Mitochondrial Dysfunction in Primary Ovarian Insufficiency

Abstract

Primary ovarian insufficiency (POI) is defined by the loss or dysfunction of ovarian follicles associated with amenorrhea before the age of 40. Symptoms include hot flashes, sleep disturbances, and depression, as well as reduced fertility and increased long-term risk of cardiovascular disease. POI occurs in ∼1% to 2% of women, although the etiology of most cases remains unexplained. Approximately 10% to 20% of POI cases are due to mutations in a single gene or a chromosomal abnormality, which has provided considerable molecular insight into the biological underpinnings of POI. Many of the genes for which mutations have been associated with POI, either isolated or syndromic cases, function within mitochondria, including MRPS22, POLG, TWNK, LARS2, HARS2, AARS2, CLPP, and LRPPRC. Collectively, these genes play roles in mitochondrial DNA replication, gene expression, and protein synthesis and degradation. Although mutations in these genes clearly implicate mitochondrial dysfunction in rare cases of POI, data are scant as to whether these genes in particular, and mitochondrial dysfunction in general, contribute to most POI cases that lack a known etiology. Further studies are needed to better elucidate the contribution of mitochondria to POI and determine whether there is a common molecular defect in mitochondrial function that distinguishes mitochondria-related genes that when mutated cause POI vs those that do not. Nonetheless, the clear implication of mitochondrial dysfunction in POI suggests that manipulation of mitochondrial function represents an important therapeutic target for the treatment or prevention of POI.

Primary ovarian insufficiency (POI), which is also commonly referred to as premature ovarian failure, is defined by the loss or dysfunction of ovarian follicles associated with amenorrhea before the age of 40 (1). POI is a major cause of female infertility, with a prevalence between 1% and 2%. It can occur spontaneously or be the result of medical interventions such as removal of the ovaries, chemotherapy, or radiation treatment. Symptoms of POI are similar to menopause and commonly include hot flashes, sleep disturbances, vaginal dryness, and night sweats (2). Other less commonly reported symptoms include decreased energy, dry eyes, hair loss, urinary incontinence, cold intolerance, cognitive changes, depression, and joint pain, among others (3). Longer-term consequences associated with the premature decrease in estrogen levels include increased risk of cardiovascular disease and decreased bone mineral density (4, 5). Symptom severity can be correlated with ovarian function, as POI encompasses a range of decreased ovarian functions from individuals with partial function who retain fertility to those who completely lack ovarian function and are infertile. Treatment will vary depending on the symptoms, but it can include hormone replacement therapy, fertility management, and psychosocial support, as well as annual screenings of thyroid and adrenal function (2).

In cases of POI that are not induced by chemotherapy or radiation, the etiology is determined for only ∼20% of all cases (6). Approximately 5% of all POI cases are of autoimmune origin, and 15% of cases have a clear genetic origin, either in the form of a chromosomal abnormality or a mutation in an individual gene (7, 8). Among the chromosomal abnormalities that are frequent causes of POI are a number of X chromosome defects, including Turner syndrome, triple X syndrome, and fragile X syndrome. A number of monogenic disorders resulting in POI have also been identified, with variants in >50 genes having been associated with POI (8). The genes that have been implicated in POI fall within a number of pathways that are critical for ovarian development and function, including DNA repair, meiosis, germ cell recruitment, and steroidogenesis. Additionally, mutations in many genes involved in mitochondrial function have been identified as causes of POI. Thus, this review summarizes the role of mitochondria in oocytes as well as discusses the genetic causes of POI related to mitochondrial dysfunction.

Mitochondrial Composition and Function

Mitochondria are maternally inherited double-membrane organelles that are best known for their role as the powerhouses of a cell, based on their generation of ATP via the process of oxidative phosphorylation (OXPHOS). The mitochondrial genome is ∼16.7 kb and encodes 13 proteins that function within the OXPHOS pathway, as well as 22 tRNAs and 2 rRNAs. However, estimates of the complete human mitochondrial proteome suggest the presence of at least 1500 proteins, and thus the vast majority of proteins localized to mitochondria are encoded by the nuclear genome (9). The proteomic complexity of mitochondria enables them to perform many critical cellular functions beyond energy production such as macromolecule biogenesis (i.e., protein and nucleic acids), lipid synthesis, regulation of cell death, calcium handling/homeostasis, generation of reactive oxygen species (ROS), and antioxidant protection (10).

OXPHOS is the process by which nutrients are oxidized by a series of five multisubunit protein complexes (complex I to complex V) within the mitochondrial inner membrane, ultimately resulting in the conversion of ADP to ATP. During the process of ATP generation, mitochondria also release ROS in the form of superoxide. Although typically thought of as a byproduct of OXPHOS that can damage cells, ROS possess antimicrobial properties, act as signaling molecules, and regulate autophagy (11). Nonetheless, excessive or mislocalized ROS can have deleterious effects on the cell and can induce oxidative damage to mitochondrial DNA (mtDNA), the latter of which may be particularly sensitive to damage due to the relative absence of DNA repair enzymes in mitochondria (12). ROS are eliminated in the mitochondria by three superoxide dismutases that convert superoxide to hydrogen peroxide, which can then be further converted to water by peroxiredoxins and glutathione peroxidases.

Mitochondria are not static organelles, rather their numbers are tightly regulated by balancing the processes of mitochondrial biogenesis and clearance (13). Mitochondrial clearance can be mediated by the process of mitophagy, which describes the degradation and recycling of mitochondria via autophagy. Triggers for mitophagy include mild increases in ROS (14) and the inability to maintain an electrochemical gradient that can result from the accumulation of mtDNA damage (15). Mitochondrial biogenesis is then used to repopulate the cellular mitochondrial population with more functional mitochondria. The process of mitochondrial biogenesis involves both fusion and fission events, and it is regulated by a cascade of transcription and translation factors, both nuclear and mitochondrial encoded, which serve to generate the materials necessary for organelle duplication (16). Mitochondrial fusion describes the merging of the outer and then inner membranes of two previously separate mitochondria, whereas mitochondrial fission is the process by which a single mitochondrion divides into two or more independent mitochondria. Mitochondrial fusion, fission, and mitophagy collectively serve not only an important quality control function, but can link changes in mitochondrial abundance to cellular energy demands (17). For example, nutrient deprivation prevents mitochondrial fission, which together with other molecular and structural changes to the mitochondria serves to increase ATP synthesis capacity, whereas thermogenesis or AMP-activated protein kinase activation serves to increase fission (18–20). Nutrient excess inhibits fusion, resulting in an increase in fragmentation, which can cause mitochondrial dysfunction and increase ROS production (21). OXPHOS activity is positively correlated with fusion, although the causal nature of this relationship remains unclear (22). Mitophagy can be triggered by energy stress via activation of AMP-activated protein kinase, hypoxia, and glucose deprivation (22). Collectively, the regulation of mitochondrial dynamics ensures that these organelles can help the cell respond and adapt to a variety of cellular states to match the bioenergetic and other mitochondrial functions with the cellular energy needs.

Mitochondrial Function in Ovarian Somatic Cells

Beyond the role of mitochondria in OXPHOS, ROS homeostasis, apoptosis, and thermogenesis, mitochondria have an important role in steroidogenesis. Steroidogenesis occurs in the cumulus granulosa and theca cells, which surround the oocyte, according to the two-cell, two-gonadotropin model. The first step in steroidogenesis, which is also the rate-limiting step, involves the transport of cholesterol from the outer mitochondrial membrane to the inner mitochondrial membrane by the steroid acute regulatory protein (STAR) and other accessory proteins (23). Once inside the mitochondria, cholesterol is metabolized to pregnenolone by cytochrome P450 family 11 subfamily A member 1 (CYP11A1, also known as P450SCC), which is then further metabolized in the endoplasmic reticulum to form the final steroid products, including estradiol and progesterone (24, 25).

Mitochondrial Function in Oocytes

Oogenesis, the process by which a haploid oocyte is formed within the ovary, originates from primordial germ cells that arise in the extraembryonic mesoderm, which then migrate to the genital ridge where they proliferate by mitosis and transform to oogonia (26). Following the arrest of mitosis, human germ cells initiate meiosis in the fetal ovary at 11 to 12 weeks of gestation and become primary oocytes (27). Oocytes enter meiosis in the prophase stage, during which time homologous chromosomes undergo synapsis and recombination. Subsequent to the crossing over of genetic material during recombination, oocytes progress to the diplotene stage, which is defined by a protracted arrest. During this resting state, oocytes become surrounded by pregranulosa cells to form primordial follicles, although most oocytes undergo apoptosis during this process and do not survive to form follicles (28). Oocyte survival appears in part to depend on the transfer of organelles, including Golgi and mitochondria, from nurse cells connected by intercellular bridges that are associated with the formation of a Balbiani body in the oocyte, a structure that is densely packed with Golgi, endoplasmic reticulum, and mitochondria (29, 30). The Balbiani body is a transient formation and disappears in late-stage oocytes. A subset of primordial follicles are further induced via intra-oocyte and extra-oocyte factors involving the phosphoinositide 3-kinase and mammalian target of rapamycin pathways to transition to primary follicles (31). The primary follicles then further develop to form secondary follicles, which are surrounded by at least two layers of granulosa cells and an outermost layer of theca cells. The theca cells contain many mitochondria with vesicular cristae (32). Further development preceding ovulation includes the formation of a cavity called the antrum and the selection of dominant follicles that will eventually ovulate. Reinitiation of meiosis in the dominant follicles can follow a surge of endogenous LH during puberty, but these follicles can also remain in the dormant state from the time of their formation during fetal development until menopause, a period of ∼50 years (33).

There is a tremendous increase in the number of mitochondria per cell that occurs during oogenesis, from the 10 to 100 mitochondria found in primordial germ cells to the >100,000 mitochondria found in a mature preovulatory oocyte (34). The small number of mitochondria in primordial germ cells contributes to the establishment of an inheritance bottleneck, in which rapid shifts in the inheritance of mitochondrial variants can occur between generations based on the small number of mitochondria that form the basis of the rapid expansion (35). Coupled with a dramatic reduction in mtDNA in primordial germ cells, there exists a strong selective pressure, including influence from the nuclear genome, to eliminate deleterious mitochondrial variants that are frequently present at low levels (36–38).

Although the numbers of mitochondria rapidly increase during oogenesis, they maintain a state of relatively low activity until the blastocyst stage. Mitochondria found within oocytes have a morphology that is more rounded than the typical rod-shaped morphology found in somatic cells, and there are fewer cristae found in the inner membrane (39). Oocyte mitochondria also have a relatively lower rate of oxygen consumption, together suggesting that mitochondria within the oocyte remain in a state of relatively low activity. It has been hypothesized that the large numbers of mitochondria functioning at minimal levels are sufficient to provide just the right amount of energy to the oocyte while minimizing the frequency of mtDNA mutations and the production of ROS (40). This activity is nonetheless critical for oocyte maturation because glycolysis is limited during oogenesis, creating a heavier reliance on ATP generated by OXPHOS. In addition to the energy generated by oocyte mitochondria, energy transfer from the surrounding granulosa and cumulus cells, in the form of extracellular pyruvate and glutamine, also contributes toward the energetic requirements of oocyte maturation and early embryogenesis (41). The relatively large numbers of “quiet” mitochondria during oogenesis further suggests that other functions of mitochondria beyond ATP production may be important at this stage, including the production of the reduced form of nicotinamide adenine dinucleotide phosphate and tricarboxylic acid (TCA) cycle intermediates, which may be used to decrease oxidative stress and provide substrates for other biosynthetic pathways (42).

In addition to dramatic changes in the numbers of mitochondria during oogenesis, the subcellular localization of these organelle is remodeled to suit the changing energy demands of the cell during maturation. The most significant shift in subcellular localization occurs when >40% of mitochondria accumulate to surround the forming meiosis I spindle, remaining with the spindle during its movement from the oocyte center to the periphery (43). This association of mitochondria with the developing spindle is dependent on their trafficking along microtubules in a dynein-dependent process. Following the first meiotic division, there is an asymmetric distribution of mitochondria among the daughter cells favoring the oocyte over the polar body, which is of particular importance given the paucity of mitochondrial biogenesis from this stage of oogenesis through implantation (44). Mitochondria again associate with the meiosis II (MII) spindle during its formation; however, following the completion of the MII spindle formation and arrest at the MII stage, the mitochondria no longer completely encase the spindle, but do remain somewhat enriched in the immediate vicinity (44).

Mutations in Mitochondrial Proteins in POI

Mitochondrial disease represents a group of clinically heterogeneous disorders that can typically be grouped together based on their common feature of having a primary defect in OXPHOS (45). The prevalence of mitochondrial disease in children is ∼6 of 100,000 and in adults is ∼23 of 100,000 (46, 47). Mitochondrial diseases can be caused by mutations in either mtDNA or nuclear-encoded genes, with mutations in >350 genes having been reported as a cause (48). The clinical presentations are varied in their onset, inheritance pattern, and features, but they are often found in organs with high energy demands, including the brain, skeletal muscle, and heart. Frequent features include hypertrophic cardiomyopathy, heart conduction defects, myopathy, sensorineural deafness, cerebellar ataxia, epilepsy, and peripheral neuropathy, among others (49). Nonsyndromic presentations of mitochondrial disease include mutations in the 12S rRNA that cause nonsyndromic deafness and certain individuals with Leber optic atrophy, which is caused by mutations in multiple genes, that present with visual loss as the only clinical feature (50, 51). This brings up one of the more perplexing features of mitochondrial disorders, which is that despite the fact that mitochondria are critical for the function of every cell in the body, their clinical presentation can often be organ or cell specific. Additionally, not all patients with mitochondrial disease have OXPHOS defects, as evidenced by patients with TCA cycle defects due to mutations in aconitase 2 (ACO2) who present with infantile cerebellar-retinal degeneration. ACO2, which catalyzes the conversion of citrate to isocitrate within the TCA cycle, is not consistently associated with defects in OXPHOS in patient-derived tissue samples, but rather is characterized by abnormal TCA metabolite levels and mtDNA depletion (52–54).

In addition to the frequently observed defects in the heart, brain, and muscle, the presentation of mitochondrial dysfunction often presents with a range of endocrine features, including diabetes mellitus, GH deficiency, hypogonadism, thyroid disease, and ovarian dysfunction, among others (55, 56). Ovarian dysfunction can be seen as part of a complex syndromic presentation, but it also includes disorders where it is the only clinical finding. Mitochondrial mutations that lead to ovarian dysfunction have been linked to mtDNA, mitochondrial RNA, and mitochondrial protein synthesis, with functions in multiple distinct biological pathways (57). Much of what is known about the pathophysiology underlying ovarian dysfunction has been learned from the identification of disease genes identified in rare monogenic disorders (58). Below is a discussion of eight genes that when mutated give rise to either isolated POI or POI as part of a syndrome (Table 1), as well as three additional genes that have recently been described in a single patient or family, suggesting a potential link between those genes and POI as well.

Table 1.

Mitochondria-Related Genes in Which Mutations Can Cause POI

Gene	Inheritance	Clinical Presentation	Gene Function	Molecular Function in Mitochondria
POLG	AR, AD	Progressive external ophthalmoplegia	DNA polymerase	mtDNA replication and maintenance
		Mitochondrial DNA depletion syndrome
		Mitochondrial recessive ataxia syndrome
TWNK	AR	Perrault syndrome	mtDNA helicase	mtDNA replication and proofreading
		Mitochondrial DNA depletion syndrome
		Progressive external ophthalmoplegia
LARS2	AR	Perrault syndrome	Leucine tRNA	mRNA translation
HARS2	AR	Perrault syndrome	Histidine tRNA	mRNA translation
AARS2	AR	Ovarioleukodystrophy	Alanine tRNA	mRNA translation
		Combined OXPHOS deficiency
CLPP	AR	Perrault syndrome	Protease	Protein degradation
LRPPRC	AR	Leigh syndrome	RNA binding protein	Gene expression
MRPS22	AR	Isolated POI	Ribosomal subunit	mRNA translation
		Combined OXPHOS deficiency

Abbreviations: AD, autosomal dominant; AR, autosomal recessive.

Rare Monogenic Mitochondrial Disorders Associated With POI

DNA polymerase γ, catalytic subunit

DNA polymerase γ, catalytic subunit (POLG) is the catalytic subunit of the mitochondrial DNA polymerase that is responsible for replication of the mitochondrial genome (59). Mutations in POLG cause depletion of the mtDNA and/or the accumulation of deletions in the mtDNA and are the most commonly inherited form of monogenic mitochondrial disease (60). Mutations in POLG cause a spectrum of disorders, including myocerebrohepatopathy spectrum, Alpers–Huttenlocher syndrome, and progressive external ophthalmoplegia, among others, but they lack a clear cut genotype–phenotype correlation between the >300 described pathogenic mutations and the clinical presentation (61). Among the disparate clinical features associated with POLG mutations are epilepsy, ataxia, parkinsonism, hearing loss, cataracts, as well as POI (60, 62). Although the clinical presentation of POLG mutations can vary, POI has typically been described as part of a syndrome including the previously mentioned neurologic features, and not in isolation (55). However, a recent case report described identifying the first case of nonsyndromic ovarian dysfunction associated with a mutation in POLG (p.R964C), although caution must be taken in interpreting this study until additional cases are identified (63). Finally, although candidate gene–based studies have previously failed to find evidence of POLG mutations as a common cause of POI (64, 65), a recent meta–genome-wide association study analysis identified an association between a single-nucleotide polymorphism located near POLG (rs1054875) and the age at natural menopause, suggesting that variation in POLG may contribute to the more common polygenic forms of POI (66).

Twinkle mtDNA helicase

Twinkle mtDNA helicase (TWNK; also known as C10ORF2) encodes a mitochondrial helicase that together with the above-mentioned POLG, the DNA polymerase γ accessory subunit POLG2, and the mitochondrial single-stranded DNA-binding protein SSBP1 form the core factors required for mtDNA replication (67, 68). Similar to POLG, mutations in TWNK can cause a spectrum of genetic disorders, including progressive external ophthalmoplegia, mtDNA depletion syndrome 7 (hepatocerebral type), and Perrault syndrome, with clinical presentations that can contain POI in association with other neurologic findings. For example, 2 unrelated families each had 2 individuals present with ataxia, neuropathy, hyporeflexia, progressive hearing loss, and ovarian dysgenesis, with affected individuals in both families found to be compound heterozygous for missense mutations in TWNK (69). Given the functional overlap between POLG and TWNK in mtDNA replication, it is perhaps not surprising that mutations in both genes can cause a similar constellation of clinical features, including POI, and together clearly implicates the importance of mtDNA replication in ovarian development. However, thus far mutations in the other core components of mtDNA replication, POLG2 and SSBP1, have not been associated with ovarian dysfunction, although mutations have been observed in POLG2 that, similar to POLG, cause progressive external ophthalmoplegia (70).

Leucyl-tRNA synthetase 2, mitochondrial and histidyl-tRNA synthetase 2, mitochondrial

Histidyl-tRNA synthetase 2, mitochondrial (HARS2) and leucyl-tRNA synthetase 2, mitochondrial (LARS2) both encode aminoacyl-tRNA synthetases (aaRSs) that are used for the translation of mitochondrial-encoded genes. Mitochondrial aaRSs can be categorized into two major classification groups, class I and class II, based on their structural motifs, with LARS2 encoding a class I aaRS based on the presence of a Rossmann fold that serves as an ATP biding region and HARS2 encoding a class II aaRS based on three structural motifs referred to as 1, 2, and 3 (71). Mutations in both HARS2 and LARS2 are associated with Perrault syndrome, which describes a rare recessively inherited condition composed of sensorineural hearing loss in both males and females and ovarian dysfunction, including ovarian dysgenesis and primary amenorrhea, and can include neurologic features including ataxia, learning disability, and neuropathy (72–74). Testicular dysfunction has not been observed in males with the same mutations that cause ovarian dysfunction in females (72). However, the number of deaf males with Perrault syndrome mutations whose sperm function has been clinically evaluated is small, and a mouse model of Perrault syndrome due to complete Clpp deficiency (discussed below) is associated with infertility in both males and females (75, 76). Among the 19 mitochondrial aaRSs, it remains unclear why mutations in these two particular genes result in hearing loss and ovarian dysfunction, whereas mutations in other mitochondrial aaRSs result in multiorgan systemic disorders (77). However, although most disease-causing mutations in aaRSs occur at residues that are not evolutionarily conserved, mutations in both HARS2 and LARS2 have been described at conserved positions, suggesting that the mutations in these genes may similarly affect the canonical functions of these proteins (78).

Alanyl-tRNA synthetase 2, mitochondrial

Alanyl-tRNA synthetase 2, mitochondrial (AARS2), similar to HARS2 and LARS, is a mitochondrial aaRS; however, mutations in AARS2 do not cause Perrault syndrome, but instead are associated with recessively inherited ovarioleukodystrophy syndrome in females (79–83). Mutations in AARS2 have also been linked to infantile mitochondrial cardiomyopathy and primary pulmonary hypoplasia resulting in early childhood lethality (84, 85). The ovarioleukodystrophy syndrome consists of neurologic features stemming from neurologic deterioration, including ataxia, spasticity, tremor, and cognitive decline, with ages of onset ranging from childhood to adulthood, in the absence of cardiomyopathy. Ovarian failure occurs in the 20s or 30s and has been reported with both primary amenorrhea or secondary amenorrhea. AARS2 mutations that lead to infantile cardiomyopathy reduce protein stability with minimal effects on aminoacylation, whereas the molecular basis underlying the ovarioleukodystrophy syndrome remains unknown (86). However, it has been hypothesized that the ovarioleukodystrophy is associated with a reduction in aminoacylation efficiency, although this remains to be experimentally validated (87).

Caseinolytic mitochondrial matrix peptidase proteolytic subunit

Caseinolytic mitochondrial matrix peptidase proteolytic subunit (CLPP) is an ATP-dependent protease that is localized to the mitochondrial matrix where, together with caseinolytic mitochondrial matrix peptidase chaperone subunit (CLPX), it is part of the quality control system that degrades oxidized and denatured proteins (88). A number of CLPP substrates for degradation have been identified and include proteins involved in electron transport, metabolic processes, the TCA cycle, and mitochondrial mRNA, among others (89–91). Consistent with the identified substrates, CLPP deficiency in experimental model systems leads to reduced mitochondrial respiration, increased ROS, and impaired mitochondrial protein synthesis associated with defects in assembly of the mitochondrial ribosome (89, 92). Mutations in CLPP cause recessively inherited Perrault syndrome type 3 associated with ovarian dysfunction, hearing loss, and neurologic defects such as lower limb spasticity, epilepsy, microcephaly, and learning difficulties (93–96). Functional characterization of the identified mutations in CLPP resulted in a variety of functional defects, including decreased peptidase activity, inhibition of CLPX binding, prevention of oligomerization, and enhanced peptidase activity, suggesting that despite the clustering of most CLPP mutations near the CLPX-docking site or the active site of the peptidase, there may not be a common defect in the molecular function of CLPP (97). A mouse knockout of Clpp has also been generated, which similarly demonstrated hearing loss and infertility due to the failure of ovarian follicular differentiation in females, as well as growth retardation, decreased activity, impaired survival, and disruption of spermatogenesis in males, providing a model to better study the cellular and molecular defects associated with CLPP deficiency (76).

Leucine-rich pentatricopeptide repeat containing

Leucine-rich pentatricopeptide repeat containing (LRPPRC) is an RNA-binding protein that plays a critical role in regulating mitochondrial gene expression. LRPPRC together with the interacting protein SRA stem-loop interacting RNA binding protein (SLIRP) directly bind to and coat nearly all mitochondrial mRNAs, thereby regulating their structure to improve their translational fidelity and increase their stability (98, 99). Although SLIRP and LRPPRC both depend on the other for stability, they also have unique functions within their complex, with only LRPPRC being required for polyadenylation of mitochondrial mRNAs, whereas SLIRP functions to target mRNAs to the mitochondrial ribosome (100). Nuclear functions have also been ascribed to LRPPRC, including as a cofactor for eukaryotic translation initiation factor 4E (eIF4e) that regulates nuclear export and translation, and a component of the PGC1α transcription regulation complex that controls the expression of key metabolic genes such as phosphoenolpyruvate carboxykinase (PEPCK) and glucose-6-phosphatase catalytic subunit (G6PC) in addition to genes involved in mitochondrial biogenesis (101, 102).

Mutations in LRPPRC cause a rare type of Leigh syndrome referred to as French-Canadian, because most patients originate from the Saguenay-Lac-Saint-Jean region of Québec. A p.A354V founder mutation is present in this region at a relatively high allele frequency, resulting in an incidence of ∼1 of 2000 births (103–106). Leigh syndrome, French-Canadian variety, has an autosomal recessive inheritance pattern with clinical features that include developmental delay, failure to thrive, characteristic facial features, and hypotonia. Patients are at high risk of death due to neurologic or acidotic crises that can be instigated by triggers, including infection, stress, diet, illness, or exercise. Although life expectancy is typically <5 years, patients can survive to adulthood (107). Female patients who survive until at least adolescence, in addition to the neurologic features, present with POI, including elevated FSH, small ovaries that lack follicles, and primary amenorrhea (106). Although LRPPRC clearly has a significant effect on expression of mtDNA-encoded genes, caution must be taken when interpreting this as conclusive support for the role of mitochondrial dysfunction in POI, given that LRPPRC also regulates nuclear gene expression.

Mitochondrial ribosomal protein S22

Mitochondrial ribosomal protein S22 (MRPS22) is a component of the 28S small subunit of the mitochondrial ribosome that enables the translation of mtDNA-encoded polypeptides (108). MRPS22 is also required for maintaining the stability of the mitochondrial ribosomal complex (109). Mutations in MRPS22 have been identified to cause autosomal recessive inheritance of POI (110, 111). Unlike the syndromes mentioned above, mutations in MRPS22 caused isolated POI in the absence of other neurologic or other syndromic features. Surprisingly, MRPS22 deficiency did not cause defects in OXPHOS, suggesting that the etiology of POI was related to the nonbioenergetic functions of mitochondria. In Drosophila, germ cell–specific deficiency of the MRPS22 ortholog resulted in infertility, suggesting a cell-autonomous role for MRPS22 in germ cell development (110). Mutations in MRPS22, which are presumably more deleterious to protein function, have been identified that impair OXPHOS activity and lead to systemic disease, including cardiomyopathy, hypotonia, brain abnormalities, and in certain instances lethality (112–115). Thus, variants in MRPS22 are associated with a phenotypic spectrum of disorders ranging from isolated POI to infantile mitochondrial lethality (110).

Mitochondria-Related Candidate Genes for POI Identified in Single Families or Probands

Mitochondrial ribosomal protein S7, protein only RNase P catalytic subunit, and required for meiotic nuclear division 1 homolog

Caution must be exercised when interpreting mutations that have thus far been identified in only a single family, as the sheer number of potentially deleterious variants present in all individuals makes it difficult to identify definitively the true causal variants. Nonetheless, the initial identification and reporting of these mutations may facilitate other investigators who encounter mutations in those genes and is thus worthwhile. As such, mutations have been reported in protein-only RNase P catalytic subunit (PRORP) as a novel cause of Perrault syndrome (116). A single consanguineous family was identified with three individuals presenting with sensorineural hearing loss and POI that were each homozygous for a missense variant in PRORP (p.A485V). Functional studies in patient-derived fibroblasts supported the presence of mitochondrial dysfunction as evidenced by impaired processing of mitochondrial RNA transcripts and an accompanying decrease in protein levels of multiple OXPHOS components (116). Similarly, another consanguineous family was identified with two female siblings who presented with congenital sensorineural deafness and lactic acidemia. One sibling had a significantly more severe presentation, including failure to thrive, hepatomegaly, and hypoglycemia, and ultimately died at 14 years of age. The sibling has hypoglycemia in early childhood that resolved, mild learning difficulties, and was found at age 16 to have POI (117). Both siblings were homozygous for a mutation in the mitochondrial ribosomal protein S7 (MRPS7) at a highly conserved amino acid residue (p.M184V), and functional studies in patient-derived fibroblasts identified OXPHOS defects consistent with a mitochondrial disorder (117). Another study identified a homozygous missense mutation in required for meiotic nuclear division 1 homolog (RMND1; p.N238S) that had previously been reported as a disease causing mutation associated with neurologic, muscle, and kidney defects (118, 119). However, the proband in this study carrying the RMND1 variant had the characteristic features of Perrault syndrome, including POI that was observed at 10 years of age, as well as kidney and growth defects (118). Mutations in RMND1 have previously been associated with defects in mitochondrial mRNA translation (120, 121). Although these studies each describe a strong candidate gene for POI, with the studies for MRPS7 and PRORP in particular elegantly combining both genetic and functional data, caution must be taken until additional patients are identified to more concretely establish the genotype–phenotype correlations.

Mitochondrial Dysfunction Associated With Common POI

Rare monogenic forms of POI have been instrumental into opening a window into the key genes and pathways that are important for ovarian function (Fig. 1) (122). However, with ∼1% of women affected by POI, it is not a rare disorder, and even cumulatively, the rare monogenic causes of POI fail to account for most cases. Nonetheless, that so many genes involved in mitochondrial function, and mitochondrial mRNA translation in particular, have been implicated as causes of rare syndromic or isolated POI suggests that this organelle may be involved in the more common and genetically and environmentally complex presentations of POI that account for most cases. Toward this end, pilot studies of POI patients with no known etiology have identified an increased frequency of mutations in the mitochondrially encoded ATP synthase 6 gene (MT-ATP6) and the mitochondrially encoded cytochrome c oxidase I gene (MT-CO1); however, in both cases the number of patients screened was small (n = 24 cases and n = 63 cases, respectively) (123, 124). The need for caution in interpreting small studies for a common disease is evident, as the conclusions linking MT-ATP6 variants to POI have been called into question, based on potential ancestry differences between the case and control populations analyzed in that study (125). In addition to genetic mutations in mitochondria, elevated levels of ROS and decreased levels of OXPHOS activity and ATP production in oocytes have also been correlated with POI, although again these studies are limited in size (123, 126–128). Interestingly, treatment of mice with coenzyme Q10, a product of the TCA cycle that controls multiple aspects of mitochondrial biology, including transcription and succinate dehydrogenase activity, prevented the onset of POI in both genetic- and aging-based mouse models of POI (129). Although these preliminary studies are intriguing, future studies with increased power and further studies in model organisms will be needed to determine whether mitochondrial dysfunction is a common contributing factor to the onset of POI.

Figure 1.

Molecular functions of mitochondria-related proteins implicated in POI. A cartoon depicts the various molecular functions in mitochondria of the proteins in which mutations have been identified that can cause POI.

Summary and Future Directions

It is clear from studies of rare monogenic disorders that functional mitochondria are necessary to prevent ovarian dysfunction and infertility. However, larger and better powered studies are necessary to access the impact and causal role of dysfunctional mitochondria in most cases of POI. Nonetheless, we have learned a tremendous amount about the genetics and pathophysiology of POI by studying the rare disorders in cases where a disease-causing mutation has been identified (Fig. 1). These studies, which have identified mutations in genes including MRPS22, HARS2, LARS2, and AARS2, suggest that one putative therapeutic target for treating or preventing POI will be to improve mitochondrial mRNA translation. Proof-of-principle experiments have shown that supplementation with l-cysteine is able to improve OXPHOS activity in cell lines from a patient with a genetic defect that impairs mitochondrial mRNA translation (130). In other cellular studies, overexpression of mitochondrial tRNAs were able to correct function defects in mRNA translation and cellular respiration (131–134). Treatment with acetyl-l-carnitine has also long been known to have beneficial effects on mitochondrial function in vivo, including improved efficiency of mitochondrial mRNA translation, and has been proposed as a treatment of a variety of mitochondrial disorders (135). Treatment with acetyl-l-carnitine improves in vitro maturation of oocytes and reduces the number of abnormal mitochondria, although it is unclear what molecular mechanism underlies these improvements and whether these finding will translate to in vivo treatment or prevention of POI (136–138). These studies have shown that it is possible to manipulate the mitochondrial mRNA translational machinery, although it remains to be seen whether this can be accomplished on a therapeutic basis and what, if any, clinical benefits would be seen.

The discoveries that mutations in the mitochondria-related genes described above and in Table 1 can cause POI were largely made possible by advances in DNA sequencing technologies that greatly reduced the cost of whole-exome sequencing (WES) (139). The initial studies linking mutations in 8 of the 11 genes to POI used WES in the process of identifying the causal variant. However, given the rarity of mutations in these genes as a cause of POI, the current guidelines for diagnosing and treating nonsyndromic POI do not recommend testing for mutations in these genes or other monogenic causes of POI (140). Nonetheless, this may change as WES analyses continue to become more integrated into routine clinical practice. There is mounting evidence from numerous clinical settings demonstrating benefits in terms of both cost reductions and improved clinical care by incorporating WES early into the diagnostic process (141–146). However, to make this a clinical reality for POI, more accurate predictions of whether variants are likely to be pathogenic will be required. Fortunately, progress can be made toward this end with limited sample sizes. The studies utilizing WES discussed above were each based on analyses of individuals with POI from just one to three families, illustrating the progress that can result from even small-scale studies.

Despite all that we have learned, many pressing questions remain unanswered. (i) Why do mutations in the genes described above cause POI, whereas mutations in many more related mitochondrial genes do not? (ii) Why do mutations in genes such as MRPS22 cause a spectrum of disorders, ranging from isolated POI to multiorgan systemic disease, and can we predict the genotype–phenotype correlations? (iii) Why does mitochondrial dysfunction preferentially affect the ovary? (iv) Is mitochondria-associated POI due to cell-autonomous defects in the oocyte, or rather defects in ovarian somatic cells or other systemic changes? (v) Is ovarian dysfunction in rare monogenic cases of POI strictly due to impaired cellular respiration or are there nonbioenergetic defects that underlie the disorder? Answers to these questions will provide insight into the underlying pathophysiology of POI and may identify new therapeutic targets to improve the diagnosis and treatment of POI.

Glossary

Abbreviations:

aaRS

aminoacyl-tRNA synthetase

AARS2

alanyl-tRNA synthetase 2, mitochondrial

CLPP

caseinolytic mitochondrial matrix peptidase proteolytic subunit

CLPX

caseinolytic mitochondrial matrix peptidase chaperone subunit

HARS2

histidyl-tRNA synthetase 2, mitochondrial

LRPPRC

leucine-rich pentatricopeptide repeat containing

MII

meiosis II

MRPS22

mitochondrial ribosomal protein S22

mtDNA

mitochondrial DNA

OXPHOS

oxidative phosphorylation

POI

primary ovarian insufficiency

POLG

DNA polymerase γ, catalytic subunit

PRORP

protein-only RNase P catalytic subunit

ROS

reactive oxygen species

TCA

tricarboxylic acid

TWNK

twinkle mtDNA helicase

WES

whole-exome sequencing

Additional Information

Disclosure Summary: The authors have nothing to disclose.

Data Availability:

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.