Statistics from Altmetric.com
Congenital malformations or complex malformation syndromes are frequently associated with growth failure and have been the subject of much research and discussion in the paediatric literature. The less common overgrowth syndromes (OGSs) have until recently received little attention. The disordered growth in OGSs is, however, a primary anomaly and, unlike growth failure, is not explained away as a secondary phenomenon as is the case with many other complex syndromes. OGSs may therefore provide a fascinating window into the mechanisms of growth and the consequences of the failure of this regulation.
Ancient literature has many references to giants such as Goliath, Polyphemus, Gargantua, or the Patagonian giants. Whether real or fictional, these reports show that such patterns have been present throughout history and serve to highlight two of the central issues—what is a “true overgrowth syndrome and how many overgrowth syndromes exist?”
Previously, overgrowth patterns were often categorised as primary or secondary. In primary disorders, the growth would be an intrinsic (unexplained) feature of the condition secondary to cellular hyperplasia, whereas in secondary disorders an identifiable cause, often endocrinological, would be expected to result in growth excess.1 The limitations of this rather simplistic differentiation have been illustrated by the identification of “novel” growth factors in a number of OGSs, such as Beckwith-Wiedemann syndrome (BWS) and Simpson-Golabi-Behmel syndrome (SGBS). In these disorders, abnormalities of insulin-like growth factor II (IGF II) and glypican 3 have been implicated.2 3 It would seem that if the term secondary growth excess is still relevant, it should be limited to situations dependent on extrinsic growth promoters, such as fetal macrosomia secondary to maternal diabetes and hyperglycaemia and subsequent fetal hyperinsulinaemia. For the foreseeable future most “primary” OGSs will be classified by a process of clinical assessment and/or laboratory exclusion with the possible exceptions of BWS and SGBS, where clinical application of molecular tests might be feasible within a few years.
The constraints of this review preclude the discussion of another group of overgrowth patterns: those syndromes exhibiting regional/tissue specific overgrowth (table 1). These may provide further clues to specific growth promoters, and will have to be considered when developing a model of overall growth control.
Endocrinological and metabolic disorders with an overgrowth component
Abnormalities of the hypothalamic pituitary adrenal axis may result in early overgrowth. A number of these are autosomal recessive and the most common is congenital adrenal hyperplasia due to 21-hydroxylase deficiency. In untreated cases, the production of excess 7-hydroxyprogesterone and androgenic steroids results in increased growth from birth and precocious puberty secondary to the anabolic effects of the steroids.4 Secretory tumours within this axis may also result in similar overgrowth patterns. A much rarer recessive adrenal disorder is adrenocorticotrophic hormone receptor deficiency. This may present with symptoms suggestive of adrenal failure or overgrowth, or both, and prompt treatment of the adrenal insufficiency is of obvious importance.5
The aetiology of the overgrowth and endocrinological confirmation of each of the above is clear. This has not always been the case for disorders with proved endocrinological abnormalities, for example, Seip-Berardinelli lipodystrophy syndrome. This autosomal recessive condition appears to show insulin resistance, but normal pituitary and adrenal function on formal testing. The growth excess can be striking (considerably more than 97th centile), and appears to be more than would be expected from the postulated anabolic component of the syndrome. After some 30 years of research, the mechanism of overgrowth and its link to insulin resistance is only now becoming clearer.6
One further condition that may rarely present with overgrowth, which is important not to miss because of the neurological and genetic implications, is the recessive condition, mucopolysccharidosis type III (Sanfilippo’s syndrome).7 Early literature indicates that the excess growth is only apparent in the first two to three years of life, and that the regression and mental handicap is obvious by this stage. This is in fact frequently incorrect, and I have personal experience of a 6 year old patient over the 90th centile for all growth measurements, who was then only starting to show a decline of intellectual abilities into the range of mild mental handicap.
Congenital malformation syndromes with overgrowth as the major intrinsic component
The majority of conditions with “primary” overgrowth are poorly understood but are part of an ever increasing list of possible diagnoses. The current London dysmorphology database lists 283 conditions with either macrocephaly, obesity, increased birth weight, or excess stature.8 Personal experience in excess of 300 children, is that as many as 50% do not easily sit within the diagnostic categories currently recognised.
Any diagnostic assessment must include taking an adequate family history and measurements of the nuclear family, as the largest single factor is without doubt familial large stature or early maturation. The presence of some additional sign or symptom within an individual does not necessarily require a syndrome diagnosis because single malformations, or some degree of learning difficulty, is present in up to 10% of the population.9
All possible diagnostic considerations will not be covered within this annotation, however the reader is referred to the excellent review of Cohen.1 Tables 1 and 2 (modified from Cole and Hughes10) also list some of the range of regional and generalised overgrowth disorders.
BWS and SGBS—two OGSs but a common aetiological pathway
BWS, first described in 1963,11 has provided much of the recent impetus for research into overgrowth patterns. It has a reported frequency of one in 13 700 from a study in Jamaica.12 I would, however, be concerned of the influence of the high frequency of umbilical hernias (also known to be associated with an increased familial frequency of exomphalos) in this population.13 This might have led to over diagnosis, or might indicate that the population as a whole had a higher than average frequency of some of the predisposing mechanisms of BWS.
In BWS, the three main components are abdominal wall defects (exomphalos=E), macroglossia (M), and increased birth weight or growth (G) present in 80%, 88%, and 97% respectively, resulting in the alternative name EMG.14 The paediatric relevance of BWS is maintained because approximately 4–7.5% of patients develop childhood tumours, most commonly Wilms’ tumours.1 14Hemihypertrophy, present in 13–24% of patients with BWS,1 14 is reported in 40% of cases with tumours,1 therefore the presence of hemihypertrophy appears to be associated with an increased risk of tumourigenesis.
Growth in BWS is usually most marked during the first few years of life. This is compatible with current evidence that IGF II, the likely growth factor in BWS, has its greatest influence during this period. By contrast, final height in many individuals with BWS is within the normal range, although there are exceptions. There is even a suggestion of a bimodal pattern, with a small group showing excessive height. It will be of interest to see if specific molecular mechanisms are associated with different growth patterns.
Despite neonatal symptoms, which can include feeding and respiratory difficulties and very rarely neonatal death, the large majority of survivors have few severe non-malignant health problems. Intellectual development is believed to show the same distribution as the general population, with the possible exception of patients with extreme or untreated hypoglycaemia or with unbalanced chromosomal translocations.
Of particular genetic interest are the mechanisms underlying BWS. It is estimated that approximately 15% of families show autosomal dominant inheritance,14 and these show linkage to the chromosome region 11pl5.5, including the insulin IGF II gene loci.15Further studies, reviewed by Mannens et al, showed that paternal duplications, maternal translocations, and paternal uniparental disomy (the presence of only paternal chromosome material) involving 11pl5.5 result in BWS.16 There is evidence that the phenotype results from overexpression of IGF II, a gene usually only expressed on chromosomes of paternal origin.2 Recent papers have shown that mutations within p57KIP2,17 or disruption of the gene KVLQT1,18 which both map proximal to IGF II at 11pl5.5, will result in the BWS phenotype. In both instances the maternally inherited gene is affected. At least one family with a translocation disrupting KVQLT1 had biallic IGF II expression, while p57KIP2 may be involved in a common pathway with IGF II. A p57KIP2 gene “knock-out” mouse,19 and subsequent mutation studies of KIP2 in humans, also results in a BWS phenotype, but once again only if the maternal allele is affected.20 There also appears to be some early evidence for genotype/phenotype correlations with p57KIP2 mutation carriers having a high frequency of abdominal wall defects (five out of seven had exomphalos, one out of seven had an umbilical hernia) but no embryonal tumours,20 whereas uniparental disomy 11pl5.5 may be associated with a slightly greater risk of tumourigenesis.16
This phenomenon of imprinting (the regulation of gene expression dependent on parent of origin) is believed to be important in numerous different conditions with disordered growth including Prader-Willi syndrome, Angelman’s syndrome, Russell-Silver syndrome, and certain tumours, and has therefore been the subject of intense investigation.21 22
During research into BWS, clinicians noted the striking phenotypic overlap with SGBS, an X linked overgrowth disorder (table3).23 24 In the report of Thorburn et al the diagnosis of SGBS in some, or all, of the three boys (all of whom died in infancy and had additional malformations recognised in SGBS) remains a possibility, and is one further source of error which could result in overestimation of the incidence in BWS in their paper.12
After the identification of glypican 3 mutations as the cause of SGBS,3 subsequent antibody studies showed there was cross reaction of binding between the ligands for glypican and IGF II receptors.3 25 This in turn might explain the striking phenotypic overlap and also raised the possibility of further diagnostic errors in the literature or clinical setting. The latter has significant implications for genetic counselling.
An additional consideration is the existence of an undiagnosed group of patients, as reported in the paper by Morrison et al.26 They described “non”-BWS patients with overexpression of IGF II. I would suggest that the presence of overgrowth and nephromegaly in these patients might represent the mild end of the spectrum of BWS. An alternative explanation is that many macrosomic babies have some underlying mechanism which results in abnormalities of IGF II expression, the BWS phenotype being just one of the resultant outcomes. If the outcome is not the BWS phenotype, it raises the fascinating possibility that overexpression of IGF II might explain the excess of neuroblastoma and Wilms’ tumours in babies weighing over 4000 g.27 As we start to unravel the mechanisms behind BWS and SGBS it appears that a more intriguing set of questions are raised.
While molecular studies for glypican 3 and IGF II are still at an early stage, it seems likely that within the next 2–3 years these will become part of the routine investigation for BWS and SGBS in an effort to confirm the diagnosis and the recurrence risk, and perhaps even help predict the individual natural histories.
Sotos’ syndrome (SS), Weaver-Smith syndrome (WSS), and other overlapping phenotypes
My own experience, and that within our department, is that referrals of SS and BWS are approximately equal although this figure may be biased by my own research interests. However, it is probable that SS is at least the second most common overgrowth disorder after BWS.
SS was first reported in 1964,28 although I suspect the first literature case may be the report of Schlesinger in 1932.29 The clinical features are of prenatal and postnatal overgrowth, developmental delay, characteristic facial appearance, and an advanced bone age. Apart from the subjective assessment of facial gestalt (table 4) and relative frequencies of developmental delay, the remaining diagnostic criteria are not only features of BWS and SGBS, but also the WSS and Marshall-Smith syndrome (MSS).30 31 Furthermore, a number of less frequent signs such as hernias, minor skeletal anomalies, and tumour predisposition are probably present to varying degrees in each of the above disorders. Currently the majority of opinion splits SS, WSS, and MSS into three distinct syndromes.
The growth in SS is most marked during the first 4–5 years of childhood, although may be “masked” during the first 36 months of life if complications secondary to prematurity, hypotonia, poor feeding, or respiratory difficulties are present. The two growth variables most significantly increased are length and occipitofrontal circumference, whereas weight is often lower than would be expected for the patients stature.32 Final height is commonly below or close to the +2 SD. This appears to be particularly true of girls whose growth often stops earlier than their peers after a relatively early puberty. This pattern appears more variable in boys, but still very few exceed or even approach + 3 SD (Agwu et al in preparation). It is the authors experience that while careful monitoring of growth, bone age, and pubertal status is advisable, therapeutic intervention is almost never necessary to obtain an “acceptable” final height.
SS was originally incorrectly reported to have a high frequency of mental handicap (88%).33 It is now clear that the figure is lower, and Finegan et al reported that 78% of children have an IQ above 70.31 This still remains one of the highest figures for OGSs and is often associated with significant behavioural problems.34 Other significant medical complications are individually uncommon.32
Despite the similarities with WSS and MSS, these two conditions appear to exhibit higher frequencies of severe medical complications, including skeletal anomalies and infant death. Surprisingly, personal experience of the three older cases of WSS (two adults and one adolescent) and the two oldest literature cases of MSS is that intellectual abilities have been within, or near to, the normal range in all these individuals.35 This might be fortuitous, as significant intellectual impairment has been documented in both disorders.
The range of final heights in WSS remains unclear, however three literature reports of adults with WSS, and my own experience of two adult patients, may suggest that heights, significantly in excess of +2 SD, are more common than in SS.30 32 It should be stressed that ascertainment bias could be very relevant in four out of five of these cases.
The genetics of OGSs
The genetics of BWS is that the majority are sporadic, but approximately 15% follow autosomal dominant inheritance. In my experience of approximately 200 cases of SS, three cases, and one of 15 cases of WSS segregated in an autosomal dominant fashion,30 the remainder are sporadic. Unlike BWS, to date there has been no consistent region of chromosomal alteration in the other OGSs, and a study of uniparental disomy in SS was negative within the limitations of the study.37 The genetics therefore remain obscure in the majority of OGSs.
The clinical similarities among the OGSs should intrigue, yet alert, the researcher to the issue of whether the OGSs should be “lumped” or “split”. It seems certain that only as the molecular basis unfolds will it become clear if the disorders are differing ends of the same spectrum or separate conditions. Could it be that they will be allelic variations in the same gene, such as in the RET oncogene which results in the different multiple endocrine neoplasia II syndromes and Hirschsprung’s disease,38 or mutations in the same fibroblast growth factor gene associated with the different craniosynostosis syndromes such as Apert’s, Pfeiffer’s, Crouzon’s, and Jackson-Weiss?39 Or perhaps the different OGSs will be due to mutations within a group of genes like the fibroblast growth factor I, II, and III genes, each associated with syndromes including the feature of craniosynostosis,39 or the different fibrillin genes which are the cause of Marfan’s syndrome (fib 1),40 congenital contractural arachnodactyly (fib 2),41 and Shprintzen-Goldberg (fib 1).42While these are possible mechanisms, the only precedent in OGSs is in BWS and SGBS. In these two conditions, none of the above mechanisms apply, but rather two different genes interact on the same pathway.25
Different overgrowth patterns—different genetic mechanisms
AUTOSOMAL DOMINANT MACROCEPHALIES
The autosomal dominant macrocephalies are one further overgrowth spectrum which appear to show a perplexing degree of pleiotropy and heterogeneity.43 They may also overlap with intriguing sporadic macrocephaly conditions such as the macrocephaly, telangiectasia, and cutis mamorata syndrome. Unravelling the mechanisms behind features associated with macrocephaly, such as vascular anomalies43 or autism,44 could be a further fruitful area of research related to overgrowth disorders.
This text so far has only considered conditions that I believe to be non-mosaic mutations. Disorders such as Proteus’ syndrome and McCune-Albright syndrome, are most likely, or have already been proved to be, due to somatic mutations, and if present in a non-mosaic syndrome it is postulated that they would be lethal in most situations.
Molecular investigations of patients with McCune-Albright syndrome have confirmed the presence of mosaic mutations in Gsα gene.45 Patients with McCune-Albright syndrome show quite remarkable early overgrowth. Much of this appears to be secondary to the effects of a strikingly early puberty. It is clear, however, that the G proteins are a family of signal transducers and have a regulatory role on many different endocrinological and growth factor pathways. The growth and endocrine role and the recognition of de novo mutations makes this group of genes interesting potential candidates for OGSs such as SS and WSS.
OGSs offer the opportunity to study prenatal and postnatal growth in normal tissues, and may show how abnormal growth may predispose to tumour development. It is also of great interest that many of the genes involved in growth disorders are regulated and (for example, imprinted) dependent upon their parent of origin, which may be highly significant in fetal growth.
The recognition of new OGSs is still occurring and yet as many as 50% of cases do not appear to fall neatly into any recognised patterns. However, the aetiology of the clinical overlap between diagnosed and currently undiagnosed syndromes may become apparent as the underlying genetic mechanisms become clearer. As this molecular understanding remains incomplete, it is important to try and identify on clinical grounds the specific “syndrome pattern” where possible, as the natural history and genetics may vary considerably.
The author would like to acknowledge the support provided by Action Research which enabled collection of much of his data on overgrowth syndromes and stimulated a lasting interest and Professor Eamonn Maher for his many helpful comments on the annotation.