Clinical manifestations

Presentation and course

Prader-Willi syndrome is a sporadic condition characterized by neonatal hypotonia, hypogonadism, genital abnormalities, hyperphagia, obesity, developmental delay, characteristic facies, small hands and feet (acromicria), and a cytogenetic or molecular abnormality of the long arm of chromosome 15. It is the most common genetic cause of marked obesity (18). Diagnostic criteria based on the age of the patient and on symptoms have been published (84).

Classically, the course of Prader-Willi syndrome is divided into 2 stages. The first stage is characterized by neonatal hypotonia, temperature instability, hypogenitalism, and feeding difficulties. The second stage begins after the first year of life and is marked by the onset of hyperphagia with obesity, behavioral problems, and developmental delay. Seven distinct nutritional phases have been described, with increased appetite occurring in phase 2b (4.5 to 8 years), followed by classic hyperphagia in phase 3 (122).

Prenatal characteristics include decreased fetal movements (reported in 27% to 83% of pregnancies), preterm birth (reported in 15% to 30% of pregnancies), and high rates of Caesarian section (42% to 67%) (04; 71). Birth asphyxia, in which newborns required assistive breathing, occurred in 23% of deliveries (174). Prader-Willi syndrome commonly presents in the newborn period with profound hypotonia. The infant is floppy with little spontaneous movement and a weak cry; dyskinetic movements are sometimes noted (118). Sucking and feeding difficulties have been reported in 85% to 97% of newborns, and 50% of children require tube feedings because of poor weight gain (04; 71). Some infants have severe central hypoventilation and require mechanical ventilation for days to weeks (107).

The hypotonia is nonprogressive, central in origin, and improves between 9 to 12 months of age. Motor milestones, in spite of the improving tone, are often delayed, with independent sitting occurring at 13 months, crawling at 16 months, and walking at 28 months (18; 84).

Hypogonadism and hypogenitalism are identified early in boys but with difficulty in girls. Micropenis, scrotal hypoplasia, and cryptorchism are seen in 80% to 100% of boys in the neonatal period (18). Girls may be normal or manifest labial hypoplasia. The cause of hypogonadism in males has long been thought to be of hypothalamic origin, but evidence suggests that the majority of males with Prader-Willi syndrome have primary testicular failure (155).

Temperature instability is a less common feature of the first stage but can be problematic when it occurs, resulting in multiple hospitalizations and extensive diagnostic evaluations. Both hyperthermia and hypothermia have been reported, and several children have presented cyclic elevations in temperature without any other symptoms and no cause found (150).

Many infants with Prader-Willi syndrome have swallowing dysfunction, including silent aspiration events; all infants with Prader-Willi syndrome should undergo comprehensive feeding and swallowing evaluations (149).

Obesity is a major feature of Prader-Willi syndrome and marks the onset of the second stage of the condition (18; 84). Nearly all patients manifest marked obesity without intervention, and one third of Prader-Willi syndrome patients weigh more than 200% of ideal body weight (18). The onset of this phase occurs after the first year and certainly before 6 years of age (84). It results from hyperphagia with persistent uncontrolled hunger. Food foraging and hoarding, theft of food, pica, and other behaviors indicative of food obsession develop. The combination of excessive caloric intake and decreased physical activity results in obesity.

Around the time that hyperphagia develops, other behavioral and cognitive problems become evident. Children show delayed motor, cognitive, and language development. The average intelligence quotient is in the 60s. In a meta-analysis of psychological testing results in Prader-Willi syndrome from 13 published studies, Yang and colleagues reported that 15q11-q13 deletion subjects have a significantly lower full-scale IQ and verbal IQ than subjects with maternal uniparental disomy (179). In contrast, uniparental disomy subjects have lower performance IQs and are at greater risk for psychiatric illness. Many of the individuals with low to normal intelligence have a scattering of abilities consistent with learning disabilities (161; 42; 61).

Language function is particularly impaired in children with Prader-Willi syndrome, out of proportion to verbal intelligence levels (54). Dimitropoulos and colleagues also reported that deletion subjects in their cohort had equal degrees of impairment in expressive and receptive language, whereas uniparental disomy subjects had higher expressive language function than receptive (54). In contrast, visual spatial processing may be a particular strength. Individuals with Prader-Willi syndrome have an unusual skill with jigsaw puzzles (32).

Behavior problems include temper tantrums, violent outbursts, and obsessive-compulsive behaviors. Patients have been described as argumentative, oppositional, rigid, manipulative, and perseverative; difficulties with social interaction are common. Indeed, as summarized in a meta-analysis, 26.7% of individuals with Prader-Willi syndrome meet the criteria for autism spectrum disorder (07). Notwithstanding this result, Dykens and colleagues have argued that the true incidence of autism spectrum in individuals with Prader-Willi syndrome is probably not that high largely because, in most studies, the diagnosis of autism spectrum has been based on data accrued from third-party questionnaires, especially from parents (64). In their study of 146 children with Prader-Willi syndrome aged four to 21, extensive videotaped testing centered on the Autism Diagnostic Observation Schedule-2 (ADOS-2) and performed by an expert clinical team yielded a diagnosis of autism spectrum disorder in 18 (12.3%) out of 146 patients.

One important reason for the presence of autistic behaviors in Prader-Willi syndrome patients may be a significant deficit (in comparison with the normal population) in the ability to recognize emotional states from an analysis of facial expression (176). Similar deficits in facial expression analysis in Prader-Willi syndrome were observed using event-related potential techniques by Key and colleagues (90). In a later study by the same group again using event-related potentials, Dykens and colleagues found that participants with uniparental disomy (n=13) were unable to recognize familiar faces, whereas those with deletions (n=15) had no such difficulty (64). Given that the capacity to recognize faces as familiar is crucial to normal social interaction, these observations may explain why the percentage of Prader-Willi syndrome individuals with uniparental disomy who meet the formal criteria for autism spectrum is twice as high as it is in those with deletions (07).

Self-injury, especially skin picking, is commonly reported and may be secondary to pain insensitivity (84). Rectal picking may lead to rectal ulceration and bleeding, and if unrecognized, unnecessary testing and interventions (148). Maladaptive and obsessive-compulsive behaviors have been reported to occur more frequently in patients with deletions (particularly large deletions) than in those with uniparental disomy (20).

In adolescence or early adult life, a recurrent, rapid-cycling psychosis may develop (172). Episodes of psychosis may be preceded by transient, acute somatic symptoms such as pain, flu-like symptoms, diarrhea, or urinary frequency. Cumulative evidence strongly suggests that psychotic features are more likely to develop in cases with uniparental disomy than in those with deletions (62; 172; 156; 179). Later in life, for those who have developed cyclical psychotic episodes, satisfactory control of symptoms is typically established with antipsychotic medications, and the tendency to have further psychotic episodes markedly declines (99).

In comparison with Angelman syndrome, unprovoked epileptic seizures are relatively uncommon in Prader-Willi syndrome, but, nonetheless, are seen far more frequently than in the general population. Reported incidence rates have varied from 4% to 6%, 18%, and 26% (67; 169; 70; 164). Seizure patterns may be generalized or focal. Reported series have varied widely with respect to relative frequencies of generalized versus focal seizure patterns. Percentages of cohorts with generalized seizures range from 8% (169) to 60% (67). It is likely that the ratio of generalized to focal seizure patterns in the largest cohort reported to date (n=38; 55.2% generalized, 44.8% focal) comes closer to representing a more accurate estimate for this disorder (170). In the series by Fan and colleagues, 6 of the 10 epileptic patients had generalized seizures (67). In contrast, focal-onset seizures predominated in the series published by Vendrame and colleagues, Takeshita and colleagues, and Gilboa and Gross-Tsur: 92%, 77.8%, and 80%, respectively (169; 70; 164). In general, although the number of subjects in the four reported series was too small to permit any statistical validity, unprovoked epileptic seizures are seen more often in deletion patients than in those with uniparental disomy. Febrile seizures also occur more frequently than in the general population (167; 169; 164; 170).

Excessive daytime sleepiness frequently occurs in Prader-Willi syndrome. Some patients have sleep-disordered breathing associated with severe obesity, whereas others have narcoleptic features with sleep-onset REM sleep. In the majority of individuals, however, sleep studies suggest that the hypersomnolence results from primary hypothalamic dysfunction (112). Central and obstructive sleep apnea are frequently encountered in persons with Prader-Will syndrome, occurring in over half (33; 134). Central apnea is more common in younger patients, whereas obstructive apnea is more common in older children (33). Although obesity undoubtedly plays a role in the pathogenesis of obstructive sleep apnea, on the whole, there is a poor correlation between obstructive sleep apnea and body mass index (134).

Parents of Prader-Willi syndrome patients have long reported that their children appear to have decreased sensitivity to pain (84). This suspicion was confirmed by Priano and colleagues using a quantitative measure of sensory thresholds to different modalities in 14 adult Prader-Willi subjects (137). The authors found significant elevations in heat and pain thresholds in comparison with nondiabetic obese subjects and age-matched nonobese controls.

Physical features include dolichocephaly in infancy, a narrow face, almond-shaped eyes, and a small-appearing mouth with a thin upper lip and downturned corners (84). No clear differences in facial appearance have been noted between subjects with deletions and those with uniparental disomy (167). Thick saliva, crusted at the corners of the mouth, has been noted. Hypopigmentation with fair skin and hair is often seen in individuals with deletions. The presence of small hands and feet is frequently noted, but the usefulness of this feature as a diagnostic criterion has been difficult to assess (23; 21; 88; 84). A typical metacarpophalangeal pattern profile has been noted (24). The hands are narrow with straight ulnar borders.

A variety of orthopedic problems have been reported; the most prominent are scoliosis, kyphosis, flat feet, knock knees, and hip dysplasia (173). Osteoporosis at all ages has been reported and can result in fractures following minor trauma (84). Bone mineral density is typically reduced, apparently due to increased bone turnover (171).

Later in childhood, a variety of endocrinologic deficiencies have been noted, all seemingly of hypothalamic origin. First, delayed and incomplete sexual maturation is seen. In adults, 92% of males had a small penis, and 88% had undescended testicles. One half of men had some facial hair, but two thirds had little or no other body hair. Sixty percent of women with Prader-Willi syndrome reported primary amenorrhea. Menarche, when it occurred, was often late; the average onset was at 17 years of age. Sixty-four percent of women had an abnormally small amount of breast tissue, whereas the rest had a normal amount of breast tissue (73).

Second, short stature, although not apparent at birth, becomes manifest postnatally. This may not be a striking feature until the second decade when there is a failure of the pubertal growth spurt (23). The majority of patients with Prader Willi syndrome show reduced growth hormone, or insulin growth factor 1 (IGF-1) levels, or both. Prevalence studies suggest growth hormone deficiency is present in 75% to 85% of all children and IGF-1 values decreased in virtually all patients (53; 75).

Clinical hypothyroidism has also been reported and is present in 24.4% of Prader-Willi patients (53).

Finally, there is some evidence that adrenocortical insufficiency may be an important feature of Prader-Willi syndrome. Using a metyrapone suppression test, de Lind van Wijngaarden and colleagues found that the majority of their subjects had an impaired ability to generate ACTH (49). In another study by the same group, patients with central adrenal insufficiency tended to have an increased frequency of sleep-related breathing problems in comparison with those having normal ACTH production (48). In contrast, Connell and colleagues measured serum cortisol responses to insulin-induced hypoglycemia and found no differences in subsequent cortisol peak levels between Prader-Willi patients and the reference range (34). Head-to-head comparison of these 2 study groups is rendered difficult by the differences in methodology. Nevertheless, it is conceivable that central adrenal insufficiency may help explain the high rate of unexpected death in Prader-Willi syndrome, especially in the context of infection-related stress.

Prognosis and complications

Overall mortality rates in children and young adults are much higher in Prader-Willi syndrome than in the general population. Butler and colleagues report that the average life expectancy for individuals with Prader-Willi syndrome is 29.4 years, with a 20% mortality rate by 20 years of age, 50% by 29 years of age, 75% by 42 years of age, and 99% by 60 years of age (22). Death in childhood was more likely due to respiratory failure or aspiration; in adults the most common causes of death were cardiac disease, pulmonary thromboembolism, accidents, and sepsis (22).

Obesity is the major contributing factor to the morbidity and mortality of this condition, and there is a clear relationship between the degree of obesity and the frequency of medical complications, which may correlate with mortality (73). Type 2 diabetes is seen in 19% to 41% of patients; hypertension in 18% to 32% of patients; and recurrent cellulitis, phlebitis, and edema in 27% of patients. Restrictive lung disease is seen in 7 out of 8 adults. Obstructive sleep apnea may occur, but cor pulmonale rarely develops (73; 29; 112).

A significant number of adults (86%) continue to have behavior problems, with 23% requiring psychiatric hospitalization and 45% requiring medication for behavioral or psychiatric problems (29; 62). Most adults with this condition require some supervision.

Clinical vignette

The patient, a boy aged 9, was the product of a term gestation that was uncomplicated except for the detection of oligohydramnios at 36 weeks. Cardiac decelerations occurred during labor, prompting an emergency cesarean section. Birth weight was 2.87 kg with Apgar scores of 6 and 7 at one minute and 5 minutes, respectively. The baby was given oxygen but required no further resuscitative efforts. He was noted to be markedly hypotonic and fed poorly. There was no muscle weakness or wasting, and tendon reflexes were brisk. The testes were undescended, and the scrotum was underdeveloped; otherwise, there were no obvious dysmorphic features.

Extensive blood work (including CBC, electrolytes, glucose, calcium, magnesium, creatinine, ALT, AST, ammonia, lactate, venous gases, and creatine kinase) was normal. LDH levels were increased at 1549U/L. A urine metabolic screen was negative, as was an organic acid spectrographic screen. Cranial imaging (head ultrasound and CT) was normal.

Family history for neonatal hypotonia, neuromuscular disease, and developmental delay was negative

The baby’s feeding ability gradually improved to the point of allowing him to be discharged from the hospital. The working diagnosis at that time was central hypotonia of undetermined etiology.

At 8 months of age, the patient was reassessed at the Neurology Clinic. His development was delayed; he was unable to roll over, sit unaided, or crawl, and he had no interest in helping with feeding. Examination showed height, weight, and head circumference below the third percentile, whereas all had been previously normal. The boy was noted to have almond-shaped eyes and a downturned mouth. As before, he showed marked hypotonia with a head lag but normal limb movement against gravity and intact tendon reflexes.

The diagnosis of Prader-Willi syndrome was suspected on clinical grounds and was confirmed by the detection of a microdeletion at 15q11-q13, using fluorescent in-situ hybridization with a probe for the SNRPN locus. Giemsa banding revealed no sign of a macroscopic deletion.

The patient was referred to a developmental treatment program and had a carefully supervised diet. At the age of 6 years, he was significantly delayed from language and cognitive standpoints, functioning, in general, at about a 3- to 3.5-year-old level. Although his height was in the 98th percentile at 4 years of age, there was a subsequent gradual deceleration in growth velocity. At 8.5 years of age, his height and weight are in the region of the second percentile.

Biological basis

Etiology and pathogenesis

The etiology of this condition is the lack of a paternally inherited gene or genes in the 15q11-13 locus. This occurs most commonly by deletion of the paternal copy (accounting for 65% to 75% of cases) but can also occur via maternal uniparental disomy (20% to 30%) whereby both copies of the region are present but are of maternal origin and inactivated (110; 145; 115; 130). A cytological deletion of 15q11-q13 is detected in more than 50% of cases. A small number of patients have imprinting center deletions or translocations involving the 15q11-q13 region. In several reports, the relative percentages of patients having uniparental disomy has significantly increased relative to those with deletions.

Gene imprinting. The phenomenon of gene imprinting appears to be central to the biological mechanisms involved in the pathogenesis of Prader-Willi syndrome. Imprinted genes are active in a given chromosome derived from one parent and inactive in the homologous chromosome from the other parent. This phenomenon appears to be common in genes involved in growth and development and has conferred an evolutionary advantage across species (131). Although there are imprinted genes on several chromosomes (6, 7, 11, 14, 15, and 20), the proximal segment of the long arm of chromosome 15 harbors a number of imprinted genes and is the site of the deletions common to Prader-Willi and Angelman syndromes (78).

The mechanism of gene imprinting is not entirely clear but may involve one or more of the following: methylation of DNA cytosine within the gene, changes in the configuration of chromatin, or the adduction of a protein (68). When a gene is methylated, the regional DNA strand is more tightly coiled, and, thus, the gene is not sufficiently exposed to permit RNA transcription.

Genetic findings in Prader-Willi syndrome. Within the 15q11-q13 segment implicated in Prader-Willi and Angelman syndromes, there are a series of genes that are active in the chromosome derived from the father and inactive (methylated) in the maternally derived chromosome: MKRN3, MAGEL2, NECDIN (NDN), C15orf2, SNURF/SNRPN, and a group of small nucleolar RNA genes (snoRNAs) (131; 78; 11; 89). All of these genes are potentially involved in the pathogenesis of Prader-Willi syndrome. Just downstream from the snoRNAs are at least 2 genes (UBE3A and ATP10C) that are active in the maternally derived chromosome and inactive in the chromosome from the father. Mutations in the UBE3A gene have been shown to be associated with Angelman syndrome, a distinct disorder characterized by developmental delay, microcephaly, and seizures (116).

Prader-Willi syndrome results from the absence of activity of the paternally imprinted genes in the 15q11-q13 segment of the affected individual. This lack of active (unmethylated) genes occurs most frequently (approximately 70% of cases in most series) due to a deletion in the 15q11-q13 region in the chromosome derived from the father. The maternally derived chromosome is intact, but the implicated genes have appeared to be completely inactive (methylated); however, this inactivation process is not always 100%. The mechanism of deletion is uncertain but may reflect a “hotspot” because the breakpoints cluster in the same region for both Prader-Willi and Angelman syndromes, resulting in the size of the deletion being similar for both paternal and maternal (Angelman syndrome) deletions (98).

There are 5 major breakpoints in the 15q11-q13 region (BP1-5). The majority of deletions in Prader-Willi syndrome occur between BP2-3 (type 2 deletion), whereas most of the remainder are between BP1-3 (type 1 deletion). In general, patients with type 1 deletions are more compromised than those with type 2. This difference may reflect absence of function in one or more unimprinted genes located between BP1 and BP2 (TUBGCP5, CYFIP1, NIPA1, NIPA2). Deletions confined to this region have been associated with intellectual impairment, speech and language delays, and autism spectrum disorder (26). Larger deletions, especially BP2-BP5, include non-imprinted genes such as CHRNA7 (implicated in the so-called 15q13.3 microdeletion syndrome) and are associated with dysmorphic features not seen in most patients with Prader-Willi syndrome (91).

Of those Prader-Willi patients lacking obvious deletions, the majority (approximately 25% of the total) have uniparental disomy, with both chromosome 15s derived from the mother (76). This phenomenon is believed to occur most often from a combination of chromosome 15 nondisjunction in the egg and the subsequent loss of the paternal copy of chromosome 15 in the embryo, in which case the implicated genes are inactive in both chromosomes. The risk of uniparental disomy from nondisjunction increases with advanced maternal age (145). A study suggests that in countries where the average maternal age is increasing, the proportion of Prader-Willi cases due to uniparental disomy is also increasing. Whittington and colleagues reported that in the United Kingdom, of children aged 5 or younger with Prader-Willi syndrome, 50% were found to have maternal isodisomy versus 44% with deletions (175).

An alternative explanation for the unexpectedly high percentage of uniparental disomy cases reported by Whittington and colleagues has been proposed by Sinnema and colleagues (157). In a Dutch cohort of 102 adults with Prader-Willi syndrome, these authors found 54% with a paternal deletion, 43% with uniparental disomy, and 3% with imprinting center defects. Sinnema and colleagues suggested that the high percentage of uniparental disomy cases reflected a delay in the diagnosis of milder (ie, uniparental disomy) cases until later in life; earlier series were based on genetic diagnoses made in young patients with a more obvious phenotype. Multiple studies have found that the age of Prader-Willi diagnosis is significantly delayed in patients with uniparental disomy compared to those with deletions, lending further credence to the postulation that uniparental disomy results in a milder phenotype that may take longer to diagnose (71).

Two other mechanisms for Prader-Willi syndrome have been identified in the remaining 5% of patients. Some have been found to have microdeletions in the “bipartite imprinting center,” a region adjacent to the SNURF/SNRPN gene, with 2 components: AS-IC (Angelman syndrome imprinting center), located just upstream from SNURF/SNRPN, and PWS-IC (Prader-Willi syndrome imprinting center), which is embedded within exon1 of SNRPN (140). The normal function of PWS-IC is to unmethylate the inactive genes on the 15th chromosome derived from the mother during gametogenesis in the male and to methylate (render inactive) the maternal UBE3A and ATP10C genes. Conversely, during oogenesis, the AS-IC inactivates the relevant genes on the paternally derived chromosome 15 and activates UBE3A and ATP10C (15). This mechanism allows for any of the 15th chromosomes in the gametes to function in the eventual zygote as appropriate male-derived or female-derived chromosomes, the ultimate raison-d’etre for the imprinting process. A deletion in the imprinting center of the maternally derived chromosome 15 in the sperm leads to a failure of conversion to the “male pattern” and ultimately results in 50% of the man’s offspring having 2 sets of inactive genes: the normally inactive genes from the mother and the nonactivated genes from the father. All of these individuals have Prader-Willi syndrome and represent the only exception to the rule that Prader-Willi syndrome is normally a sporadic disorder (15).

The remaining identified mechanism for Prader-Willi syndrome is a balanced translocation through the 15q11-q13 region with no identifiable deletion. Two patients have been described with balanced translocations involving chromosome 15 with chromosome 19 and chromosome 4 respectively, with the breakpoint between exon 2 and exon 3 of the SNURF/SNRPN gene. Both had a classic Prader-Willi phenotype (96).

Possible roles of the imprinted genes. The individual roles of the inactivated paternally derived genes at the 15q11-q13 deletion site in the pathogenesis of Prader-Willi syndrome are far from clear. Although null mouse studies offer some insights into potential molecular mechanisms contributing to the somatic, behavioral, and metabolic symptoms of Prader-Willi syndrome, no mouse models are able to fully replicate the human Prader-Willi phenotype (143). Of the genes located within the Prader-Willi region, the SNORD116 cluster appears to be critically implicated in the pathogenesis of Prader-Willi syndrome; other genes may contribute to, but cannot fully account for, the Prader-Willi phenotype.

Individuals with microdeletions confined to the snoRNAs site in the 15q11-q13 region, particularly the SNORD116 cluster, show typical features of Prader-Willi syndrome (147; 59; 09). The SNORD116 gene locus produces a primary RNA transcript that is processed into 2 noncoding RNAs: Snord116 snoRNAs, which localize to the nucleolus of mature neurons, and the Snord116 host gene (38). Most snoRNAs function by methylating target rRNA or snRNA molecules; however, SNORD116 and SNORD115 snoRNAs have no identified rRNA targets. Instead, they may actually undergo further processing into smaller snoRNA molecules (psnoRNAs), which may perform a variety of regulatory functions (eg, splicing factors, activating proteins, stabilizing RNA) (66). In collaboration with SNORD115, the SNORD116 gene promotes production of at least 23 other genes, suggesting the presence of Prader-Will syndrome symptoms in individuals with deletions confined to the SNORD116 cluster may effectively result from downregulation of a number of other genes whose products are essential for normal brain development and function (65).

Snord116-null mice recapitulate the main features of Prader-Willi syndrome including low birth weight, hyperphagia, obesity, and endocrine abnormalities (139). The SNORD116 gene itself is expressed ubiquitously in mice but splicing and processing of the Snord116 transcript into snoRNAs only occurs in brain tissues (38). Using a selective SNORD116 knockout mouse model, Qi and colleagues demonstrated that elimination of Snord116 expression from neuropeptide Y cells of the arcuate nucleus in the hypothalamus, but not other brain tissues, reproduced the murine Prader-Willi condition, suggesting that Snord116 plays an important role in food and hunger control through regulation of neuropeptide Y neurons in the hypothalamus (139). Using a transgenic mouse model, Coulson and colleagues discovered that adding an exogenous Snord116 transgene to Snord116 knockout mice was not sufficient to rescue these mice from the Prader-Willi phenotype; additional cellular machinery present in wild type but not Snord116 knockout mice is necessary to process SNORD116 into functional snoRNAs (38). In addition, researchers have shown dysregulation of diurnal methylation patterns in the cortex of Snord116 knockout mice, including altered expression of diurnally methylated genes involved in neurogenesis, circadian entrainment, and metabolic pathways implicated in obesity (39). This work may provide new mechanistic insights into epigenetic influences on circadian rhythms and how alterations in regulation of circadian rhythm contribute to manifestations of Prader-Willi syndrome including sleep disorders, hyperphagia, and metabolic/endocrine dysfunction.

The SNORD115 gene produces an RNA transcript that binds to and potentially promotes alternative splicing or editing of the 5-HT2C serotonin receptor mRNA, which is expressed predominantly in the brain (66). A mouse model of altered serotonin 2C receptor RNA editing demonstrates many features of human Prader-Willi syndrome: failure to thrive, decreased somatic growth, neonatal hypotonia, and initially poor food consumption followed by post-weaning hyperphagia (126). However, both patients and knockout mice with SNORD115 gene cluster deletions lack typical Prader-Willi syndrome features (55; 146). It remains unclear whether downstream regulation of 5-HT2C via Snord115 (or perhaps other unidentified factors or cofactors) plays any significant role in the manifestation of Prader-Willi syndrome.

Nectin plays a role in axonal outgrowth, possibly through interactions with other neuronal proteins including Magel2, which help regulate cytoskeleton organization (101). Nectin appears to regulate respiratory rate. Most necdin-deficient mice (-/-) die in the neonatal period from respiratory insufficiency akin to the respiratory problems seen in neonates with Prader-Willi syndrome. Heterozygous necdin-deficient mice with the deletion only on the paternal allele have less severe respiratory difficulties and are much more likely to survive to adulthood (144). In theory, this improved outcome in heterozygotes should not be happening as the maternal allele is presumably silenced. As demonstrated by Rieusset and colleagues, however, deletion of the paternal origin necdin allele in mice is accompanied by varying degrees of residual expression of necdin mRNA derived from the maternal allele. This finding also appears to apply to human Prader-Willi syndrome as these authors also found NECDIN mRNA expression in the paraventricular and supraoptic nuclei of the hypothalamus in 8 postmortem brains of Prader-Willi patients, whether of the deletion or isodisomy subtypes.

The respiratory insufficiency in necdin-null mice appears to be related to rhythm instability in the pre-Botzinger complex, the respiratory rhythm-generating center (141). Necdin-deficient mice surviving to adult life have reduced numbers of oxytocin-producing and gonadotropin-releasing hormone neurons in the hypothalamus (128; 124). Clinically, they demonstrate frequent skin-scraping activities and superior spatial learning, both of which are seen in humans with Prader-Willi syndrome (128). The former behavior may be related to increased neuronal apoptosis in sensory ganglia and a paucity of substance P-containing neurons in necdin-deficient mice (97); in turn, these findings may be relevant to the elevated pain thresholds seen in Prader-Willi subjects (137).

The MAGEL2 gene encodes a protein expressed in multiple brain areas including the hypothalamus and appears to be involved in brain structure and development, reproductive function and fertility, and multiple homeostatic processes including circadian rhythm, appetite and weight gain, and bone metabolism (10; 119; 05). A cohort of 78 patients with truncating mutations confined to the paternal allele of MAGEL2 have been described, and although they share similar early symptoms with Prader-Willi patients, they lack many of the typical morphological features and do not develop hyperphagia and obesity in later childhood and are classified as a distinct syndrome, Schaaf-Yang syndrome (117). Autism spectrum disorder (ASD) has a much higher prevalence in patients with Schaaf-Yang syndrome, suggesting Magel2 may play a role in the development of autism spectrum disorder symptoms (117).

Magel2-null mice demonstrate different but complementary features to those seen in Prader-Willi syndrome: neonatal growth retardation, impaired reproductive function with infertility, altered bone homeostasis, changes in circadian rhythm, and dysregulation of feeding and weight leading to excessive weight gain after weaning and increased adiposity in adult life (10; 119; 05). With respect to regulation of food intake and weight gain, Magel2-null mice more closely follow a Prader-Willi trajectory than that of Schaaf-Yang patients and, thus, have provided insights into the feeding impairments seen in Prader-Willi syndrome. Magel2-null mice show progressive leptin insensitivity and reduced volumes of anorexigenic axons in the arcuate nucleus of the hypothalamus, a crucial center for appetite control (136; 111).

MKRN3 encodes a protein (makorin ring finger protein 3) involved in the regulation of puberty and may play a role in the hypogonadism and infertility seen in Prader-Willi syndrome (36).

Kanber and colleagues describe 2 patients with atypical deletions in the Prader-Willi region on the paternal chromosome 15 that include C15orf2 and the SNURF-SNRPN locus (which encompasses the SNORD116 gene cluster), but not MKRN3, MAGEL2, and NDN genes (89). Both patients show classic Prader-Willi syndrome, indicating the MKRN3, MAGEL2, and NDN genes are not sufficient to cause Prader-Willi syndrome.

The proposed role of SNURF/SNRPN in Prader-Willi syndrome comes largely from human evidence, particularly the production of the classic Prader-Willi phenotype by the disruption of exon 2 and exon 3 of the SNURF component in balanced translocations (96). Contradictory evidence comes from mouse models in which neither snurf-deficient nor snrpn-deficient mice demonstrate features in common with Prader-Willi syndrome (128).

Finally, there are reports of cases with typical features of Prader-Willi syndrome related to microdeletions confined to the snoRNAs site in the 15q11-q13 region, particularly the SNORD116 cluster (147; 59). The snoRNAs include SNORD64 (1 copy), SNORD107 (1 copy), SNORD108 (1 copy), SNORD109 (2 copies), SNORD116 (29 copies), and SNORD115 (48 copies) (139). An in vitro study by Falaleeva and colleagues revealed that SNORD116, in collaboration with SNORD115, promotes the production of mRNAs from at least 23 other genes, suggesting that the presence of Prader-Will syndrome symptoms in individuals with deletions confined to the SNORD116 cluster may result from, effectively, downregulation of a number of other genes whose products are essential for normal brain development and function (65).

Snord116-null mice recapitulate key features of Prader-Willi syndrome including low birth weight, hyperphagia, and obesity (139). Interestingly, Qi and colleagues demonstrated that selective deletion of Snord116 from neuropeptide Y (NPY) neurons resulted in the same phenotype as the global deletion (139). In addition, researchers have shown dysregulation of diurnal methylation patterns in the cortex of Snord116 knockout mice, including altered expression of diurnally methylated genes involved in neurogenesis, circadian entertainment, and metabolic pathways implicated in obesity (39). This work may provide new mechanistic insights and understanding how genetic factors, through effects on methylation, contribute to manifestations of Prader-Willi syndrome including sleep disorders, hyperphagia, and metabolic/endocrine dysfunction.

Genotype-phenotype correlations. Larger deletions in the 15q11-q14 region have been associated with additional clinical manifestations. Deletion of the human P gene, downstream from UBE3A, results in a decrease in pigmentation. This phenomenon can be seen in both Prader-Willi and Angelman syndromes as the P gene is not imprinted (130). Such patients typically have central visual impairment, in part related to retinal hypopigmentation, as is seen in albinism (41). Inclusion in a deletion of one or more of the 3 nonimprinted GABA-receptor genes at the 15q11-q13 site may account for some of the cases of febrile seizure and epilepsy in Prader-Willi syndrome (67; 169). Windpassinger and colleagues described a 2-year-old boy with Prader-Willi syndrome caused by a large 15q11-q14 deletion resulting from an unbalanced translocation t(3; 15). In addition to depigmentation and central visual impairment, this patient also demonstrated relative macrocephaly, retrognathia, preauricular tags, and bilateral club feet, presumably related to the absence of other nonimprinted genes in the 15q13-q14 region (178).

Some molecular genetic evidence has emerged that may help explain why there are phenotypic differences (particularly cognitive and behavioral) between subjects having deletions and subjects with maternal disomy. Bittel and colleagues evaluated transcripts for 73 genes in the 15q11-q13 region (12). Although paternally expressed genes from Prader-Willi syndrome cell lines were equally suppressed in deletion and maternal disomy cases, 2 maternally expressed genes (UBE3A, ATP10C) showed significantly increased expression in disomy cell lines in comparison with controls and Prader-Willi deletion cases (12). On the basis of available evidence, there is a strong possibility that overexpression of UBE3A and ATP10C in disomy cases is related to the observed cognitive and behavioral differences between subtypes.

Evidence of a second mechanism for the autistic features seen in Prader-Willi syndrome has emerged, a mechanism that does not directly involve the paternally imprinted genes in the 15q11-q13 region. Nagarajan and colleagues evaluated the expression of MeCP2 (the gene product defective in Rett syndrome) in the frontal cortex of Rett syndrome patients and of patients with a variety of other genetic disorders in which autistic features commonly occur. They found that MeCP2 expression was significantly reduced not only in Rett syndrome patients but also in subjects with Prader-Willi syndrome, Angelman syndrome, Down syndrome, and idiopathic autistic spectrum disorder (129). These results suggest the possibility that the autistic symptoms seen in Prader-Willi, Angelman, Down, and Rett syndromes may relate to MeCP2 deficiency, the mechanism in the non-Rett cases being an epigenetic phenomenon.

It is, thus, possible that other clinical features of Prader-Willi syndrome are secondary to relatively remote downstream effects on other genes or genetic mechanisms. A mouse model of altered serotonin 2C receptor RNA editing (a process mediated by snoRNA genes) demonstrates many features of human Prader-Willi syndrome: failure to thrive; decreased somatic growth; neonatal hypotonia; and initially poor food consumption followed by post-weaning hyperphagia (126).

The hypothalamus in Prader-Willi syndrome. From the clinical perspective, there are grounds for suspecting hypothalamic dysfunction in Prader-Willi syndrome. The hypogonadism is associated with low gonadotropin levels, low gonadal steroid levels, the presence of Sertoli cells, and the variable numbers of Leydig and germinal cells in the testes (18). Growth hormone secretion is low in both obese and nonobese patients and has shown diminished response to provocative tests (37). The temperature instability problems sometimes seen in the first stage of the disorder also presumably stem from hypothalamic dysfunction (150). Finally, the characteristic insatiable hyperphagia suggests a defect in the hypothalamic satiety center, a hypothesis supported by the findings of Maillard and colleagues in the Magel2-null mouse (111).

A putative dysfunction in the hypothalamic satiety mechanism has been supported by an fMRI study of regional blood flow changes following an oral glucose load. Shapira and colleagues found that a postglucose decrease in blood flow to the ventromedial frontal cortex, nucleus accumbens, and hypothalamus was significantly delayed in Prader-Willi subjects in comparison with lean and obese controls (153). Using a different paradigm, Holsen and colleagues found that, in response to pictures of food shown prior to a meal, Prader-Willi subjects had higher metabolic activity in the nucleus accumbens/amygdala and lower activity in the hypothalamus/hippocampus than did obese and normal-weight controls (85). In the postprandial state, the Prader-Willi subjects continued to show enhanced activity in the subcortical regions, whereas the obese (but not the healthy) controls had higher activation of dorsolateral prefrontal and orbital-frontal cortices, the sites of inhibitory control over food intake.

Although these fMRI results do not clearly localize the origin of hyperphagia in Prader-Willi syndrome, they do suggest a fundamental alteration in neural mechanisms of satiety.

Hyperphagia is also observed in Prader-Willi mouse models (44; 111; 139). Interestingly, excess food consumption in these mice does not appear to be determined by the sweetness of food items, but by their caloric content (44).

Neuropathological studies. Neuropathologic studies in Prader-Willi syndrome are few, and the results are inconsistent (80; 142). There are no characteristic gross or microscopic abnormalities in the hypothalamic region or elsewhere. Swaab reported a specific reduction in oxytocin-containing and vasopressin-containing neurons in the paraventricular nucleus in comparison with controls (162); oxytocin neurons are important in the regulation of food intake. A similar study of the infundibular nucleus by the same group showed that growth hormone-releasing hormone neurons were not reduced in Prader-Willi syndrome in comparison with obese controls (72).

Swaab’s observations concerning oxytocin neuron depletion in Prader-Willi syndrome are interesting given the accumulating evidence that oxytocin is an important neuromodulator in facilitating facial recognition of emotional states, peer recognition, and interpersonal bonding behavior (57; 56).

Down-regulation of oxytocinergic activity has, in fact, been demonstrated in Magel2-deficient mice (120). Interestingly, daily administration of subcutaneous oxytocin to these mice in the first postnatal week permanently protects the mice against the later evolution of deficits in social behavior and learning (120).

Hayashi and colleagues reported a pathological study of a deletional Prader-Willi patient who had a selective loss of cholinergic neurons in the pedunculopontine tegmental nucleus, a finding that could account for the patient’s hypotonia and REM sleep abnormalities (81). Cholinergic neurons in the nucleus basalis of Meynert, in contrast, were normal, as were GABAergic interneurons in the cerebral cortex and catecholaminergic neurons in the brainstem. These are intriguing findings that have yet to be replicated.

Postmortem transcriptional analysis of hypothalamic tissue in Prader-Willi patients reported by Bochukova and colleagues identified increased expression of genes that signal hunger and proinflammatory genes (13). Upregulated genes were located predominantly in microglial cells. There was downregulation of neuronal genes, including genes that signal fed state and genes associated with neurogenesis, synaptic plasticity, and neurotransmitter release. Specifically, the authors noted decreased levels of oxytocin mRNA (supporting Swaab’s 1997 neuropathological study) and decreased levels of brain derived neurotrophic factor (BDNF) and its receptor, trkB (NTRK2), mutations to which have previously been associated with symptoms that overlap with the Prader Willi phenotype. Addition of BDNF to a SNORD116 deficient human neuroblastoma cell line partially improved neuronal growth, suggesting decreased BDNF/NTRK2 expression may contribute to the neurodysgenesis and/or neurodegeneration associated with Prader-Willi syndrome (13).

MRI-based neuroanatomical studies. To date, four voxel-based morphometric MRI studies comparing Prader-Willi patients to controls have been published, showing heterogenous results. Ogura and colleagues analyzed 12 adults with Prader-Willi syndrome, finding decreased gray matter volumes in orbitofrontal cortex, inferior temporal gyrus, caudate nucleus, supplementary motor area, pre- and post-central gyri, and cerebellum (132). After correction for total gray matter volume, only the orbitofrontal cortical volume reduction remained, a finding that is potentially relevant to many of the behavioral disturbances in Prader-Willi syndrome.

A study comparing 20 adults with Prader-Willi syndrome to age-matched controls found subjects with Prader-Willi syndrome showed no difference in total brain or grey matter volumes but had multiple localized grey matter abnormalities including areas of both decreased grey matter (ventromedial prefrontal areas, right lateral prefrontal cortex, and bilateral temporal poles) and increased grey matter volume (prefrontal cortex, cingulate cortices, insula, portions of the parietal and temporal cortices, caudate, putamen, and thalamus) (113). Increased grey matter was attributed to increased cortical thickness. This diffuse range of structural brain abnormalities may be the result of an early systemic dysregulation of neurodevelopmental processes leading to the development of abnormal brain structures, which may provide the biological basis for abnormal brain function and many of the symptoms associated with Prader-Willi syndrome (113).

There are 2 voxel-based, morphometric studies in which comparisons have been made among Prader-Willi subjects with deletions – those with uniparental disomy and sibling controls (86; 106). In both studies, the number of subjects was small (deletion n = 15, 11; disomy n = 8, 9; control n = 25, 11, respectively), and the results must be interpreted with caution. In addition, the study by Honea and colleagues largely focused on cerebral cortical gray and white matter and was performed in adults, whereas that of Lukoshe and colleagues was in children and also evaluated central gray matter structures, ventricular size, and brainstem. In both studies, gray matter volumes were reduced in deletion subjects, but not in disomy subjects with respect to controls. Results of comparisons for white matter volumes, however, were discordant, possibly due to differences in methodology.

Another MRI-based technique for the assessment of cortical (and, thus, gyral) development is the local gyrification index, referring to the ratio between the total surface area of the brain (including that present in the sulcal regions) and the surface perimeter area; the ratio can be calculated for any cortical region under consideration. In a second study of 24 children with Prader-Willi syndrome by Lukoshe and colleagues, local gyrification indices in large areas of both cerebral hemispheres were lower than those found in age- and sex-matched siblings and in healthy controls (104). Although there were some differences between groups, these findings applied more or less equally to subjects with deletions and those with uniparental disomy. Relative cortical underdevelopment in the Prader-Willi subjects was particularly prominent in the frontal, parietal, and temporo-limbic areas. Although total cortical surface area was diminished in these regions, cortical thickness in the same areas did not differ from controls (104). The authors hypothesized that their findings could be consistent with a degree of simplification, or relative lack of complexity, in the regions of concern, which could be related to the developmental deficits observed clinically. More studies are needed to sort out current inconsistencies and further elucidate the nature and functional implications of volumetric differences in the grey matter of Prader Willi patients.

Lukoshe and colleagues have also evaluated MRI functional connectivity differences between 27 Prader-Willi participants and 28 typically developing children (105). They found that, in comparison to controls, Prader-Willi syndrome participants demonstrated increased connectivity between the hypothalamus and the lateral occipital cortex. The latter regions are important for visual processing, object recognition, and evaluating the caloric value of foods (105). These authors also found that, although the volume of the hypothalamic region in Prader-Willi subjects was equal to that of controls, pituitary gland volumes were decreased by 50%.

Endocrinological aspects. There has been speculation that the hyperphagia in Prader-Willi syndrome might be linked to aberrations in levels of 2 closely related circulating peptides, ghrelin and obestatin. Ghrelin is a stomach-derived orexigenic compound that induces fat deposition by increasing food intake and decreasing fat utilization, particularly in conditions of starvation and cachexia (50). In contrast, obestatin decreases food intake and decelerates gastric emptying, reducing body weight (19; 133). Both compounds are derived from the posttranslational cleavage of the same gene product, preproghrelin. In principle, either excessive levels of ghrelin or inadequate levels of obestatin might lead to hyperphagia and obesity.

In fact, plasma ghrelin levels are consistently elevated in older subjects with Prader-Willi syndrome in comparison with obese controls (50; 165; 19); ghrelin levels are also elevated in a mouse model of Prader-Will syndrome (44). Beauloye and colleagues showed that, in children and adults with Prader-Will syndrome, the elevated ghrelin levels are primarily in the acetylated form (06). In contrast, infants with the disorder have low levels of acetylated ghrelin and high levels of unacetylated ghrelin, suggesting that the latter form may be anorexigenic (06). Contrary to expectation, obestatin levels in Prader-Willi subjects were not found to differ from those in obese controls, regardless of age, either in the fasting state or following a glucose load (19; 133).

These somewhat conflicting findings suggest that ghrelin may not be the direct cause of hyperphagia in Prader-Will syndrome but that its elevated plasma concentrations may be secondary to a more fundamental defect that has not yet been identified. This conclusion is supported by a controlled trial of octreotide, a potent suppressor of ghrelin production. As expected, Prader-Willi subjects taking octreotide had a significant and sustained reduction in ghrelin levels; there were no significant changes, however, in food-seeking behaviors and body mass index (51).

Hyperphagia in Prader-Willi syndrome has also been linked to low circulating levels of brain-derived neurotrophic factor (BDNF) and to a failure of BDNF blood levels to rise following a glucose load (77; 14). BDNF knockout mice also demonstrate hyperphagia and obesity (77). Bueno and colleagues also demonstrated that plasma leptin levels are elevated in Prader-Willi syndrome, as they are in obese typical adults (14). Leptin is a peptide produced by adipocytes, and it normally stimulates anorexigenic neurons in the hypothalamus. Again, these BDNF and leptin results may be secondary to a more fundamental hypothalamic disturbance (14).

Also, indicative of a fundamental disturbance in hypothalamic regulation in Prader-Willi syndrome is the finding that orexin A levels are also elevated in this disorder (114). Orexin A (hypocretin1) is involved in arousal, regulation of the sleep/wake cycle, metabolism and energy expenditure, and reward seeking behavior (114). Finally, there is also a suggestion that the obesity in Prader-Willi syndrome may not result exclusively from overeating. Those patients who manage to lose weight require lower caloric intakes, approximately 60% of the norm for their age, to maintain reduced weight (18). Energy expenditure appears to be below normal, but studies measuring metabolic rates have been inconclusive. Thyroid hormone, lipid profiles, insulin, and glucocorticoid levels are comparable to those of obese individuals (18). Fat cells are larger than in controls, but the number of fat cells and the uptake of fat and fatty acid composition are not increased (08; 18).

An alternative mechanism for fat production in Prader-Willi syndrome was suggested in a study of dispersed cell cultures of pre-adipocytes (17). Prompted by extensive previous work in Wevrick’s laboratory on the necdin null mouse, these authors found that necdin overexpression inhibited the conversion of pre-adipocytes to mature fat cells, whereas reduced necdin levels in tissue culture enhanced this conversion.

Dysregulation of the growth hormone/insulin-like growth factor 1 axis is prominent in the majority of patients with Prader Willi syndrome. Pituitary growth hormone reserve levels decrease with age, which may, at least in part, contribute to the changes in growth velocity observed in older children and adolescents (75).

Epidemiology

Prevalence rates vary by study and location. Butler estimates incidence at 1 in 25,000 (18). An epidemiological survey performed in North Dakota showed a prevalence of 1 per 16,602 with an equal sex distribution (16). In Sweden, using a standard criterion, the population prevalence for individuals from newborn to 25 years of age was 12 per 100,000 with a male to female ratio of 1.6:1. The slight difference in sex ratio probably reflects the difficulty of diagnosing hypogonadism in prepubertal girls. In the same study, the sex ratio was 1:1 in pubertal and postpubertal cases (02). Prevalence in the United Kingdom is estimated at 1 in 45,000 to 1 in 52,000 (177).

Prevention

A small recurrence risk (less than 1 in 1000) has been cited for families with one child with Prader-Willi syndrome (27). In order to ascertain the risk of recurrence, it is essential to characterize the molecular genetic basis for Prader-Willi syndrome in a given patient. For those with deletions, uniparental disomy, and de novo balanced translocations, the recurrence risk is low. Where the proband is found to have an imprinting center mutation, however, the recurrence risk is 50% (15). In such situations, further cases can be prevented via prenatal diagnosis.

A second scenario exists in which accurate prenatal diagnosis can prevent the birth of a child with Prader-Willi syndrome. On occasion, chorionic villus sampling for advanced maternal age has revealed trisomy 15, whereas amniocentesis has revealed a fetus with 46 apparently normal chromosomes, a phenomenon referred to as confined placental mosaicism (28; 127). Unfortunately, such pregnancies may produce a child with Prader-Willi syndrome secondary to maternal disomy, with loss of the paternal chromosome 15 present in the trisomic chorionic villus cells. In pregnancies where trisomy 15 is identified on chorionic villus sampling, the fetal cells should be evaluated for the presence of maternal disomy for chromosome 15.

In a study by Chang and colleagues, they found that routine second trimester karyotyping of amniocytes in high-risk mothers (for advanced maternal age, abnormal maternal serum screening, or abnormal fetal ultrasound findings) may reveal absence of the normal band at 15q12 (27 out of 26,041 samples) (30). This suggests the possibility of identifying both Prader-Willi and Angelman syndromes prior to the achievement of potential fetal viability. In their study, however, only one out of 43 samples with this scenario had Prader-Willi syndrome by methylation testing, whereas 2 out of 43 had Angelman syndrome (30).

There is preliminary evidence to suggest that it may be possible to detect the presence of Prader-Willi syndrome in early third trimester fetuses on the basis of a signature ultrasound phenotype. Two studies have suggested a constellation of ultrasound findings that are individually nonspecific but that, in combination, could lead one to suspect the diagnosis of Prader-Willi syndrome during pregnancy: decreased fetal movement, breech presentation, severe intrauterine growth retardation, increased head/abdomen circumference ratio (indicative of asymmetrical intrauterine growth), and polyhydramnios (69; 74). Given a fetus with such a combination of findings, a search for 15q11-q13 methylation defects could be undertaken via amniocentesis in jurisdictions where pregnancy termination after 22 weeks’ gestation is permitted.

Differential diagnosis

Confusing conditions

The differential diagnosis varies depending on the age of presentation. For a newborn with hypotonia, it should include those disorders that can give rise to profound hypotonia and both central and peripheral causes. Hypoxic-ischemic encephalopathy, prematurity, intraventricular hemorrhage, congenital brain anomalies, spinal muscular atrophy, myasthenia gravis, myotonic dystrophy, congenital myopathies, Zellweger syndrome, other chromosomal defects, and other conditions must be considered.

Normal electromyograms and nerve conduction velocities, creatine phosphokinase levels, muscle biopsy, metabolic screens, and the absence of cerebral pathology on imaging studies are seen in Prader-Willi and will exclude most of these other conditions.

It is important to recognize that infants with Prader-Willi syndrome may not manifest the characteristic dysmorphic features of the disorder, including hypogonadism (125). For this reason, it is worthwhile to consider DNA methylation screening for all neonates with undiagnosed central hypotonia.

Differentiating between Prader-Willi and Angelman syndrome in a hypotonic infant with a deletion of 15q11-q13 may be difficult, especially in girls, in whom genital anomalies are frequently absent. If blood from both parents is available, determination of the parent of origin may be possible. On the basis of probability, Angelman syndrome rarely presents with neonatal hypotonia, whereas this finding is characteristic of Prader-Willi syndrome.

Patients with Schaaf-Yang syndrome, which is caused by truncating mutations to the MAGEL2 gene located within the paternally imprinted Prader-Willi 15q11-q13 locus, show similar early symptoms, including neonatal hypotonia, developmental delays, and poor feeding and growth, but patients typically do not develop hyperphagia and often lack the dysmorphic features associated with Prader-Willi syndrome. Schaaf-Yang syndrome was previously described as Prader-Willi-like syndrome; however, the characterization of this syndrome has established distinct clinical features justifying its classification as its own distinct syndrome (117).

In the older child, the differential diagnosis for the untreated syndrome will focus on other causes of marked obesity as well as developmental delay and intellectual disability. Laurence-Moon syndrome, Bardet-Biedl syndrome, Cohen syndrome, and Börjeson-Forssman-Lehman syndrome as other genetic causes of obesity should be considered, but these usually can be eliminated by history and physical features.

In addition, at least 3 other chromosomal deletion syndromes present with neonatal hypotonia, hyperphagia, obesity, mental deficiency, and dysmorphic features: 6q16.2 (SIM1 gene) (168), 1p36 (43), and 9q34 (35). As well, a Prader-Willi syndrome-like phenotype has been described in individuals with maternal uniparental disomy of chromosome 14 (87). These syndromes should be considered in patients with Prader-Willi-like features but lacking a 15q11 methylation defect.

Associated or underlying disorders

Prader-Willi syndrome is associated with obesity; multiple endocrinologic disorders [growth hormone deficiency (40% to 100%), type 2 diabetes (25%), hypothyroidism (25%), central adrenal insufficiency, sleep disturbances]; orthopedic complications (scoliosis, kyphosis, hip dysplasia, osteoporosis); and increased rates of behavioral conditions, including autism spectrum disorder (12% to 27%), obsessive-compulsive disorder, attention-deficit/hyperactivity disorder, and psychosis.

Diagnostic workup

Genetic testing to confirm the clinical diagnosis of Prader-Will syndrome involves the detection of the lack of the male methylation pattern in the paternal chromosome in the SNRPN region of the 15q11-q13 site. This is currently accomplished using methylation-specific multiplex ligation-dependent probe amplification (ie, DNA methylation assay), which is a method capable of accurately and rapidly diagnosing all types of 15q11-q13 region defects, including microdeletions, uniparental disomy, and deletions in the imprinting center (138; 82). Chromosomal array and DNA fluorescence in situ hybridization (FISH) testing can also be used to detect 15q11-q13 deletions; FISH can also be used to detect translocations and inversions (31). Alternatively, single nucleotide polymorphism (SNP)-based chromosomal microarray analysis for the detection of copy number variations and uniparental disomy can be used to make a rapid genetic diagnosis with high reliability (102). Rarely, deletions in the SNORD116 cluster will be negative on DNA methylation testing. Thus, if Prader-Willi syndrome is suspected but DNA methylation testing is negative, chromosomal array testing is indicated (31).

Electromyograms, nerve conduction velocities, creatinine phosphokinase, and light microscopy studies of muscle are routinely performed as well (18). Despite the typical history of reduced sensitivity to pain, motor and sensory nerve conduction studies, sympathetic skin responses, and somatosensory evoked potential studies are normal (137). Muscle histochemistry may demonstrate type II fiber atrophy (01).

Management

Introduction. At the present time, there are no identified treatment modalities that are able to address the fundamental genetic defect in individuals with Prader-Willi syndrome: the lack of key functional gene expression in the Prader-Willi region of chromosome 15q11-q13. The various “treatment” modalities described below only address some of the symptoms of Prader-Willi syndrome, with greater or lesser degrees of success.

Neonatal care. The floppy newborn with feeding problems will often require gavage feeding. Ventilatory support may be necessary, sometimes for several weeks (107). Cryptorchidism and other genital anomalies may be corrected surgically.

Dietary management. Children and adults require behavioral and nutritional management to control the hyperphagia and subsequent obesity. The goal when possible is the prevention of obesity before it occurs, but weight reduction in the already overweight individual may result in a significant improvement in associated medical conditions (84).

In children, food intake must be closely monitored to prevent excessive weight gain. The actual caloric needs of the individual appear to be lower than those of children of the same age. Caloric intake should be set at 60% to 80% of the age-appropriate daily allowance, with vitamin supplementation (31). Control of the environment by locking food away, an increase in activity, and management of food-related behaviors such as tantrums and food stealing are important components of treatment. Exercise programs should accompany dietary therapy.

Normative growth charts for children with Prader-Willi syndrome not receiving growth hormone therapy have been published. Because these individuals are short in stature, weight should be adjusted for height rather than age to determine ideal body weight (83; 25).

Surgical management of obesity (eg, gastric bypass, gastroplasty, biliopancreatic diversion) is associated with poorer outcomes in Prader-Willi patients than normal obese patients and is not routinely recommended (151).

Hormone treatment. Growth hormone therapy has been used in children and adults with Prader-Willi syndrome since the early 2000s. Consensus guidelines published by the Growth Hormone Research Society recommend considering recombinant human growth hormone therapy for patients with genetically confirmed Prader-Willi Syndrome in conjunction with dietary, environmental, and lifestyle interventions (45).

The Growth Hormone Research Society recommends initiating treatment before onset of obesity, which can occur as early as 2 years of age (45). There are many published trials showing the benefits of growth hormone therapy when started as early as 4 to 6 months of age. Most endocrinologists start growth hormone supplementation during infancy (31). Positive effects of growth hormone therapy are seen in patients with or without measured deficiencies in growth hormone levels, thus, from a medical perspective growth hormone level testing is not necessary prior to starting therapy (45). Growth hormone therapy requires close surveillance and monitoring. Growth hormone therapy is contraindicated in all patients with severe obesity, uncontrolled diabetes mellitus, untreated severe obstructive sleep apnea, active cancer, or psychosis (45).

Prior to starting growth hormone, all patients should undergo comprehensive multidisciplinary evaluation including endocrine assessment; referral to sleep clinic for sleep oximetry, or polysomnogram, or both; ear nose throat assessment for causes of upper airway obstruction; and scoliosis evaluation if indicated (45). Growth hormone therapy can exacerbate underlying endocrine disorders, including central adrenal insufficiency and worsening of diabetic symptoms due to insulin antagonism (31). At least 28 unexpected deaths have been reported in patients receiving growth hormone, predominantly in morbidly obese individuals with coexisting upper airway obstruction and sleep apnea or respiratory infections or both (40; 03). Growth hormone has not been directly linked to any deaths (03). Further study is needed to better understand long-term effects of mortality and morbidity, including effects on bone health, lipid profiles, and cardiovascular risk.

Testosterone in males and estrogen in females has been successfully used to induce sexual development during adolescence. This can lead to fertility in female cases; thus, sex education is essential for patients receiving estrogen. Symptomatic glucocorticoid and thyroid hormone treatment are indicated for adrenal and thyroid deficiencies, respectively.

Pharmacologic interventions. The use of medication to control appetite or improve satiety in this condition has been, in general, unsatisfactory (52). The spate of interest in the potential role of ghrelin prompted trials of agents capable of reducing levels circulating ghrelin. Octreotide and somatostatin both lower ghrelin levels but do not change eating behavior or induce weight loss (79). Exenatide, a glucagon-like peptide-1 receptor agonist used to reduce appetite and induce weight loss in patients with type 2 diabetes mellitus, was shown to increase satiety and reduce glucose and insulin levels in a small group of patients (n=8) with Prader-Willi syndrome (163). A proof-of-concept pilot trial involving 13 Prader-Willi patients indicated safety and efficacy for diazine choline-controlled release tablets in reducing hyperphagia and aggressive behaviors (93). Diazine choline-controlled release is a KATP channel agonist used to treat hyperinsulinemic hypoglycemia that enhances the effects of leptin in the hypothalamus to decrease hyperphagia signals (93). Larger controlled trials are needed to further study this approach.

Other behaviors such as aggression and obsessive behaviors have responded to pharmacological and behavioral therapies (166; 62). The most promising medications include risperidone (60) and topiramate (159). A potential advantage of the latter agent, effective in mood stabilization, is appetite suppression with resultant weight loss, but such an effect was not confirmed in a study by Shapira and colleagues (154). Serotonin reuptake inhibitors have shown some efficacy in treating obsessive-compulsive behaviors (31).

Based on animal research, there is a possibility that oxytocin may be useful as a therapeutic intervention to improve social interaction skills as well as behavioral difficulties shown by individuals with Prader-Willi syndrome. In a randomized, double-blind, placebo-controlled cross-over study of 25 children with Prader-Willi syndrome, Kuppens and colleagues found that, for children less than 11 years (n=17) of age, twice daily intranasal oxytocin for 4 weeks resulted in significant improvement in social and food-related behaviors (94). A similar improvement was not seen in children aged 11 or older.

Skin picking, a common and potentially major cause of severe morbidity in Prader-Willi syndrome, may be amenable to pharmacotherapy. In an open-label study of N-acetylcysteine in 35 subjects with severe skin picking, Miller and Angulo found an improvement in skin-picking behaviors in all 35 patients, 25 of whom had complete resolution of their skin lesions (121). N-acetylcysteine is believed to have produced this effect either by modulation of NMDA receptors or by elevating levels of glutathione.

With respect to the management of excessive daytime sleepiness, De Cock and colleagues demonstrated a significant improvement in this symptom in 9 Prader-Willi subjects treated with modafinil, an agent used to treat narcolepsy (47). The authors recommended that the drug not be used in patients with severe obstructive sleep apnea. Central sleep apnea in infants with Prader-Willi syndrome can be markedly diminished with the use of nocturnal oxygen therapy at relatively low rates of flow (0.25 to 1 L per minute) (33). Why oxygen supplementation diminishes central sleep apnea is unclear. Cohen and colleagues speculated that chemosensitivity to oxygen and carbon dioxide levels in Prader-Willi syndrome are diminished and that mild degrees of low oxygen saturation may be sufficient to dysregulate control of breathing (33).

Developmental and educational interventions. Children with Prader-Willi syndrome require appropriate developmental surveillance and intervention. Many children will benefit from early childhood intervention programs and often need speech language, occupational, and other therapies. Children will also require evaluation for educational placement paired with psychoeducational testing to assess for intellectual and learning disabilities and attention issues. Vocational training is recommended for the older child.

Treatments on the horizon. A mouse model of Prader-Willi syndrome delineates a possible rational approach to the human disorder: restoration of expression of the silenced genes at the 15q11-q13 site in the maternal-origin chromosome. Kim and colleagues identified 2 compounds (identified as UNC0638 and UNC0642) that are selective inhibitors of euchromatic histone lysine N-methyltransferase-2 (EHMT2 or G9a), a gene product that participates (via the Prader-Willi syndrome imprinting center) in the suppression of the maternal copies of the Prader-Will-specific genes from SNRPN to SNORD116 (92). Using a tissue culture of human fibroblasts containing a 5 to 6 Mb deletion of the paternal copy of the 15q11-q13 region, the authors demonstrated that UNC0638 and UNC0642 reactivated the maternally inactivated genes in the same region, in particular SNORD116. Next, utilizing a mouse model of Prader-Willi syndrome containing a paternal deletion from snrpn to ube3a, the authors gave intraperitoneal injections of UNC0642 from postnatal day 7 (P7) to postnatal day 12 (P12). Normally all of these mice would expire in the postnatal period before weaning was completed (P16). In comparison to sham-treated mice, survival was significantly improved in the UNC0642-treated mice, with 6 out of 40 (15%) being still alive at P90 (92). No adverse effects were seen in similarly treated wild-type mice.

Although much work remains to be done to translate this work safely to human subjects, these results provide preliminary evidence for further exploration of epigenetic-based regulatory factors (in this case activation of the maternally suppressed genes at the 15q11-q13 locus) as a treatment strategy for Prader-Willi syndrome.

Outcomes

Dietary measures and effective weight management play an essential role in preventing complications of obesity, including type 2 diabetes, cardiac disease, respiratory problems such as sleep apnea, fatty liver, and gastric paresis. Implementation of balanced restricted-calorie diets have shown positive effects in body composition and weight in children with Prader-Willi syndrome (123). Schmidt and colleagues showed that children with Prader-Willi syndrome who received a well-balanced, restrictive diet starting at 14 months of age had normal BMIs at 10 years of age (152).

There are now a multitude (more than 30) of published, small, randomized control trials suggesting that growth hormone therapy provides a range of physical benefits, including linear growth, reduced body fat, increased muscle mass, improved resting energy expenditure and exercise tolerance, and improvements in motor abilities (45). There is some evidence to suggest that growth hormone therapy may improve cognition and adaptive functioning (103; 63; 58). In a large cohort study of 173 children with Prader-Willi syndrome, Dykens and colleagues reported that those receiving growth hormone treatment had significantly higher verbal and composite IQs as well as adaptive communication and daily living skills compared to those not receiving hormone therapy (63). Children who began treatment before the age of 1 showed greater improvements in cognitive functioning than those initiating treatment after the age of 1.

Growth hormone therapy may also be useful in adults with Prader-Willi syndrome. In a relatively large, randomized, placebo-controlled crossover study (n=39), Sode-Carlsen and colleagues found that growth hormone treatment was accompanied by a significant increase in muscle volume, a reduction of abdominal subcutaneous fat, and an increase in lean body mass (160). Beneficial effects on cognition have been shown to persist in adults even after growth hormone treatment has been discontinued (95).

Testosterone supplementation can exacerbate behavioral problems, including aggression. Estrogen supplementation can lead to osteoporosis and possible fertility in female patients.

Special considerations

Pregnancy

Pregnancy with a Prader-Willi syndrome child may result in breech presentation and decreased fetal movements.

There is no information available concerning pregnancy in women with this condition; it rarely occurs, as fertility is probably decreased due to hypogonadism.

Anesthesia

Patients with Prader-Willi syndrome are placed at risk during anesthesia by problems including disturbances in body temperature, preoperative diabetes, intraoperative arrhythmias, cor pulmonale, and reduced pulmonary reserves secondary to obesity. There may be problems with venous access and apnea during the postoperative period (46). No cases of malignant hyperthermia have been reported, but newborns with hypotonia are considered to be at risk for this condition. Patients may be at increased risk of aspiration of gastric contents (158).

Careful medical assessment, including pulmonary evaluation and attention to glucose control and body temperature intraoperatively, is advised (108).