A comprehensive study of the sperm head defects in MMAF condition and their impact on embryo development in mice

Abstract

Among rare cases of teratozoospermia, MMAF (multiple morphological abnormalities of the flagellum) syndrome is a complex genetic disorder involving at least 70 different genes. As the name suggests, patients with MMAF syndrome produce spermatozoa with multiple flagellar defects, rendering them immobile and non-fertilizing, leading to complete infertility in affected men. The only viable treatment option is ICSI. What is less understood is the presence of the various types of head defects in the spermatozoa, which are consistently present. Due to the involvement of numerous genes and the limited number of patients with MMAF syndrome, research on head defects and their impact on embryonic development remains insufficiently explored. To address these questions, a comparative study was conducted under controlled experimental conditions using four knockout (KO) mouse lines targeting Cfap43, Cfap44, Armc2, and Ccdc146 genes, all associated with MMAF syndrome in humans and mice. Each KO line underwent a detailed examination of nuclear defects, including morphology, DNA compaction, chromosomal architecture, and ploidy. The study revealed significant heterogeneity among the four lineages, with the extent of defects varying depending on the lineage, ranked as Ccdc146^−/− > Cfap43^−/− > Armc2^−/− ≈ Cfap44^−/−. The developmental potential of sperm from males in each lineage was assessed by injecting them into wild-type oocytes, and embryo development was monitored up to the blastocyst stage. Sperm from all KO lines exhibited a marked decrease in supporting embryo development compared to the wild-type, with developmental failure rates ranked as follows: Ccdc146 > Cfap43 > Armc2 > Cfap44-deficient sperm. The degree of developmental failure thus correlated with the severity of nuclear defects, and zygotes produced with sperm from Ccdc146^−/− and Cfap43^−/− mice showed the highest rates of developmental impairment. These findings from preclinical models highlight the heterogeneous nature of MMAF syndrome, both in terms of sperm nuclear defects and developmental potentials. Genetic characterization in humans is therefore crucial for improving therapeutic counselling in affected individuals.

Introduction

Teratozoospermia, a condition marked by abnormal sperm morphology, is a major contributor to male infertility. Sperm morphology relates to the size and shape of sperm cells, crucial for successful fertilization. Abnormalities can affect the head, flagellum, or both. The causes of teratozoospermia are diverse and multifactorial, including genetic factors, environmental influences, and lifestyle choices. However, genetic factors are particularly significant due to the involvement of numerous genes in spermatogenesis stages (Uhlén et al., 2016). Specifically, DNA compaction, head formation, and the development of sperm-specific organelles such as the acrosome and flagellum necessitate the precise expression of numerous genes. These genes are typically enriched or specifically expressed in the testes and mutations in these genes often lead to sperm morphological abnormalities and infertility. Research has categorized several types of teratozoospermia, each linked to specific gene mutations (Coutton et al., 2015). Our team has identified over 24 genes implicated in a condition that results in defects in the sperm flagellum. This condition is characterized by sperm displaying a short, coiled, or absent flagellum, which leads to reduced motility. Because of the diversity in these abnormalities, this phenotype is now referred to as multiple morphological abnormalities of the sperm flagellum (MMAF) (Touré et al., 2021). Approximately 70 genes have been linked to the MMAF phenotype in humans (Wang et al., 2022b; Giordani, 2024), all of which encode structural proteins of the axoneme or components of organelles like outer dense fibers and the fibrous sheath, as well as proteins involved in intra-flagellar transport (IFT). Some of these genes were initially identified in the context of primary ciliary dyskinesia (PCD), due to the striking structural resemblance between motile cilia and the flagella’s axonemal structure. However, it is important to note that this structural similarity does not necessarily indicate molecular similarity. Out of the 22 PCD-related genes identified thus far, only about half (10 genes) are conclusively associated with male infertility (Sironen et al., 2020). In our cohort of 167 patients harboring a MMAF phenotype, we previously identified biallelic deleterious variants in the 24 following genes: AK7 (Lores et al., 2018), ARMC2 (Coutton et al., 2019), CFAP206 (Shen et al., 2021), CCDC34 (Cong et al., 2022), CCDC146 (Muroňová et al., 2024), CFAP251 (Kherraf et al., 2018), CFAP43 and CFAP44 (Coutton et al., 2018), CFAP47 (Liu et al., 2021a), CFAP61 (Liu et al., 2021c), CFAP65 (Li et al., 2020), CFAP69 (Dong et al., 2018), CFAP70 (Beurois et al., 2019), CFAP91 (Martinez et al., 2020), CFAP206 (Shen et al., 2021), DNAH1 (Ben Khelifa et al., 2014), DNAH8 (Liu et al., 2020b), FSIP2 (Martinez et al., 2018), IFT74 (Lorès et al., 2021), QRICH2 (Kherraf et al., 2019), SPEF2 (Liu et al., 2020a), TTC21A (Liu et al., 2019b), TTC29 (Lorès et al., 2019), and ZYMND12 (Dacheux et al., 2023).

Spermatozoa from MMAF patients exhibit morphological defects not only in the flagellum but also in the head. Unlike conditions such as globozoospermia or macrozoospermia, where the head defects are more distinctive (Harbuz et al., 2011; Coutton et al., 2015), the head abnormalities in MMAF are diverse and less characteristic, but always present whatever the localization of the protein involved. For instance, head abnormalities were identified for genes coding for proteins involved in centrosome (Sha et al., 2017; Jin et al., 2023), inner dynein arms (Shao et al., 2024), outer dynein arms (Liu et al., 2020b), peri-axonemal structure (Liu et al., 2023a), IFT (Liu et al., 2019b), or unknown functions (Liu et al., 2021a, 2023b). These head defects are also observed in preclinical mouse models of MMAF and have been evidenced in many mice lacking individual MMAF genes (Coutton et al., 2018, 2019; Yu et al., 2021; Martinez et al., 2022; Muroňová et al., 2024). Furthermore, these head defects are practically the sole abnormalities observed in mice carrying multiple heterozygous mutations in genes associated with MMAF syndrome, as illustrated in an article linking MMAF and oligogenism (Martinez et al., 2022). This finding underscores the intrinsic connection between these head defects and the MMAF phenotype. However, there is limited understanding regarding the impact of these defects on fertilization and embryonic development.

ARTs, including IVF and ICSI, have been thoroughly investigated for their potential effects on the health of offspring. Although ART has revolutionized fertility treatment and enabled many couples to conceive, studies indicate a slightly elevated risk of certain health issues in children conceived through ART compared to those conceived naturally (Davies et al., 2012; Parazzini et al., 2015; Chen and Heilbronn, 2017; Luke et al., 2021). The underlying cause of this increased risk in children conceived through ART remains unclear, and the proportion attributed to the techniques themselves (such as ovarian over-stimulation) versus that linked to sperm defects is not well-defined. In the case of MMAF syndrome, the use of spermatozoa from patients with this condition for procreation may initially have seemed low risk, as the syndrome primarily involves defects in the flagellum, an organelle traditionally thought to have no direct role in embryonic development, and besides, sperm flagella are broken prior to human sperm injection. However, as research into MMAF progressed, it became evident that this syndrome is also associated with significant defects in the sperm nucleus across many of the studied genes. Given that head defects, such as abnormal shape, aneuploidy, and poor compaction, can impact embryo development (D’Occhio et al., 2007; Colaco and Sakkas, 2018), one of the most important questions is how to assess the risk of miscarriage or the risk of birth defects after conception with sperm from MMAF patients. Estimating these risks associated with assisted reproduction in the context of MMAF is challenging due to several factors: first, overall, the risk associated with assisted reproduction techniques remains low; second, MMAF syndrome is rare, making it difficult to gather a large cohort of patients for study; and third, MMAF syndrome is genetically complex, involving more than 70 genes, so far.

It is important to emphasize that the molecular pathway governing spermiogenesis is highly conserved across mammals. Mice have indeed served as outstanding models for investigating the molecular mechanisms underlying flagellum biogenesis. Remarkably, all strains of mice lacking genes known to be associated with MMAF in humans demonstrate a characteristic MMAF phenotype (Coutton et al., 2018, 2019; Liu et al., 2019a,b, 2020b, 2021c; Jin et al., 2023; Muroňová et al., 2024). For these reasons, we believe that thoroughly characterizing head defects and studying their impact on embryo development in mice could yield valuable insights applicable to humans. To pursue this goal, we have chosen four validated preclinical mouse models of MMAF. Two of these models lack CFAP43 or CFAP44, which are axonemal proteins located at the tether (Fu et al., 2018), a structure linking inner dynein arms, one is deficient for CCDC146 another axonemal protein and likely localized inside the microtubule doublet as a microtubule inner protein (MIP) (Muroňová et al., 2024), and the last model is deficient for ARMC2, an IFT protein involved in radial spoke transport in Chlamydomonas (Lechtreck et al., 2022). The reproductive phenotypes of all these knock-out mice have been previously described (Coutton et al., 2018, 2019; Muroňová et al., 2024) and sperm from all these mice present morphological defects in both head and flagellum.

In this article, we first aimed to characterize in detail the head defects, and second, we aimed to study and compare the fertilization and pre-implantation development of embryos obtained by ICSI with sperm from these different models.

Materials and methods

Composition of media used for ICSI and sperm handling

HCZB medium was used for ICSI. Its composition, in g/l is: sodium chloride (#1112-A, Euromedex-France, Souffelweyersheim, France) 4.771; potassium chloride (#26764.298, VWR-France, Rosny-sous-bois, France) 0.360; potassium phosphate, monobasic (#P0662, Sigma Aldrich, Saint-Quentin-Fallavier, France) 0.161; magnesium sulfate (anhydrous) (#M9397, Sigma Aldrich) 0.291; sodium bicarbonate (#27778.293, VWR) 0.420; calcium chloride⋅2H₂O (#22317.297) 0.250; EDTA (#E9884, Sigma Aldrich) 0.041; L-glutamine (#35050038, Life Technologies, Villebon-sur-Yvette, France) 0.146; sodium lactate (60% syrup—d = 1.32 g/l) (#L7900, Sigma Aldrich) 3.139; sodium pyruvate (#11360039, Life Technologies) 0.030; glucose (#G8270, Sigma Aldrich) 1.000; penicillin (#P4333, Sigma Aldrich) 0.050; streptomycin (#P4333, Sigma Aldrich) 0.070; HEPES (#H0887, Sigma Aldrich) 5.200; PVA 30 000–70 000 (#P8136, Sigma Aldrich) 0.100.

KSOM medium was used for oocyte preservation and embryo development. Its composition in g/l is: sodium chloride (#1112-A, Euromedex) 5.550; potassium chloride (#26764.298, VWR) 0.190; potassium phosphate, monobasic (#P0662, Sigma Aldrich) 0.050; magnesium sulfate (anhydrous) (#M9397, Sigma Aldrich) 0.050; sodium bicarbonate (#27778.293, VWR) 2.100; calcium chloride⋅2H₂O (#22317.297) 0.250; EDTA (#E9884, Sigma Aldrich) 0.004; L-glutamine (#35050038, Life Technologies) 0.146; sodium lactate (60% syrup—d = 1.32 g/l) (#L7900, Sigma Aldrich) 1.870; sodium pyruvate (#11360039, Life Technologies) 0.020; glucose (#G8270, Sigma Aldrich) 0.040; penicillin (#P4333, Sigma Aldrich) 0.050; streptomycin (#P4333, Sigma Aldrich) 0.070; albumin, bovine fraction V (#A3803, Sigma Aldrich) 1.000; NEAA (#11140050, Life Technologies) 0.5%; EAA (#11130036, Life Technologies) 1%.

NIM medium was used for sperm head preparation. Its composition in g/l is: sodium chloride (#1112-A, Euromedex) 0.152; potassium chloride (#26764.298, VWR) 9.320; potassium phosphate, monobasic (#P0662, Sigma Aldrich) 0.191; sodium phosphate dibasic, 2H₂O (#71643, Sigma Aldrich) 1.220; EDTA (#E9884, Sigma Aldrich) 0.876.

Animals and ethics

Generation of Cfap43 (Cfap43^−/−) and Cfap44 (Cfap44^−/−) knock-out (KO) mice was described in Coutton et al. (2018), generation of Armc2 (Armc2^−/−) KO mice was described in Coutton et al. (2019), and generation of Ccdc146 (Ccdc146^−/−) KO mice was described in Muroňová et al. (2024). Pronuclear injection and embryo transfer were performed by the transgenic core facility at the Faculty of Medicine, University of Geneva, Switzerland. Guide RNA, TracrRNA, and Cas9 were purchased from Integrated DNA Technologies. For each KO strain, mice were maintained in the heterozygous state, and males and females were crossed to produce wild-type (WT) or deficient (⁻/⁻) animals for subsequent generations. Heterozygous and homozygous animals were selected following PCR screening, and the primers used for each strain are indicated in Supplementary Table S1.

Animals were housed and bred at ‘Plateforme de Haute Technologie Animale’ UGA core facility hTAG, Inserm US46, CNRS URA2019 (La Tronche, France), EU0197, Agreement D38-516 10 006, under specific pathogen-free conditions, temperature-controlled environment with a 12-h light/dark cycle, and ad libitum access to water and diet. Animal housing and procedures were conducted in accordance with the recommendations from the Direction des Services Vétérinaires, Ministry of Agriculture of France, according to European Communities Council Directive 2010/63/EU and according to recommendations for health monitoring from the Federation of European Laboratory Animal Science Associations. Protocols involving animals were reviewed by the local ethic committee ‘Comité d’Ethique pour l’Expérimentation Animale #12, Cometh-Grenoble’ and approved by the Ministry of Research (APAFiS #7128 UHTA-U1209-CA).

Intracytoplasmic sperm injection

Oocytes were obtained from superovulated B6D2 females. Females were stimulated by a first injection of 7.5 UI of PMSG (pregnant mare serum gonadotropin, Med-Vet, syncro-part PMSG, 600 UI), followed 48 h later by a second injection of 7.5 UI of HCG (human chorionic gonadotropin, Med-Vet, Chorulon 1500 UI). Fourteen hours later, females were euthanized by cervical dislocation, and cumulus–oocyte complexes (COCs) were harvested from the ampulla. COCs were incubated 10 min in 500 µl M2 (#5910, Sigma Aldrich)/0.1% hyaluronidase (#H3884, Sigma Aldrich) and cumulus-free oocytes collected, and rinsed twice with 500 µl KSOM medium, then incubated in KSOM medium under mineral oil (#M8410, Sigma Aldrich) until injection at 37°C/5% CO₂.

Spermatozoa were harvested by dilaceration of the cauda epididymides in 1 ml of NIM medium, then washed twice by centrifugation with 1 ml fresh NIM medium for 5 min at 500g. Finally, the sperm heads were separated from the flagella by sonication for 5 s at 60% on ice using a QSONICA sonicator, model Q125. Sperm were incubated at 4°C in an NIM medium until injection.

For injection, oocytes were transferred to 500 µl of HCZB medium and incubated for 10 min at room temperature. Spermatozoa were then diluted in viscous NIM/12% polyvinylpyrrolidone (PVP) medium (#PVP360-100G, Sigma Aldrich) to facilitate sperm handling. An injection dish was prepared containing (i) a 20-µl drop of NIM/12% PVP medium for washing the injection pipette, (ii) a 20-µl drop of HCZB medium for washing the holding pipette, (iii) a 20-µl drop of NIM/6% PVP medium containing the spermatozoa, and (iv) a 60-µl drop of HCZB medium for oocyte injection. All drops were covered with mineral oil (#M8410, Sigma Aldrich). Under a Nikon Eclipse TE200 microscope (Nikon, Champigny-sur-Marne, France), WT or KO sperm were injected into WT oocytes using a microinjector CellTram^®4r Oil (Eppendorf, Montesson, France) attached to a Narishige MN-188NE micromanipulator. The injection pipette (BioMedical Instruments, Zöllnitz, Germany) was installed on a piezo (PiezoXpert Eppendorf) following the method described previously (Yoshida and Perry, 2007). Five minutes after injection, oocytes were washed three times with KSOM medium and then incubated in a droplet of 500 µl of KSOM covered by mineral oil at 37°C/5% CO₂.

Embryo development

Injected oocytes were layered in a droplet of 500 µl of KSOM in a petri dish (#16004, Vitrolife, Paris, France) covered by mineral oil and incubated at 37°C/5% CO₂. Embryo development was followed every 24 h under a heated binocular for 5 days. For zygotic development, 4–6 injected oocytes were layered in an embryoslide imaging dish (#16450, Vitrolife) in KSOM, covered by mineral oil as described previously (Pierré et al., 2022). Embryo development was then monitored by time-lapse imaging under a holographic microscope and images were recorded every 10 min for 48 h. From the acquired images, a movie was made with ImageJ (https://imagej.net/ij/), and the times of occurrence of different events including second polar body extrusion, pronuclei appearance, and fading and first mitosis, were identified.

Analysis of nuclear morphology

Nuclear morphology was precisely evaluated by the Nuclear Morphology Analysis Software (NMAS) (version 1.19.2, https://bitbucket.org/bmskinner/nuclear_morphology/wiki/Home), according to the analysis method described in Skinner et al. (2019a,b). The software processed images of DAPI-stained nuclei were captured with a Zeiss Imager Z2 microscope, using a CoolCube 1 CCD camera, with a 100×/1.4 Zeiss objective and Neon software (MetaSystems, Altlussheim, Germany). Nucleus detection settings were Kuwahara kernel: 3, and flattening threshold: 100, for preprocessing; canny low threshold: 0.5, canny high threshold: 1.5, canny kernel radius: 3, canny kernel width: 16, gap closing radius: 5, to find objects; and min area: 1000, max area: 10 000, min circ: 0.1, max circ: 0.9, for filtering. After acquisition of images of nuclei, landmarks were automatically identified using a modification of the Zahn–Roskies transform to generate an angle profile from the internal angles measured around the periphery of the nuclei. Angles were measured at every individual pixel around the original shape. This method combines data from every possible polygonal approximation into a single unified trace, from which landmark features can be detected (under-hook concavity, tail socket, caudal bulge and base, acrosomal curve, etc.). Angle profiles are measured as angle degrees according to the relative position of each pixel along the perimeter, and variability profiles use the interquartile range (IQR, difference between the third and first quartile) as a dispersion indicator to measure the variability of values obtained for each point. Sperm shape populations were then clustered using a hierarchical ward-distance method without reduction, based on angle profiles.

Sperm fluorescence in situ hybridization (FISH) experiments

Epididymal sperm were washed in PBS and fixed in Carnoy’s solution (3:1 methanol/acetic acid) for 1 h. Cells were then spread on slides and processed according to the FISH procedures as described previously (Martinez et al., 2023). Briefly, after a wash in saline-sodium citrate solution (SCC-2X), sperm cells were decondensed by a 30-min incubation in a dithiothreitol solution (10 mM DTT in 0.1 M Tris-HCl with 0.1% Triton X-100, pH 8). Sperm cells were then washed again in SSC-2X, dehydrated and hybridized with commercial MetaSystems whole-chromosome painting probes XCP 11 (D-0311), and XCP 19 (D-0319) according to the manufacturer’s protocol in a HYBrite system (Abbott Laboratories Chicago, IL, USA). Nuclei were counterstained with a DAPI II solution (Abbott Laboratories) and scoring was performed according strict criteria (Martinez et al., 2013).

For chromosomal area analysis, pictures were acquired on a Zeiss Imager Z2 microscope using a CoolCube 1 CCD camera with a 100×/1.4 Zeiss objective and Neon software (MetaSystems, Altlussheim, Germany). Pictures were processed with the NMAS (version 1.19.2, https://bitbucket.org/bmskinner/nuclear_morphology/wiki/Home), according to the analysis method described in Skinner et al. (2019a,b). Nucleus detection settings were: Kuwahara kernel: 3, and flattening threshold: 100, for preprocessing; canny low threshold: 0.5, canny high threshold: 1.5, canny kernel radius: 3, canny kernel width: 16, gap closing radius: 5, for object finding; min area: 1000, max area: 10 000, min circ: 0.1, max circ: 0.9, for filtering.

Aniline blue and chromomycin A3 staining

Aniline blue coloration was performed on epididymal sperm as described in Yassine et al. (2015). Cells were washed in PBS, fixed in 3% glutaraldehyde solution and incubated in a succession of coloration baths: water, 5% aniline blue (in 4% acetic acid solution), water, ethanol solutions and, finally, toluene. Slides were then analyzed using a transmitted light microscope using a 100× objective with oil. Positive cells displayed dark blue staining whereas negative cells displayed light or very light staining.

For chromomycin A3, epididymal sperm were washed for 5 min at 500g in PBS (1×) and fixed in Carnoy’s solution for 45 min. Fixed cells were dropped onto slides and treated as follows: 20 min in staining solution (0.25 mg chromomycin in McIlvaine buffer, pH 7), rinsed twice for 2 min each in McIlvaine buffer (pH 7), stained for 3 min in Hoechst solution (0.5 µg Hoechst in PBS (1×), and rinsed for 3 min in PBS (1×). Slides were then mounted with Dako anti-fading solution and analyzed with an epifluorescent microscope. Negative nuclei fluoresced slightly pale green, and positive nuclei fluoresced strongly bright green.

Statistical analyses

The experimental units are: for Figs 1, 2, 3, and 4C and D, single sperm cell; for Fig. 4A and B, sperm from a single male, for Figs 5 and 7, single zygote/embryo; and for Figs 6 and 7 and Supplementary Figs S1 and S2, batch of zygotes/embryos with oocytes coming from several females and injected with sperm from a single male.

The statistics relating to nuclear morphology were automatically calculated by the NMAS. This analysis relied on a Mann–Whitney U test with Bonferroni multiple testing correction. P-values were considered significant when ≤0.05.

All other data were treated with GraphPad Prism software (version 10; Boston, MA, USA). Data are represented as box plots and the median or histograms with mean ± standard deviation, and the statistical significance of differences was assessed by applying a statistical test as specified in the legends of the figures. The statistical test was chosen according to the following rules: non-parametric test if n < 3 or if the distribution was not Gaussian. Otherwise, a parametric test was used. Statistical tests with two-tailed P-values ≤0.05 were considered significant.

Results

Due to the high heterogeneity of the head defects, we used the NMAS to perform a comprehensive study of the head abnormalities (Skinner et al., 2019a,b). From the four KO mouse lineages, the NMAS identified 24 different head shapes from head angle profiles (Fig. 1). While in some forms, the morphological defects remained limited (Forms # 1–8), in others, the head was severely deformed and the hook was no longer distinguishable (Forms # 15–24). The four distributions in the different categories were not homogeneous and differed between the different lineages. For example, while Armc2^−/− males had the lowest rate of severely altered forms (8.72% for forms # 15–24), this rate reached almost 30% for Cfap44^−/− males, showing that almost 1 in 3 heads had significant alterations. Cfap43^−/− and Ccdc146^−/− males showed intermediate rates at 22.93% and 22.59%, respectively. To objectivize the heterogeneity of sperm shapes among the four lineages, we measured the variation of the sperm angle profile (Fig. 2). This variation was assessed as the root-mean-square difference between the per-nucleus angle profile and the median angle profile for the dataset, after interpolation to a fixed length. As expected for WT mouse sperm, the scatter plot showed that the angle profiles were mostly similar (Fig. 2A). In contrast, this measure evidences the large heterogeneity of the angle profiles for KO sperm (Fig. 2B–E). To compare the variation of the sperm angle profiles between the different KO lineages, we split the graphs into three areas, corresponding to almost normal head shape, malformed and severely distorted (Fig. 2F), and compared the percentage of sperm found in each area (Fig. 2G). Consistent with Fig. 1, the lineages presenting the most distorted sperm shape were Ccdc146^−/− > Cfap44^−/− > Cfap43^−/− > Armc2^−/−.

The chromosomal organization in the sperm nucleus is not random and chromosomal territories have been identified in both human and mouse sperm (Zalenskaya and Zalensky, 2004; Skinner et al., 2019a,b). The study of chromosome localization in sperm cells is of importance since the activities of genes are influenced dynamically by their chromosomal and nuclear position. This chromosomal arrangement has never been assessed in the context of MMAF and, for this purpose, FISH experiments were performed and the localizations of chromosomes 11 and 19 were investigated. Sperm nucleus was divided in four areas (apical, ventral, dorsal, and basal; Fig. 3A) and for each sperm, the localization of the chromosomes 11 and 19 was allocated in one of these areas (Fig. 3B). Chromosome 11 was predominantly positioned in basal and ventral areas in WT sperm. The absence of one of the different MMAF proteins did not change this organization despite minor differences (Fig. 3C). For chromosome 19, no preferred location was observed in WT sperm and similarly, the absence of one of the four MMAF proteins had no impact on chromosome 19 location.

We next assessed the level of aneuploidy in the context of MMAF by counting the number of chromosomes 11 and 19 per sperm nucleus. The counts of chromosomes 11 and 19 were determined by FISH (Fig. 3D). Small increases of aneuploidy were observed for all lineages, from 0% for WT sperm to 0.2–0.4% for KO lineages.

The study of sperm shapes from the different KO lineages clearly showed that numerous shapes present a larger area, suggesting that compaction occurring during spermiogenesis may be hampered. Sperm DNA compaction, based on histone replacement by protamine, can be assessed by staining sperm with acidic aniline blue and chromomycin A3 (CMA3), which are positive (stained) when histones are retained inside the nucleus and when protamines are absent, respectively (World Health Organization, 2010). Positive sperm are thus a marker of poor compaction. Cfap43^−/− and Ccdc16^−/− males presented a significant increase of positive sperm for aniline blue and CMA3 tests (Fig. 4A and B). In contrast, Armc2- or Cfap44-deficient sperm did not show a significant increase of aniline blue or CMA3 positive cells (Fig. 4A and B). The compaction of chromosome 11 and 19 were specifically studied by FISH (Fig. 4C and D). We observed for sperm from Cfap43^−/− and Ccdc146^−/− lineages that areas of chromosome 11 and 19 are significantly larger than that of WT, therefore confirming the aniline blue and CMA3 results. The situation is more complex for Cfap44^−/− and Armc2^−/−-deficient sperm. For chromosome 19, the results of both methods are consistent, no statistical differences were observed between the area of the WT chromosome 19 and the areas of Cfap44^−/− and Armc2^−/− chromosome 19. In contrast, for chromosome 11, a discrepancy was noticed: whereas no significant compaction defect was observed with aniline blue and CMA3 (Fig. 4A), the areas of chromosome 11 of Armc2 or Cfap44-deficient sperm presented a significantly larger area than that measured from WT.

Because it has been demonstrated that sperm head defects lead to embryo developmental defects (D’Occhio et al., 2007; Colaco and Sakkas, 2018), the potential of MMAF sperm to support development was evaluated. WT oocytes (B6D2) were collected from WT females after hormonal stimulation and fertilized by ICSI either with WT sperm or with sperm from the different KO lineages. Then, zygotes were maintained in a CO₂ incubator and their preimplantation development was monitored up to the blastocyst stage. For some batches of embryos, their development was monitored by a time-lapse system up to the two-cell stage to measure zygotic events.

We first focused our study on the embryo survival curve between sperm injection and the blastocyst stage. For WT embryos and those produced with sperm from the four lineages, a survival curve was calculated and Log rank Mantel Cox and multiple comparison to WT was used for comparing the survival experience of all groups (Fig. 5A). For the four lineages, the embryonic survival curves were significantly different to the control, demonstrating that using abnormal sperm negatively impacts embryo development (Fig. 5B). To better characterize this impaired development, we next studied the rate of two-cell embryos and compared the percentage of zygotes reaching the two-cell stage (Fig. 6). Sperm from the four KO lineages showed a significant reduction in their potential for normal development: the percentage of live injected oocytes reaching two-cell embryos dropped from 82.8% for WT sperm to yield between 21.6% for sperm from Ccdc146^−/− males and 59.9% for sperm from Cfap44^−/− males. Significant differences in developmental potential of sperm from the four lineages were actually observed and ranked in the following order, Cfap44^−/− > Armc2^−/− > Cfap43^−/− > Ccdc146^−/− (Fig. 6A). We observed that a fraction of zygotes reached the two-cell stage later at 48 h post injection, and we wondered if sperm from the MMAF lineages were associated with a delayed development. To address this question, we compared the rate of zygotes reaching the two-cell stage at 48 h, for zygotes injected with WT sperm and from the 4 MMAF lineages (Fig. 6B). The rate of embryos reaching the two-cell stage at 48 h was not different between the WT (mean 9.7%) and the four different lineages (mean in between 3.3% and 7.2% for Cfap43^−/− and Armc2^−/−, respectively). Several defects, including fragmentation, asynchronous or asymmetric divisions were observed during zygote development, preventing zygotes from reaching two-cell embryos, and in particular for Armc2, Cfap43, and Ccdc146-deficient sperm (Fig. 6C). For sperm from Cfap44^−/− males, we did observe an increase of such defects, but the difference was not significant (Fig. 6C). The mean percentage of fragmented zygotes was 1.5% for WT sperm after 48 h whereas it rose to 9.1% for sperm from Cfap44^−/−, 23.1% for Cfap43^−/−, 25.0% for Armc2^−/−, and 27.4% for Ccdc146^−/− males (Fig. 6C). Some examples of compromised embryos obtained with sperm from Ccdc146 deficient males are illustrated in Fig. 6D.

To better understand why zygote development aborted during the first mitotic division, their development was followed with a holographic time-lapse system to monitor several morphokinetic parameters. Following sperm injection, the oocyte completes the second meiotic division and male and female pronuclei (PN) start to migrate towards each other to form the first mitotic spindle. In the WT zygote, male and female PN appeared visible with our system at 5.09 h and 5.42 h, respectively. We followed their movement towards each other, and we measured the moment of pronuclei fading, corresponding to nuclear envelope breakdowns, prior entry in first mitosis (Suzuki et al., 2021). In the WT, this event was observed at 16.4 h post-injection. We also registered the time between pronuclei fading and first mitosis (mean WT: 1.8 h) and time of the first division (mean WT: 18.2 h) (Fig. 7A). Next, we compared these values to those obtained with zygotes produced with sperm lacking the different MMAF proteins (Fig. 7B–F). We measured the morphokinetic parameters only in zygotes that had an early ‘normal’ development including at least pronuclei formation and fading. Many zygotes that showed strong cortical deformation waves, and eventually fragmenting, were discarded from the morphokinetic analyses because pronuclei formation was not observable. The percentages of such zygotes are indicated in Fig. 6A. Taking into account all zygotes (those reaching two-cell stages and those eventually failing to divide), we observed that none of the parameters were affected for zygotes produced with Cfap44 and Ccdc146-deficient sperm, an unexpected result since Ccdc146-deficient sperm resulted in the highest rate of zygote abortion. For zygotes produced with Cfap43-deficient sperm, the time for pronuclei fading (mean Cfap43^−/−: 27.4 h), the interval between pronuclei fading and first mitosis (mean Cfap43^−/−: 5.25 h), and time for first mitosis (mean Cfap43^−/−: 30.4 h) were significantly increased. Finally, all kinetic parameters were delayed for zygotes produced with Armc2^−/− sperm (time for appearance of male and female pronuclei) (mean Armc2^−/−: 5.95 h and 6.29 h for male and female PN, respectively), time for pronuclei fading (mean Armc2^−/−: 19.7 h), and first division (mean Armc2^−/−: 21.8 h). We next analyzed the PN formation and PN fading separately in the zygotes that end up dividing from those that never cleave. Unfortunately, there were few zygotes that failed to divide at the end of the recording and their number for each KO lineage was between 2 and 5, making an individual lineage analysis impossible. The only possible statistical analysis concerned the appearance time of male pronuclei for Armc2^−/− lineage; no statistical difference was observed between zygotes that end up dividing from those that never cleave, and both experimental groups were statistically different from WT zygotes (Fig. 7G).

Next, the embryonic development between the two-cell stage and the blastocyst stage was studied (Supplementary Fig. S1). The transitions from two-cell to four-cell embryos and from four-cell to blastocyst embryos were analyzed individually. First, we measured the yield of the transition of two-cell to four-cell embryos for those produced with WT sperm and 77% of two-cell embryos developed into four-cell embryos. Next, we compared this rate with those obtained with sperm from the four deficient lineages and the following yields were obtained: Armc2^−/− = 75%, Cfap44^−/− = 61%, Cfap43^−/− = 70%, and Ccdc146^−/− = 43%; these yields were not significantly different in comparison to the control (Supplementary Fig. S1A). For the next transition, four-cell to blastocyst, the development of embryos obtained with Cfap44 and Armc2-deficient sperm was not statistically different to the development of WT embryos (50% and 59% of four-cell embryos obtained with Cfap44 and Armc2-deficient sperm developed into blastocysts, respectively, compared to 59% for WT embryos). On the other hand, the development of four-cell embryos obtained with Cfap43 and Ccdc146-deficient sperm in blastocyst was significantly hampered compared to the development of WT embryos and 19% and 26% of four-cell embryos obtained with Ccdc146 and Cfap43-deficient sperm developed to blastocysts, respectively (Supplementary Fig. S1B). When considering the entire developmental process from the two-cell stage to the blastocyst stage, the embryonic development was not significantly different from control for embryos produced with sperm from Cfap44^−/− and Armc2^−/− lineages. In contrast, the development post two-cell stage is hampered for embryos produced with Ccdc146 (10%) and Cfap46-deficient sperm (19%) (Supplementary Fig. S1C).

Finally, the mean percentages of live injected oocytes reaching the four-cell and blastocyst stages are presented in Supplementary Fig. S2A and B and statistical analyses were performed. These bar graphs show that the mean percentage of live injected zygotes reaching the 4-stage stage were 72%, 44%, 31% 18%, and 18% for sperm from WT, Cfap44^−/−, Armc2^−/−, Cfap43^−/−, and Ccdc146^−/− lineages, respectively. For the percentage of live injected zygote reaching the blastocyst stage, the mean percentages were 46%, 25%, 13%, 7%, and 6% for sperm from WT, Cfap44^−/−, Armc2^−/−, Ccdc146^−/−, and Cfap43^−/− lineages, respectively. For both endpoints, the yields of embryos obtained with sperm from deficient lineages were significantly different to that obtained with WT sperm.

Discussion

Couples affected by MMAF typically encounter difficulties achieving natural conception, often relying on assisted reproductive techniques, particularly ICSI, to overcome the low motility of teratozoospermic sperm. A pivotal question regarding MMAF is whether head defects increase the risk of developmental failure. Currently, there are over 20 reports documenting successful births of newborns conceived using ICSI with sperm from MMAF patients (Li et al., 2019; Ni et al., 2020; Liu et al., 2021a; Sha et al., 2021; Wang et al., 2022a; Shao et al., 2024). However, there are also conflicting reports showing either an impairment of embryo development (Long et al., 2024) or no successful births at all (Sha et al., 2017; Liu et al., 2021b), which complicates the analysis of the situation. Assessing the efficacy and associated risks of using MMAF sperm in ICSI from these clinical cases is challenging because each report typically involves only one or a few couples, and therefore a low number of embryos, making a statistical analysis impossible. A recent study by Ferreux et al. included 25 MMAF patients, but molecular causes were identified for only 13 patients (Ferreux et al., 2021), casting doubt on conclusions regarding the impact of MMAF on embryo development. Moreover, many reports focus on males or couples with mutations in different genes making the results difficult to analyze. It is worth noting that even within the same gene, different mutations can lead to varying impacts on transcription and induce different phenotypes. For example, various variants of ZMYND15 have been associated with distinct reproductive phenotypes (Kherraf et al., 2022). Furthermore, the fertility status of the female partner is often undisclosed, adding another layer of complexity to result analysis. Moreover, the validity of some control cohorts in previous reports is questionable, as they consist of oligo-astheno-teratozoospermic patients whose molecular causes are unknown. Due to the rarity of MMAF, assembling a substantial cohort of patients with the same mutation is challenging, which hinders obtaining definitive answers to the aforementioned questions in humans. Hence, we opted to address this issue using a relevant animal model, employing genetically controlled wild-type males (B6D2) as our comparative control. These four KO lineages, with the absence of the corresponding protein, reflect the human MMAF condition because patients have identical molecular defects, with absent or non-functional proteins, as demonstrated in our previous human study (Coutton et al., 2018, 2019; Muroňová et al., 2024).

This study represents the first comprehensive investigation into sperm head defects and their influence on embryo development in the context of MMAF. We conducted a comparative analysis of four mouse lineages displaying typical MMAF characteristics under standardized experimental conditions within the same laboratory (using identical equipment, methods, experimenters, and genetic background), enabling direct comparisons between lineages. We produced between 147 and 228 zygotes for each lineage, making this analysis statistically relevant. Initially, we examined their morphology, nuclear DNA compaction, and the incidence of aneuploidy, which are critical parameters known to impact embryonic development prediction (D’Occhio et al., 2007; Colaco and Sakkas, 2018). Firstly, we meticulously delineated the morphological defects specific to each lineage, a crucial parameter highlighted by the intracytoplasmic morphologically selected sperm injection (IMSI) technique (Mangoli and Khalili, 2020). IMSI, which employs a high-magnification analysis of nuclear head shape and structure, indeed underscores that using perfectly formed spermatozoa enhances pregnancy success rates. Our findings reveal that the prevalence of defects varies depending on the gene studied, with the Armc2^−/− lineage exhibiting the lowest percentage of malformed sperm (20%), contrasted by the Ccdc146^−/− lineage with the highest (55%). Additionally, we conducted a detailed assessment of compaction levels using two types of stains that distinguish the presence and absence of histones and protamines. Numerous studies underscore that compaction quality serves as a pivotal predictive factor for pregnancy success, as reviewed extensively in the literature (D’Occhio et al., 2007). For instance, aniline blue staining was shown as capable to discriminate between fertile and infertile men (Auger et al., 1990). Based on this parameter, we show that two lineages have a significant increase of badly compacted sperm, Cfap43^−/− and Ccdc146^−/−, whereas Cfap44^−/− and Armc2^−/− lineages exhibit a non-significant increase. We confirmed that compaction is hampered in Cfap43^−/− and Ccdc146^−/− lineages using CMA3 staining and a measure of the surface of chromosome 11 (only for Cfap43-deficient sperm) and 19. We observed a significant but limited increase in the level of aneuploidy for the four lineages. Although this increase is weak, this result suggests that the absence of these proteins not only impacts spermiogenesis but also meiosis. Finally, we did not observe an impact on chromosome localization. From these results, it is evident that sperm from two lineages, Ccdc146^−/− and Cfap43^−/−, are particularly affected by the absence of proteins associated with MMAF, displaying high levels of sperm deformities and defective DNA compaction. The Armc2^−/− and Cfap44^−/− lineages exhibit an intermediate profile, with Armc2^−/− showing the lowest level of sperm deformities and minor DNA compaction defects observed only with FISH, while Cfap44^−/− shows comparatively an intermediate level of sperm deformities and minor compaction defects also observed with FISH. Notably, we did observe an increase in compaction defects using aniline blue and chromomycin stains for sperm from Armc2^−/− and Cfap44^−/− males (Fig. 4A and B) compared to WT, although these increases were not statistically significant with a non-parametric test. The discrepancy between FISH and stain methods may be due to the lower number of replicates in the stain method (only three), thereby influencing the observed outcome.

CFAP43 and CFAP44 are two sister proteins located at the inner dynein arm tether (Fu et al., 2018); nevertheless, their phenotype is remarkably different, and these differences observed between sperm from Cfap43^−/− and Cfap44^−/− lineages raise questions. It suggests that both proteins play different functions beyond their involvement in the tether. CFAP43 has been involved in the intra-manchette transport (Yu et al., 2021) and it would be interesting to check whether CFAP44 is also involved in this process. If it is, it would confirm the importance of the manchette is sperm shaping. Overall, the most important result of this first part is the heterogeneity of head defects among the different lineages, suggesting that the developmental competence of MMAF-protein deficient sperm should also be different.

In the second part of our study, we addressed this hypothesis by focusing on the rate of two-cell embryos following injection of knockout (KO) sperm. Across all four lineages, we observed significant decreases in the rate of two-cell embryos compared to WT sperm. However, notable differences were observed between the lineages, as anticipated. The most severely affected lineages, Ccdc146^−/− and Cfap43^−/−, showed a substantial drop in the rate of two-cell embryos from approximately 82.8% (WT sperm) to 21.6% and 22%, respectively. Sperm from Armc2^−/− have a lower impact, with a two-cell embryo rate around 42%. Conversely, sperm from Cfap44^−/− males displayed a milder decrease, with a two-cell embryo rate around 59.9%, still a 27% reduction. Several studies on embryo development, particularly in MMAF mouse models following ICSI, have reported similar heterogeneity in the rate of two-cell embryos. For example, for the Cfap47 (Liu et al., 2021a) and Cfap52 (Jin et al., 2023) genes, no negative impact has been described, whereas, for Dnah8 (Liu et al., 2020b) and Cfap206 (Shen et al., 2021) genes, a 50% decrease in the two-cell rates has been published. It is evident that MMAF syndrome should be analyzed on a gene-by-gene basis concerning embryonic development rather than as a global condition. Unexpectedly, the developmental kinetics of zygotes reaching the two-cell stage were minimally disturbed, suggesting healthy two-cell embryos were obtained. To investigate this further, we examined the two-cell to four-cell transition and subsequently the four-cell to blastocyst transition. Firstly, no differences were observed in the two-cell to four-cell transition among the lineages. However, regarding the four-cell to blastocyst transition, results varied significantly among the four lineages: embryos obtained with sperm from Cfap44^−/− and Armc2^−/− lineages showed no significant differences compared to the control, indicating relatively normal progression. In contrast, embryos obtained with sperm from Ccdc146^−/− and Cfap43^−/− lineages exhibited significantly impaired four-cell to blastocyst transitions. Impaired blastocyst formation was also reported for two other MMAF genes, Cfap206 and Dnah8 in a mouse model (Liu et al., 2020b; Shen et al., 2021), demonstrating that impaired development beyond the two-cell stage is more common than expected. These results hold significant implications for human infertility for several key reasons. Firstly, our results show that the quality of embryos obtained by ICSI with teratozoospermatozoa is greatly reduced for certain genes, contrary to what has been published previously (Terriou et al., 1997). Second, our report demonstrates that the two lineages exhibiting the most adverse embryonic development also exhibit significant morphological and compaction defects. These parameters thus appear predictive of poor embryonic development (Fig. 8) and warrant thorough characterization in therapeutic management. Third, morphokinetic parameters of zygote development are not useful for three of the four lineages, making zygote observation through time-lapse incubation systems not useful for MMAF couples. Fourth, in humans, embryo transfer at the morula stage (Day 3) is increasingly common (Garbhini et al., 2023). However, our findings suggest this strategy may be risky in some MMAF cases, as certain genes impair embryo development post-four-cell stage. Therefore, based on these preclinical results and assuming that the rates of development of blastocysts into fetuses are similar for all lineages, we suggest to perform embryo transfer at the blastocyst stage to exclude embryos at risk of developmental arrest between the four-cell and blastocyst stages.

In conclusion, MMAF syndrome presents a significant challenge in male infertility due to abnormal sperm morphology and nuclear defects. Our results suggest that these defects are gene-dependent and therefore, understanding their molecular causes is essential for improving diagnosis and management options, which are crucial for effective treatment and support for couples facing infertility. Our findings highlight that head defects observed in sperm from MMAF patients correlate with increased developmental abnormalities and arrests, compromising the success of assisted reproduction technologies. Couples affected by MMAF syndrome should be informed about these risks to prepare them adequately. Moreover, these results should prompt clinical trials to validate sperm nuclear defects in MMAF and their consequences on embryo development. Additionally, based on our findings, embryo transfer at the blastocyst stage may enhance success rates, and would deserve confirmation in human studies.

Supplementary data

Supplementary data are available at Molecular Human Reproduction online.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding authors.

Acknowledgements

We thank the zootechnicians of Animal core facility (UAR2019 hTAG) for animal housing and care.

Authors’ roles

G.M. and C.A. analyzed and interpreted the data and wrote the manuscript; S.N., G.C., and C.L. generated and managed the mouse lineages; J.M. and E.L. performed and analyzed experiences on embryonic development; C.T. and Z.W. measured and analyzed sperm head abnormalities; M.C., J.E., Z.-E.K., and C.C. performed data interpretation and statistical analyses. G.M., P.F.R., and C.A. designed the study, supervised all laboratory work, had full access to all of the data from the study, and took responsibility for the integrity of the data and its accuracy. All authors contributed to the manuscript by critically revising it. All authors approved the final version and committed to ensuring the accuracy and integrity of the work presented.

Funding

This work was supported by INSERM, CNRS, Université Grenoble Alpes, the French Agence Nationale pour la Recherche (ANR) grants ‘MAS-Flagella’ (ANR-19-CE17-0014), and ‘FLAGELOME’ (ANR-19-CE17-0014) to P.F.R., ‘MIP-MAP’ (ANR-20-CE13-0005) to C.A., the Direction Générale de l’Offre de Soin (DGHOS) for the program PRTS 2014 to P.F.R., the Fondation Maladies Rares (FMR)—grant ‘Whole genome sequencing of subjects with Flagellar Growth Defects (FGD)’ financed by for the program Séquençage à haut débit 2012 to P.F.R.

Conflict of interest

The authors declare no conflict of interest.

References

Author notes