Transcriptome of Sterile Testes in dnd-Depleted Atlantic Salmon (Salmo salar L.) Highlights Genes Involved in Gonadal and Brain Development

1. Introduction

Development of gametes is initiated by speciation of primordial germ cells (PGC) in early embryonic development followed by their migration to the genital ridge to form gonad anlagen together with somatic cells. The somatic cell lineages then develop into Sertoli cells and Leydig cells in the testis, and germline stem cells sustain gamete production giving rise to the sperm. PGC and descending cells express a suite of genes required for maintenance of pluripotency, prevention of premature differentiation and preservation of genome integrity. The RNA-binding protein dead end (dnd) protects PGCs from reprogramming into other cell types by inhibiting the interaction of micro RNAs with the target messenger RNAs [1]. Accordingly, dnd-depleted PGCs in zebrafish lost their characteristic morphology and adopted specific somatic cell fates, whereas only a few were found to undergo apoptosis [2]. PGC-specific expression of dnd has been shown in various fish species [3,4,5,6]. Germ cells are a prerequisite for female development in zebrafish (Danio rerio), medaka (Oryzias latipes), and mackerel (Scomber australasicus) as ablation of PGCs results in phenotypic, though sterile, males only [7,8,9,10,11]. In comparison, PGC ablation failed to affect the sexual fate of gonadal somatic cells in loach (Misgurnus anguillicaudatus), goldfish (Carassius auratus), grass puffer (Takifugu alboplumbeus), and salmonids, which could develop as either phenotypic males or females [12,13,14,15,16,17].

PGC have been a focus of developmental biology since their discovery in XIX century as reviewed in [18,19]. Commercial aquaculture requires effective methods of sterilization to prevent early maturation, improve flesh quality and reduce gene flow from farm escapees to wild fish [20]. Promising results have been achieved by transient degradation of dnd transcripts or by permanent inactivation of the dnd gene using CRSPR Cas [12,21,22]. Sterilization of Atlantic salmon by temporal knock down of dnd did not block development of the somatic testes, and the GC ablated fish maintained endocrine regulation by the hypothalamus – pituitary – gonad (HPG) axis during maturation entrance, but without any further progression of the reproductive cycle [23]. In addition to achieving a practical goal, GC-ablated gonads may serve as a powerful research tool to highlight genes specifically expressed in GC and structures that critically depend on interactions with the germline. In this study, we present a transcriptomic analysis of GC-ablated Atlantic salmon males. Genes that were silenced or differentially expressed in the testes of sterile males were categorized based on their involvement in normal developmental processes, such as gonadal growth and the initiation of sexual maturation. The dataset was integrated with transcriptomic data from testes of males entering the first reproduction cycle, obtained from a study that utilized the same genome-wide Atlantic salmon microarray platform [24].

2. Materials and Methods

2.1. Fish Material

Gene expression analyses were performed on testes from Atlantic salmon described in detail in [23]. In brief, dnd knockdowns were produced by microinjection of “Gapmer” antisense oligonucleotides targeting the dnd mRNA at the one-cell stage [25]. Embryos, alevins and juveniles were raised under standard hatchery conditions [23]. Sterility caused by germ cell ablation was assessed by regular histology, in situ hybridization, qPCR and immunohistochemistry of the germ cell marker vasa (ddx4) during the embryonic (c. 500 degree days) and several juvenile stages [23]. Fish used in the present work were kept at a photoperiod of daily 6 h light (6L:18D) and water temperature of 5–6°C for eight weeks followed by continuous light and 10°C for six weeks according to established protocols of smoltification control. Samples of dnd knockdowns (referred to as sterile in text) and fertile controls were collected at the onset of constant light (fish weight of 220-250g) and six weeks later (fish weight of 350-400g) are referred to as w0 and w6, respectively. The study did not involve any treatments that could cause suffering. For sampling purposes, fish were euthanized using an overdose of tricaine (Merck, Rahway, NJ, USA).

2.2. Gene Expression

Transcriptome analyses were performed on the testis, ovary, brain, pituitary, gills, and head kidney. Although primary focus was on the testis, some results from the other tissues are included to support the key findings. Total RNA was extracted from tissue samples stored in RNALater (Thermo Fisher Scientific, Waltham, MA, USA) using the automated Biomek 4000 (Beckman Coulter, Brea, CA, USA) (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. The concentration, purity and integrity of RNA (RIN values > 7.5) were checked using a NanoDrop ND-8000000 Spectrophotometer (NanoDrop Technologies, Wilmington, DE, US) and 2100 Bioanalyzer with RNA Nano Chips (Agilent Technologies, Santa Clara, CA, US). Salgeno-2, the 44 K Atlantic salmon DNA oligonucleotide microarray (GPL28080) was fabricated by Agilent Technologies; reagents and equipment were purchased from the same provider. RNA was amplified, labelled, and fragmented using a One-Color Quick Amp Labelling Kit, and a Gene Expression Hybridization kit. The total RNA input for each reaction was 220 ng. After amplification and labelling, 1650 ng of cRNA was loaded on the array. After overnight hybridization in an oven (17 h, 65℃, rotation speed 0.01 × g), the arrays were washed with Gene Expression Wash Buffers 1 and 2 and scanned with SureScan, Agilent Technologies. Results were analyzed with bioinformatic pipeline STARS [26] and submitted to NCBI GEO Omnibus (GSE297806). To investigate the developmental profiles of genes affected by dnd silencing, this dataset was integrated with microarray analysis results from Atlantic salmon males undergoing sexual maturation [24]. Differentially expressed genes (DEG) were selected by criteria: log2-ER (expression ratio) > 0.8 (1.75-fold) and p < 0.05. For enrichment analysis, the number of genes associated with each term was compared between the list of DEGs and the full microarray platform, with statistical significance assessed using Yates’ corrected chi-square test. Functional categories and pathways were ranked based on enrichment (fold), p-values, and the number of DEGs. Functional groups of genes (STARS annotation) with coordinated expression differences were identified by significant deviation of mean log₂-ER from zero baseline. Additionally, a body map of normalized transcriptomes generated using the Salgeno microarray [27] was utilized to assess gene expression distribution in 21 different cell types and tissues of Atlantic salmon. Expression levels were quantified as dSI (normalized relative signal intensity). Expression of 12 DEGs in sterile and fertile testes was verified with PCR. Primers (Table 1) were designed using Primer 3. The PCR products were purified and sequenced by Eurofins to confirm amplification of correct targets. The transcript abundance was analysed using a QuantStudio5 instrument (ThermoFisher, MA, USA), following the MIQE guidelines [28]. Each reaction was run in duplicates with non-template and non-RT enzyme controls for each gene. The PCR master mix was prepared with 0.5 μl forward, 0.5 μl reverse primer (Table 1), final concentrations of 0.5 μM and 5 μl SYBR Green (Applied Biosystems, Waltham, MA, USA). This master mix was added to a 4 μl of a 1:10 dilution of cDNA per well, resulting in a total reaction volume of 10 μl. The PCR protocol included an initial incubation at 95 °C for 20 seconds, followed by 40 cycles of denaturation at 95 °C for 1 second and annealing at 60 °C for 20 seconds, finally one last step of extension with an increase of 0.075 °C per second until 97 °C. Primer efficiency was assessed using 10-fold serial dilutions of cDNA for each primer set, and the specificity of amplification was verified by melting curve analysis. Rpol2 and eif-3 were used as reference genes showing stable expression in Atlantic salmon [29].

3. Results

3.1. Gene Markers of Germ Cells

Expression of nine GC gene markers was detected in sterile testes (Figure 1A). Five of these genes were expressed in all analyzed samples, whereas tdrd6 and piwil2 were silenced in 76.5% and 82.5% of cases, respectively. Tdrd6 is involved in spermatogenesis through RNA processing [30], while piwil2 plays a role in germline stem cell maintenance and transposon silencing [31]. Similar functions are attributed to tdrd1, tdrd9, tdrkh, and mov10l1, while phf1 is involved in chromatin modification and epigenetic regulation [32]. Six of these GC gene markers were analyzed with RT-qPCR (Figure 1B), which did not detect dnd transcripts in sterile testes, whereas the remaining genes showed significantly higher expression in fertile testis, with differences ranging from 2.4- to 570-fold. Nanos, which prevents untimely maturation of GC, was upregulated in growing testes but showed decreased expression at maturation (Figure 1A). In contrast, tdrd6 and phf1 were stimulated during this period. Other GC gene markers, such as piwil2, tdrd1, tdrd9, vasa, mov10l1 and tdrkh did not exhibit developmental changes above the threshold for differential expression. Vasa, a key regulator of RNA metabolism in GC, displayed the greatest tissue specificity, with a dSI of 11.2 – one of the highest among all Atlantic salmon genes. Other GC markers shown in Figure 1A are characterized by preferential expression in the gonads although they were also active in other tissues. Temporal dnd knockdown led to relatively small but significant transcriptome changes in the brain, pituitary, gill and head kidney. Genes that were downregulated across different tissues (Table S1) can be molecular targets of dnd. These genes have diverse biological functions, and their similar expression patterns in dnd-knockdown salmon appears to be the only common feature.

3.2. Genes Silenced or Downregulated in Sterile Testes

3.2.1. Silenced Genes

Microarray analyses in Atlantic salmon typically reveal decreased transcript levels rather than a complete absence of signal. A distinctive feature of this experiment was the silencing of a substantial number of genes in GC-ablated gonads, in both males and females. Transcripts of 260 genes were undetected in at least 75% of sterile testes and ovaries, and 165 of these genes were also inactivated during the maturation of fertile testes (Table S2). Nearly all GO terms enriched among this gene set are associated with brain and nervous system development (Table S3), and 38% of genes exhibited preferential expression in the Atlantic salmon brain (dSI > 3). Of note was the large number of genes encoding proteins with homeobox domains (45 genes). Genes with essential roles in brain development and function are shown in Figure 2. Several genes are specifically involved in early brain development and regionalization, including atp, bsx, dlx1, nkx1a, nk6.1, gbx, gsc and otp [33,34,35,36]. Others control neural patterning at various levels: zic2 and zic4 are required for neural tube closure [36], dmbx1 is crucial for spinal cord formation [37], and tlx2 regulates neural crest development [39]. Genes essential for forebrain, midbrain, hindbrain and cerebellum development include march1, foxg1, gbx2, dmbx1, foxg1 and cbln1 [40,41,42]. Additionally, pax6a is a key regulator of eye development [43]. Several genes contribute to neurogenesis and neural differentiation: sox21 maintains neural stem progenitor cells [44], while gpr3, neurod1 and 2 regulate neurogenesis. Scrt is involved in neuronal differentiation. Isl1 is critical for motor neuron development [45], nkx2-2 for oligodendrocyte differentiation [46], and sp5/9 for interneuron formation. Bdnf, nptx2 and sybp participate in excitatory neurotransmission and synaptic plasticity [47], whereas viiat is responsible for packaging GABA (γ-aminobutyric acid) and glycine into synaptic vesicles in inhibitory neurons [48]. Schip1 regulates myelinization, while bsx is implicated in hypothalamic development and behavioral control [49]. Fosl1 is a marker of neuronal activity [50]. Several genes have endocrine functions. Crh and trhr play a central role in the hypothalamic-pituitary-adrenal (HPA) axis [51]. Tac1 encodes a precursor to several neuropeptides and trh is involved in neuroendocrine signaling. RT-PCR analysis was performed on six selected genes (Table 1) to validate microarray results. As expected, transcripts of viaat, hr6, soho, gbh, dbx1 and crf1 were detected in all fertile but in none of the sterile testes (gel figure is not shown).

3.2.2. Genes Downregulated in sterile GC-Depleted Testes

This group comprises 229 genes that were downregulated, but not silenced, in sterile testes and showed increased expression during gonadal growth. It was also enriched for genes associated with the nervous system, including homeobox genes (Figure 3). Among these, six6, lhx4, homez, and irx play crucial roles in the development of the hypothalamus, pituitary gland, hippocampus, and thalamus [52,53,54]. Lhx4 regulates retinal neurons differentiation and vision [55]. Dlx4a is important for differentiation of GABAergic sensory neurons [56]. Genes involved in neuronal and neuroendocrine signaling (highlighted in the upper part of Figure 3) exhibit notably high relative expression in the brain. Genes with key roles in the neuroendocrine regulation of reproduction include vt1, olig2, nmu, nrn1, sv2, scg2, scg3, and scg5 [57,58,59,60,61]. A hallmark of this group was the high proportion of zinc finger (zf) proteins, accounting for 21.4% of the total. Unlike many homeobox genes, zf did not show preferential expression in testis and brain of Atlantic salmon (mean dSI respectively 2.6 and -0.3). Several genes linked to specific reproductive and developmental processes including smc1a and nipbl – 2 genes (meiosis), lift88 and cfap99 (cilia formation) and efcab2 (sperm motility) were downregulated but not silenced in sterile GC-ablated testes.

3.3. Genes with Stable Expression During Gonadal Growth and Maturation

Stable expression without significant changes in normal testes during development was observed in 1,067 genes that were downregulated in sterile gonads. Enrichment analysis primarily identified functional groups involved in core cellular processes, including chromosome maintenance, DNA replication, cell division, transcription, and RNA degradation, as well as nucleotide and energy metabolism (Table 2). Additionally, processes such as double strand break repair, base excision repair, and gene silencing by RNA are likely associated with spermatogenesis.

3.4. Genes with Expression Changes at Maturation

Increased and decreased expression was observed in 790 and 689 genes, respectively, when comparing maturing to immature testes. The enriched functional terms with highest ranks are listed in Table 3. This group included genes involved in notochord and brain development that were downregulated but not silenced in GC-depleted and maturing testes. The term “DNA replication” was enriched in both up and downregulated genes, reflecting changes in the repertoire of genes involved in this process during the transition from immature to mature testes. Maturation was primarily associated with cell division, particularly meiosis, as well as DNA repair, chromosome maintenance, and segregation. Enrichment of homologous recombination (KEGG pathway) in downregulated genes indicates that this process might be principally completed by the time of sexual maturation. Maturation was marked with a decline in the activity of growth factors and developmental processes. A suite of genes linked to spermatogenesis were inactivated in more than half of sterile GC-depleted testes (Figure 4). Daz and boule are members of Deleted in Azoospermia multigene family that regulate mRNA translation in germ cells, influencing the expression of genes critical for meiosis, spermatid differentiation, and sperm maturation [62,63]. Prdm9 and rad21 participate in homologous recombination [64,65]. However, transcripts of 35 genes involved in meiosis were detected in sterile GC-depleted testes, and only two out of 48 genes associated with cilia were silenced in half or more of sterile males.

3.5. Genes Upregulated in Sterile Testes

The upper section of Figure 5 highlights testis-specific genes (dSI > 4 or approximately 8-fold higher than the average level) that showed decreased expression at maturation. This small group includes amh, a key inhibitor of sexual maturation, and the immune related irf9-like gene named sdy, which regulates male sex determination by preventing ovarian development in salmonids [66]. Vip may play a role in regulating sex hormone secretion, whereas the reproductive functions of nodal and nog1, members of the TGF-β and BMP families, respectively, remain unknown. Notably, tgfb1 and bmp8B were among the most upregulated genes, but the greatest expression changes were observed in proteolytic enzymes. Sterility appeared to enhance processes that counteract maturation, but the primary effect of dnd- silencing was developmental reprogramming. The major trends included increased expression of immune genes, activation of developmental pathways unrelated to reproduction, and enhanced extracellular matrix deposition (Figure 6A). Specifically, sterile testes showed stimulation of chemokine and eicosanoid signaling, antigen presentation, humoral and cellular immune responses, and an increase in B and T lymphocyte populations. At the same time, enhanced differentiation of bone, cartilage, and endothelium was indicated. These functional groups were downregulated in testes at maturation. Enrichment analysis (Figure 6B) assigned the highest ranks to transcription modules (TMs) associated with bacterial responses and inflammation – gene sets identified through meta-analysis of transcriptome databases [67]. Additionally, genes related to cell adhesion (integrin binding) and extracellular matrix remodeling were markedly enriched (Figure 6B). On average, genes upregulated in sterile testes showed the highest expression levels in the skin and gill, while their expression was lowest in immature ovary (Figure 6C).

4. Discussion

Comparison of sterile and fertile testes in Atlantic salmon revealed numerous genes affected by dnd knockdown. Combining these results with profiles from normal gonadal development may provide insights into the fate of germ cells and the regulation of Atlantic salmon reproduction. The upregulated genes indicated the altered fate of the GC-ablated gonad. Sterile testes showed activation of functional groups of genes that are typically downregulated during sexual maturation, primarily those related to immune function. This might suggest gonadal degeneration and resorption, followed by fibrosis, as indicated by increased expression of genes involved in collagen synthesis and other extracellular matrix components. However, no clear transcriptional signature of acute inflammation was observed. Notably, transcripts of genes involved in adaptive immunity, specifically those associated with lymphocytes, B and T cells, and hematopoiesis, were highly expressed. Reprogramming toward an alternative role, with a predominance of immune functions, appears to be the most likely scenario in the development of sterile testes.

As expected, many genes involved in spermatogenesis were silenced or downregulated in sterile testes. However, a suite of GC markers retained low-level expression, along with genes involved in meiosis and the formation and function of flagella. This suggests the presence of a small population of GC-derived cells in sterile testes that have lost the capacity to differentiate into sperm. While temporal dnd knockdown did not completely abolish the expression of genes directly related to reproduction, the most pronounced effect was observed in a large set of genes associated with the nervous system, including those expressed in the brain. A significant fraction of these genes was completely silenced in sterile testes. Interestingly, a striking similarity between the adult human brain and testes was recently revealed through proteomic comparison [68]. Our results suggest that the same genes may be involved in fundamental developmental processes of both organs in Atlantic salmon, with their expression dependent on dnd and possibly other germ cell – related genes. Most of the genes involved in brain structure specification that were silenced in GC-depleted testes are transcription factors, including 45 genes containing homeobox domains. While the roles of homeobox proteins and other developmental regulators may vary across taxa, their high expression levels in the Atlantic salmon brain suggest functional conservation. Many of the genes downregulated or inactivated in sterile testes have diverse roles, including differentiation of brain regions, maintenance of neural stem cells, neuronal differentiation, neurotransmission and synaptic plasticity, sensory perception as well as paracrine and endocrine regulation. Collectively, these findings indicate the presence of a complex nervous structure in normal Atlantic salmon testes. The silencing of the same genes in GC-ablated testes and ovaries underscores their potential importance in reproduction. To the best of our knowledge, such a structure has not been previously described. Therefore, we can only speculate on its nature, highlighting potential directions for future research. We see three aspects that can be considered: PGC migration and homing, environment of GC residence and differentiation, and entering first sexual maturation.

The migration and homing of PGC in both mouse and zebrafish are controlled by chemokines [69,70], cell adhesion molecules and extracellular matrix interactions [71]. Studies in zebrafish highlighted the role of myosin contraction in PGC in response to chemokine SDF-1 [72]. Similar mechanisms of cell migration are crucial in brain development, ensuring proper neuronal regionalization and the formation of functional networks. It is reasonable to anticipate that the same molecular players can be involved in both gonadal and nervous system development. PGCs possess unique properties that distinguish them from somatic cells, notably their ability to differentiate into any cell type. This developmental potential is orchestrated by a set of genes that regulate cell specification, suppress somatic recombination, maintain genomic integrity, and protect against mutations and transposable element integration. The inactivation of even a single key gene, such as dnd, disrupts gametogenesis entirely. In addition to intrinsic regulation, the maintenance of PGCs requires a specialized niche that provides essential signals and structural support to preserve their identity and prevent premature differentiation [73]. The concept of a “niche” emerged from studies on interactions between PGCs and somatic cells in invertebrate models like nematodes and Drosophila and has since been extended to other animals and humans, particularly through advances in germline transplantation [74,75,76,77]. Despite growing interest, research in this area remains in its early phase, largely due to the complexity and variability of the PGC microenvironment across developmental stages [76] and species [78]. Inability of cells expressing GC markers and gamete-specific genes to differentiate in spermatozoa indicates that the hypothesized structure can be an essential part of GC niche. The absence of this structure in sterile dnd-deficient testes may imply that it originates from PGCs themselves, which possess the potential to differentiate into multiple lineages, including neural cells. The increasing application of single-cell sequencing in Atlantic salmon research is expected to clarify this issue by determining whether GC- and brain-specific genes are co-expressed within the same cellular lineages.

An interesting finding was the downregulation and silencing of brain-specific genes in testes at the onset of first sexual maturation. This suggests that the neural-like structure within the gonad ceases to function once the reproductive cycle begins. Unlike Pacific salmon of the genus Oncorhynchus, which reproduce only once, Atlantic salmon undergo multiple spawning events throughout their lives. The regulation of initial versus subsequent maturations may differ. What triggers the initial decision to enter the reproductive cycle remains unclear, particularly because Atlantic salmon males can mature at markedly different body sizes. While the age of sexual maturity is genetically determined, as demonstrated by successful selective breeding and identification of a master gene vgll3 [79], the physiological mechanisms mediating this genetic control are unknown. The primary signal of reproductive readiness may originate in the testes, and a provisional neural tissue could, in theory, mediate communication between the gonads and the brain. In contrast to the first maturation, subsequent reproductive cycles occur annually and no longer require a decision point, as the testes come under full external control. Increased daylight stimulates the hypothalamic-pituitary axis (HPA), which triggers the production of sex hormones that drive spermatogenesis.

5. Conclusions

A transcriptome comparison between fertile and sterile dnd-knockdown Atlantic salmon, combined with gene expression profiles from normal male gonads, revealed genes involved in sexual maturation and the regulation of this process. While most findings aligned with existing knowledge, several unexpected results were also obtained. The most dramatic effect of dnd knockdown was observed in genes with the highest expression in the brain, which are known to play key roles in central nervous system development and function. The origin and role of this structure, which ceases activity at maturation, remains to be investigated.