853576173551351440245259675027010925176303213724524563504656992728327910LPDRPHLPHRMHLMH245151.9156.80.00.0287215.6205.858.868.6259210.7220.568.673.5287196.0186.249.029.4301274.4245.098.0127.4315225.4235.2112.7117.6301210.7235.2102.9127.4287215.6205.888.288.2301205.8225.4117.6102.9329220.5215.6117.6127.4343235.2225.4137.2127.4315264.6254.8156.8137.2301235.2220.5137.2147.0322274.4245.0156.8156.8301245.0264.6147.0171.5343254.8249.9176.4171.5329264.6245.0156.8147.0329254.8245.0161.7161.7315274.4294.0176.4186.2315259.7245.0161.7132.3371279.3274.4176.4176.4357240.1245.0166.6166.6RPD196217245287245LPDRPHLPHRMHLMH210137.2137.219.624.5231127.4137.29.80.0259147.0147.039.29.8280161.7176.468.698.0252151.9161.729.498.0APPENDIX2BALB/cBYJCBS 1 2 3 4 5 6 7 8 9 10 11 12 13H14 16 17 18 19 20 21 22RPD 245301259273301301287287287343315301287315301329343329315329357357C57BL/6JCBSTSHRWTRMXPLMXPRPPLPPRMLM113264241541542174622569274343213264211611542346124078137031326426161168279202881221951374132642620318235191356034870857551326423182182270973100741185756142642622421040954408171042712485301301200.9205.8117.6112.771427227196196325163951358318369273315147.0156.858.878.48142642717516126273237351303548259245137.2132.339.224.5914264252031894129732104100154870301287205.8156.8102.978.41014264251961963683838347871210152273301210.7181.3117.6102.911142643120320342326421201378820923301287215.6215.6137.2171.51214264281822103512338759123489741259287186.2205.8127.4122.513152723422423845687491861646419070371371166.6176.4117.6132.314152723823826656114544682558723598343357225.4205.8156.8166.615152723422423848568519301797320099329329205.8215.6137.2147.016152723519616842394418461275913651329329166.6156.898.098.017152643825223841091481571200520991301329196.0205.8117.6147.01815264252242244129739513919211593273287225.4205.8117.6112.719152642723822442943441781372014680273287245.0225.4156.8147.020162703325225241640371122009921814273287205.8205.8156.8156.821162703925225251998552912346130115273287274.4294.0196.0215.622162703628028058035598872723430115301343284.2254.8196.0196.023162752621021044521500781927624696315329196.0215.6137.2166.624162643225225246510452072044221128273301254.8245.0176.4176.425162703723823845550492541852222363329287245.0254.8166.6186.22616272282382384040546785891814954259301196.0215.698.0137.2A/J OBSTSHRWTRMXPLMXPRPPLPPRMLMRPDLPDRPHLPHRMHLMH1132762110598227061680700175154181.3161.70.00.021327722119119273022078500231189176.4156.80.00.031327626154147255873189901166217252166.6176.40.029.441327723140140258622352900252245147.0147.00.00.051427626147133358092929217830245217225.4200.968.60.06142812313314722363251765480217231137.20.019.60.07142772716817530115286744111234273259166.6166.629.449.08142772416116123735247645480245245147.0147.039.20.0OBS 1 2 3 4 5 6 7 8 9 10 11 12TSHR132791327913281132791328014272142721428014279142811428015279LM0 0 0 0 0 0274 028810 03224RPDLPDRPHLPH189175166.6151.97717563.744.1203210132.3147.0161175102.9102.916118998.0122.528725998.0132.3287259127.4142.1273252166.6117.6259245196.0191.1280273127.4147.0231231117.6122.5301287181.3196.09142772416114736289291553292686287273191.1156.873.529.4101427726126119229121872700224217166.6137.20.00.0111427526154140298413203600259273166.6166.60.00.0121527724175154365633059516462949273259196.0176.449.068.6131527526182161447953779887124596343315196.0186.288.278.4141527633189175389643903351453773259273225.4210.7107.878.4151527722189175391704335510975831329315181.3205.844.198.0161527722196182469224061135672332329315205.8200.988.268.61715281361961892949833751480343245280176.4166.634.324.51815276361751893759243561980912073259287225.4235.2132.3132.3191527632196182448644287598099123315294220.5215.6127.4107.82016277292452315810449392187279398357343254.8240.1142.1122.521162772922423156800506952071716258343343235.2240.1171.5161.7221627725196196380044520785757683259287235.2235.2137.2137.223162772923123145481468531008415503343329225.4225.4122.5156.8241628130196203385534802074779809287315196.0220.5102.9156.8251627930112126617102900495619.639.20.00.0A/WySnWTRMXPLMXPRPPLPPRM2312612621197172870167056288149390221121122771417424018119112125531255302098105112501543502913311919756208540281191332599024901028154140292921941302518217532379306641372241401612798826959023112126194131975602618218935054335450RMH0.00.00.00.00.00.00.00.049.0 0.00.00.0LMH0.00.00.00.00.00.019.6 0.073.5 0.00.00.013 14 15 16 17 18 19 20 21 22152761527615272152781527816272162721628116281162722538 0 0 27439786997 7542058 06997329287196.0171.5259245171.5137.2329273147.0147.0259231166.6181.3301315186.2205.8301315225.4230.3301315215.6191.1301329205.8181.3287273220.5166.6266259235.2240.1LMRPDLPDRPHLPH0214949.029.40301189137.268.6013313373.568.60231203122.5122.50329329181.3151.9018924588.2132.3018920393.183.30147161117.668.60301273137.2122.5072149.078.4031525988.288.20301273132.3122.501613558.839.20287245122.5117.60217245117.6137.20301315166.6176.40203161166.6137.20266252147.0156.827189182378673087024012798126261361413103317515434505291557542616814730046262730301751823793542943102929224231452074335511113HCL/FrCBS 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18261891893834739650281891824061137387332031963649532379282102034150339582WrRMXPLMXPRPPLPP27567782310293615491299091077026914959685488361471472030518247241891823985634643249810512210235292684911344514680371129811113740825168168297722325528426368615092712611917561163953216810530321240102910512650761097271261332627321609361401401879624696291681823395740748361681192545017493261401612586225930TSHR1327813272132721328113280132781328114281142771428014279142801427814279142821428014272152812812336106791 RM0 0 0 06174 0 0 0 411 0 0 0 0 0 0 0 0 058.8 0.034.3 0.039.2127.483.373.5 0.0102.9RMH0.00.00.00.078.4 0.00.00.019.6 0.00.00.00.00.00.00.00.00.053.9 0.00.019.678.498.034.358.8 0.0117.6LMH0.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.0273217122.598.00.00.021718998.088.20.00.016177181.3107.80.00.0315147142.149.00.00.0273301156.8166.60.049.0273259181.3196.019.649.0259273186.2196.068.688.2329301176.4137.278.40.0301301186.2196.083.398.0343315156.8147.053.90.017516149.0117.60.00.0287287215.6196.088.20.0301301230.3176.4137.278.416120398.098.00.00.0259315205.8215.688.2117.6287301205.8176.4127.449.021727398.0107.80.00.0287343166.6176.483.3107.8273273147.0166.60.039.219152783618915421814142000020152812911298155031193600211527824154119175615625002215277331681472819437040023162824021017527302318300686241628234182196338193505427412342516278432101892963533888267559682616280321821824088528263507602716279331892033841636358596810976281627734189168399253292818520291628030989127441262200301628133196210423263560312340311627931210231428753375111936219532162773010516173401344500331628138217224325854198346647408H3416280272101964054236083122792332351628122105112166691927600361627831196196331334253219891296537162782918218227028300460891APPENDIX 3BALB/cByJPEARSON CORRELATION COEFFICIENTS / N =22TS WT RMXP RPP RM RPD RPH RMHTS 1.00 0.69 0.85 0.81 0.88 0.74 0.77 0.90WT 1.00 0.69 0.64 0.74 0.73 0.53 0.74RMXP 1.00 0.90 0.86 0.88 0.83 0.90RPP 1.00 0.88 0.92 0.90 0.89RM 1.00 0.85 0.75 0.92RPD 1.00 0.70 0.79RPH 1.00 0.85RMH 1.00TS WT LMXP LPP LM LPD LPH LMHTS 1.00 0.69 0.83 0.80 0.86 0.72 0.80 0.85WT 1.00 0.73 0.78 0.87 0.77 0.66 0.79LMXP 1.00 0.88 0.82 0.85 0.77 0.80LPP 1.00 0.91 0.92 0.86 0.90LM 1.00 0.87 0.77 0.90LPD 1 .00 0.65 0.78LPH 1.00 0.91LMH 1.00C57BL/6JPEARSON CORRELATION COEFFICIENTS / N =26TS wr RMXP RPP RM RPD RPH RMHTS 1.00 0.68 0.84 0.79 0.83 0.49 0.74 0.82‘NT 1.00 0.71 0.76 0.79 0.64 0.58 0.71RMXP 1.00 0.86 0.83 0.52 0.84 0.86RPP 1.00 0.95 0.74 0.85 0.92RM 1.00 0.65 0.83 0.94RPD 1.00 0.37 0.57RPH 1.00 0.93RMH 1.00TS WT LMXP LPP LM LPD LPH LMHTS 1.00 0.68 0.81 0.78 0.83 0.59 0.77 0.81‘NT 1.00 0.72 0.77 0.79 0.63 0.63 0.72LMXP 1.00 0.87 0.87 0.67 0.83 0.87LPP 1.00 0.92 0.83 0.82 0.89LM 1.00 0.69 0.87 0.96LPD 1.00 0.41 0.62LPH 1.00 0.92LMH 1.00149NJPEARSON CORRELATION COEFFICIENTS / N =25TS WT RMXP RPP RM RPD RPH RMHTS 1.00 0.51 0.72 0.47 0.68 0.32 0.26 0.74WT 1.00 0.42 0.17 0.40 0.03 0.18 0.45RMXP 1.00 0.86 0.78 0.81 0.69 0.83RPP 1.00 0.80 0.93 0.89 0.83RM 1.00 0.61 0.64 0.90RPD 1.00 0.79 0.65RPH 1.00 0.74RMH 1.00TS WT LMXP LPP LM LPD LPH LMHTS 1 .00 0.51 0.76 0.56 0.71 0.41 0.37 0.77Wt 1.00 0.49 0.33 0.45 0.20 0.29 0.41LMXP 1.00 0.86 0.84 0.80 0.62 0.87LPP 1.00 0.79 0.94 0.80 0.87LM 1.00 0.64 0.66 0.94LPD 1.00 0.71 0.72LPH 1.00 0.73LMH 1.00NWySnPEARSON CORRELATION COEFFICIENTS / N =22TS WT RMXP RPP RM RPD RPH RMHTS 1.00 0.74 0.83 0.85 0.59 0.74 0.83 0.68WT 1.00 0.69 0.77 0.27 0.88 0.63 0.34RMXP 1.00 0.91 0.64 0.74 0.88 0.73RPP 1.00 0.60 0.85 0.91 0.70RM 1.00 0.31 0.60 0.92RPD 1 .00 0.65 0.43RPH 1.00 0.70RMH 1.00TS WT LMXP LPP LM LPD LPH LMHTS 1.00 0.74 0.84 0.83 0.56 0.84 0.74 0.57Wt 1.00 0.68 0.68 0.30 0.75 0.59 0.31LMXP 1.00 0.94 0.72 0.81 0.91 0.68LPP 1.00 0.72 0.87 0.91 0.74LM 1.00 0.50 0.78 0.90LPD 1.00 0.68 0.52LPH 1 .00 0.76LMH 1.00150CL/FrPEARSON CORRELATION COEFFICIENTS / N =37TS WT RMXP RPP RM RPD RPH RMHTS 1.00 0.28 0.54 0.42 0.36 0.34 0.50 0.49WT 1.00 0.49 0.26 0.01 0.21 0.34 0.15RMXP 1.00 0.89 0.53 0.80 0.90 0.64RP 1.00 0.59 0.86 0.91 0.70RM 1.00 0.35 0.62 0.90RPD 1.00 0.67 0.43RPH 1.00 0.73RMH 1.00TS WT LMXP LPP LM LPD LPH LMHTS 1.00 0.28 0.64 0.50 0.42 0.43 0.58 0.55WT 1.00 0.32 0.24 0.28 0.17 0.30 0.40LMXP 1.00 0.83 0.49 0.73 0.82 0.64LPP 1.00 0.53 0.91 0.94 0.63LM 1.00 0.39 0.52 0.89LPD 1.00 0.78 0.46LPH 1.00 0.67LMH 1.00151

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0086710/manifest

Comment

Related Items