Postgraduate Medical Journal (August 1977) 53, 517-522.

The embryological basis of craniofacial dysplasias

D. E. POSWILLOD.D.S., D.Sc., F.D.S., F.I.Biol., M.R.C.Path.

Department of Teratology, Royal College of Surgeons of England, andQueen Victoria Hospital, East Grinstead

SummaryCraniofacial dysplasia of a syndromic pattern can

usually be classified into one of two easily identifiablegroups. In the first group are those malformations ofthe craniofacial skeleton and soft tissues that are

asymmetrical in form and in the other, those that areprincipally symmetrical. Clinical studies have de-monstrated that affected subjects in the symmetricalgroup frequently improve in terms of facial appearanceas growth and development proceed to maturity,while those with asymmetrical defects often deterio-rate in this respect. Embryological studies on animalmodels of these malformations have shown thatasymmetrical lateral facial dysplasia and symmetricalmandibulofacial dysplasia exhibit discrete and widelydisparate causal mechanisms of malformation.Analysis of these mechanisms and their effects on

subsequent growth and development has suggestedsignificant variations in the timing and technique ofreconstructive procedures which will enable thesurgeon to produce the most effective results when usedfor the rehabilitation of the afflicted.

THE congenital craniofacial dysplasias fall into twomain groups: one with a strong familial history andusually with autosomal dominant transmission and a

group in which the abnormality appears as an

isolated event with no relevant family history.Many of these craniofacial dysplasias, familialor otherwise, have very similar clinical features andyet they appear to respond in very different ways toreconstructive surgery. This paper will describe someembryological investigations into the causal mechan-isms which initiate these congenital craniofacialdeformities and show how these studies havecontributed to an understanding of the growth anddevelopment of these dysplasias, and the way inwhich they respond to surgical measures.

Lateral facial dysplasiaThe first group is called lateral facial dysplasia

and was known in the pre-Columbian era, c. 100A.D., when the characteristic facies was recognized.This condition has no family history; it appears once

in every 3500-4000 births and usually presents with aunilateral malformation which particularly affectsthe oto-mandibular structures. There is a deficiencyof the malar bone and the zygomatic arch of the earwhich may be reduced to just a small auricular tag,and a considerable skeletal deficiency, which isbilateral in about 30yo of cases-it is invariablymore severe on one side than on the other. It isnever a symmetrical malformation (Fig. 1).On examination, there is a defect in the body and

ramus of the mandible behind the molar teeth,extending often in severe cases, right up to thecoronoid process and the condyle, a defect of themalar bone, usually absence of the zygomatic arch,and often considerable deficiency in the area of thepetrous temporal.There is another closely comparable condition

known as Goldenhar's syndrome, where there arenot only the skeletal and facial defects seen in lateralfacial dysplasia, but often a co-existing coloboma,usually of the upper eyelid and sometimes anomaliesof the cervical vertebrae. It is now believed that thisGoldenhar syndrome is a variant of lateral facialdysplasia. Work by Ross (1975) at the TorontoChildren's Hospital has shown that by and largethese patients do not always have the combination ofcoloboma and vertebral anomalies as expected, andit is very likely, in view of the similarities of Golden-har syndrome and lateral facial dysplasia, that theyare essentially the same disorder. There is no familialhistory in these cases, but as the patients grow olderand reach puberty, so their condition on the affectedside often deteriorates. Growth and developmenton the affected side lags behind growth on theunaffected or less affected side, and so the cosmeticdefects become considerably worse.The major defects in the craniofacial skeleton are

found in the reductions in the temporal bone, in theramus and body of the mandible, in the zygomaticarch and the malar bone; the auditory ossicles mayeither be absent, abnormal or fused together. Thereare also soft tissue defects. The masticatory muscles,the temporal and masseter are particularly involved;the parotid gland may be small or absent, and

Protected by copyright.

on Novem

ber 11, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.53.622.517 on 1 August 1977. D

ownloaded from

518 D. E. Poswillo

i b...

::sys~ ~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

FIG. l(a). Facial appearance of left-sided lateral facial dysplasia showing facial paresis, malar flattening and auricularmalformation. (b). Radiograph of lateral facial dysplasia seen in Fig. Ia. Note skeletal defects in mandible andzygomatico-malar regions on affected side.

occasionally there is not complete facial paralysis,but a weakness of the facial muscles on the affectedside.

This dysplasia was called the 'first and second archsyndrome' for many years, but this was not aparticularly good title, hence the name was changedto lateral facial dysplasia. Not all the structuresinvolved are derived from the first and secondbranchial arches, e.g. the squamous temporal isunlikely to be so derived. Yet other structures whichare derived from these arches, such as the hyoidbone which arises from the second, is never involvedin this condition. In the U.S.A. the term 'facialmicrosomia' is used for this disorder.

Recently an animal model of this condition hasbeen described (Poswillo, 1973) which, in terms of itsexternal features, compares very closely with thehuman condition. In these experimental animals,approximately 3040o/ showed bilateral defects,but these were never symmetrical. There wereanomalies of the ear, gross reduction of the pinna,blind endings to the meatal tubes, defects of themalar bone and ramus of the mandible, demon-strated by comparison with the normal rodent.In the cleared and stained specimens, instead of thenormal zygomatic arch and the condylar and coro-noid processes of the mandible, there were grossdefects in the ramus, in the malar bone, zygoma andtemporal bone. Not only were there soft tissuedefects comparable with man in these animal modelswhich were produced by anti-folate drugs, but therewere also skeletal defects very similar to those oflateral facial dysplasia. Having developed a model in

which 100% of animals were deformed, it was thenpossible to study the sequence of events leading upto this malformation.

It was observed in the rodent at day 14, whichcorresponds roughly with day 32 or 33 in man, thatthe stapedial arterial system was beginning todevelop. This replaces the first and second aorticarches after they shut down; from the stapedialarterial stem, which arises in the neck by the junctionof the ascending pharyngeal and hyoid arteries,develop the supra-orbital, infra-orbital, maxillary,mandibular and hyoid branches. Later on, about day42, these are annexed by the carotid system and thestapedial stem itself disappears. It was found that atthe time of emergence of the stapedial stem in therodent model, a haemorrhage appeared just at thedeveloping stapedial anastomosis. This spread intothe tissues, sometimes very small, sometimes verylarge, destroying large tracts of differentiated tissuedestined to form the mandibular ramus, the middleand external ear, and all the structures in the vicinityof the oto-mandibular region (Fig. 2).One could demonstrate a small, localized haema-

toma disrupting tissue destined to develop into theendaural cartilage, auditory ossicles and part of theexternal ear, also spreading to involve the condyleof the mandible. When there is a very large haema-toma it displaces much larger tracts of tissue. Later,there is consequent phagocytosis of the extravasatedblood and the defect is repaired by mesenchymeonce again; but the delay is such that completedifferentiation does not take place in that area, orif it does catch up at all, then primitive ears and

Protected by copyright.

on Novem

ber 11, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.53.622.517 on 1 August 1977. D

ownloaded from

Embryological basis of craniofacial dysplasias 519

LI'.J ,

FIG. 2(a). Frontal section of otomandibular regions of normal fetus showing differentiation of ear and jaw at approximatetime of focal haematoma formation. (b). Asymmetrical bilateral haematomas in otomandibular regions of fetus developinglateral facial dysplasia. Note that the haematoma is deflected away from the inner ear by the otic capsular cartilage.

mandibular rami develop. Therefore many structuresare incompletely formed, leading not only to skeletaldefects but also to deficiency of the muscles. In arelatively minor example, comparing the affectedwith the normal side, the former shows a muchsmaller body of mandible, masseter and temporalmuscle, showing that there has been damage both tomuscle and the skeleton. This damage to the musclesplays an important part in the reduction of growthassociated with this particular problem. Obviously,with a very large haematoma, local damage is moreextensive, leading to very severe defects, particularlyin the masticatory muscles.With unequal damage on both sides, there is

clearly a considerable difference in the residualeffect, with marked asymmetrical deficiency of thebody of the mandible and molar teeth and completeabsence of the zygomatic arch and masseter muscleon the severely affected side. Hence, with absence ofthe muscular components of the functional matrixresponsible for growth, it is understandable why, inthese cases, as growth proceeds towards puberty thegreatly affected side lags behind and the conditionbecomes much worse as age progresses.Owing to the absence of adequate musculature,

most attempts at surgical reconstruction during thegrowth period are foredoomed to failure in thissituation. This is why surgeons have had such poorsuccess in attempting to restore to normal the facialappearance while growth is still active on the normalside and static on the severely affected side. Whenreconstructive surgery re-establishes facial sym-metry during the growth period, this unilateral

growth leads once again to progressive asymmetryand deformity.Our understanding of the causal mechanism helps

us to realize that in severe cases it is useless to tryreconstructive surgery during the growth period.In lateral facial dysplasia, it is desirable to postponeany attempt to establish facial symmetry until growthon the affected or lesser involved side is almostcomplete.

Mandibulofacial dysplasia (Treacher Collins syn-drome)The second group of facial dysplasias include

those of the Treacher Collins type, where there isstrong familial tendency; it is interesting to comparethe changes in the dsyplastic facial skeleton andthe differences in the causal mechanisms in the twogroups. The Treacher Collins group was also re-cognized by the pre-Columbian Indians who under-stood the familial pattern; it is now known thatabout 50% of these cases have an autosomaldominant form of transmission.

In this disorder the skeletal changes are sym-metrical in the craniofacial skeleton. There is somereduction in the ramus and condyle of the mandible,deficiencies of the zygomatic arch and malar bone(Fig. 3). But these are always symmetrical, asopposed to lateral facial dysplasia where they areinvariably asymmetrical. The Hallermann-Streiffsyndrome is another condition with a similar geneticbackground to the Treacher Collins type, withsimilar defects-hypoplasia of the malar bones,defects of the zygomatic arch, abnormalities of the

Protected by copyright.

on Novem

ber 11, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.53.622.517 on 1 August 1977. D

ownloaded from

520 D. E. Poswillo

ci. b 4, w

FIG. 3(a). Facial appearance of mild Treacher Collins syndrome showing symmetrical defects of eyes,malar regions and ears. (b). Lateral radiograph of Treacher Collins syndrome showing characteristicdysplasia of mandible (with lower border curvature) open bite and zygomatic-malar insufficiency.

ear and somc loss of the overall growth of themandible. In Hallermann-Streiff syndrome there issymmetrical hypoplasia of the middle and lowerthirds of the face.To the embryologist this symmetry is most in-

teresting for it indicates a different causal mechanismfor the malformations, although the facial appear-ance is very similar to that of lateral facial dysplasia.One must seek for some other explanatory patho-genesis, operating very early in embryogenesis,unrelated to focal haematoma formation, and yetcapable of producing symmetrical hypoplasia withdefects in the skeleton. Johnston and Listgarten(1972) showed that a large proportion of the middleand lower thirds of the face arose in mesenchymeor ectomesenchyme which migrated into the branchialarches from the neural crest just after closure of theneural tube. They showed that many cells migratedunder the ectoderm into the branchial archeswhere, in combination with the mesoderm, theycontributed to the development of the skeleton of themiddle and lower thirds of the face.

In attempting therefore to produce a suitableanimal model of mandibulofacial dysostosis studywas made of teratogenic agents which could in-fluence the migration of the ectomesenchyme of theneural crest leading to greatly diminished branchialarches. These investigations were eventually success-ful (Poswillo, 1975) and revealed a further interestingfeature on comparison of the normal animal withthe Treacher Collins animal model. In the normal

animal the otocyst lay adjacent to the secondbranchial arch. In the Treacher Collins model, theotocyst had drifted up, as a result of the spatialrearrangement of tissues which followed the deathof the neural crest ectomesenchyme to the firstbranchial arch territory, locating the developing earover the angle of the mandible instead of in itsnormal situation (Fig. 4). In the Treacher Collinssyndrome we characteristically find the ears asym-metrically located low down, close to the angle of themandible in first arch territory rather than in theirnormal second arch region. From the animal modelof the Treacher Collins syndrome, it appears thatthe damage results from the early death of cells dueto migrate into the branchial arches, so interferingwith the development of the facial skeleton. Asection of this region in the Treacher Collins modelshows the otocyst high in the region of the firstbranchial arch with an area of dead tissue represent-ing death of ectomesenchymal cells which shouldmigrate into the branchial arches to produce thefacial skeleton (Fig. 4).The animal model shows the changed position of

the ear which has been described, not only thespatial change to the region of the angle of themandible, but abnormalities of the pinna which alsoappear in Treacher Collins syndrome, together withdefects in the malar region and the mandible (Fig. 5).These are symmetrical deficiencies. In the matureanimal skeleton of the Treacher Collins model,defects are seen in the zygomatic arch, the malar

Protected by copyright.

on Novem

ber 11, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.53.622.517 on 1 August 1977. D

ownloaded from

Embryological basis of craniofacial dysplasias 521

FiG. 4(a). Normal rat at term showing position of auricle and malar-mandibular contours. (b). Rat model of TreacherCollins syndrome showing malformed and malposed ear, malar flattening and altered craniofacial proportions.

4Al

I 21~~~4

I /-

a~~~~~~~~~~~~~~~~~~~~~

FIG 5(a). Sagittal section ofnormal rat embryo at time ofdevelopment of the branchialarches. Note dimensions of 1st and 2nd arches (1, 2) and neural crest cells in mesen-cephalon (arrow). (b). Sagittal section of rat model of Treacher Collins syndromeinduced by vitamin A, showing cell death in vicinity of neural crest (arrow) andpaucity of mesenchyme in branchial arches 1 and 2.

Protected by copyright.

on Novem

ber 11, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.53.622.517 on 1 August 1977. D

ownloaded from

522 D. E. Poswillo

bone, the lateral wall of the orbit and dysplasia ofthe coronoid and condyle region of the mandible,by comparison with the normal morphology. Theexperimental model showed an attempt to producea mandible that was functional; this was notcompletely successful as it did not have the normalangular, coronoid and condylar processes. It seemedthat the result of the deficiency of neural crestectomesenchyme was incomplete production of thenormal, genetically-controlled facial skeleton.

This is, of course, exactly what is found in theTreacher Collins craniofacial dysplasia. The ears areabnormally placed, there are defects in the malarregion with alterations in the slope of the eyelids,often colobomata of the lower lids (which aoainmay be due to a deficiency in neural crest ecto-mesenchyme); there is an open bite and a facewhich, although recognizable as human, is notquite normal. Although we have these similardefects in the Treacher Collins syndrome andlateral facial dysplasia, they don't behave in thesame way with respect to subsequent growth.In Treacher Collins syndrome, as the child growstowards puberty, the face likewise grows. Atadolescence this progress may be measured andfound to be symmetrical, but the growth patternhas not been quite normal; a cephalometric tracingof an afflicted child shows there are still differencesof morphology with excessive downward and for-ward growth of the dysplastic bones of the facialskeleton when compared with a normal skull ofabout the same size (Fig. 6). The reason for thisvariation is that there is abnormal but symmetricalmusculature attached to the mandible and posterioraspect of the maxilla in the Treacher Collins syn-drome on both sides. Thus the motor units aresymmetrical and although there is some reductionin the facial bones-hypoplastic bones beneathslightly abnormal muscles-together they do form afunctional matrix which provides for symmetricalgrowth of the face. Hence, one may operate toreconstruct a child with Treacher Collins syndromeduring the period of active growth, confident thatgrowth will continue to be symmetrical. This im-portant fact has made a considerable difference tothe clinical management of these cases. In lateralfacial dysplasia, as mentioned above, surgery duringthe growth period is usually contra-indicated. Theanimal studies of these two craniofacial dysplasiashave therefore not only helped to distinguish themin terms of their causal mechanisms but have alsocontributed greatly to a more scientific approach tothe timing and techniques of surgical reconstructiveprocedures.

I

FIG. 6. Tracing of lateral cephalogram of normal childsuperimposed on tracing from head films of TreacherCollins child of similar age. Observe the variation indownward and forward growth patterns in the TreacherCollins tracing.

A similar experimental approach to the study ofanimal models of Apert and Crouzon syndromes iscurrently in progress. Early investigations areencouraging and suggest that the models may leadto further information about their causal mechanism.Already, some of the exploratory surgical workbased upon embryological studies seems to indicatethat much earlier management of these disorders bycraniectomy may lead to more positive facialdevelopment than the later sophisticated recon-structive operations introduced by Tessier (1967)during the last few years, procedures which have amuch greater morbidity. Embryological studiesdesigned to explore the mysteries of congenitaldysplasias thus contribute not only to the bodyof knowledge concerning patterns of normal andabnormal development, but also to the rehabilita-tion of the handicapped child.

ReferencesJOHNSTON, M.C. & LISTGARTEN, M. (1972) In: Developmental

Aspects of Oral Biology (Ed. by H. C. Slavkin and M. C.Bavetta), p. 53. Academic Press, New York.

POSWILLO, D. (1973) The pathogenesis of the first and secondbranchial arch syndrome. Oral Surgery, Oral Medicine,Oral Pathology, 35, 302.

POSWILLO, D. (1975). The pathogenesis of the TreacherCollins syndrome. British Journal of Oral Surgery, 13, 1.

Ross, R.B. (1975) Lateral facial dysplasia (first and secondbranchial arch syndrome, hemifacial microsomia). BirthDe.fects, 11, 7, 51.

TESSIER, P. (1967). Osteotomies totales de la face syndrome deCrouzon, syndrome d'Apert. Annales de chirurgie plastique,12, 273.

Protected by copyright.

on Novem

ber 11, 2020 by guest.http://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.53.622.517 on 1 August 1977. D

ownloaded from