Spermatogonial stem cells replicate and produce sperm throughout adult life, while oocyte precursors complete all of their mitotic cell divisions during fetal development and primary oocytes are arrested in meiosis at birth. Because many more rounds of cell division occur in spermatogenesis than in oogenesis, the incidence of germline mutations, and particularly of single base substitutions, is higher in sperm than in ova (Crow, 2000). Most single base substitutions occur roughly four or five times more frequently in sperm than in eggs, and the incidence of these mutations increases modestly with paternal age. Against this background, a handful of diseases, known as paternal age effect (PAE) diseases, stand out. PAE diseases, of which achondroplasia is the most common and best known, have several unusual features. Almost all (more than 95%) of the new mutations that cause these diseases occur in sperm; the incidence of these diseases increases significantly with paternal age, such that the fathers of affected patients are on average several years older than the fathers of unaffected babies in the same populations; and the genes underlying these diseases have unusually high apparent mutation rates. The mutations that cause PAE diseases are dominant, gain of function mutations in genes related to the RAS signal transduction pathway. Achondroplasia is due to mutations in fibroblast growth factor receptor 3 gene (FGFR3), while Costello syndrome, another PAE disease, is caused by mutations in the HRAS proto-oncogene.
Anne Goriely and Andrew Wilkie have marshaled evidence in favor of the hypothesis that the epidemiologic features of PAE diseases are due to positive selection of spermatogonial stem cells (Goriely and Wilkie, 2012). Spermatogonial stem cells, like other stem cells, normally divide asymmetrically. One daughter cell retains the stem cell phenotype while the other differentiates, undergoes several rounds of cell division, and ultimately gives rise to spermatocytes. Spermatogonial, or germline, selection is thought to result from mutations that cause the stem cells to undergo occasional symmetrical division. Symmetrical cell division produces two daughter stem cells and thus would increase the frequency of these mutations in the stem cell population and, ultimately, the frequency of affected sperm (Crow, 2012). Spermatogonial stem cells replicate about 23 times a year, so these cells undergo many hundreds of rounds of replication during a man’s adult life. Even a small selective advantage could lead to a large increase in the abundance of mutations in the testes and sperm of older men.
Several recently published studies have provided strong support for the hypothesis of germline selection. In one heroic study, Deepali Shindi, Dominik Emer, and their colleagues determined the spatial distribution of the most common achondroplasia-associated FGFR3 mutation in 192 tissue samples taken from the testis of an unaffected 80 year old man (Shinde et al., 2013). The mutations were not distributed randomly throughout the testis but rather were concentrated in a few discrete areas. Ninety-five percent of the mutations were found in 27% of the testis. The distribution of FGFR3 mutations could not be accounted for by a high mutation rate but was best fit by a model in which affected spermatogonial stem cells had a proliferative advantage over unaffected cells. Computer simulations suggest that only a small probability of symmetric cell division, on the order of 1%, would be sufficient to account for the distribution of FGFR3 mutations.
In complementary and also technically challenging experiments, Eleni Giannoulatou, Gilean McVean, and their colleagues quantified the levels of HRAS mutations in sperm samples of 89 men, ages 22–74, and in a smaller number of blood samples from men of different ages (Giannoulatou et al., 2013). The frequency of HRAS mutations in sperm, but not in blood, increased dramatically with age. The pattern of increasing frequency of HRAS mutations with age is inconsistent with the hypothesis of a mutation hot spot and is best accounted for by a model of germline selection. Again, a selection coefficient of less than 1% could account for the observed increase in frequency of mutations with age. Together, these two new studies provide compelling evidence that germline selection plays a role in the epidemiology of PAE diseases.
Germline selection is a fascinating and still relatively unexplored level of selection. Although spermatogonial stem cells are germline cells, they have the properties of somatic stem cells. Germline selection is analogous to the somatic selection that occurs in oncogenesis. Indeed, the mutations that result in PAE diseases may, on occasion, lead to spermatocytic seminomas. The PAE diseases that have been identified are caused by dominant mutations that have marked effects of growth. Almost certainly, other mutations will be found that confer a smaller selective advantage on affected spermatogonia and that have more subtle phenotypic effects. The contributions of germline selection to human variation and disease remain to be elucidated.
Crow, J. F., 2000, The origins, patterns and implications of human spontaneous mutation: Nat Rev Genet, v. 1, p. 40-7.
Crow, J. F., 2012, Upsetting the dogma: germline selection in human males: PLoS Genet, v. 8, p. e1002535.
Giannoulatou, E., G. McVean, I. B. Taylor, S. J. McGowan, G. J. Maher, Z. Iqbal, S. P. Pfeifer, I. Turner, E. M. Burkitt Wright, J. Shorto, A. Itani, K. Turner, L. Gregory, D. Buck, E. Rajpert-De Meyts, L. H. Looijenga, B. Kerr, A. O. Wilkie, and A. Goriely, 2013, Contributions of intrinsic mutation rate and selfish selection to levels of de novo HRAS mutations in the paternal germline: Proc Natl Acad Sci U S A, v. 110, p. 20152-7.
Goriely, A., and A. O. Wilkie, 2012, Paternal age effect mutations and selfish spermatogonial selection: causes and consequences for human disease: Am J Hum Genet, v. 90, p. 175-200.
Shinde, D. N., D. P. Elmer, P. Calabrese, J. Boulanger, N. Arnheim, and I. Tiemann-Boege, 2013, New evidence for positive selection helps explain the paternal age effect observed in achondroplasia: Hum Mol Genet, v. 22, p. 4117-26.