ORIGINAL RESEARCH article

Front. Genet., 04 November 2026

Sec. Livestock Genomics

Volume 17 - 2026 | https://doi.org/10.3389/fgene.2026.1945517

Loss of bmp15 gene function does not abolish fertility in male Atlantic salmon

  • 1. Research Group Reproduction and Developmental Biology, Institute of Marine Research, Bergen, Norway

  • 2. Research Group Reproduction and Developmental Biology, Institute of Marine Research, Matredal, Norway

  • 3. Research Group Population Genetics, Institute of Marine Research, Bergen, Norway

  • 4. Research Group Reproduction and Developmental Biology, Institute of Marine Research, Storebø, Norway

  • 5. Reproductive Biology Group, Division Developmental Biology, Department Biology, Science Faculty, Utrecht University, Utrecht, Netherlands

Abstract

Introduction:

Bone morphogenetic protein 15 (BMP15), a transforming growth factor β (TGF-β) superfamily member, is well known for its crucial involvement in regulating female reproduction in mammals, and recently also in fishes. However, the biological function(s) of Bmp15 in male reproduction are largely unknown in vertebrates in general.

Methods:

In this study, we used CRISPR/Cas9 technology to generate Atlantic salmon bmp15 mutants to investigate its potential role in testis development, spermiation rates and sperm motility.

Results:

No apparent phenotypic differences were observed at either of the two sampling points (18 and 22 months of age) carried out prior to the spawning season, in any of the mutant groups. However, at 22 months of age, mutant males showed a higher frequency of individuals in the rapid testicular growth phase than wild-type (WT) controls (63.6% versus 33.3%). Then, males that remained sexually immature and homozygous for the frameshift mutations (bmp15 fs/fs) were significantly shorter and lighter than WT controls from 26 months of age (at spawning) onwards; they also exhibited a higher incidence of deformed vertebrae and significantly lower testicular transcript levels for two growth factors (amh, gsdf) when examined at the age of 26 months. Regarding testis maturation, 26 month-old fs/fs mutant males showed a higher frequency of spermiating fish than WTs. However, further analyses revealed variation in vgll3a genotype - linked to age at maturity - among the fish investigated, as well as a vgll3a genotype-dependent association with spermiation probability. No significant differences were found regarding the plasma levels of 11-ketotestosterone, the main androgen in fish, irrespective of the maturation stage. Also regarding sperm motility and trajectory, no significant differences were detected between mutants and WTs in the two following reproductive seasons, under the conditions of this study.

Conclusion:

This is the first mutant model in fish investigating Bmp15-mediated action in the male gonad and, similarly to the only other genetic model available in vertebrates to date–the pig –, no remarkable testicular phenotypes were revealed. Future studies will be required to determine whether these effects are evolutionary conserved across vertebrates, or whether distinct or additional phenotipic consequences emerge in other fishes or tetrapods.

1 Introduction

Transforming growth factor β (TGF-β) superfamily members are known to play important roles in the regulation of the reproductive function in a wide array of organisms (Monsivais et al., 2017). Over the last decades, numerous TGF-β superfamily members have been extensively investigated due to their profound involvement in mediating both testicular (Young et al., 2015) and ovarian development (He et al., 2025) across several species. Within the ovary, locally produced oocyte factors received considerable attention since the 90s for their crucial functions regulating various stages of mammalian oogenesis (Dube et al., 1998; Eppig, 2001; Erickson and Shimasaki, 2001). Among those factors, the ovarian function(s) of bone morphogenetic protein 15 (BMP15) are well characterized in various mammalian species, including for example key roles controlling caprine follicular development (Celestino et al., 2011), oocyte maturation (Su et al., 2004) and ovulation (Yan et al., 2001) in mice, and fertility in sheep (Galloway et al., 2000). Also, natural mutations of BMP15 have been associated with primary ovarian insufficiency in humans (Afkhami et al., 2022; Rossetti et al., 2020) and fertility in pigs (Flossmann et al., 2021). Therefore, these species-specific aspects reported in mammals for BMP15-mediated effects seem related to the species’ reproductive strategy, as suggested previously by Otsuka and colleagues (Otsuka et al., 2011). In teleost fish, the zebrafish and Atlantic salmon bmp15 knockouts are the only mutant models available to date. While zebrafish bmp15 mutant females sex-reverse and become fertile males (Dranow et al., 2016; Zhai et al., 2023), loss of bmp15 function results in ovulatory dysfunction in female salmon (Crespo et al., 2024). However, except for stating fertility and apparently normal testicular development, sex-reversed and genetic male bmp15 knockout zebrafish remain largely uncharacterized.

The overall limited understanding regarding potential BMP15 testicular functions primarily relies on its pattern of expression, predominantly localized in ovarian (germ) cells at both gene and protein levels. Yet, BMP15 protein has been detected in pachytene spermatocytes and round spermatids in pig (Tang et al., 2020) and rat testes (Nicholls et al., 2009); in the latter study, however, only the closely related paralog growth differentiation factor 9 (GDF9) regulated Sertoli cell function (Nicholls et al., 2009). On the other hand, catfish Bmp15 protein was detected in somatic (Sertoli and interstitial) cells (Yadav and Lal, 2017). Testicular BMP15/Bmp15/bmp15 mRNA expression has been also identified in various mammalian (Ciller et al., 2016; Nicholls et al., 2009; Aaltonen et al., 1999) and fish species including the Atlantic salmon (Clelland et al., 2006; Crespo et al., 2024; Halm et al., 2008; Kleppe et al., 2017b; Kleppe et al., 2020). Nevertheless, the potential involvement of BMP15 in testis physiology has been rarely investigated in vertebrates. The only male BMP15 genetic model characterized to date is the pig, where constitutive knockdown of BMP15 suggested potential–mild–effects on sperm quality (Tang et al., 2020). However, the mechanism of action remains to be elucidated, while no effects on testis-specific gene expression, downstream signaling or testicular steroidogenesis were identified in this transgenic model (Tang et al., 2020). In Chinese Holstein bulls, a single nucleotide polymorphism (C5697T) in intron 1 of the BMP15 gene altered sperm motility (Sun et al., 2014), and in Rasa Aragonesa sheep, rams carrying the BMP15 FecXR allele showed enhanced massal motility and a greater proportion of rapid sperm than non-carrier individuals (Abecia et al., 2020). In the latter model, no differences in scrotal circumference or testicular length and diameter were found, although FecXR rams produced a higher proportion of pregnant ewes after artificial insemination (Abecia et al., 2020). In contrast to these observations in mammalian vertebrates, no potential bmp15-mediated effects on sperm quality or functionality have been reported in teleost fish to date. Nonetheless, in zebrafish, pharmacological inhibition of Bmp signaling inhibited the differentiation of spermatogonia (Fernandes da Costa et al., 2024), as did mutating the type IB Bmp alk6b receptor (also known as bmpr1bb), which resulted in impaired germ cell differentiation and testicular hypertrophy (Neumann et al., 2011; Yue et al., 2026).

Our previous work in Atlantic salmon has shown a female-biased and germ cell-specific expression of bmp15, and that bmp15 transcript was also found moderately expressed in the male gonads (Crespo et al., 2024; Kleppe et al., 2017b; Kleppe et al., 2020). The germ cell-specific expression of bmp15 in both sexes pinpointed Bmp15 as a candidate target to induce sterility in Atlantic salmon, which was in fact confirmed for females by CRISPR/Cas9-induced loss of bmp15 function (Crespo et al., 2024). The search for alternative sterility models in the salmon farming industry may provide approaches to possible solutions for major challenges in finfish aquaculture related to fish welfare as well as environmental impacts, e.g., concerns of unwanted precocious sexual maturation in males (Taranger et al., 2010) and escapee-mediated introgression into wild populations (Bolstad et al., 2021). Hence, the aim of the present study was to investigate the potential function of bmp15 in testis development and fertility in male Atlantic salmon.

2 Materials and methods

2.1 Ethical statement

Atlantic salmon used in this study were reared under standard conditions and sampled at Matre Aquaculture Research Station (Matredal, Norway). All use of animals for the present study was carried out in accordance with the Norwegian Animal Welfare Act of 19th of June 2009, in force from 1 January 2010. The experiments were approved by the Norwegian Food Safety Authority (https://www.mattilsynet.no, permit numbers 14865 and 22082).

2.2 Guide RNA (gRNA) and Cas9 mRNA production

The bmp15 CRISPR target site and suitable oligonucleotides for targeting were predicted using an online tool provided by http://zifit.partners.org/ZiFiT/ (Hwang et al., 2013), as published previously (Crespo et al., 2024). The Atlantic salmon reference genome was screened for regions sharing homology with the CRISPR target sequence using BLASTN (Altschul et al., 1990) to ensure that there were no genomic regions likely to be subjected to unwanted off-target cleavage. The gRNA was designed to target exon 1 of the bmp15 gene (GenBank ID 106578107, chromosome 18), in a region covering the putative cleavage site of the Bmp15 propeptide (amino acids 310-313, RLRR; (Crespo et al., 2024)), as previously described (Duckert et al., 2004; Halm et al., 2008). Subsequently, gRNA and Cas9 mRNA were produced following the protocol described by Gagnon and colleagues (Gagnon et al., 2014), with minor modifications (Andersson et al., 2024). In addition to bmp15, slc45a2 was also mutated, causing lack of pigmentation which was used as a screening tool for mutation efficiency of the target gene (Edvardsen et al., 2014; Wargelius et al., 2016).

2.3 Microinjections and fish rearing

Fish rearing and all experiments took place at the Institute of Marine Research, Matre Research Station, Matredal, Norway. Atlantic salmon eggs and sperm were obtained from AquaGen AS (Trondheim, Norway). In November 2015, fertilized eggs were genome edited using CRISPR/Cas9 to target the bmp15 gene. Briefly, bmp15 and slc45a2 gRNAs, and Cas9 mRNA were simultaneously co-injected as previously described (Edvardsen et al., 2014). The gRNAs directed the resulting Cas9 protein to induce a double stranded cut in slc45a2 and bmp15 genes, causing mutations. The resulting F0 generation was reared in indoor tanks under standard rearing conditions until spawning in Autumn 2018.

To generate an F1 generation, eggs from two highly mutated females were fertilized with sperm from four highly mutated males at two different dates. An overview of the time points, crosses, egg and sperm volumes used, as well as broodstock bmp15 mutation rate and type were previously described in (Crespo et al., 2024). Non-edited eggs and sperm were used to generate a WT control group in November 2018. In March 2019, all the bmp15 mutated fish and WT controls were combined in a common tank and introduced to external feeding. Throughout their life cycle, the fish were reared in indoor tanks (tank size was increased with increasing fish size) in freshwater (parr stage), followed by seawater (after smoltification) and finally brackish water. Light and temperature were natural and ambient, respectively, and the fish were fed ad libitum with a standard commercial diet throughout the study period. In June 2020, all individuals were PIT-tagged and fin clipped for DNA analysis.

2.4 DNA isolation, sex evaluation and mutation analysis

From each fish, genomic DNA was isolated from fin tissue using the DNAdvance Kit and a Biomek i5 automated workstation (Beckman Coulter Life Sciences), according to the manufacturer’s instructions. Genetic sex determination was carried out by qPCR genotyping of the sexually dimorphic on the Y-chromosome (sdY) locus (Ayllon et al., 2019). Vestigial-like family member 3 (vgll3a), associated with age at maturation in Atlantic salmon (Ayllon et al., 2015; Barson et al., 2015), was genotyped using a qPCR-based allelic discrimination assay as previously described (Ayllon et al., 2019). For bmp15 mutation analysis, extracted DNA was amplified in a two-step nested barcoding PCR as described by Gagnon et al. (Gagnon et al., 2014). Detailed information on gene-specific primers for the first PCR, library sequencing and filtering and assembly of the sequenced reads can be found in (Crespo et al., 2024). Reads were then individually mapped to the reference amplicon sequence using Muscle (v3.8.1551 (Edgar, 2004)) and divided into categories based on presence of indels resulting in frameshift or in-frame mutations, or no indels.

2.5 Measurement of body metrics and tissue sample collection

Throughout the experiment, body length and weight were measured and tissue samples collected at different time points as shown in Figure 1A. Samplings of body length and weight were carried out after anesthetizing the fish with 100 mg/L Finquel vet. (tricaine methanesulfonate). At larger fish sizes, clove oil (1 mL/100 L Aqui-S vet.) was used as a sedative. Prior to all end-point samplings, the fish were anesthetized with 200 mg/L Finquel vet., sacrificed by cutting the medulla oblongata and blood and tissue samples collected. Blood samples were centrifuged at 16,000 x g at 4 °C for 2 min, plasma was collected and stored at −80 °C until further analyses (see Plasma steroid hormone levels). Testis tissue samples for gene expression analyses were placed in RNAlater (ThermoFisher Scientific, Waltham, MA) and stored at −20 °C until use. For histological analysis, testis tissue was fixed in 4% glutaraldehyde or Bouin’s solution at 4 °C overnight, rinsed and stored in 70% ethanol until further processing.

FIGURE 1

2.6 RNA isolation, cDNA synthesis and qPCR analysis

Total RNA was extracted from testis tissue pieces using the Maxwell HT simplyRNA Kit (Promega), according to the manufacturer’s instructions. For pipetting, a BioMek 4000 instrument (Beckman Coulter, Brea, CA) was applied. The samples were DNase-treated as part of the RNA extraction procedure. For each sample, RNA concentration and purity were checked using a NanoDrop spectrophotometer (ThermoFisher Scientific).

cDNA was synthesized from 500 ng RNA, using the VILO cDNA synthesis kit (Invitrogen, Waltham, MA). qPCRs were performed in duplicates in 384-well optical plates in a QuantStudio 5 Real-Time PCR system (ThermoFisher Scientific) using default settings. 2 μL of cDNA (1/40 diluted) was used in 6 μL Fast TaqMan or Fast SYBR Green master mixes. All primer sets used for qPCR were previously published (Crespo et al., 2024). No-template controls for each gene were run in all qPCR plates. The 2−ΔΔCt method (Livak and Schmittgen, 2001) was used to calculate the expression of each gene relative to the reference gene elongation factor 1 alpha (ef1a (Olsvik et al., 2005)). Samples from mature fish (spermatid and spermatozoa stages; see below) were not included for gene expression analysis due to poor and variable RNA quality and quantity.

2.7 Plasma steroid hormone levels

Sex steroids were extracted from blood plasma by a method modified from (Pankhurst and Carragher, 1992). The extracted and dissolved steroids were stored at −20 °C until analysis and 11-ketotestosterone (11-KT) levels were then measured by an enzyme-linked immunosorbent assay (ELISA (Cuisset et al., 1994)), previously validated for Atlantic salmon (Vikingstad et al., 2016).

2.8 Analysis of vertebra health

Since members of the BMP family are known to play a role in bone metabolism (Chen et al., 2021; Kamiya and Mishina, 2011), we examined deformities in the vertebral column (which are associated with reduced welfare and growth in salmon (Hansen et al., 2010)). Radiographs of sedated live male post smolts at 26 months of age were taken with a computed radiography system (CR35 VET, DÜRR NDT GmbH & Co. KG) using a 35 cm × 43 cm image plate (IP 35/43, DÜRR NDT GmbH & Co. KG) and a portable X-ray unit (HiRay PLUS, model Porta 100 HF, JOB Corporation, Yokohama, Kanagawa, Japan) at 85 cm distance with 40 kV and 4 mAs. Each radiograph was examined for vertebra deformities, and the type of deformity was assessed visually as described elsewhere (Witten et al., 2009), by an observer blinded to the genotype of each fish. For all groups, the proportion of fish showing 1 ≤ deformed vertebra (DV) and/or malformations in neural and haemal spines or ribs (DS) was calculated for statistical analysis.

2.9 Testicular histology

Testis tissue pieces for histological analysis were dehydrated and embedded in resin (Technovit 7100, Kulzer) or paraffin when fixed in 4% glutaraldehyde (22 months-old sampling) or Bouin’s solution (all other samplings), respectively, according to standard protocols. Histological examination was carried out using 4 µm sections after toluidine blue staining (resin) or 3 µm sections after haematoxylin–eosin (paraffin), and slides were scanned using a digital slide scanner (NanoZoomer, Hamamatsu Photonics, Hamamatsu, Shizuoka, Japan) in ×40 mode (4.53 pixels/µm). Male fish were then classified according to the stage of spermatogenesis, as previously described (Luksenburg et al., 2026; Melo et al., 2014). To this end, sections were analyzed qualitatively for the presence and prevalence of the different germ cell types. Regarding spermatogonia, we distinguished undifferentiated type A spermatogonia (Aund; apparently single cells with no or very little nuclear heterochromatin, one or two prominent nucleoli, and a nuclear diameter of ∼14 μm), differentiating type A spermatogonia (Adiff; at least two or small groups of cells with some heterochromatin and a nuclear diameter of ∼10 μm; males showing Aund and/or Adiff as furthest developed germ cell type were assigned to SPG A stage), and type B spermatogonia (groups of cells with a nuclear diameter of ∼8 μm and a high amount of heterochromatin; SPG B stage). Primary and secondary spermatocytes were combined as spermatocytes (SPC stage). Germ cells during spermiogenesis were collectively referred to as spermatids (SPT stage). Spermatozoa found free in the lumen of spermatogenic tubules indicated the completion of the spermatogenic process for a fully mature cyst that opened to release spermatozoa (i.e., spermiation) into the tubular lumen (SPZ stage). The analysis resulted in assigning a specific stage to a testis sample, according to the furthest developed germ cell type found.

2.10 Evaluation of sperm motility at spawning

Sperm motility was used as a proxy for sperm quality, based on the procedures described in Lorillard and colleagues (Lorillard et al., 2026), with modifications. The milt of bmp15-edited and WT male salmon was collected at two consecutive spawning seasons (Autumn of 2021 and of 2022) corresponding to 36 and 48 months of age. Salmon sperm was activated by adding 2 mL of distilled water to a tube containing 2.5 µL of milt. The exact time of activation was recorded as an audio cue in the corresponding video recording. Immediately after activation, 2.5 µL of the sperm suspension was transferred to a Leja 20 µm two-chamber slide and examined using a Leica DM1000 microscope equipped with a 10× objective, negative phase contrast, and an Infinity5-5C camera. Videos were recorded at 60 frames s−1, 1280 × 720-pixel resolution, with a calibrated image resolution of 1.365 pixels µm−1.

For each recording, the time interval between sperm activation and the first stable video frames containing spermatozoa was determined. Based on these observations, 8 s post-activation was selected as the standardized starting time for motility analysis across all recordings. All recordings were performed in a climate-controlled room at approximately 8.1 °C. Sperm motility was analysed using ImageJ v1.53 and the OpenCASA v2 plugin (Alquezar-Baeta et al., 2019). For each recording, video frames corresponding to the interval from 8 to 9 s post-activation were analysed. The parameters examined included the proportion of motile spermatozoa (motility), curvilinear velocity (VCL), straight-line velocity (VSL), average path velocity (VAP), linearity (LIN), wobble (WOB), and straightness (STR). Objects with an area >30 μm2 were included as spermatozoa, and those with a VCL ≥10 μm s−1 were classified as motile. The minimum track length was set to 10 frames, and the maximum permitted displacement between consecutive frames was 10 µm. A four-frame window was used for calculation of VAP.

2.11 Statistical analysis

GraphPad Prism 9.0.0 package (GraphPad Software, Inc., Boston, MA) was used for selected statistical analyses. Datasets were first tested for normal distribution using a D’Agostino-Pearson normality test, unless the number of individuals investigated (n) was insufficient to perform the test. For datasets with normal distribution, significant differences between groups were identified using unpaired Student’s t-test or one-way ANOVA followed by Tukey’s test for multiple group comparisons, as appropriate. For datasets with no normal distribution or with too low n (n < 8) to test for normality, the non-parametric Mann–Whitney test or Kruskal–Wallis test followed by Dunn’s multiple comparisons test was applied, respectively. Results are represented as mean ± SEM. Regarding the prevalence of vertebra deformities and spermiation frequencies, two-sided Fisher’s exact tests were applied to test for significant difference between the WT and bmp15-edited fish groups using 2 × 2 contingency tables for pairwise comparison.

To evaluate the independent effects of bmp15 fs mutations on spermetation and whether the effect of vgll3a was consistent with the previously described sex-biased dominance pattern in male Atlantic salmon (Ayllon et al., 2015; Barson et al., 2015), we first fitted univariate logistic regression models with spermiation status as the response and bmp15 and vgll3a genotypes (fs/fs, if/fs, if/if, WT and EE, EL and LL, respectively) as the predictor. Pairwise comparisons among the different genotypes were subsequently obtained from estimated marginal means using the emmeans package (Lenth and Piaskowski, 2026). The joint effects of bmp15 and vgll3a were then assessed, with spermiation status as the binary response and bmp15 knockout allele status (KO, presence versus complete absence of an fs allele), vgll3a genotype (EE/EL versus LL) and their interaction as explanatory variables:where denotes the probability of spermiation for individual . In all cases, a Firth´s logistic regression model was fitted using the package logistf (Heinze et al., 2025). A penalized likelihood approach was used, which likelihood approach reduces small-sample bias and provides stable estimates in the presence of sparse genotype combinations and quasi-separation (Firth, 1993; Heinze and Schemper, 2002). Model fit was assessed using simulation-based residual diagnostics implemented in the DHARMa R package (Hartig, 2026). All logistic regression analyses were conducted in R (R Core Team, 2025). A complete, knitted version of the regression analysis is included as Supplementary Material.

3 Results

3.1 bmp15 F1 mutant males: mutation characterization and body metrics

We have reported previously in Atlantic salmon the preferential ovarian and germ cell-specific expression of bmp15, however, bmp15 transcript was also found significantly expressed in the testis (Crespo et al., 2024; Kleppe et al., 2017b; Kleppe et al., 2020). The present study revealed a transient increase of testicular bmp15 levels in 22 months-old males (Figure 1B), a few months before spermiation in the experimental conditions investigated. However, due to poor and variable RNA quality and quantity, we cannot exclude changes in bmp15 transcript levels during more mature stages.

To investigate the potential involvement of Bmp15 in regulating salmon spermatogenesis, we generated bmp15 mutants using CRISPR/Cas9. Detailed information on gRNA design, F0 crispant crosses for establishing a F1 bmp15 mutant generation and rearing conditions can be found in (Crespo et al., 2024). In that previous study, we reported that crosses of highly mutated bmp15 broodstock fish resulted in elevated F1 embryo mortality, with certain crosses yielding no viable offspring. Therefore, the number of F1 male individuals for investigation was limited in the current study. A total of 53 WT and 76 bmp15 mutant males were studied, falling into three subgroups of 18 if/if, 26 if/fs and 32 fs/fs mutant males, respectively (Figure 1C; Supplementary Table S1). Additionally, two heterozygous bmp15 mutant males were identified, however due to their low number this subgroup did not receive further attention. Regarding mutation type analysis, F1 bmp15 mutants displayed a mix of genotypes (see Supplementary Table S1 for detailed information) due to the mosaic nature of the F0 founders used (Crespo et al., 2024). The most common bmp15 variants identified in F1 fin clips were an in-frame deletion of three nucleotides (del3) or of six nucleotides (del6). However, a range of different variants was found in bmp15 fs/fs males (Figure 1D). Body length and weight of the whole fish batch were measured at three different time points (18, 22 and 26 months of age). While no effects were observed in 18 and 22 months-old–all immature–fish, only bmp15 fs/fs mutant males showed deviant phenotypes (significantly smaller and thinner) compared to WT controls at 26 months of age in their first spawning event (Figures 1E,F). Further stratified–maturation-matched–statistical comparisons of body length and weight revealed that the body size difference remained significant within the immature subgroup exclusively (Supplementary Figures S1, S2), an effect consistently found in all samplings from 26 up to 40 months of age in bmp15 fs/fs mutant males (Supplementary Figures S1, S2). In contrast, a sampling at 48 months of age showed an opposing maturation-dependent pattern, with body length and weight significantly reduced only in spermiating bmp15 fs/fs mutant males (Supplementary Figures S1, S2), however the limited number of immature fs/fs fish precluded formal testing for this subgroup.

Moreover, gonadal samples were collected at 18, 22 and 26 months of age to investigate whether editing of the bmp15 gene affected testis development in Atlantic salmon (see below). Since the only other BMP15 genetic (male) model characterized in vertebrates to date–the pig–displayed no notable testicular phenotypes apart from reduced sperm viability (Tang et al., 2020), we decided to utilize most fish available to investigate the potential effects of bmp15 gene loss in late stages of sexual maturation.

3.2 Preliminary evaluation of post smolt bmp15 mutant males at 18 months of age

First, we sampled randomly a limited number of immature F1 males after transfer to sea water, i.e., the post smolt stage. However, the findings reported here are preliminary and require validation in larger cohorts or subsequent filial generations. While no obvious effects were recorded for body length and weight, certain–not significant–discrepancies in plasma 11-KT and testicular transcript levels were found within 18 months-old bmp15 mutant males. The if/if mutant group showed increased 11-KT plasma levels (Supplementary Figure S3A) and high testicular igf3 and low amh transcript levels compared to WT counterparts (Supplementary Figure S3B). On the other hand, if/fs mutant males exhibited an opposite tendency for the parameters investigated (Supplementary Figures S3A,B). Nevertheless, analysis of testis tissue sections suggested that both 18 months-old if/if and if/fs mutant groups displayed germ cell-poor testes compared to WT controls that showed the normal germ cell numbers of immature males, or were transiting (increased mitotic activity) into puberty (Supplementary Figure S3C). The morphological observations seemed in line with lower transcript levels of the germ cell-specific factors vasa and gdf9 in both if/if and if/fs mutant testes (Supplementary Figure S3B). In contrast, all bmp15 fs/fs fish showed testes transiting into pubertal development, in agreement with the levels of testicular transcripts investigated (high vasa and igf3, low amh) and 11-KT in plasma (Supplementary Figures S3A,B), and an apparent increase in the preponderance of interstitial (somatic) cells (Supplementary Figure S3C). All mutant groups showed lower gsdf transcript levels compared to WT controls (Supplementary Figure S3B).

3.3 First indication of a potentially increased rate of maturation in bmp15 mutants

In a later sampling (i.e., 22 months of age), body length and weight, and GSI values did not differ in sampled WT and bmp15 mutants, irrespective of the maturation stage (Figures 2A–C; Supplementary Figures S4A,B). Expression levels of selected genes were also unaffected in all groups investigated (Figure 2D). However, the fraction of sampled males exhibiting well developed testes and spermatogenesis until the spermatid (SPT) stage, was higher in mutants compared to WT controls (63.6% versus 33.3%; Figures 2E,F). The effect seemed clearest in the bmp15 fs/fs group, where individuals at SPT stage represented 75% of all fish (yellow dots in Figures 2A–C). However, the total number of individuals available for these analyses was low and precluded reaching statistical significance. Histological analysis of testis tissue showed that also immature fish were present in all groups (showing type A spermatogonia only; Figure 2F), unlikely to mature during this season.

FIGURE 2

3.4 Loss of bmp15 function did not directly affect spermiation rates at spawning

All remaining fish were examined by abdominal massage for strippable sperm (scoring non-spermiating versus spermiating) in December of 2021, at 26 months of age. The estimated odds of spermiation showed a gradual increase from WT through if/if and if/fs to fs/fs males (Figure 3A; Supplementary Figure S5; Supplementary Table S2A), consistent with an increased effect of bmp15 loss-of-function fs allele dosage, with 3-fold change in the spermation odds for homozygous fs/fs individuals compare to WTs (adjusted CI = 0.76–12.02). However, this analysis did not account for potential effects of, or interactions with, the vgll3a genotype, associated with age at sexual maturaty in Atlantic salmon (Ayllon et al., 2015; Barson et al., 2015). Therefore, all individuals were genotyped for vgll3a, which revealed variation in vgll3a genotype among the fish included in this study (Supplementary Figure S6). In the subsequent univariate logistic regression analysis, the vgll3a genotype was significantly associated with marked differences in spermiation probability (Supplementary Table S2). Thus, EL and EE males showed 5.23-fold (adjusted CI = 1.30–21.10) and 4.63-fold higher odds of spermiation, (adjusted CI = 1.17–18.42) than LL individuals, respectively. No differences in the odds of spermation were detected between EL and EE individuals (OR = 1.13, adjusted CI = 0.35–3.60), supporting the subsequent grouping of EE and EL into a single early maturation category as opposed to LL or late maturation category. The joint effects of bmp15 KO allele status (presence [fs/fs and if/fs groups] versus complete absence of an fs allele) and vgll3a maturation category were then evaluated. The effect of the vgll3a early allele showed a dominat influence on the spermation probability in both WT and KO (fs carrier) groups (Supplementary Figure S7). For example, among the WT fish, carriers of the E allele showed 4.1 higher odds (adjusted CI = 0.8004–20.98) of spermation than their LL counterparts. Interestingly, among the E allele carriers, WT individuals had an estimated 20% lower odds of spermiation than fish carrying at least a bmp15 fs allele (OR = 0.8, 95% CI = 0.21–3.03). There was no statistical evidence that the effect of bmp15 fs status differed between the two vgll3a maturation categories (interaction OR = 0.84, 95% CI = 0.03–137.8). Thus, although the overall model was significant and the observed spermiation frequencies were higher in bmp15 fs allele carriers, the adjusted analysis did not provide statistical evidence for either a bmp15 fs effect for the vgll3a genotypes or a bmp15 × vgll3a interaction.

FIGURE 3

In all groups, spermiating fish exhibited the typical external appearance in terms of coloration and body shape (i.e., darker skin pigmentation, smaller size and soft abdomen; Figure 3B), and fully developed testes (Figure 3C). Also, all fish were radiographed to evaluate potential differences in the development of the vertebral column. The prevalence of fish showing vertebral and/or spinal and rib deformities was higher in all bmp15-edited groups compared to WT controls, reaching statistical significance in the bmp15 if/fs and fs/fs groups (p = 0.0041 and p = 0.0060, respectively; Figure 3D). A similar trend was found when splitting the data into immature and spermiating groups, however, only the WT-if/fs comparison reached statistical significance when considering immature fish (p = 0.0136; Figures 3E,F).

We also randomly selected mutant and WT control males for a terminal sampling. In agreement with the effects found when examining the whole fish batch (Figures 1E,F; Supplementary Figures S1, S2), we again observed reduced body length and weight in this subsample of bmp15 fs/fs fish (Figures 4A,B). However, after data stratification by maturation stage, the significant difference in body size was observed only in the immature fs/fs mutant subgroup (Supplementary Figure S8A,B). Mean GSI values and plasma 11-KT levels did not differ between genotypes when combining all maturation stages (Figures 4C,D), nor when splitting the data into immature and spermiating groups (data not shown). However, a slightly higher proportion of spermiating fish was recorded in bmp15 mutants compared to WT controls (51.7% versus 44.0%; identified by red dots in Figures 4A–D). Regarding the non-spermiating fish, most males showed testes filled with SPG A, either displaying a low, or a high level of mitotic activity (referred to as immature and transiting, respectively), independent on the genotype (Figure 4E). Nonetheless, a higher proportion of transiting testes was observed in bmp15 fs/fs fish (50.0% versus 14.3% in the WT control group; Figure 4E). Only one fish showed a deviant stage of spermatogenesis, a bmp15 if/if male exhibited large, SPT stage testes (Figure 4E). Interestingly, significantly lower levels of the Sertoli cell-specific transcripts amh and gsdf (and higher igf3, although not significantly) were found exclusively in non-spermiating bmp15 fs/fs mutants; expression levels of the germ cell-specific transcripts vasa and gdf9 were unaffected (Figure 4F).

FIGURE 4

3.5 Editing of the bmp15 gene did not affect sperm motility parameters

To investigate whether mutation of bmp15 gene affects sperm motility and viability, we examined different parameters in milt samples collected in two consecutive spawning seasons (Autumn of 2021 and of 2022) corresponding to 36 and 48 months of age. In our analyses of sperm motility and trajectory parameters, no statistically significant differences were detected between any of the genotypes or time points, however, high variability was observed within the samples examined in all groups (Figures 5A,C; Supplementary Figure S9A,B). That includes several milt samples, collected from both 36 and 48 months-old spermiating males, that showed too low sperm concentrations for proper analysis (red dots overlapping with the X-axis; Figures 5A,C; Supplementary Figure S9A,B). Examples of frames recorded during sperm activation are shown in Figures 5B,D.

FIGURE 5

4 Discussion

4.1 Generation and analysis of F1 bmp15 mutants in Atlantic salmon

Targeting the salmon bmp15 gene generated a mosaic array of mutations (including in-frame and loss-of-function frameshift variants) in F0 fish. While no obvious reproductive phenotypes were observed up until spawning, offspring from crosses using gametes from highly mutated F0 broodstock fish showed low, and in case of a highly mutated F0 female even no embryo survival (Crespo et al., 2024). Survivors from non-lethal crosses allowed establishing an F1 generation exhibiting normal growth and sex ratio (Crespo et al., 2024) different from the only other bmp15 mutant model described in fish to date. Genetic removal of bmp15 resulted in female-to-male sex change in zebrafish following the failure to upregulate ovarian estrogen production (Dranow et al., 2016; Zhai et al., 2023). Thus, the germ cell-derived factor Bmp15 does not seem involved in sex differentiation in Atlantic salmon, possibly because sex differentiation of the gonadal somatic cells follows the genetic sex and is not affected by the loss of germ cells or endogenously produced sex steroids (Kleppe et al., 2017a; Wargelius et al., 2016; Yano et al., 2012).

Equivalent numbers of male salmon carrying if/if, if/fs or fs/fs genotypes (ranging from 18 to 32 fish per group) were identified in the F1 offspring generated from bmp15 crispant founders. Analysis of DNA samples revealed that while multiple bmp15 allele variants were present in loss-of-function (fs/fs) mutants, a moderate number of variants was observed in fish carrying in-frame mutations. Yet, the two most frequent bmp15 mutations within the F1 generation were in-frame (del3 and del6 deletions) but did not induce obvious consequences in male reproductive biology, as previously observed in female salmon (Crespo et al., 2024). Overall, bmp15 if/fs mutants showed relatively few phenotypic differences from WT controls across the parameters investigated. However, potential gene-dose effects, and/or partial dysfunction of the if allele itself, cannot be excluded for certain phenotypes, including an increased incidence of spinal deformities observed in 26 months-old if/fs males. Consequently, the if alleles should not necessarily be considered functionally equivalent to the WT allele. Yet, based on the findings reported here together with our previous work on bmp15 mutant females (Crespo et al., 2024), we suggest that homozygosity for frameshift mutant variants of the bmp15 gene may be required to induce any effect during salmon spermatogenesis (see next sections). Nonetheless, the effects of in-frame mutations on synergy and heterodimerization with the closely related TGF-β paralog growth differentiation factor 9 (GDF9) are unknown in Atlantic salmon, and further molecular and functional characterization of individual mutant alleles are required to establish solid genotype–phenotype interpretations. In mice and human ovaries, BMP15:GDF9 heterodimers are indeed the most potent activator, compared to homodimers, of downstream signaling pathways (Peng et al., 2013). Here, we designed a gRNA targeting a region preceding the putative cleavage site of the Bmp15 propeptide and our results suggest no potential Gdf9-mediated compensatory effects in salmon since gdf9 transcript levels remained unaffected in all bmp15 mutant groups and most time points compared to WT controls (except for the–not significantly–reduced gdf9 expression in mutants at 18 months of age). In agreement with this observation, testicular GDF9 mRNA (and protein) expression was unaffected by constitutive knockdown of BMP15 in a transgenic pig model (Tang et al., 2020). Nonetheless, gdf9 expression levels in WT salmon testes seem to mimic those observed for bmp15 (i.e., transient peak at 22 months of age; data not shown).

4.2 bmp15 and its potential effects on spermatogonial development

While found at lower levels than in ovarian tissue, bmp15 transcript is present in Atlantic salmon testis tissue, and comparison of sexually immature (SPG A only) WT and germ cell-free (dnd mutant) gonads revealed its germ cell-specific expression in the testis (Crespo et al., 2024; Kleppe et al., 2017b; Kleppe et al., 2020). These results are compatible with a potential function for Bmp15 in regulating testicular physiology in salmon. In zebrafish, pharmacological inhibition of all Bmp signaling nullified the stimulatory effect of Fsh on spermatogonial differentiation in primary cultures of adult testis tissue, such that the proportion of areas occupied by type Aund and type Adiff spermatogonia increased and decreased, respectively (Fernandes da Costa et al., 2024). It therefore appears that part of the Fsh effects on spermatogenesis may be mediated by Bmp signaling in this model species. The relevance of Bmp signaling for zebrafish spermatogenesis is also supported by a mutant model, in which the loss of the type IB bone morphogenetic protein receptor alk6b impaired germ-cell differentiation and caused accumulation of undifferentiated gonial cells resembling type Aund spermatogonia (Neumann et al., 2011; Yue et al., 2026). On the contrary, in immature catfish testes, Bmp15 protein showed high levels of expression in Sertoli and interstitial cells (unlike mammals and salmon) before the start of spermatogonial proliferation, and the authors suggested an alternative role for Bmp15 in promoting renewal of spermatogonial cells (Yadav and Lal, 2017). The loss of bmp15 did not result in a spermatogenesis phenotype in salmon, akin to the effects obtained with the Bmp signaling inhibitor, or akin to the loss of alk6b in zebrafish. Contrarily, our observations on bmp15 fs/fs mutants suggest that the spermatogenic process was accelerated along the complete developmental trajectory. Since this seems incompatible with previous observations in zebrafish, one hypothesis (although speculative) is that Bmp15 in salmon does not bind solely to Alk6b and may use other Tgf-β ligand type I and/or type II receptors, yet to be identified and not sensitive for the pharmaceutical inhibitor used in the above-mentioned zebrafish study. In rodent ovaries for example, while BMP15:BMP15 homodimers are inactive, BMP15:GDF9 heterodimers bind to a BMPR2-ALK4-ALK6 receptor complex to regulate granulosa cell function (Peng et al., 2013). A study in rats found co-expression of BMPR2, ALK5 and ALK6 in the seminiferous epithelium, limited to Sertoli cells and round spermatids, but only GDF9 seemed to modulate key Sertoli cell functions (Nicholls et al., 2009). Moreover, constitutive knockdown of Bmp15 did not induce changes in BMP receptor expression in boar testes (Tang et al., 2020). Thus, there is currently a lack of conclusive information regarding the BMP15-induced mechanism of action within the vertebrate testis and species-specific effects cannot be ruled out, as previously reported for female gonads (Otsuka et al., 2011).

In our study, at 18 months of age (SPG A only stage), bmp15 fs/fs fish generally showed reduced testicular transcript levels of the Sertoli cell-specific Tgf-β factor gsdf (Sawatari et al., 2007) and the spermatogonial proliferation and differentiation inhibitor amh (Morais et al., 2017; Skaar et al., 2011) compared to WT controls. In contrast, testicular igf3 and plasma 11-KT levels, factors known to stimulate spermatogonial proliferation and differentiation in fish (Crespo et al., 2019; Melo et al., 2015; Skaar et al., 2011), were equal or higher to those observed in transiting WT counterparts. These observations are in agreement with the testicular histology of fs/fs fish, all showing activated mitotic activity and increased numbers of interstitial (including steroid-producing Leydig) cells. While similar results on growth factor expression and steroid levels were found in the if/if mutant group, if/fs fish showed opposite trends, and both groups exhibited underdeveloped (germ cell-poor) testes in sharp contrast to the transiting testes found in loss-of-function fs/fs fish. Future studies in subsequent filial generations are required to clarify the impact of selected if mutations only and in combination with fs mutations, and their correlation with the differing phenotypes observed. Overall, the number of individuals available for analysis at this sampling point appears insufficient to draw firm conclusions on potential effects of Bmp15-mediated action on spermatogonial development in Atlantic salmon.

4.3 Spermiation rates and sperm quality are not directly affected by bmp15 mutation

Consistent with the activated testes found in the immature stage, fs/fs mutants exhibited a higher tendency to proceed with spermiogenesis 4 months later compared to WT controls (75% versus 33.3%). Interestingly, time-course analysis at transcript level revealed that bmp15 expression peaked during early stages of spermiogenesis in WT male salmon; bmp15 levels remained low in the immature–type A spermatogonia only–stage (18 months-old testes) until a slight but not significant upregulation during early spermiogenesis (22 months-old testes), and then declined to basal levels at spawning (26 months-old testes). In pig and rat testes, highest BMP15/Bmp15 transcript expression was localized in pachytene spermatocytes and round spermatids (Nicholls et al., 2009; Tang et al., 2020). Based on these observations, we speculate that Bmp15 might be involved in gate-keeping the entry into the haploid phase of spermatogenesis in salmon, thereby allowing first large numbers of spermatogonia/spermatocytes to accumulate, until Bmp15 downregulation promotes the individual cysts to complete development. Here, we did not investigate the specific subcellular expression of bmp15 but its transcript expression peaks in a stage of development, for fish showing maturing testes, characterized by the abundant presence of haploid germ cells (see lower panels in Figure 2F). Four months later, at spawning, significant effects on increased spermiation rates and reduced body length and weight were observed in fs/fs mutants compared to WT controls. Nonetheless, stratified statistical comparisons of body length and weight accounting for maturation status revealed that the body size difference remained significant within the immature non-spermiating fs/fs mutant subgroup exclusively. This fish group also showed a higher proportion of transiting testes (50.0% versus 14.3%) characterized by low amh/gsdf and high igf3 transcript levels, similar to the above-mentioned preliminary findings observed in (the limited number of) immature fs/fs mutant males at 18 months of age. Together, our results indicate that the loss of a germ cell-specific gene (i.e., bmp15) elicited the reduced expression of Tgf-β factors in Sertoli cells, in turn suggesting a germ-to-soma (Sertoli) cell signaling involved in the regulation of spermatogonial development.

Since the fish used in this study were not full-siblings and differed in their genetic background, we genotyped the vgll3a locus known to explain 33%–39% phenotypic variation in age at maturity in Atlantic salmon (Ayllon et al., 2015; Barson et al., 2015). While WT controls were mostly homozygous for the late-maturing (L) variant, most of the bmp15 mutant males carried the early-maturing (E) variant (see Supplementary Figure S6). Thus, the difference observed in spermiation frequency may be influenced by the over-representation of LL individuals among the WTs, still showing the expected vgll3a genotype-dependent pattern with significantly higher maturation probability in EE and EL individuals than in LL individuals, as described by Barson et al. (2015). In contrast, bmp15 KO (fs) allele had a non significant mild effect on spermation rates when accounting for the vgll3a genotypes, suggesting that the increased spermiation frequency found in bmp15 fs/fs fish is at least partly attributable to the loss of bmp15 gene function although the relative contributions of bmp15 and vgll3a cannot be resolved conclusively from the present dataset.

Despite the apparent increase in spermiation rates recorded in 26 months-old fs/fs mutants, no notable effects on 11-KT plasma levels were found prior to or during spawning. Regardless of the bmp15 genotype, 11-KT measurements consistently reflected the stage of maturation for all fish investigated (see Figure 4D; Supplementary Figures S8A,B). This seems in agreement with data reported in pigs, where constitutive knockdown of BMP15 did not affect testicular levels of steroidogenic enzymes involved in testosterone synthesis (Tang et al., 2020). Nonetheless, while in salmon sperm motility and trajectory parameters in following reproductive seasons showed equally variable values in all groups and no potential effects associated with Bmp15 signaling, sperm viability (although not density or fertility) was decreased in the transgenic pig model (Tang et al., 2020). One limitation of our study is the lack of data on sperm functionality in terms of the ability to fertilize eggs and generate (viable) offspring. However, we have previously shown that four individual salmon males carrying high frequencies of bmp15 fs mutations (i.e., F0 crispants) could fertilize eggs from a bmp15 crispant female generating the F1 male offspring studied here and the F1 female offspring reported by Crespo and colleagues (Crespo et al., 2024). Furthermore, F2 bmp15 mutants were generated from certain crosses using F1 fs/fs males and females, albeit showing high mortality and incidence of malformations (Crespo et al., 2024). These observations suggest that sperm from bmp15 fs/fs salmon males retain the capacity to fertilize eggs and generate offspring to a certain extent, while future studies will have to properly quantify sperm quality parameters using standard WT eggs. Overall, the biological role of BMP15/Bmp15 appears to be less critical for the regulation of male reproduction and fertility than for females in both salmon and pigs (Crespo et al., 2024; Qin et al., 2019), as well as in several other vertebrates (Dranow et al., 2016; Yan et al., 2001).

4.4 Increased incidence of vertebral deformities in post smolt bmp15 mutant males

Although bmp15 mutant salmon appeared phenotypically normal throughout juvenile and adult development, radiographic analyses revealed an increased prevalence of vertebral, spinal, and rib deformities relative to WT controls. This effect was observed in both if/fs and fs/fs mutant genotypes ∼6 months after sea water transfer (i.e., 26 months-old salmon) but reached statistical significance only in immature males (see Figure 3), which may partly reflect differences in sample size rather than a genotype-specific biological effect. Also in females, we have previously reported an increased frequency of malformations in all mutant fish, significantly for the if/if group (Crespo et al., 2024). Regardless, these findings raise the possibility that loss of bmp15 function influences skeletal integrity either directly or indirectly during development in salmon. To our knowledge, comparable skeletal abnormalities have not been reported in BMP15/Bmp15-deficient mammalian or zebrafish models. However, it remains unclear whether the skeletal phenotype reported here is unique to Atlantic salmon or has simply been overlooked in other vertebrate species since detailed assessments of the vertebral column have not been described in those models.

In contrast to several BMP family members that have well-established roles in skeletal development and bone homeostasis (Chen et al., 2021; Kamiya and Mishina, 2011), BMP15 has not previously been implicated in vertebrate bone biology to date. A study in frogs demonstrated that Bmp15 is required for normal embryogenesis (Di Pasquale and Brivanlou, 2009), while increased BMP15 transcript abundance has been associated with improved embryo quality in humans (Li et al., 2014). Our previous observations are consistent with these reports, as fertilizations using bmp15 fs/fs female gametes resulted in either complete embryonic lethality or markedly reduced survival, and surviving F1 embryos displayed severe developmental defects prior to and after hatching supporting a role for Bmp15 during early development in salmon (Crespo et al., 2024). Here, we did not observe further anomalies regarding sperm production and motility in bmp15 mutant males, however, further investigations on sperm cells functionality and their capacity to fertilize standard WT eggs are required for definite conclusion. The nature of the observed skeletal abnormalities, including duplicated neural and haemal spines, is consistent with perturbations arising during early embryogenesis (Mommens et al., 2015). Nonetheless, these observations do not distinguish between a maternal effect of bmp15 gene loss on oocyte quality and early embryogenesis, and a potential direct role of Bmp15 in skeletal development or post-embryonic bone physiology. We unfortunately did not assess the occurrence of skeletal defects also in swim-up larvae (e.g., whole-mount staining), which would have partly answered some of the uncertainties raised. Analyses of subsequent filial generations will be necessary to disentangle the relative contributions of these mechanisms.

Statements

Data availability statement

The original contributions presented in the study are publicly available. This data can be found here: https://doi.org/10.6084/m9.figshare.34886139.

Ethics statement

The animal study was approved by the Norwegian Food Safety Authority (https://www.mattilsynet.no, permit numbers 14865 and 22082). The study was conducted in accordance with the local legislation and institutional requirements.

Author contributions

DC: Writing – review and editing, Formal Analysis, Visualization, Writing – original draft, Data curation, Investigation. PF: Funding acquisition, Formal Analysis, Writing – review and editing, Methodology, Conceptualization, Investigation. TH: Writing – review and editing, Investigation, Funding acquisition, Conceptualization. EK-S: Methodology, Writing – review and editing, Formal Analysis. FA: Writing – review and editing, Formal Analysis. KS: Writing – review and editing, Investigation. AT: Writing – review and editing, Methodology, Formal Analysis. BN: Data curation, Methodology, Writing – review and editing, Formal Analysis. RE: Conceptualization, Methodology, Writing – review and editing, Funding acquisition, Investigation. EA: Writing – review and editing, Investigation. RS: Formal Analysis, Writing – review and editing, Data curation, Conceptualization, Funding acquisition, Investigation. AW: Funding acquisition, Project administration, Conceptualization, Writing – review and editing, Investigation, Methodology. LK: Writing – review and editing, Data curation, Investigation, Conceptualization, Project administration, Formal Analysis, Supervision, Funding acquisition.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This study was funded by the Norwegian Research Council projects SALMOSTERILE (2015-2017, n° 267610), STERWELL (2017-2020, n° 245979) and STERFEMSAL (2020-2023, n° 302532).

Acknowledgments

The authors thank Anne Torsvik, Grethe Thorsheim, Audun Østby Pedersen, Anne Hege Straume, Sven Leininger and Sara Olausson for expert technical assistance, and Ivar Helge Matre, Lise Dyrhovden, Ann-Kathrin Jakobsen Kroken, Arnor Gullanger, Christine Sørfonn, Florian Sambraus, Simon Flavell and Tone Knappskog for producing and maintaining the fish used in this study. We also would like to thank Hanne Sannæs for the assistance in running the MiSeq instrument.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The authors DC, RS declared that they were an editorial board member of Frontiers at the time of submission. This had no impact on the peer review process and the final decision.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fgene.2026.1945517/full#supplementary-material

References

  • 1

    AaltonenJ.LaitinenM. P.VuojolainenK.JaatinenR.Horelli-KuitunenN.SeppaL.et al (1999). Human growth differentiation factor 9 (GDF-9) and its novel homolog GDF-9B are expressed in oocytes during early folliculogenesis. J. Clin. Endocrinol. Metab.84 (8), 2744–2750. 10.1210/jcem.84.8.5921

  • 2

    AbeciaJ. A.MacíasÁ.CasaoA.BurilloC.MartínE.Pérez-PéR.et al (2020). Semen quality of Rasa Aragonesa rams carrying the FecXR allele of the BMP15 gene. Animals-Basel10 (9), 1628. 10.3390/ani10091628

  • 3

    AfkhamiF.ShahbaziS.FarzadiL.DanaeiS. (2022). Novel bone morphogenetic protein 15 (BMP15) gene variants implicated in premature ovarian insufficiency. Reprod. Biol. Endocrinol.20 (1), 42. 10.1186/s12958-022-00913-6

  • 4

    Alquezar-BaetaC.Gimeno-MartosS.Miguel-JimenezS.SantolariaP.YanizJ.PalacinI.et al (2019). OpenCASA: a new open-source and scalable tool for sperm quality analysis. PLoS Comput. Biol.15 (1), e1006691. 10.1371/journal.pcbi.1006691

  • 5

    AltschulS. F.GishW.MillerW.MyersE. W.LipmanD. J. (1990). Basic local alignment search tool. J. Mol. Biol.215 (3), 403–410. 10.1016/S0022-2836(05)80360-2

  • 6

    AnderssonE.SchulzR. W.AlmeidaF.KleppeL.SkaftnesmoK. O.Kjaerner-SembE.et al (2024). Loss of Fshr prevents testicular maturation in Atlantic salmon (Salmo salar L.). Endocrinology165, bqae013. 10.1210/endocr/bqae013

  • 7

    AyllonF.Kjaerner-SembE.FurmanekT.WennevikV.SolbergM. F.DahleG.et al (2015). The vgll3 locus controls age at maturity in wild and domesticated Atlantic salmon (Salmo salar L.) males. PLoS Genet.11 (11), e1005628. 10.1371/journal.pgen.1005628

  • 8

    AyllonF.SolbergM. F.GloverK. A.MohammadiF.Kjaerner-SembE.FjelldalP. G.et al (2019). The influence of vgll3 genotypes on sea age at maturity is altered in farmed mowi strain Atlantic salmon. BMC Genet.20 (1), 44. 10.1186/s12863-019-0745-9

  • 9

    BarsonN. J.AykanatT.HindarK.BaranskiM.BolstadG. H.FiskeP.et al (2015). Sex-dependent dominance at a single locus maintains variation in age at maturity in salmon. Nature528 (7582), 405–408. 10.1038/nature16062

  • 10

    BolstadG. H.KarlssonS.HagenI. J.FiskeP.UrdalK.SaegrovH.et al (2021). Introgression from farmed escapees affects the full life cycle of wild Atlantic salmon. Sci. Adv.7 (52), eabj3397. 10.1126/sciadv.abj3397

  • 11

    CelestinoJ. J.Lima-VerdeI. B.BrunoJ. B.MatosM. H.ChavesR. N.SaraivaM. V.et al (2011). Steady-state level of bone morphogenetic protein-15 in goat ovaries and its influence on in vitro development and survival of preantral follicles. Mol. Cell Endocrinol.338 (1-2), 1–9. 10.1016/j.mce.2011.02.007

  • 12

    ChenY.MaB.WangX.ZhaX.ShengC.YangP.et al (2021). Potential functions of the BMP family in bone, obesity, and glucose metabolism. J. Diabetes Res.2021, 6707464. 10.1155/2021/6707464

  • 13

    CillerI. M.PalanisamyS. K.CillerU. A.McFarlaneJ. R. (2016). Postnatal expression of bone morphogenetic proteins and their receptors in the mouse testis. Physiol. Res.65 (4), 673–682. 10.33549/physiolres.933193

  • 14

    ClellandE.KohliG.CampbellR. K.SharmaS.ShimasakiS.PengC. (2006). Bone morphogenetic protein-15 in the zebrafish ovary: complementary deoxyribonucleic acid cloning, genomic organization, tissue distribution, and role in oocyte maturation. Endocrinology147 (1), 201–209. 10.1210/en.2005-1017

  • 15

    CrespoD.AssisL. H. C.van de KantH. J. G.de WaardS.SafianD.LemosM. S.et al (2019). Endocrine and local signaling interact to regulate spermatogenesis in zebrafish: follicle-stimulating hormone, retinoic acid and androgens. Development146 (21), dev178665. 10.1242/dev.178665

  • 16

    CrespoD.FjelldalP. G.HansenT. J.Kjaerner-SembE.SkaftnesmoK. O.ThorsenA.et al (2024). Loss of bmp15 function in the seasonal spawner Atlantic salmon results in ovulatory failure. FASEB J.38 (14), e23837. 10.1096/fj.202400370R

  • 17

    CuissetB.PradellesP.KimeD. E.KühnE. R.BabinP.DavailS.et al (1994). Enzyme immunoassay for 11-ketotestosterone using acetylcholinesterase as laberl: application to the measurement of 11-ketotestosterone in plasma of Siberian sturgeon. Comp. Biochem. Physiol., Part C:Pharmacol., Toxicol. Endocrinol.108 (2), 229–241. 10.1016/1367-8280(94)90035-3

  • 18

    Di PasqualeE.BrivanlouA. H. (2009). Bone morphogenetic protein 15 (BMP15) acts as a BMP and Wnt inhibitor during early embryogenesis. J. Biol. Chem.284 (38), 26127–26136. 10.1074/jbc.M109.036608

  • 19

    DranowD. B.HuK.BirdA. M.LawryS. T.AdamsM. T.SanchezA.et al (2016). Bmp15 is an oocyte-produced signal required for maintenance of the adult female sexual phenotype in Zebrafish. PLoS Genet.12 (9), e1006323. 10.1371/journal.pgen.1006323

  • 20

    DubeJ. L.WangP.ElvinJ.LyonsK. M.CelesteA. J.MatzukM. M. (1998). The bone morphogenetic protein 15 gene is X-linked and expressed in oocytes. Mol. Endocrinol.12 (12), 1809–1817. 10.1210/mend.12.12.0206

  • 21

    DuckertP.BrunakS.BlomN. (2004). Prediction of proprotein convertase cleavage sites. Protein Eng. Des. Sel.17 (1), 107–112. 10.1093/protein/gzh013

  • 22

    EdgarR. C. (2004). MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res.32 (5), 1792–1797. 10.1093/nar/gkh340

  • 23

    EdvardsenR. B.LeiningerS.KleppeL.SkaftnesmoK. O.WargeliusA. (2014). Targeted mutagenesis in Atlantic salmon (Salmo salar L.) using the CRISPR/Cas9 system induces complete knockout individuals in the F0 generation. PLoS One9 (9), e108622. 10.1371/journal.pone.0108622

  • 24

    EppigJ. J. (2001). Oocyte control of ovarian follicular development and function in mammals. Reproduction122 (6), 829–838. 10.1530/rep.0.1220829

  • 25

    EricksonG. F.ShimasakiS. (2001). The physiology of folliculogenesis: the role of novel growth factors. Fertil. Steril.76 (5), 943–949. 10.1016/s0015-0282(01)02859-x

  • 26

    Fernandes da CostaD.de Oliveira RibeiroA.Morena Bonita RicciJ.da Silva RodriguesM.Antonio de OliveiraM.Felipe da RosaI.et al (2024). A83-01 and DMH1 effects in the zebrafish spermatogonial niche: unraveling the roles of TGF-beta and BMP signaling in the Fsh-mediated spermatogonial fate. Gene897, 148082. 10.1016/j.gene.2023.148082

  • 27

    FirthD. (1993). Bias reduction of maximum likelihood estimates. Biometrika80 (1), 27–38. 10.1093/biomet/80.1.27

  • 28

    FlossmannG.WurmserC.PauschH.TengheA.DodenhoffJ.DahintenG.et al (2021). A nonsense mutation of bone morphogenetic protein-15 (BMP15) causes both infertility and increased litter size in pigs. BMC Genomics22 (1), 38. 10.1186/s12864-020-07343-x

  • 29

    GagnonJ. A.ValenE.ThymeS. B.HuangP.AkhmetovaL.PauliA.et al (2014). Efficient mutagenesis by Cas9 protein-mediated oligonucleotide insertion and large-scale assessment of single-guide RNAs. PLoS One9 (5), e98186. 10.1371/journal.pone.0098186

  • 30

    GallowayS. M.McNattyK. P.CambridgeL. M.LaitinenM. P.JuengelJ. L.JokirantaT. S.et al (2000). Mutations in an oocyte-derived growth factor gene (BMP15) cause increased ovulation rate and infertility in a dosage-sensitive manner. Nat. Genet.25 (3), 279–283. 10.1038/77033

  • 31

    HalmS.IbanezA. J.TylerC. R.PratF. (2008). Molecular characterisation of growth differentiation factor 9 (gdf9) and bone morphogenetic protein 15 (bmp15) and their patterns of gene expression during the ovarian reproductive cycle in the European sea bass. Mol. Cell Endocrinol.291 (1-2), 95–103. 10.1016/j.mce.2008.03.002

  • 32

    HansenT.FjelldalP. G.YurtsevaA.BergA. (2010). A possible relation between growth and number of deformed vertebrae in Atlantic salmon (Salmo salar L.). J. Appl. Ichthyol.26 (2), 355–359. 10.1111/j.1439-0426.2010.01434.x

  • 33

    HartigF. (2026). DHARMa: Residual Diagnostics for Hierarchical (Multi-Level/Mixed) Regression Models. 10.32614/CRAN.package.DHARMa

  • 34

    HeY.GanM.MaJ.LiangS.ChenL.NiuL.et al (2025). TGF-beta signaling in the ovary: emerging roles in development and disease. Int. J. Biol. Macromol.306 (Pt 4), 141455. 10.1016/j.ijbiomac.2025.141455

  • 35

    HeinzeG.SchemperM. (2002). A solution to the problem of separation in logistic regression. Stat. Med.21 (16), 2409–2419. 10.1002/sim.1047

  • 36

    HeinzeG.PlonerM.JirickaL.SteinerG. (2025). Logistf: Firth’s Bias-Reduced Logistic Regression. 10.32614/CRAN.package.logistf

  • 37

    HwangW. Y.FuY.ReyonD.MaederM. L.KainiP.SanderJ. D.et al (2013). Heritable and precise zebrafish genome editing using a CRISPR-Cas system. PLoS One8 (7), e68708. 10.1371/journal.pone.0068708

  • 38

    KamiyaN.MishinaY. (2011). New insights on the roles of BMP signaling in bone-A review of recent mouse genetic studies. Biofactors37 (2), 75–82. 10.1002/biof.139

  • 39

    KleppeL.AnderssonE.SkaftnesmoK. O.EdvardsenR. B.FjelldalP. G.NorbergB.et al (2017a). Sex steroid production associated with puberty is absent in germ cell-free salmon. Sci. Rep.7 (1), 12584. 10.1038/s41598-017-12936-w

  • 40

    KleppeL.EdvardsenR. B.FurmanekT.AnderssonE.JuanchichA.WargeliusA. (2017b). bmp15l, figla, smc1bl, and larp6l are preferentially expressed in germ cells in Atlantic salmon (Salmo salar L.). Mol. Reprod. Dev.84 (1), 76–87. 10.1002/mrd.22755

  • 41

    KleppeL.EdvardsenR. B.FurmanekT.AnderssonE.SkaftnesmoK. O.Thyri SegafredoF.et al (2020). Transcriptomic analysis of dead end knockout testis reveals germ cell and gonadal somatic factors in Atlantic salmon. BMC Genomics21 (1), 99. 10.1186/s12864-020-6513-4

  • 42

    LenthR.PiaskowskiJ. (2026). “Emmeans: estimated marginal means, aka least-squares means,” in R Package Version 2.0.5. Available online at: https://rvlenth.github.io/emmeans (Accessed September 2026).

  • 43

    LiY.LiR. Q.OuS. B.ZhangN. F.RenL.WeiL. N.et al (2014). Increased GDF9 and BMP15 mRNA levels in cumulus granulosa cells correlate with oocyte maturation, fertilization, and embryo quality in humans. Reprod. Biol. Endocrinol.12, 81. 10.1186/1477-7827-12-81

  • 44

    LivakK. J.SchmittgenT. D. (2001). Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods25 (4), 402–408. 10.1006/meth.2001.1262

  • 45

    LorillardG. F.DurifC. M. F.CresciA.SkiftesvikA. B.BrowmanH. I.BjellandR.et al (2026). Effect of anthropogenic electromagnetic fields used for subsurface oil and gas exploration (controlled-source electromagnetics, CSEM) on the fertilization success and embryonic development of Atlantic cod (Gadus morhua). Mar. Environ. Res.215, 107811. 10.1016/j.marenvres.2025.107811

  • 46

    LuksenburgM. D. A.MeloM. C.AnderssonE.FjelldalP. G.BogerdJ.TarangerG. L.et al (2026). Stereological and histological characterization of spermatogenesis in Atlantic salmon. Aquaculture621, 743973. 10.1016/j.aquaculture.2026.743973

  • 47

    MeloM. C.AnderssonE.FjelldalP. G.BogerdJ.FrancaL. R.TarangerG. L.et al (2014). Salinity and photoperiod modulate pubertal development in Atlantic salmon (Salmo salar). J. Endocrinol.220 (3), 319–332. 10.1530/JOE-13-0240

  • 48

    MeloM. C.van DijkP.AnderssonE.NilsenT. O.FjelldalP. G.MaleR.et al (2015). Androgens directly stimulate spermatogonial differentiation in juvenile Atlantic salmon (Salmo salar). Gen. Comp. Endocrinol.211, 52–61. 10.1016/j.ygcen.2014.11.015

  • 49

    MommensM.StorsetA.BabiakI. (2015). Some quantitative indicators of postovulatory aging and its effect on larval and juvenile development of Atlantic salmon (Salmo salar). Theriogenology84 (1), 170–176 e172. 10.1016/j.theriogenology.2015.03.001

  • 50

    MonsivaisD.MatzukM. M.PangasS. A. (2017). The TGF-beta family in the reproductive tract. Cold Spring Harb. Perspect. Biol.9 (10), a022251. 10.1101/cshperspect.a022251

  • 51

    MoraisR.CrespoD.NobregaR. H.LemosM. S.van de KantH. J. G.de FrancaL. R.et al (2017). Antagonistic regulation of spermatogonial differentiation in zebrafish (Danio rerio) by Igf3 and Amh. Mol. Cell Endocrinol.454, 112–124. 10.1016/j.mce.2017.06.017

  • 52

    NeumannJ. C.ChandlerG. L.DamoulisV. A.FustinoN. J.LillardK.LooijengaL.et al (2011). Mutation in the type IB bone morphogenetic protein receptor Alk6b impairs germ-cell differentiation and causes germ-cell tumors in zebrafish. Proc. Natl. Acad. Sci. U. S. A.108 (32), 13153–13158. 10.1073/pnas.1102311108

  • 53

    NichollsP. K.HarrisonC. A.GilchristR. B.FarnworthP. G.StantonP. G. (2009). Growth differentiation factor 9 is a germ cell regulator of Sertoli cell function. Endocrinology150 (5), 2481–2490. 10.1210/en.2008-1048

  • 54

    OlsvikP. A.LieK. K.JordalA. E.NilsenT. O.HordvikI. (2005). Evaluation of potential reference genes in real-time RT-PCR studies of Atlantic salmon. BMC Mol. Biol.6, 21. 10.1186/1471-2199-6-21

  • 55

    OtsukaF.McTavishK. J.ShimasakiS. (2011). Integral role of GDF-9 and BMP-15 in ovarian function. Mol. Reprod. Dev.78 (1), 9–21. 10.1002/mrd.21265

  • 56

    PankhurstN. W.CarragherJ. F. (1992). Oocyte maturation and changes in plasma steroid-levels in snapper Pagrus (=Chrysophrys) auratus (Sparidae) following treatment with human chorionic-gonadotropin. Aquaculture101 (3-4), 337–347. 10.1016/0044-8486(92)90036-K

  • 57

    PengJ.LiQ.WigglesworthK.RangarajanA.KattamuriC.PetersonR. T.et al (2013). Growth differentiation factor 9:bone morphogenetic protein 15 heterodimers are potent regulators of ovarian functions. Proc. Natl. Acad. Sci. U. S. A.110 (8), E776–E785. 10.1073/pnas.1218020110

  • 58

    QinY.TangT.LiW.LiuZ.YangX.ShiX.et al (2019). Bone morphogenetic protein 15 knockdown inhibits porcine ovarian follicular development and ovulation. Front. Cell Dev. Biol.7, 286. 10.3389/fcell.2019.00286

  • 59

    R Core Team (2025). R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing. Available online at: https://www.R-project.org (Accessed September 2026).

  • 60

    RossettiR.FerrariI.BestettiI.MoleriS.BrancatiF.PetroneL.et al (2020). Fundamental role of BMP15 in human ovarian folliculogenesis revealed by null and missense mutations associated with primary ovarian insufficiency. Hum. Mutat.41 (5), 983–997. 10.1002/humu.23988

  • 61

    SawatariE.ShikinaS.TakeuchiT.YoshizakiG. (2007). A novel transforming growth factor-beta superfamily member expressed in gonadal somatic cells enhances primordial germ cell and spermatogonial proliferation in rainbow trout (Oncorhynchus mykiss). Dev. Biol.301 (1), 266–275. 10.1016/j.ydbio.2006.10.001

  • 62

    SkaarK. S.NobregaR. H.MagarakiA.OlsenL. C.SchulzR. W.MaleR. (2011). Proteolytically activated, recombinant anti-mullerian hormone inhibits androgen secretion, proliferation, and differentiation of spermatogonia in adult zebrafish testis organ cultures. Endocrinology152 (9), 3527–3540. 10.1210/en.2010-1469

  • 63

    SuY. Q.WuX.O'BrienM. J.PendolaF. L.DenegreJ. N.MatzukM. M.et al (2004). Synergistic roles of BMP15 and GDF9 in the development and function of the oocyte-cumulus cell complex in mice: genetic evidence for an oocyte-granulosa cell regulatory loop. Dev. Biol.276 (1), 64–73. 10.1016/j.ydbio.2004.08.020

  • 64

    SunL. P.SongY. P.DuQ. Z.SongL. W.TianY. Z.ZhangS. L.et al (2014). Polymorphisms in the bone morphogenetic protein 15 gene and their effect on sperm quality traits in Chinese Holstein bulls. Genet. Mol. Res.13 (1), 1805–1812. 10.4238/2014.March.17.8

  • 65

    TangT.LinQ.QinY.LiangX.GuoY.CongP.et al (2020). Effects of bone morphogenetic protein 15 (BMP15) knockdown on porcine testis morphology and spermatogenesis. Reprod. Fertil. Dev.32 (11), 999–1011. 10.1071/RD20056

  • 66

    TarangerG. L.CarrilloM.SchulzR. W.FontaineP.ZanuyS.FelipA.et al (2010). Control of puberty in farmed fish. Gen. Comp. Endocrinol.165 (3), 483–515. 10.1016/j.ygcen.2009.05.004

  • 67

    VikingstadE.AnderssonE.HansenT.NorbergB.MayerI.StefanssonS.TarangerG. L. (2016). Effects on temperature on the final stages of sexual maturation in Atlantic salmon (Salmo salar L.). Fish Physiol. Biochem42, 895–907. 10.1007/s10695-015-0183-1

  • 68

    WargeliusA.LeiningerS.SkaftnesmoK. O.KleppeL.AnderssonE.TarangerG. L.et al (2016). Dnd knockout ablates germ cells and demonstrates germ cell independent sex differentiation in Atlantic salmon. Sci. Rep.6, 21284. 10.1038/srep21284

  • 69

    WittenP. E.Gil-MartensL.HuysseuneA.TakleH.HjeldeK. (2009). Towards a classification and an understanding of developmental relationships of vertebral body malformations in Atlantic salmon (Salmo salar L.). Aquaculture295 (1-2), 6–14. 10.1016/j.aquaculture.2009.06.037

  • 70

    YadavH.LalB. (2017). BMP15 in catfish testis: cellular distribution, seasonal variation, and its role in steroidogenesis. Steroids125, 114–123. 10.1016/j.steroids.2017.07.002

  • 71

    YanC.WangP.DeMayoJ.DeMayoF. J.ElvinJ. A.CarinoC.et al (2001). Synergistic roles of bone morphogenetic protein 15 and growth differentiation factor 9 in ovarian function. Mol. Endocrinol.15 (6), 854–866. 10.1210/mend.15.6.0662

  • 72

    YanoA.GuyomardR.NicolB.JouannoE.QuilletE.KloppC.et al (2012). An immune-related gene evolved into the master sex-determining gene in rainbow trout, Oncorhynchus mykiss. Curr. Biol.22 (15), 1423–1428. 10.1016/j.cub.2012.05.045

  • 73

    YoungJ. C.WakitaniS.LovelandK. L. (2015). TGF-beta superfamily signaling in testis formation and early male germline development. Semin. Cell Dev. Biol.45, 94–103. 10.1016/j.semcdb.2015.10.029

  • 74

    YueY.ZengC.ZhangX.BianC.ZhangZ.WuK.et al (2026). Genetic and functional characterization of AMH signaling in Zebrafish - evidence for roles of Amh-Bmpr2a-Bmpr1bb pathway in controlling gonadal homeostasis. PLoS Genet.22 (3), e1011958. 10.1371/journal.pgen.1011958

  • 75

    ZhaiY.ZhangX.ZhaoC.GengR.WuK.YuanM.et al (2023). Rescue of bmp15 deficiency in zebrafish by mutation of inha reveals mechanisms of BMP15 regulation of folliculogenesis. PLoS Genet.19 (9), e1010954. 10.1371/journal.pgen.1010954

Summary

Keywords

Atlantic salmon, bone morphogenetic protein 15, CRISPR/Cas9, spermatogenesis, spermiation

Citation

Crespo D, Fjelldal PG, Hansen TJ, Kjærner-Semb E, Ayllon F, Skaftnesmo KO, Thorsen A, Norberg B, Edvardsen RB, Andersson E, Schulz RW, Wargelius A and Kleppe L (2026) Loss of bmp15 gene function does not abolish fertility in male Atlantic salmon. Front. Genet. 17:1945517. doi: 10.3389/fgene.2026.1945517

Received

22 July 2026

Revised

01 October 2026

Accepted

12 October 2026

Published

04 November 2026

Volume

17 - 2026

Edited by

Sebastian Escobar-Aguirre, Pontificia Universidad Católica de Chile, Chile

Reviewed by

Jin Gao, Hainan Academy of Ocean and Fisheries Sciences, China

Yuanri Hu, Chinese Academy of Fishery Sciences (CAFS), China

Updates

Copyright

*Correspondence: Diego Crespo, ; Lene Kleppe,

† Present address: Diego Crespo, Benchmark Genetics Norway AS, Bergen, Norway

Disclaimer

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.

Outline

Figures

Cite article

Copy to clipboard


Export citation file


Share article

Article metrics