This study was approved by the Ethical Committee of University of Science and Technology of China (USTC). Written informed consent forms from all participants were obtained at the beginning of the study.

This study recruited three Pakistani families, family 1, 2 and 3, having four (P1-P4), four (P5-P8), and three (P9-P11) infertile brothers, respectively. Semen analysis, including semen volume, sperm concentration, and sperm motility, were performed at least twice for P1-P5, P7, P9, and P10 according to the World Health Organization (WHO) guidelines (World Health Organization, 2010). Semen smears were prepared from fresh semen samples on clean glass slides, air-dried, fixed in 4% paraformaldehyde, and stored at −20°C for subsequent hematoxylin and eosin (H&E) staining or IF staining. For H&E staining, smears were washed by PBS once, stained in hematoxylin for 20 min, washed by ddH₂O once. and put into 70% ethanol with 1% HCl for 2 s. After washing in tap water for 3 min, the slides were gradiently dehydrated by ethanol. Then the smears were stained with eosin and then placed in the baths of 100% ethanol and xylene. After sealed with neutral balsam, the smears were observed under a light microscope. Morphological analysis was performed for P1-P9 with at least 200 spermatozoa for each experiment examined (for P6 and P7, two independent experiments were performed). The semen samples for fertile control were obtained from the volunteers at the First Affiliated Hospital of USTC. A detailed questionnaire regarding PCD clinical features was obtained from all available family members and all excluded the history or presence of ciliary-related symptoms, including situs ambiguous, unexplained chronic airway infections, respiratory distress at newborn, chronic middle ear infections, clubbing, hydrocephalus etc (according to the National Institute of Health, https://www.nhlbi.nih.gov/health-topics/primary-ciliary-dyskinesia).

QIAamp Blood DNA Mini Kit (QIAGEN) was used to extract the blood genomic DNA from available family members. For WES, AIExome Enrichment Kit V1 (iGeneTech, Beijing, China)-captured libraries were constructed for family members of Family 1 (I:2, II:2, II:4, II:5, II:6 and II:7), Family 2 (I:1, I:2, II:3, II:4, II:5 and II:6), and Family 3 (III:2 and IV:1) as instructed by the manufacturer. Sequencing was carried out on a Hiseq2000 platform (Illumina, San Diego, CA, United States). Variants following recessive inheritance were kept for further screening following the strategy as we described previously (Yin et al., 2019; Zhang et al., 2020), and as shown in Supplementary Figure S1. Sequences of primers used for Sanger sequencing are listed in Supplementary Table S1.

Fresh semen samples were obtained and centrifuged at 900 x g for 3 min. After washing with PBS for three times, spermatozoa were fixed in 0.1 M phosphate buffer (pH 7.4) containing 4% glutaraldehyde, 4% paraformaldehyde, and 0.2% picric acid at 4°C for at least overnight. Scanning electron microscopy (SEM) and TEM analyses were conducted as we previously described (Ma et al., 2021).

IF staining were performed as we previously reported (Zhang et al., 2020). Briefly, the fixed smears were taken out from −20°C for 1 h to allow to air dry at room temperature. After permeabilized with 0.2% Triton X-100 in PBS for 30 min, the smears were blocked with 3% skim milk, and then incubated with primary antibodies at 4°C overnight, followed by secondary antibodies at 37°C for 1 h. For negative control experiments, semen smears from a fertile man were stained with the mouse IgG antibodies (as the negative control for the anti-α-Tubulin antibodies) or rabbit IgG antibodies (as the negative control for the anti-SPAG6, anti-TSGA2, anti-DNAH1, anti-CFAP61, and anti-CFAP251 antibodies), followed by incubation with corresponding secondary antibodies under the same staining conditions (Supplementary Figure S2). The antibodies that were used are listed in Supplementary Table S2.

Eleven infertile men from three unrelated Pakistani families were investigated in the current study. Family 1 has four infertile brothers, P1, P2, P3, and P4; family 2 has four infertile brothers, P5, P6, P7, and; family 3 has three infertile brothers, P9, P10, and P11, born to a consanguineous marriage (Figure 1A). All the patients had been married and trying to conceive for at least 8 years, but were infertile. They had a normal karyotype (46; XY) and no large-scale deletions in Y chromosomes. According to WHO guidelines (World Health Organization, 2010), the semen volume was either normal or slightly decreased in these 11 patients. P3 had a low sperm concentration (6 million/ml) and sperm concentrations were all in the normal range for the rest of patients. However, all the patients had asthenozoospermia, with <40% sperm motility or <32% progressive motility, except for P8, whose sperm motility and progressive motility were found normal in the routine semen analysis (Table 1). Sperm morphological analysis was performed for P1-P9, and revealed that less than 3.3% of spermatozoa were morphologically normal according to WHO guidelines (Table 1). Sperm morphology was not evaluated for P10 and P11, who refused to do any more test. Hence, P1-P7 and P9 were diagnosed with asthenoteratozoospermia, P8 was diagnosed with teratozoospermia, while P10 and P11 were diagnosed with asthenozoospermia with sperm morphology undetermined.

The morphology of spermatozoa from P1-P9 was further assessed by H&E staining of semen smears. Comparing to the long and threadlike flagella observed in a fertile control, more than 82.2% of the spermatozoa in P1-P9 displayed malformed morphologies of flagella, including absent, short, coiled, bent, and irregular caliber, which are typical presentations of MMAF (Table 1 and Supplementary Figure S3). Moreover, the SEM analysis results of spermatozoa from patients were consistent with those of light microscopy (Figure 1B).

In addition, all the patients declared they have no ciliary-related clinical features (including situs ambiguous, unexplained chronic airway infections, respiratory distress at newborn, chronic middle ear infections, clubbing, hydrocephalus etc.) and none of them wish to participate in any further related examinations. Altogether, all the 11 patients exhibited male infertility due to asthenozoospermia and/or teratozoospermia associated with MMAF.

To explore the underlying genetic cause of MMAF in these patients, we carried out WES of P1-P4, their fertile brother II:4 and mother I:2 from family 1; P5-P8, their father I:1 and mother I:2 from family 2; P9 and his mother III:2 from family 3. Through WES data analysis, we identified a homozygous frameshift variant in CFAP61 (M1), NM_015585.4: c.451_452del (NP_056400.3: p. I151Nfs*13), in family 1, and a homozygous CFAP61 stop-gain variant (M2), NM_015585.4: c.847C > T (NP_056400.3: p. R283*) in family 2 (Supplementary Figure S1). In family 3, 17 variants in 17 genes, including the CFAP61 variant same with that in family 2, were left after variant filtration (Supplementary Figure S1). Among these 17 genes (Supplementary Table S3), the CFAP61 c.847C > T, p. R283* variant is the most likely pathogenic variant. In the gnomAD database, the allele frequencies of the CFAP61 M1 and M2 are extremely low (0 and 0.00004017, respectively). Among the seven different ethnic groups of gnomAD, the allele frequency of M2 is 0.00023 in South Asians, 0.00016 in East Asians, and was absent in the other ethnic groups. Subsequent Sanger sequencing of all the available family members confirmed that the CFAP61 variants were recessively inherited from their heterozygous parents except P11 whose sperm morphology was not yet assessed (Figures 1C–E). These two CFAP61 variants, c.451_452del and c.847C > T, occurring in exon 6 and exon 8, respectively, introduce a premature stop codon and thus are expected to produce either truncated proteins lacking the vast majority of the amino acid residues or no proteins due to nonsense-mediated mRNA decay (Figure 1F). Given that Cfap61 knockout mice was reported to have sperm morphological defects characteristics of MMAF, it is thus suggested that these two CFAP61 variants are likely pathogenic for the MMAF phenotype observed in the patients recruited by us.

Variants in CFAP61 led to severe disorganizations of the axonemal structures in patients.

Since disruption of Cfap61 in mice was associated with the abnormalities of flagellum components (Huang et al., 2020), we thus investigated the ultrastructure of spermatozoa from the patients (P1-P6, P8, and P9) by TEM. In contrast to the typical “9 + 2” axoneme arrangement in the spermatozoa from a fertile control, most of the flagellar cross-sections from spermatozoa of patients exhibited severely disorganized axoneme structure, along with the missings of MTDs, dynein arms, RSs, and/or CP (Figure 2A). On average, 63.2 ± 28.4% of midpiece, 71.9 ± 17.7% of principal piece, and 69.4 ± 21.1% of end piece cross-sections were abnormal (Table 2). Interestingly, among the cross-sections with an identifiable “9 + 2” axoneme conformation in patients’ flagella, 17.4 ± 13.9% displayed misoriented CP compared to zero in controls’ (Figure 2B).

To further investigate the ultrastructural defects observed by TEM, SPAG6, a marker of CP, TSGA2, a marker of RS, and DNAH1, a marker of IDA, were examined on the semen smears from P7. IF staining results revealed signals of SPAG6, TSGA2, and DNAH1 along the full length of sperm tails from a fertile control. In contrary to this, SPAG6 and TSGA2 signals were completely absent in the patient’s spermatozoa (Figure 2C&D). DNAH1 signals were also absent from most of the spermatozoa that were examined, and only 4.8% of spermatozoa displayed normal and 14.4% displayed weak DNAH1 signals (Figure 2E). Thus, these findings confirmed that the axoneme components, CP, RS, and IDA were severely disrupted due to the variants in CFAP61 in patients.

Taking the above results together, patients carrying the CFAP61 variants exhibited decreased sperm motility and morphological and ultrastructural anomalies of sperm flagella, consistent with the phenotype of Cfap61 knockout mice (Huang et al., 2020), indicating that these two CFAP61 variants are indeed pathogenic for MMAF in these patients.

To confirm the pathogenicity of the identified CFAP61 variants on protein expression, we performed IF staining on semen smears obtained from P1 and P7 with an anti-CFAP61 antibody which recognizes human CFAP61 amino acids 689–834. The specific signals of CFAP61, co-localizing with α-Tubulin, were detected on the sperm tails from a fertile control, but CFAP61 signals were absent in the spermatozoa of both patients (Figure 3A), indicating the absence of full-length CFAP61 proteins in the patients.

Since CFAP61 is a member of CSC, and spermatozoa from patients carrying variants in CFAP61 displayed similar defects to those carrying variants in other members of CSC (Auguste et al., 2018; Kherraf et al., 2018; Li et al., 2019; Martinez et al., 2020), we next detected CFAP251 in the spermatozoa from patients. Using IF staining, we observed no CFAP251 signal in spermatozoa from patients in contrast to the signal along the full length of the flagella in spermatozoa from control (Figure 3B). These results collectively suggested that the CSC proteins are likely absent from sperm flagella of patients.