1. Developmental Biology

A phenotypic screening platform utilising human spermatozoa identifies compounds with contraceptive activity

  1. Franz S Gruber
  2. Zoe C Johnston
  3. Christopher LR Barratt  Is a corresponding author
  4. Paul D Andrews
  1. University of Dundee, United Kingdom
  2. Ninewells Hospital and Medical School, University of Dundee, United Kingdom
https://doi.org/10.7554/eLife.51739
Share this article
Cite this article
  1. Franz S Gruber
  2. Zoe C Johnston
  3. Christopher LR Barratt
  4. Paul D Andrews
(2020)
A phenotypic screening platform utilising human spermatozoa identifies compounds with contraceptive activity
eLife 9:e51739.
https://doi.org/10.7554/eLife.51739
  2. Download BibTeX
  3. Download .RIS

Abstract

There is an urgent need to develop new methods for male contraception, however a major barrier to drug discovery has been the lack of validated targets and the absence of an effective high-throughput phenotypic screening system. To address this deficit, we developed a fully-automated robotic screening platform that provided quantitative evaluation of compound activity against two key attributes of human sperm function: motility and acrosome reaction. In order to accelerate contraceptive development, we screened the comprehensive collection of 12,000 molecules that make up the ReFRAME repurposing library, comprising nearly all the small molecules that have been approved or have undergone clinical development, or have significant preclinical profiling. We identified several compounds that potently inhibit motility representing either novel drug candidates or routes to target identification. This platform will now allow for major drug discovery programmes that address the critical gap in the contraceptive portfolio as well as uncover novel human sperm biology.

Introduction

An effective and comprehensive family planning strategy is fundamental to delivery on the United Nations 17 Sustainable Development Goals (Starbird et al., 2016; Goodkind et al., 2018), however the current contraceptive portfolio is suboptimal. For example, it is estimated that more than 214 million women in developing countries have an unmet need for contraception which results in 89 million unintended pregnancies and 48 million abortions every year, with substantial impacts on maternal and child health and poverty levels (Guttmacher Institute, 2017). To achieve the UN targets it is critical that involvement of men be considered of equal importance as women (Hardee et al., 2016). However, it is noticeable that no effective, reversible and widely available form of contraception has been developed for men since the condom and as such the burden falls largely on the woman. The development of drugs that can be used in the male would address a critical gap in the contraceptive portfolio.

Progress in developing new male contraceptives has been slow for a plethora of reasons. Despite decades of research, safe and effective hormonal approaches to prevent spermatozoa production have yet to be realised and viable alternatives remain elusive (Page and Amory, 2018). Developing a drug for any purpose is difficult but one that blocks sperm function is a particular challenge as spermatozoa don’t divide, transcribe or translate, have specialized organelles, are highly motile and continually produced in large numbers (~100 million per day). Moreover, as the production of fertilization competent cells is necessarily different between species, it is essential to use a human system early in the drug discovery cascade to increase the chances of translational success. The spermatozoa’s sole function post-ejaculation is to swim until it closes in on the oocyte. The spermatozoa are then able to undergo a unique exocytotic process called the acrosome reaction (AR) which facilitates penetration of the zona pellucida, the oocyte’s outer glycoprotein shell and allows fusion with the oocyte’s plasma membrane, delivering its DNA. Hence, a male contraceptive could work in a number of different ways – it could completely halt spermatogenesis, block an essential function of the spermatozoon (such a motility) or perturb a cellular process that results in a non-productive interaction with the oocyte (such as prematurely triggering the acrosome reaction, and/or by blocking exocytosis or inhibiting receptor binding). Low sperm motility is a common cause of male infertility (Samplaski et al., 2019) indicating that targeting this process should yield effective drugs. The options for delivery of a contraceptive are wide ranging, but our primary aim is to develop a relatively fast-acting safe drug, self-administered by the male, that specifically and effectively irreversibly blocks the sperm’s ability to get to the egg and/or fertilise it.

Though some potential sperm-specific contraceptive drug targets, such as the calcium channel CatSper, have been identified (Lishko and Mannowetz, 2018), our relatively poor understanding of spermatozoon biology makes target-agnostic, or phenotypic strategies more attractive (Barratt et al., 2017). Phenotypic drug discovery is undergoing a renaissance in a number of therapeutic areas because it can uncover novel biology in an unbiased way. Retrospective analysis of drug approvals shows the approach to be more successful in first-in-class medicine discovery than previously appreciated (Swinney and Anthony, 2011; Moffat et al., 2017). However, for male contraceptive drug discovery, a limiting barrier has been the absence of a scalable drug screening system using spermatozoa.

To address this unmet need we set out to develop the first high-throughput phenotypic screening platform for sperm where an imaging-based kinetic analysis module that tracks motility, is followed by a flow cytometry -based assay module that detects both the degree of acrosome reaction (AR) and cell viability - both modules coordinated by a fully automated robotic system. The bespoke computational pipeline quantitatively evaluates each compound’s effect on both motility and AR (and viability) thus achieving unprecedented throughput. The platform was used to identify motility inhibitors from a unique 12,000 molecule ReFRAME (Repurposing, Focused Rescue, and Accelerated Medchem) library, which represents the most comprehensive collection of drugs available, as it contains nearly all the small molecules that have achieved regulatory approval and others that have undergone varying degrees of clinical development or have had significant preclinical profiling (Janes et al., 2018). This advance opens up the possibility of accelerated male contraceptive development by allowing drug repurposing, target identification as well as screening of chemical diversity libraries for novel medicinal chemistry start points.

Results

Phenotypic assay development

Our automated platform (Figure 1A) allows for the identification of relatively fast-acting compounds with a rapid plate-processing time (~75 minutes/ 384-well plate), where the motility assay runs for ~30 min followed directly by the AR assay (~45 min). Automated cell tracking (Figure 1B and Figure 1—figure supplement 1A,B) using time-lapse images allowed kinetic parameter quantification on >200 cells/well (Figure 1C). The kinetics show good agreement with clinical systems (Figure 1—figure supplement 1C) and a comparable distribution was observed between experimental days, plates and positions (Figure 1—figure supplement 2) with any small differences between days stemming from donor pool variance. Importantly, spermatozoa were found to be tolerant to the DMSO concentrations used in our screening program (0.0625%–0.1% Figure 1—figure supplement 1D). To determine the optimal screening batch size untreated spermatozoa were dispensed every 30 min for 150 min observing only a small decrease in motility (~10%) after four plates were screened (Figure 1—figure supplement 1E). A flow cytometry-based assay measuring AR (Figure 1D) was run directly afterwards (Figure 1E and Figure 1—figure supplement 1F).

Figure 1 with 2 supplements see all
Download asset Open asset
Phenotypic Assay workflows.

(A) Graphical summary of modular screening workflow where motility measurement is followed by acrosome reaction (AR) measurement allowing a screening throughput of >1400 compounds per donor pool (B) Steps in imaging and analysis: human sperm are recorded with brightfield illumination (first panel) then sperm heads detected using a particle tracking algorithm (second panel) which are then tracked across the timelapse series of images (third panel) and subsequently classified (fourth panel). Each panel contains a zoomed-in subsection of the field. Colour coding for tracking distance and kinetic classification is shown in the panel insets: Tracking panel - rainbow gradient (showing progression over time); Classification panel - red for immotile (IM), yellow for non-progressively motile (NPM) and blue for progressively motile (PM). Scale bars: 100 μm (main images), 5 μm (insets). (C) Sperm counts per well after microscopy and detection shown in a combined violin/box plot. Colours: purple violin outline (probability density of values), yellow dots (outliers of boxplot). (D) Graphical summary of the expected populations determined by flow cytometry based on distribution of cells measured with FL3-A (Pi) vertical axis and FL1-A (PNA-488) horizontal axis: dead cells (upper left, Pi+ PNA-); dead and acrosome-reacted (upper right, Pi+ PNA+); unstained/live/non-reacted (lower left) and live acrosome-reacted (lower right, Pi-PNA+). (E) Example flow cytometry data comparing sperm treated with the Ca2+ ionophore (A23187) which induces AR (left panel) with sperm from DMSO-treated well (right panel). Colours indicate event density. (F) Combined violin/box plot data showing flow cytometry event counts per well. Colours and label as in (C).

Screening of the ReFRAME library, confirmation of hits by dose response

The ReFRAME library (11,968 compounds) was supplied in ‘assay-ready’ imaging plates and was solubilised prior to addition of sperm. High assay quality was achieved throughout with a Z’-factor between 0.4–0.8 (Figure 2B). A total of 63 putative hit compounds decreased motility to varying degrees based on a hit selection cut-off of >15% reduction in motility (Figure 2A with examples in Figure 2C,D,F,G and Video 1 and 2). 14 compounds were selected as putative AR+ hits (Figure 3A with examples in B, C). Motility hits were confirmed with resupplied material in dose-response experiments. There was a dose-dependent decrease in motility for 29 compounds (46% hit confirmation rate), with EC50 values as low as 0.05 µM and effect sizes ranging from 15–100% (Figure 2F,G; Figure 2—source data 1, Supplementary file 1). Amongst the confirmed hits were the aldehyde dehydrogenase inhibitor, Disulfiram (70% maximum reduction at 10 µM) and a putative platelet aggregation inhibitor, KF-4939 (100% max. reduction; EC50 = 0.49 μM), and a range of other compounds having very modest effect and/or showing low potency (See Figure 2—source data 1, Supplementary file 1). In order to test reversibility, wash out experiments were performed. Only a small recovery of motility was observed, after wash out of Disulfiram (Figure 2—figure supplement 1). In the AR screen, 9 of the resupplied compounds had a dose-dependent effect with EC50 values as low as 0.4 µM (See Figure 3—source data 2, Supplementary file 2). However, following the orthogonal assay triaging we implemented to eliminate assay interference compounds, none of the AR hits were found to be true positives (see examples in Figure 3—figure supplement 1D) despite the primary screen showing good assay robustness.

Figure 2 with 1 supplement see all
Download asset Open asset
ReFRAME Library Screening: Motility Assay Results.

(A) Primary screening results of the motility assay. Each dot represents a well (either compound or control well) showing % of DMSO control (Curvilinear Velocity: VCL). Positive controls (Pristimerin), negative controls (DMSO) and individual compound datapoints are shown. The dashed line showing the 15% of control (reduction in VCL) – the cut-off for primary hit selection. Total number of compounds = 11,903, excluding wells with auto-focus errors and with ‘sticky’ compounds which have been excluded from analysis. See dataset (http://doi.org/10.5061/dryad.jdfn2z36z) Figure 2—source code 1. (B) Assay robustness: the standard high throughput screening metric, Z' (see Materials and methods; Zhang et al., 1999) was used to determine the performance of the assay for all screening plates. Dashed lines indicate min/max Z' values. (C) Tracking data visualizations of 3 example wells showing sperm tracks of both imaging positions (Position 1 [Pos. 1] and Position 2 [Pos. 2] respectively) within the wells. A DMSO control well (left panels ‘Control’) shows a large number of progressively motile (PM) sperm (yellow) with few non-progressively motile (NPM) sperm (green and very few immotile (IM) sperm (purple) – this is in contrast to the shorter tracks and higher levels of NPM and IM in the middle panel (‘20% reduction’) for a compound that shows 20% inhibition of motility (i.e. 80% of control) and the right hand panel (‘80% reduction’) for a compound showing 80% inhibition (i.e. 20% of control), showing almost all cells are in the IM and NPM classes. (D) Histogram of sperm tracks quantification of the data show in (C). (E) Summary of motility assay hit rate (0.5%) and reconfirmation rate (0.24%). (F-G) Dose response confirmation of two hits. 8-point 3-fold dilution curves are shown with 10 μM as the highest concentration. Two data points per concentration (n = 2, data point is mean ± SD). Each curve is a 4-parameter logistic fit. Each plot shows estimated values Hill Slope and EC50. The chemical structure of the hit compound is shown as well as some annotation and physicochemical properties. Physicochemical properties were calculated using RDKit, Python and KNIME: SlogP = partition coefficient (Wildman and Crippen, 1999); TPSA is the Topological Polar Surface Area (Ertl et al., 2000); MW is the exact Molecular weight; QED = Quantitative Estimate of Drug-likeness (Bickerton et al., 2012).

Figure 2—source code 1

R Code for Figure 2 primary motility assay.

https://cdn.elifesciences.org/articles/51739/elife-51739-fig2-code1-v2.r
Figure 2—source code 2

R Code for Figure 2—source data 1 dose response confirmation motility assay.

https://cdn.elifesciences.org/articles/51739/elife-51739-fig2-code2-v2.r
Figure 2—source data 1

Dose response and additional data for primary motility hits.

Compound names and nominal target/functional annotation shown on left along with chemical structure and physico-chemical properties. The 8-point dose response curves for each hit with estimated Hill slope, EC50 and Efficacy [% max reduction] are shown on the right. Two data points per concentration (n = 2); data points are Mean ± SD. A 4-parameter logistic fit has been performed using R package: dr4pl. Calculated EC50 values and Efficacy (% max effect) in the motility assay are shown in the right hand section. CC50 values for HEK293 and HepG2 CellTiter-Glo cytotoxicity assays performed by CALIBR (www.reframedb.org) are shown in the right hand section. Note: 0 = inactive in cytotoxicity assay. The chemical structure of the hit compound is shown as well as some annotation and physicochemical properties (for nomenclature description see Figure 2). See Figure 2—source data 2 along with Figure 2—source code 2.

https://cdn.elifesciences.org/articles/51739/elife-51739-fig2-data1-v2.pdf
Figure 2—source data 2

Dose response confirmation data motility assay.

https://cdn.elifesciences.org/articles/51739/elife-51739-fig2-data2-v2.csv
Figure 3 with 1 supplement see all
Download asset Open asset
ReFRAME library screening: ar assay results.

(A) Results of primary screening of the library using the acrosome assay (live cells, acrosome reacting, Pi- PNA+ population). Each dot represents a well (either compound or control well). Shown is % Events (number of events in Pi- PNA+ gate relative to total events in the sampled well). Datapoints for compounds, negative controls (DMSO) and positive controls (A23187) are shown. Black dashed line = 15% threshold for primary hits selection. See Figure 3—source data 1 along with Figure 3—source code 1. (B-C) Data from two example wells: a DMSO well (left panel) and a well with 65% Pi- PNA+ population (right panel). (D) Summary of acrosome assay results before and after triage.

Figure 3—source code 1

R Code for Figure 2 primary acrosome assay.

https://cdn.elifesciences.org/articles/51739/elife-51739-fig3-code1-v2.r
Figure 3—source code 2

R Code for Figure 3—source data 2 dose response confirmation acrosome assay.

https://cdn.elifesciences.org/articles/51739/elife-51739-fig3-code2-v2.r
Figure 3—source data 1

Primary screening data acrosome assay.

https://cdn.elifesciences.org/articles/51739/elife-51739-fig3-data1-v2.csv
Figure 3—source data 2

Dose response and additional data for primary AR hits.

Compound names and nominal target/functional annotation shown on left along with chemical structure and physico-chemical properties. The 8-point dose response curves for each hit with estimated Hill slope, EC50 and Efficacy [% max induction] values are shown on the right. Two data points per concentration (n = 2); data points are Mean ± SD. A 4-parameter logistic fit has been performed using R package: dr4pl. CC50 values for HEK293 and HepG2 CellTiter-Glo cytotoxicity assays performed by CALIBR (www.reframedb.org) are shown in the right hand section. Note: 0 = inactive in cytotoxicity assay. Physicochemical properties were calculated using RDKit, Python and KNIME: SlogP = partition coefficient (Wildman and Crippen, 1999); TPSA is the Topological Polar Surface Area (Ertl et al., 2000); MW is the exact Molecular weight; QED = Quantitative Estimate of Drug-likeness (Bickerton et al., 2012). See Figure 3—source data 3 along with Figure 3—source code 2 .

https://cdn.elifesciences.org/articles/51739/elife-51739-fig3-data2-v2.pdf
Figure 3—source data 3

Dose response confirmation data acrosome assay.

https://cdn.elifesciences.org/articles/51739/elife-51739-fig3-data3-v2.csv
Video 1
Download asset
Movie was generated using a brightfield image sequence for a typical control well (left pane), a well where a compound reduced motility by 20% (middle pane) and one where it reduced it by 80% (right pane).
Video 2
Download asset
Movie was generated using a brightfield image sequence for a typical control well (left pane), a well where a compound reduced motility by 20% (middle pane) and one where it reduced it by 80% (right pane).

Tracking is overlayed in each panel. Colour coding is detailed in Figure 2.

Discussion

A major barrier in the search for new male contraceptives has been the lack of an effective high throughput phenotypic screening system. This study presents the first platform based on imaging and flow cytometry that measures two fundamental aspects of sperm behaviour – motility and AR, allowing for the rapid screening of compound collections.

A significant challenge was to create a system for the tracking of these highly motile cells that was fully automated and scalable to allow throughput at sufficient speed. Moreover, concomitant assessment of a second functional attribute, AR along with cell viability was developed in order to maximise use of the biological material. Assessment of motility required novel implementation of tracking algorithms. Importantly, the kinetic outputs were equivalent to those from the ‘industry-standard’ CASA, which is very low throughput and unsuitable for screening more than a few compounds at one time. Significant workflow optimisation resulted in good assay robustness and an acceptable hit confirmation rate following the primary screen using a medium stringency cut off for hit selection (>15% reduction). For AR detection, a flow cytometry approach was taken based on established methods (Mortimer et al., 1987) and showed good assay robustness. However, the presence of assay-interfering compounds in screening libraries (potential fluorescent compounds including DNA intercalators) necessitated further triaging steps (see Materials and methods) to eliminate false positives. This combination of triaging approaches meant that, in this screen, no AR compounds were suitable for progression. An improvement to the AR primary assay could be to use a far-red labelled lectin reagent to detect the glycoproteins. Other variants of the assay would be to either use sperm that had been capacitated, a state of increased readiness to undergo the AR process or use it to find drugs that would block the AR induced by a physiological stimulus. These latter variants are worthy of future investigation.

It is worth noting that whilst the primary goal of this programme is to identify compounds that will act as contraceptive by decreasing motility we detect a larger number of putative motility upregulating hits. These compounds are of interest for not only target identification and uncovering novel sperm biology but also their potential to improve fertility treatments. Overall, although the two assays presented here are fairly complex biological assay using primary cells, it achieved acceptable throughput with the potential to increase batch size for larger screening campaigns.

The advantage of using the ReFRAME collection was the potential to repurpose drugs that are already approved for other indications, or are at earlier stages of clinical progression. The associated compound annotations and safety data offers the potential to help accelerate the development of an effective male contraceptive. In the motility assay the hit rate of ~0.2% is typical for a cell-based assay, however surprisingly a relatively high number of hits were not deemed suitable for compound progression. Examples included potentially toxic mercury-containing compounds (phenylmercuric borate and mercufenol chloride); antiseptics and antibiotics including tyrothricin, or compounds with impacts on fundamental biology for example the microtubule stabiliser and chemotherapeutic, docetaxel. However, several hits are potentially of interest. One is Disulfiram (70% reduction at 10 μM) a drug that is well tolerated, has long been used for treating alcohol dependency and has previously been shown to inhibit sperm motility (Lal et al., 2016). Disulfiram is unlikely to be developable as a suitable contraceptive drug due to its side-effects when alcohol is consumed but has potential to facilitate mode-of-action studies. One recent report identified 3-Phosphoglycerate dehydrogenase (PHGDH) as a target for disulfiram (Spillier et al., 2019) whilst another identified NLP4, a subunit of VCP/p97 segregase (Skrott et al., 2017; Skrott et al., 2019) although the relevance of either to sperm function is unclear. The identification of KF-4939, a putative platelet aggregation factor (PAF) inhibitor, as a potent hit (100% max. reduction EC50 = 0.5 µM) may also be of interest since PAF has been used to improve sperm function in male infertility treatments (Roudebush et al., 2004). It is noteworthy that like disulfiram, KF-4939 also contains a disulphide linkage indicating that its biological effect maybe mediated by modifying sulfhydryls on proteins (Yamada et al., 1985). Further thiol-containing compounds, bismuth ethanedithiol (98% max reduction; EC50 = 2.5 μM) and the anti-fungal benzothiazole, Ticlatone (96% max. reduction EC50 = 4.4 µM) were also potent hits, suggesting free cysteines in, as yet, unidentified protein(s), may play a critical role in motility. Of interest is the observation that the Toll-like receptor 7/8 ligand (Resquimod; R848), recently shown to preferentially reduce the motility of X chromosome-bearing mouse sperm by suppressing ATP production (Umehara et al., 2019), was a hit, demonstrating that this phenotypic screening approach has the potential to uncover new aspects of sperm biology.

It was somewhat surprising that relatively few drugs or drug-like molecules in the ReFRAME collection were potent and effective inhibitors in the motility assay. This may reflect the unique biology of the spermatozoon that has evolved for a singular purpose – to reach the oocyte and fertilise it and/or that the targets within pharma’s therapeutic areas of interest over the last few decades are not similarly functional in human sperm. Conversely however, we did observe a large number of compounds (Figure 2A) that had a significant positive effect on sperm motility which may be target agonists as well as potential start points for treating men with low sperm motility. The development of an effective safe male contraceptive is challenging because a prerequisite, other than the obvious need for safety, is a level of reversibility. Other possibilities, such as a drug taken by the female or an agent for use as a topical spermicide (for example to replace nonoxynol-9) are options to consider in the future and depend to a large extent on the pharmacokinetic properties of compounds we identify and their mode of action – for example some may have poor bioavailability.

In summary, our high-throughput phenotypic platform allows for the screening of bioactives including natural products, focussed small molecule libraries and chemical diversity sets, which can provide novel start-points for male contraceptive drug development, as well as the identification of new drug targets in human spermatozoa.

Materials and methods

Key resources table
Reagent type
(species) or resource		DesignationSource or referenceIdentifiersAdditional
information
Biological sample (Homo sapien)Live spermatozoaDonated semen samplesLocal ethical approval (13/ES/0091)
AntibodyLectin PNA (Arachis hypogaea), Alexa Fluor 448 ConjugateThermoFisher ScientificThermoFisher:L21409;
RRID: AB_2315178		Stored as 1 mg/mL stock in DMSO; used at 1:1000 final dilution
Commercial assay or kitPropidium Iodide; Live/Dead Sperm Viability kit,ThermoFisher ScientificThermoFisher:L7011,Stored as 2.4 mM stock; used at a 1:2000 final dilution
Chemical compound, drugReFRAME (Repurposing, Focused Rescue, and Accelerated Medchem) libraryCALIBR at the Scripps Institute; Publication (Janes et al., 2018)www.reframedb.org
Chemical compound, drugPristimerinMerckMerck:530070Stored as 10 mM stock in DMSO; used at final concentration of 20 µM
Chemical compound, drugcalcium ionophore (A23187)Sigma-AldrichSigma-Aldrich:C7522Stored as 10 mM stock in DMSO, used at a final concentration of 10 µM)
Chemical compound, drugDisulfiramTocrisTocris:3807Stored as 10 mM stock in DMSO, used at 10 μM final concentration
Software, algorithmTrackpy v0.4.1Zenodo. (http://doi.org/10.5281/zenodo.1226458)Publication: (Crocker and Grier, 1996); Publication: (Allan, 2018)
Software, algorithmFFMPEGFFmpeg Developers (http://ffmpeg.org)RRID:SCR_016075
Software, algorithmBioconductor packagesBioconductor (https://bioconductor.org)RRID:SCR_006442flowCore, flowDensity, flowWorkspace, ggcyto
Software, algorithmHDF5HDF Group (www.hdfgroup.org)
Software, algorithmdr4plDr4pl (https://cran.r-project.org/web/packages/dr4pl/index.html)
Software, algorithmCode used for data analysisThis paperThe R code used for data analysis is included in the supplement files accompanying this paper
Software,
algorithm		KNIMEBerthold et al., 2008https://www.knime.com
Software,
algorithm		RDKitRDKit, 2018https://www.rdkit.org

Ethical approval

Request a detailed protocol

Written consent was obtained from each donor in accordance with the Human Fertilization and Embryology Authority (HFEA) Code of Practice (version 8) under local ethical approval (13/ES/0091) from the Tayside Committee of Medical Research Ethics B.

Development of methods for motility and Acrosome Reaction (AR)

Request a detailed protocol

A pre-existing automated cell-based phenotypic screening platform was adapted with the use of live human spermatozoa. The platform utilises a Yokogawa CV7000 Cell Voyager high-throughput microscope able to image 384 multiwell plates under full environmental control acquiring timelapse images at a sufficiently high speed (~48 fps)_to capture the fast-moving sperm for tracking analysis. For high throughput screening with human spermatozoa to be effective and yield meaningful hits, the system needed to replicate, as close as possible, the manual workflow currently used in an andrology lab, so that the effect of large numbers of compounds on viable human sperm could be performed. In order to achieve this goal, we investigated a number of factors that could impact the quality of the data such as: plate type; concentration and number of spermatozoa per well; logistics of spermatozoa preparation by density gradient centrifugation and handling; donor-pooling; dispense speed for getting spermatozoa into 384 well plates; timing of compound addition; temperature control; image acquisition parameters; algorithm choice for motility assessment and data management and, evaluating the overall speed of data analysis. In addition, a flow cytometry-based assay and data analysis workflow was developed to measure acrosomal status and cell viability. We chose to perform all experiments on non-capacitated sperm as this condition is the most consistent, less prone to donor-effects and is closer to the state in which the sperm will be when they would be encountering the male contraceptive drug. The goal was to be able to screen a 384-well plate in both assays consecutively within 90 min of sperm dispensing using the fully automatic robotic system. Data was analysed for inter- and intra-plate variation (both technical and biological replicates) and signal-to-background and variance measured to demonstrate the assay was reproducible and robust. The assay was shown to be scalable and all the workflows optimal. Once established, the system was validated by screening the ReFRAME library (Janes et al., 2018) and hits confirmed by dose-response analysis.

Sperm handling

Request a detailed protocol

Control semen samples were obtained from volunteer donors with normal sperm concentration, motility and semen characteristics (WHO, 2010) and no known fertility problems. Samples were obtained by masturbation after 48–72 hr of sexual abstinence. After ~ 30 min of liquefaction at 37°C, cells were prepared using a discontinuous density gradient procedure (DGC). Semen was loaded on top of a 40–80% suspension of Percoll (Sigma Aldrich, UK) diluted with non-capacitation medium using Minimal Essential Medium Eagle, supplemented with HEPES, Sodium lactate and Sodium Pyruvate to achieve a similar buffer as described previously (Tardif et al., 2014). We routinely pooled samples after DGC from 3 to 5 donors for use in each screening batch run in order to reduce donor-to-donor variability. Samples were obtained and analysed in line with suggested guidance for human semen studies where appropriate (Björndahl et al., 2016).

Motility assay

Request a detailed protocol

Sperm cells were incubated for 3 hr at 37°C after DGC, transferred to the robotic platform (HighRes Biosolutions Inc) and maintained at 37°C with gentle stirring (100 rpm) in a water bath on top of a magnetic stirrer. Approximately 10,000 spermatozoa (10 μl) were dispensed per well using a MultiDrop Combi (ThermoFisher) into pre-warmed assay-ready plates. Spermatozoa were incubated with the compounds for 10 mins (thus favouring fast acting compounds) at 37°C in the CV7000 microscope prior to the commencement of imaging.

Time-lapse imaging

Request a detailed protocol

A CV7000 Cell Voyager high-content imaging system (Yokogawa) was used as it allowed full environmental control, sufficient contrast using simple brightfield optics and a fast acquisition rate (up to 45 frames per second). Using a 20x lens with bright field illumination (0.11 ms exposure, 3% lamp power), time-lapse image series were acquired (24 frame in 0.5 s) at two positions per well (with a 400 μm gap between positions to eliminate double counting of sperm). Image acquisition across a 384-well plate with such settings takes ~ 17 min.

Sperm tracking

Request a detailed protocol

A Python implementation of the particle tracking algorithm originally developed by others (Crocker and Grier, 1996) was employed (Trackpy v0.4.1 Allan, 2018). The algorithm determines the position of every sperm head in each frame and links the position over time creating tracks. We optimized the algorithm parameters for our imaging data to exclude sperm cross-tracks and avoid detecting non-sperm particles. The image data (~17,000 files amounting to 21 GB per plate) can be processed within 30 min using a standard desktop PC (Intel Core i5-6600 CPU, 3.3 GHz, 8 Gb RAM) and can also be parallelized on a compute cluster with minimal effort.

Acrosome assay

Request a detailed protocol

A flow cytometry-based assay was developed employing an iQue Screener (Sartorius) with Alexa488-conjugated peanut agglutinin (PNA) (Mortimer et al., 1987) and general sperm viability (using propidium iodide, Pi). This assay was run directly after motility assessment. Controls (DMSO and the Ca2+ ionophore A23187) were added using an acoustic dispenser (Echo 555, Labcyte Inc) and incubated for 10 min at 37°C in the SteriStore (HighRes Biosolutions Inc). PNA488 (ThermoFisher Scientific, Cat. No. L21409, stored as 1 mg/mL stock) was then added to achieve a 1:1000 final dilution and propidium iodide (ThermoFisher Scientific, Live/Dead Sperm Viability kit, Cat. No. L7011, stored as 2.4 mM solution) was added to achieve a 1:2000 final dilution using a liquid handler (Tempest, Formulatrix). After addition of dyes, plates were incubated for 10 min at 37°C in the SteriStore before sampling using an iQue Screener (2 s sip time per well with pump setting of 45 rpm). This resulted in a sample processing time of about 25 min per 384-well plate. Cells were categorised as either: unstained cells; ‘dead’ cells (Pi+ PNA-); Acrosome-reacted but ‘dead’ cells (Pi+ PNA+); or Acrosome-reacted ‘live’ cells (Pi- PNA+). This method is capable of detecting a shift of populations (Pi+ and PNA+) upon induction of AR using A23187. The AR assay was performed as an agonist screen (scoring for induction of AR in live cells compared to DMSO controls) with the expectation that compounds could be found that induce AR beyond levels of induction with A23187.

Controls and QC criteria

Request a detailed protocol

For the motility assay DMSO was used as a vehicle control (negative control) and Pristimerin (Merck, Cat. No. 530070, stored as 10 mM stock in DMSO; final concentration of 20 µM) as a positive control. For the acrosome assay we added a calcium ionophore (A23187, Sigma-Aldrich, Cat. No. C7522, stored as 10 mM stock in DMSO, used at a final concentration of 10 µM) as a positive control and DMSO was added to wells as a negative control to be at the same final concentration as that present in the compound-containing wells (0.0625%). DMSO tolerance was tested by performing the standard motility assay exposing sperm to increasing concentrations (max 10%) of DMSO for 10 min. We calculate the standard statistical measure of assay robustness, the Z’ value (Zhang et al., 1999) for each plate using positive and negative controls for the motility assay and observed Z’ values ranging between 0.4–0.8. In addition, we performed a visual check of heatmaps for every plate to detect edge effects.

Data analysis and normalization

Motility

Request a detailed protocol

custom R scripts were written to calculate standard sperm kinetic parameters (Mortimer et al., 2015). Those parameters allow classification of sperm into standard WHO classes: progressively motile (PM) (where average path velocity: VAP > 25 um/sec AND straightness: STR > 80%); non-progressively motile (NPM) (where VAP > 5 um/sec OR straight line velocity: VSL > 11 um/sec), and immotile (IM). In addition to calculating kinetic parameters, we established a workflow to generate movies of time-lapse videos with overlapping sperm tracks using R and FFMPEG (FFmpeg Developers. Available from http://ffmpeg.org). We expressed results as % of control curvilinear velocity: VCL. This was defined as VCL_median (cpd)/VCL_median(DMSO)*100. Hit selection criteria was 15% reduction of VCL. VCL was chosen as the main kinetic measurement as it had an acceptable Z’ value and is independent of path averaging. Wells with autofocus errors or compounds which did not dissolve properly (resulting in sperm cells being stuck in one location but moving normally in the rest of the well) were excluded from analysis.

Acrosome

Request a detailed protocol

iQue Screener Data was exported as FCS format and processed using the following Bioconductor packages for analysing flow cytometry data: flowCore, flowDensity, flowWorkspace, ggcyto (see https://bioconductor.org). These packages allow handling flow cytometry data as objects, which can then be compensated, gated and visualized efficiently. In addition, we use HDF5 format (www.hdfgroup.org) for storing processed and averaged data. Percent events for each population (Pi+ PNA+, Pi- PNA+, Pi+ PNA- and Pi- PNA-), normalized to total well events, were calculated. Hits were defined as compounds which induce AR beyond 15% (maximum level of induction achieved with positive control). Wells with irregularities (low in Pi+ or Pi+PNA+ population, or below 200 events) have been excluded from analysis.

Compound screening

Request a detailed protocol

We screened the ~12,000 compound ReFRAME (Repurposing, Focused Rescue, and Accelerated Medchem) library (Janes et al., 2018; and www.reframedb.org) supplied by CALIBR at the Scripps Institute. This unique library consists of bioactives and approved drugs that have been assembled from the literature, drug databases and by patent mining. Compounds dissolved in DMSO were spotted (12.5 nL, final assay concentration ~ 6 μM) into 384-well black-sided optical imaging plates (CellCarrier, PerkinElmer) at CALIBR and were shipped on dry ice to the screening centre in Dundee and stored at −20 °C until required. Immediately prior to screening plates were thawed, controls were added using the acoustic dispenser (Echo 555, Labcyte), and 10 μl warm media added followed by shaking on a plate shaker for 10 s. Plates were then incubated for 30 min at 37°C to solubilise the compounds and prewarm the plates prior to the addition of live sperm (see above). With the addition of the media containing sperm (10 μl) the final concentration of DMSO in each well was 0.0625%.

All experiments were performed with pools of donors (3–5 per batch of 4 screening plates) and results normalised to DMSO-treated wells (16 wells per plate). Sperm was first prepared by DGC and then pooled. Where a single donor sample was found to less than 1/3 of the yield required then the contribution from the other doors was increased to compensate.

The ReFrame library has been screened for cytotox (using CellTiterGlo) in HEK293 and HepG2 cells (see www.reframedb.org) with some of the hit compounds displaying activity in one or both assays (Figure 2—source data 1), information which helps inform compound progression decisions.

Dose response experiments

Request a detailed protocol

Assay-ready 384-well CellCarrier plates with 8-point curves (10 μM highest concentration) were resupplied by CALIBR in duplicate and processed as described above. Curves were fitted using the R package dr4pl (https://cran.r-project.org/web/packages/dr4pl/index.html) using a 4-parameter logistic fit option.

Triaging of AR hits

Request a detailed protocol

Given the prevalence of a range of interfering compounds amongst the library hits, two approaches were used for triaging. These additional triaging steps, after the primary screen, allowed for the elimination of intrinsically fluorescent compounds and compounds fluorescing in the presence of biological material. Screening plates were excited at 488 nm and read at 520 nm and 670 nm in a M1000Pro multimode plate reader (Tecan) after flow cytometry. In addition, when performing dose-response curve follow up on resupply material a replicate set of compound-dosed cells were processed with no-dyes added. This identified the small number of compounds whose fluorescence is indistinguishable from a true positive. In addition, visual confirmation of the appropriate pattern of acrosome staining was performed by microscopy.

Disulfiram washout assay

Request a detailed protocol

In order to assess the ability of sperm cells to recover motility after Disulfiram removal, wash-out experiments were conducted at two timepoints using non-capacitating conditions. Spermatozoa were pooled after DGC preparation in the same manner as for screening. Pooled samples were diluted to a concentration of approximately 106/mL and incubated at 37°C under non-capacitating conditions (see above). After this initial incubation period, Disulfiram (10 μM final concentration) or vehicle control (DMSO to 0.1% final concentration) were added and incubated at 37°C for 20 min. Following exposure to Disulfiram or DMSO control, sperm cells were washed by centrifugation at 300 g for 5 min in 5 mL of the non-capacitation media. After washing, the pellet was resuspended in 500 μl of the appropriate media and incubated at 37°C for 1 hr. Samples were taken for analysis by computer-assisted-sperm analysis (CASA) [CEROS machine (version 12), Hamilton Thorne Research, Beverly, MA, USA] before treatment, after 20 mins treatment and 1 hr post washout. Wash-out experiments were performed on three separated days, utilising different donors to make up the pooled sample. Each condition was repeated in duplicate on each of these days and each CASA reading was performed in duplicate.

Data availability

Full data is available. Large files have been deposited to Dryad (http://doi.org/10.5061/dryad.jdfn2z36z).

The following data sets were generated
    1. Franz S Gruber
    2. Zoe C Johnston
    3. Christopher LR Barratt
    4. Paul D Andrews
    (2019) Dryad Digital Repository
    Data from: A phenotypic screening platform utilising human spermatozoa identifies compounds with contraceptive activity.
    https://doi.org/10.5061/dryad.jdfn2z36z

References

    1. Berthold MR
    2. Cebron N
    3. Dill F
    4. Gabriel TR
    5. Kötter T
    6. Meinl T
    7. Ohl P
    8. Sieb C
    9. Thiel K
    10. Wiswedel B
    (2008)
    Data Analysis, Machine Learning and Applications
    319–326, KNIME: The Konstanz Information Miner, Data Analysis, Machine Learning and Applications, Springer, 10.1007/978-3-540-78246-9_38.
    1. RDKit
    (2018) RDKit: Open-Source Cheminformatics and Machine Learning
    RDKit: Open-Source Cheminformatics and Machine Learning, http://www.rdkit.org.
    http://www.rdkit.org
  28. Book
    1. WHO
    (2010)
    WHO Laboratory Manual for the Examination and Processing of Human Semen (Fifth)
    Cambridge: Cambridge University Press.

Decision letter

  1. Polina V Lishko
    Reviewing Editor; University of California, Berkeley, United States
  2. Anna Akhmanova
    Senior Editor; Utrecht University, Netherlands
  3. Polina V Lishko
    Reviewer; University of California, Berkeley, United States
  4. Jean-Ju Chung
    Reviewer; Yale School of Medicine, United States

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

All reviewers agreed that this manuscript reports an important advancement in the development of the tools aimed to provide an unbiased search for novel contraceptives. The authors describe novel drug screening platform to evaluate the libraries of the compounds for their ability to inhibit sperm motility and sperm fertility. The authors demonstrate that such a platform was used successfully to screen for small inhibitory molecules that could potentially serve as templates for further development of novel contraceptive drugs. Thus, this work can have a very high practical impact.

Decision letter after peer review:

Thank you for submitting your article "Phenotypic screening platform utilising human sperm for the identification of new compounds with contraceptive activity" for consideration by eLife. Your article has been reviewed by three peer reviewers, including Polina V Lishko as the Reviewing Editor and Reviewer #1, and the evaluation has been overseen by Anna Akhmanova as the Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Jean-Ju Chung (Reviewer #2).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission. Our goal is to provide the essential revision requirements as a single set of instructions, so that you have a clear view of the revisions that are necessary for us to publish your work.

Summary:

The reviewers find that the manuscript represents a highly important advancement in the unbiased search for male-based contraceptives. The authors describe the development of a powerful drug screening platform to evaluate compounds for their ability to inhibit sperm motility and sperm fertility. The authors demonstrate that the motility screening platform was used successfully to screen out small inhibitory molecules that could potentially serve as templates for further development of novel contraceptive drugs. However, the small number of hits from the flow-cytometry based AR assays consecutive to the motility screening turn out to be all negative, indicating the system for this aspect remains to be further improved. Overall, the design and details of the current screening platforms are nice additions to the field and the properties of the candidate lead molecules for the motility inhibition will serve a useful foundation for the further endeavor to develop non-hormonal male contraceptives. The data provided are of good quality and the statistical analyses are adequate, while some additional controls will be needed (see below).

Central conclusions:

1) The development of male contraceptives addresses a critical gap in the contraceptive portfolio. Namely, the field is in critical need of a high-throughput system that is robust to variations between men, and to variations within and between plates/trials.

2) The manuscript describes a high-throughput drug-screening platform that can be used to target sperm motility and the sperm acrosome reaction (AR) with smart design of utilizing 12,000 molecules from ReFRAME repurposing library directly applied to human sperm cells from fertile donors.

3) The use of ReFRAME allows both productive uses of compounds that have already undergone lengthy/expensive testing. Of equal importance, it enables compounds to be identified as hits known to affect sperm function or would be predicted to do so.

4) Identification of disulfiram and KF-4939 provide evidence of this important attribute. Other strong positives include the long list of variables they tested to ensure the robustness of the assay.

Overall, the authors provide valuable data validating the approach and showing representative data and are to be commended for revealing the full lists of their hits.

Essential revisions:

The reviewers raise a number of concerns that must be adequately addressed before the paper can be accepted. Some of the required revisions will likely require further experimentation within the framework of the presented studies and techniques.

1) It would be helpful to explain from the beginning the rationale of the assay design and why this assay is focused primarily on the sperm motility and acrosome reaction. To appeal to a scientifically diverse cohort of eLife readers, and particularly to people who are not experts in male fertility, the proper explanation and brief introduction to sperm biology is needed.

2) Subsection “Phenotypic Assay Development” and Figure 1—figure supplement 1D: The authors stated that 'spermatozoa were found to be tolerant to DMSO up to 4%.' As the authors note, it is far above typical screening concentration 0.1% DMSO and surprising. Reviewers noted, that to their knowledge, human sperm cells are quite sensitive to DMSO. It would be nice to show the time point of measuring the motility in Figure 1—figure supplement 1D legend. Is the time-course effect of DMSO on sperm motility was tested at given concentrations? It is also recommended to show that the effect of DMSO concentration on the viability (and/or membrane integrity by PI permeability) and its own effect on the AR state of spermatozoa up to 4% DMSO to support the conclusion. Depending on the results, the reviewers recommend that the authors state the sperm tolerance specifically limited to motility or tone down. Additionally, it is needed to see DMSO alone controls on multiple men, and a couple of times, over that full range.

3) The initial screening was performed at ~ 6 μm final concentration and it would be nice to state the concentration of final DMSO concentration per each well to achieve ~ 6 μm and the final concentration of DMSO in the DMSO-treated wells.

4) Figure 2A clearly shows that primary screening results of the motility assay with the ReFRAME library had motility boost hits even if a lot more stringent cut-off is applied. While they were not of interest for a contraceptive development, it could have the potentials for infertility treatments. This reviewer suggests that the authors discuss these 'hits on the opposite side of the spectrum' at a minimum even if the authors decided not to show.

5) Figure 2C and D: do the 20% or 80% reduction in these panels mean the reduction of total motility (PM+NPM) of control? As total sperm counting per each group is different, it will be helpful to indicate the real number of sperm counted on each bar (IM, NPM, and PM) for the readers to easily grasp the calculation.

6) Subsection “Screening of the ReFRAME library, confirmation of hits by dose response” and regarding the negative results of the AR inducer screening: 29 hits were screened out by motility test from the initial 63 positives (~50%) after the dose-response test. In contrast, none of 9 potential candidates from the AR inducer screening were found to be true positives after following the orthogonal assay, which makes the reliability and/or robustness of the AR assays in the screening system in doubt. Can the authors explain the discrepancy in the discovery rate?

7) It would be nice that some alternative strategy is discussed to test the acrosome reaction that won't be hindered by fluorescent compounds. For example, different fluorophores with emissions at more extreme ends of the spectrum could be used, so as not to overlap with the fluorescence spectra of small molecules in the screen used for the flow cytometry. Of course, this will reduce the number of false-positive, but not necessarily the number of the initial hits.

8) Reversibility and less/no side effects are essential requirements for developing or screening male reproductive drugs. Can sperm motility be rescued upon drug elimination?

9) Were there any drugs in the library that are known to have impacts on sperm motility that weren't identified? Failure to detect such a compound might point to a limitation or area in which the assay could be improved.

10) Of the long list of issues that could affect the robustness of the approach, sperm handling is one of the most important. Could more information be provided in the subsection “Sperm handling”, particularly regarding the discontinuous density gradient? Were the samples from different men pooled prior to DGC, or were the washed sperm pooled after? If after, were the samples pooled using equal numbers of sperm from each man, to keep each man's contribution equal, or were all the sperm recovered from each man used? Increased information on this point could help reduce inter-assay variability.

11) Somewhere in the Discussion, authors should provide a paragraph on the potential immediate utility of this approach for identifying topical spermicides/contraceptives. The world does need an improvement over nonoxynol-9, which can increase transmission of STIs, particularly for commercial sex workers. It is worth mentioning that compounds that directly affect sperm function could then be used as a starting point to identify druggable targets that would impair that target during spermatogenesis. Without this discussion, you might be open to criticism that this approach would only work for topicals. There will, of course, be influences of metabolism, etc., but high throughput screening for targets is a novel and critically important contribution.

12) Acrosome reaction was used as one of the key parameters for screening, and yet the assessed samples of sperm cells were ejaculated spermatozoa and not the capacitated cells. According to common knowledge, sperm ability to undergo an acrosome reaction is achieved only after they are capacitated (in vivo or in vitro), and yet in this study, the capacitation was not performed (subsection “Sperm handling”). To clarify this confusion and make the text more digestible to readers, authors are encouraged to expand their very brief result and/or Discussion sections. Specifically, since the goal of the team was to search for male contraceptives, it is not clear whether the compounds are expected to target sperm while they are stored inside the male reproductive tract or post-ejaculated. As reviewers noted, the compounds that prematurely induce acrosome exocytosis could only serve as topical contraceptives. Whether in a condom, jelly, or foam, that would come into contact with only non-capacitated sperm. Premature induction of exocytosis would NOT be recommended as a means of killing the sperm inside male reproductive tract, because if they were in the rete or epididymis, that would result in a sperm granuloma and obstructive azoospermia. This might be a point worth mentioning in the Discussion.

https://doi.org/10.7554/eLife.51739.sa1

Author response

Essential revisions:

The reviewers raise a number of concerns that must be adequately addressed before the paper can be accepted. Some of the required revisions will likely require further experimentation within the framework of the presented studies and techniques.

1) It would be helpful to explain from the beginning the rationale of the assay design and why this assay is focused primarily on the sperm motility and acrosome reaction. To appeal to a scientifically diverse cohort of eLife readers, and particularly to people who are not experts in male fertility, the proper explanation and brief introduction to sperm biology is needed.

This is a very good suggestion – we acknowledge that in the desire to keep the manuscript short we neglected to set the scene and explain the logic behind our approach. We have thus expanded the second paragraph of the Introduction.

2) Subsection “Phenotypic Assay Development” and Figure 1—figure supplement 1D: The authors stated that 'spermatozoa were found to be tolerant to DMSO up to 4%.' As the authors note, it is far above typical screening concentration 0.1% DMSO and surprising. Reviewers noted, that to their knowledge, human sperm cells are quite sensitive to DMSO. It would be nice to show the time point of measuring the motility in Figure 1—figure supplement 1D legend. Is the time-course effect of DMSO on sperm motility was tested at given concentrations? It is also recommended to show that the effect of DMSO concentration on the viability (and/or membrane integrity by PI permeability) and its own effect on the AR state of spermatozoa up to 4% DMSO to support the conclusion. Depending on the results, the reviewers recommend that the authors state the sperm tolerance specifically limited to motility or tone down. Additionally, it is needed to see DMSO alone controls on multiple men, and a couple of times, over that full range.

We appreciate the reviewers concern about DMSO sensitivity on spermatozoa. We have reworded and added to the text e.g. subsections “Phenotypic Assay Development”, “Controls and QC criteria”, “Compound Screening”

and updated Figure 1—figure supplement 1D, to highlight the DMSO concentrations used in the screen and hit confirmation. We have also added data in Figure 1—figure supplement 1D for DMSO concentrations up to 10 μM, where we start to see an effect on sperm motility in our assay (10-30 min incubation with compound/DMSO). We did not want the reader to get the impression that spermatozoa are tolerant to high concentrations of DMSO, but wanted to point out that with the range of concentrations used in our assay (up to 0.1% DMSO) we did not see an effect caused by DMSO over the time course of screening. Further evidence to support this is provided by the newly added wash out experiments performed on three different donor pools using 20 min incubation of either 0.1% DMSO or 10 μM Disulfiram (new Figure 2—figure supplement 2).

3) The initial screening was performed at ~ 6 μm final concentration and it would be nice to state the concentration of final DMSO concentration per each well to achieve ~ 6 μm and the final concentration of DMSO in the DMSO-treated wells.

The final concentration of DMSO used in the primary screen was 0.0625% and was the same in both compound wells and control wells. This is detailed in the Materials and methods subsection “Controls and QC criteria” and also mentioned in the subsection “Compound Screening”.

4) Figure 2A clearly shows that primary screening results of the motility assay with the ReFRAME library had motility boost hits even if a lot more stringent cut-off is applied. While they were not of interest for a contraceptive development, it could have the potentials for infertility treatments. This reviewer suggests that the authors discuss these 'hits on the opposite side of the spectrum' at a minimum even if the authors decided not to show.

We agree it would be useful for the readers to appreciate the dual purpose. The characterisation of these hits was outside the scope of the contraceptive programme but clearly are of interest. We have added two sentences highlighting this in the Discussion (third and fifth paragraphs).

5) Figure 2C and D: do the 20% or 80% reduction in these panels mean the reduction of total motility (PM+NPM) of control? As total sperm counting per each group is different, it will be helpful to indicate the real number of sperm counted on each bar (IM, NPM, and PM) for the readers to easily grasp the calculation.

We have adjusted Figure 2D to include the numbers. The percentage values shown were indeed total motility expressed as% of control. For this we use a median VCL value of all spermatozoa tracked over two positions in each well. This median is normalized to a median VCL value of DMSO control wells. Classifying sperm tracks by WHO definitions, we can then count each group and see what changes occur upon compound addition e.g. 20% reduction shows an increase in immotile and non-progressively motile classified cells, while progressively motile-classified cells are reduced in number.

6) Subsection “Screening of the ReFRAME library, confirmation of hits by dose response” and regarding the negative results of the AR inducer screening: 29 hits were screened out by motility test from the initial 63 positives (~50%) after the dose-response test. In contrast, none of 9 potential candidates from the AR inducer screening were found to be true positives after following the orthogonal assay, which makes the reliability and/or robustness of the AR assays in the screening system in doubt. Can the authors explain the discrepancy in the discovery rate?

For the motility data we took a fairly low cut off (>15% decrease in motility) in the primary screen to ensure we identified compounds with a range of effect sizes and thus more start points for medicinal chemists to evaluate. Inevitably, some of those weakly downregulating compounds (~50%) did not confirm in dose-response experiments.

For the AR assay we chose a hit cut off based on the median effect size of our positive control A23187 (15%). Putative hits were then selected and 9/14 reconfirmed (64%). On this basis and examining plate metrics the assay is robust. However, the primary AR assay has two limitations: Firstly, the presence of fluorescent molecules gives rise to false positives in the primary screen. However, given the low number of primary hits and the relative accessibility of the assay, it is possible to screen out false positives in orthogonal assays with little effort. As shown in Figure 3—figure supplement 2D some of these “hit” molecules were compounds that fluoresce in the presence of biological material. The platform has the potential to discover compounds increasing acrosome reaction, but this was our first screen with a medium sized drug collection of limited availability. Variations on the assay could potentially offer improvements – for example changing fluorophore (as the reviewers point out in point 7 below), or offer alternative outcomes – such as screening spermatozoa that had been incubated under capacitating conditions, or finding compounds that block induction of the AR.

7) It would be nice that some alternative strategy is discussed to test the acrosome reaction that won't be hindered by fluorescent compounds. For example, different fluorophores with emissions at more extreme ends of the spectrum could be used, so as not to overlap with the fluorescence spectra of small molecules in the screen used for the flow cytometry. Of course, this will reduce the number of false-positive, but not necessarily the number of the initial hits.

We have considered this post hoc and indeed this approach should decrease the primary hit numbers but as mentioned, not necessarily the number of true positives. We have added a sentence suggesting a possible modification (Discussion, second paragraph). Other alternatives would include a microscope-based assay using fixed cells, however this is hard to combine with a live cell assay.

8) Reversibility and less/no side effects are essential requirements for developing or screening male reproductive drugs. Can sperm motility be rescued upon drug elimination?

We agree with the reviewers that the safety profile of a contraceptive is critical. As is the case for all drug development projects the toxicological side effects will need to be scrutinized further downstream and would be dependent on how the compound was administered and to whom. Indeed, reversibility is an important issue to consider and understand. Irreversible inhibitors are often highly effective drugs. In the case of a contraceptive, such a mode of inhibition would be predicted to resist natural “wash out” as the spermatozoon moves from the male to the female reproductive tract. The downside would be that to restore fertility, time would need to pass – possibly for as long as the full spermatogenic cycle. In response to the reviewers’ comments, we have investigated reversibility using Disulfiram and have shown the data from a series of wash-out experiments in Figure 2—figure supplement 2. This data indicates that the negative effect of Disulfiram on sperm motility remains after washing out (measured at 60 min after a recovery phase). We have modified the Results (subsection “Screening of the ReFRAME library, confirmation of hits by dose response”) as well as Materials and methods section (“Disulfiram washout assay”) accordingly. We have added a fourth author to reflect their input into the additional experiments that were performed – all authors agree with this addition.

9) Were there any drugs in the library that are known to have impacts on sperm motility that weren't identified? Failure to detect such a compound might point to a limitation or area in which the assay could be improved.

This is an important point. Unfortunately, as part of the MTA with Calibr we only get to see the structures of the confirmed hits (up to a maximum of 1% of the library) so we cannot directly answer this question. False negatives are a concern in high-throughput screening. The presence of false negatives can be due to limitations in the assay or stability/quality of the compounds, or by choosing a high hit selection cut-off. Other libraries that can be screened in the future will hopefully address this point.

10) Of the long list of issues that could affect the robustness of the approach, sperm handling is one of the most important. Could more information be provided in the subsection “Sperm handling”, particularly regarding the discontinuous density gradient? Were the samples from different men pooled prior to DGC, or were the washed sperm pooled after? If after, were the samples pooled using equal numbers of sperm from each man, to keep each man's contribution equal, or were all the sperm recovered from each man used? Increased information on this point could help reduce inter-assay variability.

We agree this should have been clarified. The samples were pooled after DGC and in cases were one donor was of lower yield (<1/3 of what was needed) the contribution from the other donors was increased. Pool size was between 3-5 donors/day. We have added text added in the Materials and methods (subsection “Compound screening”) to help clarify this.

11) Somewhere in the Discussion, authors should provide a paragraph on the potential immediate utility of this approach for identifying topical spermicides/contraceptives. The world does need an improvement over nonoxynol-9, which can increase transmission of STIs, particularly for commercial sex workers. It is worth mentioning that compounds that directly affect sperm function could then be used as a starting point to identify druggable targets that would impair that target during spermatogenesis. Without this discussion, you might be open to criticism that this approach would only work for topicals. There will, of course, be influences of metabolism, etc., but high throughput screening for targets is a novel and critically important contribution.

We agree that this is worthy of further discussion. We have added a paragraph (Discussion, fifth paragraph). As suggested, not every active agent will be suitable for a drug that can be taken by the male.

12) Acrosome reaction was used as one of the key parameters for screening, and yet the assessed samples of sperm cells were ejaculated spermatozoa and not the capacitated cells. According to common knowledge, sperm ability to undergo an acrosome reaction is achieved only after they are capacitated (in vivo or in vitro), and yet in this study, the capacitation was not performed (subsection “Sperm handling”). To clarify this confusion and make the text more digestible to readers, authors are encouraged to expand their very brief result and/or Discussion sections. Specifically, since the goal of the team was to search for male contraceptives, it is not clear whether the compounds are expected to target sperm while they are stored inside the male reproductive tract or post-ejaculated. As reviewers noted, the compounds that prematurely induce acrosome exocytosis could only serve as topical contraceptives. Whether in a condom, jelly, or foam, that would come into contact with only non-capacitated sperm. Premature induction of exocytosis would NOT be recommended as a means of killing the sperm inside male reproductive tract, because if they were in the rete or epididymis, that would result in a sperm granuloma and obstructive azoospermia. This might be a point worth mentioning in the Discussion.

As the reviewers point out acrosome reaction is normally investigated using capacitated sperm. Sperm capacitation in the laboratory is achieved after 2-3 hours under specific buffer conditions. The work sponsored by the BMGF was to examine the use of a compound administered to the man to affect sperm function. As such we used non-capacitating conditions, as the cells in the epididymis – where the potential site of action is proposed – would not be in a capacitated state. In our platform, non-capacitated sperm cells are incubated with compounds, motility is measured, and subsequently acrosome status is measured. This offers the potential to find drugs or compounds that induce acrosome reaction in non-capacitated cells. In a different screening workflow, one could aim to find molecules which block acrosome reaction (provided we could obtain natural agonists of the AR e.g. ZP). The AR assay could also readily be run “stand-alone” on capacitated cells.

Premature acrosome reaction in the male reproductive tract could indeed cause problems and this would of course need to be investigated. However, it is interesting to note that when men have a vasectomy and all the sperm cells are resorbed into the body (spermatogenesis continues) fertility returns in the overwhelming majority of men upon reversal indicating that even in extreme circumstances destruction of the spermatozoa is not a significant issue.

https://doi.org/10.7554/eLife.51739.sa2

Article and author information

Author details

  1. Franz S Gruber

    National Phenotypic Screening Centre, School of Life Sciences, University of Dundee, Dundee, United Kingdom
    Contribution
    Resources, Data curation, Software, Supervision, Funding acquisition, Validation, Visualization, Methodology, Project administration
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-2008-8460

  2. Zoe C Johnston

    1. National Phenotypic Screening Centre, School of Life Sciences, University of Dundee, Dundee, United Kingdom
    2. Reproductive and Developmental Biology, Division of Systems Medicine, School of Medicine, Ninewells Hospital and Medical School, University of Dundee, Dundee, United Kingdom
    Contribution
    Data curation, Investigation
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-0904-7166

  3. Christopher LR Barratt

    Reproductive and Developmental Biology, Division of Systems Medicine, School of Medicine, Ninewells Hospital and Medical School, University of Dundee, Dundee, United Kingdom
    Contribution
    Conceptualization, Funding acquisition, Methodology
    For correspondence
    c.barratt@dundee.ac.uk
    Competing interests
    Editor for RBMO, has received lecturing fees from Merck, Pharmasure and Ferring and was on the Scientific Advisory Panel for Ohana BioSciences
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-0062-9979

  4. Paul D Andrews

    National Phenotypic Screening Centre, School of Life Sciences, University of Dundee, Dundee, United Kingdom
    Contribution
    Conceptualization, Resources, Supervision, Funding acquisition, Methodology
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-7699-5604

Funding

Bill and Melinda Gates Foundation (OPP1160989)

  • Franz S Gruber
  • Christopher LR Barratt
  • Paul D Andrews

Bill and Melinda Gates Foundation (OPP1203050)

  • Franz S Gruber
  • Zoe C Johnston
  • Christopher LR Barratt
  • Paul D Andrews

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication. The funders facilitated using the REFRAME library

Acknowledgements

We are very grateful to all members of the research team for their invaluable assistance. We also want to thank all the sperm donors who took part in this study and Dr Stephen Gellatly and Ms Morven Dean for recruitment. We acknowledge the assistance of Irene Sucquart in the early phase of this study. Additionally, we want to thank Dr David Mortimer and Dr Sharon Mortimer for their helpful insights into comparisons with the CASA system and advice in AR assay development. Thanks go to Dr Steve Publicover and Professor C De Jonge for critical reading of the manuscript. Thanks are also due to NPSC lab members for help, particularly John Raynor for engineering support. We thank Mitch Hull, Emily Chen and Kelli Kunen at CALIBR for their help in library plating, logistics and supply of ReFRAME data. Thanks also go to Professor Kevin Reid, Professor Ian Gilbert and Dr Caroline Wilson of Drug Discovery Unit in Dundee for helpful discussions. Special thanks are due to Professor Andrew Hopkins (AH) for his support.

Ethics

Human subjects: Written consent was obtained from each donor in accordance with the Human Fertilization and Embryology Authority (HFEA) Code of Practice (version 8) under local ethical approval (13/ES/0091) from the Tayside Committee of Medical Research Ethics B.

Senior Editor

  1. Anna Akhmanova, Utrecht University, Netherlands

Reviewing Editor

  1. Polina V Lishko, University of California, Berkeley, United States

Reviewers

  1. Polina V Lishko, University of California, Berkeley, United States
  2. Jean-Ju Chung, Yale School of Medicine, United States

Publication history

  1. Received: September 9, 2019
  2. Accepted: January 19, 2020
  3. Accepted Manuscript published: January 28, 2020 (version 1)
  4. Version of Record published: February 27, 2020 (version 2)

Copyright

© 2020, Gruber et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 3,194
    Page views
  • 434
    Downloads
  • 9
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Franz S Gruber
  2. Zoe C Johnston
  3. Christopher LR Barratt
  4. Paul D Andrews
(2020)
A phenotypic screening platform utilising human spermatozoa identifies compounds with contraceptive activity
eLife 9:e51739.
https://doi.org/10.7554/eLife.51739

Categories and tags

Research organism

Further reading

    1. Developmental Biology

    Contraception: Stopping sperm in their tracks

    Luke L McGoldrick, Jean-Ju Chung
    Insight

    An automated high-throughput platform can screen for molecules that change the motility of sperm cells and their ability to fertilize.

    1. Developmental Biology

    Npr3 regulates neural crest and cranial placode progenitors formation through its dual function as clearance and signaling receptor

    Arun Devotta, Hugo Juraver-Geslin ... Jean-Pierre Saint-Jeannet
    Research Article Updated

    Natriuretic peptide signaling has been implicated in a broad range of physiological processes, regulating blood volume and pressure, ventricular hypertrophy, fat metabolism, and long bone growth. Here, we describe a completely novel role for natriuretic peptide signaling in the control of neural crest (NC) and cranial placode (CP) progenitors formation. Among the components of this signaling pathway, we show that natriuretic peptide receptor 3 (Npr3) plays a pivotal role by differentially regulating two developmental programs through its dual function as clearance and signaling receptor. Using a combination of MO-based knockdowns, pharmacological inhibitors and rescue assays we demonstrate that Npr3 cooperate with guanylate cyclase natriuretic peptide receptor 1 (Npr1) and natriuretic peptides (Nppa/Nppc) to regulate NC and CP formation, pointing at a broad requirement of this signaling pathway in early embryogenesis. We propose that Npr3 acts as a clearance receptor to regulate local concentrations of natriuretic peptides for optimal cGMP production through Npr1 activation, and as a signaling receptor to control cAMP levels through inhibition of adenylyl cyclase. The intracellular modulation of these second messengers therefore participates in the segregation of NC and CP cell populations.

    1. Developmental Biology
    2. Evolutionary Biology

    Mutation of vsx genes in zebrafish highlights the robustness of the retinal specification network

    Joaquín Letelier, Lorena Buono ... Juan R Martínez-Morales
    Research Article

    Genetic studies in human and mice have established a dual role for Vsx genes in retina development: an early function in progenitors’ specification, and a later requirement for bipolar-cells fate determination. Despite their conserved expression patterns, it is currently unclear to which extent Vsx functions are also conserved across vertebrates, as mutant models are available only in mammals. To gain insight into vsx function in teleosts, we have generated vsx1 and vsx2 CRISPR/Cas9 double knockouts (vsxKO) in zebrafish. Our electrophysiological and histological analyses indicate severe visual impairment and bipolar cells depletion in vsxKO larvae, with retinal precursors being rerouted toward photoreceptor or Müller glia fates. Surprisingly, neural retina is properly specified and maintained in mutant embryos, which do not display microphthalmia. We show that although important cis-regulatory remodelling occurs in vsxKO retinas during early specification, this has little impact at a transcriptomic level. Our observations point to genetic redundancy as an important mechanism sustaining the integrity of the retinal specification network, and to Vsx genes regulatory weight varying substantially among vertebrate species.