Sins of the fathers: sperm DNA damage in the context of assisted reproduction

A permanent challenge in Assisted Reproduction is the uncovering of methodologies to allow for better gamete and embryo evaluation and selection, which may in turn lead to higher success rates in general, and to more predictable evaluations of specific cases. In this regard sperm analysis holds several interesting paradoxes. While the number of male gametes ultimately required for fertilization almost always allows material to be available for analysis, it does not necessarily follow that the same cells analysed may then be of clinical use. From another perspective, while classical sperm evaluation via the standard spermiogram (WHO, 2010) provides a technically easy and affordable measure of ejaculate quality, the notion that the data obtained are relatively poor at predicting the success of any intervention is as unanimous a statement as is likely to be found in the field. Besides sperm quantity, motility and morphology (and viability) the need for better assessment tools and parameters has been clear for many years (Said and Land, 2011; Sousa et al., 2011). In addition to standard spermiogram measurements, the sperm parameter that arguably has received the most attention is DNA integrity, given its potential importance on the quality of the genetic information transmitted to a developing embryo, and thus on the likelihood that impaired sperm DNA may negatively impact the success of Assisted Reproduction. Additionally, sperm DNA status may also serve as an important indicator to evaluate both defects in spermatogenesis, or the possible effects of different environmental exposures on male reproductive function (Perry, 2008; Sousa et al., 2009; Tavares et al., 2013). Although there seems to be a general agreement as to the importance of measuring sperm DNA integrity, the literature is controversial (if not downright confusing), with some studies suggesting a modest predictive power that does not support general implementation of the parameter (Collins et al., 2008), while others show more promising outcomes. Namely, sperm DNA integrity was shown to have predictive power (higher for IUI and IVF, lower for ICSI) in terms of pregnancy outcomes (Evenson and Wixon, 2006; Simon et al., 2010, 2014) and to be useful in terms of predicting pregnancy loss (Zini et al., 2008; Robinson et al., 2012), but not in the prediction of embryo quality (Zini et al., 2011). No doubt the lack of agreement in the literature is partially due to the many variables involved, and some of the main issues that need tackling have been identified (Zini and Sigman, 2009; Barratt and De Jonge, 2010; Barratt et al., 2010; Said and Land, 2011). How is DNA integrity specifically measured, and how can measurements be robustly standardized? What does each test monitor, which is more appropriate, what controls need to be performed? How many groups of samples with varying DNA damage does each test define, and how should they be established? What are the cut-off values that identify potential problems and that should function as ‘red flags’ for clinicians in terms of therapeutic options? In which samples should sperm DNA integrity be measured (native versus prepared sperm)? Were published tests carried out with any degree of blinding, and what strategies were employed to deal with potential outliers? What exactly can sperm DNA integrity help predict (fertilization rates, embryo quality, embryo development, pregnancy, pregnancy loss)? In terms of the specific Assisted Reproduction technologies, are there differences in the predictive power of sperm DNA integrity for IUI, IVF or ICSI? When should measurements be performed, and in what samples, in order to be cost-effective in a clinical setting? What other parameters/factors should be taken into account in models of increasing complexity to allow for relevant predictions? What statistical analyses should be performed and what is their true predictive power? This final point is often overlooked; much more than a ‘mere’ P-value below 0.05 (still a ‘holy grail’ in most papers), what needs to be determined is to what extent sperm DNA evaluation can be truly relevant. When considering the tests available, it should be noted that they monitor distinct aspects of sperm DNA integrity (DNA packaging, different types of strand breaks) with varying degrees of robustness, technical complexity and cost. Among the most well established tests are the SCSA (Sperm Chromatin Structure Assay; Evenson et al., 1980), TUNEL (terminal deoxynucleotidyl transferase-mediated dUDP nick-end labelling; Gorczyca et al., 1993), Comet (single-cell gel electrophoresis assay; Hughes et al., 1996) and SCD (sperm chromatin dispersion test, commercially available as the Halosperm kit; Fernandez et al., 2003). Some of these tests (e.g. TUNEL) can be carried out both by flow cytometry and fluorescence microscopy, with concomitant changes in affordability, degree of potential subjectivity and number of cells that can be analysed; although agreement between measurements is certainly possible and the methodology adaptable (to some extent) to distinct laboratorial settings (Varum et al., 2007). Additionally, other tests take advantage of common methodologies normally used to monitor sperm