Article Text

Download PDFPDF

Association between mitochondrial DNA damage and ocular blood flow in patients with glaucoma
  1. Maki Inoue-Yanagimachi1,
  2. Noriko Himori1,
  3. Kota Sato1,2,
  4. Taiki Kokubun1,
  5. Toshifumi Asano1,
  6. Yukihiro Shiga1,3,
  7. Satoru Tsuda1,
  8. Hiroshi Kunikata1,4,
  9. Toru Nakazawa1,2,3,4,5
  1. 1 Department of Ophthalmology, Tohoku University Graduate School of Medicine, Sendai, Japan
  2. 2 Collaborative Program for Ophthalmic Drug Discovery, Tohoku University Graduate School of Medicine, Sendai, Japan
  3. 3 Department of Ophthalmic Imaging and Information Analytics, Tohoku University Graduate School of Medicine, Sendai, Japan
  4. 4 Department of Retinal Disease Control, Tohoku University Graduate School of Medicine, Sendai, Japan
  5. 5 Department of Advanced Ophthalmic Medicine, Tohoku University Graduate School of Medicine, Sendai, Japan
  1. Correspondence to Professor Toru Nakazawa, Department of Ophthalmology, Tohoku University Graduate School of Medicine, Sendai 980-8575, Japan; ntoru{at}


Background/Aims We determined the relationship between tissue mean blur rate (MT) and mitochondrial dysfunction, represented by the mitochondrial/nuclear DNA (mtDNA/nDNA) ratio. We also investigated the usefulness of these biomarkers.

Methods We assessed ocular blood flow in 123 eyes of 123 patients with open-angle glaucoma (OAG) and 37 control eyes of 37 healthy subjects by measuring MT in the optic nerve head with laser speckle flowgraphy. We measured mtDNA and nDNA with PCR, calculated the mtDNA/nDNA ratio and compared this ratio with MT using Spearman’s rank test. We used multiple regression analysis to further investigate the association between MT and glaucoma in the most severe group.

Results The control and the patients with glaucoma had significant differences in the mtDNA/nDNA ratio, circumpapillary retinal nerve fibre layer thickness and MT. There was no significant relationship between the mtDNA/nDNA ratio and MT in patients with OAG overall or the female patients with OAG, but there was a significant relationship between the mtDNA/nDNA ratio and MT, temporal-MT and superior-MT in male patients with severe OAG (r=−0.46, p=0.03; r=−0.51, p=0.02; r=−0.61, p<0.01, respectively). Furthermore, we found that the mtDNA/nDNA ratio was an independent contributor to temporal-MT and superior-MT in these patients (p<0.01 and p=0.03, respectively).

Conclusion We found that there was a significant relationship between the mtDNA/nDNA ratio and MT in male patients with severe OAG, suggesting that the mtDNA/nDNA ratio may be a new biomarker in glaucoma and may help research on the vulnerability of these patients to mitochondrial dysfunction.

  • glaucoma
  • optic nerve
  • experimental & laboratory

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Glaucoma, the second most common cause of blindness worldwide, involves the death of retinal ganglion cells (RGCs) through complicated mechanisms that are not yet fully understood. Intraocular pressure elevation is known to be the most important risk factor for glaucoma and so far is the only treatable factor.1 However, lowering intraocular pressure is not always sufficient to prevent the progression of glaucomatous optic neuropathy,2 suggesting that glaucoma is a multifactorial disease. Non-intraocular pressure risk factors for glaucoma include mitochondrial dysfunction, which a variety of studies have suggested contributes to glaucoma pathogenesis.3 Additionally, ocular blood flow, assessed with laser speckle flowgraphy (LSFG) measurements of mean blur rate, decreases very early in glaucoma.4 These findings suggest that reduced ocular blood flow in glaucoma might be associated with mitochondrial damage.

The mitochondria are known to regulate both metabolic and apoptotic signalling pathways by generating energy in the form of ATP and balancing reactive oxygen species production.5 However, the mitochondria are subject to dysfunction under a variety of conditions, and this dysfunction is associated with many diseases.6 This dysfunction reduces ATP generation to levels causing cell death and increases reactive oxygen species production causing oxidative stress. The RGCs, which have unmyelinated axons, have particularly large numbers of mitochondria and are vulnerable to these changes.7 RGC death caused by mitochondrial dysfunction underlies many diseases that cause blindness.8 Previous studies of mitochondrial dysfunction in diseases such as cancer, diabetes and renal failure have used the mitochondrial/nuclear DNA (mtDNA/nDNA) ratio as a biomarker of mitochondrial damage.5 Damaged mitochondria undergo fission and DNA replication, increasing the quantity of mtDNA.

A wide variety of work has shown that reduced ocular blood flow is implicated in glaucoma. Shiga et al 4 found that the mean blur rate was reduced in eyes with preperimetric (ie, asymptomatic) glaucoma and was associated with a greater risk of progression. Furthermore, the mean blur rate is significantly impaired in patients with autosomal dominant optic atrophy, a disease known to involve mitochondrial dysfunction.9 Taken together, previous findings thus suggest that patients with glaucoma might undergo mitochondrial damage, causing decreased optic nerve head blood flow. Therefore, we investigated the relationship between optic disc microcirculation and the mtDNA/nDNA ratio.

Materials and methods


This was a cross-sectional study of 123 consecutive Japanese patients with newly or previously diagnosed open-angle glaucoma (OAG), including 89 patients with normal-tension glaucoma and 34 patients with primary OAG. The subjects were enrolled at the glaucoma subspecialty clinic of Tohoku University Hospital between April 2013 and August 2016. If both eyes had glaucoma, the eye with the lower mean deviation in standard automated perimetry testing was used in the statistical analysis. Standard automated perimetry was performed with the Swedish interactive threshold algorithm standard strategy of the 24–2 program of the Humphrey Field Analyzer (Carl Zeiss Meditec, Dublin, California). All patients were classified as having early (>−6 dB), moderate (>−12 dB) or severe (from −12 dB to −18 dB) glaucoma, according to mean deviation.

Ophthalmological examinations included best-corrected visual acuity (logarithm of the minimal angle of resolution), fundus photography (3D OCT-2000; Topcon Corporation, Tokyo, Japan), measurement of intraocular pressure and axial length with the Zeiss IOLMaster (Carl Zeiss AG, Oberkochen, Germany), Goldmann applanation tonometry measurement of intraocular pressure, optical coherence tomography (3D OCT-2000) measurement of circumpapillary retinal nerve fibre layer thickness, Humphrey Field Analyzer measurement of mean deviation, and evaluation of the optic disc with a 90-dioptre lens by a glaucoma specialist. Subjects were excluded if they had other ophthalmic conditions, such as angle-closure glaucoma, pigment dispersion glaucoma, exfoliative glaucoma, trauma, any other type of secondary glaucoma, high myopia (below −8 D) or hyperopia (above +3 D).

Control subjects were recruited from glaucoma-free patients who attended a health screening project in the town of Taiwa, Miyagi Prefecture, in Japan. We selected the healthy contralateral eye of each subject for inclusion. If both eyes had cataracts, we selected the eye with better visual acuity.

Measurement of biomarkers of mitochondrial dysfunction

Blood samples were obtained with EDTA tubes and whole-blood genomic DNA was extracted with QIAsymphony (Qiagen, Germany). Quantitative real-time PCR was performed as previously described,10 with minor modifications. We used a 7500 Fast Real-Time PCR System (Applied Biosystems, Foster City, California, USA). TaqMan Fast Advanced Master Mix (Applied Biosystems) was used for quantitative real-time PCR to quantify mtDNA and nDNA levels with the following primers and probes:

  • Human mitochondrial genome NC_012920.




  • Human β2M accession number M17987.




MtDNA and nDNA probes were labelled with FAM and VIC, respectively. The mtDNA and nDNA PCR mix totalled 20 µL, as follows: 1× TaqMan Fast Advanced Master Mix, 150 nM of nDNA primer and 900 nM of mtDNA primer, respectively, 250 nM of TaqMan probe, and 1 ng of genomic DNA extract. The quantitative real-time PCR was performed as follows: 95°C for 20 s, followed by 40 cycles at 95°C for 15 s and 60°C for 1 min. The threshold cycle value was measured and then the mtDNA/nDNA ratio was calculated. We normalised the mtDNA level to the nDNA level, represented by β2M, a ‘housekeeping’ gene expressed at equal levels in all cells.

LSFG measurement

Before LSFG measurement, the pupils of each subject were dilated with 0.5% tropicamide and 0.5% phenylephrine hydrochloride. The details of the underlying principles of LSFG (Softcare, Fukutsu, Japan) have been described in previous reports.

Systolic and diastolic blood pressure and intraocular pressure were measured after the subjects had rested for 5 min in a sitting position in a dark room. Following these measurements, colour mean blur rate maps of the optic nerve head were obtained with the LSFG-NAVI device (Softcare; previously described in detail).11 The LSFG software automatically divides this map into the large vessel and capillary areas, and blood flow parameters are assessed separately for the vessel area (referred to as MV, ‘mean blur rate in vessel area’), the tissue area (referred to as MT, ‘mean blur rate in tissue area’) and the total area of the optic nerve head (MA, ‘overall mean blur rate’). We repeated the LSFG measurements three times in all eyes and excluded the data if there were excessive fixation losses, tracking errors or eye blinks during the measurement. The average of the stable LSFG data was then calculated. All statistical analyses of optic nerve head circulation were based on the average of three separate LSFG measurements.

Statistical analysis

Spearman’s rank correlation test was used to evaluate single correlations between the mtDNA/nDNA ratio and other variables (MT, superior-MT, temporal-MT, inferior-MT and nasal-MT). Mann-Whitney U test and Fisher’s exact test were used to determine the significance of differences between groups. A series of multiple linear regression analyses were performed to determine independent variables affecting MT. All data are expressed as mean±SD. For all tests, p values of less than 0.05 were considered statistically significant. The JMP Pro V.11 software (SAS Institute Japan, Tokyo, Japan) was used to analyse the data.


One hundred and twenty-three Japanese patients with OAG (61 male, 62 female) and 37 normal control subjects (18 male, 19 female) were recruited for this study. The clinical characteristics of the patients with OAG are shown in tables 1 and 2. The control and the patients with OAG had significant differences in the mtDNA/nDNA ratio, circumpapillary retinal nerve fibre layer thickness, MA and MT, both overall and in each quadrant (all: p<0.01). There were no differences in age, blood pressure, axial length or intraocular pressure. There were no significant differences in the incidence of hypertension, diabetes, hyperlipidaemia or a current smoking habit in the control and OAG groups (29.73% vs 32.52%, p=0.75; 5.41% vs 10.57%, p=0.31; 10.81% vs 16.26%, p=0.39; and 16.22% vs 8.94%, p=0.23, respectively). There were significant relationships between the mtDNA/nDNA ratio and MT, temporal-MT and superior-MT in male patients with severe OAG (table 3 and figure 1A–C), although there were no relationships with inferior-MT or nasal-MT (figure 1D–E), nor in patients with mild and moderate OAG (table 3). Moreover, there was no significant relationship between the mtDNA/nDNA ratio and overall MT or MT in any quadrant in female patients with OAG (table 3). We also performed a multiple regression analysis of factors independently contributing to MT in male patients with severe OAG (table 4), which showed that the mtDNA/nDNA ratio was an independent contributor to temporal-MT and superior-MT in these patients.

Figure 1

Relationship between the mtDNA/nDNA ratio and ocular blood flow in each quadrant of the optic nerve head in male patients with severe open-angle glaucoma. Scatterplots showing the relationship between the mtDNA/nDNA ratio and MT overall (A) and in each quadrant (B–E). Correlations were found in A (r=−0.46, p=0.03), B (r=−0.51, p=0.02) and C (r=−0.61, p<0.01). Correlations were not found in D or E. AU, arbitrary unit; MT, mean blur rate in tissue area; mtDNA/nDNA, mitochondrial/nuclear DNA.

Table 1

General demographics of control and patients with OAG

Table 2

Information for LSFG measurements of control and patients with OAG

Table 3

Relationship between the mtDNA/nDNA ratio and ocular blood flow in each quadrant of the ONH and in groups of patients divided by sex

Table 4

Multiple regression analysis of factors independently contributing to MT in male patients with severe OAG


Our main finding was that the mtDNA/nDNA ratio was significantly related to overall MT, temporal-MT and superior-MT in male patients with severe OAG, even though these factors were not related in an overall group of patients or in female patients. Multivariable logistic regression analysis confirmed this finding, showing that the mtDNA/nDNA ratio was a statistically significant indicator of MT in severe patients. Moreover, these severe patients showed no correlation between MT and circumpapillary retinal nerve fibre layer thickness. Taken together, we consider that these findings suggest that mitochondrial damage was a causative factor of low MT in these patients, rather than a secondary effect of other measured factors. This result reinforces previous findings that mitochondrial dysfunction is linked to glaucomatous RGC loss12 and that optic nerve head blood flow is reduced in eyes with glaucoma.4 This is the first study to suggest that mitochondrial damage may cause impaired ocular blood flow in severe glaucoma.

We evaluated male and female subjects separately because of well-known sex-based differences in hormones, such as oestradiol and oestrogen, which may protect against neurodegenerative diseases in women, most likely via activation of the antioxidant defence system.13 Moreover, oestrogen may have a neuroprotective effect against RGC death.14 Furthermore, Nakano et al 15 revealed that the urinary level of 8-hydroxy-2′-deoxyguanosine, a biomarker of oxidative DNA damage, was lower in premenopausal female subjects than in male subjects. This may be due to the loss of iron caused by menstruation, which can cause chronic iron deficiency. Iron can cause oxidative stress by generating hydroxyl radicals. Additionally, oestrogen has many roles in the central nervous system, including the brain and retina. The cholesterol side-chain cleavage enzyme (P450) is present in the retina. This enzyme synthesises pregnenolone (P5) from cholesterol. P5 is the basis of sex hormones and may be involved in the production of neurosteroids.16 The majority of oestradiol in the serum is thought to bind with sex hormone binding proteins and cannot pass through the blood–retinal barrier.17 However, androstenedione is converted to oestradiol by the effects of aromatase in the synthesis of sex hormones, and when injuries occur in the brain, reactive astrocytes are said to manifest aromatase and contribute to the production of oestradiol.18 Nishikawa et al 19 showed that high concentrations of oestradiol are present in the vitreous body of female subjects even after menopause, suggesting that oestradiol might be synthesised locally in the eye. Further multifaceted studies are needed to elucidate sex differences in the eye and identify ocular cells that produce oestradiol.

Interestingly, we found a significant relationship between the mtDNA/nDNA ratio and MT, temporal-MT and superior-MT only in male patients with severe OAG. The finding that only this group showed a significant relationship may be explained by the likelihood that RGC death can be induced by factors other than mitochondrial dysfunction,20 and that these other factors predominate in other groups. Moreover, ischaemia may also affect the RGC mitochondria and glial cells (astrocytes, microglial and Müller cells) in other groups, contributing to glaucoma pathogenesis. In particular, glial cells are activated during ischaemia and release a variety of chemicals, such as tumour necrosis factor-alpha, nitric oxide and glutamate. Thus, individual RGCs, which have differing receptor profiles, that is, with different numbers and types of receptors, might be stimulated at different times by these chemicals. This would cause the influx of calcium that ultimately results in the collapse of mitochondrial function and death to occur at different time points in different RGCs. It is thus likely that RGC death is induced not simply by mitochondrial dysfunction, but that mitochondrial dysfunction is a component of many other phenomena. In addition, the origin of the mitochondria may affect the mtDNA/nDNA ratio. Whole-blood genomic DNA can contain white blood cells and vascular endothelial cells, and mitochondrial dysfunction in these cells may affect the ratio. Fetterman et al 21 showed that damage to the endothelial cells was associated with increased mtDNA damage, and that this may be related to microvascular dysfunction. Our hypothesis is that general mitochondrial dysfunction led to the accumulation of oxidative stress, damaging the endothelial cells in the blood vessels. Mitochondria are the major site of reactive oxygen species. Oxidative stress is thought to be a major contributor to endothelial damage, as it can decrease nitric oxide production, which promotes inflammation and disturbs normal endothelial cell functions. This change results in decreased blood flow in the optic nerve head, affecting the status of glaucoma in patients.

However, the present study only identified general correlations with mtDNA. Therefore, additional work is required to elucidate the specific contribution of mtDNA damage in vascular endothelial cells.

Mitochondria are integral to optic nerve head biology. RGCs have particularly high metabolic demands and a high requirement for ATP, as evidenced by the dense distribution of axonal mitochondria around the optic nerve head.22 Mitochondrial dysfunction renders the RGCs particularly susceptible to programmed cell death and glaucomatous injury, because it decreases the energy available for repair processes. Moreover, the anatomy of the papillomacular bundle further increases its vulnerability. The papillomacular bundle is located in the area leading to the temporal optic nerve head and is composed of RGC axons with relatively small cross-sectional areas, which contain relatively few mitochondria, decreasing the mitochondrial reserve of the papillomacular bundle. Various substances taken orally might reach the RGC mitochondria in sufficient quantities to have a positive therapeutic influence by reducing oxidative stress and/or simultaneously acting as electron carriers to modulate mitochondrial electron flow. Free radical scavengers, such as idebenone and coenzyme Q10, that improve mitochondrial function may be useful as therapies to ameliorate cell death in various neurological disorders. In the case of Leber’s hereditary optic neuropathy (LHON), where mitochondrial complex 1 is affected, patients treated with idebenone appear to benefit from use of the drug.23 Importantly, coenzyme Q10 and its derivatives are able to cross both mitochondrial membranes and the blood–brain barrier. Treatment with coenzyme Q10 may lead to greater ATP production and a beneficial reduction of oxidative stress in the RGCs.24

Our study had several limitations. First, our results showed that the mtDNA/nDNA ratio was not closely related to the presence of glaucoma (area under the curve: 0.69, data not shown), and we therefore cannot conclude that it is a useful biomarker of glaucoma. Moreover, previously, Bianco et al 25 revealed that increased mtDNA may be a protective response to mitochondrial damage in LHON, a hereditary disease, at least in a subset of patients who were unaffected carriers of the disease. Mitochondrial dysfunction may have quite different roles in LHON and glaucoma. Moreover, previous papers have supported the use of the mtDNA ratio as a biomarker in other diseases. Second, we did not evaluate oestrogen levels in patients with OAG. Oestrogen has been reported to have a protective effect in patients with glaucoma.13 Thus, we could not accurately determine the influence of oestrogen on OAG in these patients. Nevertheless, we did not find a significant association in female patients with OAG. Therefore, we consider that our results lend sufficient support to our conclusions. Third, although our finding that temporal-MT was correlated with the mtDNA/nDNA ratio was reasonable, considering that temporal-MT reduction is seen in patients with autosomal dominant optic atrophy,9 we have no findings to explain the relationship between superior-MT and the mtDNA/nDNA ratio. There are no previous reports on this subject, so more investigations are needed to confirm the association between superior-MT and mitochondrial dysfunction. Finally, our study was limited by a lack of certainty as to why there was an especially significant relationship between the mtDNA/nDNA ratio and MT in severe OAG. Previously, we found that patients with severe OAG had higher accumulated systemic oxidative stress. Thus, these patients may be more vulnerable to further oxidative stress caused by mitochondrial damage. Our study is a cross-sectional study; therefore, a prospective, multicentre study with a greater number of patients is needed to confirm our findings.

The most important finding of this study was the significant relationship we observed between the mtDNA/nDNA ratio and the mean blur rate level in patients with glaucoma. This finding should shed new light on the relationship between mitochondrial dysfunction and optic nerve head blood flow in glaucoma, and we hope that it will lead to new therapeutic strategies and improve understanding of the connected risks of glaucoma and mitochondrial dysfunction. Nevertheless, evidence for the association between mitochondrial dysfunction and glaucoma remains inconclusive, and in the future we hope to investigate this association in greater detail, particularly the relationship between the precise level of mitochondrial dysfunction and glaucoma severity.


We thank Mr Tim Hilts for reviewing and editing the language of the manuscript.



  • Contributors MI-Y designed the study; contributed to collection, analysis and interpretation of data; drafted the manuscript. NH made substantial contributions to the conception of the study and analysed the data. KS designed the study; provided technical or material support; contributed to critical revisions. TK, TA, YS and ST collected, assembled and analysed the data. HK contributed to critical revisions for important intellectual content. TN contributed to the conception and design of the study and critical revisions for important intellectual content.

  • Funding This work was supported by a JSPS KAKEN Grant-in-Aid for young scientists (NH: 16K20299) and a JST grant from JSPS KAKENHI Grants-in-Aid for scientific research (B) (TN: 26293372).

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval This study followed the tenets of the Declaration of Helsinki and was approved by the Ethics Committee of the Tohoku University School of Medicine (study 2016-1-708, 2017-1-254, 2016-260).

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles

  • At a glance
    Keith Barton Jost B Jonas James Chodosh