Chromosome Abnormalities and Genetic Counseling , 3rd Edition

15.Variant Chromosomes and Abnormalities of No Phenotypic Consequence

In this chapter we deal with three kinds of chromosomal variation having the quality in common that they carry no implications for abnormality in the person who has such a chromosome. First, there is the matter of normal variation. Normal chromosomes do not necessarily look exactly alike in different individuals, and some chromosomes show a remarkable degree of variation in their morphology. Obviously enough, it is crucial that the cytogeneticist distinguish normal variation from abnormality. Generally, there is no point in reporting a particular variant to the referring practitioner or to the patient. But sometimes it is necessary to pursue the matter with family studies when it is not clear whether a particular finding is a normal variant or an abnormality, or when a previous report has sown a seed of doubt in the patient's mind. The study of normal variants is also a research activity in its own right.

Second, we bring together in this chapter a group that grows somewhat out of the first, comprising structural rearrangements due to the translocation of harmless material, such as heterochromatin and nucleolar organizing regions, from one chromosome to another. Finally, we treat a group categorized as the “euchromatic abnormality without phenotypic effect.” These are chromosomal differences that do not comfortably accommodate the expressions “normal” and “variant,” and yet abnormality may convey too strong a sense. Anomaly may be a better word, or if “abnormality” is to be used, it should retain its quotation marks in the reader's mind. These are deletions and duplications that might at first sight have been thought to be abnormalities that would be associated with some clinical defect—deletions averaging 7–8 Mb and duplications as large as 14 Mb in size—but which are in fact observed in normal persons.

BIOLOGY

CHROMOSOME VARIANTS

If, in an individual, one chromosome of an homologous pair looks normal and the other has a different cytogenetic appearance, the latter may be said to be heteromorphic. Cytogenetic differences across the whole pool of chromosomes are referred to as variants.¹ Variants may be frequent or rare. The classical chromosome variants present a continuous spectrum, rather than a bimodal distribution. For example, in considering the differing lengths of the short arms of the acrocentric chromosomes, these could be arbitrarily classified as being of approximately typical length, somewhat shorter, somewhat longer, very short, or a lot longer. They are sufficiently variable that they were exploited as markers in research (in the days before DNA markers had been well developed). The distribution within a family is bimodal: the variant is either present or absent (barring introduction of a similar variant by an unrelated spouse). Some are so rare they qualify as private variants, recorded in just a single family. We consider chromosome variation in these three areas: the centromeric heterochromatin, the acrocentric short arms, and the fragile sites.

C-Band Size, Position, and Staining Properties

C-band heterochromatin comprises, by definition, permanently inactive DNA (constitutive heterochromatin), and is usually located adjacent to the centromere (C for centromeric). It stains darkly on C-banding. The four originally described variant forms are 1qh, 9qh, 16qh, and Yqh (qh for long arm heterochromatin); the differences in size were great enough to have been detected on solid-stain chromosomes in the prebanding era. C-bands vary in size, and for those chromosomes in which the material is centromerically placed, there is variation in position relative to the centromere (Craig-Holmes, 1977). The position of the centromere within the C-band–positive heterochromatin block of the 1qh, 9qh, and 16qh may vary from one end to the other. Variants at these extremities are sometimes referred to as inversions of the heterochromatin. The observed frequencies vary according to the precision of staining and the criteria of the observer (Kaiser, 1988).

The most common of these polymorphic inversions—and indeed the most common chromosomal variant in the human race—is the placement of 9q heterochromatin into 9p, immediately adjacent to the centromere. This high frequency presumably reflects the existence within the pericentromeric region of a number of hotspots for recombination (Starke et al., 2002). In about half of no. 9 chromosomes, a small amount (less than one-third) of the heterochromatic block is sited in the short arm. In about 10%, there is a “partial” inversion, with about one-third of the heterochromatin in the short arm, inv(9)(p11q12). In 0.6%, all the heterochromatin is in the short arm—a total inversion, inv(9)(p11q13). Rivera et al. (1999a) and Verma (1999) conducted a slightly testy debate about how many variant forms there actually are in the correspondence columns of Human Genetics. The wide possible range is presented in Starke et al. (2002). A single case of a somatically arising inv(9qh) in a patient with essential thrombocythemia may be more of a curiosity than a clinical relevance (Wan et al., 2000). Possibly the most impressive 9qh variant is that recorded in Lukusa et al. (2000), in which the segment 9p11–q13 underwent a tandem duplication. The abnormal chromosome at first sight may have offered a diagnosis for the mental defect of the woman in whom it was discovered. But her normal sister had the same chromosome, effectively exonerating it. Partial inversions of the heterochromatic region of chromosome 1, inv(1)(p11q12), are quite frequently seen; while total inversions, inv(1)(p13q21), are uncommon enough to be considered rare heteromorphisms (Fig. 15-1).

Whether an “inversion variant” chromosome can influence its own disjunction is speculative. Willatt et al. (1992) observed that in 11 reported cases of the rare mosaic trisomy 9 syndrome, four occurred in the setting of maternal heterozygosity for the inv(9)(p11q12) variant, and suggested a causal link; and partial trisomy 9 has been associated with a parental inv(9)(p11q12/3) (Kaiser, 1984; Stamberg and Thomas, 1986). These tiny numbers need to be seen against the background of the hundreds of thousands of inv(9) heterozygotes who have not come to cytogenetic attention for such a reason. Fortuitous coincidence remains a perfectly acceptable explanation.

Variation of C-band material can be demonstrated using other staining techniques, such as Q-banding, which reveals differing intensity of the fluorescence. Most notably, this variation is seen in the C-bands of chromosomes 3 and 4, which range from very dull to very bright. The staining of the C-band material, after one round of replication in BrdU, varies in the pattern of lateral symmetry (Angell and Jacobs, 1978). Heteromorphic staining of the centromeric regions of all chromosomes except chromosome 8 has been demonstrated using various restriction endonucleases to treat the chromosomes prior to Giemsa staining (Babu et al., 1988). Jabs and Carpenter (1988) showed that the extra C-band material adjacent to the centromere on the short arm of a variant chromosome 6 is due to increased amounts of a chromosome-specific alphoid DNA repeat sequence.

Other unusual C-band variants are reported for chromosome 3 (Petrovic, 1988), chromosome 5 (Fineman et al., 1989), chromosome 11 (Aiello et al., 1994), chromosome 18 (Pittalis et al., 1994), chromosome 20 with increased C-band material in the short arm (Fryns et al., 1988), and chromosome 20 long arm (Romain et al., 1991). The 20q variation reflects different amounts of chromosome-specific alphoid DNA. A duplication of the centromere itself is a different entity (Callen et al., 1990). Till et al. (1991), for example, report a duplicated no. 11 centromere, which they interpreted to be of no clinical consequence.

Morphological variants of the Y chromosome are in two categories. Continuous variation in the amount of C-band positive heterochromatin can range from a virtual absence, in which case the Y chromosome may appear to be only about half the size of chromosome 22, to a large amount such that the chromosome is about the same size as chromosome 13. Paternal chromosome study is worthwhile to confirm that very small chromosomes are in fact variants (short Y chromosomes with breakpoints proximal to the Yq11/12 heterochromatin interface are pathogenic deletions; Salo et al., 1995). For very large ones, C-banding is adequate to confirm that the increased size is due to heterochromatin. Discontinuous variation in the Y chromosome is expressed as a metacentric appearance, presumably because of pericentric inversion. Satellites on the end of the long arm are another variant; these are presumably due to translocation from one of the acrocentric chromosomes and have been documented to segregate in large kindreds over centuries (Genest, 1973) (and see below under Nuclear Organizing Region Translocation). These discontinuous variants are normally without phenotypic effect but, again, paternal chromosome study is warranted if there is any doubt.

Clinical Significance. Many studies have purported to show that variant chromosomes involving C-band size and position are associated with congenital malformations, malignant disease, habitual abortion, and infertility. There have also been many studies that report no such association. Carothers et al. (1982) conclude that “reproductive fitness of carriers of heterochromatic variants of the human karyotype is normal,” and that is the view we would espouse unless and until persuasive documentation to the contrary is adduced. In the specific case of the inv(9) variants, no increased reproductive risk exists that we can usefully measure in people in whom such a variant is discovered.

Acrocentric Short Arm Morphology

The short arms of the acrocentric chromosomes show a range of morphology, reflecting variation in three components of the short arm: the centromeric heterochromatin, the satellite stalk, and the satellite material (Figs. 15-2 and 15-3). These three components correlate with the three bands p11, p12, and p13. The p11 band contains several types of tandem DNA repeats, including satellites I, II, III, and beta satellite; band p12 contains multiple copies of genes coding for rRNA; and band p13 incorporates beta satellite (Piccini et al., 2001; Bandyopadhyay et al., 2001b). The nucleolus of the cell is formed by an aggregation of rRNA; thus, the stalk (band p12) is sometimes called the “nucleolar organizing region” (NOR). This stalk stains darkly with silver nitrate (Ag-NOR staining) on those acrocentric chromosomes that have an active NOR; and the picture varies from no uptake through marked uptake (darkness) of stain. A 3.3 kb repeat sequence, having homology to the repetitive DNA D4Z4 located at distal 4q, is found on the short arms of all the acrocentric chromosomes and has the characteristics of heterochromatin (Lyle et al., 1995).

Length and satellite size and number are distinguishing features. At one extreme, a short arm may seem to be absent; at the other extreme, it may be so long that a D-group chromosome is of C-group appearance, and a G-group chromosome has an F-group resemblance. Possibly the largest acrocentric short arm ever seen is the chromosome 15 described by Friedrich et al. (1996) in which the “short” arm was actually longer than the long arm. Molecular analysis of one chromosome 14 with an apparently absent short arm showed loss of satellite III DNA (Earle et al., 1992). Satellites vary widely in appearance: they can be apparently absent, small or large, and single or double. Variation can occur within the one individual (Ozen et al., 1995). A meiotic exchange can lead to a parent's variant satellite appearing on another chromosome in the child (Farrell et al., 1993)—the “jumping satellites of Gimelli.”

An apparent large acrocentric short arm can rarely mislead. Benzacken et al. (2001b) report a “14p+” chromosome discovered at late-pregnancy amniocentesis (36 weeks), a fetal macrosomy having been seen on ultrasonography; the 14p+ was initially considered to be a normal variant. When the child was born, a diagnosis of Beckwith-Wiedemann syndrome (BWS) was confirmed clinically, and the in-fant's karyotype was reassessed. The appearance of the NORs was judged on this occasion to be not quite typical. Analysis by FISH with 11p probes convincingly showed a t(11;14) (p15;p13), demonstrating that the abnormal chromosome was a der(14), and that the BWS was due to a partial 11p trisomy. The parents' chromosomes were normal. De Pater et al. (2000)describe a similar circumstance, in which a prenatally diagnosed de novo 14p+ “polymorphism”, which was negative on NOR staining, turned out actually to be a t(14;17) (p11;p11), and the child, who died as a neonate, thus had 17p trisomy.

Clinical Significance. Hassold et al. (1987) were able to demonstrate no association of NOR variants with an increased risk for trisomic conception, a possibility that had previously been postulated. An increased risk of having a child with Down syndrome had previously been proposed for the person with a “double NOR” on an acrocentric chromosome. This was a reasonable postulate, given the in vitro cytogenetic observation of “satellite association” and the possibility that a similar phenomenon could happen in vivo in meiosis, predisposing the chromosomes to aberrant segregation, and that this would be more likely with larger satellites. But the actual observation is, in fact, of no such increased risk (Serra and Bova, 1990).

Fragile Sites

A fragile site is a point on a chromosome that is liable to show gaps and breaks. The location of the fragile site is the same in all cells in a particular individual or family (Sutherland and Hecht, 1985). Fragile sites are classified on the basis of their frequency in the population and the conditions of tissue culture that are required to induce them.

In terms of frequency (ignoring fragile XA and fragile XE) there are three categories of fragile site. (1) Rare sites are present only in one per several hundred individuals. (2) Two fragile sites are of intermediate frequency: fra(10q25) is seen in 2.5% of individuals, and fra(16q22) is seen in 1%–5% of individuals. (3)So-called common fragile sites are universal and form part of normal chromosomal architecture. Only the rare and intermediate fragile sites are classified as chromosome variants (Fig. 15-4). The common fragile sites, being universal, by definition do not vary, although the proportion of metaphases in which they are seen can vary, depending on conditions, from 0% to 20%. The fragile sites at Xq27.3 (FRAXA) and Xq28 (FRAXE) are not, of course, normal variants, but pathological changes associated with mental retardation. These fragile sites and others on the distal long arm of the X chromosome are discussed in Chapter 14.

FRA11B is the fragile site at 11q23.3, within the proto-oncogene CBL2, which displays a molecular behavior similar to that of FRAXA (Jones et al., 1994). Jacobsen syndrome comprises variable deletions of 11q, and it may be that mothers of some affected children have a cytosine-cytosine-guanine (CCG) triplet expansion at FRA11B. However, in the majority of these children the deletion breakpoint is at a site other than 11q23.3 (Penny et al., 1995; Michaelis et al., 1998).

Clinical Significance. Fragile sites are harmless (except, of course, fragile XA, fragile XE, and, with remote possibility, FRA11B). The fragile sites of rare and intermediate frequency have been described variously as being associated with congenital malformations, sporadic chromosome abnormalities, and a predisposition to malignant disease. But there is no convincing evidence of any of these associations existing other than by chance (Sutherland and Baker, 2000). The CCG repeat at 11q23.3 might conceivably be a vulnerable point allowing de novo deletion of 11q (Jacobsen syndrome), but the evidence for this is by no means compelling. The fra(10)(q23) may lead to a del(10)(q23) being detected at prenatal diagnosis; this observation appears to be without phenotypic consequence (Zaslav et al., 2002). There is no apparent relation between autosomal fragile sites and mental retardation (Fryns and Petit, 1987). A theoretical reservation is that the fragile sites known as the autosomal folate-sensitive group (Table 15-1) have been seen only in heterozygotes. The fact that homozygosity for some folate-in-sensitive fragile sites (10q25, 16q22, 17p12) is on record and that the individuals studied had no consistent phenotypic abnormality offers further reassurance of the fundamental biological harmlessness of these morphologic curiosities.

TRANSLOCATIONS OF HETEROCHROMATIN AND NUCLEOLAR ORGANIZING REGIONS

Y Heterochromatin Translocations

In the sense that they cause no phenotypic abnormality, the so-called t(Y;15) and t(Y;22) translocations, in which the C-band material from the long arm of the Y has been translocated onto the short arm of the acrocentric, can be regarded as examples of normal heteromorphism (Friedrich and Nielsen, 1972; Cohen et al., 1981; Neumann et al., 1992). Once formed, these variant chromosomes are stable (see also p. 115). In one reported family, a girl with Prader-Willi syndrome had a de novo 15q11–q13 microdeletion on a familial t(Y;15)(q12;p11) chromosome. Her father, brother, uncle, and two aunts also carried the intact t(Y;15) (Eliez et al., 1997). Only a slender argument could be made that the t(Y;15) may in some way have predisposed to the formation of the deletion.

A segment of Y chromosome heterochromatin can be insertionally translocated into another chromosome, and this may be without any phenotypic effect. Ashton-Prolla et al. (1997) report detecting such an anomaly at prenatal diagnosis, with the karyotype only interpretable after the father had been studied. He had a segment from Yq12 inserted into chromosome 11 at band 11q24: 46,XY,der(11) ins(11;Y)(q24;q12q12). He had inherited this ins(11;Y) from his mother, and her normality, and subsequently his infant daughter's normality, attest to the innocuousness of this variant chromosome. A similar case is described by Spak et al. (1989) of a harmless C-band–positive insertion at 11q23.2; the origin of the heterochromatin could have been an autosome or the Y chromosome.

Clinical Significance. Unless otherwise proven, these Y heterochromatin translocations are to be regarded as variants without phenotypic consequence.

Nucleolar Organizing Region Translocation

The NOR can be translocated to another chromosome, usually to a terminal region; the nomenclature of ps or qs denotes a satellite (that is, a NOR) at the tip of the arm in question. Willatt et al. (2001a) record examples of this phenomenon, either as terminal or as interstitial NORs, in chromosomes 1, 2, 4 (several cases), 6, 7, 8, 10, 12, 17, and 22. Faivre et al. (1999) discuss the recorded cases identified at prenatal diagnosis, noting that in only 1 of 13 was there an otherwise unexplained abnormal phenotype, and acknowledge that this could have been a coincidental association.

The anomalous chromosome can arise somatically, as Storto et al. (1999) show in a father with 46,XY,10qs/46,XY mosaicism, whose (normal) child was diagnosed with the 10qs at amniocentesis. Wilkinson and Crolla (1993) showed in the Yqs chromosomes segregating in three families that the NOR on the Y in each had arisen from a 15p to Yq translocation, as did two of the four cases of Kühl et al. (2001). These Yqs chromosomes have lost their PAR2 region, but this is of no discernible clinical consequence. Reddy and Sulcova (1998b) undertook detailed molecular dissections on three NOR translocations: a chromosome 21 into the short arm of which segments of chromosome 15 beta-satellite elements were inserted, making it “tricentric”; a chromosome 7 with acrocentric beta-satellite DNA inserted at 7p13; and a pseudodicentric chromosome 2 with beta-satellite inserted at the terminal tip of 2q.

Notwithstanding the typical circumstance of harmlessness, the NOR translocation can, in rare instances, be pathogenic. This applies particularly in the setting of the translocation process having led to a concomitant deletion at the tip of the recipient chromosome. Chen et al. (2000b) report their experience with a “4ps” in a mother and Wolf-Hirschhorn syndrome in her child, but in fact this reflected unbalanced segregation from a reciprocal translocation t(4;15)(p16;p11.1), and this circumstance should not unduly taint the reputation of the satellited chromosome. Their second case of an Xqs chromosome is more telling, in which a 46,X,Xqs mother had an abnormal 46,Y,Xqs son. They proved the point that there was a molecular distal Xq deletion, and so the abnormal chromosome might more accurately be described as Xpter → Xq28::sat, rather than Xpter → Xqter::sat. A notable example of a possibly pathogenic insertional NOR translocation is that described by Tamagaki et al. (2000) (and see p. 168), in which an NOR inserted interstitially at Xq11.2 was segregating with a spastic paraplegia in the males. The cosegregation could be coincidental, but equally might reflect disruption of a causative X-borne locus due to the NOR insertion.

Clinical Significance. In general, these satellited chromosomes are to be regarded as harmless heteromorphisms. Faivre et al. (2000a) reasonably comment, in their study of a familial 4qs ascertained through a child with a cerebellar ataxia, that “genetic counselling should be reassuring” in the setting of an NOR translocation identified at prenatal diagnosis. We note above the rare exceptions to this rule, in which the translocated NOR may have caused genomic disruption at the site of translocation. The lesson from these cases is that any satellited chromosome deserves detailed study (especially an X chromosome NOR not known to be carried by a phenotypically normal male in the family), while retaining the perspective that very probably the conclusion will be that it truly is a harmless variant.

CHROMOSOME ANOMALIES

Euchromatic Abnormalities of No Phenotypic Consequence

The G-band pattern is generally constant, and the relative sizes of G-bands are similar among the karyotypes of any members of the human race. The fact of being a diploid species is useful, enabling one homolog to be a control for the other. Most of the banding pattern variation observed by the cytogeneticist is artifactual. The level of resolution, which is a function of the degree of compaction of the chromatin, will determine the number of bands seen in any metaphase, and even within a metaphase the number of bands on homologous chromosomes may vary (“homolog asynchrony”). True anomalous forms exist from which, to the best of our understanding from the clinical assessments that have been done, no phenotypic consequences flow. They may be classified, following Barber (2000), according to the following three categories:

1. Euchromatic “variants” due to constitutional cytogenetic amplification

2. Euchromatic deletions and duplications without phenotypic effect

3. Euchromatic deletions and duplications of uncertain phenotypic effect.

A distinction between one of these anomalies and a pathogenic cytogenetic abnormality may not always be easy. It is generally considered that if, in a family study, all the individuals (with the usually allowable exception of the index case) having the unusual chromosome are phenotypically normal, then the chromosome really is a euchromatic abnormality of no phenotypic consequence. Some chromosome anomalies, judged to be harmless by this criterion, might have a euchromatic deletion or duplication that is actually of similar extent to some of those which are undoubtedly pathogenic. How can there be such a difference in outcome: a normal or an abnormal individual? What matters is the nature of the loci in the segment of interest. Is this a region of low gene content? Deletions of G-band dark chromatin may typify the harmless deletion: these regions are poor in CpG islands and thus presumably accommodate fewer genes. Are the genes functional or merely pseudogenes, fossilized relics of evolution preserved in the chromatin? If the latter, it is to be presumed that no harmful consequence would arise from their loss or surplus. Has there been a position effect, whereby the new location of the genetic material leads to its inactivation? Are the genes dosage-sensi-tive? If a gene, or a set of genes, functions equally well at 50% as at 100%, then a deletion on one chromosome will be without any effect on phenotypic expression. Such genes could be thought of as functioning “haplosufficiently.” (Presumably, a 0% amount of a block of genes would cause abnormality, and the anomalous chromosome would have become a pathogenic one. The rarity of these anomalies has not yet permitted such an observation to be made.) Finally, the interpretation of phenotypic normality needs to made with some confidence. While perhaps unlikely, it could be that a deletion or duplication in a gene-sparse region might lead to only a very subtle compromise of form and function, just within the range of normality, and the chromosome abnormality is actually pathogenic.

Euchromatic “Variants” due to Constitutional Cytogenetic Amplification

Three chromosome regions are known in which an additional band or bands may infrequently be discerned, this band reflecting the presence of an amplified DNA cassette. When the amplification reaches a certain degree, the extra band is detectable cytogenetically, and the observation can be described as a “constitutional cytogenetic amplification.” The additional material comprises solely or substantially pseudogenes. It will not be surprising if three further variant forms (all of chromosome 9) turn out to be in the same category. The regions concerned are 8p23.1, 9p12, 9q12, 9q13, 15q11.2, and 16p11.2. To the best of current knowledge, no clinical significance attaches to the discovery of any of these variants (Barber, 2000).

Chromosome 8p23.1. Several families with a duplication of 8p23.1 have been described, in which there has been transmission of the chromosome through numerous phenotypically normal family members (Barber et al., 1998; Barber, 2000; Engelen et al., 2000b). Abnormality in the index patients in the series of Tsai et al (2002) may reflect ascertainment bias. The anomalous band appears as a fine G-dark band in the center of 8p23.1, and could be designated 8p23.12. The additional material comprises two or three extra copies of a domain that includes defensin and olfactory receptor loci.

Chromosome 9. The best studied of the three no. 9 candidates for “euchromatic variant” is the chromosome in which an extra G-band is inserted within the C-band heterochromatin of the long arm (9q12 or 9qh). Wojiski et al. (1997) have demonstrated that the source of the additional material is from the short arm, the (euchromatic) segment 9p12. The likely mechanism for its being harmless in the duplicated state is that it has been “heterochromatinized,” inactivated by virtue of being sandwiched between segments of 9qh—a possible example of a position effect (Macera et al., 1995). Otherwise on chromosome 9, there is a short arm variant in which an extra dark G-band appears to be inserted into 9p12 (Fig. 15-5), and a long arm variant in which there is extra G-band–positive C-band–negative material at band 9q13–q21 (Knight et al., 1993a).

Chromosome 15q11.2–q13. Several authors have reported individuals with apparent duplication of 15q11.2–q13 euchromatin but without any consistent phenotypic effect (and which can make the other homolog seem, by comparison, to have a Prader-Willi/Angelman deletion). Barber et al. (1998) summarize the history. These workers and Ritchie et al. (1998) and Fantes et al. (2002) have shown that the duplicated region in at least some of these chromosomes is due to a variably increased number of a cassette of nonfunctional pseudo-genes. Pseudogenes of neurofibromatosis type 1, immunoglobulin heavy chain, a γ-amino-butyric acid receptor (GABRA5), and a B-cell lymphoma gene (BCL8) comprise the cassette. Note that this particular dup(15) is to be contrasted with pathogenic duplications of 15q13 which include the Prader-Willi/Angelman critical region (PWACR). Here the chromosome abnormality is of itself the basis of an abnormal phenotype (p. 290). The distinction can be made by probing with a centromere and a PWACR probe: in the euchromatic anomaly, the distance between these two markers is increased, while in the pathogenic duplication, it is not (Barber, 2000).

Chromosome 16p11.2. Numerous families have been reported with a variant of chromosome 16 in which there is an extra dark G-band, C-band negative, inserted into 16p11.2 (Barber et al., 1999). This is euchromatic material, consisting of additional copies of a pseudogene cassette that includes nonfunctional sequences and evolved from a chromosome 14 immunoglobulin locus, an X chromosomal creatine transporter gene, and the adrenoleukodystrophy domain.

Euchromatic Duplications and Deletions Without Phenotypic Effect

Thinking purely in mechanical and structural terms, these duplications and deletions may be no different from a pathogenic abnormality of similar size. That is, a similar process has led to the addition or removal of a small segment of chromatin. As discussed above, it is likely that the nature of the segment in question is what determines harmlessness or pathogenicity: put simply, a gain or a loss of genes that does or does not matter. Or, if it does matter, it may matter very little, and the effect of the gain or loss may be so mild that the phenotype is only subtly influenced, which thus remains well within the sphere of normality.

Quite a number of these duplications and deletions have been reported (and several more unpublished examples are likely known in a number of cytogenetics laboratories). Barber (2000) has undertaken a review, and Table 15-2 sets out the harmless deletions and duplications he recorded. Most of these unusual chromosomes were detected in normal persons, and the same then in normal relatives, prenatal diagnosis being a frequent route of ascertainment. In other cases (Table 15-3), the chromosome was first discovered in an abnormal individual, but then the same chromosome was identified in other phenotypically normal relatives. This observation of normality in other family members allowed the presumption of the chromosome being simply a variant. In a few cases, the child's abnormality may have been the direct result of the chromosomal “variant,” the parent's normality notwithstanding. Possible explanations for this discordance include an effect due to imprinting and, in cancer, the operation of “first and second hits.” In terms of the latter, a normal parent with a del(13)(q14) “variant” had a child with a retinoblastoma (Cowell et al., 1988), and it would be supposed that the tumor arose following a second hit, in a retinal cell, at the RB locus on the child's other chromosome 13. Presumably the unaffected heterozygous parent had the good fortune of not undergoing a second hit in early life. And presumably that par-ent's normality and the child's otherwise could be explained on the basis that a 50% output from any other loci involved in the deletion was functionally sufficient.

The distinction between a euchromatic abnormality of no phenotypic consequence and one of slight consequence may be a very subtle one. Consider the case of a girl described in Bonaglia et al. (2002) who had a karyotype ordered on fairly slender grounds at age 18 months (she was in hospital at the time for an ear infection), but as a 6½ year old had only the most minor signs (one ear ½ cm longer than the other, and crowded teeth, for example), and IQ testing at 5½ years had shown an average intelligence. She had a de novo dup(9)(p22 → p13), proven to have arisen following an asymmetric sister chromatid exchange at paternal meiosis. Would she have been slightly different in appearance, her dentition less imperfect, and her IQ a little higher, if she had been 46,XX? Or would she have been just the same, and the dup(9p) was truly a euchromatic abnormality of no phenotypic consequence? Since dental crowding is seen in the classical 9p duplication syndrome, perhaps this aspect truly is due to the karyotypic imbalance. This discussion again illustrates the point that, at the fundamental level of cytogenetic mechanism, at least some of the abnormalities with and some without phenotypic effect may have arisen in essentially the same way, and the difference depends upon whether or not the genes duplicated or deleted were dosage-sensitive (and recognition naturally depends upon whether their effects are observable at clinical examination).

The X Chromosome

X chromosome anomalies have the special quality that the other normal X in the female may suffice to endow phenotypic normality, while in the male, nullisomy for the segment may be of severe effect. For example, Schinzel (2001) lists two cases of 46,X,del(X)(q13.3 q21.3), one in a normal woman, and the other in a severely retarded and dysmorphic girl. The former had been ascertained through having had a son with multiple malformations. We have seen a girl who is physically quite normal but has epilepsy, learning difficulty, and dysarthria, with this same karyotype, de novo in her. The inconsistency of phenotype, and the fact of one case being normal, suggests that the abnormalities in the other two may have been due to different causes. The del(X)(q13.3q21.3) may be a “deletion without phenotypic effect” in females.

Subtelomeric Deletions and Duplications

Some “deletions” and “duplications” identified on subtelomeric FISH probing may well be true findings, but have no effect on the phenotype (Riegel et al., 2001). Ballif et al. (2000b) record, for example, the following (sub)telomeric variants: ish add(1)(13qtel+) using PAC probe 163C9; ish del(2)(qter) with PAC 1011O17 and P1 210E14; ish del (9)(pter) with PAC 43N6; and ish del(X)(pter) with cosmid CY29. The harmlessness of these variants, in all probability, was demonstrated by showing the same thing in a parent. Many more such variants are likely to be discovered as cytogeneticists continue to focus attention on the subtelomeric regions of the chromosomes. Intelligent interpretation of these findings will require a familiarity with the current literature, an awareness of the best probes to use in order to distinguish between variants and true pathological abnormalities, and good technical skill. Subtle subtelomeric rearrangements may be associated with a less abnormal phenotype that will make patient selection for such studies difficult, and more examples of familial transmission are likely to come to light (Baker et al., 2002b).

MEIOSIS

Variant chromosomes, being normal chromosomes, behave normally at meiosis; and other things being equal, 1:1 segregation occurs. Hence any individual with a variant chromosome transmits it to, on average, half of his or her offspring. In the specific case of the inv(9)(p11q13), an essentially 1:1 ratio was confirmed on a sperm study (Colls et al., 1997). An exception is the autosomal folate-sensitive fragile sites. Unexpectedly, when these are transmitted by a male, only one-quarter of the offspring appear to have them (Sherman and Sutherland, 1986). This may be an effect of unstable DNA sequences, which are amplifications of CCG trinucleotide repeats, and of which these fragile sites are comprised (Nancarrow et al., 1994). The expanded CCG repeats behave differently when transmitted by males and females, as best studied for FRAXA, but the expansion shows 1:1 segregation. It is the degree of expansion that determines whether a fragile site is present, and this does not follow Mendelian inheritance.

GENETIC COUNSELING

A person carrying a variant chromosome has, by definition, no increased risk for having abnormal offspring, abortion, or any other reproductive problem. Some see it as at best pointless and at worst counterproductive even to mention to the individual that a variant chromosome has been found; others feel obliged to pass on the observation. If it is discussed, it must be made clear that it is a normal finding—perhaps interesting, but of no practical importance. Some patients may be intrigued to learn that they are of interest to genetic researchers. For the size variants (C-band and NOR), the point can simply be made that some chromosomes come in short, medium, and long forms, and where a chromosome happens to fit in this continuum is without significance. There is considerable potential for iatrogenic anxiety, whereas in reality the biology of the supposed anomaly has no pathogenic implication. The counselor may thoroughly understand the presumed harmlessness of a variant chromosome, but the person in whose family it has been discovered may react “non-scientifically.” The worst possible response might be for a couple to choose to terminate a pregnancy because of an overinterpreted variant chromosome. “Primum non nocere”—first do no harm.

From time to time, children with delayed neurodevelopmental progress with or without minor dysmorphic features are identified as having a subtle deletion or duplication; and in most, of course, the karyotypic abnormality will have been the cause of the phenotypic defect. Occasionally, parental studies will come up with the surprising result that one parent has the same karyotype; and a reinterpretation of “harmless variant” may be made. The counselor will need to make sure that this reinterpretation is well understood; and that the label of “chromosome abnormality” does not attach to the child and stifle other clinical investigation. These unusual examples should ideally have a full family investigation to solidify the case for or against the innocuousness of the chromosomal finding.

It is necessary to emphasize (and see Chapter 3) the confusion and misinterpretation that can be due to ascertainment bias. Practically by definition, a child having a chromosome test has to be abnormal; therefore it follows that a variant chromosome newly discovered in this context is associated with phenotypic defect. Association, however, does not necessarily mean there is a causal link; and if a normal parent or other relative has the same unusual chromosome it should be seen as being harmless unless proven otherwise, notwithstanding various esoteric scenarios that any imaginative geneticist could devise. One author should, in honesty, acknowledge an early paper that exemplifies this type of error: in this case, an overinterpretation of the clinical relevance of the 1qh variant (Gardner et al., 1974).

Malignant Disease. Persons who have variant fragile sites (we all have the common fragile sites) may be advised, if the question is raised, that there is no evidence they are at increased risk for developing any malignant disease, despite suggestions that have been made to the contrary. These suggestions have arisen from observations of an apparent correlation between the location of fragile sites and the chromosome bands within which the characteristic breakpoints of some tumor chromosomal rearrangements occur (Hecht and Sutherland, 1984; LeBeau and Rowley, 1984; Yunis and Soreng, 1984). The case for such a correlation has largely been refuted (Simmers et al., 1987; Sutherland and Simmers, 1988). There is emerging evidence that some of the common fragile sites have within them tumor suppressor genes that can be disrupted in some cancers (Arlt et al., 2003). But there is no indication that any individual in whom a particular common fragile site may be prominent (in terms of the proportions of metaphases in which it is seen) is at any increased risk of cancer.

Note

1. The word polymorphism is also used as a cytogenetic colloquialism in this context, but, to be precise, it is an inaccurate usage. The genetic term polymorphism refers to an allele that has a frequency of <1% in the population, and although some chromosome variants may be more common than this, each one, e.g., 9qh+, is not homogeneous but a collection of many “alleles” at this locus.