Genetics of human situs abnormalities

NOMENCLATURE AND CLINICAL CONSIDERATIONS

The normal left-right anatomic arrangement is called situs solitus. Mirror-image reversal of all asymmetric structures has been given a variety of labels, most commonly situs inversus, complete or total situs inversus, and situs inversus totalis. The resulting anatomy is more of a curiosity than a hazard. When the entire anatomic left–right axis is neither normal nor mirror-image reversed, the resulting phenotype has been called situs ambiguus, partial situs inversus, heterotaxy or heterotaxia (sometimes accompanied by the adjective visceral), laterality or isomerism sequence, and Ivemark, asplenia, or polyasplenia syndrome. All of these terms describe the same general phenotype: discordant left–right anatomy within and among the lateralized structures of the chest and abdomen. Complex, often fatal heart malformations are common (although not invariable), as are intestinal malrotation and abnormalities of spleen position and/or number.

Rough estimates place the incidence of left–right malformations at 1/5000 births, with cases divided equally between situs inversus and situs ambiguus [Ferencz et al., 1993; Afzelius and Mossberg, 1995]. This figure may underestimate the actual incidence of each. Complete left–right reversal (situs inversus) may escape detection because it poses no detriment to the individual, and cases of situs ambiguus with normal hearts or those with clinically silent cardiac malformations also may not come to medical attention. For example, in a retrospective review of 18 patients with intestinal malrotation, seven were found also to have polysplenia, and six of these also had either interrupted or double inferior vena cava [Zissin et al., 1999]. None of these patients had been determined previously to harbor left–right axis malformations.

Many familial cases of situs abnormalities have been described in which also one or more family members have only cardiac malformations that would not be described by most cardiologists as being the consequence of abnormal left–right development [Alonso et al., 1995; Casey, 1998]. However, given the family background and the fact that these isolated heart malformations also occur among individuals with obvious situs ambiguus, one might entertain the hypothesis that these and perhaps other cases of isolated heart malformations may be the result of abnormal laterality. Recently, Bouvagnet and colleagues [Mégarbané et al., 2000] have provided experimental evidence in support of this hypothesis (see below).

GENES AND ENVIRONMENT

Clearly, both genes and environment can contribute to the development of left–right axis malformations. Retinoic acid exposure, for example, can induce laterality defects in a variety of vertebrates, including Homo sapiens. Much more important from a public health standpoint, however, is the increased risk of left–right malformations in the offspring of mothers with (nongestational) diabetes mellitus [Splitt et al., 1999]. Whether or not left–right malformations develop in the offspring may depend on his or her genetic background. Morishima and colleagues have developed lines of nonobese diabetic (NOD) mice with a high incidence of situs abnormalities among offspring if the dam is hyperglycemic early in gestation [Morishima et al., 1991]. The incidence is highest (65%) if the sire is from the NOD strain, decreases to 24% when the sire is from strain ICR (from which NOD was derived), and decreases to background levels with C57Bl sires [Morishima et al., 1996]. These results suggest that the genetic background of the embryo lowers the threshold for malformations induced by environmental agents.

CYTOGENETICS

Recognition of familial situs abnormalities during the past several decades has led to the obvious conclusion that a strong genetic predisposition underlies these anomalies, at least in some cases. Cytogenetic abnormalities in association with situs malformations support this conclusion. Trisomies and monosomies (complete and partial), translocations (balanced and unbalanced), inversions, deletions, and even isodisomy have all been described [Kosaki and Casey, 1998]. Presumably these cytogenetic alterations provide clues to the location of genes involved in left–right axis development.

Unfortunately, no such gene has been discovered by exploiting a cytogenetic rearrangement. Iida and colleagues recently reported that UVRAG, a gene previously implicated in the pathogenesis of xeroderma pigmentosum, was disrupted in the 11q inversion found in a case of situs ambiguus [Iida et al., 2000]. This gene has not been implicated in left–right development of nonhuman vertebrates, and no mutation has been found in an unrelated case, so it remains unclear whether or not disruption of this gene is causally related to the accompanying phenotype.

A more convincing connection between left–right genes and cytogenetic abnormalities is found in the case of a de novo interstitial deletion in chromosome 10q22 [Carmi et al., 1992]. Situs ambiguus and midline malformations are the accompanying phenotype. The murine gene nodal and its counterpart in other vertebrates has been shown to be required for normal left–right axis development. Human NODAL maps to 10q22 and, based on polymorphic microsatelllite analysis, is deleted in the 10q22 deletion chromosome [B. Casey, unpublished results].

LINKAGE MAPPING

Cytogenetic abnormalities associated with left–right malformations are quite uncommon, and when identified have not been helpful to date in identifying new genes associated with left–right axis development. Linkage studies have for the most part been equally unhelpful. Several reasons account for this. Large families with many affecteds and obligate carriers are rare, in part because the malformations are often lethal before the individuals reach reproductive age. Given the number of genes involved in left–right development (and therefore potentially involved in left–right malformations), the underlying genetic defect is unlikely to be the same in any two or more large families.

Morelli and colleagues have detected linkage to chromosome 3p in a single family with LR malformations, and Vitale and colleagues have provided suggestive (but not convincing) evidence for linkage of a left–right gene to chromosome 6p [Vitale et al., in press; Morelli et al., 2001]. The results point to very large regions within their respective chromosomes, suggesting that it will be very difficult to identify the causative genes by traditional positional cloning without additional data to narrow the critical regions.

CANDIDATE GENES

Despite the difficulties, some data have accumulated connecting phenotype with underlying genotype in human left–right malformations. Spectacular success in the study of left–right development among vertebrate model organisms has yielded a large number of candidate genes for study in human left–right malformations. Mutations in human homologues of the mouse genes nodal, ActRIIb, and lefty have been described [Bassi et al., 1997; Kosaki K. et al., 1999; Kosaki R. et al., 1999] Most recently, Muenke and colleagues have reported loss-of-function mutations in the human homologue of cryptic, a gene essential for normal left–right development in mouse [Muenke, personal communication].

ZIC3: MUTATION, MOUSE, AND MIDLINE

Mutations in ZIC3, an X-linked gene encoding a transcription factor, have been identified among both sporadic and familial cases of situs abnormalities [Gebbia et al., 1997]. Males hemizygous for loss-of-function mutations manifest situs ambiguus and, occasionally, midline abnormalities as well. In fact, situs ambiguus and midline anomalies (particularly of the posterior neural tube or of the hindgut) suggest the presence of an underlying ZIC3 mutation. Usually carrier females are free of malformations, but in one family, some females heterozygous for a ZIC3 mutation are situs inversus, whereas the affected males are situs ambiguus.

Two recent studies suggest that ZIC3 may be involved in malformations not necessarily thought to be directly associated with normal left–right development. Purandare and colleagues (personal communication) have reported discovery of an interstitial deletion in Xq26 that includes ZIC3. Several family members have inherited the deletion, including an affected male with features of VACTERL-H but without characteristic situs abnormalities [B. Casey, unpublished results]. Bouvagnet and colleagues have described a family harboring a ZIC3 mutation in which affected males have transposition of the great vessels and midline anomalies but no obvious left–right malformations.[Megarbane et al., 2000]. Furthermore, there is a male in this family who harbors the mutation but is anatomically normal.

Midline malformations associated with ZIC3 mutations suggest that this gene may be essential for normal midline development rather than directly involved in left–right axis development per se. Targeted disruption of murine Zic3 supports this hypothesis. All Zic3-null mice that survive intrauterine development have a kinked tail, and approximately 10% manifest other malformations. Among these are heart anomalies, altered lung lobation, and gut malrotation, all characteristic of abnormal left–right specification. Malformations of the neural tube and axial skeleton appear independently of left–right-axis anomalies and with equal frequency. Skeletal abnormalities include asymmetric, homeotic transformations.

Symmetric nodal expression at the node in Zic3-deficient mice begins appropriately but disappears rather than becoming asymmetric as in the wild type. Subsequent expression of nodal and the downstream gene Pitx2 in the lateral plate mesoderm is randomized rather than consistently left-sided. These results suggest that Zic3 functions in murine left–right development by helping to maintain nodal expression at the node.

CONSIDERATIONS FOR THE FUTURE

Most of what we know about vertebrate left–right asymmetry has been discovered only in the last few years. Chick, mouse, frog, and zebrafish are proving to be powerful, complementary systems in which to sort out the biology of this developmental process. However, the application of this new knowledge to understanding the etiology of human left–right malformations is proving to be complicated. Already, more than 20 genes have been implicated in vertebrate left–right asymmetry, and surely there are many more to be uncovered. All are reasonable candidates for human disease, but none is likely to be associated with more than just a small percentage of cases, given the genetic complexity of the system. Furthermore, work in mouse has shown us that normal left–right development is sensitive to appropriate gene dosage, such that mice doubly-heterozygous for some gene knock-outs manifest left–right malformations, whereas the single heterozygotes do not. This observation, along with the NOD mouse experiments, suggests the somewhat daunting hypothesis that the underlying genetics of many human cases may be quite complex, just as it is for adult-onset diseases such as hypertension and diabetes. The unusually rich supply of mouse models for human left–right malformations will prove to be invaluable in sorting out this hypothesis.