Transplantation experiments in mice reveal that the increased risk of congenital heart disease in the pups of older mothers is not conferred by ageing eggs, but by the mothers' age, and can be mitigated by exercise. See Letter p.230
Congenital heart disease is the most common developmental malformation in humans, and a leading cause of death in infancy. About 1 in 100 children have minor congenital heart disease, and 1 in 1,000 children will need heart surgery. Epidemiological studies1,2 indicate that there are many risk factors for congenital heart disease: infections, genetics, environmental factors such as toxins or maternal diabetes, and higher maternal age, to name just a few. In this issue, Schulkey et al.3 (page 230) question whether, in terms of age, it is that of the egg or of the mother herself that affects the risk of congenital heart disease in offspring of older mothers. Their answer is clear — the risk lies, surprisingly, in the age of the mother.
The heart is the first organ to form in the vertebrate embryo, following a precisely orchestrated and evolutionarily conserved developmental process4. In mice, septation of the heart into a four-chambered structure consisting of two atria and two ventricles begins with the formation of the interatrial and interventricular septa5. A master regulator of this process is Nkx2-5, a cardiac homeobox transcription factor. The gene that encodes this protein is frequently mutated in humans with septation defects6.
Schulkey and colleagues took advantage of mice that have only one functional copy of the Nkx2-5 gene, instead of the usual two, and that as a result are more likely to develop ventricular septal defects (VSDs)7. The authors performed reciprocal ovarian transplants, in which ovaries from young mice were transplanted to older mice and vice versa, and convincingly showed that the risk of VSD increases with the advancing age of the mother, but not of the ovaries — and therefore, not of the egg (Fig. 1). Next, to rule out the main environmental factors implicated in epidemiological studies of congenital heart defects (maternal diet, body mass and glucose intolerance), pregnant mice were fed a high-fat diet. A higher body weight and altered glucose metabolism did not change VSD incidence in young mothers, and did so only marginally in older mothers.
Where, then, does the VSD risk originate, and can it be mitigated? To examine whether genetic factors are also involved in risk, the authors bred their mice with those from other strains with different genetic make-ups (genetic backgrounds). Changes in genetic background modified the risk of maternal-age-associated VSD, thus providing evidence for a complex gene–environment interaction. Finally, and importantly, Schulkey et al. demonstrated that VSD risk can be considerably decreased by placing mice in running wheels. Small amounts of exercise were insufficient, but beyond a threshold amount, exercise had a protective effect.
This study is relevant for several reasons. Because the authors' mouse model recapitulates the complex nature of human congenital heart disease, it makes individual contributors of disease risk quantifiable. The reciprocal transplant protocol will be instrumental in studying whether maternal-age-associated risk is also a factor in mice harbouring other genetic mutations, and whether the risk is identical in subsequent pregnancies. Schulkey and colleagues' model will also enable us to identify the genetic and environmental modifiers associated with risk, and to devise intervention strategies to make human pregnancies safer.
At present, we can only speculate about the nature of the factors that increase the maternal-age-related risk of VSDs. On the basis of what is known about heart development, and of constraints imposed by the design of the current study, we can assume that the elusive factor or factors must be embryo-independent, maternally based, influenced by the maternal genome and affected by exercise with a threshold effect. Two possibilities spring to mind: metabolic and epigenetic factors.
Epigenetic factors alter gene expression, without changing the underlying DNA sequence, in a permanent manner that can be passed down through generations of cells and, in sperm and eggs, through generations of organisms. Such transgenerational epigenetic inheritance has been described in the pups of mice with reduced activity of the gene Mtrr — developmental heart defects arise not only in mutant offspring, but also in up to four generations of wild-type descendants8. Transfer of embryos habouring the Mtrr mutation into wild-type mothers revealed two separable traits: growth defects caused by the uterine environment, and congenital abnormalities passed on through epigenetic inheritance. The underlying molecular processes are unclear. However, the defects brought about by Mtrr mutation can be mitigated by supplementing the maternal diet with folate, and evidence9 suggests that folate supplementation might work by altering the patterns of an epigenetic modification called methylation. Whether the influence of age and exercise on epigenetic effects is independent of, or under the control of, folate levels could be tested experimentally using Schulkey and colleagues' model. Controlled interventional studies10,11 show that exercise can change the methylation levels of several genes and pathways, including those involved in metabolism, signalling and transcription.
Several questions remain to be answered. First, what is the precise role of Nkx2-5 in the maternal placenta and in those tissues derived from the fertilized egg that do not form the embryo, such as the yolk? A step towards answering this question would be to determine whether the outcome remains the same when embryos instead of ovaries are transplanted — although, admittedly, the sheer magnitude of such an undertaking seems prohibitive. Second, what about paternal mutations, which were not analysed in this study? Third, what is the role of other types of age-related genetic factor, such as mutations resulting in placental growth disturbances? A wide range of factors, including age-related decline in metabolic or placental functions, could conceivably contribute to such anomalies. But on the basis of both Schulkey and co-workers' model and another recent study12, it seems clear that these factors must be under genetic control.
Future studies will need to be large, and to focus on multigenerational families, if they are to provide sufficient statistical evidence for genetic risk. Studies similar to Schulkey and colleagues' should be designed to better quantify epigenetic risk factors. Subsequently, the fraction of risk attributable to different factors needs to be defined, in particular those amenable to interventional strategies, such as the use of folate supplements. In light of the increasing average maternal age at childbirth, and the advent of egg freezing as an employee benefit in some countries, the authors' study provides a timely opportunity to improve risk stratification and management of the stages immediately before and after conception, to prevent birth defects.
Miller, A., Riehle-Colarusso, T., Siffel, C., Frías, J. L. & Correa, A. Am. J. Med. Genet. A 155, 2137–2145 (2011).
Ferencz, C., Rubin, J. D., Loffredo, C. A. & Magee, C. A. (eds) The Epidemiology of Congenital Heart Disease: The Baltimore-Washington Infant Study 1981–1989 (Futura, 1993).
Schulkey, C. E. et al. Nature 520, 230–233 (2015).
Lescroart, F. et al. Nature Cell Biol. 16, 829–840 (2014).
De La Cruz, M. V., Giménez-Ribotta, M., Saravalli, O. & Cayré, R. Am. J. Anat. 166, 63–72 (1983).
Kasahara, H., Seidman, C. E. & Izumo, S. J. Clin. Invest. 106, 299–308 (2000).
Winston, J. B. et al. Circ. Cardiovasc. Genet. 5, 293–300 (2012).
Padmanabhan, N. et al. Cell 155, 81–93 (2013).
Ionescu-Ittu, R. et al. Br. Med. J. 338, b1673 (2009).
Nitert, M. D. et al. Diabetes 61, 3322–3332 (2012).
Rönn, T. et al. PLoS Genet. 9, e1003572 (2013).
Mitchell, L. E. et al. Hum. Mol. Genet. 24, 265–273 (2015).