Inversions are intrachromosomal structural rearrangements. The most common is the simple (or single) inversion. If the inversion coexists with another rearrangement in the same chromosome it is a complex inversion. The simple inversion comprises a two-break event involving just one chromosome. The intercalary segment rotates 180° and reinserts, and the breaks unite (Fig. 8-1). The rearranged chromosome consists of a central inverted segment and flanking distal or noninverted segments. If the inverted segment includes the centromere, the inversion is pericentric; if it does not, it is paracentric. Figure 8-2 depicts two different pericentric inversions of chromosome 3. Note that the pericentric inversion has one break in the short arm and one in the long arm, whereas in the paracentric inversion both breaks occur in the same arm. Thus, when reading cytogenetic nomenclature, one can readily tell which is which—for example, 46,XX,inv(3)(p25q21) is pericentric and 46,XY,inv(11)(q21q23) is paracentric (inv = inversion). The clinical relevance of inversion chromosomes is that they can set the stage for the generation of recombinant (rec) gametes that may lead to abnormal pregnancy.
The heterozygote is, other things being equal, a phenotypically normal person. The reorientation of a sequence of genetic material apparently does not influence its function, and breakage and reunion at most sites do not perturb the smooth running of the genome. Some inversions of the X may be an exception to this rule: a breakpoint involving the X long arm within the critical regions can cause gonadal insufficiency. Some pericentric breakpoints occur at preferential sites, including 2p13, 2q21, 5q13, 5q31, 6q21, 10q22, and 12q13 (Kleczkowska et al., 1987), and certain para-centric breakpoints are likewise overrepresented (Madan, 1995). An inversion may not necessarily be detected on routine study, and knowing when to mount a directed search requires clinical acumen. Thus, Yokoyama et al. (1997) discovered an inv(17)(p13.1q25.1) in a father whose child had lissencephaly, a particular type of severe brain malformation. At first glance, the inverted chromosome looked normal. They noted a family history of similarly affected children, suspected a diagnosis of Miller-Dieker syndrome (which is due to 17p13.3 deletion; p. 285), and went on to demonstrate the cytogenetic abnormality using FISH with a Miller-Dieker chromosome 17 cosmid probe.
In the event that a breakpoint occurs within a gene, the inversion could be directly pathogenic. Rare examples include a de novo inv(16)(p13.3q13) disrupting the Rubinstein Taybi syndrome locus, a de novo inv(17) (q12q25) disrupting SOX9 and causing campomelic syndrome, a familial inv(X)(p11.4q22) compromising the Norrie syndrome gene, a de novo inv(20)(p12.2p13) with one breakpoint occurring between exons 5 and 6 of the JAG1 gene, causing Alagille syndrome, and (as potential examples) a familial inv(5)(p15.1q11.2) cosegregating with benign neonatal convulsions, and an inv(3)(p14q21) with a develop-mental-behavioral phenotype (Maraia et al., 1991; Lacombe et al., 1992; Pettenati et al., 1993; Stankiewicz et al., 2001b; Concolino et al., 2002; Efron et al., 2003). A de novo inv(2)(q35q27.3) provided, in fact, the entrée to the mapping of the Waardenburg syndrome locus to 2q35 (Ishikiriyama et al., 1989). Likewise, an inv(15)(q11.2q24.3) transmitted from a normal mother to her Angelman syndrome daughter led to the cloning of the UBE3A gene, absence or dysfunction of which is the basis of the syndrome (Greger et al., 1997). On the X chromosome, Xu et al. (2003) describe a family with congenital androgen insensitivity (see p. 296) segregating an inv(X)(q11.2q27). Pre-sumably, the break at Xq11.2 compromised the integrity of the androgen receptor locus. An inversion X chromosome with gene damage at both breakpoints was reported by Saito-Ohara et al. (2002). A mother with the karyotype 46,X,inv(X)(p21.2q22.2) had a severely retarded 46,Y,inv(X)(p21.2q22.2) son with Duchenne muscular dystrophy, these effects being due to disruption of the dystrophin gene at Xp21.1 and of a novel gene RLGP at Xq22.2 (there was also duplication of the proteolipid protein gene at Xq22.2).
|
|
|
Figure 8-1. Structure of the pericentric (left) and paracentric (right) inversions. The inverted segment is cross-hatched. Asterisks provide landmarks at each end of the inversion segment. |
|
|
|
Figure 8-2. Two pericentric inversions of chromosome 3. Both of the noninverted segments are small in one (a) and one non-inverted segment is large in the other (b), (cases of N. A. Monk and L. M. Columbano-Green). |
New methodology enables the recognition of inversions or recombinants that might otherwise have escaped detection. A submicroscopic deletion may be generated during the formation of a paracentric inversion, so that an apparently balanced rearrangement at the microscopic level is in fact genetically unbalanced. Langer-Giedion syndrome (LGS) is due to a deletion at 8q24.11–q24.13 (p. 282), and Sasaki et al. (1997) studied a child with LGS who had a de novo inv(8)(q13.1q24.11). Molecular analysis revealed a 4 Mb deletion encompassing the LGS region; presumably this segment had been deleted as part of the process that generated the inversion. A familial inv(18)(q21.1q23), in which a gene for brain myelination and presumably some adjacent genes were deleted, led to some features of the 18q syndrome in a mother and daughter (Keppler-Noreuil et al., 1998). Subtelomeric probing showed a cryptic deletion at 6qter in a de novo paracentric inv(6)(q22.1q27), apparently balanced on initial cytogenetics, in a child with developmental delay, microcephaly, and facial dysmorphism, reported by Lorda-Sanchez et al. (2000). Chia et al. (2001) studied a girl with an apparent del(2)(q37) on high-resolution analysis. Using subtelomeric probes to clarify the nature of the deletion, they were surprised to see a 2p signal at each end of the chromosome. Thus, the “deletion chromosome” could be seen for what it really was, a recombinant inversion chromosome, the essential genetic consequence of which was a deficiency of distal 2q. Since the short arm breakpoint was right at the tip of the chromosome cytogenetically, at 2p25.3, there may have been little or no duplication of functional 2p genetic material.
“Inversions” having a breakpoint within the heterochromatic regions of chromosomes 1, 9, 16, and Y are frequently seen and are to be thought of as variants, not abnormal chromosomes (see Chapter 15). The most common inversion in humans not involving centromeric heterochromatin is the inv(2)(p11.2q13); just two recorded cases in the world are known of a possibly related pathogenic recombination (see below, p. 153). Other presumed harmless inversion variants include the following: inv(3) (p11q11) and inv(3)(p11q12), inv(3)(p13q12), inv(5)(p13q13), and inv(10)(p11.2q21.2). The inv(10) has been rather extensively studied by a collaborative group of five laboratories in the United Kingdom, who between them had 33 families available for investigation (Collinson et al., 1997). They found no excess of infertility or spontaneous abortion among carriers.
|
|
|
Figure 8-3. Inversion loop in meiosis, direct observation. Left, inversion loop in a mouse study (courtesy Y. Rumpler). Right, spermatocyte study of a man with inv(6)(p22q22.2) (From de Perdigo et al., 1989, courtesy Y. Rumpler and with the permission of Springer-Verlag). |
Excluding these variant forms, inversions are a fairly uncommonly recognized rearrangement. Estimates of frequency range from about 0.12‰ to 0.7‰ (pericentric) and about 0.1‰ to 0.5‰ (paracentric) of individuals (Van Dyke et al., 1983; Kleczkowska et al., 1987; Worsham et al., 1989; Pettenati et al., 1995). Madan (1995) has commented that the paracentric inversion is “the most common form of chromosomal polymorphism found in nature,” and suspects that many small examples remain undetected.
Pericentric Inversion
BIOLOGY
AUTOSOMAL PERICENTRIC INVERSION
Meiosis
The inversion heterozygote may produce chromosomally unbalanced gametes and consequently suffer reproductive pathology. The chromosomal imbalance is a result of the formation of a recombinant (rec) chromosome. This is “aneusomie de recombinaison,” or aneusomy due to recombination. Recombination occurs within the inverted segment if there is a crossover between the inversion chromosome and the normal homolog.
Synapsis and Recombination
Classically, crossing-over follows the reversed loop model (Figs. 8-3 and 8-4). This configuration of the bivalent allows as complete as possible alignment and pairing of matching segments of the inversion chromosome and its normal homolog (homosynapsis). One (or an uneven number of) crossover(s) within the inversion loop, between a chromatid of the normal homolog and a chromatid of the inversion chromosome, leads to the production of two complementary recombinant chromosomes. One of these has a duplication of the distal segment of the short arm and a deletion of the distal segment of the long arm (chromosome c-c′ in Fig. 8-4); the pattern is the other way around in the other rec chromosome (d-d′ in Fig. 8-4). Thus, the conceptuses that result have either a partial trisomy for one distal segment and a partial monosomy for the other, or vice versa. Typically, only one of these, the least monosomic, is ever viable. Consider the recombinant 7 due to a paternal inversion illustrated in Figure 8-5. There is a duplication of the substantial segment 7p14.2 → pter, and a deletion of only the tiny segment comprising the distal-most sub-band of 7q (7q36.3 → qter). The countertype form having a monosomy for 7p14.2 → pter (and trisomy 7q36.3 → qter) would, we suppose, cause a miscarriage.
The cytogenetic nomenclature to describe the recombinant karyotype is straightforward. In the above case, for example, we have
Parent: 46,XY,inv(7)(p14.2q36.3)
Recombinant offspring (c-c′): 46,XY,rec(7)dup(7p)inv(7)(p14.2q36.3)
It is not necessary to include “dup(7p) del(7q)”—the complementary deletion is taken as read. More fully, the nomenclature is 46,XY,rec(7)dup(7p)inv(7)(pter → p14.2::q36.3 → p14.2::q36.3 → qter)pat
|
|
|
Figure 8-4. Inversion loop in meiosis, theoretical recombinant outcomes (based on the inv(3) shown in Fig. 8-2a). Both sister chromatids are shown. The inversion (centromeric) segment is cross-hatched, the long arm noninverted segment is stippled, and the short arm noninverted segment is open. The four possible gametic outcomes following one crossover within the inversion loop are depicted. Chromosomes a–a′ and b–b′ are the intact homolog and the inversion, respectively; chromosomes c–c′ and d–d′ are the dup p and dup q recombinant chromosomes. Compare with the actual observation in Figure 8-3, right. |
This complex twisting of the chromosomes to form a loop may not necessarily take place. In an inversion with a short inverted segment (Fig. 8-6a), a partial pairing may occur. Both distal segments, or sometimes just one, align in homosynapsis. The inverted segment and the corresponding part of the normal homolog either “balloon out” (asynapsis of the inversion segment) or lie adjacent but unmatched (heterosynapsis) (Gabriel-Robez and Rumpler, 1994). Thus, no crossing-over can happen within the inverted segment, and recombinant products do not form. Conversely, some inversions with long inverted and very short distal segments may undergo synapsis of the inverted segment only, with the distal segments at each end remaining unpaired (Fig. 8-6b). Recombination can occur in this setting. The quality of the chromatin may of itself have an influence. If both breakpoints are in G-light bands, the lack of homology is detected at synapsis, and the chromosomes respond by formation of a loop, achieving a complete homosynapsis. If, however, one or both breakpoints are in a G-dark band, non-homology may not be recognized, and heterosynapsis is not prevented (de Perdigo et al., 1989; Ashley, 1990). In this state, recombination is suppressed (Jaarola et al., 1998). Recombination in the inverted segment has apparently been suppressed when an inversion in balanced form is transmitted except in the rare theoretical situation of a double crossover within the segment. With specific reference to some X inversions, it may be that they have a lesser propensity to engage in recombination within the inverted segment (Shashi et al., 1996).
|
|
|
Figure 8-5. Pericentric inversion 7 in father (left) of an abnormal child with a recombinant 7 (right). The recombinant chromosome has a duplication of just over half of 7p, and a minuscule deletion involving the distalmost sub-band of 7q. The child has a triple amount of the segment p14.2 → pter. The karyotypes are 46,inv(7)(p14.2q36.3) and 46,rec(7)dup(7p)inv(7)(p14.2q36.3)pat. (Case of S. M. White.) |
Sperm Studies
Sperm studies in a handful of inversion heterozygotes give an indication of the frequency with which recombination happens, at least in male gametogenesis (Guttenbach et al., 1997; Anton et al., 2002). Dual-color FISH methodology, with one color (say, green) for the p arm and another (say, red) for the q arm of the inversion chromosome can show nicely whether a sperm is recombinant or not. Sperm with non-recombinant chromosomes would show one red spot and one green spot. A recombinant chromosome, with, for example, two red spots, would reveal the dup(q)/del(p) state while vice versa the dup(p)/del(q) chromosome would have two green spots. The advantage of this approach is that large numbers (thousands) of sperm can be analyzed.
We can separate the studied cases into those with a long inversion segment and those in which it is short. In four examples with longer inversion segments, inv(3)(p25q21), inv(6) (p23q25), inv(7)(p13q36), and inv(8)(p23q22), the proportions of dup(p)/del(q) and dup(q)/ del(p) recombinant chromosomes were 31%, 38%, 24%, and 13% respectively. The fractions of each recombinant type were precisely determined using the two-color FISH approach. In the inv(8), for example, about equal numbers of sperm showed the del(p)/dup(q) state (which is viable), and the dup(p)/del(q) state (which is not), 6.9% and 6.3% respectively (Martin, 1993). The fractions of the two types were also essentially 1:1 in the inv(6) case, with 19% of each (Anton et al., 2002). Recombination was however rare in an inv(1)(p31q12): only 23 recombinants seen in 5966 sperm, a fraction of 0.4% (Jaarola et al., 1998). This reflects a near-complete suppression of recombination (see above). No recombinants at all were seen in inversions with a short (or a very short) inversion segment, including an inv(3)(p11q11), an inv(9)(p11q13), and an inv(20)(p13q11.2), the former two being “normal variant” pericentric chromosome inversions (Anton et al., 2002). Thus, from first principles and from observing outcomes in these few gametes, we suppose that, other thing being equal, the longer the inverted segment, the more likely it is that recombination will happen. Other things may not be equal, however, as evidenced by the inv(1) noted above with a long inversion segment and almost no recombination. Theoretically, a “correcting” second crossover might take place in a long inversion segment, but this is not a concept we can usefully apply to risk estimation.
|
|
|
Figure 8-6. Alternative models for meiotic pairing, in which only a partial synapsis is achieved. Synapsis of (a) both distal segments, and (b) the inverted segment. One crossover is shown in each. |
Segment Content and Viability
While a long inversion segment can set the stage for recombination, what determines the viability of the recombinant conceptus is the functional content of the noninverted (distal) segments. We speak of a “genetically small” content if the combined effect of a duplication and deletion does not cause lethality during the earlier part of pregnancy but allows development to proceed well through the pregnancy and possibly to live birth. Thus, only those heterozygotes who have inversions with genetically small distal segments will ever have a chromosomally unbalanced phenotypically abnormal liveborn child. The inversion shown in Figures 8-2a and 8-5 illustrates this case. Inversion heterozygotes in whom one or both distal segments are genetically large (e.g., Fig. 8-2b) cannot have an abnormal recombinant child, although they may well have an increased risk for miscarriage. Any recombinants produced by such a person would impart a degree of imbalance that would be lethal in utero.
Genetic content corresponds fairly well to chromosome length. In inversion families in which recombinant children have been born, the distal (noninverted) segments together comprise, on average, only 35% of the total chromosome length; whereas in families having no known recombinant offspring, the figure is 62% (Kaiser, 1988). Nevertheless, if the distal segments comprise genetically small material, a larger fraction would not necessarily exclude a reproductive risk. Consider the inv(13)(p11q14) and inv(13)(p12q13), in which the distal segments comprise as much as 75% of the chromosome length (Kaiser, 1988). Although the imbalance in the recombinant is large in terms of haploid autosomal length, the result in the dup(q) form is in effect a partial trisomy 13 (the partial monosomy for 13p being without phenotypic influence). This, of course, is well known to allow intrauterine and postnatal survival. Similarly, an inversion in chromosome 18 can have distal segments that may be long relative to a short inversion segment, but they are still small genetically, and the dupdel combination can be viable (Schmutz and Pinno, 1986; Ayukawa et al., 1994). With specific reference to chromosome 4, Stipoljev et al. (2002) reviewed 20 reported familial cases and showed that recombinant forms have never been seen in those with smaller inversions, but are frequently seen with large inversions.
As noted above, it is typically the case that only one recombinant form is ever viable. This is rather impressively illustrated by Allderdice et al. (1975) in a kindred with the inv(3) (p25q21). Numerous cases of known or suspected dup(3q) children had been born, but none with the countertype del(3q). There is not even an increase in the miscarriage rate, suggesting that the del(3q) is lethal very early in pregnancy and causes occult abortion. Viability with both recombinant forms from the same inversion, the dup/del and the reciprocal del/dup, is infrequently seen. Kaiser (1984) records this only in the cases of inv(5)(p13q35), inv(13)(p11q22), and inv(18)(p11q21), and Hirsch and Baldinger (1993) add an inv(4) (p15.32q35). These four instances have this quality in common: the noninverted segments are not only short but remarkably small in terms of genetic content.
|
|
|
Figure 8-7. An inversion inv(4)(p15.32q35) with small noninverted segments, in which each of the two recombinant possibilities is viable. The del(4p)/dup(4q) karyotype (left recombinant offspring) produces a Wolf-Hirschhorn–like picture, and in the dup(4p)/del(4q) case (right recombinant offspring) the phenotype resembles the partial 4p trisomy syndrome. The normal chromosome 4 contributed by the other parent is shown grayed out. The 4q segment is so small (indicated by the dots) that it appears to make no major contribution, whether duplicated or deleted, to the phenotypes. (From the case of Hirsch and Baldinger, 1993.) |
It is instructive to consider the inv(4) (p15.32q35) in Hirsch and Baldinger's (1993) study, in which recombinant offspring could be del(4p)/dup(4q) or dup(4p)/del(4q) (Fig. 8-7). The four separate segmental imbalances are all well known individually to be viable. Distal 4p is, of course, the basis of the Wolf-Hirschhorn syndrome; and distal 4p trisomy has syndromic, if not eponymic, status. The distal 4q segment is small cytogenetically (0.25% HAL) and functionally, and duplication1 and deletion are quite well tolerated. So the respective imbalances in the combined states, the del(4p)dup(4q) and the dup(4p)del(4q), remain sufficiently small to be viable, at least much of the time. The index case, with the former imbalance, is a severely retarded child with a Wolf-Hirschhorn phenotype; and an aunt, having the latter combination, had rather minor dysmorphism and mental retardation. The inverted segment is very long: 87% of the total length of no. 4. Therefore, a crossover within the inverted segment is, we assume, very likely to take place. Surely, the genetic risks to heterozygotes for this inv(4) would be high. An even higher risk might apply to the inv(13)(p11q22) described by Williamson et al. (1980), in a family with several documented, suspected, or possible recombinant abnormal offspring. Here the contribution of 13p imbalance to the two recombinant states—the del(13p)dup(13q), and the dup(13p)+del(13q)—has no phenotypic effect, and the effective “single-segment” imbalances of dup(13)(q22 → qter) and del(13) (q22 → qter) are each well known to be viable. Applying the principles of private segregation analysis set out in Chapter 3, the risk for a recombinant form in this family comes to a high 50%. We emphasize again the point that, while the length of the inverted segment may influence the likelihood of recombination happening, it is actually the combined genetic content of the distal segments that is the direct determinant of viability of the recombinant form.
During the period 1981–1995, over 50 papers were published that reported the birth (or prenatal diagnosis) of offspring having a recombinant chromosome that derived from a parental pericentric inversion. In their review of this body of literature, and adding a family of their own, Ishii et al. (1997) determined the involvement of specific chromosomal segments. Figure 8-8, which is taken from their paper, depicts the combinations of dup-del genotypes that have been associated with viability. A few of these, which are shown asterisked, were identified at prenatal diagnosis, and in those with no known postnatal case, viability through to term remains unproven. A glance at the figure is enough to see that the gaps— that is, the noninverted segments—are generally longer, and usually a lot longer, than the sum of the lengths of the two inverted segments. This serves to illustrate again the point that inversions with large non-inverted segments are, as a rule, the ones with the greatest genetic risk. It is also to be observed that the thick bars (representing duplications) are mostly longer than the thin bars (deletions), a reflection of the preferential viability of the least monosomic combination. The individual autosomal inversions from this review are recorded in Table 8-1.
Inversions with very small distal segments may stretch the limits of cytogenetic detection (as also noted in the introductory section). Biesecker et al. (1995)describe an inv(22) with the long arm breakpoint in subtelomeric 22q, with the terminal 23–30 centimorgans of 22q now attached to 22p, which required molecular analysis with microsatellite markers and then FISH with a distal 22q cosmid probe for its identification. Because of the relative lack of G-band landmarks in 22q and the normal variation that occurs with 22p, the defect was not recognized on a 450-band cytogenetic study. The mother carrying this inversion would presumably have had a risk approaching 50% to have a further abnormal recombinant child.
Effect on Fertility
Uncommonly, the inversion heterozygote can be infertile (Groupe de Cytogénéticiens Français, 1986b; De Braekeleer and Dao, 1991). Abnormal synapsis of the chromosome pair can affect cellular mechanics at meiosis in the male, more likely so if the inversion involves a larger chromosome, consequently arresting spermatogenesis (Gabriel-Robez and Rumpler, 1994). For example, Meschede et al. (1994), describe azoospermic brothers, one with histologically documented arrest at the level of the primary spermatocyte, and each heterozygous for an inv(1)(p34q23) inherited from their mother.
Parental Mosaicism
Mosaicism for a (balanced) inversions is rare indeed. Lazzaro et al. (2001) describe a mother with 46,XX,inv(21)(p12q21.1)[19]/46,XX[11] on blood karyotyping, who had a child with a partial form of Down syndrome. The childs karyotype was nonmosaic 46,XX,rec(21) dup(21q)inv(21)(p12q21.1)mat. Given the fact that the mother's karyotype was from a peripheral blood sample and she had a recombinant child, clearly this is a case of somatic-go-nadal mosaicism.
Pericentric Inversions Frequently Innocuous
Many pericentric inversions are not associated with any discernible reproductive problems. The families of Voiculescu et al. (1986) and Rivas et al. (1987) are in this respect not atypical: an inversion chromosome was transmitted through several generations, numerous carriers having been identified, and there was no difference between the offspring of carriers and those of noncarriers in the incidences of abortion and neonatal death.
Interchromosomal Effect Is Unlikely
Some pericentric inversions have been discovered in the setting of a child with an aneuploidy such as trisomy 21, and “interchromosomal effect” has been invoked (Groupe de Cytogénéticiens Français, 1985b).
More likely, these associations are fortuitous; sperm studies endorse this inference (Martin, 1993; Anton et al., 2002). It is intriguing to note that one case of cystic fibrosis due to maternal uniparental isodisomy 7 occurred in the setting of a maternal pericentric inversion variant for chromosome 7. This link seems more likely coincidental than causal (Voss et al., 1989). In one instance, intrachromosomal effect in an inv(21) is considered to have caused trisomic and monosomic 21 conceptions (Gabriel-Robez and Rumpler, 1994).
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Figure 8-8. Viable recombinants from 55 recorded parental pericentric inversion chromosomes. The pairs of bars alongside each chromosome ideogram, one thick and one thin, show the inverted segments. The thick bars indicate the duplicated segment and the thin bars, the deleted segment in the recombinant offspring. The detail of the actual breakpoints is set out in Table 8-1. The inversions are grouped according to those chromosomes in which a dup(q)+del(p) is consistently seen in the recombinant offspring (above left), those in which a dup(p) del(q) is consistently seen (above right), and those in which either pattern may be observed (below). Most of these recombinants have been reported in only one or a few cases, with the notable exception of the inv(8)(p23q22), observed on 54 occasions. Asterisks indicate that a case was diagnosed prenatally; the inv(4)(p13q28) and the inv(5) (p13q33) have been seen only at prenatal diagnosis, so viability to term is not proven in these cases. (From Ishii, F., Fujita, H., Nagai, A., Ogihara, T., Kim, H.-S., Okamoto, R. and Mino, M., Case report of rec(7)dup(7q)inv(7)(p22q22) and a review of the recombinants resulting from parental pericentric inversions on any chromosomes, Am. J. Med. Genet. 73, 290–295, © 1997 Am. J. Med. Genet., courtesy F. Ishii, and with the permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc.) |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Table 8.1. Autosomal Pericentric Inversions Associated with the Birth of a Recombinant Offspring, Listed in Numerical Order |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Rare Complexities
Rarely, an inversion may create a setting that allows for the formation of a different type of rearrangement. Navarro et al. (1993) identified 2 sperm out of 140 examined from an inv(7)(p13q36) carrier to have an “inversion within inversion.” A submicroscopic alteration could occur in parent–child transmission, with a locus or small number of loci being disrupted or removed; unequal crossover within the inversion loop is a possible mechanism. An example is a familial inv(15)(p11q13) which, when transmitted from mother to child, underwent loss of the region that contains the putative Angelman syndrome (AS) locus (Webb et al., 1992). The loss was not detectable cyto-genetically—the child appeared to have the same inversion that his mother and grandfather carried—but was revealed on molecular analysis. The child had AS. Kähkönen et al. (1990) likewise describe a child with Prader-Willi syndrome and a 15q11 deletion whose father and grandmother were 46,inv(15)(p11q12). More speculatively, Urioste et al. (1994a), suggest this type of mechanism in a child with a short rib/polydactyly syndrome, the mother and child having the same apparently balanced inv(4)(p16q13.2). It remains entirely possible that associations like this are fortuitous (p. 54).
Collectors of remarkable cases will find fascinating the report of Allderdice et al. (1991). They studied a kindred (mentioned also above) with a segregating inv(3)(p25q21) that originated from a couple marrying in 1817 and that was quite widely spread over the maritime provinces of Canada and other parts of Eastern Canada and the Northeastern United States. In the course of the study, a normal man was found to have two recombinant 3 chromosomes: one with a dup(q)+del(p) and the other with a complementary dup(p)+del(q), such that his karyotype was balanced. Probably, both of his parents were inv(3)(p25q21) heterozygotes, and one produced one recombinant gamete and the other, the other. Theoretically, he could have produced a sperm with a “correcting” recombination that would (as the reader may care to check) make a balanced inversion or a normal chromosome; in fact, he was infertile.
PERICENTRIC INVERSION X
Pericentric inversions of the X are rare indeed, and in 1997 Madariaga and Rivera were able to review the observations in less than 30 published families. The X inversion forms in the same way as an autosomal inversion, but the implications are different. This is because (1) breakpoints in certain parts of the X (its critical region) may have an influence on the phenotype of the female; (2) X chromosomal imbalance in the 46,X,rec(X) female may be mitigated by selective inactivation of the abnormal X; and (3) the 46,Y,rec(X) conceptus will have a partial X nullisomy and functional X disomy. The inv(X) can be transmitted by both males and females. For example, Baumann et al. (1984) and Schorderet et al. (1991) describe families with an inv(X) transmitted through four generations, with all carriers—female heterozygotes and male “hemizygotes”— being phenotypically normal. The female and male inv(X) carrier need to be looked at separately.
Female inv(X) Heterozygote
Outwardly, the female heterozygote appears normal. The concept of position effect is of practical importance in the context of X rearrangement. If the long arm breakpoint lies within the segment Xq13 → q22 or Xq22 → q26, gonadal dysfunction may occur (Therman et al., 1990). There may be primary amenorrhea; or, after a fertile period in early adulthood, the menopause comes prematurely. Meiosis would be expected to proceed according to one of the preceding scenarios (see Figs. 8-4 and 8-6), with recombination within the inverted segment being a possibility. While there is not much practical experience to go on, we presume that an ovum with a normal X or the intact inv(X) would produce a normal child, whether male or female. In the case of the male, this would require there to have been no compromise of loci at the breakpoints, and evidence of normality in the male in another family member would be reassuring. A hemizygous son would typically be of normal fertility (Madariaga and Rivera, 1997). If in the family the balanced inversion is associated with normal gonadal function in the female, a daughter would be expected to likewise have normal puberty, fertility, and menopause at the usual time. This family information may not be accessible (or may not exist). In the family of Soler et al. (1981), for example, a hemizygous father had three sons and three daughters, each daughter, of course, being an obligate heterozygote. He apparently had no gonadal deficiency, but his two older daughters had menopause at 37 and 34 (the youngest was only 30). There was no family history recorded antecedent to him.
An ovum carrying a recombinant X would have two very different results, depending on whether it is fertilized by an X- or a Y-bearing sperm, as follows.
The 46,X,rec(X) conceptus. In their review, Madariaga and Rivera (1997) record outcomes in recombinant cases in 10 families. In female offspring, the del(Xq)/dup(Xp) combination is characterized by normal or tall stature and ovarian dysgenesis. The countertype, del(Xp)/ dup(Xq), is associated with short stature and, in some, intact ovarian function. These phenotypes presumably reflect the loss of stature genes (such as SHOX; Rao et al., 1997a,b) located on Xp and ovarian genes located on Xq, respectively. Any effect of the concomitant duplication is presumably mitigated by selective inactivation of the recombinant X chromosome. There is no obvious effect upon intellect.
Consider the case presented by Buckton et al. (1981) (Fig. 8-9). One of the breakpoints is at the tip of the short arm, and the other is in proximal Xq. The recombinant chromosome, with a deficiency of the tip of Xp and a duplication of distal Xq (Fig. 8-9, lower right) was associated only with shortness of stature in this family. The partial Xq trisomy made no discernible contribution to the phenotype. A 26-year-old mother with the rec(X) herself had a rec(X) daughter—unarguable evidence that oogenesis had not (at least by age 26) been compromised.
The 46,Y,rec(X) conceptus. In this instance there will be a nullisomy for the deficient X segment. If this segment constitutes any but the tiniest length of chromatin, the conceptus will not be viable. Nullisomy for a tiny telomeric segment may be viable, but with major dysmorphogenesis and severe neurodevelopmental compromise. Further, the concomitant disomy X is functional, not being subject to inactivation, and therefore of itself produces a major deleterious effect (Groupe de Cytogénéticiens Français, 1986b).
|
|
|
Figure 8-9. X chromosome inversion. The mother (above) has the karyotype 46,X,inv(X)(p22q13). Shown below are the two possible unbalanced reproductive outcomes in daughters, following recombination within the inverted segment; the normal X on the left in each is contributed by the father. Each type of daughter would have a variant form of Turner syndrome. Male recombinant conceptuses are not shown: the combination of X nullisomy and functional X disomy in the 46,Y,rec(X) conceptus would in this instance be lethal in utero. (From the case in Buckton et al., 1981.) |
Male inv(X) Hemizygote
In the male carrier, the rearrangement apparently has no effect on phenotype or reproduction. Meiosis proceeds unperturbed (rather obviously, there can be no recombination within the inverted segment). All his daughters will be heterozygotes. Many will have normal gonadal function, although a family history of premature ovarian failure might predict the same problem (see also above). Sons receive his normal Y and their mother's X chromosome.
PERICENTRIC INVERSION Y
A pericentric inversion of the Y, inv(Y) (p11q13), is not uncommon in the general population (Verma et al., 1982; Tóth et al., 1984). It has no phenotypic effect and implies no risk for having an abnormal child. It may be regarded as a normal variant. Meiosis proceeds as it would in the 46,XY male.
GENETIC COUNSELING
AUTOSOMAL PERICENTRIC INVERSION
Variant Forms
The not uncommon inv(2)(p11q13) is practically always innocuous (MacDonald and Cox, 1985; Groupe de Cytogénéticiens Français 1986b; Daniel et al., 1989). Two possible exceptions are on record to belie its reputation: two abnormal children, one with a 2p duplication and the other a 2p deletion, the proximal boundary at or adjacent to 2p11.2, and the fathers being inversion heterozygotes, described as inv(2)(p11.2q12.2) and inv(2)(p11.2q13) respectively (Magee et al., 1998a; Lacbawan et al., 1999). It may be that the configurations adopted by the no. 2 homologs led to an unequal crossing-over, and hence the duplication or deletion. With only two such observations of recombination in the decades of history of clinical cytogenetics, some circumspection is required, and Lacbawan et al.'s comment, “at this point, it seems premature to recommend prenatal diagnosis of all couples in this situation,” is well taken.
No genetic risks are known to be associated with the other inversion variants noted in the Biology section—“inversions” of 1, 9, 16, and Y heterochromatin, inv(3)(p11~13q11~12), inv(5)(p13q13), and inv(10)(p11.2q21.2). Concerning the inv(10)(p11.2q21.1), Collinson et al. (1997) offer the practical advice that “family investigation of carrier status is not warranted in view of the unnecessary concern this may cause family members.” We exclude these inversion variants from the discussion below.
Risks of Having an Abnormal Child
Ascertainment via Recombinant Child
Identification of a family through a recombinant individual proves the viability of at least one of the two recombinant chromosomes. Table 8-1 lists a large number of different inversions for which a carrier is known to have had a recombinant child. There have been various empiric estimates of the overall level of risk to the heterozygote in families ascertained through an abnormal child. From a number of studies, a consensus range for the usual risk of having a liveborn abnormal child due to recombination is 5%–10% (Groupe de Cytogénéticiens Français, 1986b; Sherman et al., 1986; Stene, 1986; Daniel et al., 1988). As a general rule, the longer the inversion seg-ment—and, consequently, the shorter the distal segments—the greater the risk of producing a viable recombinant gamete. Very long inversions, such as that in Roberts et al. (1989), an inv(10) that comprised 80% of the whole chromosome, would imply the highest risks: in this particular case, two out of the carrier fa-ther's three children were recombinant. For the majority of families, there is probably no risk difference depending on sex of heterozygote (Kaiser, 1984; Stene, 1986); but in some families, the female heterozygote may run a greater genetic risk (Sutherland et al., 1976; Pai et al., 1987). Indeed, for the inv(21)(p12q21.1), recombinant children (with dup(21q) and thus a partial form of Down syndrome) have been seen only where it is the mother who is the carrier parent (Lazzaro et al., 2001).
Daniel et al. (1989) pooled American, Canadian, and European data to derive risks for an unbalanced karyotype at amniocentesis and derived (from rather small data) a figure of 10%–15% for inversions with small distal segments in a family in which a previous affected child had been born.
Each individual inversion carries its own individual risk. This figure may be arrived at by analyzing the patient's family, studying the literature, and assessing the degrees of imbalance potentially arising in the recombinant conceptuses. A specific figure has been derived for one relatively common inversion, the inv(8)(p23q22): the risk for liveborn recombinant offspring, all of whom would have the del(p)/dup(q) form, is 6.2% for both maternal and paternal transmission (Smith et al., 1987). This compares closely with the figure of 6.9% of sperm with the del(p)/dup(q) form, attesting to an essentially uncompromised viability. In contrast, the countertype dup(p)/del(q) recombinant, which is seen in 6.3% of sperm, is never seen in liveborn offspring, reflecting zero viability. With inversions of 18, the breakpoints at p11 and at q11, q12, or q21, a group risk of 8% applies (Ayukawa et al., 1994). In due course, figures may be determined for other inversions seen in more than one family, such as the inv(3)(p25q21), inv(4)(p14q35), inv(10) (p11q25), inv(13)(p13q21), and inv(21) (p12q21.1).
No Family History of Recombinant Form
For families identified by means other than through the birth of an abnormal child (e.g., discovered fortuitously at prenatal diagnosis), the overall risk is around 1%. The individual risk, which is what really matters, depends on the actual inversion. Is the inversion chromosome known to be associated with viable imbalance (Table 8-1)? Or, does the inversion segment include and extend beyond the inversion segment of one of these recorded cases? In that circumstance, a significant risk surely does apply (see above). Is the inversion segment much shorter in length than any of those listed in Table 8-1? Here the risk may be as low as zero. The level of risk can be assessed from a study of the family, noting the reproductive histories of other heterozygotes, and from a consideration of the degrees of potential imbalance in a conceptus. As a rule, any chromosome with a short inversion segment (less than one-third of the chromosome's length) is most unlikely ever to lead to a viable recombinant product (Kaiser, 1988).
Nevertheless, one should determine the composition of the theoretically possible recombinant gametes and gauge whether the resulting partial trisomy and partial monosomy might be viable. This applies in particular to inversions of chromosomes 13, 18, and 21, partial trisomies and partial monosomies of these chromosomes being well recognized as viable. If in any inversion chromosome one breakpoint is very close to the telomere, one recombinant form will impose very little partial monosomy. The contribution of the duplication can then be assessed on its own, and reference to the viability of this segment in other cytogenetic contexts (translocation, de novo rearrangement) will likely provide a valid comparison. For example, had the father in Figure 8-5 been identified before he had children, we could have deduced that the rec(7) dup(7p) genotype might survive to term, knowing that the databases of Stene and Stengel-Rutkowski (1988) and Schinzel (2001) record a viable phenotype for trisomy 7p14 → pter.
Prenatal diagnosis2 should be offered to the following individuals:
a. Any heterozygote in whose family a recombinant child has been born.
b. A heterozygote for any of the inversions listed in Table 8-1.
c. A heterozygote for an inversion involving a segment longer than, but including, a region listed in Table 8-1.
d. Any other heterozygote for whom the theoretical recombinant product(s) might be viable. Many inversions of chromosomes 13, 18, and 21 will fall into this category.
e. Molecular analysis to exclude deletion in the Prader-Willi/Angelman region of 15q11–13 may be appropriate in an inversion having a breakpoint within or adjacent to this segment.
Of the phenotypically normal offspring, approximately half will have normal chromosomes and half will be inversion heterozygotes (Groupe de Cytogénéticiens Français, 1986b).
Whether inversion heterozygotes have a risk for having children with other categories of chromosome abnormality is unclear. Do carriers of inversions of 13, 18, and 21 run an increased risk for trisomy of the respective chromosome? If they do, the rarity of reported cases indicates that the risk must be very small. A suggested interchromosomal effect for an inversion of any chromosome that would impart an increased risk for, in particular, trisomy 21 is unconfirmed (Groupe de Cytogénéticiens Français, 1986b; Kleczkowska et al., 1987). The possibility of deletion within the inversion loop or at its boundaries is exemplified by the Angelman case of Webb et al. (1992), but for most inversions at least, the risk for this is probably very small, bordering on negligible. Except in a very few special cases, and bearing in mind proximal 15q, it would be neither feasible nor reasonable to attempt detection of such deletions in routine practice.
INVERSION X
The female heterozygote could have a premature menopause if the long arm breakpoint is in the critical region and if there is a family history of early ovarian failure, and practical advice would be to have children sooner rather than later. But normal reproductive function is perfectly possible. Recombination may be less likely than for an autosomal inversion (Pinto Leite and Pinto, 2001), although a risk of producing an abnormal daughter with a recombinant X nevertheless exists. To some extent, the abnormality is predictable according to the deleted segment, Xp or Xq: short stature is typically seen in del(Xp), and ovarian failure in del(Xq). Hemizygous sons would be expected to be normal, and reassurance in this respect may be drawn from the observation, if it can be made, of normality in a male relative. For the most part, no risk exists for having an abnormal son because recombinant male conceptuses, having partial X nullisomy and disomy, would be nonviable. Only when the breakpoints are very close to the telomere is male viability possible, and such a child would have major physical abnormalities and mental retardation, probably severe.
All daughters of the male heterozygote would be inv(X) heterozygotes. Other things being equal, they will be phenotypically normal. If the long arm breakpoint is in a critical region and if heterozygous female relatives have had ovarian deficiency (e.g., primary amenorrhea, premature menopause), they may develop the same problem. All sons would have a 46,XY karyotype.
Inversion Y
All the sons of the inv(Y) carrier are themselves inv(Y) carriers. They are all normal and, other things being equal, have normal gonadal function. All the daughters would be 46,XX.
Paracentric Inversion
BIOLOGY
Meiosis
AUTOSOMAL INVERSIONS
Classical theory has it that the heterozygote for an autosomal paracentric inversion cannot produce a viable unbalanced progeny. If a recombinant gamete is formed following a crossover in the inverted segment, the chromosome would be either acentric (lacking a centromere) or dicentric (Fig. 8-10). An acentric chromosome is never viable, since it lacks a point of attachment to the spindle fibers. The dicentric chromosome is generally considered a lethal impediment, as it is attached to spindle fibers pulling in opposite directions, with the chromosome thus suspended between the daughter nuclei at telophase and excluded from either cell. If the dicentric were to rupture, however, the possibility theoretically exists for a product (this might be, effectively, a dupdel chromosome) to enter the zygote and be viable. Alternatively, if the dicentric were to be included in the nucleus of a gamete, Mc-Clintock's classical breakage–fusion–bridge cycle might impose an eventually insuperable obstacle to continuing cell division as the chromosome is tugged in two directions by its two centromeres in succeeding mitoses after formation of the zygote. The possible scenarios are more fully dealt with in Madan (1995).
What are the findings on direct observation of gametes? Two sperm studies, one of a man with 46,XY,inv(7)(q11q22) and the other of 46,XY,inv(2)(q14.2q24.3), gave 0% and 0.8% levels of recombination, respectively (Martin, 1986; Devine et al., 2000). In the former, no recombinant forms were seen in 94 cells analyzed, and in the latter, in only 4 sperm (2 dicentric, 2 acentric) out of 496. Sperm typing by amplification of microsatellite markers in the region of the inversion is a powerful instrument with which to study meiotic recombination. Brown et al. (1998a) analyzed 282 sperm from a man with a paracentric inversion, 46,inv(9)(q32q34.3), whose wife had had a number of miscarriages (they had also had two children). Recombination was suppressed in the inversion segment, but notably, of the five recombination events within the segment that were observed, each involved at least two crossovers. Brown et al.suggested the following mechanism. Synapsis, which starts at the telomere, advances along the chromatids, and then encounters a region of heterosynapsis and “stalls”. This stalling allows an increase in recombination in the chromatid regions that are already synapsed. Synapsis eventually advances past the inversion segment, and continues towards the centromere. But within the inversion segment itself, an active search for homology goes on, which may require the chromatids taking on a particular configuration (such as a microloop), and this may set up a hot spot for recombination. Only rare double recombinants from this setting would be able to form morphologically normal chromosomes, with sperm that would then be able to continue along their process of maturation; sperm with a single recombination would be acentric or dicentric.
|
|
|
Figure 8-10. Theoretical recombinant products from classical crossover in paracentric inversion. One is acentric (ace) and the other, dicentric (dic). The inversion segment is shown cross-hatched, and the different directions of cross-hatching indicate the parts proximal and distal to the crossover point. |
Meiosis in oogenesis commences during fetal life, and its study therefore requires access to fetal tissue. Cheng et al. (1999) analyzed ovarian tissue from a 19-week termination of pregnancy in which a de novo inv(7) (q11.23;q21.2) had been shown at amniocentesis. By using a FISH probe for the Williams syndrome critical region (WSCR), which is at 7q11.23, they were able to determine whether the inverted segments were aligned alongside each other (homosynapsis) or not (heterosynapsis). Most cells showed the no. 7 homologs lined up side by side, but with the WSCR signals off from each other, thus the inversion segment was unaligned. A classical inversion loop was seen in only 10% of cells. This example, concerning a small inversion segment, offers an explanation for the rarity with which recombinant forms are seen: the necessary prerequisite of homosynapsis may not often be attained. As noted in the introductory comments on the pericentric inversion (p. 146), the particular breakpoints may also be important in determining this blocking of homosynapsis (and note that the two inv(7)s discussed above involve very similar segments).
Recombination/Reunion with Viable Products
Classical theory remains valid in essence, some exceptions notwithstanding. The abnormal process of “U-loop recombination” (Feldman et al, 1993; Mitchell et al., 1994) is a mutational event, not a predictable consequence of a “normal” meiotic process (albeit in a chromosome that is abnormal). “Reunitant” may be a better word than recombinant. The crossover within the inversion loop, instead of continuing on in the same direction along the chromatid, reverses upon itself as a U-loop. The mechanism is illustrated in Figure 8-11. According to this construction, the resulting reuniting chromosomes would have either a duplication of that part of the inversion loop proximal to the crossover, and a deletion of that part distal to it, or the reverse. A crossover (or, rather, chromatid breakage with abnormal reunion) at one of the entry points to the loop would produce a duplication alone, or a deletion alone. Feldman et al. (1993) review the inversion duplication (inv dup) chromosome, and notably, of the six familial cases on record, five may have been due to presumed U-type reunion from a maternal paracentric inversion. Chia et al. (1992) describe a case which quite probably reflected the same mechanism, a man with 46,inv(18)(q12.1q23) who had a child with a duplication/deletion 18q syndrome due to a presumed rec(18)(pter → q21.3::q21.3 → q12.1::q23 → qter) chromosome, as shown in Figure 8-11. Another very similar inv(18) case is noted in Hani et al. (1995). In their exhaustive review, Pettenati et al. (1995) collected about a dozen similar cases. These cases represented offspring in 3.8% of their series of 446 paracentric inversions. But since all of these offspring were probands, and some we actually doubt were truly paracentric reunitants, we presume the actual reproductive risk due to U-loop reunion would be a much smaller figure (Sutherland et al., 1995). Madan and Nieuwint (2002) have pursued this question and showed that indeed most “paracentric inversions” found through a recombinant child were really insertions.
Classical theory also needs to accommodate the phenomenon of centromere suppression, which, extremely rarely, can allow the basically dicentric recombinant to function stably as, in effect, a monocentric. The chromosome attaches to the spindle fiber of only one daughter nucleus. At this writing, “extremely rarely” could be defined as four recorded cases from an autosomal paracentric inversion. Mules and Stamberg (1984) described an infant dying as a neonate with a rec(14) whose mother had an inv(14)(q24.2q32.3); Worsham et al. (1989) studied in considerable detail a child with a rec(9) from a maternal inv(9)(q22.1q34.3); Whiteford et al. (2000) reported on a dysmorphic infant with growth and neurodevelopmental retardation and a major heart defect, with the karyotype 46,XY,rec(15)(pter-q26.3::q11.2–pter)inv(15)(q11.2q26.3)mat; and Lefort et al. (2002) described an abnormal child in whose dicentric rec(14) chromosome one centromere could be demonstrated to have been inactivated, the mother's karyotype being 46,XX,inv(14)(q13q32.2). These four cases share the features of a large inversion involving most or almost all of a long arm, with the short arms (14p, 9p, 15p, and 13p, respectively) being genetically small. In other words, the dup pq/del q combination might not impose a lethal imbalance. Only in this setting, and if the dicentric chromosome were stable, could recombination cause an imbalance that would be viable and allow the birth of an abnormal child.
A mechanism reminiscent of paracentric inversion U-loop reunion may be the cause of some isochromosome Xq Turner syndrome (Wolff et al., 1996). Two zinc-finger genes (ZXDA and ZXDB) in proximal Xp, just above the centromere, have about 98% homology and transcribe in opposite directions. In X-to-X synapsis in some meioses, a small inversion loop in proximal Xp might enable ZXDA (the more centromeric locus) in one Xp to match up with ZXDB on the other Xp, and vice versa. Then, a breakage and U-loop reunion between the two ZXD loci would generate an isodicentric chromosome Xqter → cen → ZXDA::ZXDA → cen → Xqter. Similar events at other loci may underlie other Xq isochromosomes or supposed Xq isochromosomes (Giglio et al., 2000).
|
|
|
Figure 8-11. Above, parent with paracentric inversion and child with recombinant (“reunitant”) chromosome. Father has paracentric inversion of 18q, inv(18)(q12.1q23). The inverted segment is shown cross-hatched (cross-hatching changes slope at q21.3). Child has duplication of the segment q12.1 →21.3 on the reuniting chromosome (shown cross-hatched) and deletion q21.3 → q23 (Courtesy N. L. Chia and L. R. Bousfield). Below, proposed mechanism of U-loop exchange depicted; asterisk indicates point of U-loop. The position of the point of exchange within the inversion loop (in this case, q21.3) determines the nature of the imbalance. There is duplication of chromatin proximal to the crossover point (q12.1 → q21.3) and deletion of distal chromatin (q21.3 → q23), as in the child's rec(18)′, and the reverse in the complementary product, rec(18). (An alternative interpretation is that the father's rearrangement is a within-arm insertion of 18q, rather than an inversion, in which case the karyotype of the child would have been derived from recombination in the inserted segment.) |
A minuscule number of cases of other sorts of viable recombinant offspring are known (Worsham et al., 1989). A dicentric recombinant chromosome, pulled in two directions, may rupture and yield a deletion. This may be the mechanism in the case of Courtens et al. (1998), in which a mother with 46,XX,inv(18) (q21.1q22.3) had monozygotic twins with a deletion of the segment distal to the inversion (q22.3–qter) and duplication of a small part proximal to it (q12.1–q21.1). Although the scale is not right for this example, in Figure 8-10 it could be imagined that a break occurred in the dicentric recombinant chromosome just above the lower centromere. An alternative mechanism is that the abnormal synapsis may set up a milieu that encourages some other type of rearrangement to form, such as excision of an inversion loop, and unequal crossing-over at the base of an inversion loop, a format initially proposed by Hoo et al. (1982). Yang et al. (1997) propose such a scenario in a family in which the index child had the deletion 46,XY,del(17)(p11.2p11.2), while the father and two aunts carried the paracentric inversion 46,inv(17)(p11.2p13). The deletion removed the Smith-Magenis region. This was “the first unequivocal demonstration by molecular analysis that a parent who carries a paracentric inversion is capable of having a viable child with an unbalanced monocentric recombinant chromosome.” Phelan et al. (1993) report the unique case of a father with an inv(9)(p13p24) having a child with a rec(9) containing a tandem duplication, which they propose came from breakage and reunion between sister chromatids within the inversion loop. An inversion with a breakpoint in the vicinity of 15q12 may lead to a rearrangement that would cause Prader-Will syndrome or Angelman syndrome, as mentioned above.
Some inversion carriers have been ascertained through their having had many miscarriages (Madan, 1995). In most of these, surely, the discovery was fortuitous. One family with an inv(10) was widely studied, and 19 carriers in three generations had only 1 miscarriage out of 36 pregnancies (Venter et al., 1984). The report by Devine et al. (2000) mentioned above of two brothers with 46,XY,inv(2)(q14.2q24.3) presenting with reproductive pathology may suggest a link, but other causes are quite possible. One brother's wife had three miscarriages, and at IVF in the other brother, five of ten fertilized eggs failed to cleave, and progression in the remaining five failed at the blastocyst stage. No karyotyping was done of any of these several products of conception. In a very few cases, theoretical dicentric recombinant products might convey a genetic imbalance that could allow at least some weeks of in utero growth before miscarrying (Bocian et al., 1990; Bell et al., 1991). The nine miscarriages suffered by the carrier grandmother in the family in Worsham et al. (1989) might have been due (at least some of them) to recombinant gametes, the dicentric state having been proven in her index grandchild.
Other Mechanisms
Mendelian loci can be vulnerable when chromosomal rearrangement happens, because of position effect, epigenetic influence, or direct disruption. We have seen, for example, a family in which an inv(7)(p22.2p21.2) was associated with Saethre-Chotzen syndrome, this being a Mendelian disorder due to the TWISTgene. Two heterozygous children showed major craniosynostosis, but the father and grandfather had only the subtlest facial, auricular, and digital signs. It may be that Saethre-Chotzen syndrome due to position effect has a milder phenotype than in the case of point mutation (Rose et al., 1997). The intriguing observation of Cornelia de Lange syndrome (CDLS) in a child with a maternally inherited balanced paracentric inv(3q), in which the 3q27 breakpoint is in the vicinity of a postulated CDLS locus, awaits clarification (Rizzu et al., 1995). As for a possible epigenetic mechanism, Norman et al. (1992) described a family in which a mother had one child with Beckwith syndrome and a presumably affected fetus, all three carrying an apparently balanced inv(11)(p11.2p15.5). The normal imprinting state of the Beckwith region on distal 11p may have been perturbed.
Iida et al. (2000) studied a child with heterotaxy (variably abnormal left–right organ placement) having an inv(11)(q13.5;q25)pat. The existence of a gene that was proven to be disrupted at one of the breakpoints (UVRAG at 11q13.5) naturally raised the suspicion that it could have had a causative role. The normal father may have produced a barely sufficient amount of UVRAG protein, while his child may have been only just below the threshold. This is an attractive speculation, but, as indeed the authors acknowledge, the association with the inversion may have been coincidental. As we have had occasion to remark a number of times elsewhere in this book, in this sort of case the child had to be abnormal (in whatever respect) to have been karyotyped.
X CHROMOSOME INVERSIONS
If a paracentric inv(X) is associated elsewhere in the family with normality, no defect would be anticipated in future heterozygotes or hemizygotes (Neu et al., 1988a). Breakpoints in the critical region in Xq, however, might compromise gonadal integrity. For example, Dar et al. (1988) report a woman with a de novo inv(X)(q13q24) who had ovarian dysgenesis with primary amenorrhea and no spontaneous pubertal development, and Németh et al. (2002) describe an infertile man with a Kline-felter-like phenotype having an X inversion with rather similar breakpoints, 46,Y,inv(X) (q12q25). A woman with a somatotype of Turner syndrome having an Xp inversion (p11.2p22.1) is described in Dahoun (1990). A breakpoint might damage a Mendelian locus, and Briault et al. (1999)report a family in which the FG syndrome (mental defect, facial dysmorphism, hypotonia, anal abnormality) cosegregated with a paracentric X inversion, inv(X)(q12q28). One FG locus has been mapped to Xq12–q21.31, and so it is plausible that the q12 breakpoint in the inversion may have been at the site of the FG gene.
Y CHROMOSOME INVERSIONS
Only two paracentric inv(Yq) cases are on record (Madan, 1995; Liou et al., 1997). In Liou et al.'s three-generation family, the normal grandfather and father were 46,X,inv(Y) (q11q21), and the child with the same karyotype had ambiguous external genitalia with müllerian structures internally and intraabdominal testes. The inversion Y may have been coincidental; alternatively, in the child there may have been a position effect whereby the expression of the testis-determining gene SRY had been compromised.
Paracentric Inversions Usually Innocuous
The above compendium notwithstanding, the observed facts attest to the general innocuousness of the autosomal paracentric inversion, concerning either the heterozygous state per se or a risk for chromosomally unbalanced offspring. Madan (1995) reviewed 184 cases of autosomal paracentric heterozygosity. Many were ascertained fortuitously and, including those discovered during the course of investigation for recurrent miscarriage, 58% were identified in a normal person. Several had an abnormal phenotype, but this was, of course, the reason they had the chromosome test done in the first place; by definition, they had to be abnormal. No clear consistent pattern among phenotypes of presenting cases is apparent. As Madan comments, there may have been a bias in choosing cases for publication, and editors of journals might not find compelling a paper describing an “uninteresting” inversion discovered incidentally in a normal individual (a series of a dozen or so cases might stand a better chance). In their review, Pettenati et al. (1995) were confident about a causal association with a specific phenotype only in the paracentric inv(X), and not with any of the autosomal inversions.
The Groupe de Cytogénéticiens Français (1986a) note that the reproductive fitness of heterozygotes in 32 French families was normal. Two quite common inversions seen in a number of families in more than one part of the world are the inv(3)(p13p25) and the inv(11)(q21q23) (Madan, 1995). No abnormalities directly attributable to these inversions have been documented. It may be founder effect or a recurring mutation that is the basis for their frequency. In the one sperm study on a paracentric inversion heterozygote, having the relatively common inv(7)(q11q22), Martin (1986) found no recombinants. The smallness of the inversion segment may have been a factor in preventing formation of a synaptic loop. (It is not without interest to note that a similar inversion is the norm in the gorilla no. 7, and so this human form could be thought of as a “back mutation” to that of the ancestral primate.)
Interchromosomal Effect is Unlikely
Watt et al. (1986) raise the possibility that the paracentric inversion might have an “interchromosomal effect.” They note an apparently high level of reported associations, within families, of an inversion plus some other chromosomal defect. We suspect that this is artifactual; as these authors note, ascertainment and publication biases are potential confounders in this setting. Pettenati et al. (1995) reached a similar conclusion.
Technical Comment
Paracentric inversions can be technically difficult to detect. Gross chromosome morphology is not altered, and unless major landmark bands are shifted, the rearrangement may go unnoticed. Only with the use of good-quality, high-resolution banding are paracentric inversions likely to be detected regularly. These cytogenetic difficulties may be the reason why relatively few cases of this type of inversion have been published. Also, for technical reasons, reported cases of recombination in the literature should be regarded with caution; as mentioned above and noted below, some “inversions” are likely actually to be intrachromosomal insertions (“paracentric shifts”). The cytogenetic distinction can be difficult to make, especially for chromosomal regions without distinctive banding patterns or where the inverted segments are very small (Callen et al., 1985; Madan, 1995). We have seen a family in which the index case seemed to have an unbalanced translocation at distal 4p, but the normal mother and grandfather had the same anomaly that could then be reinterpreted as the minimum inversion detectable on routine cytogenetics, a one-band paracentric inversion, in this case inv(4)(p15.3p16.3) (Smith et al., 1992).
GENETIC COUNSELING
On practical grounds, the reassuring point to note is that practically all paracentric inversion heterozygotes identified have been discovered fortuitously, and not through the birth of a child with an abnormality attributable to the parental inversion (Madan, 1995). We agree with Madan: “the vast majority of paracentric inversions are likely to be harmless.” Apparently, the genetic risks to offspring are extremely small. In the U.S. collaborative study described by Daniel et al. (1988), there were no unbalanced karyotypes in 30 prenatal diagnoses. The sex chromosomes warrant separate attention, and it may be that some X and Y paracentric inversions have an effect on gonadal development in the intact (that is, unrecombined) state.
A tiny handful of abnormal offspring (noted above) refute a complete harmlessness in the parental paracentric inversion—whether due to classic recombination with a dicentric chromosome, the generation by U-loop “reunion” of an inverted duplication (inv dup) chromosome, deletion from an unequal crossover at the base of the loop, or the rupture of a dicentric recombinant that produces a viable monocentric product. Whether this would warrant prenatal diagnosis when a parent is a carrier is a matter for debate. Even where the new chromosome from a classic recombinant or U-loop reunion might on theoretical grounds be viable, the risk for one to be generated, while its exact magnitude is unknown, is surely extremely small. Submicroscopic molecular damage is possible but almost unknown (Greger et al., 1997), although Madan (1995) does note that subtle duplications or deletions resulting from rearrangement within or at the extremities of the inversion segment might not be readily recognizable, and a pregnancy diagnosed with apparently the same parental inversion may need to be carry a slight (we think very slight) reservation.
We suggest that an offer of prenatal diagnosis be discretionary in the case of a fortuitously discovered inversion in the family; and would regard it as not inappropriate if the offer were declined. A firmer stance may be appropriate if there has been a previous history of an apparently associated reproductive abnormality. Inversions on record with a demonstrated recombinant would oblige the offer of prenatal diagnosis. These include those noted above: inv(9)(p13p24), inv(9)(q22.1q34.3), inv(14)(q24.2q32.3), inv(17)(p11.2p13), inv(18) (q12.1q23), and inv(18)(q21.1q22.3). In the particular case of inversions having breakpoints in distal 11p, proximal 15q, and at 17p13, a case could be made for prenatal diagnosis with attention to the possibilities of Beckwith-Wiedemann syndrome, Prader-Willi or Angel-man syndrome, and Smith-Magenis syndrome, respectively.
As mentioned above, a diagnosis of a para-centric inversion might be incorrect, and the rearrangement is actually a paracentric insertion, which carries a high genetic risk (p. 172). Since the distinction in the routine laboratory can be difficult, a practical view might be to risk overinterpreting subtle paracentric inversions as potential insertions in those cases where the cytogeneticist is not absolutely certain. The true picture may emerge by determining the order of a number of FISH probes across the relevant region.
Notes
1. Duplication for a considerably longer segment, 4q31.3 → qter, comprising 1.15% of HAL, is viable, as the children in the frontispiece photograph illustrate.
2. Escudero et al. (2001) see a place for preimplantation diagnosis, and Anton et al. (2002) report a successful PGD outcome from a man with inv(6)(p23q25).
