In vitro storage of boar spermatozoa increases the demand of adenosine triphosphate for reactivation of motility

Prolonging the shelf‐life of liquid‐preserved semen without compromising its fertilizing capacity may increase the efficiency of artificial insemination in pigs. Many fertilization‐relevant processes are adenosine triphosphate dependent. The impact of semen storage and rewarming to body temperature on the energy status of spermatozoa is as yet unknown.


Conclusion:
Storage of boar spermatozoa increases the demand of adenosine triphosphate for reactivation of spermatozoa towards fast, non-linear, and hyperactivationlike motility patterns upon rewarming. Maintenance of glycolysis seems to be decisive for sperm function after long-term storage in vitro.

K E Y W O R D S
adenylate energy charge, cluster analysis, mitochondria, motility, semen preservation INTRODUCTION Artificial insemination (AI) is the most widely used and efficient assisted reproductive technique in pigs. Currently, AI is practiced in over 90% of sows bred in the major pork-producing countries except for China. 1 Despite an already high reproductive performance, a further increase in AI efficiency is a main goal to accelerate genetic progress and to enhance economic benefit. 1 On the level of semen production, attempts are made to reduce the number of spermatozoa per insemination dose and to prolong the shelf-life of liquid-preserved semen without compromising its fertilizing capacity. [2][3][4] The challenge in liquid semen preservation is to reduce sperm metabolism without comprising the sperm's integrity and function.
Boar spermatozoa for AI are commonly preserved in the liquid state at temperatures between +15 • C and +18 • C. 1,5 Cooling and storing semen doses are supposed to slow down all intracellular adenosine triphosphate (ATP) consuming and producing processes and concomitantly result in a complete immobilization of boar spermatozoa.
Observations were made that ATP levels in stored semen samples declined, whereas after rewarming, the percentage of motile and viable spermatozoa reached a high level. 6 This implies that liquid preservation of boar spermatozoa leads to profound imbalances in their energy status.
Fluctuations in ATP levels are typically interpreted as a mirror of the metabolic status of a cell or cell population. A more robust estimate of the actual energy status is the adenylate energy charge (EC or AEC), which considers not only the available ATP but also the ratio of the available ATP, adenosine diphosphate (ADP), and adenosine monophosphate (AMP) molecules in the adenylate pool. 7 The EC is indicative of whether a majority of microbials, somatic cells, or spermatozoa in a sample can be considered vital. [8][9][10] Manipulation of spermatozoa, such as a simple dilution with a semen extender, can already result in a pronounced decrease in the EC. 11 The extent of such imbalances in ATP concentrations and EC during semen storage and whether the energy status can be partially or fully re-stored after warming the semen samples to body temperature have yet to be investigated.
Porcine spermatozoa are able to generate ATP through oxidative phosphorylation in the midpiece and through glycolysis at the fibrous sheath and in the sperm head. 12 Although boar spermatozoa predominantly rely on glycolysis to utilize glucose and produce ATP, 13 mitochondrial function and the related ATP production are indispensable for a range of ATP-dependent cell functions, such as calcium home-ostasis, motility, or redox signaling, 14,15 which in turn are involved in sperm capacitation, hyperactivation, or volume regulation. 16,17 Sparks of elevated mitochondrial ATP production are especially required during specific fertility-relevant events, for example, capacitation and acrosome reaction. 18,19 The mixture of available extracellular substrates determines which pathway for ATP production is preferably used. For example, exposing boar semen to extenders with high glucose concentrations shifts the metabolism of boar spermatozoa predominantly toward glycolysis. 20,21 In addition, natural energy substrates in seminal plasma, such as fructose, citrate, inositol, lactic acid (c.f. Table S1) might influence energy metabolism at storage time, even though their concentration is low and is further reduced by semen dilution. It is thus pivotal to gain knowledge on the extent to which deficits in mitochondrial function, ATP levels or EC are primed by liquid semen preservation and how they contribute to impaired sperm function under storage conditions. The aim of our study was to determine the energy status of boar spermatozoa during storage and subsequent rewarming, and to reveal the potential role of mitochondrial function for the reactivation and maintenance of sperm motility. To this end, the mitochondrial transmembrane potential was monitored in viable spermatozoa, and cluster analysis was used to monitor changes in the movement patterns in subsets of motile cells. The implications of our findings for progress in semen preservation concepts are discussed.

Experimental design
Boar semen samples (n = 7) were diluted to 20 × 10 6  The percentage of viable spermatozoa with high mitochondrial transmembrane potential and motility parameters was only determined in samples incubated at 38 • C. Any procedures leading to sperm selection, for example, density gradient centrifugation, were not applied in the experiments.
In additional experiments, the putative accumulation of L-lactate in the extracellular medium over storage time and during incubation at 38 • C was assessed. Furthermore, the effect of inhibiting glycolysis on sperm function was tested. To this end, the D-glucose in BTS was fully replaced with 2-deoxy-D-glucose while maintaining identical pH and osmolality.

Motility
An aliquot of 2 ml diluted semen was transferred to a 10 ml tube and incubated in a water bath at 38 • C. After 15, 30, 60, 120, and 180 min, an aliquot (3 µl) was loaded into one chamber of a pre-warmed (38 • C) four-chamber slide (Leja Products B.V., Nieuw Vennep, the Netherlands) with a chamber depth of 20 µm for analysis. At the same time, an aliquot of each semen sample (100 µl) was processed for assessment of ATP concentration and EC.
The sperm motility parameters were determined by a computerassisted semen analysis (CASA) system (SpermVision, Minitüb GmbH).

Viability and acrosome integrity
In parallel to motility assessments, another aliquot of 2 ml diluted semen was incubated in a water bath at 38 • C.

Mitochondrial transmembrane potential in viable spermatozoa
The principle of this measurement is based on the properties of the dye JC-1. Depending on the electrochemical gradient across the mitochondrial membrane, JC-1 reversibly transforms from a green fluorescent monomer (low transmembrane potential; emission peak approximately 529 nm) to an aggregated form (J-aggregates) emitting orange to red fluorescence (high transmembrane potential; emission peak approximately 590 nm). Combined staining with PI allowed differentiation of viable and dead spermatozoa.

Inhibition of glycolysis in the presence and absence of seminal plasma
In an additional experiment, the effect of inhibiting glycolysis on sperm function was evaluated. To this end, the predominant component of BTS, that is, D-glucose, was fully replaced with 2-deoxy-D-glucose (product no. D8375, Sigma-Aldrich Production GmbH) while maintaining identical pH and osmolality. 2-Deoxy-D-glucose inhibits glycolysis because of the formation and intracellular accumulation of 2-deoxy-D-glucose-6-phosphate, thus inhibiting the function of hexokinase and glucose-6-phosphate isomerase. 24 Seminal plasma contains glucose and other energy substrates, 25,26 which may interfere with the experiment and therefore was removed. For this, aliquots of fresh semen samples (15 ml) were centrifuged (750 g, 10 min, room temperature).
The sperm pellet was diluted to 20 × 10 6 cells/ml in either variant of the BTS medium (seminal plasma-free samples). In addition, diluted samples with both BTS variants were supplemented with 10% (v/v) autologous seminal plasma (seminal plasma containing samples). Small aliquots (1.5 ml diluted semen in 1.5 ml Eppendorf cups) were kept for 90 min at room temperature and subsequently stored for 24 h at 17 • C before being rewarmed to 38 • C and assayed for motility, viability, and mitochondrial membrane potential.

2.9
ATP and energy charge assay 2.9.1 Sample preparation and nucleotide extraction ATP content and EC in spermatozoa were determined by recently described methods. 10 The protocols were based on modifications of previously published assays. 27 (1)

L-Lactate assessment
Semen samples were centrifuged (3360 g, 10 min at room temperature or 38 • C) to obtain a sperm-free supernatant. 30 The supernatant (0.1 ml) was mixed with 0.5 ml 0.36 N HCIO 4 for deproteinization.
After 10 min of incubation, the samples were centrifuged (10,000 g, Cluster analysis was performed as described previously. 31   . Whenever two or more parameters correlated ≥0.90, only one was chosen to enter the final clustering procedure to reduce the number of variables. Chosen variables were standardized to a mean of 0 and a standard deviation of 1 to avoid any bias in the clustering procedure because of different parameter scales. Hierarchical clustering was performed using squared Euclidian distance as a distance measurement and the "centroid" algorithm for cluster fusion (PROC CLUSTER). The choice of a suitable solution from the clustering procedure was guided by the cubic clustering criterion, pseudo-F statistics, and pseudo-t 2 values. The aim was to obtain a solution that explained as much of the variance in the data set and contained as few as possible major clusters. To test whether the distribution of spermatozoa between the revealed cluster differed between the incubation periods (pooled data set) or between days at a given incubation time, a chi-square test for homogeneity was performed (PROC FREQ). It was tested whether the storage period had a confounding effect on the results from the pooled data set using the Cochran-Mantel-Haenszel test (PROC FREQ). Cramer's V (ranging from 0 to 1) was used to measure the effect size that storage or incubation had on the distribution of spermatozoa to the different clusters, that is, the size of sperm subpopulations with distinct motility patterns.
For interpretation, the following guidelines were followed as suggested by Cohen: 32 V < 0.10 = no effect, 0.10 < V ≤ 0.30 = slight effect, 0.30 < V ≤ 0.50 = moderate effect, and V > 0.50 = strong effect. Spearman's correlation coefficients were calculated for selected parameters (PROC CORR). The significance level was set at p < 0.05.
ATP levels and EC were higher at the day of semen collection, that is, at 23.7 ± 0.  Figures 2A and S1A,B). A gradual decrease in PM was noted at Days 3, 5, and 7, respectively, when samples were incubated for 60 min or more ( Figure 2A and S1C-E; p < 0.05). The average percentage of viable, acrosome-intact spermatozoa never dropped below 70% throughout the experiment (Table S2).
In contrast to PM, the percentage of viable spermatozoa with hMMP steeply declined with increasing incubation time irrespective of storage time ( Figure 2B). After a 60-min incubation period, fewer than 40% of the spermatozoa were viable with a high mitochondrial transmembrane potential (max: 38.9% at Day 1; min: 13.5% at Day 7).
After a 180-min incubation period, viable spermatozoa with a high mitochondrial transmembrane potential were virtually absent in the samples ( Figure 2B, Table S3). The absence of this sperm population was not because of an increased permeability of the spermatozoa for PI but because of a consistently decreasing fluorescence intensity of the J-aggregates ( Figure S2). Changes in EC showed the same pattern as for ATP concentration ( Figure 2D). In addition to significant gradual changes in ATP content at Days 0 and 5, the EC indicated a significant stepwise decrease in ATP and/or ADP during incubation at Day 1 ( Figure 2D). Only at Day 0, when samples had still not been cooled below 20 • C prior to rewarming . Different letters (a-d) on a given day indicate significant differences between incubation times (p < 0.05). The relative change for each parameter on a given day of storage is depicted in Figure S1 at 38 • C, was a steep decline in the percentage of viable spermatozoa with high mitochondria accompanied by a gradual decline in the EC and a delayed drop in ATP concentration. In addition to the aforementioned parameters, the AMP to ATP ratio was calculated ( Figure S3) because the AMP/ATP ratio serves as an additional indicator of energy stress. 33 The AMP/ATP ratio was significantly increased after 120 and 180 min of incubation on all days of the experiment ( Figure S3).

Sperm motility, viability, and hMMP after inhibition of glycolysis in the absence and presence of seminal plasma
In seminal plasma-free samples, inhibition of glycolysis by replacement of D-glucose with 2-deoxy-D-glucose in the semen extender BTS pre-vented the reactivation of motility and the establishment of a hMMP in viable spermatozoa during a 30-min incubation period at 38 • C (p < 0.001). Viability and acrosome integrity were not affected by inhibited glycolysis (Figure 3A; p > 0.05). In the presence of seminal plasma, the inhibition of glycolysis reduced the activating effect of incubation on motility (p < 0.05) and hMMP (p < 0.01; Figure 3B).
These data suggest that firstly, reactivation of sperm motility after storage in a glucose-rich extender relies on glycolysis, and secondly, that seminal plasma contributes to sperm reactivation by providing energy substrates for mitochondrial ATP production.

L-Lactate concentration in the extracellular medium in stored semen samples during incubation at 38 • C
The L-lactate concentrations in the extracellular medium, that is, semen extender plus seminal plasma, did not differ between Days 1 and 7 (p > 0.05; Figure 4A Figure 4C). This indicates that thermic stress induces excess L-lactate accumulation, which is no longer counterbalanced by the mitochondria and thus leads to externalization of L-lactate.

Changes in sperm motility patterns in stored semen samples and during subsequent incubation at 38 • C
The solution that was chosen from the clustering procedure consisted of eight major clusters and explained 67.6% of the variation in the data set ( Figure 5A). The eight major clusters accounted for 94.9% Sperm in the major clusters were characterized by significantly different motility characteristics (p < 0.05; Figure 5A, Table S5), which were reflected in the corresponding sperm trajectories ( Figure 5B).
The chi-square test indicated a significant relation between storage time and the distribution of spermatozoa to the different clusters  (Table S4).
Two groups of spermatozoa were present at consistent levels at all time points. These were a group of fast spermatozoa with moderate ALH and high LIN (Cluster 6; 6%-9%) and a group of very fast spermatozoa with high ALH and moderate LIN (Cluster 7; 9%-14%; Figure 5C). This indicates that some spermatozoa did not rely at all on active mitochondria to (transiently) show these highly active motility patterns. In contrast, very fast spermatozoa with highly erratic movement (wide ALH and low LIN; Cluster 13) were almost exclusively present after a 15-min incubation period (4% of all spermatozoa) and reduced to 1% thereafter ( Figure 5C). The cell chi-square indicated that deviations in sperm distribution for Clusters 3 and 7 after a 15-min incubation

Relation of ATP levels and energy charge with sperm motility parameters and mitochondrial activity
Based on all semen samples (n = 175), the ATP content was positively correlated with TM (r = 0.56), PM (r = 0.48), the percentage of spermatozoa with high MMP (r = 0.58), and the percentage of viable, acrosome-intact spermatozoa (r = 0.51; all p < 0.001). Sperm velocity and other motility descriptors were not correlated or weakly correlated with the ATP concentration (Table 1). Instead, the ATP concentration in a sample was positively correlated with the amount of spermatozoa with low LIN and moderate VCL, that is, Cluster 4 (r = 0.53, p < 0.01, n = 105; Table 2). At the same time, high ATP concentrations inversely correlated with the number of spermatozoa with high LIN and low to moderate VCL, that is, Clusters 1-3, respectively (Table 2). Together, the data indicate that a high MMP, ATP supply and sperm motility are only moderately correlated with each other.
EC correlated highest with the percentage of spermatozoa with high MMP (r = 0.51), the percentage of viable, acrosome-intact spermatozoa (r = 0.48; both p < 0.001) and, to a lesser extent, with TM (r = 0. 28) and PM (r = 0.23; both p < 0.01). Correlations with the average motility descriptors for progressively motile spermatozoa were weak ( Table 1).
The EC of a given sample showed a comparable or stronger positive correlation than the ATP concentration with the amount of moderately active spermatozoa from Cluster 4 (r = 0.49), Cluster 5 (r = 0.48), or hyperactivation-like moving cells from Cluster 13 (r = 0.34; all p < 0.01; Table 2). At the same time, a high EC in a sample inversely correlated with the number of spermatozoa with high LIN and low to moderate VCL, especially with the size of Cluster 1 characterized by relatively slow-moving spermatozoa (r = -0.51, p < 0.01; Table 2).
The percentage of viable spermatozoa with high MMP showed a positive correlation with the TM (r = 0.43), PM (r = 0.40), and the percentage of viable, acrosome-intact spermatozoa (r = 0.49; all p < 0.001; Table 1). The amount of spermatozoa with low to moderate LIN and moderate to high VCL, that is, Cluster 4 (r = 0.70), Cluster 5 (r = 0.63), and Cluster 13 (r = 0.44; all p < 0.001; Table 2

DISCUSSION
The present study demonstrates that long-term preservation of boar spermatozoa in a liquid state does not affect ATP levels and EC, but with ongoing storage time, the cells spend an increasing degree of ATP to reactivate motility during rewarming.
Consistent with the minor fluctuations in ATP and EC levels during storage at 17 • C, TM, and sperm viability remained at a high level throughout the experiment. Although our results seem to be conflicting when compared with those of earlier studies reporting decreasing ATP levels during storage time in chilled and liquid-preserved samples, 34,35 it is important to note that in the latter studies, the lowered ATP levels coincided with markedly reduced sperm motility and viability.
Our results strengthen observations from preliminary data that viable spermatozoa can maintain their energy metabolism in balance for several days at 17 • C. 10 Each cluster represents spermatozoa with a distinct motility pattern (c.f. Figure 3). *p < 0.05. **p < 0.01. ***p < 0.001.
to inactivation of enzymes in the pathways of glycolysis and/or oxidative phosphorylation should be considered. Interestingly, the present study shows that boar spermatozoa in a glucose-rich semen extender apparently rely heavily on functional glycolysis, not only for reactivating motility after chilling and storage but also for re-establishing a high mitochondrial transmembrane potential. However, the micromolar levels of energy substrates from 10% seminal plasma were sufficient to outperform the general motility-inhibiting effect of a glycolysis block by 2-deoxy-D-glucose in a subset of spermatozoa. Citrate was present in extender variants with 2-deoxy-D-glucose either combined with or without seminal plasma. Therefore, citrate may be excluded as a "rescuing" molecule. The available energy substrates in seminal plasma, fructose, glucose, and sorbitol can probably also be excluded as key molecules that rescued motility and mitochondrial membrane potential because they would require the functional availability of the putatively inhibited hexokinase to become phosphorylated and enter glycolysis. 36 A fructokinase activity for direct phosphorylation of fructose has not yet been detected at a significant level in boar spermatozoa. 37 Thus, the most likely candidate seems to be L-lactate from seminal plasma, which, independent from glycolysis, can efficiently be metabolized by spermatozoa in the mitochondria. 37 One possible factor influencing the expense of ATP for reactivating stored spermatozoa might include the activation of ATP-dependent transporters that are involved in intracellular ion homeostasis, particularly in the regulation of cytosolic calcium levels. 38,39 It has been demonstrated that sperm reactivation after chilling-induced silencing coincides with a release of calcium from intracellular stores, resulting in a hyperactivation-like motility pattern in a subset of spermatozoa 5 min after starting the warming process. 40 Thereafter, the population of hyperactivated spermatozoa declined, which indicates a time-delayed regulation of the free intracellular calcium levels. 40 Notably, in the present study, changes in sperm kinematic patterns during rewarm- ing and storage, excess lactate in the spermatozoa contributes to a gradual intracellular acidification (pH 6.2). 11 The significantly reduced ATP concentration and the reduced EC after 24 h of storage that we observed suggest that the net production of ATP is slightly negative until a temporary equilibrium is reached. A combination of factors, that is, a low intracellular pH, the presence of ATP, and probably the intracellular presence of citrate (from the semen extender), are finally all inhibitory to the phosphofructokinase activity and thus glycolysis. 41,42 In line with these assumptions, no increase in the glycolysis product L-lactate in the extracellular medium was observed between Days 1 and 7 of storage. Consequently, the "resting" level for the EC between Days 1 and 7 (0.70-0.77) at 17 • C was always lower than for freshly diluted semen at Day 0 or when compared with raw semen 10,11 or freshly isolated epididymal spermatozoa. 43 The concept of lactate/pHmediated silencing of sperm motility is comparable to the physiological situation in sperm storage tubules in birds, where these factors lead to pH-dependent reduction in ATPase activity in the spermatozoa. 43  with distinct movement patterns than with the general percentage of total and progressive motile spermatozoa. In particular, the decrease in the major sperm cluster, that is, Cluster 4, consisting of a sperm subpopulation with moderate speed, wide ALH, and moderate LIN, was highly positively correlated with the decline in mitochondrial activity (r = 0.70), while the increase in spermatozoa with low motility (Cluster 3) was inversely correlated (r = -0.55).
Despite the general notion that boar spermatozoa generate energy mainly by glycolysis, 13,50 the role of mitochondrial ATP production in modulating the LIN of sperm movement and velocity is emerging. 15,51 However, the degree to which either of the two pathways contributes to the sperm's energy balance is highly dependent on medium composition, especially the provided energy substrates. 20,21 In the present study using a typical glucose-rich semen extender, sperm subpopu- Future directions of semen preservation might aim at designing semen extenders, which support the replenishing of the ATP pool during the rewarming phase. In stallion spermatozoa, supplementation with pyruvate, which can efficiently be metabolized by mitochondria with a high ATP yield, has already been successfully evaluated to increase the longevity of the spermatozoa. 62 An alternative strategy would be not to cure the symptom of increased ATP depletion but to prevent the increased investment of ATP for re-establishing sperm functions.
Semen extenders with drastically lowered glucose levels may be a step forward in reducing putative glycolytic stress in liquid-preserved semen. 21 In conclusion, long-term storage of liquid-preserved boar spermatozoa increases the demand of ATP for reactivation of spermatozoa toward fast, non-linear, and hyperactivation-like motility patterns upon rewarming. The energy status is not affected by the storage length.
Initiation of glycolysis and re-establishment of full mitochondrial activity, particularly in the initial phase of rewarming of spermatozoa in the female genital tract, might be decisive for sperm function after longterm storage in vitro. To this end, future research focused on improving semen preservation should be directed at the pathophysiological ATPconsuming mechanisms that are induced by rewarming after sperm storage in vitro.