On methods for the detection of reactive oxygen species generation by human spermatozoa: analysis of the cellular responses to catechol oestrogen, lipid aldehyde, menadione and arachidonic acid

Authors

  • R. J. Aitken,

    Corresponding author
    • Discipline of Biological Sciences and Priority Research Centre in Reproductive Science, Faculty of Science and IT, University of Newcastle, Callaghan, NSW, Australia
    Search for more papers by this author
  • T. B. Smith,

    1. Discipline of Biological Sciences and Priority Research Centre in Reproductive Science, Faculty of Science and IT, University of Newcastle, Callaghan, NSW, Australia
    Search for more papers by this author
  • T. Lord,

    1. Discipline of Biological Sciences and Priority Research Centre in Reproductive Science, Faculty of Science and IT, University of Newcastle, Callaghan, NSW, Australia
    Search for more papers by this author
  • L. Kuczera,

    1. Discipline of Biological Sciences and Priority Research Centre in Reproductive Science, Faculty of Science and IT, University of Newcastle, Callaghan, NSW, Australia
    Search for more papers by this author
  • A. J. Koppers,

    1. Discipline of Biological Sciences and Priority Research Centre in Reproductive Science, Faculty of Science and IT, University of Newcastle, Callaghan, NSW, Australia
    Search for more papers by this author
  • N. Naumovski,

    1. Discipline of Biological Sciences and Priority Research Centre in Reproductive Science, Faculty of Science and IT, University of Newcastle, Callaghan, NSW, Australia
    Search for more papers by this author
  • H. Connaughton,

    1. Discipline of Biological Sciences and Priority Research Centre in Reproductive Science, Faculty of Science and IT, University of Newcastle, Callaghan, NSW, Australia
    Search for more papers by this author
  • M. A. Baker,

    1. Discipline of Biological Sciences and Priority Research Centre in Reproductive Science, Faculty of Science and IT, University of Newcastle, Callaghan, NSW, Australia
    Search for more papers by this author
  • G. N. De Iuliis

    1. Discipline of Biological Sciences and Priority Research Centre in Reproductive Science, Faculty of Science and IT, University of Newcastle, Callaghan, NSW, Australia
    Search for more papers by this author

Correspondence:

Robert J. Aitken, Discipline of Biological Sciences, University of Newcastle, Callaghan, NSW 2308, Australia. E-mail: John.Aitken@newcastle.edu.au

Summary

Oxidative stress is known to have a major impact on human sperm function and, as a result, there is a need to develop sensitive methods for measuring reactive oxygen species (ROS) generation by these cells. A variety of techniques have been developed for this purpose including chemiluminescence (luminol and lucigenin), flow cytometry (MitoSOX Red, dihydroethidium, 4,5-diaminofluorescein diacetate and 2′,7′-dichlorodihydrofluorescein diacetate) and spectrophotometry (nitroblue tetrazolium). The relative sensitivity of these assays and their comparative ability to detect ROS generated in different subcellular compartments of human spermatozoa, have not previously been investigated. To address this issue, we have compared the performance of these assays when ROS generation was triggered with a variety of reagents including 2-hydroxyestradiol, menadione, 4-hydroxynonenal and arachidonic acid. The results revealed that menadione predominantly induced release of ROS into the extracellular space where these metabolites could be readily detected by luminol-peroxidase and, to a lesser extent, 2′,7′-dichlorodihydrofluorescein. However, such sensitivity to extracellular ROS meant that these assays were particularly vulnerable to interference by leucocytes. The remaining reagents predominantly elicited ROS generation by the sperm mitochondria and could be optimally detected by MitoSOX Red and DHE. Examination of spontaneous ROS generation by defective human spermatozoa revealed that MitoSOX Red was the most effective indicator of oxidative stress, thereby emphasizing the general importance of mitochondrial dysregulation in the aetiology of defective sperm function.

Introduction

The generation of reactive oxygen species (ROS) by human spermatozoa is thought to make a significant contribution to the aetiology of defective sperm function (Jones et al., 1979; Aitken & Clarkson, 1987; Alvarez et al., 1987; Aitken & Curry, 2011; Gong et al., 2012). The measurement of ROS generation by human spermatozoa is therefore an important investigative tool in the diagnosis of male infertility (Aitken et al., 2012a). A majority of the early studies on ROS production by human spermatozoa were conducted using luminol- or lucigenin-dependent chemiluminescence. Although such techniques are sensitive and apparently predictive, they suffer from the disadvantage that they cannot be standardized because the response characteristics of the photomultipliers employed in such systems vary from instrument to instrument and their read out is in arbitrary units (Aitken et al., 2004). In addition, such chemiluminescence assays can be heavily influenced by the presence of leucocytes, which are many times more active than spermatozoa in the generation of ROS (Aitken & West, 1990; Aitken et al., 1995; Henkel et al., 1997; Whittington & Ford, 1999). A solution to this problem came with the use of flow cytometry in combination with dihydroethidium (DHE) to specifically focus on ROS production by spermatozoa (De Iuliis et al., 2006). Chemical detection of the fluorochrome, 2-hydroxyethidium, in these incubations confirmed that the major oxygen metabolite detected in this assay was the superoxide anion (O2ֹֿ). It was originally proposed that the O2ֹֿ detected by DHE originated from a non-mitochondrial source because its generation could not be disrupted by mitochondrial inhibitors such as rotenone and CCCP (carbonyl cyanide m-chlorophenylhydrazone) (De Iuliis et al., 2006). However, this hypothesis was refuted when a derivative of DHE was subsequently developed, MitoSOX™ Red (MSR), which contained a positively charged triphosphonium cation that allowed the probe to concentrate in the mitochondrial matrix. Using this probe, a significant proportion of the O2ֹֿ generated by human spermatozoa was shown, in fact, to originate from the mitochondria (Koppers et al., 2008). An additional flow cytometry probe, dichlorodihydrofluorescein (H2DCFDA), has also been employed in recent studies to focus attention on the generation of powerful oxidants such as hydrogen peroxide (H2O2) and peroxynitrite (ONOO−) by human spermatozoa, rather than O2ֹֿ (Guthrie & Welch, 2006). Given this sudden availability of methods for the diagnostic analysis of ROS production by human spermatozoa, this is an ideal time to compare their performance characteristics in the presence of reagents that stimulate the redox activity of these cells. This study has addressed this question employing the following compounds:

2-Hydroxyestradiol (2OHE2) – Catechol oestrogens are major products of oestrogen metabolism in the testes. These steroid metabolites are known to have disruptive effects on the germinal and epididymal epithelia as well as the male germ line, inhibiting soluble adenylyl cyclase, cross linking sperm chromatin and promoting ROS generation (Braun, 1990; Seegers et al., 1991a,b; Bennetts et al., 2008). Recent studies in breast tissue, where catechol oestrogens are thought to be involved in the aetiology of cancer (Okoh et al., 2011), have suggested that these compounds stimulate H2O2 production from the mitochondria (Felty et al., 2005), making it a particularly good reagent for assessing methods purported to detect mitochondrial ROS production in spermatozoa.

Arachidonic acid (AA) – Cis-unsaturated fatty acids such as arachidonic acid are known to stimulate O2ֹֿ generation by human sperm mitochondria via mechanisms that can be inhibited by the flavoprotein inhibitor, diphenylene iodonium and appear to be dependent upon the amphiphilic properties of these molecules (Aitken et al., 2006; Koppers et al., 2010).

4-hydroxynonenal (4HNE) – Lipid aldehydes such as 4HNE are by-products of lipid peroxidation that have been shown to stimulate mitochondrial ROS generation in human spermatozoa via mechanisms that involve the alkylation of succinate dehydrogenase (Aitken et al., 2012b).

Menadione – Quinones such as menadione have been shown to redox cycle in human spermatozoa under the influence of quinone oxidoreductase, generating significant oxidative stress in the process (Hughes et al., 2009; Mitchell et al., 2011).

Employing these reagents, we have systematically examined the ability of a variety of assays, employing flow cytometry, luminometry and spectrophotometry as core methodologies, to detect the ROS generated by human spermatozoa in different subcellular compartments (see Fig. 1 for an overview). We have then examined the ability of these assays to detect the elevated levels of ROS generation associated with defective sperm function. The results have significant implications for our interpretation of these assays and for their diagnostic potential in cases of male infertility.

Figure 1.

Schematic representation of the study design. In this analysis, we have used four different stimuli to provoke ROS generation by human spermatozoa at different subcellular sites: 2-hydroxyestradiol (ROS primarily released into the mitochondrial matrix), arachidonic acid and 4-hydroxyalkenal (ROS primarily released into the mitochondrial intramembranous space) and menadione (ROS generated at or close to plasma membrane as a result of quinone oxidoreductase activity). The ROS generated in response to these various reagents have been monitored with a variety of diagnostic assays including MitoSOX Red (MSR), dihydroethidium (DHE), dichlorodihydrofluorescein diacetate (H2DCFDA), nitroblue tetrazolium (NBT) luminol-peroxidase and lucigenin. We have also compared the ability of the four stimulatory reagents employed in this study to induce oxidative damage to sperm DNA, as monitored by 8-hydroxy-2′-deoxyguanosine (8OHdG) formation. Finally, we have assessed the ability of these assays (along with 4,5-diaminofluorescein diacetate [DAF-DA], a probe for nitric oxide generation) to detect the increased oxidative stress associated with spermatozoa recovered from the low-density region of Percoll gradients.

Materials and methods

Sperm preparation

This research was based on a cohort of unselected healthy donors to our reproductive research program, a majority of whom were normozoospermic students of unknown fertility status. Scientific use of these samples for research purposes was approved by our Institutional Human Ethics Committee and the State Minister for Health. After at least 48 h abstinence, semen samples were produced by masturbation and collected into sterile sample containers, which were delivered to the laboratory within 1 h of ejaculation. Purification of human spermatozoa was achieved using 44 and 88% discontinuous Percoll (GE Healthcare, Castle Hill, Australia) centrifugation gradients, as described (Mitchell et al., 2011). After 30 min centrifugation at 500 g, purified spermatozoa were recovered from the base of the 88% Percoll fraction and washed with Biggers, Whitten and Whittingham medium containing 1 mg/mL polyvinyl alcohol instead of bovine serum albumin (BWW; Biggers et al., 1971). These cells were then pelleted by centrifugation at 500 g for a further 15 min and finally resuspended at a concentration of 6 × 106 cell/mL in BWW. Where indicated, low-density Percoll fractions were also prepared using exactly the same procedures. Motility was determined using phase-contrast optics while vitality was assessed using the eosin exclusion test (Aitken et al., 2010).

Leucocyte removal

Where indicated, all residual traces of leucocyte contamination in the sperm suspensions were removed using magnetic beads (Dynabeads, Dynal, Oslo, Norway) coated with a monoclonal antibody against the common leucocyte antigen, CD45 (Invitrogen, Carlsbad, CA, USA). Following Percoll isolation, 5 × 106 cells in 100 μL BWW were added to pre-washed antibody-bound Dynabeads and then placed on a rotator for 30 min. Following incubation, each sample was placed in a magnetic holder to separate leucocyte-bound Dynabeads from purified sperm cells in BWW. Luminol–HRP-mediated chemiluminescence was then used to confirm the removal of leucocytes from each sperm suspension; for this purpose, 20 μL of zymosan, opsonized with autologous serum, was added to each 400 μL sample, 5 min from the beginning of the luminometry run (Aitken et al., 1996).

Exposure to reagents designed to increase cellular ROS generation

The isolated spermatozoa were exposed to a variety of reagents at 37 °C to induce the cellular generation of ROS. The dose and duration of exposure for each reagent was selected to give a dynamic range of responses that would facilitate comparison of the various assays employed in this study. Arachidonic acid (0–50 μm), menadione (0–50 μm) and 2OHE2 (0–500 μm) all used a 15 min duration of exposure while for 4HNE (0–400 μm) the exposure time was increased to 1 h. Vehicle controls were included in each run using the amount of vehicle included at the highest dose of reagent. At the end of the exposure period, the cells were pelleted by centrifugation at 600 g and assessed using the various ROS detection protocols described below.

Flow cytometry assays

MitoSOX™ Red and DHE assays were performed by flow cytometry incorporating SYTOX G green as a vitality stain, as described previously (Koppers et al., 2008, 2011); 2.5 × 106 spermatozoa were used per sample and these probes were incubated with the spermatozoa for 15 min at 37 °C. Non sperm-specific events were gated out and 10 000 cells were examined. The results were expressed as the percentage of viable cells staining positively with each probe.

Mitochondrial membrane potential was assessed with JC1 in combination with propidium iodide, using carbonyl cyanide m-chlorophenylhydrazone (CCCP) at a final concentration of 10 μm to create a negative control (Koppers et al., 2011); results were expressed as the percentage of live cells exhibiting a high membrane potential.

For the H2DCFDA assay, stock concentrations of this probe (10 mm) were diluted in DMSO and stored in aliquots for single use under nitrogen at −20 °C in the dark. Immediately before use, stock H2DCFDA was diluted to a final concentration of 10 μm in BWW containing 2 × 106 spermatozoa in 200 μL BWW. The cell suspensions were then incubated for 1 h at 37 °C. Propidium iodide (PI) was used in conjunction with H2DCFDA as a vitality stain and was added to the cells at a concentration of 7 μm immediately prior to flow cytometry analysis. Results are presented as the percentage of live cells exhibiting a response.

The cellular generation of nitric oxide (NO) was determined by flow cytometry using DAF-DA (4,5-diaminofluorescein diacetate) as the probe. DAF-DA was initially diluted to a stock concentration of 1 mm in DMSO and frozen in aliquots at −20 °C. On the day of analysis, further dilution was achieved using BWW to generate a final concentration of 10 μm. Spermatozoa were incubated with this probe for 30 min at 37 °C. Propidium iodide was used in conjunction with DAF-DA as a vitality stain and was added to the cells at a concentration of 7 μm immediately prior to flow cytometry analysis.

The presence of the oxidized DNA base adduct, 8-hydroxy-2′-deoxyguanosine (8OHdG), was also assessed by flow cytometry using the OxyDNA Assay (Calbiochem, CA, USA) in combination with LIVE/DEAD fixable dead cell stain (Molecular Probes, OR, USA) as described by Koppers et al. (2011).

All flow cytometry analyses reported in this study was conducted on a FACS-Calibur flow cytometer (Becton Dickinson, CA, USA) with a 488 nm argon laser. Forward scatter and side scatter measurements were taken to generate a scatter plot, which was used to gate for sperm cells only, excluding any larger contaminating cells. All data were acquired and analysed using CellQuest Pro software (Becton Dickinson) and a total of 10 000 events were collected per sample.

Nitroblue tetrazolium assay

The NBT assay for oxidative stress was adapted from Tunc et al. (2009). For the assay, 5 × 106 cells in 200 μL BWW were incubated at 37 °C with ROS-generating reagents for the periods of time stipulated above (15 min for all treatments other than 4HNE for which a 1 h incubation period was used). The cells were then incubated for 1 h in 1.25 mm NBT in BWW at 37 °C in the dark. They were then washed three times, made up in 110 μL with BWW and then 110 μL of KOH (4 mm in DMSO) was added to lyse the cells. After vortexing, 200 μL aliquots were removed to a microtitre plate (FLUOstar Optima; BMG Labtechnologies, Durham, NC, USA) and the absorbance read on an ELISA plate reader at 630 nm. Each assay was run in triplicate.

Chemiluminescence

For lucigenin-dependent chemiluminescence, 4 × 106 spermatozoa in 400 μL BWW were supplemented with 4 μL lucigenin (25 mm) and the samples were then run for 5 min at 37 °C in a Berthold AutoLumat luminometer LB-953 (Berthold, Bad Wildbad, Germany) to stabilize the chemiluminescent system. A given ROS-inducing agent was then added and the resulting chemiluminescence was monitored for 30 min and the results expressed as integrated counts. The luminol-peroxidase assay was performed on 5 × 106 spermatozoa in 400 μL BWW and was essentially the same as the lucigenin assay except that the cells were supplemented with 4 μL luminol (25 mm) and 8 μL horseradish peroxidase (HRP, 11.52 U/mL). For both chemiluminescence assays, media blanks were run for every treatment to ensure that the signals recorded were not because of the spontaneous activation of the probe. The values obtained in these media-only control incubations were subtracted from those obtained in the presence of spermatozoa.

Sperm movement characteristics

Once the cells had been incubated with ROS-inducing reagents their movement characteristics were assessed using a Hamilton-Thorn motility analyzer (HTMA IVOS; Hamilton-Thorn Research, Danvers, MA, USA). The settings for human spermatozoa were: negative phase-contrast optics, recording rate 60 frames/sec, minimum contrast 80, minimum cell size 3 pixels, low size gate 1.0, high size gate 2.9, low intensity gate 0.6, high intensity gate 1.4, non-motile head size 6, non-motile head intensity 160, progressive VAP (average path velocity) threshold value, 25 μm/sec, slow cells VAP cut off, 5 μm/sec, slow cells VSL (straight line velocity) cut off, 11 μm/sec and threshold STR (straightness) >80%. Progressive cells were those exhibiting a VAP of >25 μm/sec and a STR of >80%.

Coupling menadione to BSA

One milligram of bovine serum albumin (BSA) (Sigma Aldrich, Castle Hill, NSW, Australia) was reduced by 0.8 mol/equiv Tris-(2-carboxyethyl)phosphine hydrochloride (TCEP) (Invitrogen, Mulgrave, VIC, Australia) under nitrogen (Burns et al., 1991). Menadione was prepared as a stock solution (100 mm) in DMSO (Sigma Aldrich) and then diluted (6.61 mm) in deionized water (Sigma Aldrich). Menadione was then gradually added to the reduced BSA at 37 °C under nitrogen over a 2 h period. The reaction product was then purified by solid phase extraction, using 30% acetonitrile to elute the conjugated protein after equilibration in 0.1% formic acid. The eluent was then dried using a centrifugal vacuum evaporator set at 45 °C. The dried product was then resuspended in water (100 μL).

Statistical analysis

All experiments were replicated at least three times on independent samples and the results analysed by one- and two-way anova using the SuperANOVA program (Abacus Concepts Inc, CA, USA) on a MacIntosh G4 Powerbook computer; post hoc comparison of group means was by Fisher's PLSD (Protected Least Significant Difference). Paired comparisons were conducted using a paired t-test using the Statview program (Abacus Concepts Inc). Differences with a p value of <0.05% were regarded as significant.

Results

Responsiveness of various detection systems following the induction of ROS formation in vitro

2-Hydroxyestradiol – Addition of 2OHE2 to populations of human spermatozoa led to a dramatic dose-dependent loss of motility and progressive motility following 15 min exposure (< 0.001; Fig. 2A). The loss of progressive motility was evident at a dose of 31.25 μm and by 125 μm, overall motility was reduced to less than 10%. Importantly, this loss of motility was achieved without any change of cell viability within the 15 min time course of this experiment (Fig. 2B). The loss of sperm function was exactly mirrored by the dose-dependent induction of mitochondrial O2ֹֿ generation detectable by MSR, which was significant at a dose of 12.5 μm and at 500 μm involved more than 90% of the sperm population (Fig. 2C). The redox probe, DHE, was also able to detect this O2ֹֿ generation, generating a dose response curve that paralleled the MSR signal (= 0.993).

Figure 2.

Analysis of the responses of human spermatozoa to a 15 min exposure to 2OHE2. (A) Changes in motility (closed bars) and progressive motility (open bars). (B) Maintenance of cell vitality following exposure as determined by the SYTOX Green assay in the course of assessing MSR (closed bars) or DHE (open bars) responsiveness. (C) Significant dose-dependent change in MSR signal. (D) Significant dose-dependent change in DHE signal. (E) Significant dose-dependent change in H2DCFDA signal although the percentage of responsive cells is less than half of that observed with MSR or DHE. (F) NBT reduction was an insensitive method for detecting mitochondrial ROS generation in response to 2OHE2, only producing a significant response at the highest does tested. (G) Similarly, lucigenin-dependent chemiluminescence only responded at the highest dose tested. (H) Luminol–HRP-dependent chemiluminescence was not responsive to the mitochondrial ROS generated in the presence of 2OHE2. Data corresponds to mean values ±SEM;= 3 independent samples for each panel. *< 0.05, **< 0.01. ***< 0.001 for differences compared with control.

2OHE2 also stimulated an increase in cellular oxidant production as detected by H2DCFDA (Fig. 2E), although a significant response was only observed in around 30% of the sperm population, compared with the near universal responses elicited with 2OHE2 when DHE and MSR were used as the probes (Fig. 2C,D). NBT and lucigenin reduction also provided insensitive measures of the ROS generation stimulated by 2OHE2 with a significant change only being observed at the highest doses of compound assessed (Fig. 2F,G). Nevertheless, the correlations between MSR and NBT (= 0.661; < 0.01) as well as between MSR and lucigenin-dependent chemiluminescence (= 0.787; < 0.001) were statistically significant. By contrast, the luminol–HRP system actually exhibited a fall in chemiluminescence in the presence of 2OHE2, possibly reflecting the intrinsic antioxidant properties of this molecule (Ruiz-Larrea et al., 1994; Fig. 2H). Furthermore, luminol-dependent chemiluminescence exhibited no significant correlation with the MSR or DHE signals in the presence of 2OHE2.

These results clearly indicated that the various assays used to detect ROS generation by human spermatozoa responded very differently when 2OHE2 was used to induce free radical generation by the sperm mitochondria and cytoplasm. In light of these results, the same probes were examined to determine how they would react to ROS generated close to the plasma membrane by the redox cycling quinone, menadione.

Menadione – Addition of the redox cycling quinone, menadione, to populations of human spermatozoa induced a significant decline in percentage sperm motility (< 0.01) and progressive sperm motility (< 0.001) following 15 min exposure (Fig. 3A), without any change in sperm vitality (Fig. 3B). This loss of sperm movement was associated with a weak statistically significant increase in O2ֹֿ levels in the mitochondrial compartment as detected by MSR (< 0.01; Fig. 3C) and in the cells as a whole, as detected by DHE (= 0.05; Fig. 3D). When H2DCFDA was used as the probe, a more statistically significant change was observed (< 0.001), which was clearly evident at doses as low as 12.5 μm, suggesting an efficient conversion of the O2ֹֿ generated in response to menadione to H2O2 or ONOO- (Fig. 3E). Despite a high background value for NBT reduction, this probe also effectively detected the ROS response to menadione (< 0.05), generating values that were dose-dependent and significantly elevated at 12.5 μm (Fig. 3F). Lucigenin-dependent chemiluminescence similarly generated signals that increased with the level of menadione exposure, but the variance associated with these measurements resulted in a lack of statistical significance (Fig. 3G). On the contrary, luminol–HRP-dependent chemiluminescence was extremely responsive to the ROS response elicited by menadione, generating a highly significant change (< 0.001), which was clearly dose-dependent and gave group mean values that were statistically different from control incubations at doses of 12.5 μm and above (Fig. 3H). While all the ROS detection assays showed an increase with menadione dose, the read out of the DHE and MSR were not significantly correlated as had been observed with 2OHE2. By contrast, the luminol signal was highly correlated with NBT in the presence of menadione (= 0.752; < 0.001) whereas these ROS assays had shown no correlation with 2OHE2.

Figure 3.

Analysis of the responses of human spermatozoa to a 15 min exposure to menadione. (A) Changes in motility (closed bars) and progressive motility (open bars); = 4. (B) Maintenance of cell vitality following menadione exposure as determined by the SYTOX Green assay in the course of assessing MSR (closed bars) or DHE (open bars) responsiveness; = 4. (C) The dose-dependent change in MSR was significant overall (< 0.01), although comparison of group means revealed that only the highest dose was significantly different from the control; = 9. (D) Significant dose-dependent changes in the DHE signal, although the responses were variable and only significantly different from control values at the highest dose tested; = 9. (E) Significant dose-dependent change in H2DCFDA signal; = 3. (F) Significant dose-dependent NBT response to menadione; = 4. (G) Lucigenin-dependent chemiluminescence exhibited a dose-dependent trend, but the variability in this data set precluded attainment of statistical significance; = 7. (H) Luminol–HRP-dependent chemiluminescence revealed a highly significant dose-dependent change in redox activity following exposure of human spermatozoa to menadione; = 3. Data corresponds to mean values ±SEM. *< 0.05, **< 0.01. ***< 0.001 for differences compared with control.

These data suggested that the ROS signals generated in the presence of menadione and 2OHE2 were very different. Whereas the latter appeared to generate primarily a mitochondrial signal, with little escape of H2O2 into the extracellular space, menadione appeared to generate primarily an extracellular H2O2 signal with highly variable penetration into the mitochondria. The contention that menadione primarily elicited an extracellular H2O2 signal was supported by the fact that the luminol-HRP response to menadione was completely eliminated following the addition of ROS scavenging enzymes such as SOD and catalase, to the extracellular space (Fig. 4A,B). Furthermore, the luminol response to menadione was not influenced by the presence of several inhibitors of mitochondrial function [rotenone (10 μm), tri-thenoyl fluoroacetone (10 μm), antimycin A (10 μm) KCN (1 μm), CCCP (10 μm); Fig. 4C]. To test the possibility that the menadione might be redox cycled at the sperm surface this quinone was covalently linked to bovine serum albumin. This protein-menadione construct was found to be extremely active in generating ROS signals that were detectable with luminol-HRP (Fig. 4D), suggesting that this quinone was indeed being redox cycled at, or very near, the surface of the cell.

Figure 4.

Evidence that the redox cycling activity induced by menadione involves the release of ROS into the extracellular space. (A) The luminol-HRP generated in response to menadione (Men) was completely quenched by the presence of extracellular ROS scavenging enzymes such as catalase (Cat) and SOD;= 3; response monitored for 2 h. (B) A representative chemiluminescence trace demonstrating the effectiveness of the scavengers (Scav) SOD and catalase in disrupting the redox response induced by menadione. (C) Inability of a variety of mitochondrial inhibitors to interfere with the chemiluminescence response precipitated by menadione. The inhibitors used were rotenone (Rot; 10 μm), Tri-thenoyl fluoroacetone (TTFA; 10 μm), antimycin (Ant; 10 μm), potassium cyanide (KCN; 1 μm) and CCCP (10 μm). Reponses monitored over a 2 h period; = 3. (D) Demonstration that menadione covalently linked to BSA is still able to redox cycle in the presence of spermatozoa, indicating that the cycles of quinone reduction and oxidation must be occurring at or near to the surface of the cell. Data corresponds to mean values ±SEM or representative continuous traces.

Arachidonic acid – AA has been established as an extremely powerful activator of mitochondrial ROS in human spermatozoa (Aitken et al., 2006; Koppers et al., 2010). In contrast with the other activators of ROS generation examined in this study, a brief 15 min exposure to AA had no impact on any aspect of sperm motility and similarly did not influence sperm vitality (Fig. 5A,B). To be certain that AA was generating ROS and creating the anticipated level of oxidative stress, measurements of 8OHdG were made during the 24 h pursuant to exposure. This analysis revealed that AA was probably the most active ROS-generating reagent assessed in this study because, within 1 h of exposure, AA-treated cells showed a significant increase in 8OHdG formation that was not exhibited by any of the other compounds assessed (Fig. 6A). Furthermore, the ability of AA to create oxidative stress in human spermatozoa was sustained over the following 24 h period (Fig. 6B,C). The generation of mitochondrial ROS was also indicated by a highly significant (< 0.001) and dose-dependent stimulation of MSR fluorescence, which was significantly elevated above control levels at a dose of 6.25 μm (Fig. 5C). A less significant (< 0.01), but dose-dependent increase in the DHE signal was also observed with AA, which reached significance at 25 μm (Fig. 5D). None of the other probes responded in a dose-dependent manner to AA exposure with the exception of lucigenin (Fig. 5E–H). However, even with lucigenin, the variability in chemiluminescence meant that no statistically significant dose-dependent effect was observed.

Figure 5.

Analysis of the responses of human spermatozoa to a 15 min exposure to AA. (A) No change in motility (closed bars) or progressive motility (open bars) was observed following exposure to free unesterified AA;= 4. (B) Cell vitality was also unchanged following AA exposure as determined by the SYTOX Green assay in the course of assessing MSR (closed bars) or DHE (open bars) responsiveness; = 4. (C) Highly significant dose-dependent change in MSR signal elicited by AA;= 4. (D) Significant dose-dependent change in DHE signal elicited by AA;= 4. (E) No dose-dependent change in the H2DCFDA response to AA was observed, moreover the number of responsive cells was a fraction of that observed with MSR or DHE;= 4. (F) No significant change in NBT reduction was induced on exposure to AA;= 5. (G) Lucigenin-dependent chemiluminescence exhibited a dose-dependent trend in response to AA, but the data set lacked statistical significance; = 4. (H) Luminol–HRP-dependent chemiluminescence was not responsive to the mitochondrial ROS generated in the presence of AA;= 4. Data corresponds to mean values ±SEM. *< 0.05, **< 0.01. ***< 0.001 for differences compared with control.

Figure 6.

Analysis of oxidative stress induced by activators of ROS generation as reflected in the generation of the oxidized DNA base adduct, 8OHdG. Measurements of oxidized DNA base damage were recorded after (A) 1 h, (B) 4 h and (C) 24 h exposure to ROS-inducing agents for 15 min [AA, 2OHE2 and menadione (Men)] or 1 h (4HNE). The data clearly indicate that a brief exposure to AA rapidly creates high levels of oxidative stress, as does a longer exposure to 4HNE. Both these reagents trigger mitochondrial ROS generation at the inner mitochondrial membrane whereas reagents that cause little overt oxidative DNA damage generate ROS in the mitochondrial matrix (2OHE2) or near the plasmalemma (Men). Data corresponds to mean values ±SEM;= 3 independent samples. **< 0.01. ***< 0.001 for differences compared with control.

With this AA data set, the DHE and MSR signals were significantly correlated with each other, but not any of the other methods of detecting ROS employed in this study. Furthermore, the relationship between these probes more closely fitted a logarithmic rather than a linear function (r = 0.857; < 0.001), reinforcing the notion that MSR was more responsive to AA than DHE and in keeping with the mitochondrial origin of the ROS signal.

4-hydroxynonenal – A 1 h exposure to 4HNE was sufficient to induce highly significant (< 0.001) declines in overall motility as well as progressive motility in the absence of any concomitant change in sperm vitality (Fig. 7A,B). The ability of 4HNE to activate mitochondrial ROS generation (Aitken et al., 2012a,b) was evidenced by the highly significant dose-dependent changes observed using MSR and DHE as probes (< 0.001; Fig. 7C,D). Moreover, the analysis of 8OHdG formation revealed that 4HNE was second only to AA in generating high levels of oxidative stress in these cells (Fig. 6B,C). As with the other activators of mitochondrial ROS (2OHE2 and AA), the MSR and DHE responses were highly correlated with each other (= 0.884; Fig 2F), but not with any other measurement of ROS generation. Significant changes were also observed with H2DCFDA as the probe (Fig. 7E), but not with NBT or either of the chemiluminescence probes used in this analysis (Fig. 7F–H).

Figure 7.

Analysis of the responses of human spermatozoa to a 1 h exposure to 4HNE. (A) 4HNE induced a highly significant dose-dependent decline in motility (closed bars) or progressive motility (open bars). (B) Cell vitality was unchanged following 4HNE exposure as determined by the SYTOX Green assay in the course of assessing MSR (closed bars) or DHE (open bars) responsiveness. (C) A significant dose-dependent change in MSR signal was elicited by 4HNE. (D) A parallel dose-dependent change in DHE signal was also observed. (E) H2DCFDA also revealed a dose-dependent change in redox activity although the number of responsive cells was half that observed with MSR or DHE. (F) No significant change in NBT reduction was induced by 4HNE. (G) No significant change in lucigenin-dependent chemiluminescence was induced by 4HNE. (H) Luminol–HRP-dependent chemiluminescence was also not responsive to the mitochondrial ROS generated in the presence of 4HNE. Data corresponds to mean values ±SEM;= 3 independent samples for all panels. *< 0.05, **< 0.01. ***< 0.001 for differences compared with control.

Detection of spontaneous ROS generation by human spermatozoa

To compare the various assays for their ability to detect spontaneous ROS generation by human spermatozoa, analyses were conducted on sperm populations recovered from the high- (good quality) and low- (poor quality) density regions of discontinuous Percoll gradients. This analysis revealed that these populations of cells could be discriminated on the basis of MSR, DHE, H2DCFDA and luminol assays, all of which detected the significantly elevated levels of ROS generation characteristic of low-density Percoll fractions (Fig. 8A,B). On the other hand, a probe for nitric oxide, DAF-DA, generated consistently high signals that did not differentiate between the high and low-density Percoll fractions (Fig. 8A).

Figure 8.

Analysis of spontaneous ROS generation by spermatozoa isolated from the high- and low-density regions of Percoll gradients. (A) All the flow cytometry probes analysed in this study (MSR, DHE, H2DCFDA) successfully detected the difference in redox activity of high- and low-density sperm populations. In contrast, DAF-DA generated a consistently high signal in both Percoll fractions; = 13. (B) Although luminol–HRP-dependent chemiluminescence could also differentiate between these two sperm populations, such discrimination was completely lost when leucocytes were removed using CD45-coated Dynabeads; = 6. (C) Impact of Dynabead treatment on the ROS signals generated by low-density Percoll sperm populations. Such treatment significantly increased the MSR and DHE signals, but reduced the H2DCFDA signal to very low levels, while the DAF-DA signal was unchanged; = 11. (D) Following the removal of contaminating leucocytes all the probes except for DAF-DA could discriminate the high- and low-density Percoll populations, the most effective probe being MSR, in keeping with the key role that mitochondria play in the aetiology of defective sperm function; = 12. (E) H2DCFDA could not detect the significant increase in ROS generation induced by exposing human spermatozoa to inhibitors of the electron transport chain such as rotenone (10 μm) or antimycin (10 μm) which are known to trigger levels of ROS generation in human spermatozoa that can be detected by both MSR and DHE;= 7 (Koppers et al., 2008).

Because low-density Percoll sperm populations are known to be contaminated by leucocytes (Aitken & West, 1990), which confound the analysis of ROS generation, Percoll-prepared sperm populations were subjected to treatment with Dynabeads coated with an anti-CD45-coated antibody to remove these cells (Aitken et al., 1996). Such treatment reduced the luminol-HRP signals generated by both high- and low-density sperm populations to background levels (Fig. 8B). Leucocyte removal also diminished the H2DCFDA signal, but actually enhanced the MSR and DHE signals detected in the low-density Percoll fractions, while having no effect on DAF-DA reactivity (Fig. 8C). Following removal of all detectable traces of contaminating leucocytes, MSR, DHE and H2DCFDA could still significantly discriminate the difference in spontaneous redox activity between the high- and low-density Percoll populations (Fig. 8D). However, the MSR signal was significantly greater than those observed in the presence of DHE and H2DCFDA in both the low- (< 0.01) and high- (< 0.001) density Percoll fractions (Fig. 8D). The particular sensitivity of MSR for detecting oxidative stress in human sperm populations is probably related to the fact that this probe is designed to detect changes in mitochondrial ROS and it is the latter which largely defines the redox status of these cells. Probes such as H2DCFDA are clearly redox sensitive, but are relatively insensitive to the changes in mitochondrial ROS generation triggered by reagents such as antimycin and rotenone (Fig. 8E), whereas such treatments are known to profoundly influence the activity detected with MSR (Koppers et al., 2008).

Discussion

This study has shed light on the responsiveness of the current range of assays used to detect ROS generation by human spermatozoa and also increased our understanding of how specific reagents increase oxidative stress in the male germ line. In the following discussion, we shall begin by considering the assays themselves and then review the mechanisms by which different classes of reagent elicit a free radical response from human spermatozoa.

Comparison of probes for assessing ROS generation by human spermatozoa

MSR and DHE – These probes are both reduction products of ethidium bromide, but MSR has been chemically modified to give the molecule a positive charge that results in its concentration in the mitochondrial matrix. From a diagnostic perspective, it is important to emphasize that commercial preparations of these reagents are contaminated with small amounts of the parent ethidium compound, which can result in the spurious staining of non-viable cells. It is therefore imperative that these assays are run with vitality markers such as Sytox green, the emission characteristics of which permit simultaneous assessment of cell vitality and ROS-generating activity by flow cytometry (Koppers et al., 2008, 2011). In this study, all the reagents which acted primarily on the mitochondria to stimulate ROS production (2OHE2, AA, 4HNE), generated MSR and DHE signals within the live cell population that were highly correlated with each other. In general, the relationship was linear, but in the case of AA in particular, the MSR signal was the more responsive (Fig. 5C,D). As the physiological induction of ROS generation in defective human spermatozoa is thought to involve the ability of free, unesterified unsaturated fatty acids such as AA to trigger electron leakage from the mitochondrial electron transport chain, this result emphasizes the particular clinical utility of the MSR assay (Koppers et al., 2010). This conclusion was reinforced by the analysis of spontaneous redox activity in defective human spermatozoa recovered from the low-density region of Percoll gradients. Although H2DCFDA, DHE and MSR were all capable of detecting the enhanced redox activity associated with defective, low-density human spermatozoa, the MSR signals were significantly more intense than those observed with any other probe (Fig. 8D).

H2DCFDA – This probe becomes fluorescent on oxidation and is purported to measure cellular H2O2 production. In reality, this oxidant has no effect on H2DCFDA fluorescence unless it is accompanied by peroxidase activity, which in a highly compartmentalized, cytoplasm-deficient cell such as a spermatozoon, may be a rate limiting factor in the genesis of activity. Other ROS such as ONOO- and the hydroxyl radical can also directly oxidize this probe and might make significant contributions to the positive signals observed in defective human spermatozoa (Myhre et al., 2003; Mahfouz et al., 2010). In practice, the signals generated by H2DCFDA were most effective when the oxidants were being generated in the cytoplasm at or near the plasma membrane, as when menadione was used to enhance cellular redox activity (Fig. 3E). Leucocytes were also found to have a confounding effect on the H2DCFDA signal (Fig. 8C). When these cellular contaminants were removed, H2DCFDA could only detect an intense redox signal in around 3% of defective low-density human spermatozoa, whereas MSR detected excessive ROS generation in 30% of such cells (Fig. 8D). Furthermore, when mitochondrial ROS was triggered by incubation with rotenone or antimycin, no change was observed in the H2DCFDA signal (Fig. 8E) whereas MSR is known to detect the increased ROS generated under such circumstances (Koppers et al., 2008).

Luminol and lucigenin – Like H2DCFDA, these chemiluminescent probes appeared to be relatively insensitive to the mitochondrial ROS generated on exposure to 4HNE (Fig. 7G,H), AA (Fig. 5G,H) or 2OHE2 (Fig. 2G,H). In contrast, luminol-HRP was, by far, the most sensitive probe for detecting the ROS generated in response to menadione, where a significant proportion of the ROS are released to the outside of the cell. In the extracellular space, HRP is able to catalyse the oxidation of luminol to generate a luminol radical (Lֹ). The latter interacts with ground state oxygen to produce O2ֹֿ which then participates in the oxygenation of Lֹ to create an unstable endoperoxide, which breaks down with the release of light (Aitken et al., 2004, 2012a). Again, like H2DCFDA, this method is also highly vulnerable to the confounding influence of leucocytes, which are clearly capable of releasing ROS into the extracellular space (Aitken & West, 1990; Hipler et al., 1998; Aitken et al., 2004). When these cells were removed from the sperm suspensions with anti-CD45-coated Dynabeads, the ability of the luminol system to discriminate between spermatozoa from high- and low-density Percoll fractions was lost (Fig. 8B).

Lucigenin could detect the mitochondrial ROS generated in response to 2OHE2, but only at the highest dose of reagent used. It also detected the increased ROS activity generated in response to menadione and AA treatment, but the data were so variable that statistical significance was not achieved. Lucigenin appeared to be insensitive to the increased redox activity associated with 4HNE exposure. The problem with this probe is that the chemistry of its chemiluminescence is very complex. Activation of the probe requires a one-electron reduction, rather than the one-electron oxidation associated with luminol-dependent chemiluminescence (Aitken et al., 2012a). This one-electron reduction creates a radical (LucH+ֹ) from lucigenin (Luc2+) that rapidly gives up its electron to ground state oxygen to create O2ֹֿ and return the lucigenin to its parent state. The LucH+ֹ generated from the one-electron reduction of lucigenin then combines with O2ֹֿ to produce the dioxetane that, in turn, decomposes with the generation of light. The O2ֹֿ involved in the last reaction could come from an independent cellular source, such as the mitochondria, in which case the chemiluminescence recorded would reflect the generation of O2ֹֿ. However, an unknown proportion of the O2ֹֿ involved in this reaction is an artefact created by the reaction between LucH+ֹ and ground state oxygen. Chemiluminescence created by the cellular generation of O2ֹֿ or the redox cycling of lucigenin cannot be readily distinguished, as both sources of ROS are suppressible by SOD. Furthermore, we have clearly demonstrated that lucigenin chemiluminescence in the presence of NADH or NADPH does not represent O2ֹֿ production, but rather the respective abilities of cytochrome b5 reductase, and cytochrome P450 reductase to reduce Luc2+ to LucH+ֹ and artificially trigger a redox cycle that generates O2ֹֿ as a by-product. In light of these considerations, lucigenin is not regarded as a suitable probe for ROS generation by human spermatozoa (Baker et al., 2004, 2005).

A final comment about chemiluminescence is that the photomultipliers that transduce the light emitted in such reactions into a digital signal are not standardized, so the response characteristics of every instrument will be different. Furthermore, the chemistry of chemiluminescence is complex and the read out will be influenced by minor differences in such factors such as pH, temperature, presence of transition metals, oxygen tension, tube agitation, air bubbles in the reaction chamber etc. (Aitken et al., 2004). As a result, the chemiluminescence signals will vary significantly between runs creating so much variance that statistical differences between group means are difficult to attain, even when clear dose-dependent trends are evident. These sources of variation also mean that the calculation of numerical threshold values for chemiluminescence in a diagnostic context is an unachievable objective (Desai et al., 2009). The technology is really best suited to longitudinal studies where time-dependent changes in chemiluminescence are monitored in response to the presence of a particular reagent (e.g. Fig. 8B).

NBT – The problems with NBT reduction as a probe for ROS are similar to those described above for lucigenin. Thus, any enzyme capable of effecting NBT reduction using NADH or NADPH as an electron donor, will generate a response that masquerades as a ROS signal (Baker et al., 2004, 2005). So, while O2ֹֿ is theoretically capable of reducing NBT, the same response can be generated by a number of reductases using alternative electron donors. The result is that NBT exhibits high background levels of reduction and although the technique was capable of detecting the mitochondrial ROS generated by 2OHE2 as well as the release of extracellular ROS in response to menadione, it failed to detect the mitochondrial redox activity elicited with either AA or 4HNE.

DAF-DA – It was evident from the experiments with DAF-DA that human spermatozoa are constantly generating NO as long as they are in a viable state (Fig. 8C,D) and that this probe is not capable of detecting the differences in redox activity between low- and high-density Percoll fractions (Fig. 8A,C,D).

Mechanisms for the activation of ROS generation

The results obtained in this study have also shed some light on the complex cellular mechanisms by which the various activators of ROS generation exert their biological effects on spermatozoa. Thus, 2OHE2 elicited ROS generation from the mitochondria, generating powerful responses that could be readily detected when MSR and DHE was used as the probe. For 2OHE2 to redox cycle, it has be oxidized to the corresponding semiquinone by oxidoreductases including cytochrome P450 isoforms that are located on the inner mitochondrial membrane, with their active sites facing the mitochondrial matrix (Zhang et al., 2007; Sangar et al., 2010). In the case of AA or 4HNE, the site of action is thought to be electron transport within the inner mitochondrial membrane, which results in a significant discharge of O2ֹֿ into the inter-membranous space which, again, can be readily detected by MSR or DHE (Fig. 5C,D; Fig. 7C,D).

Menadione was different from all of the other compounds tested in that it redox cycled at or close to the surface of the spermatozoa generating significant quantities of ROS in the extracellular space, even when covalently bound to a protein scaffold that would prevent its passage across the plasma membrane (Fig. 4D). Such activity suggests that plasma membrane redox systems must be present in the sperm plasma membrane, which are capable of transmitting electrons across the plasma membrane to acceptor molecules on the sperm surface. It is presumably in this manner that spermatozoa are able to maintain the surface expression of surface thiols as they engage the capacitation process (de Lamirande & Gagnon, 1998; Gualtieri et al., 2009).

Finally, AA was different from the other inducers of ROS generation in that while it clearly stimulated mitochondrial ROS generation, it did not induce a change in sperm motility. This was not because this fatty acid did not create a state of oxidative stress. On the contrary, AA stimulated higher levels of 8OHdG formation than any of the other ROS-inducers within the first 4 h (Fig. 6A,B). The most likely explanation for this difference is while oxidative stress can precipitate motility loss, the lipid peroxidation process that underpins this process takes a matter of hours to be maximally effective (Aitken et al., 2006). On the other hand, electrophiles can adduct proteins in the sperm axoneme and induce a complete loss of motility within minutes (Hughes et al., 2009). All the reagents inducing rapid motility loss in this study were either electrophiles (menadione, 4HNE) or have the potential to rapidly transform into electrophiles (2OHE2) capable of paralysing sperm movement by adducting to axonemal proteins. In contrast, AA could only influence sperm movement by generating ROS and this mechanism alone was incapable of significantly influencing sperm motility within the 15 min observation period, although it could precipitate long-term DNA damage (Fig. 6).

Clinical relevance

Free radical generation by human spermatozoa is a complex process involving several different mechanistic pathways which may result in the generation of ROS in the mitochondria, cytoplasm or cell surface (Koppers et al., 2010; Musset et al., 2012). The current consensus is that mitochondria constitute the major source of ROS in defective human spermatozoa and that MSR is the method of choice from a clinical diagnostic perspective. Luminol–peroxidase-dependent chemiluminescence is an excellent probe for detecting ROS in the extracellular space and represents the most sensitive method for detecting seminal leucocytes, particularly when combined with a leucocyte-specific agonist such as opsonized zymosan. Conversely, lucigenin, H2DCFDA and NBT appear to lack the sensitivity, specificity and, in lucigenin's case, the repeatability needed to effectively monitor ROS generation by human spermatozoa, while flow cytometry employing DAF-DA cannot discriminate between good and poor quality spermatozoa because NO is constitutively generated by viable cells. This study should help achieve a measure of rationalization and standardization in the methods used to detect the ROS generation that underpins a significant proportion of male infertility.

Acknowledgements

We are very grateful to Jodi Powell for the management of our panel of human semen donors and the NHMRC (Program grant #494802) and ARC (DP 110103951) for financial support.

Conflict of interest

None of the authors has to declare a conflict of interest.

Ancillary