Turner Syndrome

Chromosomal Origins

In 1959 TS was linked to monosohmy X in a 14-year-old girl by using a karyotype of colchicine-arrested metaphase bone marrow cells. In a subsequent study of 307 individuals with gonadal dysgenesis and sex chromosome anomalies, the short arm of the X chromosome was identified as the critical region responsible for TS .

Today, 6 decades after the first karyotyping of TS, the molecular mechanisms that lead to monosomy X and TS are still not completely understood. Based on other aneuploidies, especially the trisomies, such as Down syndrome, the most popular hypothesis for the molecular basis of TS is an error in meiosis. A maternal X is present in 70% to 80% of individuals with TS, suggesting a male predisposition to meiotic errors. Fluorescent in situ hybridization (FISH) studies in sperm support a male predisposition to meiotic errors in the sex chromosomes as autosomal disomy occurs in 0.1% and sex chromosome disomy occurs in 0.27%. If indeed TS results from an error in meiosis, unlike for trisomy investigations, the stage of meiosis that contributes to monosomy X cannot be determined. For example, in Down syndrome, nonsegregation is known to occur in meiosis II, when two chromosome 21s are homologous.

Besides an error in meiosis, the other possibility explaining monosomy X is a postzygotic nonsegregation in mitosis. Postzygotic nonsegregation is an attractive hypothesis considering that approximately half of individuals with TS are mosaic. However, there are no empirical data in humans to support this hypothesis. However, there is evidence in the mouse model that the Y chromosome is predisposed to nondisjunction during early cleavage division, and more research in humans should provide further answers.

In addition to nonsegregation in meiosis and mitosis, there are a number of structural anomalies of the X chromosome, including isochromosome X [46,X,i(Xq)] and partial deletion of Xp and Xq. The X isochromosome is the most common isochromosome in humans, and it is likely caused by chromosome breakage and recombination in the proximal Xp, resulting in an isodicentric X i(Xq) chromosome.

Epidemiology

TS is a common finding in newborn females with a birth incidence of 1 in 2500. However, TS is diagnosed more frequently prenatally. Using the Danish Cytogenetic Central Register, Gravholt et al. compared the prenatal prevalence of TS to its postnatal prevalence. A postnatal prevalence of 32/100,000 was found compared with 176/100,000 in diagnoses by amniocentesis and 392/100,000 by chorionic villus sampling (CVS). Wharton and Hook, in their study of the New York state registry, also found a higher prevalence in prenatal diagnoses: 22.2/100,000 and 85/100,000 in postnatal age and prenatally by amniocentesis, respectively.

Turner Karyotypes

Most karyotype determinations are made from mitotic peripheral lymphocytes that have been arrested in metaphase or prometaphase. The limitation of the peripheral blood karyotype is that only blood is tested and no other tissues, thus mosaicism cannot be tested reliably. A second tissue, such as skin fibroblasts, or buccal mucosa cells, or bladder epithelial cells, should be evaluated when a peripheral blood karyotype is normal and TS is still suspected. A large Danish study has set the fractions of various peripheral blood karyotypes and found that in 781 individuals with TS: 45% were found to have 45,X; 11% had a karyotype that included an isochromosome, such as 46X,i(Xq) or 45,X/46X,i(Xq); and the remaining 44% were mostly mosaic, such as 45,X/46,XX. Supporting further a high rate of mosaicism in TS, a review of 532 live born girls and women with TS found that approximately half were mosaic.

From a clinical and genetic counseling perspective, it is important to consider why and when a karyotype is obtained, as the timing of karyotype analysis often correlates with the phenotype severity in TS. Many prenatal karyotypes diagnosing TS have been done for advanced maternal age and compared with postnatal diagnoses, these individuals have decreased cardiac anomalies and decreased incidence of phenotypic features, such as neck webbing and low hairline, while characterized by a higher incidence of mosaicism. For the purpose of genetic counseling of 45,X/46,XY mosaics, it is important to highlight that in a study of 92 cases of prenatally diagnosed 45,X/46,XY mosaic cases, 95% had normal male genitalia.

A significant finding on a karyotype is the presence of Y chromosome material, which has implications in clinical management. A Y chromosome detected by karyotype is a risk for gonadoblastoma. In a study of 114 females with TS, 14 (12.2%) had Y material found by polymerase chain reaction (PCR), of which seven had detectable Y chromosome material on karyotype. Interestingly, of the 10 individuals in this study who underwent ovariectomy, only one (10%) had a gonadoblastoma. In another study of 171 individuals with TS, 14 (8%) were found to have Y chromosome material by Y chromosome repeat markers, and four of the 12 individuals who had a gonadectomy had gonadoblastoma detected. In a more recent study of 130 individuals with TS, three individuals were found to have Y chromosome material by karyotype and six by reverse transcription-PCR. Of these nine individuals with Y chromosome material, all underwent gonadectomy and one was found to have a gonadoblastoma. Although the aforementioned studies are heterogeneous in their methods, the risk of gonadoblastoma based on the best evidence available is roughly 10% if Y chromosome material is present. Current recommendation is for gonadectomy to be performed in TS individuals with Y chromosome material found on a standard karyotype analysis ( Fig. 17.2 ).

X Chromosome Genes and Turner Syndrome

Genotype-Phenotype Associations

Over 160 million years ago, the X and Y chromosomes were homologous, until the Y chromosome acquired the sex determining gene SRY and slowly lost many of its genes because of isolation from recombination by inversion. A dosage compensation system evolved to equalize expression of the X chromosome genes on the 46,XX female and 46,XY male. The compensation mechanism is X chromosome inactivation (XCI), which is achieved by the ribonucleic acid (RNA) coding gene XIST expressed on the inactive X chromosome; XIST then coats the inactive X chromosome making it inactive. However, 20% to 30% of X-linked genes escape X inactivation and are expressed on the inactive X chromosome. Pseudoautosomal region 1 (PAR1) located at the distal tip of the short arm of the X chromosome (Xp) and the short arm of chromosome Y (Yp) contains 24 genes that escape this X inactivation. The pseudoautosomal region (PAR) is divided into two regions, PAR1 at the terminus of the short arm of X and Y, and PAR2 on the long arm. The PARs pair and recombine during meiosis. There are four genes on PAR2 on both the X and Y chromosomes, and these genes, interestingly, do not escape X inactivation on the X chromosome. There are further 17 X-Y ancestral homologs, of which also 14 escape X inactivation. However, the remaining majority of X-linked genes that escape XCI do not have homologs on the Y chromosome. Genes that escape XCI are haploinsufficient in TS and have historically been candidates for genotype-phenotype investigations. However, few X genes have been correlated with the TS phenotype. The short stature homeobox containing gene on the X chromosome ( SHOX) is the most cited example ( Fig. 17.3 ).

SHOX is located at Xp22 on the pseudoautosomal region of the X chromosome. Both males and females have two copies of SHOX. SHOX , a transcription factor, has been implicated in short stature and the skeletal anomalies associated with TS. In a study of over 1600 prepubertal non-TS children with short stature, deletions or pathogenic variants in SHOX were found in approximately 4% of individuals. SHOX heterozygous variants, most commonly a deletion, may cause isolated short stature or other anomalies, such as mesomelia or the Madelung wrist deformity, observed in Leri-Weill dyschondrosteosis. The concept of haploinsufficiency of a gene, such as SHOX located in the PAR, is an attractive explanation for part of the TS phenotype. However, there is currently little evidence of any other monogenic variants on the PARs associated with the TS phenotype.

In addition to short stature, infertility and primary ovarian insufficiency are phenotypes for which a genetic etiology has been sought. Mapping partial X chromosome deletions in women, with and without TS, has provided loci associated with primary ovarian insufficiency (POI). Starting with the short arm of the X chromosome (Xp), TS phenotypes including POI have been mapped to Xp11.2-Xp22.1. In non-TS women, there is one gene located in this locus that has been associated with POI. Bone morphogenetic protein 15 ( BMP15 ) , located on Xp11.2, is an oocyte growth factor that stimulates folliculogenesis and granulosa cell (GC) growth, and has been implicated in POI. Di Pasquale et al. presented two sisters with a heterozygous missense variant in BMP15 (p.Y235C) and functional studies suggested a dominant negative mechanism. In a more recent study of 300 women with POI, six variants were found in BMP15 with three of the variants leading to marked decrease in protein production. Interestingly, there is a report of a small 554kb duplication involving the genes BMP15 and SHROOM4 in a woman with TS and spontaneous menarche, adding more evidence to BMP15’s association with POI.

On the long arm of the X chromosome (Xq), a number of X chromosome-autosome translocations in non-TS women have identified two additional loci or critical regions associated with POI: Xq13-Xq21 (critical region I) and Xq23-Xq27 (critical region II). In these critical regions, a small number of women without TS have shown associations with genetic variants and POI. In an X-autosome translocation [t(X;11)(q24;q13)] found in both a mother and daughter with POI, disruption in the gene progesterone receptor membrane component-1 (PGRMC1) was found, based on reduced gene expression levels. PGRMC1 mediates antiapoptotic action of progesterone and several cytochrome P450 reactions in ovarian cells. Mansouri et al. screened 67 females with POI and found one person with a p.H165R variant that was shown to interfere with cytochrome P450 7A1 (CYP7A1) binding to PGRMC1. In a large database of over 43,000 presumed healthy individuals, this particular variant was found in 0.2%. Located in Xq21, a report of five daughters with POI to a consanguineous couple identified a homozygous variant in premature ovarian failure protein 1B (POF1B ) . This variant was also shown to affect nonmuscle myosin binding. One prior report in a woman with POI demonstrated a translocation involving chromosome 1 and the X chromosome [t(X;1)(q21;p34)] with a breakpoint in the third intron of POF1B . POF1B is of interest as it escapes X inactivation and is located in POI1. Also located in critical region I, dachshund homolog 2 ( DACH2) was first identified by fine-mapping the breakpoint of an autosome-X chromosome translocation in a patient with POI. Two missense variants in DACH2 have been associated with women with POI. However, there is no functional studies' evidence to support DACH2’s role in POI.

Although not in the two critical loci noted earlier, two missense variants in androgen receptor gene (AR ) on Xq12 have been associated with POI in Indian women with POI, but the study did not evaluate these variants with functional studies. The androgen receptor is of potential interest as deficiency of AR in the mouse model results in POI.

A gene on the X chromosome worthy of special mention is fragile X mental retardation (FMR1) on Xq27.3. A premutation, 50–200 CGG repeats in the 5' untranslated CGG repeat, of FMR1 on Xq27.3 leads to methylation-coupled silencing of FMR1 and absence of the fragile X mental retardation protein (FMRP), causing the classical Fragile X syndrome. Fragile X syndrome is a common X-linked cause of cognitive impairment in males when the CGG repeat is greater than 200. Although haploinsufficiency of FMR1 is unlikely to cause POI in TS, a premutation is a common cause of POI, explaining 5% of sporadic POI cases and 10% to 15% of familial cases. The mechanism of POI in FMR1 premutation is thought to be an increase in messenger RNA (mRNA).

Short stature and POI are the most studied phenotypes in TS, and cognitive impairment will only be discussed briefly. The gene XIST has been associated with cognitive impairment in TS. Cognitive impairment is infrequently associated with TS and has mostly been observed in individuals with mosaic (short) ring chromosome [45,X/46,X,r(X)]. Cognitive impairment was first described in nine of 21 individuals with such a ring chromosome. The initial hypothesis for cognitive impairment was thought to be absence of the XIST gene on the ring chromosome, resulting in the failure of the X-inactivation of the ring chromosome. However, Leppig noted that in nine individuals with an r(X), eight had XIST genes present, questioning the original “absence of XIST ” hypothesis.

It must be emphasized that genotype-phenotype relationships in TS remain largely unknown. New approaches, such as methylation- and RNA expression studies are beginning to provide new answers. In a lymphocyte methylation and RNA expression study comparing TS women and 46,XX and 46,XY controls, women with TS were found to have genome wide hypomethylation altered expression in genes on both the autosomes and sex chromosomes compared with controls. Other studies have identified X chromosome microRNA that escapes X-inactivation and may play an important role in disease differences between males and females.

X Chromosome Genomic Imprinting

Imprinting studies for TS have shown mixed results. Skuse et al. proposed that individuals with 45,X who inherited a paternal X chromosome had a higher social cognition, meaning a significant difference in verbal and higher-order executive function skills. This study was based on 80 individuals with TS, of whom 55 had inherited a maternal X. In a smaller study of 39 participants with TS, Bishop et al. found that immediate and delayed recall was in the normal range for girls with TS. A study that looked at brain morphology in 40 girls with 45,X (23 girls with a maternal X and 17 with a paternal X) found differences in brain morphology based on X chromosome origin: those who inherited a paternal X had a thicker cortex in the temporal regions, and those who inherited a maternal X had enlargement of gray matter in superior frontal regions. Lepage et al. also noted that girls with a paternal X had a lower full-scale IQ than those that had inherited a maternal X. In contrast to these studies, a study of 50 girls with TS showed no association between parental X and IQ scores, as determined by the Wechsler Intelligence Scale for Children. Finally, for the most recognized phenotype of short stature, there is no evidence of X chromosome imprinting based on a small study of 25 individuals with TS.

Diagnostic Tests

There are multiple testing options available for TS diagnosis, both prenatally and postnatally. As noted earlier, the first diagnosis was made in 1959 by karyotype testing and this method continues to be the standard of care. A karyotype can be made from any metaphase or prometaphase cell that is actively dividing, including peripheral blood lymphocytes, amniocytes, and fibroblasts. Chromosomal microarray has the same level of accuracy as a karyotype, but has the advantage of more precise chromosomal breakpoints in the case of partial deletions and translocations. Although not a first-line test for TS, next generation sequencing can accurately diagnose TS and has the advantage of being able to simultaneously diagnose single gene variant conditions such as Noonan syndrome. FISH may be used to diagnose TS and has the advantage of being faster than karyotype testing, and can be performed on nondividing cells, such as those derived from buccal swabs.

Indications for Karyotype Testing

There is a distinct biphasic pattern in the diagnosis of Turner syndrome with a substantial proportion being diagnosed around the time of birth, and another large group being diagnosed around 12 years of age ( Fig. 17.4 ). When there is a clinical suspicion of TS, the American College of Medical Genetics recommends a 20-cell karyotype that should be able to detect mosaicism greater than 11%. For a prenatal 45,X karyotype diagnosis, TS should be confirmed with a postnatal karyotype .

Differential Diagnosis

When TS is suspected prenatally or postnatally, other genetic etiologies should be considered. In the prenatal period, TS may be diagnosed with noninvasive prenatal testing (NIPT) or CVS. Confined placental mosaicism, when the placenta is 45,X and the fetus is euploid, is in the differential diagnosis and should be resolved with amniocentesis. In addition to NIPT and CVS, prenatal ultrasound findings may suggest TS and other syndromes ( Fig. 17.5 ). The following prenatal ultrasound findings may be found in other genetic syndromes: hygroma colli, fetal hydrops, coarctation of the aorta, and increased nuchal translucency. In first trimester ultrasound of 185 singleton pregnancies with cystic hygroma colli, the most frequent chromosomal abnormality was TS ( n = 49; 26.5%), followed by trisomy 21 ( n = 32; 17.3%), and trisomy 18 ( n = 27; 14.6%). In this same study, 15 euploid fetuses were tested for Noonan syndrome and six (40%) were found to have it. In 65 cases of hydrops fetalis, six (9.2%) were diagnosed with TS, three (4.6%) with trisomy 21, and two cases with Noonan syndrome (3%).

Postnatally, TS may be suspected in childhood because of short stature, or during the pubertal age because of (primary) amenorrhea. Short stature affects 3% of children and there are numerous genetic etiologies. Common etiologies on the differential diagnosis include Noonan syndrome and haploinsufficiency for the SHOX gene. Although Noonan syndrome has been labeled “male TS” by some in the past, the facial, cardiac, and development features are quite different between these two syndromes.

Prenatal Diagnosis

NIPT has changed the timing and accuracy of genetic prenatal screening, especially for aneuploidy syndromes, such as TS. NIPT is based on small fragments of circulating cell-free fetal deoxyribonucleic acid (ccfDNA) that originate from placental cells that are continuously released, and can be detected as early as 5 weeks of gestation. Approximately 10% to 15% of cell free DNA in the maternal circulation is from the placenta, when measured between 10 and 20 weeks of gestation, and the half-life of ccfDNA is less than a day. There are a number of methods using NIPT to screen for TS prenatally, and most techniques count the number of copies of sequenced DNA fragments using next generation sequencing techniques.

In a Cochrane database systematic review, a pooled analysis of over 7000 high-risk pregnancies involving 45,X, in 12 studies using massive parallel shotgun sequencing, showed a clinical sensitivity of 91.7% (95% confidence interval [CI], 78.3–97.1) and a clinical specificity of 99.6% (95% CI, 98.9–99.8). The Cochrane database study found similar results for roughly 1000 high-risk individuals in four studies, using targeted massive parallel sequencing, which only evaluates select loci of the genome. Even with high sensitivity and specificity, the positive predictive values for NIPT for TS may be as low as 23% to 26%, which strongly reinforces the need for confirmatory testing. The American College of Obstetricians and Gynecologists recommends that all women with a positive cell free DNA result have a diagnostic procedure before clinical measures, such as pregnancy termination is taken.

The diagnosis of monosomy X is possible in embryos formed after in vitro fertilization (IVF) and is termed preimplantation genetic diagnosis ( PGD ) or preimplantation genetic screening . A biopsy of a blastomere at the cleavage stage or a trophectoderm biopsy at the blastocyst stage allows for an aneuploidy assessment, using techniques including FISH, chromosomal microarray, and next generation sequencing. PGD analysis of IVF embryos is motivated by the belief that selection of euploid embryos will improve IVF outcome. However, to date, there is no evidence that PGD improves the live birth rate after IVF. More research is needed in PGD applications to aneuploidy, especially with respect to diagnosing TS.

Lymphedema

Lymphedema is relatively common in TS and begins in utero. Fig. 17.7 demonstrates the generalized edema that can be observed in many 45,X fetuses. It is the consequence of malformations of the lymphatic vessels, resulting in lymphatic hypoplasia and obstruction. A cystic hygroma is a fluid-filled sac that often is located in the head/neck area, but may occur in other parts of the body, and results from abnormal communication between the jugular and central lymphatics draining into the heart. Peripheral lymphatic hypoplasia or aplasia has also been demonstrated. Using lymphangiography hypoplasia, and even aplasia, can also be observed in the peripheral lymphatic vessels of the extremities.

Clinical features resulting from nuchal lymphatic obstruction include webbing of neck, also called pterygium colli , and is observed in up to 25% of TS patients. Additional features that appear to result from this tenting and stretching of facial/scalp and neck skin is thought to result in low-set, rotated external ears, ptosis and downsloping of the eyelids, and the development of a low posterior hairline, with upward growing hairs (see Fig. 17.6 ). The occurrence of thick hair growth, including thick eyelashes and eyebrows, may also be, at least in part, a consequence of the distended cutaneous structures during fetal life. Swelling of the dorsum of the hands and feet at birth indicates the presence of a nonpitting type of peripheral edema. The lymphedema usually improves over the first 1 to 2 years of postnatal life. It may even resolve, but in many patients it persists to some degree. Some TS individuals demonstrate intermittent worsening of their peripheral edema, and this appears to be associated with puberty or the introduction of growth hormone (GH) or sex steroid therapy (both somewhat salt-retaining).

Growth failure, Short Stature, and Skeletal Anomalies

Short stature is the most common physical abnormality seen in TS. Poor statural growth affects essentially every patient to some degree. Their growth failure is disproportionate and they have a relatively large trunk with broad shoulders and pelvis; together with the reduced height this renders them with a “stocky” appearance. They often have a short neck because of hypoplasia of one or several cervical vertebrae. Long bone growth is preferentially impaired: patients do have relatively large hands and feet relative to their height, but the relatively short lower extremities lead to an increased upper-to-lower segment ratio. Selected bones can be more affected than others: an increased carrying angle at the elbow joint, also called cubitus valgus , is observed in nearly half of TS individuals. This carrying angle is the angle of intersection of the long axis of the upper arm with the long axis of the fully extended and supinated forearm. In adult normal females, this elbow angle is approximately 12 degrees (twice the size as seen in males). This angle is between 15 and 30 degrees ( Fig. 17.8 ) in many girls and women with TS because of abnormal development of the head of the trochlear bone.

The finding of a shortened fourth metacarpal, as shown on a bone age radiograph ( Fig. 17.9 A ) is also quite common. On physical examination, it manifests as absence or less prominence of the fourth knuckle when making a fist. Sometimes the fifth knuckle is also affected. Short fourth metatarsals can be found is some patients. The Madelung deformity is less common and is caused by a curvature of the distal radius in addition to a dorsal subluxation of the distal ulna ( Fig. 17.10 ). This anomaly also occurs as part of Leri-Weill dyschondrosteosis, and in both conditions is an indicator of a deficiency of SHOX. The previously described cubitus valgus and short metacarpals are also the result of SHOX deficiency.

At least 10%, and possibly as many as 20% of TS girls between 5 and 20 years of age develop scoliosis, defined as a lateral curve greater than 10 degrees. Kyphosis may develop in up to 40% of patients. Both findings likely are the result of vertebral development anomalies. Abnormalities of the development of facial bones and skull base result in a micrognathic and retrognathic mandible and/or high arched palate in 60% of patients.

Osteoporosis and fractures are more common in women with TS. Hand and wrist radiographs may reveal mild osteopenia ( Fig. 17.9 B), a coarse trabecular bone pattern, and ballooning of the tips of the terminal phalanges. This osteoporotic appearance is observed during childhood and adolescence, suggesting it may be related to a developmental role of SHOX inherent to TS rather than to primary estrogen deficiency, which may still be a contributory factor. The fact that cortical bone mineralization is selectively reduced in TS, independent of estrogen exposure, is consistent with this. Fractures of the long bones not associated with significant trauma are more frequently observed among girls with TS. It should be noted, however, that interpretation of studies of bone mineralization in young individuals with TS is difficult because areal instead of volumetric bone mineral density is measured. Nevertheless, prepubertal TS girls have decreased markers of bone formation because of a state of low-bone turnover and decreased bone deposition.

Poor linear growth in childhood and adolescence leads to adult TS women ending up about 20 cm shorter than their non-Turner female peers. The short stature in TS can be attributed to the deleterious effect of SHOX haploinsufficiency.

One of the first large-scale assessments of the growth failure present in TS was reported by Ranke et al. in 1983. Fig. 17.11 shows cross-sectional height and height velocity data from 150 TS children who did not receive growth-promoting therapy. Davenport and coworkers further elaborated on the different stages of impaired growth throughout a TS patient’s life. Prenatally, there is a mild degree of growth restriction resulting in a length at birth that is about 1 standard deviation (SD) below the mean. During infancy and up to 3 years of age, a further mild growth deceleration can be seen. Continued subnormal growth is observed throughout childhood, so that between ages 3 and 13 years, TS girls develop more deflection from both their target percentile and the normal height curves. There is failure to experience a normal pubertal growth spurt, and slow growth may be prolonged for several more years, potentially into the early 20s.

Height at diagnosis, ultimate height achieved, and midparental height are positively correlated. The adult height deficit is approximately 20 cm and this is irrespective of patients’ ethnic background. TS-specific growth curves have been developed by Lyon and associates, based on the data from Ranke and colleagues, as well as other European centers. These curves provide mean height and SD values for age and indicate a mean adult height of 143.1 cm (about 56.5 inches). A strong correlation between the initial height on these Turner curves and the adult height was also observed by Lyon and colleagues, independently of bone age at the time of the first height. This essentially allows for the adult height prediction of any girl with TS based on her height at an earlier age.