Functional characterization of human recessive DIS3 variants in premature ovarian insufficiency^†

Abstract

Premature ovarian insufficiency (POI) is characterized by the loss or complete absence of ovarian activity in women under the age of 40. Clinical presentation of POI varies with phenotypic severity ranging from premature loss of menses to complete gonadal dysgenesis. POI is genetically heterogeneous with >100 causative gene variants identified thus far. The etiology of POI varies from syndromic, idiopathic, monogenic to autoimmune causes the condition. Genetic diagnoses are beneficial to those impacted by POI as it allows for improved clinical management and fertility preservation. Identifying novel variants in candidate POI genes, however, is insufficient to make clinical diagnoses. The impact of missense variants can be predicted using bioinformatic algorithms but computational approaches have limitations and can generate false positive and false negative predictions. Functional characterization of missense variants, is therefore imperative, particularly for genes lacking a well-established genotype:phenotype correlation. Here we used whole-exome sequencing (WES) to identify the first case of a homozygous missense variant in DIS3 (c.2320C > T; p.His774Tyr) a critical component of the RNA exosome in a POI patient. This adds to the previously described compound heterozygous patient. We perform the first functional characterization of a human POI-associated DIS3 variant. A slight defect in mitotic growth was caused by the variant in a Saccharomyces cerevisiae model. Transgenic rescue of Dis3 knockdown in Drosophila melanogaster with human DIS3 carrying the patient variant led to aberrant ovarian development and egg chamber degeneration. This supports a potential deleterious impact of the human c.2320C > T; p.His774Tyr variant.

Graphical Abstract

Open in new tabDownload slide

Introduction

Premature ovarian insufficiency (POI), characterized by the loss or complete absence of ovarian activity, is a common cause of female infertility affecting ~1-3.7% of women under the age of 40 [1–3]. Clinical presentation can vary from premature loss of menses (secondary amenorrhea) or failure to enter menarche (primary amenorrhea) associated with complete absence of gonadal structures (streak gonads). POI is diagnosed by elevated (>25 mlU/ml) Follicle-stimulating hormone (FSH) measured twice in patient blood at least one month apart as per the European Society of Human Reproduction (ESHR) guidelines [4–6].

POI is genetically heterogeneous with >100 causative gene variants identified thus far [7, 8]. This number is rising exponentially due to high throughput, cost efficient next generation sequencing methods. Variable genetic cause in POI is partly due to its broad etiology with syndromic, autoimmune, monogenic, and idiopathic causes previously elucidated [9]. Therefore, the identification and classification of causative genes and variants is pertinent to clinical management and prevention of co-morbidities. Co-morbidities of POI include reduced fertility potential, cardiovascular disease, mental health problems, osteoporosis, and premature mortality [1]. Genes identified thus far that contribute to POI include, but are not limited to, genes encoding transcription/translation factors, proteins required for DNA repair/replication, cell differentiation/development, mitochondrial maintenance, hormone signaling, and metabolism [9–11].

Ribonucleic acid (RNA) degradation is critical to the maintenance and stability of cellular life. Stringent parameters for RNA transcript quality, abundance and maturation are regulated by the RNA exosome [12]. The RNA exosome functions as a ubiquitous 3′ - > 5′ exoribonuclease and endoribonuclease, consisting of eleven evolutionary conserved subunits [13]. While initially identified as a central ribosomal processing factor, the role of the RNA exosome was later expanded to include RNA biogenesis and the maturation/degradation of cryptic transcript species in both the nucleoplasm and cytoplasm of eukaryotic cells. More recent literature indicates that the RNA exosome is also necessary for DNA repair by homologous recombination [14].

The general architecture of the RNA exosome is conserved between bacterial, archaeal, and eukaryotic phylogenies [15]. Extensive phylogenetic conservation of the RNA exosome is indicative of the critical role the complex has in cellular life. In eukaryotes the RNA exosome is comprised of a nine- subunit core (EXOSC1-9). Six subunits (EXOSC4-9) establish a catalytically inactive barrel arranged in a hexameric (PH-like) ring wide enough (8–10 Å) to accommodate a single stranded RNA [13] (Figure 1). Due to the width of the hexamer, the barrel arrangement does not accommodate a double stranded RNA [16]. RNA substrates are unwound to single stranded RNAs (ssRNAs) before being threaded unidirectionally through the core barrel [16]. Unlike the archaeal exosome, the eukaryotic exosome core is catalytically inert, suggesting that the role of the exosome core may function as a scaffold structure that requires the association of additional co-factors, enzymes, and nucleases to aid in RNA processing [16]. The remaining three core proteins (EXOSC1, EXOSC2, EXOSC3) form a trimeric S1/KH ring cap. The core exosome proteins (EXOSC1-9) interact with two hydrolytic 3′-5′ exoribonucleases, exosome component 10 (EXOSC10) (rRNA-processing protein 6, Rrp6 in budding yeast) and DIS3 (defective in sister chromatid joining, rRNA-processing protein 44, Rrp44) that both provide enzymatic activity required to cleave RNA [17–19]. Basal DIS3 is located at the base exit pore of the exosome hexameric barrel and functions as both a hydrolytic 3′-5′ exoribonucleases and endoribonuclease [17]. While some eukaryotes, including yeast and Drosophila have one Dis3 that is present in both the cytoplasmic and nuclear exosome, humans have two paralogs with DIS3 in the nuclear RNA exosome and DIS3L in the cytoplasmic RNA exosome. DIS3 is an essential effector of RNA degradation, a process that is considered critical to posttranscriptional regulation and cellular function. DIS3 is in the RNase II/R protein family and is considered the predominant ribonuclease subunit responsible for the degradation of promoter upstream transcripts (PROMPTs) and enhancer RNAs (eRNAs) [20, 21].

Aberrant function of the RNA exosome and associated co-factors have been linked to human disease [13]. Variants in genes encoding both structural and catalytic components of the RNA exosome cause diverse pathogenic phenotypes. However, current understanding of how and why RNA exosome derived conditions are isolated to specific tissue types remains rudimentary [22]. Currently, of the eleven identified exosome subunits six have been associated with human disease [23–29]. Basally located DIS3 has been implicated in human diseases associated with aberrant mitotic control [30]. Additionally, patients with Multiple myeloma harboring pathogenic DIS3 variants were found to have reduced median survival [31, 32].

A dysregulated transcriptome has the potential to disrupt oogenesis and embryonic development post-fertilization [33]. An absence of active transcription in maturing oocytes means that the developing oocyte relies on the active degradation of RNAs to remodel the transcriptome [33]. For degradation to occur the inert six-protein barrel of the RNA exosome must be associated with the enzymatically active components EXOSC10 and DIS3. Therefore, variants impacting the function of DIS3 can have a significant impact on the RNA exosome and can disrupt transcriptome balance.

Dis3 mouse studies have indicated that DIS3 depletion significantly impacts the fertility of female mice [34]. Oocyte loss of Dis3 induces premature meiotic arrest of maturing oocytes, whereas global knockout (KO) leads to embryonic lethality [34]. Dis3 was found to significantly impact RNA degradation in oocytes and its disruption led to insufficient transcription termination paired with the accumulation of intergenic RNA species, ultimately responsible for premature meiotic arrest in DIS3-depleted oocytes [34]. Similar observations have been made in Drosophila whereby germline knockdown (KD) of DIS3 ortholog Dis3 results in complete infertility in female flies and significant atrophy of ovarian structures with no observable oocytes [35]. Similar cross species observations and significant genetic conservation alludes to a functionally conserved role of DIS3 in maintaining fertility potential.

Here we report a rare homozygous missense variant of DIS3 in a patient diagnosed with POI. The patient presented at 22 years with secondary amenorrhea, elevated FSH and atrophic ovaries. The variant identified affected a highly conserved residue of the exonucleolytic RNB domain [32, 36]. Given DIS3 is not a well-established human POI gene with only two compound heterozygous variants previously reported and not functionally characterized, we aimed to model the impact of the patient variant in yeast and Drosophila. Our data support a conserved role in fertility between human and Drosophila Dis3 and demonstrate that the human DIS3 variant identified in a POI patient has partially reduced function.

This is the second reported case of human DIS3 variants associated with POI, and the first study to functionally validate variant consequence. This work establishes DIS3 missense variants as a likely new autosomal recessive cause of POI.

Materials and methods

Patient clinical information

The patient presented with secondary amenorrhea at 22 years. Ultrasound revealed she had atrophic ovaries. Blood work taken on two separate occasions reported an elevated FSH of 58 mIU/ml (reference <25 mIU/ml, ESHR guidelines, https://www.eshre.eu/Guidelines-and-Legal/Guidelines/Management-of-premature-ovarian-insufficiency.aspx). Microarray analysis revealed a 46XX karyotype with no significant copy number variation. The patient is from a consanguineous Iranian family. No relatives were known to share the condition of the patient. DNA for other family members was not available. Patient clinical information is consistent with a diagnosis of non-syndromic POI. Information is summarized in Table 1.

Whole exome sequencing

Whole exome sequencing (WES) was undertaken on patient blood DNA samples at Victorian Clinical Genetics Services (VCGS). Exome capture with Agilent SureSelect Human All Exon V6 and sequencing on an Illumina NovaSeq 6000 were performed at an average read depth of 80x. WES data were processed using Cpipe [37] and analysed using SeqR (https://seqr.broadinstitute.org/) We used three different approaches to analyse WES data as done previously by [38]. The first phase of analysis was a gene centric approach that focused on the detection of variants in already identified POI genes. For the POI gene-centric approach, high quality rare variants (minor allele frequency [MAF] <0.005) with moderate-high priority (including missense, splicing, predicted loss-of-function) were further considered. The second phase focused on the analysis of any gene with variants fitting a potential recessive inheritance pattern. Similarly, the MAF filter was set at <0.005 and high-quality variants of moderate-high priority were further considered. The third phase involved an analysis of predicted loss of function (LoF) variants in any gene. This phase was not restricted to bi-allelic variants so a more stringent MAF of <0.0001 was used, as deleterious dominant variants are less likely to exist in a healthy population.

S. cerevisiae model

The analysis of yeast mutants was performed as previously described [39, 40]. Briefly, a strain that had the chromosomal copy of DIS3 (also known as RRP44) deleted and that expressed the essential DIS3 from a URA3 plasmid (yAv1121; Matα, ura3-∆0, leu2-∆0, his3-∆1, dis3∆::HYG [DIS#,URA3]) was transformed with one of four LEU2 plasmids: Wild-type DIS3 (pAv344) [39], dis3-H841Y (pAv1569), dis3-H841A (pAv1568), or empty vector (pRS415) [41]. Transformants were selected on SC-Ura-Leu, serially diluted and plated either on 5FOA (5-fluoro-orotic acid) containing media (selecting for cells that had lost the URA3 plasmid) or on SC-Leu media (as control). The ability to grow on 5FOA-containing media reflects that the URA3 plasmid can be lost, and thus that the dis3 allele on the LEU2 plasmid is functional. The point mutations in pAv1568 and pAv1569 were generated by the QuikChange method and verified by sequencing the whole plasmid (Plasmidsaurus).

Drosophila melanogaster Dis3 KD model

RNA interference (RNAi) line for Dis3 was obtained from the Bloomington Drosophila Stock Centre (BDSC). The Drosophila line utilized was Dis3 RNAi (BL67919). Additional Drosophila lines included control strain w¹¹¹⁸ (BL5905), nanosGAL4 (nos) driver generated from UAS-GAL4; GAL4::VP16-nos.UTR (BL4937 and BL5938) all from BDSC. The Traffic-jam GAL4 (Tj) driver used (DGRC104055) was obtained from the Kyoto Stock Centre. All long-term stock storage was maintained at 25°C on a molasses-based medium.

To establish KD of Dis3, the RNAi transgenic line was crossed to both Tj-GAL4 and nanosGAL4. Tj-GAL4 drives the KD of Dis3 in germline associated somatic cells during ovarian development and oogenesis [42, 43]. In turn, nanosGAL4 drives the KD of Dis3 in germ cells from the apical germarium until the mature oocyte [44, 45]. Controls used in this work were w1118 flies crossed with the respective driver (Tj or nanos). Drosophila crosses of RNAi lines to GAL4 drivers were raised on molasses-based medium supplemented with dried yeast. When crossed to Tj and nos drivers, Drosophila were raised at 18°C for initial cross and until larvae enclosure, and then moved to 29°C until pupal enclosure. The cross was then maintained at 29°C for 3–5 days to allow for the F1 generation to mate prior to ovary dissection, immunostaining, and fluorescent microscopy analysis. >15 ovaries were observed from >2 technical replicate crosses. Ovaries were dissected from the progeny and were immunostained using anti-Fasciclin 3 (FasIII) to visualize somatic stem cell derived follicular cells and polar cells, anti-vasa that marks the VAS protein that localizes to nucleoplasm of ovarian germ cells, and nuclear stain 4′,6-diamidino-2-phenylindole (DAPI) [46–48]. Immunostaining was done as per [49].

Generation of human DIS3 transgenic lines

DNA sequences of the open reading frame transcript corresponding to wild-type human DIS3 and human DIS|${3}^{H774Y}$| were synthesized by Azenta/Genewiz and directionally cloned into pUASTattBusing XhoI and XbaI restriction sites that were added to the 5′ and 3′ ends of the sequences, respectively. Transgenic flies were generated by BestGene, Inc. Both transgenes were inserted into the identical genomic location on chromosome 3 (attP2) using phi C31 integrase [50].

Statistical methods

Calculations for significance used either a two-proportion z-test (Figure 6) or an unpaired t-test (Figures 7, 8) with p-values for significance set at P < 0.05. Statistical analysis was undertaken in Prism10 software.

Results

Identification of DIS3 homozygous missense variants

Gene-centric and variant-centric analysis was performed on the patient. 29 rare moderate-high priority variants were identified in known or candidate POI genes. Twenty-four of the 29 POI-related variants were heterozygous and did not fit with the known mode of POI inheritance associated with the gene and/or had low pathogenicity prediction based on online algorithms and conservation scores. One gene, MTF1 had two variants of interest suggesting possible compound heterozygosity. MTF1 was not pursued for follow up due to the low pathogenicity prediction of variants in-silico (Supplementary Figure S1).

Homozygous missense variants were observed in three POI candidate genes; DIS3, RAD51B and APC2. RAD51B is a strong POI candidate gene due to its role in homologous recombination. However, this gene was not considered the strongest candidate due to the low in silico pathogenicity predictions and poorly conserved altered residue (Supplementary Table S1).

APC2 has been considered a POI-candidate gene because APC2 deficiency in mice leads to disrupted ovarian WNT signaling and female subfertility [52]. Due to benign PolyPhen in-silico pathogenicity prediction, APC2 was not considered the strongest candidate (Supplementary Table S1).

In the second phase of analysis, focusing on genes with variants that could fit recessive inheritance, variants in DIS3, MTF1, RAD51B and APC2 were again identified along with variants in MGA. MAX Dimerization Protein (MGA) acts as a tumor suppressor gene [53]. Although in silico predictions indicated this was a potentially deleterious variant, no direct link with infertility had been reported in association with this gene.

Therefore, after filtering and in-silico analysis we considered the homozygous missense variant observed in DIS3 as the top candidate. The patient has a homozygous missense variant c.2320 C > T that results in a predicted amino acid change from histidine to tyrosine in DIS3 at position 774, p.His774Tyr. The variant is in a highly conserved residue, has high pathogenicity in-silico predictions as well as an association in mice with a significant infertility phenotype [34] (Supplementary Table S1). His774 is part of the RNase II signature motif that defines the family of RNase II-like enzymes (Figure 2 A). This histidine is highly conserved in eukaryotic Dis3 and Dis3L enzymes, and even in Escherichia coli (E.coli) RNase R, but not in the namesake member of the family RNase II (where the corresponding residue is Thr494).

At the time of analysis, no other DIS3 variant had been described in a human POI patient. However, more recently, an Arg781Thr variant has been described in another human POI patient. Strikingly, this is also a highly conserved residue in the RNase II signature motif, and His774 and Arg781 are adjacent in the 3D structure of DIS3, contacting the same phosphate of the RNA substrate and possibly mediating translocation of the substrate through the active site (see discussion) strengthening our assertion that this gene and variant are causative [35, 54].

The DIS3 patient variant p.His774Tyr impairs mitotic growth in S. cerevisiae

DIS3 variants often confer mitotic defects that result in premature cell cycle arrest, unfavorable centrosome amplification and defects in spindle assemble checkpoints resulting in pathogenic aneuploidy and genomic instability [55]. In S. cerevisiae (yeast) DIS3 is an essential gene, and some point mutants have been noted to confer mitotic spindle assembly defects that are believed to impact mitotic fidelity [56, 57]. To determine whether the human DIS3 variant is deleterious, we modelled it in yeast, which has the most extensively characterized RNA exosome. Sequence alignment identified yeast His841 as the residue equivalent to His774 in human DIS3 (Figure 2 A). We therefore mutated His841 to tyrosine to mimic the patient variant. In addition, we also generated a His841 to Ala mutation. We reasoned that the polar interaction of His84 with the RNA backbone [18] may be maintained by tyrosine and that mutating His841 to Ala might better reveal the function of this residue. We used a dis3 KO yeast strain complemented with a wild-type DIS3 gene on a URA3 plasmid. This strain is unable to grow on media containing 5-Fluoro-orotic acid (5-FOA) because 5FOA selects for cells that have lost the URA3 plasmid, and consequently DIS3. This strain was transformed with either a wild-type DIS3 plasmid, a dis3-H841Y mutant, a dis3-H841A mutant or an empty vector. Transformants were then serially diluted and spotted on either 5FOA-containing media, or 5FOA-free media. As expected, the strain transformed with empty vector was unable to grow on 5-FOA media (Figure 2 B). Further, the transformants expressing either DIS3, dis3-H841Y or dis3-H841A were able to grow on the selective 5-FOA media, however growth is impaired for the mutant plasmids (Figure 2 B). This indicated that in yeast, the p.His841Tyr does not eliminate, but rather impairs mitotic growth. We and others have previously shown that mutations that eliminate catalytic activity slow growth [39, 58, 59]. Thus, it can be concluded that dis3-H841Y does not completely eliminate catalytic activity, consistent with previous biochemical characterization of the RNase II family (see discussion) [39, 58, 59].

Knockdown of Dis3 in drosophila procures a dysgenic ovarian phenotype

Drosophila Dis3 is the ortholog of human DIS3

Given that in mice oocyte specific Dis3 KO produces a sub-fertile phenotype and meiotic failure, and observations that our patient variant impairs mitotic growth in yeast we decided to further model the patient variant in Drosophila. The orthologue of human DIS3 in Drosophila, Dis3, encodes a protein with 54% identity and 70% similarity to the human DIS3 protein (Supplementary Figure S1). The affected residue at human position 774 correlating to Drosophila position 783 (Supplementary Figure S1). Information taken from (https://flybase.org/). We therefore pursued a model of our patient variant in Drosophila using a Dis3 KD rescue approach. However, prior to attempting a rescue experiment with the identified patient variant we wanted to determine the impact endogenous Drosophila Dis3 depletion had on ovarian development. This was assayed in both a germline and somatic cell context.

Dis3 KD induces an “ovary-less” phenotype in drosophila when driven in somatic gonadal precursor cells

To determine the impact of Dis3 depletion on ovarian development, we generated Drosophila with Dis3 RNAi driven under the Traffic-jam (Tj) promotor (Tj> |$Dis{3}^{RNAi}$|). Tj is expressed in somatic cells that are directly associated with germline cells of the Drosophila ovary, such as intermingled cells (ICs), cap cells, escort cells and follicle cells, throughout ovarian development and oogenesis [42, 43]. Dissection of F1 Tj> |$Dis{3}^{RNAi}$| progeny revealed no visible ovarian structures (Figure 3) or germ cells, indicating ovarian development is severely disrupted upon the depletion of Dis3. (Figure 3).

Dis3 KD in drosophila germline produces atypical ovarian development

To determine the impact of Dis3 depletion in germ cells, we generated a Drosophila line with Dis3 RNAi under the control of germline driver nanosGAL4 (nanos > Dis|${3}^{RNAi}$|⁠). nanos drives the KD of Dis3 from the apical germarium until the mature oocyte. Dissection of F1 progeny revealed the presence of ovarian structures but an absence of ovariole (chain of developing egg chambers), germarium (structure that contains undifferentiated germ cells) and germ cells (Figure 4). When compared to the nanosGAL4 > Wild-Type (WT) ovaries (Figure 4, A-C) nanos-driven Dis3 KD leads to small, poorly formed ovaries, akin to those of a rudimentary larval ovary (n = 20 ovaries) (Figure 4, D-F) [42, 60]. The absence of perinuclear VASA in immunostaining, indicates the absence of germ cells in the ovary upon depletion of Dis3 in the germline (Figure 4, F). Staining of somatic follicle hub and terminal epithelial cells via FasIII protein highlights the developing ovariole epithelia as well as polar cells at both the anterior and posterior of the developing egg chambers in WT Figure 4, A-C. The differentiation of pre-polar cells into polar cells and stalk cells maintains egg chamber separation which is essential to developing Drosophila ovarioles [61, 62]. Upon Dis3 depletion FasIII staining indicates the presence of rudimentary ovarioles (Figure 4, D-F). However, FasIII staining is disorganized and not localized to the follicular epithelial or polar cells and there is a complete absence of identifiable egg chambers or distinct germarium structures (Figure 4 E, F). In sum, depletion of Dis3 in the Drosophila ovarian germline causes dysgenic development.

The human DIS3 variant p.His774Tyr (hDIS |${\mathbf{3}}^{\mathbf{p.H774Y}}$|⁠) has reduced function in ovarian development

Somatic human DIS3 expression substantially rescues ovarian function in somatic DIS3 KD drosophila

To determine if the DIS3 c.2320C > T /p.His774Tyr variant is causative of POI in our patient we modelled the patient variant in Drosophila ovaries. To achieve this, we generated two Drosophila lines, one containing the human WT DIS3 (hDIS3) and one expressing human DIS3 with the patient variant, p.His774Tyr (hDIS|${3}^{\mathrm{p}.\mathrm{H}774\mathrm{Y}}$|)(see methods) under the UAS-promoter system. These flies were genetically recombined with BL67919 Drosophila harboring Dis3 RNAi to deplete endogenous Drosophila Dis3 (dDis3). Both lines were then crossed to somatic driver Traffic-jam GAL4 (Tj) to perform a rescue experiment.

When BL67919 Drosophila Dis3 RNAi and hDIS3 were both driven under the Tj promotor (⁠|$Tj> Dis{3}^{RNAi}, hDIS3$|) hDIS3 was able to substantially rescue the “ovary-less” phenotype observed in |$Tj> Dis{3}^{RNAi}$| (Figure 5). Ovarian structures were clearly visible with linear developing ovariole structures as well as normally oriented follicle and polar cells. Developing egg chambers were separated by a discernible series of stalk cells like that of the control (Figure 5 B, red arrowheads). These rescue data demonstrate that human DIS3 has a degree of functional redundancy in Drosophila and confirms that the ovarian phenotype observed in KD flies is indeed attributable to Drosophila Dis3 depletion and not off-target effects. Whilst rescue of endogenous Dis3 by hDIS3 is observed, there were some subtle changes when compared to the Tj > Control suggesting incomplete complementation. These are discussed in comparison with hDIS|${3}^{\mathrm{p}.\mathrm{H}774\mathrm{Y}}$| below.

h |$\mathbf{DIS}{\mathbf{3}}^{\mathbf{H774Y}}$| exhibits reduced ability to rescue ovarian loss induced by downregulation of drosophila Dis3

Similar to rescue by hDIS3, gross ovarian morphology was restored by expression of hDIS3^p.H774Y in the Drosophila ovary. This suggests that the human variant does not abolish DIS3 function. To investigate whether the p.H774Y variant may be hypomorphic, we assayed for more subtle changes in ovarian morphology, such as aberrant ovariole development.

In Tj>|$Dis{3}^{RNAi}, hDIS3$| whilst gross morphology was predominantly rescued (Figure 6 C), there were subtle changes suggesting rescue was incomplete as occasional ovarioles within Tj>|$Dis{3}^{RNAi}, hDIS3$| rescue ovaries were dysmorphic. Specifically, in the packaging and development of budding egg chambers whereby a side-by-side arrangement was observed 49% of the time in 100 ovarioles assayed Figure 6 D, G. Tj>|$Dis{3}^{RNAi}, hDIS3$| ovarioles had absent or inconsistently observed/positioned polar cells. Tj>|$Dis{3}^{RNAi}, hDIS3$| ovarioles with dysgenic polar cell arrangement also exhibited a distinct “side-by-side” development of egg chambers at stages 2–3 (S2-S3) instead of the WT linear arrangement (represented in Figure 6 A–B) (Figure 6 D, white arrowhead). Aberrant “side-by-side” arrangement of egg chambers has been previously observed and linked to varying packaging defects in the process of germline cyst encapsulation by follicle cells in the germarium [63–70]. However, fewer cases have been documented whereby paired cysts in the germarium become individually encapsulated by a follicular epithelium. Bi-layered follicular epithelium at the region of fusion of “side-by-side” egg chambers has been previously noted in Pak mutants [51]. This observation is similarly observed in Tj>|$Dis{3}^{RNAi}, hDIS3\kern0.5em$| as a defined bilayer of follicular epithelial cells can be observed between the “side-by-side” paired egg chambers (Figure 6 D). Despite “side-by-side” events occurring 34% of the times in 100 ovarioles assayed, rescue of Dis3 depletion with hDIS3 was evident with this “side-by-side” phenotype significantly exacerbated (P = 0.0294) in the presence of the H774Y variant (Figure 6). Figure 3 and Figure 5 reflect a significant conservation between human DIS3 and Drosophila Dis3.

Akin to Tj>|$Dis{3}^{RNAi}, hDIS3\kern0.5em$| a “side-by-side” arrangement of early S2-S3 egg chambers is similarly observed in Tj> |$Dis{3}^{RNAi},$| hDIS|${3}^{H774Y}$| (Figure 6 F, white arrowhead). In Tj> |$Dis{3}^{RNAi},$| hDIS|${3}^{H774Y}$| a defined bilayer of follicular epithelial cells can be observed between the “side-by-side” paired egg chambers similar to Tj>|$Dis{3}^{RNAi}, hDIS3\kern0.5em$|(Figure 6 D and F, white arrowheads). Upon closer examination, this packaging defect appears to originate in the germarium of both genotypes (Figure 6 D and F, dotted line). Here germ cells fail to aggregate into singular disc shaped cysts as seen in Figure 6 C and E, dotted line. Instead, the cysts are aberrantly paired in a germarium that is disordered and structurally shorter and wider than the Tj > Control (Figure 6 A–F).

Finally, both hDIS3 and hDIS|${3}^{\mathrm{p}.\mathrm{H}774\mathrm{Y}}$| demonstrate alternative singular and paired germarium cysts and paired/unpaired S2-S3 egg chambers within the same ovary (Figure 6, C-F). This variation in phenotype within a single ovary has been noted previously in literature [51]. Analysis of 100 ovarioles for both hDIS3 and hDIS|${3}^{\mathrm{p}.\mathrm{H}774\mathrm{Y}}$| revealed a significant (^*, P = 0.0294) increase in paired egg chambers in Tj> |$Dis{3}^{RNAi},$|hDIS|${3}^{H774Y}$| when compared to Tj>|$Dis{3}^{RNAi}, hDIS3\kern0.5em$| (Figure 6 G). This indicates that hDIS|${3}^{\mathrm{H}774\mathrm{Y}}$| likely does not function to the same extent as hDIS3, supporting the notion that the probability c.2320C > T variant is a hypomorphic allele.

Rescue with hDIS |${\mathbf{3}}^{\mathbf{p.H774Y}}$| induces premature egg chamber degradation in the drosophila ovary

Tj> |$Dis{3}^{RNAi},$| hDIS|${3}^{H774Y}$| ovaries demonstrate increased egg chamber degeneration when compared to Tj>|$Dis{3}^{RNAi}, hDIS3$| rescue flies. There are two primary checkpoints in Drosophila egg chamber development [71, 72]. The first is in region two where germline cysts are surrounded by follicle cells in the germarium to prepare for cyst budding. The second is in developmental stages 7-8 which is widely considered as a mid-oogenesis checkpoint [71, 73]. The first indication of mid-oogenesis degeneration in the germline is identifiable via irregular morphology of Nurse Cell (NC) nuclei [74]. Irregular NC nuclei were observed in Tj> |$Dis{3}^{RNAi},$| hDIS|${3}^{H774Y}$|Drosophila by the globular distribution of DAPI staining indicative of NC DNA fragmentation (Figure 7A-B, arrowheads) [75]. NC death stimulates phagocytotic activity of the surrounding follicle layer [76]. The follicular cells then transitions from a quiescent support state to a phagocytotic state for complete efferocytosis of the NC debris [77]. Once complete the remaining follicular cells undergo auto-apoptosis. Late-stage chromatin condensation is observed in hDIS|${3}^{\mathrm{p}.\mathrm{H}774\mathrm{Y}}$| rescue Drosophila (Figure 7, arrowheads). Expansion of the follicle layer is also observed in Figure 7C, arrowheads. The checkpoints for egg chambers in stages 7-8 likely reflect developmental changes in the egg chamber whereby sensitivity to aberrant function and genetic aberration is increased prior to energetically taxing vitellogenesis that occurs post stage 8 [77]. Therefore, egg chambers with incorrect follicular cell layer development, germ cell abundance and altered genetic background become targets of programmed cell death [74, 75]. A total of 14 ovaries from seven different Drosophila were dissected for both Tj>|$Dis{3}^{RNAi}, hDIS3\kern0.5em$| and Tj> |$Dis{3}^{RNAi},$| hDIS|${3}^{H774Y}$|⁠. None of the egg chambers observed exhibited germline degeneration prior to the mid-oogenesis checkpoint at stage 7-8. A statistically significantly (^**, P = 0.0052) increase in degradation at stage S7-8, however, was noted in Tj> |$Dis{3}^{RNAi},$| hDIS|${3}^{H774Y}$| ovaries with a mean occurrence of 3.570 degradation events per Drosophila ovarian pair compared to 0.1429 observed in Tj>|$Dis{3}^{RNAi}, hDIS3$| flies (Figure 7D). This indicates that the presence of p.H774Y variant increases the occurrence of egg chamber degradation indicating a likely deleterious nature of the p.H774Y variant.

Drosophila Dis3 depletion in the germline is rescued similarly by both human DIS3 (hDIS  |$\mathbf{3}$|⁠) and human variant DIS  |${\mathbf{3}}^{\mathbf{p.H774Y}}$| (hDIS|${\mathbf{3}}^{\mathbf{p.H774Y}}$|⁠)

When hDIS3  |$\mathrm{and}$|hDIS|${3}^{p.H774Y}$| are driven via nanosGAL4, in germline-Dis3 depleted Drosophila, no significant difference in ovarian morphology is observed. In both instances hDIS3|$\mathrm{and}$|hDIS|${3}^{p.H774Y}$| rescue the phenotype observed in Figure 4. Distinct ovariole and egg chamber development is observed with sequential cyst budding akin to those of the nanosGAL4 > Control ovaries Figure 8 A–C. Expression of either hDIS3 hDIS|${3}^{p.H774Y}$| rescues the lack of egg chamber development and oocytes observed in nanos-driven Dis3 depleted Drosophila. These results indicate that, in this context, the variant hDI|$S{3}^{p.H774Y}$| allows for ovarian development to proceed similarly to hDIS3. Further, upon the analysis of hDIS3  |$\mathrm{and}$|hDIS|${3}^{p.H774Y}$| driven via nanos there was no significant difference in degradation of S7–S8 egg chambers (Figure 8).

Discussion

Our data report the second unrelated case of biallelic DIS3 variants in a patient presenting with POI. Our patient was diagnosed with POI at 22 years and presented with elevated FSH, atrophic ovaries and secondary amenorrhea (Table 1). A rare homozygous missense variant in DIS3 was identified as the leading candidate for further analysis. This DIS3 variant c.2320C > T is predicted to cause an amino acid change from histidine to tyrosine in a highly conserved residue.

Prior to our analysis, no variants in DIS3 had been reported in association with human POI. Recently, however, compound heterozygous variants have been reported in two sisters with POI, a Arg781Thr missense variant and in-frame deletion of Arg624 [35]. Based on structural data of the human RNA exosome and biochemical characterization of related bacterial enzymes, His774 and Arg781 may dictate the final product of RNA digestion. The related E. coli enzymes RNase R and RNase II differ in the final product: RNase R degrades RNA from the 3′ end until the RNA is 2-3 nts long, while RNase II stops when it reaches the final 4-5 nucleotides. Both His774 and Arg781 are conserved in RNase R (His565 and Arg572) but not in RNases II (Thr494 and Lys501; Figure 2). Strikingly, mutating either of these residues in RNase R changes the end product to that of RNase II [78]. Based on these analyses, a possible effect caused by the POI mutations in DIS3 is that the same substrate RNAs are targeted, but that a longer final product is released. For capped RNAs, degradation of these final products requires DcpS, which is only active on very short RNAs [79]. Thus, we speculate that the final product of DIS3-His774Tyr (and DIS3-Arg781Thr) is a poor substrate for DcpS, leading to accumulation of short-capped mRNA fragments. One explanation for the link to POI is that over time ovaries become sensitive to overproduction of 4-5nts RNA remnants.

Johnstone et al [35] assessed the impact of DIS3 in ovarian development via the KD of Dis3 in Drosophila via a Maternal Triple Driver-GAL4. Maternal Triple Driver-GAL4 drives transgene expression in the germarium and throughout oogenesis in germline cells of the ovary [80]. However, this study lacked functional validation of their specific patient variant in their Drosophila model.

Functional validation is critical for clinical applicability of research findings. Although Dis3 is indisputably necessary for proper ovarian development, whether human missense variants disrupt DIS3 function and lead to POI requires further elucidation. We initially used yeast modelling to assay the impact of our patient variant, which indicated a small defect in mitotic replication. Similar yeast models have been used to model some other exosomopathies. Only single amino acid changes that cause subtle defects on the essential RNA exosome are compatible with a completed pregnancy and birth. These subtle effects sometimes cause effects in a yeast model, but in many cases they do not detectably affect growth [81, 82]. Given DIS3 has been implicated in both mitosis and meiosis, we considered that this yeast model may not have been optimal and pursued Drosophila modelling so we could investigate the in vivo impact of the variant on clinically-relevant ovarian development. RNAi depletion in somatic and germline ovarian cells indicates a functional requirement for Dis3 in both tissue types to facilitate normal ovarian development.

The Drosophila equivalent of mouse and human granulosa cells, in an embryological context, are the follicle cells [83]. The somatic cell driver Traffic-jam GAL4 drives the KD of Dis3 RNAi in somatic follicle cells of the Drosophila ovary. Given the striking “ovaryless” phenotype observed when Dis3 is knocked down via Traffic-jam GAL4 it is reasonable to speculate that Dis3 has a critical role in granulosa cell development.

Previous work has highlighted an important relationship between the Drosophila oocyte and its support cells, nurse and follicle, during development. When Drosophila oocyte specification is lost, nurse cells and entire egg chambers undergo apoptosis [83]. This finding is similarly observed in our data with an increased degradation of Drosophila egg chambers in the presence of the human DIS3 variant in a somatic cell context. This also alludes to a critical role of both the follicle and nurse cell populations in supporting oocytes development in Drosophila. This relationship suggests that the role of DIS3 in oocyte maturation and development, in both a somatic and germline context, may be similar in higher eukaryotes such as mice and humans given the conserved role of these cell types. hDIS3^H774Y had a reduced ability, compared to hDIS3, to rescue egg chamber degradation induced via induction of Dis3^RNAi in somatic cells of the ovary. This contrasts with the ability of hDIS3^H774Y to significantly rescue egg chamber degradation when Dis3 depletion was induced in the germline and may reflect functional differences in the two tissues or differing requirements in the levels of activity required in germline and somatic cells. This therefore suggests that human DIS3 may be more critical to somatic cell support of oocytes than the germline. However, further evidence is needed to affirm this postulation.

RNAi depletion alone gives no insight into the impact of the specific variant identified. Whether the identified missense variant destabilizes DIS3 and leads to a similar DIS3 depletion, is unknown. As observed in our work, Dis3 RNAi KD can produce differing phenotypes under various GAL4 gene drivers even within the same tissue type (nanosGAL4 vs Traffic-jam GAL4). Therefore, conclusions made on RNAi analysis alone allows for potentially egregious over/under assertions of pathogenicity for variants to be made. This can have a significant impact on patients suffering from POI and could lead to an incorrect causative gene being identified.

The variant c.2320C > T observed in our patient, when expressed in Dis3-depleted Drosophila under Tj-GAL4 substantially restored ovarian development. In contrast to Dis3 RNAi depleted Drosophila that had no visible ovarian structures, hDIS|${3}^{p.H774Y}$| transgenic expression led to ovaries with normal gross morphology. When compared to the degree of rescue achieved by hDIS3, however, there were significant differences. This included aberrant polar cell arrangement in sequentially developing egg chambers as well as the presence of paired egg chambers. Paired egg chamber development is associated with failed germline cysts encapsulation by follicle cells. This observation is significantly increased in the presence of human variant H774Y DIS3 when driven via Traffic-jam GAL4, a follicle cell driver. This is indicative of the role of the DIS3 variant in the process of germline cyst encapsulation and, in turn, somatic support cells of the Drosophila ovary. These observations indicate that the c.2320C > T DIS3 variant may be a hypomorphic allele. Such analysis is lacking for the previously reported DIS3 variant and without such data, the impact of the variant remains unknown. Our data highlight the importance of in vivo variant modelling in the assessment of gene consequence.

DIS3 variants impact oocyte integrity and meiotic progression

DIS3 is a 3′-5′ catalytic exoribonuclease effector of RNA maturation and decay [32]. Post transcriptional modification and degradation of RNA species is considered critical to the modulation of gene expression and cell survival [84]. The function of DIS3 is well conserved and human DIS3 expression can complement Drosophila Dis3 depletion (Figure 5) or a point mutation in yeast DIS3, validating our use of these model organisms [85]. In somatic cells, DIS3 is specific to the nuclear RNA exosome, while DIS3L replaces it in the cytoplasmic exosome. DIS3 participates in the maturation of ribosomal RNAs from its precursor, as well as degradation of other RNAs, including PROMPTs and eRNAs. During development mammalian oocytes arrest at the first meiotic prophase and are transcriptionally inactive [86]. In this state there would be very little maturation and degradation of these RNAs. Instead, available RNA species in the oocyte cytoplasm predetermine the maternal transcriptome for subsequent embryonic development [87]. Whether DIS3 assumes a more cytoplasmic role in this situation is unknown. The modulation of the RNA transcriptome via RNA degradation pathways is therefore essential to oocyte development [34]. Aggregation of unwanted RNA transcripts, due to genetically perturbed RNA degradation machinery, are known to negatively impact meiosis, chromosome integrity, fertilization, zygotic gene activation and chromatin remodeling [34, 88, 89]. In mice depletion of Dis3 in oocytes induces the accumulation of pervasive intergenic transcripts which is a result of both insufficient RNA degradation and defective transcription termination [34]. Together these changes prevent Dis3 KO oocytes from resuming meiosis from their transcriptionally quiescent state. Failure of the oocyte to re-enter meiosis leads to a significant reduction in fertility for KO mice [34].

Dis3 is considered critical to Drosophila multicellular development with temporal/spatial KD/KO resulting in complete loss of wing and ovarian structures and perturbed brain development [90, 91]. Dis3 depleted ubiquitously in Drosophila is larval lethal at the L2 stage of development [90]. Drosophila Dis3 participates in cell specific RNA turnover and metabolism both independently and in association with the RNA exosome [90]. In the ovary, conditional loss of exosome components, such as Dis3, result in premature germarium stage defects and a complete absence of the germline when driven under nanosGAL4 [92]. We similarly observed disrupted ovarian development in both germline and somatic cell Dis3 KD.

Akin to mice, literature demonstrates that RNA degradation plays a key role in regulating Drosophila oocyte fate and inception [92]. Aberrant persistence of RNAs that escape degradation have been shown to induce a loss of oocyte maintenance and premature egg chamber death. Expression of patient variant, DIS|${3}^{p.H774Y}$|, in Dis3-depleted somatic cells of the Drosophila ovary via Traffic-jam GAL4 induces a significant degradation of developing egg chambers when compared to wild-type hDIS3. This egg chamber arrest and degradation occurs at the mid-oogenesis checkpoint of ovariole development. Mid-oogenesis egg chamber death is also observed in Drosophila Pelota mutants, defective in “no-go mRNA decay” and in Twister mutants, defective in a cytoplasmic RNA exosome cofactor. Twister mutants show oocyte defects in a germline context when driven via nanosGAL4 and no defects when driven somatically via Traffic-jam GAL4 [92]. However, it is noted that defects in other genes involved RNA degradation pathways have varying phenotypes in Drosophila depending on if they target nuclear or cytoplasmic RNAs [92]. For example, RNA degradation pathways involved in nuclear RNA degradation (rRNA-processing 6 (rrp6) and mRNA transport 4 (mtr4) or simultaneous nuclear and cytoplasmic RNA degradation, (Dis3) result in earlier germline defects or a complete loss of the germline when compared to cytoplasmic exclusive factors [92]. This indicates that it is the persistence of unwanted RNA species in such mutants that results in a loss of oocyte maintenance and egg chamber death [92]. This suggests that at mid-oogenesis high fidelity maintenance of the maternal contribution of nuclear RNAs into the developing oocyte is required for meiotic progression and successful oogenesis. The presence of our patient variant DIS|${3}^{p.H774Y}$| perturbs this process in our Drosophila model.

Variants in DIS3 have the potential to negatively impact the 3′-5′ exoribonuclease and RNA degradation activity of the basal subunit. This may allow for pervasive RNAs to be retained within a cell which, as demonstrated in literature, is unfavorable to oocyte maintenance and results in egg chamber death. This is because germ cells exist in a specialized somatic cell niche that supports germ cell growth and maturation [93]. The nurse cells in Drosophila transfer accumulated RNAs to support oocyte and embryonic development [93]. There is some evidence that reproductive support cells, such as follicle cells, may traffic RNAs to the germline, but further details need to be elucidated [93].

In mice, somatic cell cues are essential for translation in oocytes and the regulation of such is critical to embryonic development [94]. Therefore, if RNA degradation and regulation pathways are perturbed in somatic cells this has the potential to impact the functional niche and consequently the successful development of oocytes [94]. Our data supports the necessity of RNA regulation via Dis3 in Drosophila ovarian somatic cell development and how the absence of Dis3 in somatic cells results in a complete loss of germ cells and ovarian structures.

The DIS3 variant identified DIS|${3}^{p.H774Y}$| in our patient negatively impacts egg chamber development supporting a potential pathogenic effect on RNA clearance in developing oocytes. These findings align with the presenting patient phenotype of secondary amenorrhea rather than complete loss of ovarian development and function (Table 1). Given the ubiquitous role of the RNA exosome in cellular function, it could be predicted that complete LoF could cause multisystem disease. Indeed, variants in DIS3 and exosome components have been implicated in other diseases such as cancer [31, 32]. The isolated POI phenotype of the patient described in this study and the mild functional impact of the variant could reflect the sensitivity of the ovaries to damage. It is likely that the hypomorphic allele only partially disrupts DIS3 function and this disruption is enough to cause oocyte death but is tolerated by other organs and tissues.

Although the phenotype observed in our rescue flies is mild, our data are consistent with functional impairment and a hypomorphic allele. Whether this degree of impairment translates to human POI remains uncertain. Supporting the likelihood of DIS3 being a bona fide “POI gene” is the fact that its disruption in multiple animal models leads to an ovarian phenotype and independent rare predicted-pathogenic DIS3 variants have now been identified in two unrelated individuals with POI. These data highlight the critical role of the RNA exosome and its components in fertility and indicates that variants in DIS3 should be considered a potential cause of POI.

Acknowledgment

The research conducted at the Murdoch Children’s Research Institute was supported by the Victorian government’s operational infrastructure support program.

Author contributions

EJT and AHS conceived the study. BLK, EJT, NAS, FW, CJS, LO, GR, FA, SS and JvdB acquired the data. BLK, GRH, EJT, NAS, SJ, LO, AHS, KLA, AvH, LO, FA, and SS analysed and interpreted the data. KMB, EJT performed bioinformatic and genomic analysis. SJ, FA and SS recruited patients, liaised with families, collected and reviewed clinical data BLK and EJT wrote the manuscript with contribution from all authors. All authors approved this manuscript.

Conflict of interest: On behalf of all authors, the corresponding authors states that there is no conflict of interest.

Ethics approval and informed consent

Written informed consent was obtained from each participant. All procedures were approved by the Human Research Ethics Committee of the Royal Children’s Hospital, Melbourne (HREC/22073).

Data availability

Restrictions apply to the availability of some, or all data generated or analysed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.

Footnotes

References