Oxidative stress is a common condition suffered by biological systems in aerobic conditions. Human semen also has its own molecular guard against the free radicals created by normal respiratory process or from immune reactions. The equilibrium of the creation and scavenging of free radicals is mandatory in the spermatozoa to fertilize and initiate a full-term pregnancy. The paper is a systematic review of publications that evaluate oxidative stress in semen. The Cochrane Library, Medline (1966-2003), Embase (1988-2003), SciSearch (1981-2003) and the conference papers were searched. When sperm samples from fertile and infertile males were analyzed, some of the mechanisms that determine the oxidative stress level were found to be impaired while others were unaltered. In conclusion, the literature as a whole provides contradictory findings and it is necessary to carry out more work to identify all the enzymatic and non-enzymatic systems involved in oxidative stress in the ejaculate, in order to develop new diagnostic systems and therapeutic strategies for combating detrimental free radical imbalance in the semen.

In recent years, there has been a growing concern regarding the progressive decline in male fertility. Different studies [1-3], primarily based on the microscopic analysis of semen samples, as dictated in the guidelines of the World Health Organization (WHO) [4] , support this affirmation.

However, the issue is controversial, and there is a lack of consensus that keeps the debate open. A relevant percentage of apparently normal males are unable to impregnate a woman, even when the female is also considered to be normal.

Male fertility markers have been scrutinized in order to comprehend the molecular events that can lead to subfertility and permit an accurate diagnosis and design of therapeutic protocols. Among these markers, the study of oxidative stress (OS) status in semen has emerged as a promising field [5, 6].

Figure 1. Total contribution to oxidative stress status in an ejaculate. Pro- and anti-oxidative molecules have antagonistic functions in semen.

Cells in a biological environment (such as an ejaculate) contribute to the maintenance of the oxidative homeostasis via different systems, given that a low and controlled level of ROS is necessary for the normal functioning of the cells. This control depends on antioxidants [5], molecules which are grouped into two classes depending on their nature: enzymatic and non-enzymatic.

The presence of various antioxidant systems, together with that of oxidant products of normal metabolism, has been described in spermatozoa and seminal plasma. An equilibrium between these is necessary for the preservation of sperm DNA integrity, an adequate acrosome reaction and sperm-oocyte fusion, among other functions [6].

The undermining of the protective capacity of the system by ROS results in irreparable damage to DNA molecules, while the acrosome reaction is deregulated and the sperm-oocyte fusion finally disrupted or impeded [6].

Some studies have revealed the effect of these phenomena on male fertility, since almost 40% of infertile males display abnormally increased ROS levels. It is known that human spermatozoa are extremely sensitive to ROS-induced damage due to their plasma membrane composition [8, 9].

With this review, we aim to update the current knowledge regarding the role of OS in physiological function in human semen and its implications for male fertility.

Human semen is a complex mixture of cells contained within the seminal plasma, a secretion consisting of a combination of diverse products synthesized along the whole male genital tract, including the seminiferous tubules and accessory glands.

Seminal cells include both mature and immature spermatozoa, round cells from different stages of the spermatogenic process, leukocytes and other occasional cell types, such as epithelial cells. Among these, immature spermatozoa and leukocytes are the main sources of ROS (Figure 2).

Immature spermatozoa are well characterized sources of ROS and a negative correlation between ROS production and semen quality has been documented [10]. It has been hypothesized that abnormal spermatozoa with an aberrant cytoplasmic droplet retention are important ROS producers, since they retain an excess of cytoplasmic enzymes that are involved in glucose metabolism, such as glucose-6-phosphate dehydrogenase, NADPH oxidase system and NADH dependent oxidoreductase. This metabolism occurs at two different sites: the plasma membrane and the mitochondria. Serial steps of different ROS production and combination result in the formation of distinct ROS, such as superoxide anion, hydrogen peroxide and hydroxyl radicals, and initiate molecular damage through their interaction with cellular macromolecules (Figure 2).

Moreover, it has been confirmed that, when seminal ejaculate is fractioned by a density gradient method, the layer with immature sperm and the highest percentage of immature cells produces most ROS. This ROS production by immature spermatozoa is also directly correlated with the extent of DNA damage to mature sperm, and the higher the ROS production, the lower the percentage of mature spermatozoa [10, 11].

The second ROS-producing cell type is the lineage of immune white cells in the ejaculate. They trigger their defensive role directly by ROS synthesis or indirectly by other neighboring white cells via soluble factors as cytokines. ROS synthesis in response to an infection is thought to be the first line of defense against intruding microorganisms. In this way, the semen reaction against leukospermia is a mechanism that could interfere in assisted reproduction where an infection is present, since it can indirectly cause significant collateral DNA damage in spermatozoa.

Two further aspects must be considered: first, concentrations of leukocytes that are normal, according to WHO criteria, can produce damaging ROS levels [12] and second, leukocytes can stimulate ROS production in sperm [13].

This fluid is especially relevant for protecting spermatozoa from ROS, since its cytoplasmic volume is low in comparison to cell types and the surrounding media. The diminished cytoplasm of the spermatozoa reduces its capacity to retain adequate loads of protective molecules inside the cell, which in turn protects the cell from extracellular ROS attack. Thus, protection is provided by the extracellular environment (Figure 2).

Several enzymes and enzymatic systems have previously been shown to be related with protection against OS in the seminal plasma, and a direct relationship with the fertile/infertile status of the male has been illustrated.

Different studies have demonstrated the presence of superoxide dismutase (SOD) in both seminal plasma and human spermatozoa since the initial report in 1980 [14]. The SOD family is divided into three different classes according to the catalytic metal present at the active site. Cu/Zn-SOD and Mn-SOD are the main forms present on eukaryotes, the former being found mainly on red blood cells and the latter on mitochondria. These enzymes catalyze the dismutation of superoxide into hydrogen peroxide and oxygen.

Initially, it was believed there was a direct relationship between the sperm SOD activity in spermatozoa (but not in plasma) and the sperm motility [15]. Addition of exogenous SOD (400 U/mL) to the sperm suspension significantly decreases the loss of motility over time and the increase in the concentration of malonyldialdehyde is thought to be related with the percentage of immotile cells. These data suggest a significant involvement of SOD in sperm motility, as it would seem that SOD inhibits the lipid peroxidation of human spermatozoa that causes loss of motility.

Nevertheless, the results obtained by different recently published studies do not confirmed these previous findings and the influence of SOD on seminal quality and male fertility is being readdressed. Some authors have been unable to correlate SOD activity with seminal parameters and reproductive success in vivo or in vitro [16-18].

Some authors have detected differences in the seminal catalase activity of asthenozoospermic and oligo-asthenozoospermic specimens with hyperviscosity [20], but the characteristics of catalase in the ejaculate has not yet been studied in depth.

In terms of their site of synthesis, SOD and catalase [21] are equally present in seminal plasma from fertile and vasectomized males, thus suggesting a post-testicular secretion of these enzymes.

Other ROS-scavenging enzymes, such as glutathione peroxidases (GPx), have also been measured in seminal plasma and correlated with male fertility [22].

The GPx family of enzymes is comprised of several forms and they are capable of detoxifying organic and hydrogen peroxides and converting them to stable alcohols or water. Consequently, cells are protected from oxidative damage. The GPx enzymes containing Se (as a selenocysteine on the molecule) require the presence of reduced glutathione (GSH), which restores the oxidized Se.

In conjunction with this reaction, glutathione reductase (GR) catalyzes the process by which glutathione changes from oxidized to reduced form, thereby restoring the GSH stock on which GPX depends.

Yeung et al [29] have also studied the GPX, GSH reductase, superoxide dismutase and the catalase-like enzyme activities. All were quantified in seminal plasma from normozoospermic patients, men with obstructive problems, men who were already fathers and patients undergoing in vitro fertilization treatment. In one of the most complete studies found in the literature, the authors were unable to detect glutathione and the values of said activity were comparable in vasectomized and non-vasectomized males, suggesting that the origin of these enzymes is neither testicular nor epididymal [23]. In fact, their correlation with the accessory gland markers point to the prostatic origin of all the enzymes except for the catalase-like activity.

Non-enzymatic protection was described by Alvarez and Storey in well-known experiments in which they demonstrated in vitro the protective effect against rabbit spermatozoa lipid peroxidation of molecules such as taurine, hypotaurine, epinephrine, pyruvate, lactate and bovine serum albumin [24]. All of these are supposedly present in seminal liquid.

Recently, it has been shown that taurine and hypotaurine molecules are present at different concentrations in the seminal plasma of infertile males and have also been associated with specific sperm defects [25].

Other non-enzymatic molecules are likely to be involved in the scavenging of free radicals in the seminal plasma. However, though there is plenty of information regarding their scavenging activity in other biological systems, they have not yet been studied in semen. Some examples of these are ferritin, alpha lipoic acid, L-ergothionine, ebselen, etc.

The presence of mRNA transcripts in human spermatozoa is not caused by a conserved DNA transcription during spermatogenesis, but rather a remnant of the gene expression of previous steps of spermatogenesis. Different studies have demonstrated the necessity of these transcripts for adequate sperm function and early embryo development [26].

The genetic expression in the spermatozoa of enzymes involved in ROS scavenging has also been analyzed by different groups. We have detected mRNA expression for GPX-1, GPX-4 and GR in human seminal cells, thus demonstrating the ability of these cells to synthesize the enzymes (Garridoy Meseguer, submitted). Moreover, we have also discovered the enzymatic activity in spermatozoa lysates, clearly indicating the self-protective capacity of seminal cells against ROS.

Catalase expression has only been detected in rat spermatozoa and no information is available about its presence in the humans [18]. Moreover, there is no information in the literature with respect to SOD mRNA expression in spermatozoa.

In human spermatozoa, the presence of SOD, GPX/GR and GSH activities were described by Alvarez and colleagues almost 20 years ago [27]. Among them, SOD seemed to be the most relevant enzyme in the protection of human spermatozoa from lipid peroxidation.

We firmly believe that an accurate definition of "fertility" and "infertility" in males is fundamental in order to design effective studies of the factors implicated in fertility.

Fertile men can be defined as those capable of producing an ejaculate that is sufficient for impregnating a woman with no fertility problems via sexual intercourse, during a given period of time (one year is accepted as reasonable by different authors).

In this way, infertile males are those producing ejaculates that do not result in said women conceiving, including those both with and without alterations of the basic WHO seminal parameters.

Moreover, in infertile males, the absence of disturbances in the semen and the presence of only moderate disorders indicate that some molecular aspects (namely, those leading to infertility) have been overlooked in the semen analysis.

To date, there are no molecular factors described as a standard with respect to the labeling of a semen sample as "fertile" or "infertile", or for determining if this status is permanent on time. Further complicating the definition, there is a subfertile status in which a male defined as infertile according to the above mentioned criteria is able to impregnate a woman with the help of assisted reproduction techniques.

As already explained, OS status and its markers are most certainly implicated in fertility. Physiologically, moderate amounts of ROS are necessary for a successful reproductive process and many studies have illustrated that they are important in determined phases, such as sperm capacitation, motility activation, acrosome reaction and oocyte fusion [28, 29].

Based on this knowledge, it can be hypothesized that fertile and infertile statuses are distinguished by a down-regulation of the OS defense system or an overwhelming ROS presence.

Given the difficulties in understanding the complete OS system due to its complex composition of pro-oxidative and antioxidative agents, only partial information has been provided by each publication. Moreover, contradictory results are often found. Interestingly, the study by Yeung and colleagues [29], identified no relationship between the activity of any of the enzymes analyzed in the IVF patients with their success in the respective IVF cycles. Additionally, there was no correlation of enzyme activity with the morphology or percentage of motile spermatozoa.

Our group has also studied the fertile status in terms of GPX-1, GPX-4 and GR in lisated seminal cells at mRNA and enzymatic activity levels. To this end, we compared semen samples from almost 50 infertile males whose WHO semen parameters were considered to be normal or mildly pathological with a control group of males who had already fathered healthy babies. We found a strong positive correlation between the GPx-4 activity and the percentage of normal forms and no correlation with the mRNA presence of this concrete transcript (Garrido y Meseguer submitted). This suggests a post-transcriptional control of GPx-4 function by unknown modulators.

Moreover, GPX4 activity is ten times lower in fertile than in infertile males, indicating that the OS equilibrium is somehow compromised in infertile males.

When the WHO seminal parameters were considered in terms of GPX-1, GPx-4 GR and GSH activity, GPX4 was significantly lower in males whose samples had less than 5 % normal forms in their ejaculates. GSH was also lower in the same group of patients.

Whether or not the low GPx-4 activity is a cause or a consequence of aberrant sperm morphology is yet to be experimentally determined. The dual function of GPx-4 as an active enzyme in spermatids and immature sperm and as a structural protein in mature sperm (enzymatically non-functional) described elsewhere is not confirmed in our work, since the higher the GPx-4 activity, the higher is the percentage of normal forms.

In addition, deficiencies in semen samples that are frozen and subsequently thawed are related with defects in both GPX-4 expression and activity (Garrido and Meseguer, submitted). This inability to withstand the freezing process seems to occur when samples are initially stressed in oxidation terms.

Moreover, many pathologies such as spinal cord injury and varicoceles are related with ROS increase in the ejaculate [8, 31].

Previous research has demonstrated that seminal plasma superoxide dismutase, catalase, glutathione peroxidase and sulfhydryl group levels are significantly lower in infertile patients than those in controls, thus clearly indicating their direct implication in male fertility [32].

By measuring the by-products of OS reactions, the adverse effects of OS in semen have been correlated with sperm quality in some studies.

8-hydroxydeoxyguanosine (8-OHdG) and lipid peroxides are widely used as markers to quantify OS and have been compared in the seminal plasma and spermatozoa from subfertile and fertile men. The concentrations of 8-OHdG and lipid peroxides in the seminal plasma of the subfertile group were significantly higher than those of the fertile group [33]. There were no significant differences between the values of patients with normo-zoospermia and those with asthenozoospermia. In all four fractions obtained by Percoll gradient fractionation, the lipid peroxide levels in spermatozoa recovered from subfertile males were significantly higher than those from the fertile controls.

In another attempt to bring enlightenment to this area, the ratio of ROS versus TAC (total antioxidant capacity) was evaluated in the semen sample. Infertile men with male factor or idiopathic diagnoses showed significantly lower levels than controls, and men diagnosed with male factor but who eventually initiated a successful pregnancy had significantly higher ROS-TAC scores than those who failed achieving conception [34, 35].

To date, there is increasing interest in all the factors potentially affecting male fertility, since several studies seem to point to a progressive diminution in said fertility over the last few decades.

Nevertheless, it is undeniable that there is a vast lack of knowledge regarding the molecular events that lead to sub-fertile semen.

The identification of adequate molecular markers to diagnose a semen sample, in conjunction with the design of appropriate therapeutic strategies to combat defects and deficiencies, is decisive for resolving male factor problems. Among these, OS markers are relevant factors in male fertility, since they play a significant role in sperm physiology. A fine balance of these markers could determine the reproductive success.

Although increasing efforts have been made to understand the relevance of these systems in fertility, we still have no enough data to provide solutions to determine male factor pathologies and further research is necessary.

The work was financed by grant FIS 01/0107 from the Spanish Ministerio de Sanidad y Consumo.