Primary open-angle glaucoma (POAG) is a complex and chronic eye disease characterized by progressive death of retinal ganglion cells (RGCs), which manifests itself initially as visual field loss. Untreated POAG ultimately leads to irreversible blindness (Tham et al., 2014). High intraocular pressure (IOP) is the first and major risk factor identified in patients with POAG (Shaffer, 1996), in addition other known risk factors are advanced age, myopia, ethnicity, and positive family history for POAG (Gramer and Grehn, 2012). Nowadays, the only efficacious therapy for protecting the RCGs (from which the axons form the optic nerve) in POAG is directed towards decreasing IOP. POAG is also defined as a genetically complex disease because many genes have been associated with this condition (Thorleifsson et al., 2010; Burdon et al., 2011; Ramdas et al., 2011; Wiggs et al., 2011; Cao et al., 2012; Osman et al., 2012; Liu et al., 2013; Chen et al., 2014; Gharahkhani et al., 2014; Springelkamp et al., 2015; Trikha et al., 2015; Choquet et al., 2018; Pasquale, 2019). Rearrangements in the DNA and multiple disease genes have been implicated in the pathogenesis of POAG (Davis et al., 2011; Janssen et al., 2013; Liu et al., 2014; Lo Faro et al., 2021). Known disease genes include myocilin (MYOC), optineurin (OPTN), WD repeat domain 36 (WDR36), cytochrome P450 family 1 subfamily B polypeptide 1 (CYP1B1), and TANK-binding kinase 1 (TBK1) (Janssen et al., 2013). So far, more than 120 chromosomal loci have been discovered through genome-wide association studies (Burdon et al., 2011; Ramdas et al., 2011; Choquet et al., 2018; Gharahkhani et al., 2021). Even so, the disease genes and genetic risk factors identified only explain a small proportion of POAG heritability, and contribute relatively little to understanding the pathogenetic mechanisms. To date, the majority of the genetic studies aim to identify the causes underlying POAG by focusing on the nuclear genome. We and other researchers have postulated that at least part of the remaining POAG heritability might be found in variants in the mitochondrial DNA (mtDNA) (Lascaratos et al., 2012; Osiewacz, 2014; Williams et al., 2017).

The mitochondrion is an organelle mainly involved in the production of cellular energy with distinct extrachromosomal circular and double-stranded molecules of DNA. The mtDNA is transmitted through the maternal germline and it has a unique subcellular transcription and replication machinery (Vincent and Picard et al., 2018). Mutations in the mtDNA have previously been implicated in cellular energy deficits that lead to ocular degenerative disease (Yu-Wai-Man et al., 2011). Indeed, the mitochondria can be considered as a “power plant” for the cell, then it is expected that mitochondrial disorders tend to affect more frequently tissues with high energy demand, such as the retina, brain, muscles, heart, and endocrine systems (Wallace, 2010). A characteristic of the mitochondrial genome (mtGenome) is that it accumulates mutations at a notably faster rate than the nuclear genome. As a result, mtDNA is highly polymorphic. Most likely, this characteristic can be explained by two processes: the lack of protective histones and repair mechanisms, which increase the replication errors, and, given the proximity of mtDNA with the respiratory chain complexes, the exposure to reactive oxygen species (ROS) (Wallace, 2010; Lascaratos et al., 2012).

Many mitochondrial single-nucleotide polymorphisms (mtSNPs) have become fixed in a variety of populations during human evolution (Yu-Wai-Man et al., 2011; Collins et al., 2016). Due to the exclusive maternal inheritance of mtDNA and the fact that the mtGenome does not recombine, mtSNPs are accumulating and co-transmitted through the maternal lineages (Andersen and Balding et al., 2018). This characteristic allows tracing specific, ancestral patterns of human migration that occurred millennia ago, from Africa to other continents. These specific polymorphic SNP-sets accumulating on the mtDNA, allowed researchers to classify human populations into various mtDNA ‘‘haplogroups”. There are a total of nine known haplogroups that identify individuals with European ancestry. These are named H, I, J, K, T, U, V, W, and X, where haplogroup H represents about 40–45% of the total. According to recent studies, specific haplogroups can influence the development of diseases such as POAG, primary angle-closure glaucoma, exfoliation glaucoma, as well as cancer, diabetes, and late-onset neurodegenerative conditions (Herrnstadt et al., 2002; Abu-Amero et al., 2008; van Oven and Kayser, 2009; Abu-Amero et al., 2011a; Abu-Amero et al., 2011b; Urzua-Traslavina and Carlos, 2014). The association between mitochondrial haplogroups and POAG has been investigated in few studies, with conflicting results: Andrews et al. (2006), in a case-control comparison of 140 POAG patients and 75 healthy individuals from England, did not find a difference in the haplogroup distribution between cases and controls (Andrews et al., 2006). In contrast, Collins et al. (2016) found that in African populations the haplogroups L1c2, L1c2b, and L2 were risk factors for POAG (Collins et al., 2016).

An important aspect in the pathophysiology of POAG and similar optic neuropathies is represented by the increased apoptosis of RGCs, and by the functional decay of the trabecular meshwork (TM) (Izzotti et al., 2011; Almasieh et al., 2012). RGCs contain a high number of mitochondria related to their high energy demand. They are especially vulnerable to oxidative damage caused by mitochondrial dysfunction (Sanchez et al., 2016). Apart from POAG, several other optic neuropathies show axonal RGC loss correlated with mitochondrial dysfunction (Howell, 1997; Howell, 2003). Two examples are Leber’s Hereditary Optic Neuropathy (LHON; OMIM 535000) and autosomal dominant optic atrophy (DOA; OMIM 165500). While LHON is caused by three well-known pathogenic mtDNA mutations in the MT-ND1, MT-ND4, and MT-ND6 genes, DOA is caused by pathogenic mutations within the nuclear OPA1 gene, which codes for a mitochondrial wall membrane protein (Yu-Wai-Man et al., 2009; Wallace, 2010; Yu-Wai-Man et al., 2011).

Next to mitochondrial damage in the RGCs, Izzotti and coworkers investigated potential mitochondrial damage in the TM, a tissue involved in POAG via its influence on IOP. By comparing mtDNA deletions in TM from glaucoma patients and controls by qRT-PCR, the authors observed that oxidative damage arising from mitochondrial failure plays a role in the functional decay of the TM (Izzotti et al., 2011). Another, independent line of evidence for mitochondrial involvement in POAG has come from investigations of glaucoma-prone mice: retinal levels of nicotinamide adenine dinucleotide (NAD) decreased with age, rendering neurons vulnerable to disease-related insults. Interestingly, the administration of a redox metabolite NAD+ and gene therapy of the expression of Nmnat1, a key NAD + producing enzyme, had a protective effect. At the highest dose tested, 93% of eyes did not develop glaucoma symptomatology, compared to 50% for the control group (Williams et al., 2017).

Given the hypothesis that RGCs degeneration and functional decay of TM in POAG are influenced by mitochondrial dysfunction, we aimed to explore whether POAG is associated with variations in the mtGenome. To this purpose, we conducted two different association analyses further described below. First, we analyzed mtDNA SNPs in order to detect genetic variations potentially associated with the disease. Second, we analyzed the role of major haplogroups.

A total of 2,672 individuals were included. Table 1 shows the demographics of both cohorts.

In this study, we investigated the possible involvement of genetic variation in mitochondria in POAG, by performing an association analysis for mitochondrial SNPs and haplogroups in 721 patients with POAG and 1951 healthy individuals. Based on evidence derived from combined analysis of our datasets, we concluded that two mtSNPs (rs2853496 and rs35788393) are nominal associated with POAG. Our data suggest that the A allele of rs2853496, within the MT-ND4 gene, and the T allele of rs35788393, located in the MT-CYB gene, have a protective effect. With respect to mitochondrial haplogroups, our analyses identified haplogroup K as highly associated with an increased risk of POAG (OR = 5.8; 95% CI = 2.7–13.1; p = 1.2 × 10⁻⁵).

Our findings are consistent with evidence from the literature that suggest a potential role of the mtGenome, and more specifically of the genes MT-ND4 and MTCYB in optic neuropathies or glaucoma (Cortopassi and Arnheim et al., 1990; Votruba et al., 2004; Abu-Amero et al., 2006). The MT-ND4 gene is a protein-coding gene located in the mtDNA, encoding for subunit 4 of complex I (NADH ubiquinone oxidoreductase) (MT-ND4, 2021). The complex I is the first enzyme of the respiratory chain, a vulnerable site to oxidative stress, also involved in cellular functions like apoptosis (Ferguson et al., 1976). SNPs in subunit 4 of MT-ND4 can affect the first step of the electron transport chain. Therefore, these mutations may have an impact on mitochondrial respiratory chain function and could result in an alteration of the cellular energy metabolism.

Genetic variations in the MT-ND4 are implicated in other optic neuropathies. This is the case of LHON, where one of the most prevalent variants that accounts for more than 70% of all cases is the m.11778G > A, located in the MT-ND4 gene (Yu-Wai-Man et al., 2014; Mancuso and Klopstock et al., 2019). LHON is one of the most common inherited optic neuropathies and it is characterized by bilateral optic atrophy and loss of central vision due to loss of RGCs (Sadun, 2002; Yu-Wai-Man et al., 2011). MtDNA mutations associated with LHON have also been described in animal models: mice with a mutation in the mt-Nd4 gene show nerve atrophy and RGCs degeneration. Both conditions are also characteristics of LHON in humans (Divi et al., 2007; Koilkonda and Guy, 2011). In contrast to the mitochondrial mutations identified in LHON cases, in the mitochondrial genome of POAG patients the majority of the mutations were somatic transversions (a replacement of a purine with a pyrimidine, or vice versa), caused by the accumulation of oxidative stress over time (Abu-Amero et al., 2006).

In this study, we also report an association between POAG and the mtDNA variation rs35788393 in the MT-CYB gene. The protein product of this gene is involved in the oxidative phosphorylation system. This system is composed of five complexes where the MT-CYB gene encodes for the cytochrome b of complex III (that catalyzes the transfer of electrons from ubiquinol to cytochrome c), the only one solely encoded by a mitochondrial gene (Chaussenot et al., 2018). Pathogenic mutations in the MT-CYB gene can disrupt the normal activity of the electron transport chain and affect the production of ATP by increasing the production of ROS. This leads to damage of cellular proteins, lipids and nucleic acids via oxidation reactions (West et al., 2011). So far, mutations in MT-CYB have been associated with LHON, retinitis pigmentosa and cataract (Brown et al., 1992; Wibrand et al., 2001; Schuelke et al., 2002; Ronchi et al., 2011). Since the MT-CYB gene is involved in the production of ATP in the electron transport chain, it is pivotal to explore the possible role of this gene in POAG in future studies.

Genetic variations in both MT-CYB and MT-ND4 genes are able to destabilize the so-called mitochondrial super complex: the physical interaction between mitochondrial complex-I and complex-III. The destabilization of this complex leads to the loss of complex I activity, the major entry point for electrons to the respiratory chain (Hudson et al., 2007). In transgenic mice, loss of complex-I activity showed an increase of ROS levels in the RCGs and optic nerve degeneration (Qi et al., 2003). In human, glaucomatous TM cells have been reported to be more sensitive to the inhibition of complex-I (Banerjee et al., 2013). Indeed, damages in complex-I were observed to contribute to the progressive loss of TM cells in POAG patients due to the excessive mitochondrial ROS production and to the attenuation of the mitochondrial membrane. This decrease reduces the ATP synthesis, driving the cells towards apoptosis (He et al., 2008). Another study comparing both POAG and LHON lymphoblasts found an impairment of the complex-I, where functional defects of this complex were milder in POAG than LHON. This is in accordance with the less severe development of the disease in POAG (Van Bergen et al., 2015). However, more comprehensive investigations are still necessary to define the regulatory function of complex-I that in turn might lead to the increase of the oxidative stress and favor the glaucomatous condition.

To add biological context to our study, we also evaluated bioinformatically which genes were the most highly connected with MT-CYB and MT-ND4 in the retina network (Bryan et al., 2018). Genes highly connected indicate that they are more likely to have an effect in gene regulation. By querying the retina network, we implicated the POMGNT1 gene (protein O-mannose beta-1,2-N-acetylglucosaminyltransferase-1). POMGNT1 synthesizes a unique O-mannose sugar chain on α-dystroglycan, the extracellular protein that binds laminin α2 in the extracellular matrix. Reported mutations in the POMGNT1 gene have enlightened its role in four genetic muscular dystrophy disease entities: 1) Walker–Warburg syndrome (OMIM #253280), 2) the muscle-eye-brain disease (OMIM #253280)—for which patients show ocular symptoms as retinal degeneration, optic atrophy and congenital glaucoma—3) congenital muscular dystrophy with mental retardation (OMIM #613151), and 4) retinitis pigmentosa (OMIM #606822) (Pihko et al., 1995; Yoshida et al., 2001; Godfrey et al., 2007; Clement et al., 2008; Wang et al., 2016; Xu et al., 2016). The dystroglycan gene (DAG1) encodes for α-dystroglycan and β-dystroglycan (Holt et al., 2000). In the retina, dystroglycan is highly expressed in Müller glial, rods and cones at the outer plexiform layer, and plays an important role in retinal function and survival (Schmitz et al., 1993; Montanaro et al., 1995; Blank et al., 1999; Jastrow et al., 2006). The DAG1 gene is also linked to Muscular dystrophy-dystroglycanopathy (congenital with brain and eye anomalies, OMIM #616538). Using zebrafish animal models, Gupta et al. demonstrated that dystroglycan deficiency caused abnormal development of ganglion cells, lens, and cornea (Gupta et al., 2011). In addition, in a consanguineous Israeli-Arab family a homozygous loss-of-function mutation in the DAG1 gene was detected in infants with a congenital phenotype consistent with Walker–Warburg syndrome. Ocular features in those infants included bilateral corneal opacity and glaucoma (Riemersma et al., 2015). Taken together, these findings suggest that dystroglycan deficiency is strongly correlated with eye abnormalities, including glaucoma. In our view, further studies on the role of DAG1 and POMGNT1 genes in the pathomechanisms underlying POAG are warranted.

Another part of our current investigation in POAG was focused on the analysis of potentially associated mtDNA haplogroups. In order to interpret our results, it is important to consider the variations reported in the population distributions of mitochondrial haplotypes in comparison with the frequencies identified in our study (Torroni and Wallace, 1994). A study conducted in the Netherlands in 680 individuals randomly selected identified that the most common haplogroup was H (45.3%), followed by haplogroups U (25.6%), T (11.6%), J (10.7%) and K (6.3%) (Chaitanya et al., 2016). These frequencies, when present in our POAG dataset, differed from those reported in the general population. Regarding the haplogroup H, it is important to point out that it shows a complex variation with many sub-lineages. In our study, we used data generated by a SNP-chip array, which is able to detect sites that are polymorphic in populations. Therefore, the differences in frequency reported here for the haplogroup H can be attributed to the absence of sites that allows an accurate classification (Loogväli et al., 2004; Pereira et al., 2005). In our study, haplogroup K was the most significant association with POAG. Considering that haplogroup K occurs approximately in 8% of European and 6% in Dutch individuals, we reported a higher frequency of this haplogroup in our POAG cases (24.6%) compared to controls (6.8%). In line with our findings, a meta-analysis conducted in 3,613 individuals affected by LHON from 159 European pedigrees indicated that the risk of visual loss was higher in carriers of the mitochondrial haplogroup K: individual carriers of haplogroup K were more exposed to experience visual loss, whereas individuals carriers of haplogroup H had a lower risk of visual loss (Hudson et al., 2007). The haplogroup association in our study showed a similar outcome: individuals with haplogroup K had a higher risk to develop POAG compared to individuals belonging to haplogroup H. We also observed that the UK cluster shows the same direction of risk. Interestingly, studies conducted in cybrids showed that haplogroup UK are associated with less levels of mitochondrial protein synthesis and respiratory complex IV activities than cybrids from haplogroup H (Gómez-Durán et al., 2010). Therefore, also supported by previous studies, we speculate on the possible link of the mtGenome in the pathogenesis of POAG (Abu-Amero et al., 2011b; Bosley et al., 2011; Collins et al., 2016; Singh et al., 2018). In contrast, a few studies reported that mitochondrial haplogroups did not contribute to the pathogenesis of POAG. Negative associations were found in POAG cohorts from the north east of England, Saudi Arabia and Ghana (Andrews et al., 2006; Abu-Amero et al., 2008, 2012). Combining our data with those of the literature, there is evidence that mitochondrial haplotypes K plays a role in the pathogenesis of POAG, at least in some populations.

In our study, we identified associations between mitochondrial polymorphisms in the MT-ND4 and MT-CYB genes and POAG, and reported that individuals carrying mtDNA haplogroup K were at the highest risk of developing this eye disease. Our findings support the hypothesis that mitochondria have a role in the pathogenesis of POAG. Nonetheless, further genetic and functional studies are still required to highlight the role of mitochondrial genes, in relation to POAG pathophysiology.

Raul Aguirre-Gamboa, Department of Genetics, University of Groningen, University Medical Center Groningen, Netherlands; Patrick Deelen, Department of Genetics, University of Groningen, University Medical Center Groningen, Netherlands; Lude Franke, Department of Genetics, University of Groningen, University Medical Center Groningen, Netherlands; Jan A. Kuivenhoven, Department of Pediatrics, University of Groningen, University Medical Center Groningen, Netherlands; Esteban A. Lopera Maya, Department of Genetics, University of Groningen, University Medical Center Groningen, Netherlands; Ilja M. Nolte, Department of Epidemiology, University of Groningen, University Medical Center Groningen, Netherlands; Serena Sanna, Department of Pediatrics, University of Groningen, University Medical Center Groningen, Netherlands; Harold Snieder, Department of Epidemiology, University of Groningen, University Medical Center Groningen, Netherlands; Morris A. Swertz, Department of Genetics, University of Groningen, University Medical Center Groningen, Netherlands; Peter M. Visscher, Department of Epidemiology, University of Groningen, University Medical Center Groningen, Netherlands, and Institute for Molecular Bioscience, The University of Queensland, Brisbane, Queensland, Australia; Judith M Vonk, Department of Epidemiology, University of Groningen, University Medical Center Groningen, Netherlands; Cisca Wijmenga, Department of Genetics, University of Groningen, University Medical Center Groningen, Netherlands.