Behavioral Mechanism of Human Sperm in Thermotaxis: a Role for Hyperactivation

Human Reproduction, Vol.30, No.4 pp. 884–892, 2015 Advanced Access publication on January 21, 2015 doi:10.1093/humrep/dev002

ORIGINAL ARTICLE Reproductive biology Behavioral mechanism of human sperm in thermotaxis: a role for hyperactivation

Sergii Boryshpolets, Serafı´nPe´rez-Cerezales, and Michael Eisenbach* Downloaded from https://academic.oup.com/humrep/article/30/4/884/613882 by guest on 29 August 2020 Department of Biological Chemistry, Weizmann Institute of Science, 7610001 Rehovot, Israel

*Correspondence address. E-mail: [email protected]

Submitted on September 23, 2014; resubmitted on December 2, 2014; accepted on January 2, 2015

studyquestion: What is the behavioral mechanism underlying the response of human spermatozoa to a temperature gradient in thermotaxis? summary answer: Human spermatozoa swim up a temperature gradient by modulating their speed and frequencies of hyperactivation events and turns. what is known already: Capacitatedhuman spermatozoa are capable of thermotactically responding to atemperature gradient with an outcome of swimming up the gradient. This response occurs even when the gradient is very shallow. study design, size, duration: Human sperm samples were exposed to a fast temperature change. A quantitative analysis of sperm motility parameters, flagellar wave propagation, and directional changes before, during, and after the temperature change was carried out. participants/materials, setting, methods: The swimming behavior of 44 human sperm samples from nine healthy donors was recorded under a phase-contrast microscope at 75 and 2000 frames/s. A temperature shift was achieved by using a thermoregulated microscope stage. The tracks made by the cells were analyzed by a homemade computerized motion analysis system and ImageJ software. main results and the role of chance: A temperature shift from 31 to 378C resulted in enhanced speed and a lower frequency of turning events. These were reflected in a 35 + 1% (mean + SEM) increase of the straight-line velocity, 33 + 1% increase of the average path velocity, 11 + 1% increase of the curvilinear velocity, 20 + 1% increase of the wobble, and 4 + 1% increase of the linearity. Qualitatively, the inverse trend was observed in response to a 37-to-318C shift. In addition, the amplitude of flagellar waves increased close to the sperm head, resulting in higher side-to-side motion of the head and, often, hyperactivation. This increase in the extent of sperm hyperactivation was reflected in an increase in the average (mean + SEM) fractal dimension from 1.15 + 0.01 to 1.29 + 0.01 and inthepercentageofhyperactivatedspermatozoafrom3+ 1% to 19 + 2%. These changes in hyperactivation were observed less often in sperm populations that had not been incubated for capacitation. All these changes partially adapted within 3–10 min, meaning that following the initial change and while being kept at the new temperature, the values of the measured motility parameters slowly and partially returned toward the original values. These results led us to conclude that spermatozoa direct their swimming in atemperature gradient by modulating the frequency of turns (both abrupt turns as in hyperactivation events and subtle turns) and speed in a way that favors swimming in the direction of the gradient. limitations, reasons for caution: The conclusions were made on the basis of results obtained in temporal and steep temperature gradients. The conclusions for spatial, shallow gradients were made by extrapolation. wider implications of the findings: This is the first study that reveals the behavior of human spermatozoa in thermotaxis. This behavior is very similar to that observed during human sperm chemotaxis, suggesting commonality of guidance mechanisms in mammalian spermatozoa. This study further substantiates the function of hyperactivation as a means to direct spermatozoa in guidance mechanisms. study funding/competing interest(s): The authors have no conflict of interest and no funding to declare.

Key words: human sperm thermotaxis / sperm motility / sperm guidance / sperm ﬂagella

& The Author 2015. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: [email protected] Behavioral mechanism of human sperm in thermotaxis 885

Introduction hard disk in AVI format. Analyses of the sperm recordings were performed by ImageJ software and a computer-assisted sperm analyzer (CASA) plugin Mammalianspermatozoamustbe guided in order to reach theoocyte(Eisen- (Wilson-Leedy and Ingermann, 2007) upgraded according to Purchase and bach and Giojalas, 2006). Indeed, human spermatozoa were found capable Earle (2012). The analysis was carried out second by second, with 75 frames of responding to chemoattractant gradients by chemotaxis (Ralt et al., 1994), in each second. The following parameters were measured and used in this to a temperature gradient by thermotaxis (Bahat et al.,2003), and to fluid study: curvilinear velocity (VCL, time-averaged velocity of a sperm head m flow by rheotaxis (Miki and Clapham, 2013), all three being potential guid- along its actual curvilinear path, expressed in m/s), average path velocity (VAP, velocity over an average path generated by a roaming average; mm/s), ance mechanisms. On the basis of the sperm response to a concentration straight-line velocity (VSL, the time-average velocity of the sperm head along jump of a chemoattractant, the behavioral mechanism by which capacitated a straight line from its first position to its last position; mm/s), linearity (LIN, spermatozoa respond to the gradient in chemotaxis was suggested to involve defined as VSL/VAP), and wobble (WOB, defined as VAP/VCL). Two para- modulation of the frequency of hyperactivation events and turns in response meters were used for identifying hyperactivation, employing a homemade to the gradient: Reduction when swimming up the gradient and elevation computerizedmotionanalysissystem inMatLab,based onprinciplesdescribed

when swimming down the gradient (Armon and Eisenbach, 2011). by Crocker and Grier (1996): the fractal dimension (FD) and the percentage of Downloaded from https://academic.oup.com/humrep/article/30/4/884/613882 by guest on 29 August 2020 A behavioral mechanism for rheotaxis has also been proposed (Miki hyperactivated spermatozoa. FD is an expression of the degree to which the and Clapham, 2013; Kantsler et al., 2014). In this mechanism, too, hyper- sperm trajectory fillsaplane (Mortimeretal.,1996).Ifthe trajectory isa straight activation appears to be involved, assisting spermatozoa to adjust their linewith nodeviations, its FDvalue is1.0 becauseit isonlyinthe first dimension motility according to the fluid flow (Miki and Clapham, 2013). In contrast (length). If, however, the trajectory is meandering, as in the case of a hyperac- to chemotaxis and rheotaxis, there is no information about the behavior- tivated spermatozoon, it covers more of the plane and it has a value closer to 2.0. The FD values of most of the spermatozoa in a microscope field are al mechanism of thermotaxis.Uncovering this mechanism is of special im- between 1.0 and 2.0, although a few spermatozoa may have values larger portance in view of the sperm ability to respond by thermotaxis even to than 2.0 (Katz and George, 1985). FD is thus used as a measure of the intensity extremely shallow temperature gradients (Bahat et al., 2012). Our aim in of hyperactivation (Mortimer et al.,1996). The percentage of hyperactivated this study was to reveal this mechanism. spermatozoa was determined according to the criterion of spermatozoa having both FD .1.4 and VCL .70 mm/s (Armon and Eisenbach, 2011).

Materials and Methods High-speed recording of sperm motility Ethical approval High-speed recordings of sperm motility were carried out at the Laboratory of Reproductive Physiology in Faculty of Fisheries and Protection of Water, The study was approved by the Bioethics and Embryonic Stem Cell Research University of South Bohemia, Czech Republic, using a high-speed video Oversight Committee of the Weizmann Institute of Science. Human semen camera (Olympus i-speed TR, Japan, providing 848 × 688 pixels spatial reso- samples wereobtained from healthy donors after3 days of sexual abstinence. lution) at 2000 frames/s, mounted on an inverted microscope (Olympus Informed consent was obtained from each donor. IX83, Japan). Recordings were initiated immediately after the temperature reached the final value of 378C (for positive 31-to-378C gradient) or 318C Sperm sample preparation (for negative 37-to-318C gradient). Semensampleswithnormalsperm density,motilityandmorphologyaccording to World Health Organization guidelines (2010) were allowed to liquefy for Statistical analysis 30–60 min at room temperature and then separated from the seminal Forty-four human sperm samples from nine donors in total were analyzed in plasma by centrifugation (120×g, 15 min, twice) with Flushing Medium (Med- this study. All the samples were temperature responsive. Datawere analyzed iCult, Jyllinge, Denmark). Subsequently, the sperm concentration was adjusted by Statistika software (StatSoft, Inc., Tulsa, OK, USA). Each experimental to 70 × 106 cells/ml with Flushing Medium supplemented with additional point in the graphs is an average of 10–50 spermatozoa (all the motile cells human serum albumin, bringing its final concentration to 0.3%, and the sperm- seen in the microscope field) measured for 1 s each. No statistical compar- atozoa were incubated for 2 h under an atmosphere of 5% CO at 378Cfor 2 isons were performed. capacitation (Cohen-Dayag et al.,1995). The sperm samples were then diluted with Flushing Medium (with or without methylcellulose) to a final con- centrationof3–7 × 105 cells/ml(yielding 10–50swimmingspermatozoaper Results observation field of the microscope) and placed between thin coverslips separated by a tape. The thinness of this setup increased the heat transfer from the To identify and characterize the behavioral response of human sperm- heating stage to the sample. For experiments in which the sperm response to atozoa to a temperature gradient we exposed them to a temporal gradi- temperature was compared before and after capacitation, the semen (follow- ent with the intention of later applying the conclusions to a spatial ing liquefaction) was divided to two portions. One portion was treated as gradient. To do so, we rapidly changed the temperature of the micro- described above (‘after capacitation’ sample). The other portion (‘before cap- scope’s heating stage from 31 to 378C and vice versa. The rise time acitation’ sample) was diluted (without incubation and washings) with Flushing was rather constant in all experiments and took 40 s. The temperature Medium, supplemented with human serum albumin, as mentioned above, to a drop varied between experiments (100–200 s) because the cooling rate concentration of 3–7 × 105 cells/ml, and immediately assayed for sperm depended on the ambient temperature (Fig. 1A). swimming behavior as described above.

Analysis of sperm motility Temperature-dependent changes in human Motilityof human spermatozoawas recorded under a microscope (Alphaphot sperm swimming speed YS, Nikon, Japan) at 10× using a phase-contrast condenser and a digital The response of the spermatozoa to each of these changes was imme- camera (uEye, Germany) set at 75 frames/s. Recordings were stored on a diate. All the velocity parameters increased with the temperature, 886 Boryshpolets et al. Downloaded from https://academic.oup.com/humrep/article/30/4/884/613882 by guest on 29 August 2020

Figure 2 Effect of temperature change on human sperm trajectories. (A) Representative sperm trajectories at 318C just prior to temperature shift. (B) Representative sperm trajectories at 378C just after the temperature shift. Red arrows indicate hyperactivation events.

the values of WOB and, to a lesser extent, LIN increased with the temperature (Fig. 1C). The increase in WOB and LIN means that, at higher temperatures, the cells turned less, the side-to-side motion of the head decreased, and the trajectories of their motility became smoother and more linear (Fig. 2). All these changes were reversible in the sense that the velocity parameters decreased as the temperature was restored to the original value (Fig. 1). The slope of this decrease varied between experiments in relation to the variations in cooling rate mentioned above. Figure 1 Response of human spermatozoa to temperature changes. Spermatozoa were exposed to the indicated temperature changes and Temperature-dependent changes in human their motility was recorded and analyzed. The figure shows the motility parameters curvilinear velocity (VCL), average pass velocity (VAP), sperm swimming direction straight-line velocity (VSL), linearity (LIN) and wobble (WOB), Directional changes are often characterized by hyperactivation events defined under Materials and Methods, of a representative single experi- (Armon and Eisenbach, 2011). To examine whether, indeed, the cells ment (out of 30 analyzed experiments in total). Each experimental point make more turns as the temperature drops, we followed the level of is an average of 10–50 spermatozoa (all the motile cells seen in the hyperactivation during and after the temperaturechanges. We measured microscope field) measured for 1 s each. The bars represent 0.95 con- two parameters: FD, which is a parameter that reflects the shape of the fidence intervals calculated by Statistika software (StatSoft, Inc., Tulsa, trajectory and was used previously as a criterion for hyperactivation OK, USA). (A) Heating and cooling thermogram of the microscope’s (Mortimer et al., 1996), and the percentage of hyperactivated cells heating stage. (B) Temperature-jump stimulated changes in average vel- (derived from both FD and the swimming speed—see Materials and ocity parameters. (C) Temperature-jump stimulated changes in the calculated values of linearity and wobble. Methods for definitions). At each time point the FD value was an average obtained from all the cells moving in the observation field whereas the percentage of hyperactivated spermatozoa yielded a single experimental point. Consequently, the fluctuations in the case of with VSL and VAP changing to a larger extent than VCL (35 + 1% and FD were smaller. Clearly, cells became more hyperactivated as the tem- 33 + 1% change versus 11 + 1%, respectively; mean + SEM, n ¼ 30 perature dropped, and vice versa when the temperature became higher experiments; Fig. 1B). As a matter of fact, the major change in VCL (Fig. 3A and B for FD and the percentage of hyperactivated spermatozoa, was preceded by a minor, brief change in the opposite direction, prob- respectively; Supplementary data, Movie S1). On average, upon a tem- ably reflecting the dependence of VCL on both the forward speed and perature drop, FD increased from 1.15 + 0.01 to 1.29 + 0.01 and the the side-to-side movements of the head, each being oppositely affected percentage of hyperactivated spermatozoa increased from 3 + 1% to by the temperature. As a result of these changes in VSL, VAP and VCL, 19 + 2% (mean + SEM, n ¼ 13 experiments). Notably, the average Behavioral mechanism of human sperm in thermotaxis 887

Temperature effects in human sperm populations before and after capacitation In both chemotaxis (Cohen-Dayag et al., 1995) and thermotaxis (Bahat et al., 2003), spermatozoa must be capacitated for being responsive to a spatial gradient of the stimulus. Yet, when the stimulus is strong, as in the case of a rapid, large increase in a chemoattractant concentration, even a considerable fraction of the non-capacitated spermatozoa can respond (Armon and Eisenbach, 2011). To investigate whether the behavioral changes in response to temporal stimulation, observed in this study, arerestricted to capacitated spermatozoa as in a spatial temperaturegra- dient or whether non-capacitated spermatozoa are also responsive as in the case of a temporal chemoattractant gradient, we studied the same Downloaded from https://academic.oup.com/humrep/article/30/4/884/613882 by guest on 29 August 2020 sperm samples prior to and subsequent to capacitation. At 378C there was no significant difference between the motility parameters of both populations (Table I). Following a temperature shift to 318C, the parameters of both populations changed, but those of the population that had been allowed to capacitate changed to a larger extent (Table I). Gen- erally speaking, the parameters that reflect directional changes were little altered before capacitation but changed much more after capacitation. For example, the level of hyperactivated spermatozoa changed upon cooling from0to 4% priorto capacitationand 1 to14%aftercapacitation. The changes in the velocity parameters were smaller (Table I). These results suggest that the directional changes are mainly performed by capacitated spermatozoa whereas the temperature-dependent changes in velocity are the outcome of both the response of capacitated spermatozoa to the temperature gradient and, mainly, the speed dependence of all spermatozoa on the ambient temperature. Figure 3 Adaptation of the response to temperature reflected in hyperactivation. Human spermatozoa were exposed to the indicated Partial temperature adaptation temperature changes and their motility was recorded and analyzed. Another characteristic of the response to the temperature change was The figure shows the motility parameters of a single experiment having slow partial adaptation, meaning that following the initial change and relatively high levels of fractal dimension (FD) and hyperactivated sperm- while being kept at the newtemperature, the values of the measured mo- atozoa (out of 13 analyzed experiments in total). Each point shown is an average of 10–50 spermatozoa measured for 1 s each. (A) Partial adap- tility parameters slowly and partially returned toward the original values. tation reflected in the fractal dimension (FD)*. (B) Partial adaptation This was observed for FD (Fig. 3A), percentage of hyperactivated sperm- reflected in the percentage of hyperactivated spermatozoa. The red atozoa (Fig. 3B), and the velocity-related parameters VSL and VAP line stands for 50 s moving average. *FD expresses the degree to (Fig. 4B). This adaptation was not detected in VCL (Fig. 4A) probably which the sperm trajectory fills a plane (Mortimer et al., 1996). If the tra- because of the opposite effects of temperature on the speed and jectory is a straight line, its FD value is 1.0. If the trajectory is meandering, side-to-side head displacement, each being reflected in VCL. The as for a hyperactivated spermatozoon, it has a value closer to 2.0. FD is extent and kinetics of the adaptation varied between sperm samples. therefore used as a measure of the intensity of hyperactivation. The extent was in the range of 7–70% of the response (for FD and percentage of hyperactivated spermatozoa) and 5–30% for VSL and VAP. In a few sperm samples, no adaptation of VSL and VAP could be detected. increase in FD was the outcome of both cells acquiring higher FD values The time that it took for the cells to reach maximal (but partial) adapta- and a wider distribution of FD values (Supplementary Fig. S1). Also, the tion was in the range of 3–10 min. Consistent with the conclusion drawn increase in FD was large only in a relatively small fraction of the cells. We from the dependence on capacitation in the preceding section, this also tried tracking individual cells. Qualitatively, when the temperature partial adaptation suggests that spermatozoa respond to both the tem- was shifted from 37 to 318C we observed more frequent and more pro- perature gradient and the new ambient temperature (see Discussion). longed turns and hyperactivation events. However, quantification of these changes in individual cells turned out to be problematic due to Effect of viscosity the large variability between cells and, mainly, the very short time that To further examine, at a more physiological viscosity, the conclusions each cell spent in the observation field before leaving it (especially at reached above, we repeated the experiments in a medium containing the higher temperature). Taken together, the above results suggest methylcellulose (0.5% w/v) (Ivic et al., 2002). The relatively high viscosity that in thermotaxis, as in chemotaxis (Armon and Eisenbach, 2011), a imposed by methylcellulose reduced the side-to-side motion of head negative stimulus (i.e. a decreasing temperature gradient) results in and, consequently, VCL (now reaching values similar to those of VSL; increased frequencies of hyperactivation events and turns, while a posi- Fig. 5A). At this viscosity the temperature effect on all velocity para- tive stimulus results in decreased frequencies. meters was similar. In contrast, the effect of temperature on FD 888 Boryshpolets et al.

Table I Effect of capacitation of human sperm on the parameters of motility.

Parameter Before capacitationa After capacitationa ...... 378C318C Relative change (%) 378C318C Relative change (%) ...... VCL 100 + 396+ 43+ 3 104 + 397+ 56+ 3 VAP 76 + 557+ 326+ 369+ 245+ 235+ 1 VSL 70 + 550+ 328+ 264+ 239+ 239+ 2 LIN 0.89 + 0.01 0.86 + 0.01 3 + 1 0.88 + 0.01 0.82 + 0.01 7 + 1 WOB 0.72 + 0.03 0.59 + 0.02 17 + 2 0.66 + 0.02 0.5 + 0.02 24 + 1 FD 1.1 + 0.01 1.16 + 0.01 1.12 + 0.01 1.24 + 0.02

HS 0 4 + 11+ 114+ 2 Downloaded from https://academic.oup.com/humrep/article/30/4/884/613882 by guest on 29 August 2020

VCL: curvilinear velocity(time-averagedvelocityof asperm head along its actualcurvilinear path,expressedin mm/s);VAP,average path velocity(velocityoveran averagepath generated by a roaming average; mm/s); VSL: straight-line velocity (the time-average velocity of the sperm head along a straight line from its first position to its last position; mm/s); LIN: linearity (defined as VSL/VAP); WOB:wobble (defined as VAP/VCL); FD: expressesthe degreetowhich thesperm trajectory fills a plane (if the trajectory is astraight line, its FD value is 1.0. If the trajectory is meandering, as for a hyperactivated spermatozoon, it has a value closer to 2.0. FD is therefore used as a measure of the intensity of hyperactivation); HS: percentage of hyperactivated spermatozoa. aMean + SEM of six experiments.

changed at high viscosity. First, the measured FD values were close to 1, meaning that the cells swam in rather straight lines. And second, the FD values did not change in response to temperature elevation but only changed in response to a temperature drop (Fig. 5B). The reason is probably that temperature elevation, which causes the cells to be more linear, could not change the linearity of already linearly swimming cells. In contrast, a temperature drop results in turns and this is reﬂected in elevated FD, though modest due to the high viscosity.

Temperature-dependent changes in flagellar beat To determine how the different temperature-dependent changes in motility parameters were generated, we recorded the motion of the sperm flagellum with a high-speed camera (2000 frames/s) before (Fig. 6A) and immediately after (Fig. 6B) shifting the temperature from 37 to 318C. At 318C the flagellar wave was formed closer to the sperm head and its amplitude was larger than at 378C. This larger flagellar wave amplitude was reflected in a larger side-to-side head displacement (white frames at the right of Fig. 6), occasionally leading to hyperactivated spermatozoa (Fig. 7, Supplementary data, Movie S2). Notably, hyperactivation often resulted in three-dimensional motion of the flagellum and, consequently, of the whole spermatozoon (Fig. 7,Supplementarydata, Movie S2).At highviscos- ity the motions of the flagellum and the head became more planar and with smaller amplitude (Supplementary Fig. S2), consistent with earlier studies (Kirkman-Brown and Smith, 2011).Inspiteofthese viscosity-relatedrestric- tions, we detected at 318C turning events usually accompanied by brief episodes of flagellar arrest, in which curved flagella stayed immotile for a fraction of a second (Fig. 8; Supplementary data, Movie S3). Figure 4 Partial adaptation in the response to temperature reflected in velocity parameters. Human spermatozoa were exposed to the indicated temperature changes and their motility was recorded and ana- Discussion lyzed. The figure shows the motility parameters of the experiment shown in Fig. 3 (out of 13 analyzed experiments in total). Details as in Behavioral response to a temperature change Fig. 3.(A) No adaptation in VCL. (B) Partial adaptation reflected in the VAP and VSL. In this study we demonstrated that the behavioral response of human spermatozoa to a rapid change in temperature involves two major Behavioral mechanism of human sperm in thermotaxis 889

components. One is speed enhancement as the temperature increases, expressed in all velocity parameters (Fig. 1). This enhancement is fully reversible when the temperature is restored to the original value. As in any other behavioral system, this response also involves partial adaptation. For example, when the temperature is shifted from 37 to 318C, the initial decrease in speed is followed by a slower increase (Fig. 4B). The other component of the behavioral response is increased ﬂagellar wave amplitude as the temperature drops (Fig. 6). This larger amplitude leads to enhanced side-to-side head displacement and increased frequency of turnings and hyperactivation events (Figs 2 and 3). This component of the response undergoes partial adaptation as well (e.g. Fig. 3). It is worth noting that temperature elevation and increased viscosity

affect the sperm flagella similarly, i.e. both of them reduce the wave amp- Downloaded from https://academic.oup.com/humrep/article/30/4/884/613882 by guest on 29 August 2020 litude of the flagella (Fig 6 and Supplementary Fig. S2), leading to smaller side-to-side head displacement, more linear swimming and probably a more efficient method of propagation.

What is sensed by human spermatozoa? The temperature-related changes in sperm motility could, in principle, reflect two processes. One is an effect of the new ambient temperature, as sperm motility is expected to be temperature-dependent, though not much so (Bahat et al., 2012). The other is a response to the temperature gradient per se. This responseis expectedtoinvolve adaptation, i.e. restor- ation of the unstimulated behavior after being a while in the new temperature. Our observation that the cells adapt but that the adaptation was incomplete (Figs 3 and 4) therefore suggests that both these processes are involved in the sperm response to temperature. The observation that the extent of adaptation reflected in FD and in the percentage of hyperactivated spermatozoa (Fig. 3) was larger than for the velocity para- Figure 5 Temperature-jump stimulated changes in human sperm meters (Fig. 4) suggests that the velocity changes are mainly due to the dif- motility at relatively high viscosity. The figure shows the motility para- ferent ambient temperature whereas the directional changes are mainly meters of a representative single experiment (out of 10 analyzed experi- (but not solely) the sperm response to the gradient per se. The much ments in total). Each point shown is an average of 10–50 spermatozoa measured for 1 s each. (A) Average velocity parameters. The bars re- higher temperature-stimulated changes in the level of hyperactivated present 0.95 confidence intervals. (B) The mean fractal dimension. spermatozoa after capacitation than before capacitation (Table I), taken The black line stands for 20-point moving average. together with the fact that only capacitated spermatozoa are thermotactically responsive (Bahat et al.,2003),areconsistentwiththisconclusion.

Figure 6 Representative photographs of temperature-jump stimulated changes in human sperm ﬂagellar wave propagation. The photographs are shown at 0.005 s intervals. The last frames (frames with white background at the right side of each row) show typical head trajectories for these ﬂagellar motions. (A) A spermatozoon at 378C just prior to the temperature shift to 318C. The width of each photograph frame is 17–18 mm. (B) A spermatozoon at 318C just after the temperature shift from 378C. The width of each photograph frame is 16–17 mm. 890 Boryshpolets et al. Downloaded from https://academic.oup.com/humrep/article/30/4/884/613882 by guest on 29 August 2020

Figure 7 Representative photographs of human sperm ﬂagellar wave propagation during hyperactivationcaused byatemperaturejump from 37 to 318C. The photographs are shown at 0.005 s intervals. The last frame (the frame with a white background at the bottom right corner) shows typical head trajectories for these ﬂagellar motions. The width of each photograph frame is 30–31 mm.

Figure 8 Representative photographs of ﬂagellar wave propagation of a human spermatozoon turning in medium with high viscosity (0.5% methylcellulose). The photographs are shown at 0.015 s intervals. The width of each photograph frame is 36–37 mm.

It is worth noting thatsperm thermotaxis and chemotaxis appear to be How is a temperature gradient sensed? different in this respect. In chemotaxis, there seems to be a difference Human spermatozoa can sense extremely shallow temperature gradi- between a subtle stimulation (as in a spatial chemoattractant gradient), ents (Bahat et al., 2012), raising the question of how the gradient is where only capacitated spermatozoa are responsive, and a strong stimu- sensed—over space or over time? All the temperature changes made lation (as in rapid photorelease of a chemoattractant to the medium), in this studywere over time, with temperaturekeptessentially homogen- where a signiﬁcant fraction of a sperm population, including non- ous over space. This suggests that, as in the case of human sperm chemo- capacitated spermatozoa, respond (Armon and Eisenbach, 2011). This taxis (Gakamsky et al., 2009; Armon and Eisenbach, 2011), sperm difference was explained by assuming that capacitated spermatozoa thermotaxis involves temporal gradient sensing. This means that sperm- are more sensitive to the chemoattractant than non-capacitated spermatozoa compare the temperature (or a temperature-dependent atozoa and, therefore, only they can respond to shallow gradients; when function) between consecutive time points. This, however, does not the stimulus is sufﬁciently strong, non-capacitated spermatozoa can exclude the occurrence of spatial temperature sensing in addition to respond as well (Gakamsky et al., 2009). Our observation that the in- temporal sensing. crease in the fraction of hyperactivated cells in response to a steep temporal temperaturedrop is not as large as in the fractionof responsive cells in chemotaxis and it does not exceed the known level of capacitated How do human spermatozoa navigate spermatozoa suggests that, in thermotaxis, responsiveness may indeed in a spatial temperature gradient? be restricted to capacitated spermatozoa. An alternative possibility is A major difference between a temporal temperature gradient generated that the stimulus in photorelease experiments is much more intense as in the current study and a spatial temperature gradient, like the one than the stimulus in rapid temperature change. that exists in the female genital tract (David et al., 1972; Hunter and Behavioral mechanism of human sperm in thermotaxis 891 Downloaded from https://academic.oup.com/humrep/article/30/4/884/613882 by guest on 29 August 2020

Figure 9 A model of human sperm behavior in thermotaxis. The intensity of the background color represents the temperature gradient. The dashed black line represents the trajectory of the head. The dashed red line represents the average path. See text for details.

Nichol, 1986; Bahat et al., 2005), is the magnitude: the temperature concentration gradient. Moreover, even in the third mechanism of mam- change sensed by a swimming spermatozoon is orders of magnitude malian sperm guidance, rheotaxis (Miki and Clapham, 2013), where gradi- larger in the former. As a matter of fact, temporal changes equivalent ents are not involved, directional changes occur by way of hyperactivation to those in a physiological spatial gradient are so small that they would events. This means that in all of the known human sperm guidance probably be close to the thermal noise and, therefore, difficult to mechanisms—chemotaxis, thermotaxis and rheotaxis, hyperactivation is achieve. Even if such changes could be reliably generated, the response used asa meansoffast directionalchanges. Thisisalsosimilar toEscherichia would probably be too subtle to be detected. Therefore, the best that colichemotaxis,wheretumblesfulfillthefunctionofhyperactivationevents can be done is to deduce from the observed response to a large step (Macnab and Koshland, 1972). All these point to the commonality of temperature-change what the response would be to changes as subtle behavioral mechanisms in human spermatozoa. as in a spatial temperature gradient. Since the response to a large step temperature change involved both kinetic and directional components, Physiological significance we propose the following model for sperm behavior in a spatial tempera- Beyond getting an insight into the behavioral mechanism of human sperm ture gradient. thermotaxis and into the mode of temperature sensing, and beyond According to the model (shown schematically in Fig. 9), when a capa- finding commonality in some aspects of sperm guidance mechanisms, citated spermatozoon happens to move down the gradient, it slows as just discussed, this study also has implications for hyperactivation. down. Concomitantly it turns again and again until it returns to swim This motility pattern is nowadays used as one of the means to identify up the gradient and to sense a temperature increase. The turns could capacitated spermatozoa. Yet, the findings of this study, showing that be rather subtle or generated by episodes of hyperactivation. At higher the fraction of hyperactivated spermatozoa is tightly dependent on the viscosities (as in the female genital tract) subtle turns probably dominate. temperature and, especially, on temperature changes, require extra When the spermatozoon happens to swim up the temperature gradient caution when doing such correlations. This is because hyperactivation it continues swimming rather linearly, while slightly enhancing its speed. events could be generated as a result of temperature fluctuations. The enhanced speed and linear movement likely persist as long as the With the extreme temperature sensitivity of human spermatozoa and propagation of the cell is up the gradient. All these ensure that, during their response to even subtle temperature changes (Bahat et al., sperm thermotaxis, propagation in the gradient direction is preferred. 2012), it is of utmost importance to maintain a well-controlled constant When a spermatozoon does not sense a temperature change for a temperature in studies of sperm motility. Due to the high sensitivity and while it adapts, resuming its unstimulated swimming—a rather straight the dependence on capacitation, sperm responsiveness to a temporal swimming with occasional turns. temperature shift could be a new parameter for determining sperm quality, which is easy to measure and evaluate. For being applicable, Hyperactivation and commonality however, further clinical research is required. of behavioral mechanisms This proposed mechanism is very similar to that proposed for human Supplementary data sperm chemotaxis (Gakamsky et al.,2009; Armon and Eisenbach, 2011), with the temperature gradient substituting for the chemoattractant Supplementary data areavailable athttp://humrep.oxfordjournals.org/. 892 Boryshpolets et al.

Acknowledgements Cohen-Dayag A, Tur-Kaspa I, Dor J, Mashiach S, Eisenbach M. Sperm capacitation in humans is transient and correlates with chemotactic We thank Dr Leah Armon and Dr Anat Bahat for technical assistance and responsiveness to follicular factors. Proc Natl Acad Sci USA 1995; useful comments during setting up the experiments. We thank Oshri 92:11039–11043. Afanzar for the homemade computerized motion analysis system in Crocker JC, Grier DG. Methods of digital video microscopy for colloidal MatLab, and Daniel Blair and Eric Dufresne for the development of the studies. J Colloid Interface Sci 1996;179:298–310. tracking code implemented in our MatLab tracking software (http:// David A, Vilensky A, Nathan H. Temperature changes in the different parts of the rabbit’s oviduct. Int J Gynaecol Obstet 1972;10:52–56. www.physics.emory.edu/faculty/weeks//idl/index.html). Special thanks Eisenbach M, Giojalas LC. Sperm guidance in mammals—an unpaved road to are due to Dr Martin Psenicka, Laboratory of Reproductive Physiology the egg. Nat Rev Mol Cell Biol 2006;7:276–285. in Faculty of Fisheries and Protection of Water, University of South Gakamsky A, Armon L, Eisenbach M. Behavioral response of human Bohemia, Czech Republic for allowing us to carry out measurements spermatozoa to a concentration jump of chemoattractants or intracellular with their high-speed video camera. M.E. is an incumbent of the Jack cyclic nucleotides. Hum Reprod 2009;24:1152–1163.

and Simon Djanogly Professorial Chair in Biochemistry. Hunter RHF, Nichol R. A preovulatory temperature gradient between the Downloaded from https://academic.oup.com/humrep/article/30/4/884/613882 by guest on 29 August 2020 isthmus and the ampulla of pig oviducts during the phase of sperm storage. J Reprod Fertil 1986;77:599–606. Authors’ roles Ivic A, Onyeaka H, Girling A, Brewis IA, Ola B, Hammadieh N, Papaioannou S, Barratt CLR. Critical evaluation of methylcellulose as an Conceived and designed the experiments: S.B., S.P.-C. and M.E. alternative medium in sperm migration tests. Hum Reprod 2002;17: Performed and analyzed the experiments: S.B. Wrote the paper: S.B. 143–149. and M.E. Kantsler V, Dunkel J, Blayney M, Goldstein RE. Rheotaxis facilitates upstream navigation of mammalian sperm cells. Elife 2014;3:e02403. Katz MJ, George EB. Fractals and the analysis of growth paths. Bull Math Biol Funding 1985;47:273–286. No external funding was either sought or obtained for this study. Kirkman-Brown JC, Smith DJ. Sperm motility: is viscosity fundamental to progress? Mol Hum Reprod 2011;17:539–544. Macnab RM, Koshland DE. The gradient-sensing mechanism in bacterial Conflict of interest chemotaxis. Proc Natl Acad Sci USA 1972;69:2509–2512. Miki K, Clapham DE. Rheotaxis guides mammalian sperm. Curr Biol 2013; None declared. 23:443–452. Mortimer ST, Swan MA, Mortimer D. Fractal analysis of capacitating human References spermatozoa. Hum Reprod 1996;11:1049–1054. Purchase CF, Earle PT. Modifications to the IMAGEJ computer assisted Armon L, Eisenbach M. Behavioral mechanism during human sperm sperm analysis plugin greatly improve efficiency and fundamentally alter chemotaxis: involvement of hyperactivation. PLoS One 2011;6:e28359. the scope of attainable data. J Appl Ichthyol 2012;28:1013–1016. Bahat A, Tur-Kaspa I, Gakamsky A, Giojalas LC, Breitbart H, Eisenbach M. Ralt D, Manor M, Cohen-Dayag A, Tur-Kaspa I, Makler A, Yuli I, Dor J, Thermotaxis of mammalian sperm cells: a potential navigation Blumberg S, Mashiach S, Eisenbach M. Chemotaxis and chemokinesis of mechanism in the female genital tract. Nat Med 2003;9:149–150. human spermatozoa to follicular factors. Biol Reprod 1994;50:774–785. Bahat A, Eisenbach M, Tur-Kaspa I. Periovulatory increase in temperature Wilson-Leedy JG, Ingermann RL. Development of a novel CASA system difference within the rabbit oviduct. Hum Reprod 2005;20:2118–2121. based on open source software for characterization of zebrafish sperm Bahat A, Caplan SR, Eisenbach M. Thermotaxis of human sperm cells in motility parameters. Theriogenology 2007;67:661–672. extraordinarily shallow temperature gradients over a wide range. PLoS World Health Organization. WHO Laboratory Manual for the Examination and One 2012;7:e41915. Processing of Human Semen. Geneva, Switzerland: WHO Press, 2010.