Twinning, Dizygotic

Inheritance

Weinberg (1909) suggested that hereditary twinning is transmitted only through the female line, applies only to dizygotic (DZ) twins, and is probably recessive. Observation of multiple births following use of pituitary gonadotropins suggests a pituitary mechanism for gene action (Milham, 1964).

Wyshak and White (1965) presented evidence, based on Mormon records, which they interpreted as supporting recessive inheritance. Among the children of female DZ twins, 17.1 twins per 1,000 maternities occurred as compared with 7.9 among children of male DZ twins. Female sibs of DZ twins had 17 per 1,000 twins, whereas male sibs had 13.1 per 1,000. Supposedly the gene is too frequent for one to expect increased consanguinity in the grandparents of dizygotic twins.

Taylor (1931) reported multiple sets of dizygotic twins in 4 generations. In the same family, mother and daughter did not menstruate until after their first pregnancies, at ages 20 and 22, respectively. The differentiation of multifactorial and monofactorial inheritance of twinning is difficult. Ethnic differences in the rate of dizygotic twinning is evidence of genetic factors. In interracial marriages the rate follows that of the mother's ethnic group. Furthermore, when the mother is a racial hybrid, the dizygotic twinning frequency is that of the race with the lower frequency (Morton et al., 1967).

The probability of a subsequent twin pregnancy is increased 4-fold in mothers of twins, and the risk of having dizygotic twins is roughly doubled for women whose mother or sister has dizygotic twins (Bulmer, 1970; Meulemans et al., 1996).

Meulemans et al. (1996) studied spontaneous dizygotic twinning in 1,422 3-generation pedigrees in a Flemish and Dutch population. The phenotype was consistent with an autosomal monogenic dominant model, with a gene frequency of 0.035 and a female-specific lifetime penetrance of 0.10. X-linked, recessive, polygenic, and sporadic models were rejected.

The frequency of twin deliveries varies among human populations. The highest twinning rates for Caucasian populations have been recorded on the archipelago of Aland and Aboland, in southwest Finland, whereas multiple deliveries in adjacent mainland areas are known to be rarer. Using data from the pre-industrial era (1752-1850), Lummaa et al. (1998) compared the lifetime reproductive success of mothers who had produced twins with that of mothers of singletons in these archipelago and mainland sites. When they restricted their analysis to mothers with a genetic tendency to produce twins, they found that lifetime reproductive success was maximized by having twins on the archipelago, but by having singleton offspring on the mainland. They considered this result to be consistent with the difference in twinning rate being maintained by natural selection. The 'profitability' of twinning would, in their model, depend on the level of resources. The model suggested that predictable resource levels favored the evolution of increased reproductive output. In the archipelago, the amount of food available had traditionally been relatively high and constant, with total crop failures being rare and with survival ensured by fishing. In poor mainland areas, on the other hand, crop failures and subsequent famines had been common throughout the centuries. The authors assumed that only dizygotic twinning has a genetic component (Parisi et al., 1983) and considered only the analysis of dizygotic twinning as relevant to evolution of the number of children produced at one time driven by natural selection. Results using only mothers with twins of different sexes confirmed the evolutionary inference.

Lewis et al. (1996) discussed a genetic basis of dizygotic twinning.

A curious feature of the biology of twinning is the possibly reduced cancer frequency among twins as compared to singly born individuals. This was reviewed for the incidence of cancer in children by Murphy et al. (2001) and in adults by Iversen et al. (2001).

Hoekstra et al. (2008) studied twinning and fertility in 8,222 and 5,505 women with spontaneous DZ and MZ offspring and 4,164 and 250 women with assisted reproduction techniques (ART) DZ and MZ twin pairs, respectively. Hoekstra et al. (2008) observed that women with spontaneous DZ twins more often reported female relatives with twins than those with spontaneous MZ twins. The proportion of DZ versus MZ twin offspring in relatives was also larger in women with spontaneous DZ offspring than in women with MZ offspring. The first group of women reported a shorter time to conceive. Hoekstra et al. (2008) did not observe that DZ twinning was more familial in women who had their twins before age 36 years compared to older women. Women with ART twins had fewer sibs and offspring and less often reported relatives with twins. Hoekstra et al. (2008) concluded that the mechanisms underlying spontaneous and nonspontaneous DZ twinning are different.

Mapping

Busjahn et al. (2000) found linkage of dizygotic twinning to the gene encoding peroxisome proliferator-activated receptor-gamma (PPARG; 601487). Linkage in 181 dizygotic twin pairs yielded a 6.93 lod score. They found, furthermore, that the PPARG C-to-T polymorphism (601487.0009) was far less common in monozygotic twins than in dizygotic twins. Dizygotic heterozygosity was less than expected and transmission disequilibrium test analysis gave further evidence for association. In dizygotic twins, the C-to-T substitution was not in Hardy-Weinberg equilibrium. In contrast, the pro12-to-ala PPARG polymorphism (601487.0002), which occurs in exon B and is found in the PPARG2 splice variant in white fat, was not in tight linkage disequilibrium and obeyed Hardy-Weinberg equilibrium. Twinning requires both multiple conceptions and intrauterine survival of both twins. 'Vanishing twins' result in the observation that about 40% of spontaneous dizygotic twin pregnancies result in singleton births. The authors suggested that intrauterine selection may be responsible for the Hardy-Weinberg deviation of the C-to-T substitution and that PPARG may be a gene involved in the intrauterine survival of dizygotic twins.

Duffy et al. (2001) were unable to replicate the findings of Busjahn et al. (2000) in 4 datasets of dizygotic twins, the combined dataset, or a dataset comprising mothers of twins. Duffy et al. (2001) found no evidence of linkage of the region around the PPARG locus to survival of dizygotic twins.

Natural multiple pregnancy in women leading to dizygotic twins is familial and varies across racial groups, suggesting a genetic predisposition. Mothers of DZ twins have a higher incidence of spontaneous multiple ovulation and elevated FSH (136530) concentrations. FSH release is controlled by feedback of inhibin peptides from the ovary, and immunization against inhibin alpha-subunit (INHA; 147380) results in an increased ovulation rate in animals (e.g., Hillard et al., 1995). The inhibin alpha-subunit is therefore a candidate gene for mutations that may increase the frequency of DZ twinning. Montgomery et al. (2000) examined the INHA locus for polymorphism and analyzed linkage to DZ twinning in women. Restriction digests of a PCR product from exon 1 with the enzyme SpeI detected a C/T polymorphism at bp 128 with 2 alleles of 447 and 323/124 bp. Montgomery et al. (2000) typed the polymorphism in 1,125 individuals from 326 pedigrees with 717 mothers of spontaneous DZ twins. The alpha-inhibin locus mapped within 3 centimorgans of D2S164, and linkage with DZ twinning was excluded (lod score of -2.81 at theta = 0). There was complete exclusion of linkage (lod less than -2) of a gene conferring relative risk 1.8 (lambdas, greater than 1.8) across the chromosome, except at the p-terminal region and a small peak (maximum lod score 0.6) in the region of D2S151-D2S326. Analysis using either recessive or dominant models excluded linkage with DZ twinning in this population (lod score, less than -2.5) across chromosome 2. The authors concluded that DZ twinning is not linked to variation in the alpha-inhibin locus. The authors also concluded that mutations in other candidates on chromosome 2, including the receptor for FSH and the beta-2-inhibin subunit (INHBB; 147390), cannot be major contributors to risk for DZ twinning.

Montgomery et al. (1993) mapped a gene responsible for multiple ovulation in sheep to a region of the sheep genome homologous to 4q in the human. Specifically, they showed that it was linked to markers in the region of 4q21-q25. Martin et al. (1984) described higher concentrations of follicle-stimulating hormone in mothers of dizygotic twins compared with mothers of single births, in agreement with results in the sheep for ewes carrying the Booroola fecundity gene (FECB; 134720).

Duffy et al. (2001) applied nonparametric affected sib pair methods for linkage analysis to the study of 169 pairs and 17 sets of 3 sisters (trios) who had each had spontaneous DZ twins, mostly before the age of 35. Exclusion of markers in the region 4q21-q25 led them to conclude that if there is a gene influencing DZ twinning on chromosome 4, its effect must be minor.

Derom et al. (2006) failed to identify positive lod scores greater than 2 in a genomewide linkage analysis of 14 Flemish families with 57 mothers of spontaneous dizygotic twins. The families were ascertained from the study of Meulemans et al. (1996).

Mbarek et al. (2016) performed a genomewide association study in 1,980 mothers of spontaneous DZ twins and 12,953 controls and identified significant association with 2 SNPs: rs11031006, located 5-prime of the FSHB start site on chromosome 11p13 (p = 1.54 x 10(-9)), and rs17293443, within the first intron of SMAD3 on chromosome 15q22 (p = 1.57 x 10(-8)). These findings were replicated in a large Icelandic cohort involving 3,597 mothers of twins and 297,348 controls (p = 3 x 10(-3) and p = 1.44 x 10(-4), respectively). Based on approximately 90,000 births in Iceland between 1950 and 1991, Mbarek et al. (2016) estimated that the risk of a twin birth increased by 18% for each maternal rs11031006 'G' allele and by 9% for each rs17293443 'C' allele. A higher polygenic risk score (PRS) for DZ twinning, calculated based on the reuslts of the DZ twinning GWAS, was significantly associated with DZ twinning in Iceland (p = 0.001). A higher PRS was also associated with having children (p = 0.01), greater lifetime parity (p = 0.03), and earlier age at first child (p = 0.02). The G allele of rs11031006 was associated with higher serum FSH levels, earlier age at menarche, earlier age at first child, higher lifetime parity, lower risk of polycystic ovary syndrome, and earlier age at menopause. Conversely, the C allele of rs17293443 was associated with later age at last child.

See 136435 for a discussion of a possible association between dizygotic twinning and polymorphism in the follicle-stimulating hormone receptor (FSHR).