It is now just about 50 years since the genetic background for Down syndrome (DS) was identified 1 2 3 with the most common reason being an extra free chromosome 21, trisomy 21 (T21). Long before then Penrose (as well as some other authors) had suggested that the condition could be caused by a chromosome abnormality; and at the same time he documented a strong maternal age effect with an increasing incidence of DS births to mothers at later reproductive ages 4 5 . Remarkably, a couple of years before the confirmation of the true chromosomal background he also identified a biomarker for germ line and somatic chromosomal mosaicism (the typical dermatopglyphics) in parents and sibs 6 . In the interim it has become clear, primarily by family linkage studies tracing DNA markers along the length of chromosome 21q between parents and children in DS families that the majority of T21 DS cases inherit the extra chromosome 21 from their mother (more than 90%) while in only a minority (less than 10%) the extra chromosome 21 originates from the father 7 8 9 10 11 .

Importantly, the underlying mechanism for this parental sex difference still remains unknown. Nevertheless, it has been generally accepted that the main problem is mal-segregation of chromosomes 21 in an original disomy 21 oocyte, a dogma most recently re-iterated by Oliver et al. 2008, 2009, Cheng et al. 2009, Fledel-Alon et al. 2009 and Cheung et al. 2010 10 12 13 14 15 . Thus it is thought that the segregation of chromosomes 21, taking place at ovulation and after fertilisation in women post puberty is particularly vulnerable and prone to non-disjunction dependent on abnormalities in chiasma formation leading to mechanical instability. It is also generally accepted that a number of other genetic and environmental factors may contribute to the variation in chance of having a child with T21 DS (see e.g. Hunt et al. 2008, Jones 2008, Oliver et al. 2008, Allen et al. 2009, Coppedè 2009, Driscoll et al. 2009, Garcia-Cruz et al. 2009, Ghosh et al. 2009, Hassold and Hunt 2009, Keefe and Liu 2009, Mailhes 2008, Martin 2008, Migliore et al. 2009, Vogt et al. 2009 8 9 10 16 17 18 19 20 21 22 23 24 25 26 ).

We have recently challenged this dogma by suggesting that the most likely predisposing factor in women for T21 conceptions is instead the common occurrence of fetal ovarian T21 mosaicism and in particular the net result of the behaviour of any such T21 oocytes during development from fetal life until adulthood and maturation for ovulation 27 28 . Based on the observation that all the eight fetuses investigated in this respect, where termination of pregnancy had been performed for a non-medical/social reason, showed ovarian mosaicism with an average of 0.54% T21 cells (range 0.20-0.88%, SD 0.23) we concluded that most females might be low grade T21 mosaics. On the other hand, some exceptional women, who are high grade T21 mosaics, will be predisposed to T21 conceptions already at an early reproductive age and endure an associated high recurrence risk 25 29 .

We have here explored the possibility that the low incidence of T21 of paternal origin is correlated to a lower or insignificant level of germinal/gonadal mosaicism in men in comparison to women. Our data are consistent with this hypothesis. Using fluorescence in situ hybridisation (FISH) with dual chromosome 21-specific probes, we have not found a single T21 cell nucleus in a sample of nearly 12.000 relevant cell nuclei from fetal testes, obtained from four male fetuses where termination of pregnancy had been obtained for a non-medical/social reason. These data are highly statistically different from those obtained in our previous study of more than 12.000 cells by screening eight fetal ovaries, where on average we identified one cell in 200 with T21 without any statistically significant inter-individual variation. We further find it highly unlikely that, akin to the situation in women 27 any rare fetal testicular T21 cells that have remained undetected in our study would accumulate during spermatogenesis post puberty.

Using FISH technology with two chromosome 21-specific probes and applying stringent criteria for establishing chromosome 21 copy numbers in the relevant fetal testicular cell nuclei, we could not identify a single T21 cell nucleus in any of these four apparently normal male fetuses in a total cell population of nearly 12.000 (Table 1, Figure 1). We conclude that there is a highly statistically significant sex difference in T21 germ line mosaicism with a much higher incidence in fetal ovaries than testes (p < 0.001). In a previous report we documented an average of 0.54% (range 0.20 - 0.88; SD 0.23) of T21 cell nuclei in fetal ovaries from eight female fetuses 28 .

The overriding aim of the study has been to investigate the underlying reasons for the intriguing and largely unexplored aspect of the parental origin of T21 DS, i.e. why the extra free chromosome 21 in T21 DS originates from the mother in more than 90% of cases and from the father in less than 10% 7 8 9 10 11 . It would now appear that before puberty human males hardly harbour any T21 precursor cells able to generate mature 21-disomic sperm cells. Nevertheless, bearing in mind the rare paternal origin, an incidence below one fetal T21 testicular cell per six or seven thousand could still be relevant in this respect.

It is also essential to note that there are a number of Case Reports in the literature, documenting paternal inheritance with either testicular T21 mosaicism identified per se or inferred from T21 mosaicism found in somatic tissues, most commonly blood lymphocytes. In addition, there are a number of reports demonstrating a raised incidence of disomy T21 sperm in comparison to controls in fathers of T21 cases. The characteristics of these outstanding cases are summarised in table 2  30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 . It has been generally assumed that T21 DS men are infertile, but these reports suggest that at least in cases of T21 mosaicism fertility may be restored. There are also reports of two cases of apparently non-mosaic DS men, who have fathered children 49 50 . T21 DS females, on the other hand, show impaired fertility and premature menopause, but there are many more reports of offspring to apparently non-mosaic DS mothers than DS fathers 51 52 .

The data we have here presented raise a number of additional interesting questions, including in particular:

(1) How does the sex difference in fetal germ line T21 mosaicism come about?

(2) Is there a correlation with somatic T21 mosaicism?

(2) What is the reason for the disomy 21 in sperm from normal males?

In this paper we have documented copy number counts of chromosome 21 by fluorescence in situ hybridisation (FISH) on fetal testes obtained from four apparently normal male fetuses following termination of pregnancy for a non-medical/social reason. Applying stringent criteria for identification of T21 in germ cell nuclei we could not detect a single T21 cell in a population of at least 2000 per case. We conclude that there is a substantial sex difference in incidence of fetal germ line T21 mosaicism where most female fetuses may be ovarian T21 mosaics, while males in this study do not show any detectable degree of fetal testicular T21 mosaicism. We propose that this sex difference in germ line T21 mosaicism may explain the much lower paternal origin of T21 DS than maternal. The mechanisms underlying the rare DS cases where the extra chromosome 21 does originate from the father, remains unknown and further studies in this respect are required.

As we have stressed in our previous publication on this issue 27 28 further large-scale studies will be required to find out if our model on germ line mosaicism leading to secondary meiotic non-disjunction constitutes the major source of aneuploid conceptions in the human population, or if other mechanisms might also contribute to this effect.

All procedures were performed with informed consent and ethical approval from the local ethical committee. Fetal testicular cells were obtained from four fetuses at gestational age 14-19 weeks, following termination of pregnancy for social reasons with all the fetuses having a normal phenotypic appearance. Testes were removed within a few hours post-mortem and placed in L-15 (Leibovitz) medium (Life Technologies) with 0.3% bovine serum albumin (Sigma). Pieces of testes were frozen at -80°C. Parts of the tissue samples were thawed to prepare direct imprints from the cut surface of the fetal ovary 88 and the remaining material processed by micro-spreading 89 .

Microscopy slides for FISH analysis were fixed in methanol: acetic acid (3:1 v/v) then washed in 2× standard saline citrate (SSC) and treated with pepsin (0.1 mg/ml) in 0.01 M HCl for 8 min at 37°C. After additional washing in phosphate-buffered saline (PBS), paraformaldehyde (1%) fixation and dehydration through series of alcohol the slides were left to air-dry at room temperature. Hybridisation was performed according to the manufacturers' instructions with two DNA probes positioned near the end of the long arm of chromosome 21 and labelled in SpectrumOrange and SpectrumGreen respectively (Cat No: 32-190002, Abbot Molecular Inc, USA and Cytocell, Cat No. LPT21QG/R, Cytocell Technologies Ltd. UK). A chromosome 18 centromeric probe labelled in SpectrumAqua was added to be able to differentiate between trisomy and triploidy (Cat No: 32-131018, CEP 18 (D18Z1) SpectrumAqua Probe). The DNA probes were mixed and added to the slides followed by denaturation, hybridisation and post-hybridisation washing. After dehydration slides were mounted in glycerol containing 2.3% DABCO (1, 4-diazabicyclo-(2, 2, 2) octane) as antifade and DAPI (4, 6,-diamino-2-phenyl-indole) 0.5 mg/ml for nuclear counterstaining.

Fluorescent signals were analyzed using a Zeiss Axioskop 2 microscope equipped with a cooled CCD camera (CoolSnap; Photometrics Ltd, USA) controlled by a Power Macintosh computer. Grey scale images were captured, pseudocolored and merged using the SmartCapture 2 software (Digital Scientific Ltd, UK).

In scoring chromosome 21 copy number we focussed attention in particular on the testicular germ cells, i.e. the gonocytes, the intermediate cells and the pre-spermatogonia, identified by their specific morphology 90 . The images of two cell nuclei, showing two dual chromosome 21-specific signals (red and green) and two chromosome 18 control probe signals (ice blue) are illustrated in Fig 1.

The authors declare that they have no competing interests.

MH designed the study, supervised the practical work and wrote the initial draft of the paper; EI and SP performed the FISH analysis; MW and NP obtained local ethical approval; EI, MW, NP and AMJ obtained the samples. All the authors contributed to and have approved the final version of the manuscript.