SEARCH

SEARCH BY CITATION

Keywords:

  • CHARGE syndrome;
  • pathogenesis;
  • embryogenesis;
  • multiple congenital anomalies;
  • mesenchymal;
  • epithelial;
  • induction

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ADDED NOTES
  8. REFERENCES

To be seriously considered, a theory about the pathogenesis of a multiple congenital anomaly syndrome should meet three criteria: (1) it should explain all of the anomalies associated with the syndrome; (2) it should explain why certain anomalies are not associated with the syndrome; and (3) it should predict anomalies that could be associated with the syndrome, but have not yet been described. The theory must eventually pass the ultimate test, that is, molecular confirmation of the proposed mechanism. Several theories about the pathogenesis of CHARGE syndrome have been proposed, but none of these meet the three criteria stated above. In this study, the author proposes that CHARGE syndrome is due to a disruption of mesenchymal–epithelial interaction (epithelial includes ectoderm and endoderm). The theory is tested against the major, minor, and occasional anomalies that are used to make the clinical diagnosis of CHARGE syndrome. Review of the known embryology of the organs and tissues involved in CHARGE syndrome confirms that mesenchymal–epithelial interactions are necessary for proper formation of these organs and tissues. The presence of limb anomalies in approximately one-third of CHARGE syndrome patients fulfills criteria #3 above, in that limb anomalies were not felt to be a part of CHARGE syndrome until relatively recently. It is known that some patients with chromosomal abnormalities have a phenotype that overlaps with CHARGE syndrome. Given that critical developmental pathways must be robust and redundant in order to minimize errors, it may be that disruption of more than one gene is necessary to generate the CHARGE phenotype, as has been proposed for the holoprosencephaly sequence. Mutations and deletions of CHD7 have recently been identified as causing CHARGE syndrome in more than 50% of tested patients. Given this gene classes' putative role as a general controller of developmental gene expression as well as mesodermal patterning, it would fit the hypothesized mechanisms discussed in the study. © 2005 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ADDED NOTES
  8. REFERENCES

The association between choanal atresia and other anomalies was first noted by Hall [1979]. Pagon et al. [1981] confirmed the association between choanal atresia, ocular coloboma, and other anomalies and coined the acronym CHARGE. Since that time, several hundred patients with CHARGE syndrome have been described. At this point, the pathogenesis of CHARGE syndrome is unknown. Several pathogenetic theories have been proposed that include: teratogenic exposure [Oley et al., 1988]; intrauterine insult during day 35–45 postconception [Oley et al., 1988]; generalized disturbance of the mesoderm [Warburg, 1983]; and maldevelopment of the neural crest [Siebert et al., 1985; Wright et al., 1986]. Van Meter and Weaver [1996] propose a combination of these last two theories; that is, a disruption of the interactive induction of the mesoderm and neural crest cells. A genetic cause is supported by several patients with chromosomal abnormalities that have clinical overlap with CHARGE syndrome; apparent autosomal dominant inheritance of a mild CHARGE phenotype [Mitchell et al., 1985a]; and concordance in monozygotic twins [Oley et al., 1988; Tellier et al., 1998] and monozygous triplets [Blake et al., 1989].

To be seriously considered, a theory about the pathogenesis of a multiple congenital anomaly syndrome should meet three criteria: (1) it should explain all of the anomalies associated with the syndrome; (2) it should explain why certain anomalies are not associated with the syndrome; and (3) it should predict anomalies that could be associated with the syndrome, but have not yet been described. The theory must eventually pass the ultimate test, that is, molecular confirmation of the proposed mechanism. None of the proposed pathogenetic mechanisms meet all three criteria. Exposure to a teratogen was considered, as there is some overlap between CHARGE syndrome, thalidomide embryopathy, and congenital rubella syndrome. To date, no consistent exposures have been identified in affected pregnancies. Teratogen exposure also does not explain the genetic features described above. An intrauterine insult during a critical time of development is unable to be supported, as the range of timing for the various malformations is estimated between 20 and 60 days post-conception. A generalized disturbance of the mesoderm does not explain the involvement of structures that are clearly of ectodermal or endodermal origin. The neural crest theory has much to recommend it, given the preponderance of craniofacial anomalies seen in CHARGE syndrome. However, the neural crest is not known to be involved in renal or limb development, and malformations in these structures are frequently seen in CHARGE syndrome. This argument also applies to the neural crest-mesoderm induction disruption theory, although this theory falls within the hypothesis to be proposed.

In this study, the author proposes that CHARGE syndrome is due to a generalized disruption of mesenchymal–epithelial interaction (mesenchymal includes mesenchyme and mesoderm and epithelial includes ectoderm and endoderm).

METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ADDED NOTES
  8. REFERENCES

The AI/GEN Model 2 diagnostic criteria for sporadic CHARGE syndrome were used to identify malformations for analysis [Mitchell et al., 1985b]. (NB: This model incorporates all of the acronymic major features, as well as minor diagnostic features that are termed Model 1 by Mitchell et al. [1985b]). All major, intermediate, and minor findings were analyzed. The embryology of each malformation was researched using standard texts and relevant medical literature. Additional malformations associated with CHARGE syndrome, but not included in the AI/GEN diagnostic criteria set, were identified and analyzed.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ADDED NOTES
  8. REFERENCES

Condition 1: Associated Malformations Are Explained by Hypothesis

Table I lists the AI/GEN Model 2 Diagnostic Criteria. The associated malformations are listed along with the embryologic tissues involved. These are discussed further below.

Table I. Malformations Used to Develop the AI/GEN Model 2 Diagnostic Criteria [Mitchell et al., 1985b] and the Embryologic Origin of the Tissues Involved in the Malformation
Diagnostic criteriaMalformationEpithelial anlageaMesenchymal anlageb
  • a

    Includes ectoderm, endoderm, and epithelium.

  • b

    Includes mesoderm and mesenchyme.

Major findingsOcular colobomaNeuroectoderm, surface ectodermNeural crest, placodal mesoderm
 Choanal atresia/stenosisSurface ectodermFacial mesenchyme, ? neural crest
 CHARGE external ear? Surface ectodermPharyngeal arch mesenchyme
 CHARGE hearing loss  
  CN VIIIPlacodal ectodermNeural crest
  LabyrinthPlacodal ectodermPlacode associated mesenchyme
  OssiclesSurface ectoderm, endodermNeural crest, mesoderm (paraxial and lateral plate)
 Facial paralysis (CN VII)Placodal ectodermNeural crest
Intermediate findingsCongenital heart disease (conotruncal lesions)Pharyngeal endodermLateral plate mesoderm, neural crest mesenchyme
 CNS anomaliesSurface ectoderm, neuroepitheliumNeural crest mesenchyme
  ArrhinencephalyOlfactory epitheliumNasal mesenchyme
 PituitaryOral ectoderm (Rathke's pouch)Ventral mesenchyme
 Possible CHARGE earSee aboveSee above
 Sensorineural hearing lossSee aboveSee above
Minor findingsUrinary tract malformationsMesonephric epitheliumMetanephric blastema mesenchyme
 Short statureSee textSee text
 Cleft palatePharyngeal arch ecto- and endodermNeural crest mesenchyme
 DiGeorge sequencePharyngeal pouch epitheliumNeural crest mesenchyme
 OmphaloceleSurface ectodermLateral plate mesoderm
 TE fistula/esophageal atresiaForegut associated endodermForegut associated mesoderm
Cranial nerve abnormalities

Several of the major and intermediate criteria involve cranial nerve dysfunction (hearing loss CN VIII, facial paralysis CN VII, pharyngeal incoordination CNs IX/X). In addition, arrhinencephaly has been frequently reported, which could relate to CN I, although there are no reports of patients being assessed for anosmia. Rare reports of involvement of CNs III and VI have been noted [Byerli and Pauli, 1993].

The pattern of cranial nerve involvement is quite enlightening when examined from an embryologic perspective. All CNs except I and II develop from the brainstem. The motor nuclei of CNs IV, V, VI, VII, IX, X, XI, and XII develop from pairs of rhombomeres that are of neuroepithelial origin. The segmental pattern seems to be directed by mesoderm that is organized into somitomeres beneath the neuroepithelium [Sadler, 2000]. While there is doubtless an inductive event that takes place in this segmental organization, the pattern of CN involvement in CHARGE does not match that of the motor nuclei. CNs V, VII, IX, and X are different from the other cranial motor nerves, in that they also have a sensory ganglion associated with the nerve. These sensory ganglia are derived from an interaction between neural crest mesenchyme and ectodermal placodes [D'Amico-Martel and Noden, 1983; Sadler, 2000]. The implication is that the CHARGE CN abnormalities are seen specifically in these nerves due to unique induction events necessary to generate the sensory ganglia, and that the motor abnormalities seen are secondary to this, not due to a primary problem with motor ganglion induction. If this theory is correct, then the apparent absence of involvement of CN V requires explanation. One explanation is that CN V is involved but due to difficulty of clinical diagnosis this has not been recognized. None of the reported patients have documentation of facial sensation testing, or absence of the corneal reflex. CHARGE patients are known to have persistent problems with oral feeding, which are usually attributed to difficulties with swallow coordination. It is possible that abnormal function of the muscles of mastication may also contribute to the oral feeding difficulties. Alternatively, CN V may not be involved. The sensory ganglion of CN V is different from the other CNs listed above in that the neural crest cells derive from the junction of the forebrain and midbrain, whereas the neural crest for all of the other nerves derives from the hindbrain [Sadler, 2000]. Also, the ectodermal component arises from the trigeminal placode, which does not contribute ectoderm to any of the other CNs. The supposition would be that the molecular mechanism of induction for CN V is different than that of the other CNs, most likely due to the different origin of the neural crest mesenchyme (see brain anomalies below).

Cranial nerve I

Arrhinencephaly, as well as other abnormalities of the olfactory bulb and tracts, has been reported frequently in CHARGE patients, [Lin et al., 1990]. Recent gene expression experiments, looking at the olfactory nerve pathway in rats, show a probable induction interaction between the olfactory epithelium and the surrounding nasal mesenchyme beginning at around day 14 [Astic et al., 2002]. These genes appear to be critical both for proliferation of cells, as well as directing migration along the path of the olfactory nerve. These cells interact with the olfactory bulb primordium. It is possible that disruption of these inductive events could result not only in abnormalities of CN I, but also in arrhinencephaly if cell migration is necessary for growth and differentiation of the olfactory bulb. Alternatively, studies of the molecular pathogenesis of Kallman's syndrome demonstrate a topographical relationship between gonadotropin releasing hormone (GnRH) synthesizing neurons and the embryonic olfactory system [Hardelin, 2001]. It is therefore possible that the abnormalities seen in CHARGE syndrome are secondary to hypothalamic abnormalities known to occur in CHARGE [Wheeler et al., 2000] (see below). Braddock et al. [1995] did show that arrhinencephaly does not predict absence of the olfactory receptor cells or nerves so it cannot be concluded with certainty that CN I is frequently involved in CHARGE based on the presence of arrhinencephaly alone. The studies of Rabaeus et al. [1983] and Lin et al. [1990] do confirm structural involvement of CN I, in at least some cases.

Cranial nerve VIII

This nerve develops from the facial-vestibulo-acoustic ganglionic complex [D'Amico-Martel and Noden, 1983] in the chick in conjunction with CN VII. In the mouse, while the geniculate (facial) and vestibulo-acoustic ganglia develop in close proximity, they arise from different ectodermal anlage [Wikström and Anniko, 1987]. In either case, neurons in the sensory ganglia of both nerves derive from ectoderm [epibranchial placode (CN VII) and the otic placode (CN VIII)] and neural crest mesenchyme. In addition, glial cells in these ganglia also are of neural crest derivation [Le Douarin et al., 1986]. It should also be noted that the proximal ganglion of CN VII fuses with the vestibular ganglion of CN VIII during the course of development [Le Douarin et al., 1986]. While the neural crest contribution to the vestibular ganglion of CN VIII is much greater than seen in the acoustic ganglion of the same nerve [D'Amico-Martel and Noden, 1983], there does appear to be contribution of the neural crest to the acoustic ganglia in both the chick and mouse [Deol, 1967; Wikström and Anniko, 1987], although this is disputed by other authors [Sadler, 2000]. This is important, as abnormalities of vestibular function have been recognized more recently as a significant contributor to the delayed acquisition of gross motor milestones in these patients [Abadie et al., 2000].

Alternatively, formation of the CN VIII ganglia may have little to do with the hearing and vestibular abnormalities noted in these patients. Abnormal development of the labyrinth will also result in sensorineural hearing loss and vestibular dysfunction. The structures of the labyrinth develop independently of CN VIII [Guyot et al., 1987]. Labyrinth dysplasia due to abnormal formation of the temporal bone has been frequently reported in CHARGE syndrome [Sekhar and Sachs, 1976; Wright et al., 1986; Guyot et al., 1987; Admiraal et al., 1998; Lemmerling et al., 1998]. Amiel et al. [2001] felt that this anomaly was frequent enough to warrant consideration as a major diagnostic criterion. Recent success treating the hearing loss in patients with CHARGE syndrome by using cochlear implantation [Bauer et al., 2002] lends additional support to the concept that labyrinth dysplasia, not abnormalities of CN VIII, is responsible for the sensorineural hearing loss and vestibular dysfunction seen in these patients.

Formation of the labyrinth involves an induction interaction between mesenchyme and surface ectoderm (otic placode) that leads to invagination and formation of the otic vesicle at approximately 28 days [Sadler, 2000]. The interaction between these two tissues continues when the ventral portion of the otic vesicle forms the cochlear duct (6 weeks) that penetrates the surrounding mesenchyme in a spiral fashion, completing a total of 2.5 turns by the end of the 8th week, after which time the mesenchyme differentiates into cartilage [Sadler, 2000]. At approximately the same time, the dorsal portion of the otic vesicle interacts with its surrounding mesenchyme to form the utricle and semicircular canals [Sadler, 2000].

CHARGE patients are also known to have ossicular abnormalities [Sekhar and Sachs, 1976; Wright et al., 1986; Guyot et al., 1987; Admiraal et al., 1998; Lemmerling et al., 1998]. The ossicles are derived from the first pharyngeal arch, which is composed of surface ectoderm, epithelial cells of endodermal origin, and mesenchyme from paraxial and lateral plate mesoderm, as well as the neural crest. Recent gene expression studies indicate that normal development of the middle ear is dependent on sequential interactions between the epithelia and underlying mesenchyme [Mallo, 2001]. These abnormalities contribute to the conductive component of hearing loss seen in these patients.

Cranial nerves III and VI

Byerli and Pauli [1993] report that 3 of 163 (2%) patients had involvement of CN III and 2 of 163 (1%) had involvement of CN VI. All of these were associated with other cranial nerve abnormalities. The next lowest frequency was 31% for CNs IX/X. There is no compelling evidence to suggest that there is a significant epithelial/mesenchymal interaction, other that that seen in all of the cranial motor nerves as described above. CN III does have neural crest contribution, but only for the visceral efferent parasympathetic fibers of the ciliary ganglion [Sadler, 2000]. The very low frequency of involvement suggests that these nerves are involved coincidentally, or that they may exhibit abnormalities secondary to primary visual defects caused by abnormal formation of the optic vesicle.

Central nervous system (CNS) anomalies

While not directly related to discussion of the cranial nerves, CNS anomalies are an intermediate criterion in the AI/GEN model 2. An analysis of the CNS findings in CHARGE syndrome by Lin et al. [1990] showed that 55% of CHARGE patients had a structural CNS anomaly. Of the 26 patients studied, 17 had forebrain malformations (primarily arrhinencephaly (11), but also holoprosencephaly and other anomalies), 4 had hindbrain malformations (one of which was isolated agenesis of CN VII nucleus), 2 had a combination of forebrain and hindbrain malformations, and 3 had generalized findings (2 microcephaly and 1 lissencephaly). The pathogenesis of arrhinencephaly was discussed above. The pattern of malformations is not enlightening regarding the proposed hypothesis, given the complexity of brain embryogenesis.

Coloboma

This is a major criterion in both the traditional diagnostic model and the AI/GEN Model 2. Coloboma represents a failure of closure of the optic, or choroidal fissure that should take place around week 7 [Sadler, 2000]. The eye derives from four embryological germ layers: the neuroectoderm (optic vesicle), neural crest cells (anterior chamber), surface ectoderm (lens placode), and adjacent mesoderm [O'Rahilly, 1975]. The closure of the optic fissure is dependent on participation of both neuroectodermal and mesodermal cells [Warburg, 1993]. More recent studies suggest that coloboma is related to a defect in dorsal-ventral polarity in the optic vesicle [Uemonsa et al., 2002]. Several genes (Pax2, Tbx5, Ephb2, Efnb2, and Vax2) have been shown to play a role in dorsal-ventral axis determination in the mouse [Barbieri et al., 2002]. Mutations in PAX2 are associated with the renal-coloboma syndrome [Amiel et al., 2000] but have not been found in CHARGE syndrome patients [Tellier et al., 2000], even though renal anomalies are also present in CHARGE syndrome (see below). Targeted null mutation of Vax2 causes failure of closure of the optic fissure and leads to coloboma in homozygous mice [Barbieri et al., 2002]. In either case, it appears that disruption of an interaction between ectoderm and mesenchyme results in a coloboma.

Choanal atresia

Choanal atresia is another major criterion in both CHARGE diagnostic models. While the developmental sequence of formation of the choanae is known, as are the contributing tissues (surface ectoderm, facial mesenchyme), the actual embryologic error that leads to choanal atresia has not been well characterized. A number of theories have been proposed most of which implicate persistence of either the buccopharyngeal membrane, or the nasobuccal (or oronasal) membrane. While this mechanism may explain atresia of the membranous type [Hengerer and Strome, 1982], these authors present anatomic evidence that the remainder of choanal atresia is the result of misdirection of mesodermal elements of neural crest origin. It is known that the mesenchyme induces the ectoderm of the nasal placode to invaginate, leading to formation of the nasal pit. Hengerer and Strome [1982] speculate that if the mesenchyme migrates improperly, this leads to an abnormal relationship between the nasal sacs and the developing nasal vault, which may result in failure of communication with the nasopharynx. A primary abnormality involving ectodermal/mesenchymal signaling cannot be excluded, but it would be a tautology to use this possibility as supporting evidence for the hypothesis. There is currently no molecular characterization of the developmental events that lead to the formation of the choanae. Given current understanding, all that can be said with certainty is that ectoderm and mesenchyme are involved in formation of the choanae, but the nature of the interaction, and whether abnormalities of induction cause choanal atresia are unknown.

“CHARGE” ear

Characteristic external ear anomaly is a major criterion in the AI/GEN Model 2. The typical “CHARGE” ear has a short, wide pinna, a distinctive triangular concha and discontinuity between the antihelix and antitragus [Davenport et al., 1986]. The external ear develops from six mesenchymal proliferations at the dorsal ends of the 1st and 2nd pharyngeal arches, surrounding the 1st pharyngeal cleft. These later fuse and form the auricle [Sadler, 2000]. The normal development of the pharyngeal arches is dependent on complex molecular interactions between surface ectoderm, epithelium of endodermal origin and paraxial, lateral plate, and neural crest mesenchyme [Sadler, 2000; Graham and Smith, 2001]. It is tempting to conclude that disruption of an inductive interaction between surface ectoderm and the underlying mesenchyme results in the ear malformation, but the molecular genesis of the auricle has not been characterized at this time. Other abnormalities of pharyngeal arch development, particularly those dependent on the presence of neural crest mesenchyme, have been observed in CHARGE syndrome, and will be discussed in more detail next.

Neural crest related anomalies

The hypothesis that the neural crest plays an important role in several CHARGE malformations has been recognized for nearly 20 years [Siebert et al., 1985; Wright et al., 1986]. Abnormal interactions between the cephalic neural crest and the endoderm and ectoderm of the pharyngeal arches have been implicated in the development of the characteristic CHARGE facies, coloboma, choanal atresia, cleft palate (minor finding AI/GEN Model 2), and ear anomalies [Siebert et al., 1985; Wright et al., 1986; Sadler, 2000]. Analysis of the cardiovascular anomalies has shown a predominance of conotruncal malformations (intermediate finding AI/GEN Model 2) in CHARGE syndrome [Lin et al., 1987; Wyse et al., 1993]. There is ample embryologic evidence from the chick and rat, which supports the crucial inductive role of the neural crest in formation of the conotruncus [Sadler, 2000; Kochilas et al., 2002]. The neural crest has also been shown to be necessary for normal thymus formation [Lin et al., 1987; Sadler, 2000], explaining the DiGeorge anomaly that is a minor finding in the AI/GEN Model 2. Given the general acceptance of these mechanisms, further details will not be described in this study.

Tracheoesophageal fistula (TEF), esophageal atresia, and omphalocele

These are minor findings in the AI/GEN Model 2, but are seen with some frequency in CHARGE syndrome. In addition, there are infrequent reports of other foregut abnormalities [Buckfield et al., 1971; Siebert et al., 1985]. The association between TEF and other GI malformations has been noted [Andrassy and Mahour, 1979]. These anomalies are thought to be pathogenetically related to TEF [Qi et al., 2001]. The embryogenesis of the foregut, particularly with respect to developmental abnormalities, such as TEF, is controversial. There is at least some evidence to suggest that the splanchnic mesoderm adjacent to the foregut directs the differentiation of the adjacent endoderm [Sadler, 2000]. A rat model of TEF demonstrates that Adriamycin treated animals that develop TEF have less mesenchymal cellularity around the foregut than control animals [Sasaki et al., 2001]. The authors speculate that the maldevelopment of the trachea and esophagus of treated animals is due to an abnormal interaction between the foregut and the surrounding mesenchyme. Others studying the same model have noted abnormal pattern and timing of apoptosis of both endoderm and associated mesenchyme [Williams et al., 2000] that may also suggest an interaction between these tissues. Litingtung et al. [1998] present data that a failure in endodermal sonic hedgehog signaling prevents induction of target structures in the associated mesenchyme in Shh−/− mice. In a review article by de Santa Barbara et al. [2002], the authors concluded that current molecular evidence gives a central role to Gli2 and Gli3 transcription factors that are activated by endodermal Shh and have mesenchymal Foxf1 as one of their major targets. There is also emerging evidence that the neural crest may play a role in esophageal development. A recent study by Morini et al. [2001] analyzed the pattern of cardiac malformations associated with esophageal atresia. Over 70% of the malformations were conotruncal. In addition, Cozzi et al. [1993] recognized a similar prevalence of facial anomalies and maturational dysautonomia in patients with esophageal atresia, compared to a control group with choanal atresia. They conclude that esophageal atresia “should be considered a branchial arch neurocrystopathy” [sic].

Another interesting observation is the almost complete absence of mid- or hindgut malformations in CHARGE syndrome. If the hypothesis is accepted, this would suggest that the developmental control of mid- and hindgut development is significantly different than that of the foregut.

Omphalocele is a defect in the ventral body wall. It is related to defects in formation of the somatopleure [deVries, 1980]. The somatopleure develops at about 20 days gestation and is formed when lateral plate mesoderm fuses to the overlying ectoderm [deVries, 1980]. This is followed by lateral growth of somite-derived myotomes into the somatopleure [deVries, 1980]. Omphalocele can be thought of as persistence of the body stalk in the region normally occupied by the somatopleure caused by failure of development of the somatopleure. Recent molecular studies of somatopleure development in the chick demonstrate that ectoderm overlying the lateral plate both defines what will become somatic mesoderm, and is necessary to maintain this specification [Funayama et al., 1999]. The process appears to be modulated by BMPs [Funayama et al., 1999].

Pituitary abnormalities

AI/GEN Model 2 considers pituitary dysfunction an intermediate criterion. The pituitary gland develops from an invagination of oral ectoderm known as Rathke's pouch. This ectoderm is in contact with the ventral diencephalon, infundibulum, and surrounding mesenchyme. Each of these tissues secretes signaling molecules that promote growth and differentiation of the pituitary gland [Raetzman et al., 2002]. The ventral mesenchyme secretes BMP2 and BMP7, which oppose FGF8 and BMP4 secreted by the infundibulum. This interaction defines region-specific expression of the lim homeodomain transcription factors, LHX3 and ISL1 [Ericson et al., 1998] and is essential for normal pituitary development. Short stature is a minor finding in the AI/GEN Model 2. Growth hormone deficiency has been reported in some patients with CHARGE syndrome [August et al., 1983] but is not always present in CHARGE syndrome patients with short stature using standard provocative testing [Khadilkar et al., 1999]. The etiology of the short stature/growth failure in CHARGE is not known, although it is likely multifactorial, with growth hormone deficiency playing a role in some cases. Hypogonadotropic hypogonadism has been confirmed in a number of patients [Wheeler et al., 2000]. These authors speculate that this is due to hypothalamic abnormalities leading to deficiencies in GnRH, but these levels were not measured, nor was any other evidence provided to support this contention. The proposed pathogenetic hypothesis would suggest that abnormalities of the pituitary lead to impaired secretion of the gonadotropins. This portion of the hypothesis is testable, in that levels of GnRH could be determined in CHARGE patients with hypogonadotropic hypogonadism. Structural abnormalities of the pituitary gland have not been reported, although absence of the septum pellucidum was reported in two patients [Lin et al., 1990].

Urinary tract malformations

CHARGE urinary tract malformations are a minor finding in the AI/GEN Model 2. Epithelium of the ureteric bud from the mesonephros interacts with mesenchyme of the metanephric blastema [Sadler, 2000]. The mesenchyme expresses WT1 that regulates production of glial-derived neurotropic factor (GDNF) and hepatocyte growth factor (HGF) by the mesenchyme. The receptors RET (for GDNF) and MET (for HGF) are synthesized by the epithelium of the ureteric buds, which establishes signaling pathways between the two tissues. The buds induce the mesenchyme to proliferate by the secretion of FGF-2 and BMP-7.

Condition 2: Hypothesis Predicts Malformations That Should Not Be Seen

An exhaustive review of malformations not commonly encountered in CHARGE association is impossible, both due to the large numbers of malformations this would entail, as well as a lack of understanding of the role of inductive mechanisms in many if not most of these malformations. Malformations that are thought to be derived primarily from a single embryologic tissue (ectodermal dysplasias, endodermal-derived organs such as the liver, pancreas and gut and striated muscle and dermis derived from the paraxial mesoderm) have not been reported in CHARGE syndrome with any frequency.

Condition 3: Hypothesis Predicts Malformations not Previously Recognized

This heuristic condition is difficult to fulfill, given the diligence with which the malformations in CHARGE syndrome have been studied. Limb anomalies would fulfill this condition. Neither the AI/GEN Model 2, nor the acronym-based Model 1 recognized limb anomalies as a diagnostic criterion for CHARGE syndrome. However, a review of published descriptions of CHARGE syndrome shows that it may be present in as many as 33% of individuals that meet the diagnostic criteria [Williams and Rooney, 1996; Brock et al., 2003]. The limb is one of the best-studied models of ectoderm-mesenchyme induction (a comprehensive review of the molecular embryology of the limb was recently published by Gurrieri et al. [2002] and will not be recapitulated here). Therefore, limb anomalies would be expected if the proposed hypothesis were correct. Formal testing of other functions of the cranial nerves such as smell (CN-I), facial sensation and strength of muscles of mastication (CN-V), and taste (CN-VII and IX) could confirm involvement of these cranial nerves, as the hypothesis would predict.

Current Status of Molecular Investigation of CHARGE Syndrome

Several investigators have begun to look for evidence of a molecular pathogenetic mechanism for CHARGE syndrome. As previously noted, there are several reports of patients with chromosomal abnormalities that meet the diagnostic criteria for CHARGE syndrome [Dev et al., 1985; Clementi et al., 1991; Hurst et al., 1991; North et al., 1995; Devriendt et al., 1998; De Krijger et al., 1999; Lev et al., 2000; Martin et al., 2001]. This has led to speculation that CHARGE syndrome may be associated with chromosomal microdeletions [Clementi et al., 1991]. Lalani et al. [2001] did a genome-wide screen for loss of heterozygosity (LOH) at a 10 cM resolution in 55 CHARGE syndrome case/parent trios. No deletions were identified. A higher resolution screen (5 cM) done in 10 case/parent trios ruled out a 2 Mb deletion in about 40% of the genome [Lalani et al., 2002]. Sanlaville et al. [2002] used comparative genomic hybridization on 27 CHARGE syndrome patients. They identified two cryptic chromosomal anomalies; der(9)t(9;13) derived from a paternal translocation and der(6)t(4;6) of unknown origin. These studies suggest that chromosomal rearrangements are causative in at least some CHARGE syndrome patients. Analysis of the involved loci may identify genes involved in the pathogenesis of CHARGE syndrome.

Other groups have begun to look for mutations in candidate genes that are expressed in developing tissues involved in CHARGE syndrome malformations. Tellier et al. [2000] analyzed the PAX2 gene in 34 CHARGE syndrome patients. While some polymorphisms were identified, no mutations were found. Yahagi et al. [2002] analyzed TBX1, HOXA3, HOXB3, and HOXD3 in 19 patients with CHARGE syndrome. No mutations were identified.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ADDED NOTES
  8. REFERENCES

The proposed hypothesis fulfills the three test criteria within the current limits of the known embryologic pathogenesis of CHARGE syndrome malformations. It represents the only pathogenetic hypothesis that explains all of the commonly observed malformations in CHARGE syndrome. As more definitive information on embryogenesis is obtained for specific malformations, the hypothesis will need to be re-examined to confirm that the data still supports the contention that the malformations are due to abnormalities in epithelial–mesenchymal interactions.

The negative findings of the gene analyses done in CHARGE patients to this point do not represent a problem to the hypothesis, due to the small number of patients and genes examined to this point. As is evident in the molecular embryology presented in the body of the study, there are a large number of genes potentially involved in embryogenesis of the relevant tissues, and the ones that have been studied cannot even be said to represent the most likely candidate genes available. The resolution of the microdeletion studies reported is not sufficient, nor is the coverage of the genome adequate to rule out microdeletion as a potential contributor to CHARGE syndrome.

The evidence presented in the introduction argues strongly that there is a genetic cause for CHARGE syndrome. The author would propose that it is likely that more than one gene mutation may be necessary to cause CHARGE syndrome. There are several lines of evidence that support this. Tellier et al. [1998] note that transmission of at least one CHARGE malformation was observed in 20% of families with a CHARGE syndrome patient. These include two families where the parent of a CHARGE syndrome patient had a conotruncal heart malformation, and seven families in which 1 or 2 CHARGE syndrome findings were present in 1st cousins, or more distant maternal relatives. Reports of familial CHARGE syndrome are difficult to interpret, in some cases, due to lack of sufficient information to confirm the diagnosis [Awrich et al., 1982; Brown and Israel, 1991; Tellier et al., 1998]. (It is of interest to note that three of these four patients had the combination of coloboma and/or microphthalmia, hearing loss, and patent ductus arteriosus along with mental retardation and genital hypoplasia in males, but no choanal atresia. While this could be fitted under the diagnosis of CHARGE syndrome, this could represent a separate condition.) In other cases, one sibling clearly meets the criteria, while another has some of the features, but not enough to meet the strict diagnostic criteria [Pagon et al., 1981], which may be analogous to the situation reported in relatives by Tellier et al. [1998]. The family with presumed autosomal dominant transmission reported by Mitchell et al. [1985a] has generally milder features, and choanal atresia and heart defects are not present. The mother and son reported by Metlay et al. [1987] would both meet the diagnostic criteria for CHARGE syndrome, and both had unilateral choanal atresia. The mother did have normal intellectual development, while the son died at 19 months. No cases of discordance in monozygous twins have been reported. These case reports would suggest that the presence of a single gene mutation may lead to development of some malformations seen in CHARGE syndrome, but the presence of at least one other mutation in another critical gene is necessary for CHARGE syndrome to be fully manifest.

There is also information that suggests that critical developmental pathways have built in redundancy that minimizes risk to the developing embryo. This concept has emerged from the study of familial holoprosencephaly, in which extreme variability of expression and high frequency of nonpenetrance are observed. This has led some to speculate that a single gene mutation may not be sufficient to create the severe holoprosencephaly phenotype [Wallis and Muenke, 2000]. While major genes have been identified in holoprosencephaly, the contribution of mutations or polymorphisms in other genes, and/or environmental influences have not been elucidated to this point. A similar phenomenon may explain the observations in CHARGE syndrome. The reports of several patients with different chromosome anomalies that meet the diagnosis of CHARGE syndrome would also support a multiple gene hypothesis.

Until the molecular basis of embryogenesis is well understood, no hypothesis regarding the pathogenesis of CHARGE syndrome will be able to be proved. However, it is only through the careful study of patterns of abnormal morphogenesis that insights into the common pathways of normal embryogenesis will be identified. All the more reason to “treasure our exceptions.”

ADDED NOTES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ADDED NOTES
  8. REFERENCES

Since the acceptance of this study, Vissers et al. [2004] have demonstrated deletions or mutations of the CHD7 gene in more than 50% of CHARGE patients analyzed. While the identification of an important single gene would seem to be at odds with the author's contention that more than one gene may be involved, examination of the putative function of CHD7 would confirm multigenic pathogenesis. CHD7 is a member of the chromodomain helicase DNA-binding (CHD) genes. This class of proteins is known to regulate gene expression by altering chromatin structure. Many of these proteins play a critical role in the embryogenesis of other animals by controlling expression or repression of sets of developmental genes by establishing stable epigenetic structural changes that can be maintained throughout subsequent cell division and differentiation [Cavalli and Paro, 1998]. As an example, Polycomb-M33-deficient mice show poor growth, early mortality, multiple skeletal anomalies, and proliferation defects in several cell types [Coré et al., 1997]. The skeletal malformations are homeotic in nature and imply dysregulation of Hox genes, although the authors were able to demonstrate abnormal localization of expression for Hoxa-3 only. These changes are similar to those seen in null mutants for mel-18 and bmi-1, two other members of the Polycomb group genes [van der Lugt et al., 1994; Alkema et al., 1995; Akasaka et al., 1996]. This implies potential interaction between these genes in the regulation of Hox complexes in mice. Interaction between CHD genes is supported by increasing severity of anomalies seen in Drosophila mutant for two or more polycomb genes [Jürgens, 1985], as well as observed interaction of XPOLYCOMB and XBMI-1 proteins in Xenopus [Reijnen et al., 1995].

There is currently no evidence for direct involvement of CHD proteins in induction effects such as described in this study. As such, no conclusions can be drawn as to whether this discovery disproves the proposed mesenchymal–epithelial interaction hypothesis. It should be noted that Coré et al. [1997] conclude the M33 protein, establishes the pattern in the mesoderm that defines subsequent inductive events.

While Vissers' report represents an important step forward in our understanding of CHARGE syndrome, the evidence presented does not allow the conclusion that haploinsufficiency of CHD7 is sufficient by itself to cause the CHARGE phenotype, or if interaction with mutations or polymorphisms in other genes affects the variability and penetrance of the phenotype. While expression of CHD7 cDNA was seen in tissues commonly involved in CHARGE syndrome, some expression was detected in all tissues and highest expression was noted in the brain, kidney, and skeletal muscle [Nagase et al., 2000]. Abnormalities of skeletal muscle have not consistently been noted in CHARGE syndrome [Brock et al., 2003]. Additionally, 7 of the 17 patients tested did not have identifiable mutations in the CHD7 gene suggesting that CHARGE syndrome is causally heterogenous.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. ADDED NOTES
  8. REFERENCES