Excitatory and inhibitory neurotransmitter alterations with advancing age and injury in the mouse retina

Full text

Excitatory and inhibitory neurotransmitter alterations with advancing age
and injury in the mouse retina
Katharina C. Bell
a,b,*
, Vicki Chrysostomou
b
, Markus Karlsson
c
, Bryan W. Jones
d
,
Pete A. Williams
e,1
, Jonathan G. Crowston
b,c,1
a
NHMRC Clinical Trial Centre, University of Sydney, 92-94 Parramatta Rd, Camperdown, NSW 2050, Australia
b
Neuroscience and Behavioural Diseases and Eye-ACP, SERI/SNEC, Centre for Vision Research, Duke-NUS Medical School, 8 College Road, 169857, Singapore
c
Save Sight Institute, University of Sydney, Sydney, NSW, Australia
d
John Moran Eye Center, University of Utah School of Medicine, Salt Lake City, UT 84132, United States
e
Department of Clinical Neuroscience, Division of Eye and Vision, St. Erik Eye Hospital, Karolinska Institutet, Stockholm, Sweden
ARTICLE INFO
Keywords:
Retinal ganglion cells
Retina
Metabolomics
Neurotransmitter
Glaucoma
Aging
Intraocular pressure
ABSTRACT
Increasing age and elevated intraocular pressure (IOP) are the two major risk factors for glaucoma, the most
common cause of irreversible blindness worldwide. Accumulating evidence is pointing to metabolic failure
predisposing to neuronal loss with advancing age and IOP injury. Many neurotransmitters are synthesized from
endogenous metabolites and are essential for correct cell to cell signaling along the visual pathways. We per-
formed detailed, small molecule metabolomic proling of the aging mouse retina and further explored the impact
of IOP elevation at different ages. The resultant metabolomic proles showed clear discrimination between
young and middle-aged retinas and these changes are accentuated following eye pressure elevation. Alterations
in glutamate and Gamma-aminobutyric acid (GABA) related metabolites were the most apparent changes with
advancing age with further reductions in GABA and related pathways after IOP elevation. These changes were
further conrmed using immunohistochemistry and patch-clamp electrophysiological recording experiments.
1. Introduction
Glaucoma is the most common cause of irreversible blindness
worldwide. The number of affected individuals is expected to increase to
112 million by 2040, consequent to increased life-expectancy in many
countries (GBD, 2019 Blindness and Vision Impairment Collaborators;
Vision Loss Expert Group of the Global Burden of Disease Study, 2021;
Tham et al., 2014). Glaucoma is manifest by the selective loss of retinal
ganglion cells (RGCs) that transmit the visual message from the retina to
the brain. Advancing age and intraocular pressure (IOP) elevation are
two leading risk factors for developing glaucoma (Quigley and Broman,
2006). Despite IOP lowering, vision loss can still progress and up to 20 %
of patients progress to monocular blindness (Bengtsson et al., 2024). It is
therefore likely that other factors play a role in the pathogenesis of
glaucoma.
There is accumulating evidence that links metabolic failure with loss
of RGCS. This includes disturbed glucose, pyruvate, and Nicotinamide
adenine dinucleotide (NAD) metabolism in several experimental animal
models as well as alterations in the metabolites of human tear and
aqueous humor samples taken from primary open angle glaucoma pa-
tients (Baltan et al., 2010; Casson et al., 2021; Harder et al., 2020; Myer
et al., 2020; Rossi et al., 2019; Tribble et al., 2021; Williams et al., 2017).
In addition to playing an important role in cell energy supply, metabo-
lites are also important constituents of neurotransmitters, such as
glutamate or Gamma-aminobutyric acid (GABA) that are essential for
neurotransmission along the visual pathway.
We have previously demonstrated impaired functional recovery and
increased neuronal loss in response to short term elevation of IOP. A
signicant impairment in recovery and elevated vulnerability to repeat
IOP elevations were already clearly manifest in middle aged mice (12
months of age) compared to young (3 months of age) mice
(Chrysostomou et al., 2024; Crowston et al., 2015; Fry et al., 2018; Kong
et al., 2009). Whereas full functional recovery is seen in young mice (3
months of age) by 7- days after a single IOP elevation, delayed and
* Correspondence to: NHMRC Clinical Trial Centre, University of Sydney, Camperdown NSW 2050, Australia.
E-mail addresses: katharina.bell@sydney.edu.au (K.C. Bell), v.chrysostomou@duke-nus.edu.sg (V. Chrysostomou), erikmarkus.karlsson@sydney.edu.au
(M. Karlsson), u0060967@umail.utah.edu (B.W. Jones), pete.williams@ki.se (P.A. Williams), jonathan.crowston@sydney.edu.au (J.G. Crowston).
1
Contributed equally
Contents lists available at ScienceDirect
Neurobiology of Aging
journal homepage: www.elsevier.com/locate/neuaging.org
https://doi.org/10.1016/j.neurobiolaging.2025.03.004
Received 26 July 2024; Received in revised form 23 January 2025; Accepted 2 March 2025
Neurobiology of Aging 150 (2025) 69–79
Available online 8 March 2025
0197-4580/© 2025 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (
http://creativecommons.org/licenses/by/4.0/ ).

incomplete (to 85 % of baseline) recovery is seen in 12 months of age
mice. Older, 24 months of age mice do not recover function following a
similar IOP challenge (Chrysostomou et al., 2024, 2014; Crowston et al.,
2015). The mechanisms that underpin these dramatic differences in RGC
response are still to be fully elucidated.
To further understand the underlying cellular mechanisms associ-
ated with these differences in the capacity to recover function we con-
ducted a detailed analysis of the retinal metabolome to explore
metabolic changes with age and the overlying impact of IOP elevation.
We additionally explored a publicly available dataset derived from the
DBA/2J mouse, which suffers sustained elevation of IOP and RGC loss.
We hereby present a detailed, selected-small molecule metabolomics
prole of the aging mouse retina along and additionally document the
effect of superimposed injury. These data point to both aging and IOP
elevation having a signicant impact on glutamate and GABA signaling.
2. Materials and methods
2.1. Animal strain and husbandry
All animal procedures conformed to the ARVO (Association for
Research in Vision and Ophthalmology) Statement for the Use of Ani-
mals in Ophthalmic and Vision Research. IOP elevation experiments for
metabolomics and immunohistochemistry were approved by the Sing-
Health Institutional Animal Care and Use Committee (2019/SHS/1534).
Male and female C57BL/6J mice at 3 and 12 months of age were pur-
chased from InVivos Pte Ltd, Singapore, and housed at the SingHealth
Experimental Medicine Centre (Academia, Singapore) in a temperature-
(22 ±1 ◦C), light- (12 h light, 12 h dark) and humidity-controlled
(30–40 %) environment with free access to food and water. Prior to
procedures, animals were anesthetized by intraperitoneal injection of
ketamine:xylazine (80:10 mg/kg). Topical proxymetacaine (0.5 %
Alcaine, Alcon Laboratories) and tropicamide (1 % Mydriacyl, Alcon
Laboratories) were instilled to induce local anesthesia and pupil dilation
respectively. Animal core body temperature was maintained at
37.0 ±0.5◦C on a heated platform throughout experiments. IOP eleva-
tion was performed on both the 3 months of age and 12 months of age
mice (N per group =8). Experiments for electrophysiology were per-
formed at the Florey Institute, Melbourne, under the guidelines of the
Australian Code of Practice for the Care and were approved by The
Florey Animal Ethics Committee (18–112-FINMH). Three months of age
B6.Cg-Tg(Thy1-YFP)HJrs/J mice (Stock No: 003783; The Jackson Lab-
oratory, USA), hereafter termed Thy1-YFP mice, were maintained on a
12-hour light/dark cycle and housed in a PC2 facility with ad libitum
access to food and water (N per group =4–6). Male and female mice
were used for all experiments.
2.2. Intraocular pressure elevation
RGC injury was induced by short-term elevation of IOP, a well-
characterized non-ischemic insult that has been described in detail
(Crowston et al., 2015; Kong et al., 2009). For this, the anterior chamber
of the mouse eye was cannulated with a 50
μ
m borosilicate needle
connected via polyethylene tubing to a syringe mounted on a motorized
pump (PHD Ultra CP; Harvard Apparatus, Massachusetts, USA) and a
pressure transducer (Transpac IV, Abbott Critical Care Systems, Sligo,
Ireland). The syringe and tubing were lled with sterile-ltered endo-
toxin-tested Hanks’ balanced salt solution reservoir (HBSS, JRH Bio-
sciences, Lenexa, KA, USA). HBSS was chosen as the uid in the pressure
column as its concentrations of salts and ions show a close approxima-
tion to aqueous humor. The pump was calibrated to drive uid to
maintain a constant pressure within 1 mmHg of the target. Resting IOP
was recorded immediately after cannulation of the anterior chamber
before IOP was elevated to 50 mmHg for 30 min. The contralateral eye
served as an untreated control. A schematic of the model and experi-
mental design has been depicted in the graphical abstract.
2.3. Small molecule metabolomics
Three days post-IOP injury of the 3 months of age and 12 months of
age mice, the retina from the injured eye as well as the contralateral eye
were immediately extracted in ice cold HBSS, dried and snap frozen.
Tissues were stored at −80 ◦C until shipment on dry ice to the Swedish
Metabolomics Centre. Measurements were performed as described pre-
viously (Tribble et al., 2021). Briey, 200 µL extraction buffer (80/20
v/v methanol: water) including internal standards and 1 tungsten bead
were added to each tube. After shaking (30 Hz for 3 min) and centri-
fugation (+4 ◦C, 14,000 rpm for 10 min), 170µL supernatant evaporated
to dryness and the samples were stored at −80 ◦C until analysis. Small
aliquots of the remaining supernatants were pooled and used to create
quality control (QC) samples. Re-suspension and measurements of the
samples were performed as detailed in Canovai et al. (Canovai et al.,
2023). Before randomized batch analysis, the samples were
re-suspended in 10 +10 µL methanol and elution solvent A. Each batch
of samples was rst analysed in positive mode and then the instrument
was switched to negative mode and a second injection of each sample
was performed. Chromatographic separation was performed on an
Agilent 1290 Innity UHPLC-system (Agilent Technologies, Waldbronn,
Germany). 2
μ
L of each sample were injected onto an Atlantis Premier
BEH-Z-HILIC VanGuard FIT (1.7 µm, 2.1 x 50 mm) column (Waters
Corporation, Milford, MA, USA) held at 40 ◦C. The HILIC gradient
elution solvents were A) H2O, 10 mM ammonium formate, 5µM
Medronic acid, pH 9, B) 90:10 Acetonitrile: H2O, 10 mM ammonium
formate, pH 9 and 5 µM Medronic acid. Chromatographic separation
was achieved using a linear gradient (ow rate 0.4 mL/min): min
0 =90 % B, min 6 =80 % B; min 9.5 =20 % B, min 11 =90 % B. The
ow rate was increased to 0.7 mL/min for 2 min, held at this rate for
0.5 min, and then reduced to 0.4 mL/min for 0.5 min before the next
injection. The compounds were detected with an Agilent 6546 Q-TOF
mass spectrometer equipped with a jet stream electrospray ion source
operating in positive or negative ion mode. The reference ions purine (4
μ
M) and HP-0921 (1
μ
M) were infused directly into the MS at a ow rate
of 0.05 mL min-1 for internal calibration. Gas temperature was set to
150◦C, the drying gas ow to 8 L min-1 and the nebulizer pressure
35 psi. Sheath gas temp was set to 350◦C and the sheath gas ow 11 L
min-1. The capillary voltage was set to 4000 V in both positive and
negative ion mode. The nozzle voltage was 300 V. The fragmentor
voltage was 120 V, the skimmer 65 V and the OCT 1 RF Vpp 750 V. The
collision energy was set to 0 V. The m/z range was 70–1700, and data
was collected in centroid mode with an acquisition rate of 4 scans s-1.
MSMS analysis was run on the QC samples for identication purposes.
All data pre-processing was performed using the Agilent MassHunter
Pronder version B.10.0 SP1 (Agilent Technologies Inc., Santa Clara,
CA, USA). The data pre-processing was performed in a targeted fashion.
A pre-dened list of metabolites commonly found in plasma and serum
were searched for using the Batch Targeted feature extraction in Mass-
Hunter Pronder. An-in-house LC-MS library built up by authentic
standards run on the same system with the same chromatographic and
mass-spec settings, were used for the targeted processing. One hundred
low molecular weight metabolites that could be certied with standards
were detected. The quantication of the metabolites was calculated as
area under the curve of the mass spectrometry peak and normalized to
an internal standard for negative and positive runs. Data were analysed
and graphs were made using MetaboAnalyst [version 6.0;] and R (Xia
et al., 2009; Xia and Wishart, 2011). All data were normalised by the
sample median and subjected to auto scaling (van den Berg et al., 2006).
Principal component analysis (PCA) was performed in with Metab-
oanalyst. Comparisons between groups were analysed by two-sample
t-tests with an adjusted p value using a cutoff of 0.05 considered sig-
nicant. Over Representation Analysis (ORA) was implemented using
the hypergeometric test available with Metaboanalyst to evaluate
whether a particular metabolite set is represented more than expected
by chance within the given compound list. One-tailed p values are
K.C. Bell et al.
Neurobiology of Aging 150 (2025) 69–79
70

provided after adjusting for multiple testing. Receiver operating curves
were calculated with the standard settings in Metaboanalyst (linear
support vector machines (SVM)). Quantitative pathway analysis was
performed using the Mus musculus KEGG library in MetaboAnalyst and a
background metabolome of all detected metabolites from the same
samples.
2.4. Immunohistochemistry
GABA immunoreactivity of the control and injured retina of 3
months of age mice at day 3 after IOP elevation was performed as pre-
viously described in detail (Marc et al., 1995) (N =4/group). Briey,
day 3 after IOP injury both eyes were removed and rapidly xed in 1 %
paraformaldehyde, 2.5 % glutaraldehyde, 3 % sucrose, 0.01 % CaCl2, in
0.1 M phosphate buffer, pH 7.4. Eyes were dehydrated and embedded in
epoxy resin, followed by sectioning at 250 nm onto 12-spot
Teon-coated slides. Samples were probed with IgGs targeting GABA
(anti-GABA IgG, AB_2532061, Signature Immunologics, 1:50) and
cellular Retinaldehyde binding protein 1 (CRALBP) (anti-CRALBP IgG,
AB, AB_2314227, 1:400; gift of Dr. John Saari). Primary IgG for GABA
and CRALBP was detected with goat anti-rabbit IgGs adsorbed to 1 nm
gold particles and visualized with silver intensication. Light micro-
scopic images were captured as 12-bit 1936 pixel ×1456 line frames
under constant ux light with feedback regulation and xed CCD cam-
era gain and gamma, as described previously (Marc and Jones, 2002).
Analysis of the GABA signal was performed in a masked fashion using
the manual cell counter in ImageJ. From each retinal section all GABA
positive cells were subdivided to being either GABA-hi or GABA-lo. All
cells were counted, and the length of the retinal section was measured.
The number of either GABA-hi or GABA-lo cells/182µm (1000 pixels)
were calculated and a Kruskal-Wallis test was performed with GraphPad
Prism (Version 10.2.1) with an alpha of 0.05 to determine if there were
any statistically signicant measurements.
3. Whole-cell patch-clamp electrophysiology
3.1. Retinal preparation
Eyes were removed and transferred into cold dissecting solution
(125 mM Choline Cl, 2.5 mM KCl, 0.4 mM CaCl2, 6 mM MgCl2, 1.25 mM
NaH2PO4, 26 mM NaHCO3 and 20 mM D-glucose) and saturated with
95 % O2/5 % CO2. The cornea was immediately punctured with a 23 G
needle to allow oxygenation. The retina was dissected out, hemisected
and the vitreous removed. Retinal hemisections (A1420; Sigma Aldrich,
USA) saturated with 95 % O2/5 % CO2 for one hour at room
temperature.
3.2. Whole-cell patch-clamp electrophysiology
Retinal hemisections were transferred to a recording chamber with
the RGC layer facing up and set in place with a harp. The recoding
chamber were placed on an upright microscope (Slicescope Pro 1000;
Scientica, UK) and perfused with AMES medium (22.6 mM NaHCO3)
saturated with 95 % O2/5 % CO2 at room temperature and at a rate of
2 mL/min. RGCs and bipolar cells were identied visually using a
uorescence lamp under a 40x water-immersion objective and a CCD
camera.
3.3. Electrodes
Glass electrodes (BF150–110–10 and BF150–86–7.5HP; Sutter In-
struments, USA) were pulled using a Flaming Brown micropipette puller
(Model P-1000, Sutter instruments, USA) to produce a resistance of
0.5–1 MΩ (stimulating electrodes) and 4–7 MΩ (recoding electrodes.
Silver wire was wrapped around the glass electrode and connected via
an insulated wire to a stimulating box. Glass electrodes were painted in
silver-chloride paint. Stimulating electrodes were lled with AMES
media (22.6 mM NaHCO3), fastened in a microelectrode holder con-
taining a silver chloride coated electrode. Positive pressure was added to
the stimulating electrode, which was then placed within the bipolar cell
layer. Recording electrodes were placed within the ganglion cell layer
above the stimulating electrode. Synaptic currents were recorded with
electrodes containing: 135 mM CsMeSO4, 8 mM NaCl, 10 mM HEPES, 2
Mg2ATP, 0. mM 3 Na3GTP, 7 Phosphocreatine, 10 EGTA (pH 7.3).
3.4. Inhibitory postsynaptic recordings (IPSC)
Patch-clamp recordings were made in current clamp mode using
PatchStar micromanipulators (Scientica, UK) and Acon Multiclamp
700B patch-clamp amplier (Molecular Devices, USA). Data were ac-
quired using pClamp software (v10; Molecule Devices, USA) with a
sampling date of 50 kHz and low pass bessel ltered at 10 kHz (Digidata,
1440a; Axon).
IPSCs were evoked using a single 0.5 mA, 1 ms pulse repeated 10
times every 10 s (Digital Isolator, Model BIN8–9V; Getting Instruments,
USA). Spontaneous inhibitory postsynaptic currents (sIPSCs) were
recorded using a 2-minute-long gap-free protocol. Evoked inhibitory
postsynaptic currents (eIPSCs) and sIPSCs were recorded by holding the
retinal ganglion cells at +10 mV172.
3.5. Data analysis
eIPSC recordings were analysed using AxoGraph software (Berkeley,
USA). Cells with an access resistance of <40 MΩ were analysed. All 10
recordings had their baseline subtracted from 0–250 ms. The peak eIPSC
amplitude was detected from 500–700 ms and the peak shape was
analysed to determine eIPSC onset, rise time and decay time. eIPSC
amplitude, onset, rise time and decay time were averaged for all 10
recordings per cell.
sIPSCs were analysed using AxoGraph software (Berkeley, USA).
Recordings were rst divided into episodic chunks of 12 s. Traces were
then be ltered using a sharp low pass lter at 5 kHz. A template
function for sIPSCs was be dened with a baseline of 10 ms, template
length of 20 ms, amplitude of 30 pA, rise time of 4 ms and decay time of
10 ms. Events were detected at a threshold of 2 x standard deviation.
Cumulative frequency was extracted for the detected events and saved.
All events with an amplitude over −100 pA were removed.
3.6. Statistical analysis
All data were imported into Prism (GraphPad Software, USA) for
statistical analysis. Outliers were identied using the ROUT outlier’s
method to identify up to 2 outliers (Q =1 %) and excluded from the
analysis. One-way ANOVA (alpha =0.05) and subsequent Tukey’s
multiple comparisons test was used to compare the mean of each group
with the mean of the control group. Descriptive statistics were reported
as (mean ±standard deviation, number of values).
4. Additional IOP injury dataset
Short term IOP elevation results in a reversible loss of the pSTR
changes with limited RGC loss (Chrysostomou and Crowston, 2013;
Crowston et al., 2015). To understand in greater detail the effect of IOP
elevation that precedes RGC cell death, we also analyzed a publicly
available metabolomics dataset using the DBA/2J mouse model (Harder
et al., 2020). At timepoint of the measurement of retina from C57BL/6J
(B6), DBA/2J (D2; glaucoma), and DBA/2J-Gpnmb
R150X
(D2-Gpnmb+;
control) mice were 9 months of age. DBA/2J mice at this timepoint had
developed ongoing elevated IOP, however without RGC death, allowing
for the analysis of pre-degenerative metabolomic changes in the retina.
Measurements of 9 D2 mice and 10 D2-Gpnmb+mice were included
from the publicly available dataset. Detailed methods are described in
K.C. Bell et al.
Neurobiology of Aging 150 (2025) 69–79
71

the original study (Harder et al., 2020). Analysis and statistics were
performed with Metaboanalyst [version 6.0] as described above.
5. Results
5.1. Advancing age alters the metabolomic prole of the mouse retina
One hundred low molecular weight metabolites were detected with
high accuracy and measured in both the non-injured 3 months of age
retina and 12 months of age non-injured retina (raw data, processed
and annotated data is available from Zenodo, doi: 10.5281/zen-
odo.14723294). Of these, 14 metabolites were signicantly increased
in 12 months of age retinas and 3 metabolites were signicantly
decreased in comparison to retinas from young mice (Fig. 1A). Some of
these metabolites, such as indoxyl sulfate and hippuric acid are known
to increase with age in other human and rodent tissues and serum
(Schnackenberg et al., 2007; Tan et al., 2023; Wyczalkowska-Tomasik
et al., 2017), and 3-methylhistidine has previously been demonstrated
to be increased in the aging retina (Wang et al., 2018) further validating
our approach. Pathways known to be involved in the aging process such
as the nicotinate and nicotinamide metabolism (Fig. 1B) (Covarrubias
et al., 2021) were enriched in the KEGG enrichment analysis as well as
both the degradation and biosynthesis pathways of the branch chain
amino acids valine, leucine, and isoleucine. These amino acids are
involved in a multitude of metabolic pathways as well as the regulation
of glutamate levels in the CNS (Zhang et al., 2023). Of the various
exploratory Receiver Operating Characteristic (ROC) curves calculated,
model 3 using 10 features for discrimination resulted in the highest
predictive accuracy of 95.3 % (Suppl. Fig. 1A and B) and showed high
specicity and sensitivity of the measured metabolites with aging (area
under the curve =0.987) (Fig 1). This also resulted in high predictability
of middle-aged versus young based on the metabolic signature (Fig. 1D).
Fig. 1. Metabolic signatures with aging. A: Whole retina samples of non-injured retina from 3- and 12 months of age mice underwent metabolic proling. A total of
100 metabolites were identied. Volcano plot indicating the number and the increase/decrease of signicantly changed metabolites in 12 months of age non-injured
retinae in comparison to 3 months of age non-injured retinae (p-value <0.05, red =increased, blue =decreased). Of the 100 measured metabolites 14 were
increased and 3 were decreased in retina from 12 months of age mice in comparison to retina from 3 months of age animals. N =8 mice/group. B: Summary Plot for
Kyoto Encyclopedia of Genes and Genomes (KEGG) Over Representation Analysis (ORA) metabolites demonstrates signicant pathway enrichment of pathways
related to Lysine degradation and valine, leucine and isoleucine degradation and biosynthesis C: The calculated exploratory Receiver Operating Characteristic Curve
(ROC) calculated with 10 features (model 3) shows high specicity and sensitivity of the measured metabolites with aging (Area under the curve =0.987). The 95
percent condence interval is shown as a band around the ROC curve. D: Plot of predicted class probabilities for all samples using a single biomarker model. The
classication boundary is at the center (x =0.5, dotted line). E Plot of the most important features of a selected model ranked from most to least important. 2 of the
most relevant metabolites discriminating between old and young are relevant for intraretinal synaptic signaling.
K.C. Bell et al.
Neurobiology of Aging 150 (2025) 69–79
72

Calculation of the most relevant metabolites in the biomarker model for
discrimination between middle-aged and young mice were metabolites
involved in intraretinal synaptic signaling such as sarcosine and L-lysine
(Fig. 1E).
5.2. Effect of IOP injury on the retinal metabolome in 3 months of age
mice
Metabolite proles demonstrated clear differences among IOP-
injured and non-injured retina. These differences are shown in the
principal component analysis and ROC analysis (Fig. 2A and B). Multiple
ROC curves were calculated and based on the predictive accuracies with
different features (Suppl. Fig. 2A, B and C), we chose model number 3
that used 10 features for discrimination between the samples for the PCA
and ROC curve (Fig. 2A and B). Out of the 100 measured metabolites, 18
were up-regulated and 18 were down-regulated in retina of young ani-
mals 3 days after IOP injury (Fig. 2C) and KEGG pathway analysis
(Fig. 2D) revealed enrichment of pathways related to retinal neuro-
transmission including arginine and proline metabolism, glycine, serine
and threonine metabolism and alanine, aspartate and glutamate meta-
bolism. Guanidoacetic acid and 4-guanidobutanoic acid were among the
most signicantly up-regulated metabolites in the injured and also
among the most relevant metabolites to discriminate between injured
and non-injured retina (Guanidoacetic acid: Log2(FC)=1.16, -Log10(p-
value) =7.0; 4-guanidobutanoic acid: Log2(FC)=1.04, -Log10(p-value)
=5.6) (Fig. 2A, E).
5.3. Effect of IOP injury on the retinal metabolome of 12 months of age
mice
Differences in the metabolome of non-injured and injured retina
were less obvious in the middle-aged mice. The calculated principal
component analysis was based on model 3 (Suppl. Fig. 3A and B) and
shows no or minimal discrimination between the injured middle-aged
Fig. 2. Metabolic signatures in retina of 3 months of age mice after IOP injury. A: The calculated principal component analysis (PCA) shows good discrimination
between injured and non-injured retina in the 3 months of age animals (squares represent the injured 3- months of age retina, dots represent the non-injured 3 months
of age retina). N =8 mice/group. B: The calculated exploratory Receiver Operating Characteristic Curve (ROC) calculated with 10 features (model 3) shows high
specicity and sensitivity of the measured retinal metabolites with injury in 3 months of age mice (Area under the curve =1). The 95 percent condence interval is
shown as a band around the ROC curve. C: Whole retina samples of injured (single IOP challenge) and non-injured retina from 3 months of age mice underwent
metabolic proling. Of the total of 100 identied metabolites, 18 were up-regulated and 18 were down-regulated in retinae of young animals 3 days after a single IOP
elevation injury, as demonstrated in the volcano plot (FDR <0.05, red =increased, blue =decreased). D: Kyoto Encyclopedia of Genes and Genomes (KEGG)
pathway analysis of metabolites demonstrates signicant pathway enrichment of pathways related to neurotransmission such as Arginine and proline metabolism,
Glycine, serine and threonine metabolism and Alanine, aspartate and glutamate metabolism. Another enriched pathway is the taurine and hypotaurine metabolism.
E: Plot of the most important features of a selected model (Model 3) ranked from most to least important. Guanidoacetic acid is one main metabolite in this model to
discriminate between injured and uninjured retina and is increased in injured retina after IOP a single IOP in 3 months of age animals.
K.C. Bell et al.
Neurobiology of Aging 150 (2025) 69–79
73

retinal samples versus the non-injured samples (Fig. 3A and suppl.
Fig. 3C). This is also reected in the calculated ROC curve (Fig. 3B). Of
the 100 measured metabolites, 8 were up-regulated and 11 were down-
regulated in retina of 12 months of age animals 3 days after a single IOP
elevation injury (Fig. 3C) and KEGG pathway analysis (Fig. 3D) of these
metabolites again revealed enrichment of pathways related to retinal
neurotransmission such as arginine and proline metabolism, glycine,
serine and threonine metabolism and phenylalanine, tyrosine and
tryptophan biosynthesis. Cytidine and succinic acid are among the me-
tabolites that were most frequently used to discriminate the injured
samples from the non-injured samples (Fig. 3E). Cytidine plays a role in
neuronal-glial glutamate cycling and can decrease glutamate and
glutamine levels in the CNS and thus glutamate induced excitotoxicity
(Mir et al., 2003; Yoon et al., 2009) and succinic acid is involved in the
GABA shunt, which connects the tricarboxylic acid (TCA) cycle to GABA
and glutamate metabolism (Schousboe et al., 2013).
5.4. GABA decreased after IOP elevation in young and middle-aged mice
Of the 36 metabolites differently regulated in the retina of 3 months
of age mice after injury and the 19 metabolites differently regulated in
the 12 months of age injured retina, 13 metabolites were detected in
both the middle-aged and the young retina after injury (Fig. 4A). All
these metabolites showed the same direction of change in both age
groups after a single short-term IOP elevation (Fig. 4A). GABA, the main
inhibitory neurotransmitter in the retina was decreased after injury in
both the middle-aged and the young mice and in addition, shows a
general decrease in signal intensity with aging in the non-injured retina
(Fig. 4B). Taurine, which is known to promote neuroprotection of RGCs
after injury (Froger et al., 2012), was also decreased after injury in both
the middle-aged and the young (Fig. 4B). Further investigation of
GABA-positive cells in the retina was performed with immunohisto-
chemistry. As expected, diffuse GABA staining was localized in the IPL
and GABA staining of cell bodies was found in the INL and GCL of the
retina. As CRALBP stains Mueller cells and the retinal pigment epithe-
lium (RPE), it is shown here to allow the reader to localize the GABA
Fig. 3. Metabolic signatures in retina of 12 months of age mice after IOP injury. A: The calculated principal component analysis (PCA) fails to demonstrate
discrimination between injured and non-injured retina in the 12 months of age animals (squares represent the injured 12 months of age retina, dots represent the non-
injured 12 months of age retina). N =8 mice/group. B: The calculated exploratory Receiver Operating Characteristic Curve (ROC) calculated with 10 features (model
3) shows lower specicity and sensitivity of the measured metabolites in 12 months of age retina after IOP elevation injury (Area under the curve =0.811). The 95
percent condence interval is shown as a band around the ROC curve. C: Whole retina samples of injured (single IOP challenge) and non-injured retina from 12
months of age mice underwent metabolic proling. Of the total of 100 identied metabolites, 8 were up-regulated and 11 were down-regulated in retinae of old
animals 3 days after a single IOP elevation injury, as demonstrated in the volcano plot (FDR <0.05, red =increased, blue =decreased). D: Kyoto Encyclopedia of
Genes and Genomes (KEGG) pathway analysis of metabolites demonstrates signicant pathway enrichment related to neurotransmission and pathways related to
Phenylalanine, tyrosine and tryptophan biosynthesis as well as taurine and hypotaurine metabolism. E Plot of the most important features of a selected model (Model
3) ranked from most to least important.
K.C. Bell et al.
Neurobiology of Aging 150 (2025) 69–79
74

Fig. 4. Comparison of injury response in both young and middle-aged mice. A: 36 metabolites were signicantly differently regulated in the retina of 3 months of age
(young) mice after injury in comparison to the non-injured 3 months of age retina, whereas 19 metabolites were signicantly differently regulated in the 12 months
of age (old) injured retina in comparison to the 12 months of age non-injured control. Of these metabolites, 13 metabolites were found to be differently regulated in
both the old and the young retina after injury. All these metabolites showed the same mode of change (increased or decreased levels) in the 12 months of age (black)
and the 3 months of age (orange) retina after a single short term IOP elevation injury. N =8 mice/group. B shows the signal intensity of GABA and Taurine, which are
both found in signicantly lower levels after IOP injury in the retina of 3 months of age and 12 months of age mice Violin plots show the non-normalized signal
intensities. The black bar shows the median, the dotted lines show the 1st and the 3rd quartile. C shows localization of GABA immunoreactivity within the retina.
CRALBP staining (green) highlights especially Mueller cells and the retinal pigment epithelium and allows to identify the different retinal layers. GABA immuno-
reactivity (purple) is seen in the IPL, INL and RGC layer. Exemplary grey scale images of GABA immunoreactivity in the CTRL and at Day 3 after IOP elevation are
shown on the right. Arrows with small arrowheads point towards GABA-hi cells and arrows with large arrowheads point towards GABA-lo cells. Scale bar repre-
sents182µm D shows the GABA-hi and the GABA-lo cell numbers/182µm in the CTRL and Day-3 post-IOP elevation retina (D3) (* p <0.05). N =3 mice/group. E
Example traces of eIPSC recordings from 3 months of age mice. eIPSC traces from uninjured controls and 3-days post IOP elevation. Scale =200 ms x 100 pA.
Quantication of the eIPSC amplitude demonstrates a signicant reduction in IPSC amplitude 3-days post IOP elevation compared to controls (* p <0.05). F
Quantication of the eIPSC decay time demonstrates a signicantly faster IPSC decay 3-days post IOP elevation compared to controls (* p <0.05). G Example traces
of spontaneous IPSC recordings from retinal ganglion cells in 3 months of age mice. sIPSCs were recorded 3-days post IOP elevation as well as from uninjured
controls. Scale =10 ms x 100 pA. Quantication of sIPSC amplitude demonstrates no signicant change following IOP elevation (p >0.05). Quantication of sIPSC
frequency demonstrates a signicant decrease in IPSC frequency 3-days post IOP elevation compared to controls (** p <0.01). N =4–6 mice/group.
K.C. Bell et al.
Neurobiology of Aging 150 (2025) 69–79
75

staining within the retina (Fig. 4C left). In both the non-injured control
retina and at day 3 after IOP elevation, cells containing high (hi)
amounts of GABA (darker cells) and cells containing lower (lo) amounts
of GABA were present (Fig. 4C right). GABA-hi and GABA-lo cells were
counted in a masked fashion and calculated as cells per 182µm. Retinas
at day 3 after IOP elevation injury shows a signicant 40 % decrease in
GABA-hi cells (CTRL mean value 4.6, SD 0.36, D3 mean value 2.8, SD
0.9; p-value <0.05) and a trend of decrease in the GABA-lo cells (CTRL
mean value 5.8, SD 1.6, D3 mean value 4.5, SD 1.2) (Fig. 4D). GABA is
the main inhibitory neurotransmitter in the retina. To study inhibitory
synapse function in the retina, we therefore performed whole-cell
patch-clamp electrophysiology to measure the evoked and sponta-
neous inhibitory potentials in the non-injured control retina as well as
retina at day 3 after IOP elevation in 3 months of age mice. eIPSC
(evoked inhibitory postsynaptic current) amplitude decreased signi-
cantly 3-days post IOP elevation compared to uninjured controls
(P =0.029; control 198.3 ±149.5 pA, n =9 cells; 3-days post IOP
elevation 55.25.1 ±24.04 pA, n =7 cells) (Fig. 4E). As eIPSC amplitude
was reduced following IOP elevation in 3 months of age mice, we also
investigated the effects of IOP elevation on eIPSC kinetics. eIPSC onset
(control 539.4 ±2.95 ms, n =9 cells; 3-days post IOP elevation 543.8
±6.326 ms, n =7 cells) and rise time (control 3.39 ±2.32 ms, n =7
cells; 3-days post IOP elevation 3.13 ±1.69 ms, n =7 cells) were both
unchanged at day 3 after IOP elevation. However, eIPSC decay was
signicantly faster day 3 after IOP elevation in comparison to the
non-injured controls (control 16.91 ±14.22 ms, n =9 cells; 7-days post
Fig. 5. Metabolic changes in other IOP models prior to RGC neurodegeneration. A: Whole retina samples of 9 months of age DBA/2J mice displaying elevated IOP
without obvious RGC or axonal loss and 9 months of age Gpnmb+control mice (CTRL) underwent metabolic proling. In total 86 metabolites were detected. Of
these, 28 metabolites were found in signicantly higher levels in the retina DBA/2J mice (D2) and 14 metabolites were found with signicantly lower levels in the
retina of DBA/2J mice in comparison to the CTRLs, as demonstrated in the volcano plot (p-value <0.05, red =increased, blue =decreased). 44 of the measured
metabolites were not signicantly altered. B: Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis of metabolites demonstrates signicant pathway
enrichment e.g. of pathways related to Purine metabolism, Nicotinate and nicotinamide metabolism, taurine and hypotaurine metabolism as well as Arginine
metabolism. C: Venn diagram comparing signicantly changed metabolites found in 12 months of age injured retina in comparison to 12 months of age non-injured
retina, 3 months of age injured retina in comparison to 3 months of age non-injured retina and the metabolites signicantly differently regulated in D2 mice vs CTRL.
Three metabolites were detected in all 3 injury groups and are shown in D, E and F. D: Guanosine is found in decreased levels in all 3 groups with IOP elevation. E:
Uridine is found in decreased levels after an acute IOP and chronic elevation and F: Proline was increased in the retinae of both young and old mice after single IOP
injury and after chronic IOP increase. Violin plots show the non-normalized signal intensities. The black bar shows the median, the dotted lines show the 1st and the
3rd quartile.
K.C. Bell et al.
Neurobiology of Aging 150 (2025) 69–79
76

IOP elevation 3.29 ±3.52 ms, n =7 cells, p-value <0.05) (Fig. 4F).
sIPSCs were also measured in the non-injured control retina as well as
day 3 post IOP elevation. Whereas the amplitude remained stable be-
tween the two groups (control 41.92 ±8.73 pA, n =4 cells; 3-days post
IOP elevation 33.26 ±2.48 pA, n =4 cells, p-value=0.23), a signicant
decrease in sIPSC frequency was recoded at day 3 after IOP elevation
(control 12.89 ±4.47 Hz, n =4 cells; 3-days post IOP elevation 3.096
±1.44 Hz, n =4 cells; p-value <0.01) (Fig. 4G).
5.5. Comparison of metabolic prole in DBA/2J mice vs. short-term
experimental IOP elevation
The DBA/2J (D2) mouse develops chronic IOP elevation with age
due to a progressive iris disease. We assessed whether there are key
differences in the retinal metabolome of whole retina samples of 9
months of age D2 mice, displaying elevated IOP without obvious RGC or
axonal loss, and 9 months of age D2-Gpnmb+control mice (dataset can
be found in the source publication (Harder et al., 2020)). Of the 86
metabolites, 28 were signicantly higher in D2 mice and 14 metabolites
signicantly lower (Fig. 5A). 20 metabolites were found in both the
single-IOP elevation datasets and the D2 dataset but were not signi-
cantly altered in all three. Examples are glutamine, which was increased
with IOP elevation in the D2 mice as well as the 3 months of age mice
and adenosine and AMP, which are both down-regulated in 3 months of
age mice and D2 mice with IOP elevation. KEGG pathway analysis of
metabolites demonstrates signicant pathway enrichment e.g. of path-
ways related to purine metabolism, nicotinate and nicotinamide meta-
bolism, taurine and hypotaurine metabolism (Fig. 5B). The comparison
of these altered metabolites with metabolites changed in the retina of 12
months of age or 3 months of age mice after a single IOP injury showed
that 3 metabolites were altered in all 3 injury groups (Fig. 5C). Gua-
nosine and uridine were signicantly decreased in all injury groups
(Fig. 5D and E) and proline was signicantly increased (Fig. 5F).
Taken together, these data indicate that retinal neurotransmitters are
altered with aging and following IOP elevation. The underlying mech-
anisms and impact of this on age-related vulnerability to IOP and
glaucoma are still to be determined.
6. Discussion
Advancing age induced clear changes in the metabolite proles of
the mouse retina with signicant changes in the glutamate and GABA
signaling pathways. IOP elevation led to a further reduction in GABA
levels and altered glutamate metabolism both after short-term IOP
elevation as well as in response to long term IOP elevation in the DBA/2J
mouse. This is especially interesting as glutamate and GABA are the
main neurotransmitters that drive excitatory and inhibitory inputs into
RGCs.
Indoxyl sulfate was the metabolite most changed with age. Indoxyl
sulfate has been shown to promote neuroinammation and oxidative
stress but also decreases GABA receptors in the brain (Sun et al., 2021).
This was supported by reduced levels of GABA in the 12 months of age
retina. We also observed signicant alterations in the biosynthesis and
degradation pathways of the branch chain amino acids valine, leucine
and isoleucine which were enriched in the aging retina. These amino
acids are involved in glutamate production and maintenance, and dis-
balance can lead to oxidative stress in the retina (Lieth et al., 2001; Ola
et al., 2019). These amino acids can also promote neuroprotection after
traumatic brain injury by possibly buffering and maintaining correct
levels of synaptic glutamate and GABA (Cole et al., 2010). In addition, a
recent study demonstrating a reduction of GABAergic amacrine cells in
the aging retina, underlining our analysis of lower levels of GABA in the
aged retina (Zhou et al., 2024).
Sarcosine and L-lysine levels were also decreased with age and
impact glutamate signaling. Sarcosine is an inhibitor of glycine 1
transporter and enhances NMDA receptor signaling by increasing
extracellular glycine (Hanuska et al., 2016). By limiting excitotoxic
glutamate, sarcosine was shown to promote neuroprotection in a model
of oxygen/glucose deprivation of the rat hippocampus (Pinto et al.,
2012). L-lysine is relevant for glutamine metabolism and is a glutamate
precursor (Xiao et al., 2018). It therefore can inuence glutamate levels
in the CNS (Papes et al., 2001), and also inhibit excitotoxic glutamate
signaling (Kondoh et al., 2010). Lysine can additionally serve as fuel
source for glucose deprived brain astrocytes (Harders et al., 2024).
GWAS analysis studying two glaucoma cohorts have linked several
genes involved in the lysine degradation pathway to primary open angle
glaucoma (GLT25D2, EHMTA, ALDH2 and ALDH3A2) (Bailey et al.,
2014).
IOP elevation was seen to further impact the effect of advancing age
on retinal metabolites. Both GABA and taurine expression was signi-
cantly decreased after IOP elevation. GABA is the main inhibitory
neurotransmitter of the retina with taurine, the most abundant amino
acid in the retina, acting as GABA mimetic (Garcia-Ayuso et al., 2024).
GABAergic signals play an important role in shaping neuronal response
patterns and reducing RGC output frequency and RGC excitability
(Wang et al., 2007). We found signicantly lower levels of GABA in the
injured retina (compared to non-injured retina) of both 3 months of age
and 12 months of age mice. This was further evidenced by a reduced
numbers of GABA-expressing cells in the immunohistochemistry anal-
ysis. We detected GABA positive cells in the RGC as well as the ONL layer
of the retina. GABA-hi cells resemble starburst amacrine cells, which
normally reside within the INL, however can also be displaced and then
appear in the GCL, where they account for approximately 12 % of the
total somas (Jeon et al., 1998). GABA-lo cells within the GCL are most
likely RGCs, that can show a varied range of GABA signal within the cell
population (Marc and Jones, 2002), whereas GABA-lo cells in the INL
represent different types of amacrine cells that can be found in the retina
(Anderson et al., 2011). Reduction of eIPSC amplitude and sIPSC fre-
quency day 3 after IOP injury further point towards reduced levels of
GABA and its related receptors. There is a direct relationship between
IPSC amplitude and GABA receptor numbers (Aizenman et al., 1998).
Previous studies have also demonstrated a reduction in GABA
A
receptors
with chronic IOP elevation (Lam et al., 2003; Zhou et al., 2017). The
reduction of sIPSC frequency further points to a decrease in GABA
release following IOP elevation, and similar changes have been
demonstrated following chronic IOP elevation (Zhou et al., 2019).
Interestingly, untargeted analysis of 2 large glaucoma GWAS datasets
showed alterations in genes involved in GABA metabolism in primary
open angle glaucoma (Bailey et al., 2014). In addition, a recently pub-
lished study evaluating neural specicity in the visual cortex showed
decreasing levels of GABA in the visual cortex of glaucoma patients
which could be correlated to deteriorating neural specicity in the visual
cortex (Bang et al., 2023).
Changes in taurine levels have also been demonstrated in several
glaucoma animal models, and studies analyzing the metabolome of the
aqueous humor found decreased levels of taurine in glaucoma patients
(Buisset et al., 2019; Froger et al., 2012). Mice subjected to increased
IOP with supplements including homotaurine, a taurine analog, signif-
icant improvement of the inner–retina derived photopic negative
response (PhNR) signal in the ERG was detected (Cammalleri et al.,
2020). Our analysis further shows the increase guanidoacetic acid and
4-guanidinobutanoic acid in the retina of 3 months of age, but not 12
months of age mice after IOP elevation. Both guanidoacetic acid and
4-guanidinobutanoic acid can act as a GABA mimetic and can evoke
neuronal GABA
A
receptor mediated currents (Meera et al., 2023). We
hypothesize that this could be a compensatory mechanism involved in
the retina of young mice, which is absent in the retina of 12 months of
age mice after IOP injury.
The changes of metabolites found in all three metabolite measure-
ments after IOP elevation (in the 3 months of age mice and the 12
months of age mice data generated here and the publicly available
metabolomics dataset using the DBA/2J mouse model (Harder et al.,
K.C. Bell et al.
Neurobiology of Aging 150 (2025) 69–79
77

2020)) point towards a potential increase in glutamate in the retina of
eyes subjected to either short or long-term IOP elevation. Guanosine,
which was found in reduced levels, is a purine nucleoside which has
neurotrophic and anti-apoptotic effects, mediated by its ability to
stimulate neuroprotective factors from astrocytes. Guanosine is also
involved in the clearance of excess glutamate from the synaptic cleft into
astrocytes. This mechanism is understood to be age-dependent (Bettio
et al., 2016). Uridine, also reduced in the retina after IOP elevation, is
also known to alter glutaminergic synaptic transmission and plasticity
and promotes neuroprotective effects. Although the exact mechanism is
still unknown, uridine has shown protective and positive effects on
diseases related to excess glutamate such as depression or traumatic
brain injury and stroke (Chang et al., 2019). L-Proline on the other hand
was increased in the retina after IOP elevation and is a non-essential
amino acid neuromodulator and a known neural metabotoxin (Nadler
et al., 1988). It is an agonist of glycine and glutamate receptors (Henzi
et al., 1992) and additionally can induce oxidative stress in the CNS
(Delwing et al., 2003).
Our results demonstrate signicant alterations of retinal metabolites
with age and after IOP injury. One limitation of this study is that it was
restricted to a pre-dened set of metabolites. We therefore cannot pro-
vide a comprehensive overview of all metabolic pathways in the retina.
Although measuring the whole retina can bring value, performing cell
type specic metabolomics would provide greater resolution with
respect to the identication of disease specic pathways. Despite this
limitation, we could detect alterations to glaucoma-relevant neuro-
transmitters such as GABA and we could conrm our results with 2 in-
dependent methods. Of course, these experiments were performed in
mouse models and can therefore only always in parts represent the
human disease.
7. Conclusions
We believe our data will provide a useful resource for researchers
studying the metabolomic consequences of age and IOP injury. Good
discrimination between young and middle-aged retina was possible with
the panel of metabolites measured in this study. The changes detected
here support an alteration/decrease of glutamate and GABA signaling
with aging, as well as in pre-apoptotic, functionally impaired, but fully
recoverable RGCs following IOP injury. Reduction of GABA after IOP
injury points towards a loss of inhibitory neurotransmission in the
retina, which possibly is compensated in the young mice by increasing
GABA-mimetics in the retina.
Ethics approval
All breeding and experimental procedures were undertaken in
accordance with the Association for Research for Vision and Ophthal-
mology Statement for the Use of Animals in Ophthalmic and Research
and were approved by the SingHealth Institutional Animal Care and Use
Committee (2019/SHS/1534). Experiments for electrophysiology were
performed at the Florey Institute, Melbourne, under the guidelines of the
Australian Code of Practice for the Care and were approved by The
Florey Animal Ethics Committee (18–112-FINMH).
Funding
PAW is supported by Karolinska Institutet in the form of a Board of
Research Faculty Funded Career Position, by St. Erik Eye Hospital
philanthropic donations, and Vetenskapsrådet 2022–00799. JGC is
supported by the EOLAS foundation, Australia. KB was supported by a
NMRC Clinician scientist individual research grant (CS-IRG NIG) grant.
JCG is supported within a CRP grant National Research Foundation
Singapore (CRP24–2020–0077).
CRediT authorship contribution statement
Jones Bryan W: Writing – review & editing, Validation, Methodol-
ogy, Data curation. Williams Pete A: Writing – review & editing, Su-
pervision, Methodology, Data curation, Conceptualization. Crowston
Jonathan G: Writing – review & editing, Supervision, Methodology,
Funding acquisition, Conceptualization. Bell Katharina C.: Writing –
review & editing, Writing – original draft, Investigation, Formal anal-
ysis, Data curation, Conceptualization. Chrysostomou Vicki: Writing –
review & editing, Methodology, Conceptualization. Karlsson Markus:
Methodology.
Declaration of Competing Interest
The authors declare that they have no competing interests.
Acknowledgements
The authors would like to thank the Swedish Metabolomics Centre
for performing the metabolomics analysis.
Appendix A. Supporting information
Supplementary data associated with this article can be found in the
online version at doi:10.1016/j.neurobiolaging.2025.03.004.
Data availability
All raw data is available from Zenodo, doi: 10.5281/zen-
odo.14723294. Supplementary Dataset 1 available from the same doi
contains the processed and annotated data.
References
Aizenman, C.D., et al., 1998. Polarity of long-term synaptic gain change is related to
postsynaptic spike ring at a cerebellar inhibitory synapse. Neuron 21 (4), 827–835.
https://doi.org/10.1016/s0896-6273(00)80598-x.
Anderson, J.R., et al., 2011. Exploring the retinal connectome. Mol. Vis. 17, 355–379.
Bailey, J.N., et al., 2014. Hypothesis-independent pathway analysis implicates GABA and
acetyl-CoA metabolism in primary open-angle glaucoma and normal-pressure
glaucoma. Hum. Genet 133 (10), 1319–1330. https://doi.org/10.1007/s00439-014-
1468-7.
Baltan, S., et al., 2010. Metabolic vulnerability disposes retinal ganglion cell axons to
dysfunction in a model of glaucomatous degeneration. J. Neurosci. 30 (16),
5644–5652. https://doi.org/10.1523/JNEUROSCI.5956-09.2010.
Bang, J.W., et al., 2023. GABA decrease is associated with degraded neural specicity in
the visual cortex of glaucoma patients. Commun. Biol. 6 (1), 679. https://doi.org/
10.1038/s42003-023-04918-8.
Bengtsson, B., et al., 2024. Comparison of disease severity in glaucoma patients
identied by screening in the 1990s and in routine clinical care in the 2010s in
Sweden. Acta Ophthalmol. 102 (2), 238–245. https://doi.org/10.1111/aos.15777.
Bettio, L.E., et al., 2016. Guanosine and its role in neuropathologies. Purinergic Signal 12
(3), 411–426. https://doi.org/10.1007/s11302-016-9509-4.
Buisset, A., et al., 2019. Metabolomic proling of aqueous humor in glaucoma points to
taurine and spermine deciency: ndings from the eye-d study. J. Proteome Res 18
(3), 1307–1315. https://doi.org/10.1021/acs.jproteome.8b00915.
Cammalleri, M., et al., 2020. A dietary combination of forskolin with homotaurine,
spearmint and B vitamins protects injured retinal ganglion cells in a rodent model of
hypertensive glaucoma. Nutrients 12 (4). https://doi.org/10.3390/nu12041189.
Canovai, A., et al., 2023. Pyrroloquinoline quinone drives ATP synthesis in vitro and in
vivo and provides retinal ganglion cell neuroprotection. Acta Neuropathol. Commun.
11 (1), 146. https://doi.org/10.1186/s40478-023-01642-6.
Casson, R.J., et al., 2021. Retinal energy metabolism in health and glaucoma. Prog. Retin
Eye Res 81, 100881. https://doi.org/10.1016/j.preteyeres.2020.100881.
Chang, E.H., et al., 2019. Assessment of glutamatergic synaptic transmission and
plasticity in brain slices: relevance to bioelectronic approaches. Bioelectron. Med 5,
6. https://doi.org/10.1186/s42234-019-0022-2.
Chrysostomou, V., et al., 2014. Forced exercise protects the aged optic nerve against
intraocular pressure injury. Neurobiol. Aging 35 (7), 1722–1725. https://doi.org/
10.1016/j.neurobiolaging.2014.01.019.
Chrysostomou, V., et al., 2024. A new model of axon degeneration in the mouse optic
nerve using repeat intraocular pressure challenge. Exp. Eye Res. 238, 109722.
https://doi.org/10.1016/j.exer.2023.109722.
Chrysostomou, V., Crowston, J.G., 2013. The photopic negative response of the mouse
electroretinogram: reduction by acute elevation of intraocular pressure. Invest
Ophthalmol. Vis. Sci. 54 (7), 4691–4697. https://doi.org/10.1167/iovs.13-12415.
K.C. Bell et al.
Neurobiology of Aging 150 (2025) 69–79
78

Cole, J.T., et al., 2010. Dietary branched chain amino acids ameliorate injury-induced
cognitive impairment. Proc. Natl. Acad. Sci. USA 107 (1), 366–371. https://doi.org/
10.1073/pnas.0910280107.
Covarrubias, A.J., et al., 2021. NAD(+) metabolism and its roles in cellular processes
during ageing. Nat. Rev. Mol. Cell Biol. 22 (2), 119–141. https://doi.org/10.1038/
s41580-020-00313-x.
Crowston, J.G., et al., 2015. An acute intraocular pressure challenge to assess retinal
ganglion cell injury and recovery in the mouse. Exp. Eye Res 141, 3–8. https://doi.
org/10.1016/j.exer.2015.03.006.
Delwing, D., et al., 2003. Proline induces oxidative stress in cerebral cortex of rats. Int J.
Dev. Neurosci. 21 (2), 105–110. https://doi.org/10.1016/s0736-5748(02)00109-0.
van den Berg, R.A., et al., 2006. Centering, scaling, and transformations: improving the
biological information content of metabolomics data. BMC Genom. 7, 142. https://
doi.org/10.1186/1471-2164-7-142.
Froger, N., et al., 2012. Taurine provides neuroprotection against retinal ganglion cell
degeneration. PLoS One 7 (10), e42017. https://doi.org/10.1371/journal.
pone.0042017.
Fry, L.E., et al., 2018. The coma in glaucoma: Retinal ganglion cell dysfunction and
recovery. Prog. Retin Eye Res 65, 77–92. https://doi.org/10.1016/j.
preteyeres.2018.04.001.
Garcia-Ayuso, D., et al., 2024. Taurine: a promising nutraceutic in the prevention of
retinal degeneration. Neural Regen. Res 19 (3), 606–610. https://doi.org/10.4103/
1673-5374.380820.
GBD 2019 Blindness and Vision Impairment Collaborators, Vision Loss Expert Group of
the Global Burden of Disease Study, 2021. Causes of blindness and vision
impairment in 2020 and trends over 30 years, and prevalence of avoidable blindness
in relation to VISION 2020: the Right to Sight: an analysis for the Global Burden of
Disease Study. Lancet Glob. Health 9 (2), e144–e160. https://doi.org/10.1016/
S2214-109X(20)30489-7.
Hanuska, A., et al., 2016. Some operational characteristics of glycine release in rat retina:
the role of reverse mode operation of glycine transporter Type-1 (GlyT-1) in ischemic
conditions. Neurochem. Res 41 (1-2), 73–85. https://doi.org/10.1007/s11064-015-
1713-z.
Harder, J.M., et al., 2020. Disturbed glucose and pyruvate metabolism in glaucoma with
neuroprotection by pyruvate or rapamycin. Proc. Natl. Acad. Sci. USA 117 (52),
33619–33627. https://doi.org/10.1073/pnas.2014213117.
Harders, A.R., et al., 2024. Exogenous substrates prevent the decline in the cellular ATP
content of primary rat astrocytes during glucose deprivation. Neurochem. Res 49 (5),
1188–1199. https://doi.org/10.1007/s11064-024-04104-0.
Henzi, V., et al., 1992. L-proline activates glutamate and glycine receptors in cultured rat
dorsal horn neurons. Mol. Pharm. 41 (4), 793–801.
Jeon, C.J., et al., 1998. The major cell populations of the mouse retina. J. Neurosci. 18
(21), 8936–8946. https://doi.org/10.1523/JNEUROSCI.18-21-08936.1998.
Kondoh, T., et al., 2010. Lysine and arginine reduce the effects of cerebral ischemic
insults and inhibit glutamate-induced neuronal activity in rats. Front. Integr.
Neurosci. 4 (18). https://doi.org/10.3389/fnint.2010.00018.
Kong, Y.X., et al., 2009. Functional changes in the retina during and after acute
intraocular pressure elevation in mice. Invest. Ophthalmol. Vis. Sci. 50 (12),
5732–5740. https://doi.org/10.1167/iovs.09-3814.
Lam, D.Y., et al., 2003. Neurochemical correlates of cortical plasticity after unilateral
elevated intraocular pressure in a primate model of glaucoma. Invest Ophthalmol.
Vis. Sci. 44 (6), 2573–2581. https://doi.org/10.1167/iovs.02-0779.
Lieth, E., et al., 2001. Nitrogen shuttling between neurons and glial cells during
glutamate synthesis. J. Neurochem. 76 (6), 1712–1723. https://doi.org/10.1046/
j.1471-4159.2001.00156.x.
Marc, R.E., et al., 1995. Pattern recognition of amino acid signatures in retinal neurons.
J. Neurosci. 15 (7 Pt 2), 5106–5129. https://doi.org/10.1523/JNEUROSCI.15-07-
05106.1995.
Marc, R.E., Jones, B.W., 2002. Molecular phenotyping of retinal ganglion cells.
J. Neurosci. 22 (2), 413–427. https://doi.org/10.1523/JNEUROSCI.22-02-
00413.2002.
Meera, P., et al., 2023. Guanidinoacetate (GAA) is a potent GABA(A) receptor GABA
mimetic: Implications for neurological disease pathology. J. Neurochem. 165 (3),
445–454. https://doi.org/10.1111/jnc.15774.
Mir, C., et al., 2003. CDP-choline prevents glutamate-mediated cell death in cerebellar
granule neurons. J. Mol. Neurosci. 20 (1), 53–60. https://doi.org/10.1385/JMN:20:
1:53.
Myer, C., et al., 2020. Differentiation of soluble aqueous humor metabolites in primary
open angle glaucoma and controls. Exp. Eye Res. 194, 108024. https://doi.org/
10.1016/j.exer.2020.108024.
Nadler, J.V., et al., 1988. Toxicity of L-proline toward rat hippocampal neurons. Brain
Res. 456 (1), 168–172. https://doi.org/10.1016/0006-8993(88)90358-7.
Ola, M.S., et al., 2019. Gabapentin attenuates oxidative stress and apoptosis in the
diabetic rat retina. Neurotox. Res. 36 (1), 81–90. https://doi.org/10.1007/s12640-
019-00018-w.
Papes, F., et al., 2001. The essential amino acid lysine acts as precursor of glutamate in
the mammalian central nervous system. FEBS Lett. 488 (1-2), 34–38. https://doi.
org/10.1016/s0014-5793(00)02401-7.
Pinto, M.C., et al., 2012. Pharmacological induction of ischemic tolerance in
hippocampal slices by sarcosine preconditioning. Neurochem. Int. 61 (5), 713–720.
https://doi.org/10.1016/j.neuint.2012.06.018.
Quigley, H.A., Broman, A.T., 2006. The number of people with glaucoma worldwide in
2010 and 2020. Br. J. Ophthalmol. 90 (3), 262–267. https://doi.org/10.1136/
bjo.2005.081224.
Rossi, C., et al., 2019. Multi-omics approach for studying tears in treatment-naive
glaucoma patients. Int. J. Mol. Sci. 20 (16). https://doi.org/10.3390/ijms20164029.
Schnackenberg, L.K., et al., 2007. Metabonomics evaluations of age-related changes in
the urinary compositions of male Sprague Dawley rats and effects of data
normalization methods on statistical and quantitative analysis. BMC Bioinforma. 8 7
(7), S3. https://doi.org/10.1186/1471-2105-8-S7-S3.
Schousboe, A., et al., 2013. Astrocytic control of biosynthesis and turnover of the
neurotransmitters glutamate and GABA. Front. Endocrinol. 4, 102. https://doi.org/
10.3389/fendo.2013.00102.
Sun, C.Y., et al., 2021. Indoxyl sulfate caused behavioral abnormality and
neurodegeneration in mice with unilateral nephrectomy. Aging (Albany NY) 13 (5),
6681–6701. https://doi.org/10.18632/aging.202523.
Tan, J.K., et al., 2023. Metabolomics proling of age-associated metabolites in malay
population. Oxid. Med. Cell Longev. 2023, 4416410. https://doi.org/10.1155/
2023/4416410.
Tham, Y.C., et al., 2014. Global prevalence of glaucoma and projections of glaucoma
burden through 2040: a systematic review and meta-analysis. Ophthalmology 121
(11), 2081–2090. https://doi.org/10.1016/j.ophtha.2014.05.013.
Tribble, J.R, et al., 2021. Nicotinamide provides neuroprotection in glaucoma by
protecting against mitochondrial and metabolic dysfunction. Redox Biol. 43,
101988. https://doi.org/10.1016/j.redox.2021.101988.
Wang, C.T., et al., 2007. GABA(A) receptor-mediated signaling alters the structure of
spontaneous activity in the developing retina. J. Neurosci. 27 (34), 9130–9140.
https://doi.org/10.1523/JNEUROSCI.1293-07.2007.
Wang, Y., et al., 2018. Metabolic signature of the aging eye in mice. Neurobiol. Aging 71,
223–233. https://doi.org/10.1016/j.neurobiolaging.2018.07.024.
Williams, P.A., et al., 2017. Vitamin B(3) modulates mitochondrial vulnerability and
prevents glaucoma in aged mice. Science 355 (6326), 756–760. https://doi.org/
10.1126/science.aal0092.
Wyczalkowska-Tomasik, A., et al., 2017. Age-dependent increase in serum levels of
indoxyl sulphate and p-cresol sulphate is not related to their precursors: tryptophan
and tyrosine. Geriatr. Gerontol. Int. 17 (6), 1022–1026. https://doi.org/10.1111/
ggi.12811.
Xia, J., et al., 2009. MetaboAnalyst: a web server for metabolomic data analysis and
interpretation. Nucleic Acids Res. 37 (WebServer), W652–660. https://doi.org/
10.1093/nar/gkp356.
Xia, J., Wishart, D.S., 2011. Web-based inference of biological patterns, functions and
pathways from metabolomic data using MetaboAnalyst. Nat. Protoc. 6 (6), 743–760.
https://doi.org/10.1038/nprot.2011.319.
Xiao, C.W., et al., 2018. Dietary L-lysine supplementation altered the content of
pancreatic polypeptide, enzymes involved in glutamine metabolism, and beta-actin
in rats. Amino Acids 50 (12), 1729–1737. https://doi.org/10.1007/s00726-018-
2648-x.
Yoon, S.J., et al., 2009. Decreased glutamate/glutamine levels may mediate cytidine’s
efcacy in treating bipolar depression: a longitudinal proton magnetic resonance
spectroscopy study. Neuropsychopharmacology 34 (7), 1810–1818. https://doi.org/
10.1038/npp.2009.2.
Zhang, X., et al., 2023. Branched-chain amino acids metabolism and their roles in
retinopathy: from relevance to mechanism. Nutrients 15 (9). https://doi.org/
10.3390/nu15092161.
Zhou, X., et al., 2017. Differential modulation of GABA(A) and NMDA receptors by an
alpha7-nicotinic acetylcholine receptor agonist in chronic glaucoma. Front. Mol.
Neurosci. 10, 422. https://doi.org/10.3389/fnmol.2017.00422.
Zhou, X., et al., 2019. Quercetin enhances inhibitory synaptic inputs and reduces
excitatory synaptic inputs to OFF- and ON-type retinal ganglion cells in a chronic
glaucoma rat model. Front Neurosci. 13, 672. https://doi.org/10.3389/
fnins.2019.00672.
Zhou, Y., et al., 2024. Dysregulated energy and protein homeostasis and the loss of
GABAergic amacrine cells in aging retina. Exp. Eye Res. 245, 109985. https://doi.
org/10.1016/j.exer.2024.109985.
K.C. Bell et al.
Neurobiology of Aging 150 (2025) 69–79
79