Peripheral blood mononuclear cell respiratory function is associated with progressive glaucomatous vision loss

Glaucoma is a chronic, progressive optic neuropathy leading to irreversible vision loss and is projected to affect ~80 million people by decade’s end. Elevated intraocular pressure (IOP) and age are key risk factors, yet many patients, including a high proportion with normal-tension glaucoma (NTG), progress despite IOP within the normal range and despite IOP-lowering therapy. This highlights the need for additional, potentially modifiable risk factors and biomarkers to identify individuals at higher risk of deterioration. Prior studies suggest altered mitochondrial bioenergetics and nicotinamide adenine dinucleotide (NAD) metabolism are linked to glaucoma susceptibility and resistance. However, whether systemic mitochondrial function is associated with glaucoma progression, and its potential predictive value, remains unclear. This study tests the hypotheses that PBMC mitochondrial respiratory function is reduced in glaucoma, relates to VF progression rate, and that PBMC NAD levels are reduced and correlate with mitochondrial function, thereby serving as biomarkers of progressive disease.

Multiple lines of evidence implicate mitochondrial dysfunction in glaucoma: lymphocyte and fibroblast studies show impaired mitochondrial respiration in POAG and NTG; animal models (for example DBA/2J mice) demonstrate age- and IOP-related retinal NAD decline that increases RGC susceptibility; and genetic associations involve mitochondrial pathways (e.g., OPA1, optineurin, TBK1). Nutritional and metabolic links include lower serum nicotinamide in POAG, associations between lower dietary niacin intake and glaucoma (particularly NTG), and preclinical and early clinical evidence that boosting NAD (e.g., high-dose nicotinamide) supports RGC function. These findings motivate evaluating systemic mitochondrial respiratory function and NAD as potential biomarkers for glaucoma progression and as targets for therapy.

Study design: Observational study in accordance with the Declaration of Helsinki with ethics approvals in the UK. Participants were recruited (Oct 2018–Dec 2021) after consent. The study comprised three parts: (1) association of PBMC mitochondrial function with glaucoma diagnosis; (2) association of PBMC mitochondrial OCR with VF progression rates; (3) association of PBMC total NAD with diagnosis and OCR. A separate untreated glaucoma reference cohort (placebo arm of UKGTS) was analyzed to assess the association between IOP and VF progression without treatment confounding. Participants: POAG patients (HTG: pretreatment IOP ≥21 mmHg; NTG: pretreatment IOP ≤21 mmHg) were recruited from glaucoma clinics; age-similar controls from cataract clinics with no glaucoma, no first-degree family history, and IOP ≤21 mmHg. Part 1 included 99 NTG, 69 HTG, and 50 controls. Part 2 VF progression analysis included 229 eyes of 139 patients (144 NTG eyes, 85 HTG eyes) with ≥6 reliable VFs over ≥3 years before any glaucoma surgery; for surgical eyes, VF observation was censored at listing for surgery. At OCR measurement, 93% were on at least one IOP-lowering drop; 50.7% of eyes had undergone IOP-lowering surgery. Part 3 NAD analysis measured PBMC NAD in 54 subjects (25 controls, 10 HTG, 19 NTG). Exclusions included secondary, angle-closure, congenital glaucoma; conditions/treatments affecting lymphocyte or mitochondrial function. Clinical measures: Visual fields were 24-2 SITA Standard on Humphrey Field Analyzer; VF mean deviation (MD) was extracted. IOP was measured by Goldmann applanation tonometry over the VF observation period. Central corneal thickness (CCT) and demographic/systemic data, including comorbidities and medications, were collected. PBMC isolation: Venous blood in EDTA tubes; PBMCs isolated by density gradient using Lymphoprep within 3 hours and used immediately for OCR and NAD measurements. Mitochondrial respiration (OCR): Seahorse XFe24 Mito Stress Test was used with live PBMCs (8×10^4 cells/well; four technical replicates). Seahorse Base medium with glucose, pyruvate, and glutamine was used. Sequential injections: oligomycin (1.5 µM), FCCP (1.5 µM), antimycin A/rotenone (1.6 µM/16 µM). Parameters calculated per manufacturer: basal OCR, ATP-linked OCR, maximal OCR, and reserve capacity, expressed as pmol min−1 per 100,000 cells. Intra-assay CV for basal OCR was 10%. Test-retest repeatability was assessed (median 217 days between tests; 95% limits of agreement approximately −6.5 to +6.2 pmol min−1 per 100,000 cells). PBMC subpopulation assessment: Lymphocyte and monocyte populations were quantified by MoxiGo II flow cytometry and analyzed in FlowJo. ddPCR gene expression for markers (CD56/NCAM1, CD14, CD19, CD8A, CD4) was performed on 30 random samples (10 per group) to assess PBMC subset composition. PBMC total NAD assay: Total NAD (NAD+ + NADH) was quantified using Promega NAD/NADH-Glo luciferase assay, normalized to protein content, and expressed as pg NAD per mg protein. Measurements were performed in a convenience sample (25 controls, 10 HTG, 19 NTG) with at least two technical repeats; median time between OCR and NAD assays was 0 days (IQR 0–123). Intra-assay CV was 9%. Statistics: Group comparisons used ANOVA with Tukey’s HSD post hoc tests (no multiple comparison adjustment for other tests). Associations between OCR metrics and demographic/clinical characteristics used univariable and multivariable linear regression. Candidate covariates included age, sex, diagnosis (control/HTG/NTG), smoking, alcohol, diabetes, systemic hypertension, hypercholesterolemia, migraine, number of systemic illnesses, BMI, vasospasm, vitamin D, thyroid disease, and medications (ACE inhibitors, ARBs, systemic β-blockers, calcium channel blockers, diuretics). LASSO regression selected covariates for final multivariable models; certain known factors (e.g., type 2 diabetes) were included regardless of prevalence if relevant. VF progression modeling used linear mixed-effects models with random intercepts (nested eye within patient) and random slopes for time; dependent variable was MD at each visit. Fixed effects included follow-up time, covariates (e.g., basal or other OCR parameters, mean/peak/SD IOP, CCT, baseline age), and interactions with time to estimate effects on progression rate. Variables were standardized (zero mean, unit SD) to compare effect sizes. Partial R² for a given predictor (e.g., OCR or IOP) was computed as the incremental reduction in SSE of the random effect between full and reduced models. Sensitivity analyses examined surgery vs no-surgery groups, triple interaction OCR×IOP×time, and separate NTG/HTG analyses. A reference analysis in the untreated UKGTS placebo arm used analogous mixed models (covariates: mean IOP, CCT, baseline age) to estimate IOP’s contribution to VF progression without treatment confounding. Two-sided P<0.05 was considered significant. Analyses were conducted in R.

• PBMC mitochondrial respiration is reduced in glaucoma: Basal OCR was lowest in NTG (mean 19.2±4.1 pmol min−1 per 100,000 cells), intermediate in HTG (21.6±3.9), and highest in controls (26.8±5.5); NTG and HTG vs controls P<0.001; NTG vs HTG P<0.01. Maximal OCR, ATP-linked OCR, and reserve capacity showed similar trends.
• PBMC composition did not explain OCR differences: No differences in lymphocyte/monocyte ratio or ddPCR marker expression (CD56, CD14, CD19, CD8A, CD4) among groups; basal OCR remained lower in glaucoma groups (P<0.001) in the subset tested.
• Basal OCR is strongly associated with faster VF progression in treated glaucoma: In 229 eyes of 139 patients, lower basal OCR was associated with faster MD decline; per 1 SD reduction in basal OCR (4.3 pmol min−1/100k cells) the MD rate was 0.19 (SE 0.04) dB/yr worse (P<0.001). Similar associations in NTG (0.18±0.05 dB/yr; P<0.001) and HTG (0.20±0.06 dB/yr; P=0.002). Basal OCR explained 13% of the variance in progression rate (partial R²).
• Age and IOP effects in the treated cohort: Older age was associated with faster progression in the whole cohort (−0.09±0.03 dB/yr per 10 years; P=0.002) and HTG (−0.20±0.05; P<0.001), but not NTG. Higher mean IOP was weakly associated with faster progression in the whole cohort (−0.07±0.03 dB/yr per 1 SD = 2.8 mmHg; P=0.04), explaining 1% of variance (partial R²). A 1 SD difference in basal OCR was equivalent to a 7.6 mmHg IOP difference and a 21-year age difference in this cohort. The OCR×IOP×time interaction was not significant (P=0.97), suggesting additive, not multiplicative, effects across the IOP spectrum.
• Untreated reference cohort (UKGTS placebo arm): Higher mean IOP was strongly associated with faster VF progression (−0.25±0.05 dB/yr per 1 SD; P<0.001), explaining 16% of variance (partial R²). Thicker CCT was associated with slower progression (0.18±0.09 dB/yr; P=0.046); age was not significant.
• PBMC NAD levels are reduced in glaucoma and correlate with OCR: Total NAD levels differed across groups (P<0.001); lower in NTG (0.4±0.2 pg/mg protein) and HTG (0.6±0.3) versus controls (0.9±0.3). Higher NAD levels were associated with higher basal OCR in multivariable models (estimate 11.6±2.1 pmol min−1/100k cells per 1 pg/mg NAD; P<0.001), explaining 16% of basal OCR variance (partial R²). Trends were similar within subgroups but did not reach significance due to smaller sample sizes. NAD was not significantly associated with VF progression rates in the smaller subset analyzed (P=0.25).
• Covariate effects on OCR: In controls, type 2 diabetes was associated with lower maximal OCR (−52.0±20.8 pmol min−1/100k; P=0.016) and reserve capacity (−48.8±19.1; P=0.014). Systemic/topical β-blocker use was not associated with basal OCR. No meaningful differences in PBMC subpopulations accounted for OCR differences.

The study demonstrates that systemic mitochondrial respiratory function, measured as PBMC OCR, is reduced in glaucoma, particularly in NTG, and is strongly associated with the rate of glaucomatous VF deterioration in patients under IOP-lowering treatment. This suggests that mitochondrial dysfunction represents a meaningful biomarker of progression risk, independent of and additive to IOP-related risk. The stronger OCR–progression association compared with IOP within the treated clinical cohort likely reflects clinical management reducing IOP variability and its correlation with progression. In contrast, the untreated UKGTS cohort confirms the substantial effect of IOP on progression when not modulated by therapy. Lower PBMC NAD levels in glaucoma and their strong correlation with OCR link systemic metabolic status to mitochondrial respiratory function, aligning with preclinical and clinical evidence implicating NAD depletion and metabolic vulnerability in RGC injury. The absence of PBMC subset composition differences argues against immune cell mixture as a confounder. Together, findings support a model in which non-IOP factors (e.g., genetic, nutritional, metabolic) influence mitochondrial function and contribute to glaucoma susceptibility and progression, potentially more prominent in NTG where IOP is normal. Clinically, PBMC OCR and NAD may enable risk stratification, personalized monitoring, and inform the development and targeting of mitochondria- or NAD-augmenting therapies.

PBMC mitochondrial respiration is reduced in POAG, most notably in NTG, and lower systemic OCR is strongly and independently associated with faster VF progression in treated glaucoma. PBMC total NAD levels are reduced in glaucoma and correlate with OCR, supporting their roles as systemic biomarkers of disease susceptibility and progression. These findings have practical implications for risk stratification and personalized care, potentially guiding treatment intensity and selection for mitochondria-targeted interventions. Prospective studies, including ongoing randomized trials of high-dose nicotinamide (e.g., NAMinG and TGNT), will clarify causality, generalizability, and the therapeutic utility of targeting NAD metabolism and mitochondrial function in glaucoma.

• VF progression estimates are inherently imprecise, likely attenuating observed associations with OCR and IOP.
• OCR and NAD were measured after the VF observation window in many patients (especially those with prior surgery), although analyses stratified by surgery status and time interval showed similar associations, suggesting relative temporal stability.
• The glaucoma cohort was enriched for patients at higher risk of progression, which may overestimate biomarker effect sizes relative to unselected populations.
• Recruitment and sample size were impacted by the COVID-19 pandemic; controls were recruited later than most glaucoma patients. Test–retest analyses showed no drift in OCR values over ~2.5 years.
• Observational design precludes causal inference; mitochondrial dysfunction and/or NAD depletion may be cause, consequence, or both.
• Potential confounders such as physical activity, depression/anxiety (which can influence mitochondrial function) were not measured; NTG had slightly worse baseline VF loss than HTG, which could correlate with lifestyle differences.
• OCR and NAD assays have moderate repeatability and are currently labor-intensive, limiting immediate clinical implementation until technological improvements enhance practicality and precision.