Sperm movement is necessary for reproduction and high sperm motility is associated with increased fertilization rates (1,2). Asthenozoospermia (low sperm motility) can impede fertilization and affects ~81% of men attending for diagnostic semen analysis (3). Current therapeutic options for asthenozoospermia are limited, with intracytoplasmic sperm injection (ICSI) being the main treatment option. However, only 21.1% of ICSI treatments across Europe result in a live birth and it is, on average, more expensive than conventional IVF (4). During assisted conception sperm are often washed using silica bead density gradient centrifugation to select those of higher quality with increased motility, and lower quality sperm are discarded (5). A greater understanding of the metabolic processes that underpin sperm quality and motility could lead to novel therapeutic options for asthenozoospermia. In this study, we examined differences in metabolite profiles between sperm of high and low quality, obtained from different fractions of the density centrifugation sperm washing, in order to identify possible intracellular biomarkers of sperm quality.Normozoospermic ejaculates from 7 research donors (recruited with LREC approval) were collected and analyzed for sperm count and progressive motility. Samples were washed using 40:80% (v/v) Percoll-PBS gradient and sperm from both the 80% pellet (higher quality) and the 40:80% interface (lower quality) were collected. Leukocyte depletion was performed on both sperm populations by magnetic separation after incubation with Dynabeads CD45 (Invitrogen). Samples were examined at 37°C on a 9.4T MR spectrometer to acquire spectra using a {¹H} water-gate solvent suppression sequence (SW=20ppm, NS=1024, AQ=0.5s, D1=4). Each spectrum was phase and baseline corrected. Metabolite peaks were identified in the ¹H spectrum and integrated: they are presented per million sperm. Additionally, the mean integral from two empty regions of the spectrum (10-11 ppm and -1-0 ppm) was subtracted from each individual metabolite integral to remove any additional baseline bias. All integrals were normalised with the maximum integral set as 100. Wilcoxon matched-pairs signed rank test was performed and (P<0.05) taken as significant. In a subset of four samples, time dependent changes were monitored by acquiring sequential ¹H spectra (acquisition parameters as above) over a period of ~20 hours. These time course spectra were integrated in bins of 0.05ppm between 0 and 10 ppm. Each bin for the time course was fitted to a linear regression model to determine relative rates of change within the spectrum. All data reported as mean ±SE (n=7), unless otherwise stated.Higher quality sperm recovered from the 80% pellet had higher progressive motility compared to lower quality sperm from the 40:80% interface (44.5% ±5.5 vs. 31.0% ±3.7, P<0.05). Comparing the metabolite integrals for the samples showed that low quality sperm had significantly higher concentrations of choline (56.6 ± 8.8 vs. 26.0 ± 3.6), methyls (22.7 ±6.2 vs. 5.6 ±1.1), citrate (16.4 ±5.3 vs. 3.7 ±1.0), lactate (6.7 ±1.8 vs. 2.1 ±0.7) and aromatics (16.27 ±8.2 vs. 2.0 ±0.9) compared to high quality sperm (P<0.05) (figures 1-3). Acetyl-carnitine integrals were not significantly different (7.1 ±2.2 vs. 3.2 ±0.8). Minimal changes were observed for most of the integral bins in the time course spectra, with exception those containing choline, acetate and methyl integrals which increased in concentration after ~5hrs (n=5).

Sperm motility was significantly different between the two fractions and fell either side of the WHO lower reference limit (32% progressively motile sperm) (6). However, motile sperm were still found in the low quality fraction. This may be due to other factors pertinent to sperm density, such as sperm morphology and DNA anomalies, having a large influence on sperm gradient separation (7,8).

Metabolites were at higher concentrations in low quality sperm. There were significant differences between high and low quality sperm integrals for choline and methyls, both of which are incorporated into the cell membrane. These differences indicate low quality sperm have altered sperm morphology and/or increased cytoplasmic vacuoles. Though it is unclear why citrate, the first product of the TCA cycle, was increased in low quality sperm it may indicate decreased oxidative phosphorylation, leading to accumulation of TCA cycle products. Also, lactate was at higher concentrations in lower quality sperm and may indicate increased glycolysis, perhaps to compensate for a lack of oxidative phosphorylation. The metabolite changes observed over time suggest that sperm may undergo apoptosis after 5 hours under these conditions.