Soluble guanylate cyclase deficiency drives retinal ganglion cell neurodegeneration with age in female mice through disrupted oxidative metabolism

Abstract

Dysfunctional cGMP signaling is implicated in multiple neurodegenerative diseases of the central nervous system (CNS), including glaucoma, an optic neuropathy and leading cause of irreversible blindness. Female mice lacking the alpha catalytic subunit of soluble guanylate cyclase (sGCα1^−/−), the active site that binds nitric oxide (NO) to produce cGMP, exhibit progressive retinal ganglion cell (RGC) degeneration with age. Yet, the role of sGC in age- and sex-dependent RGC function remains uncharacterized. We investigated how preventing NO binding to sGC influences RGC function in the context of aging and sex by combining bulk and single-cell RNA sequencing, Western blotting, mitochondrial ultrastructural analysis, visual acuity measurements, and in vivo measurements of retinal oxidative metabolism. We found that global sGCα1 deletion impairs visual function and RGC health in aging female mice, while male mice remained unaffected. Glucose uptake was significantly disrupted in female sGCα1^−/−retinas with age, and accompanied by reduced retinal expression of the glucose transporter, GLUT1. Aged sGCα1^−/− females also exhibited dysregulated retinal mitochondrial gene and protein expression and increased nitrosative stress localized to the RGC layer. RGC mitochondria in male mice increased in size with age, while female mitochondria did not. Furthermore, retinal metabolic analysis showed decreased oxygen consumption rate in aged female but not male sGCα1^−/−retinas, suggesting impaired oxidative metabolism. These findings reveal a potential sex-specific role for cGMP signaling in maintaining retinal metabolic integrity and RGC function with age. Our results point to a possible mechanistic link between impaired cGMP signaling and age-related retinal neurodegeneration in females, highlighting the sGC-cGMP signaling pathway as a promising therapeutic target for glaucoma and other CNS neurodegenerative diseases.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-34243-5.

Introduction

The nitric oxide (NO)-soluble guanylate cyclase (sGC)-cyclic guanosine monophosphate (cGMP) signaling axis plays a well-established role in vascular regulation, including the modulation of vascular tone, regional blood flow, and platelet function^1–3. However, evidence indicates that cGMP signaling also exerts critical functions beyond the vasculature. Studies suggest that cGMP can be neuroprotective in the central nervous system (CNS) through its direct anti-apoptotic effect on neurons^4–12. The action of cGMP on the survival of retinal cells seems to be location-dependent; elevated levels of cGMP in rod and cone photoreceptors can be detrimental to their survival^13,14 whereas cGMP signaling in RGCs appears to promote survival⁴. NO, synthesized by nitric oxide synthase enzymes, binds to sGC leading to the generation of cGMP which has a broad spectrum of cellular targets including protein kinase G, and ion channels^15,16. Importantly, dysregulation of the cGMP signaling pathway has been implicated in the pathogenesis of several neurodegenerative diseases, including Alzheimer’s disease^17–19, Huntington’s disease^20,21, as well as in glaucoma, an age-related optic neuropathy^4,22–24. Aging and cognitive decline is commensurate with a decline in cGMP levels in the CNS; sGC activity and cGMP levels are reduced in the aging brain^25,26 and decreased cGMP correlates with cognitive decline in Alzheimer’s patients²⁷. These findings suggest that cGMP plays a neuroprotective role in the aging CNS and may represent a promising therapeutic target for neurodegenerative diseases.

Similar to Alzheimer’s and Parkinson’s, glaucoma incidence also increases with age. Glaucoma is the leading cause of irreversible blindness worldwide^28,29. The disease progresses through sensitivity to intraocular pressure (IOP), which is conveyed to retinal ganglion cell (RGC) axons at the optic nerve head³⁰. Across glaucoma’s complex etiology, age and IOP are the primary risk factors with IOP remaining the only modifiable risk factor. Despite IOP-lowering therapies including topical medications and surgery, 40–50% of patients continue to experience vision loss³¹, highlighting the urgent need for IOP-independent treatment strategies. Impairments in cGMP signaling have been observed in glaucoma patients; cGMP levels are reduced in aqueous humor³², and single-nucleotide polymorphisms in genes regulating the pathway are associated with increased glaucoma risk in females with early vision loss^33–35. We demonstrated that global knockout of the α1 subunit of sGC (sGCα1^−/−) leads to systemic and retinal vascular dysfunction and age-related RGC degeneration in female mice³³. Furthermore, pharmacologic preservation of cGMP pools using tadalafil (an FDA-approved phosphodiesterase-5 inhibitor), prevented RGC loss independently of IOP in both sGCα1^−/− mice and an ocular hypertension model⁴. These findings suggest an IOP-independent, cGMP-mediated mechanism contributing to retinal neurodegeneration with age.

In this study, we investigated the sex- and age-related mechanisms underlying sGCα1^−/−-driven RGC degeneration. We report that global loss of sGC promotes RGC degeneration and loss of visual acuity in female mice with age, accompanied by impaired retinal metabolism and mitochondrial dysfunction. Male mice, however, do not exhibit this trend. Specifically, we show that: (1) female, but not male, sGCα1^−/− mice exhibit age-dependent RGC dysfunction and visual decline; (2) female sGCα1^−/−retinas have reduced GLUT1 expression and impaired glucose uptake with age; (3) nitrosative stress increases in the RGC layer of aged female sGCα1^−/−mice; (4) mitochondrial gene and protein expression are dysregulated in female sGCα1^−/−retinas with age, while preliminary results show that morphology is unchanged; and (5) mitochondrial dysfunction occurs in aged, female sGCα1^−/− animals, but not male animals. Furthermore, co-expression of sGC and GLUT1 across multiple retinal cell types outside of the vasculature suggests cGMP signaling plays a critical role in retinal metabolic homeostasis.

In summary, our findings reveal a potential sex-specific vulnerability to RGC degeneration with age driven by disrupted cGMP signaling and altered retinal metabolism. These results point to a possible mechanistic link between cGMP signaling and retinal neurodegeneration with age in females and underscore the potential of targeting the sGC-cGMP signaling pathway as a therapeutic strategy for glaucoma and other CNS degenerative diseases.

Methods

Animals

All animal studies were conducted in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals and were approved by the IACUC committee at Vanderbilt University Medical Center (IACUC animal protocol number: M2300049). For the studies described, age-matched female and male wild type Sv/129S6 (WT) and sGCα1^−/− mice on an Sv/129S6 background^36–38 were bred and housed at the Vanderbilt University Medical Center. Genotyping of all progeny was performed by Transnetyx Inc., which utilizes duplicate sample processing (real-time TaqMan PCR), ensuring the accuracy of each mutation. The following probe sequences were used to confirm deletion of the 6th exon of the alpha subunit of sGC in all mice used: WT - CACCTAGGTGTCTGTCTAGACACTGAG, GAACAAAACTGGGAGGACTGG and KO - CATATTCATTTCAAAGTTAGCTAGCGGATC, GAGGACTGGAGATTGTAGTTGTTTCCC. Young mice were aged 10–12 weeks old and aged mice were 60–62 weeks of age. Weight ranges of mice used in the study were as follows: young male WT (24 ± 2.3 g), young male sGC KO (24.8 ± 1.9 g), aged male WT (32 ± 3.5 g), aged male sGC KO (27 ± 1.6 g), young female WT (22 ± 1.3 g), young female sGC KO (19 ± 3.1 g), aged female WT (28 ± 1.5 g), and aged female sGC KO (24 ± 1.5 g). Mice were maintained on a 12 h light/dark cycle with ad libitum access to standard mouse chow and water. All mice were euthanized with 10 mg intraperitoneal pentobarbital injection. All authors complied with ARRIVE guidelines; observers were masked to animal genotype, experimental group and performed all data acquisition and analyses described.

Tissue preparation

Mice were sacrificed through intraperitoneal pentobarbital injection followed by transcardial perfusion of 1X PBS. Eyes were immediately enucleated, and the retinas dissected fresh. For Western blotting, retinas were transferred to ice-cold RIPA buffer and processed immediately. For whole retinal RNA sequencing analysis, fresh retina were transferred to an RNase-free 1.5 mL tubes on ice for processing. For wholemount immunohistochemistry, retinas were transferred to a solution of collagenase (LS005273, Worthington Biochemical) and hyaluronidase (LS002592, Worthington Biochemical) in Ame’s media (A1372-25, US Biologic) for 10 min at room temperature to digest the vitreous. Following this incubation, vitreous was removed as a single sheet using forceps. Relief cuts were made to divide the tissue into quadrants and the retinas were then transferred to a solution of 4% paraformaldehyde in 1X PBS (Electron Microscopy Science; 15710) to shake at room temperature for 1 h. Retinas were stored in 1X PBS with 0.2% sodium azide at 4 °C until used. For retinal tissue sections, following transcardial perfusion of 1X PBS and 4% paraformaldehyde in 1X PBS, eyes were embedded in paraffin and sectioned by Vanderbilt Vision Research Center histology core and stored at room temperature for immunolabeling.

Cholera toxin B (CTB) assessment of RGC axon transport

Mice were anesthetized with 2.5% isoflurane and injected both eyes intravitreally with 1.5 µL of 1 mg/mL solution of cholera toxin subunit B (CTB) conjugated to Alexa-555 (Molecular Probes, Eugene, OR, USA) following the published protocol^39–41. Two days later, the mice were transcardially perfused with 1X PBS followed by 4% paraformaldehyde, and dissected tissues were cryoprotected in 30% sucrose. We prepared coronal midbrain sections (50 μm thick) on a freezing sliding microtome and photographed sections of the superior colliculus using a Nikon Ti Eclipse microscope (Nikon Instruments Inc., Melville, NY, USA). The intensity of the CTB signal (intact transport) within the collicular retinotopic map was quantified using a custom ImagePro macro (Media Cybernetics, Bethesda, MD, USA) as described^40,42.

Retinal paraffin section immunohistochemistry

Whole eye paraffin sections were heated on a slide warmer set to 60 °C for 20 min followed by deparaffinization and rehydration series in a xylene and ethanol gradient. Slides were incubated in sodium citrate (10 mM, pH 6) for 30 min at room temperature for optimal antigen retrieval. Slides were washed in 1X PBS and then covered in 0.1% sodium borohydride in 1X PBS for 30 min to minimize retinal auto-fluorescence. Slides were washed 3x in 1X PBS for 5 min and then sections blocked in a solution containing 5% normal donkey serum (NDS) in 0.1% Triton X-100 in 1X PBS for 2 h at room temperature. Slides were incubated overnight at 4 °C in primary antibodies against 3-Nitrotyrosine (rabbit 1:200; EMD Millipore ab5411), and GLUT-1 (rabbit 1:400; abcam ab115730). Slides were covered by parafilm to prevent evaporation of primary antibody solution. The following day slides were washed in 1X PBS and incubated at room temperature with appropriate secondary antibodies (Jackson ImmunoResearch Laboratories Inc.) for 2 h. Slides were cover slipped and mounted with DAPI fluoromount-G (Southern Biotech; 0100 − 20), dried, and sealed with nail polish for imaging.

Wholemount RGC quantification

Retinal RGC counts were carried out as described previously for females⁴. For males, RGC counts were carried out as follows: Retinas were removed from 1X PBS with 0.2% sodium azide and transitioned through a sucrose gradient (10% to 20% to 30%). The retinas were then frozen and thawed for a total of three cycles. Retinas were blocked in a solution containing 5% NDS in 0.1% Triton-X 100 in 1X PBS for 2 h at room temperature while shaking. Whole retinas were immunolabeled using the following primary antibodies: rabbit anti-RBPMS (GeneTex, GTX118619; 1:500). Retinas were incubated in primary antibodies for four days at 4 °C. Secondary antibodies from Jackson ImmunoResearch Laboratories Inc conjugated to fluorescent fluorophores were incubated for 2 h at room temperature. Retinas were then washed in 1X PBS for a total of three five-minute washes. Finally, retinas were washed three times in 1X PBS for 5 min each and then mounted with with DAPI fluoromount-G (Southern Biotech; 0100 − 20), cover-slipped and sealed with nail polish for imaging. Retinas were montage imaged with a 20X objective on a Nikon Ni Eclipse fluorescence microscope. Local contrast of the resulting images was enhanced using the CLAHE function embedded in Fiji-ImageJ⁴³. These enhanced images were passed into the opensource software RGCode for automated RGC counting⁴⁴. RGC count data was normalized to WT controls so as to compare the counts between published results and new data included in this manuscript.

Fluorescent imaging and analysis

Retinal section images were acquired using Nikon Ti-E Spinning Disk confocal microscope, Nikon Ni Eclipse fluorescence microscope, and CSU-W1 SoRa microscope. For imaging of retinal wholemounts, 20x montages were taken of the retina to orient and closer images were then acquired at 60x magnification. Z-stacks of equal thickness were taken at a step size of 0.3 μm. Z-stacked images were combined using the maximum intensity Z-stack option in Fiji ImageJ. For retinal sections, 20x and 40x images were obtained on the Nikon Ni Eclipse fluorescence microscope as well as the CSU-W1 SoRa microscope. On the CSU-W1 SoRa microscope, Z-stacks were taken to show all staining seen in the width of the retinal section. Z-stacked images were then combined using the maximum intensity Z-stack option in Fiji ImageJ. For comparative quantification of immunohistochemistry (IHC), analyses were carried out on samples stained in the same experiment and images taken using the same imaging parameters. Total integrated density as well as mean gray value for each channel were quantified using Fiji software⁴³.

Western blotting

Retinal tissue was placed into ice-cold RIPA buffer (50 mM Tris HCl pH 7.2, 150 mM NaCl, and 1% Triton X-100) containing 1x Halt protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific Cat #78442) and chopped using forceps. Retinal pieces were then sonicated until homogenous. Retinal samples were then centrifuged at 4 °C at 12,000 RPM for 10 min and supernatant was then transferred to a new prechilled 1.5 mL tube. 4X Laemmli buffer (Bio-Rad; Cat#1610747) containing 2-β-mercaptoethanol was added to the samples and heated for 5 min at 95 °C. Total protein concentration was measured using a BCA (Pierce) Assay (Cat #23225; ThermoFisher Scientific, Waltham, MA, USA) and 20 µg of protein was loaded per retina along with protein ladder (Precision Plus Western Blotting Standards; Bio-Rad #1610376) in 4–20% polyacrylamide gels (Bio-Rad #4561093). Proteins were then transferred onto Immobilon PVDF Membrane (Millipore Sigma IPPFL00010) with the Mini-PROTEAN Tetra Cell System (Bio-Rad #1658005EDU). Transfer was verified using Sigma Ponceau S. (Sigma-Aldrich #P7170) Membranes were then blocked in 5% Dry Milk Powder (Research Products International #M17200-500.0) in 1X TBS (Corning #46-012-CM) for 1 h at room temperature while rocking. Membranes were incubated in primary antibodies against sGC specific to both WT and mutant protein (1:500; Sigma Aldrich Cat# G4280), TOM20 (Rabbit 1:1000; Proteintech Cat# 11802-1-AP), NDUFS3 (Mouse 1:1000 Abcam Cat #ab110246), COX4-I1 (Goat 1:1,000; R&D Systems Cat# AF5814), GLUT-1 (Rabbit 1:1,000; Abcam Cat# ab115730), Phospho-CREB Ser 133 87G3 (Rabbit 1:1000; Cell Signaling Technology Cat# 9198), Nitrotyrosine (Rabbit 1:1000; EMD Millipore Cat# ab5411). For each condition, membranes were incubated in the appropriate loading control GAPDH (Rabbit 1:500; Cell Signaling Technology Cat# 14C10) and alpha-tubulin (Mouse 1:1000; Sigma-Aldrich Cat# 05–829). Membranes were washed in 0.1% Tween in 1X TBS and then incubated in secondary antibodies from Jackson ImmunoResearch Laboratories Inc. for 1 h at room temperature. Membranes were washed and imaged on a Bio-Rad ChemiDoc system. Band intensity was measured in Fiji. The same ROI was used for all bands within a blot to obtain the mean gray value of a band, which was then normalized with the appropriate loading control by dividing band intensity measurements by the measurements of the loading control.

sGC activity assays

Fresh retinal tissue was harvested from WT and sGCα1^−/− mice (n = 3 per group) and placed into 20 volumes (mL of buffer/gram of tissues) of buffer containing 50 mmol/L tris(hydroxymethyl)aminomethane (Tris)-HCl (pH 8), 1 mmol/L EDTA, 1 mmol/L dithiothreitol (DTT), and 1:100 Halt Protease & Phosphatase Single-Use Inhibitor Cocktail (100X) (Cat #78430; ThermoFisher Scientific, Waltham, MA, USA). The samples were sonicated in the lysis buffer and subsequently centrifuged at 20,000 x g for 20 min at 4 °C. Protein supernatants were aspirated and placed into separate reactions: a 3-isobutyl-1-methylxanthine (IBMX; Cat #13347, Cayman Chemical Company, MI) reaction mixture containing 0.5 mM IBMX and 50 mM Tris-HCl and an IBMX reaction mixture containing 0.5 mM IBMX, 50 mM Tris-HCl, and 1 mM PAPA NONOate. All samples were incubated in a 37 °C water bath for 30 min to allow release of NO. The reactions were terminated by adding a 1:1 dilution of 5% TCA. cGMP in the reaction mixture was measured using a commercial enzyme-linked immunosorbent assay (Cayman Chemical Company, MI; Cat # 581021). Results were normalized to total ug protein per sample.

Retinal glucose uptake assay

Retinal uptake of the fluorescent glucose analog was done using 6-NBDG as described⁴⁵. Mice were euthanized via cervical dislocation and retinas were immediately dissected and placed in cold neurobasal-A medium (Gibco #12349-015). Retina were then incubated in 1.5 mL tubes in 200µL of neurobasal A media for 20 min in a 37 °C water bath. Retinas were then incubated 500µM of 6-NBDG (Invitrogen #N23106) in 37 °C water bath for 60 min. Retinas were washed with 500µL of cold neurobasal-A media very carefully as to not tear or rupture the retinas for a total of 3 washes and added to new 1.5 mL microcentrifuge tubes with 100µL of cold neurobasal-A media and chopped. Retinas were sonicated and centrifuged for 15 min at 15,000 RPM. Standards were made up from a concentration of 0–40µM and 100µL of each standard and sample was pipetted into black/clear bottom 96 well plate (Thermo Scientific Cat# 265301). The plate was then read on a SpectraMax M2 plate reader using SoftMax Pro’s fluorescence endpoint assay with an excitation of 483 nm, emission of 550 nm, and a cut off of 530 nm. The samples were then collected and frozen at −80 °C to be used for protein quantification using a BCA (Pierce) Assay (Cat #23225; ThermoFisher Scientific, Waltham, MA, USA). Fluorescence values were then normalized to total sample protein content (µg).

Whole retina bulk RNA sequencing

Fresh retina were dissected from age-matched WT and sGCα1^−/− mice (n = 5 retinas per group) following sacrifice. RNA was extracted using TRIzol (Invitrogen, Cat# 15596026) and treated with deoxyribonuclease (DNase I; Worthington, Cat# LS006333). Experiments were performed in collaboration with the Vanderbilt Technologies for Advanced Genomics core at Vanderbilt University Medical Center. RNA quality was assessed using the 2100 Bioanalyzer (Agilent Technologies) and samples with integrity values greater than 6 were used to generate polyA (mRNA)-enriched libraries, using stranded mRNA sample kits with indexed adaptors (New England BioLabs, Ipswich, MA, United States). Library quality was assessed using the 2100 Bioanalyzer (Agilent Technologies) before being quantitated using KAPA Library Quantification Kits (KAPA Biosystems, Wilmington, MA, United States). Pooled libraries were subjected to 100-bp paired-end sequencing according to the manufacturer’s protocol (Illumina NovaSeq6000). Bcl2fastq2 Conversion Software (Illumina, San Diego, CA, United States) was used to generate de-multiplexed Fastq files. Analysis of RNAseq results was performed in collaboration with the Vanderbilt Technologies for Advanced Genomics Analysis and Research Design core at Vanderbilt University. Reads were aligned to the GENCODE GRCm38.p5 genome using STAR v2.5.3a. GENCODE vM12 gene annotations were provided to STAR to improve the accuracy of mapping. Quality control on raw reads was performed using FastQC. FeatureCounts v1.15.2 was used to count the number of mapped reads to each gene. Significantly differentially expressed genes with adjusted p value < 0.05 and absolute fold change > 2 were detected by DESeq2 v1.14.

MtDNA quantitative PCR analysis

To quantify mtDNA copy number by determining the mtDNA/nDNA ratio, mice were euthanized and eyes were promptly enucleated followed by retinal dissection. Retinas were transferred to 1.5 mL microcentrifuge tubes containing 180 uL buffer ATL and 20 uL proteinase K from DNeasy Blood &Tissue Kit (Qiagen #69506) and then incubated overnight at 56 °C. DNA was then purified in accordance with the Qiagen protocol for purification of total DNA from animal tissues. DNA concentration was then measured on the nanodrop and samples were diluted to be a final concentration of 2ng/uL. HK2 and ND1 primers were used for quantification. All samples were run in duplicate for each primer. Master mixes were made containing forward and reverse primers as well as nuclease-free water and Quanta Bio Perfecta SYBR green fast mix. Mitochondrial DNA was then measured by qPCR using an Applied Biosciences 7300 real-time PCR system (Waltham, MA).

Optomotor response test of visual acuity

Visual acuity was assessed using Optokinetic Nystagmus/Optomotor response (OKN/OMR). For female mice, animals were placed unrestrained on an elevated platform centered among four adjoining LCD screens (OptoMotry; Cerebral Mechanics Inc., Lethbridge, AB, Canada). Spatial frequency thresholds were measured by assessing the oculomotor response to drifting sinusoidal gratings at 100% contrast. Grating spatial frequency (cycles/degree) was systematically adjusted based on the optomotor response noted by a naïve experimenter. Mice were tested at least twice to determine baseline spatial acuity. For male mice, visual acuity was assessed as described⁴⁶. Mice underwent testing using the Optodrum system (Striatech Inc., Germany). Four surrounding monitors (brightness: 250 cd/m²) displayed a high-contrast (99.72%) sinusoidal grating pattern rotating at a fixed speed of 12°/s. The pattern’s resolution was systematically reduced to determine the threshold where mice failed to exhibit optokinetic nystagmus (OKN), indicating the limit of their visual resolution. The Optodrum system automatically tracked head movements, enabling quantitative OKN response analysis. The resolution threshold where OKN ceased was used to calculate visual acuity (cycles/degree) for each eye independently. This distinction was possible because of the differential sensitivity of each eye to pattern rotation direction (right eye more sensitive to counterclockwise, left eye to clockwise). To compare measurements in female and male animals, we normalized sGCα1^−/− data with respect to WT controls for each group.

Electron microscopy

Freshly dissected optic nerves were fixed with 2% glutaraldehyde and 2% paraformaldehyde in 0.1 M sodium cacodylate buffer for 4 h at room temperature. Tissue samples were sequentially post-fix stained with 1% tannic acid, 1% osmium tetroxide, and en bloc 1% uranyl acetate for one hour each. Samples were dehydrated in a graded ethanol series and infiltrated with Epon-812 resin using propylene oxide as the transition solvent. Samples were polymerized at 60 °C oven for 48 h and sectioned on a Leica ultramicrotome at 70 nm nominal thickness and collected onto 300 mesh Ni grids. Samples were examined on a JEOL 2100 Plus transmission electron microscope operated at 200 KeV equipped with an AMT Nanosprint 15 CMOS camera. Automated data collection was performed using SerialEM and processed using the IMOD software suite. Mitochondrial morphology was analyzed utilizing a custom Python segmentation neural network trained on over 2500 images. The resulting predicted segmentation edges were smoothed using a Gaussian filter with radius 10.0 and then binarized with a constant threshold. Mitochondria area, perimeter, length, and width were measured from the contour of the segmentation. All mitochondria segmentation measurements were done using the ImageJ Batch Process tools.

Retinal and optic nerve oxygen consumption rate

Oxygen consumption rate (OCR) was measured in fresh isolated retina and optic nerve tissue using a Seahorse XFe24 Analyzer (Agilent, Santa Clara CA, USA) and a modified version of the mitochondrial stress test assay. Each Seahorse XFe24 Sensor Cartilage (Cat # 103518-100; Agilent, Santa Clara CA, USA) was hydrated in a calibrant solution then placed in a 37 °C CO2-free incubator overnight the day before the assay. Media (DMEM; Cat # 103575-100; Agilent, Santa Clara CA, USA) was prepared on the day of the assay and supplemented with 25 mM glucose, 4 mM glutamine, and 0.5 mM pyruvate (Cat #s 103577-100, 103579-100, and 103578-100, respectively; Agilent, Santa Clara CA, USA) for retinal tissue or 0.6 mM glucose, 0.5 mM glutamine, and 0.12 mM pyruvate (Cat #s 103577-100, 103579-100, and 103578-100, respectively; Agilent, Santa Clara CA, USA) for the optic nerve tissue. Islet mesh inserts (Cat # 103518-100; Agilent, Santa Clara CA, USA) were hydrated with a chicken clot to aid in tissue adhesion using 15uL chicken plasma (Cat #P3266; Millipore Sigma, Burlington MA, USA) and 10uL Thrombin Bovine (Cat #605157; Millipore Sigma, Burlington MA, USA). Injection materials were prepared and loaded in two ports: Port A contained 75uL of 50 μm BAM15 (Cat # SML1760; Millipore Sigma, Burlington MA, USA) and Port B contained 75uL of 50 μm Rotenone/Antimycin A (Cat# R8875/A8674; Millipore Sigma, Burlington MA, USA). Note that optic nerve and retinal injection compounds were prepared using the media that corresponds to the tissue type. Injection materials were fed to tissue in order A to B. Oligomycin was omitted due to known detrimental effects of the inhibition of ATP synthesis on retinal tissue⁴⁷. The senor cartilage was placed in the Seahorse XFe24 to run an automatic calibration and equilibration.

Mice were anesthetized with 2.5% isoflurane, then cervically dislocated. Left eyes were enucleated and carefully dissected. Retinas were removed whole and 1–2 1 mm biopsy punches were made per retina (Cat #10185-238; VWR, Radnor, PA, USA). Biopsy punches were placed on the mesh islet inserts then secured into the Islet Capture Microplates (Cat # 103518-100; Agilent, Santa Clara CA, USA). One whole optic nerve was cut into 4–5 equal parts per eye, lined up on the center of the mesh islet insert then secured into the Islet Capture Microplate. A total volume of 500uL of the appropriate prewarmed media was added to each well. The microplate was covered and placed in a 37° C CO₂-free incubator for 5 min to allow for tissue temperature to warm to the temperature of the Seahorse.

The Seahorse Mitochondrial Stress Assay protocol consisted of three baseline measurements, four measurements after the injection of BAM15, and seven measurements after the injection of Rotenone/Antimycin A. Each measurement includes mix (3 min), wait (2 min) then measure (3 min). Total run time, including the equilibration (12 min), was 02:04:00. After the assay, mesh inserts were removed from the plate and placed in a 50uL of Halt protease inhibitor (Cat #78430; ThermoFisher Scientific, Waltham, MA, USA) to prepare for BCA (Pierce) Assay (Cat #23225; ThermoFisher Scientific, Waltham, MA, USA). Results from the OCR assay were normalized to total ug protein per sample.

Statistics

All data are presented as mean ± standard deviation (S.E.M) unless otherwise stated. Graphs were made and statistical analyses were performed using GraphPad Prism version 9.0.0 (GraphPad Software, San Diego, California USA). Parametric statistics were performed (t-tests, ANOVAs) if data passed normality and equal variance tests; otherwise, we performed non-parametric statistics. Post hoc p-values less than 0.05 were considered statistically significant. Number of measurements and specific p-values are indicated in results or figure legends. Post-hoc power analyses for all results are shown in Supplemental File 1.

Results

sGCα1^−/− mice express mutant sGC protein that is insensitive to NO in the retina

To confirm that sGCα1^−/− mice indeed express mutant sGC protein that is insensitive to NO, i.e., unable to produce elevated cGMP levels in the presence of NO, we carried out sGC activity assays and Western blotting of WT and mutant sGC protein from retinal samples (Supplemental Fig. 1). We detected cGMP in the retina across all animal groups (Supplemental Fig. 1 A). Interestingly, female WT animals had a slightly higher baseline level of cGMP than WT males (^$p = 0.04). Addition of an NO-releasing compound (PAPA NONOate) resulted in a 147% and 126% increase in cGMP levels in WT male and female animals, respectively (Supplemental Fig. 1 A; ****p < 0.0001). However, addition of PAPA NONOate failed to increase cGMP levels in sGCα1^−/− mice of either sex. Our results confirm NO-insensitivity of sGC in the retina of both male and female sGCα1^−/− mice.

Fig. 1.

Female sGCα1^−/− mice lose visual acuity and exhibit ATP-dependent axonal transport deficits and RGC loss with age. (A) Visual acuity in aged female sGCα1^−/− mice is reduced compared to age-matched WT mice (*p = 0.01; additional data added are shown by triangle symbols). All data have been normalized, and re-published with permission²⁴; data shown as means ± SEM; One-way ANOVA Tukey’s multiple comparisons test). (B) Female sGCα1^−/− animals develop deficits in anterograde transport of cholera toxin-B (CTB) to the superior colliculus (SC). Heatmap plots demonstrating degree and diversity of CTB transport deficits in sGCα1^−/− mice (Red = 100%, Blue = 0%) as well as representative coronal sections of SC where CTB is labeled in green (black directional arrows indicate anatomical directions: L = lateral, M = medial, P = posterior, A = anterior). White arrows indicate unique regions of sectoral deficit in sGCα1^−/− mice. From left to right, maps show increasing severity of transport deficits to the SC. (C) Quantification of CTB transport shows a significant reduction in aged female mice with age (*p = 0.02; One-way ANOVA Tukey’s multiple comparisons test; data plotted as means ± SEM). Each point represents a single lateralization of SC from mice with bilateral CTB injections. (D) RGC soma counts are reduced with age compared to WT (**p = 0.002; Mann-Whitney non-parametric t-test; data normalized and re-published with permission⁴.

To assess levels of sGC mutant and WT protein, we carried out whole retinal Western blotting across animal groups (Supplemental Fig. 1B, C). Using an antibody known to detect WT sGC (at 78 kDa) and sGCα1 mutant protein (at 59 kDa)³⁶, we confirmed expression of WT protein in male and female WT retina. In sGCα1^−/− retina however, the WT sGC band is absent at 78 kDa, and in its place a band slightly lower at 59 kDa, representing the mutant form of sGC. Protein loading controls (α-tubulin at 50 kDa) are indicated below.

Quantification of bands showed that levels of sGC WT protein were slightly higher in female mice compared to male mice (p > 0.05), which may account for the higher levels of cGMP observed in females in the presence of NO. The level of mutant sGC protein did not differ between sexes. Our results together indicate that sGCα1^−/−mice express mutant sGC protein in the retina, and that this protein is insensitive to NO-catalyzed cGMP production.

sGC deficiency leads to loss of visual acuity, reduced ATP-dependent axonal transport, and RGC Soma loss in aged female mice

We previously reported ~ 20% loss of RGCs in female sGCα1^−/− mice with age⁴ and reduced visual acuity (Fig. 1A)²⁴. To explore whether age-related changes in female animals were due to early RGC dysfunction, we measured anterograde cholera toxin B (CTB) transport from the retina to the superior colliculus (SC) in both young and aged female WT and sGCα1^−/− mice. CTB transport along RGC axons is ATP-dependent making it a good proxy for RGC function (Fig. 1B)^40,48. The ATP-dependent transport deficits observed in sGCα1^−/− mice varied from animal to animal; a range of severities across fluorescent SC maps are shown in Fig. 1B; complete anterograde transport to the SC from the retina is indicated in red, while absence of transport is shown in blue. We did not find functional deficits in young, female sGCα1^−/− mice (Fig. 1C; p > 0.05). With age however, RGC transport of CTB decreased to as low as 50% in female sGCα1^−/− mice indicating that loss of visual acuity is in part due to RGC dysfunction and some degeneration (Fig. 1C; *p = 0.02). Finally, as reported previously, RGC soma counts were reduced in aged female sGCα1^−/− mice (Fig. 1D; ~20%; **p = 0.03)⁴.

Male sGCα1^−/− mice do not exhibit loss of visual acuity or RGCs with age

While it is well established that female sGCα1^−/− mice exhibit degeneration of RGCs with increasing age^4,22,49, males were excluded from previous studies since male sGCα1^−/− mice develop hypertension; a potential confound for studies of glaucoma³⁶. We therefore measured visual acuity in aged male WT and sGCα1^−/− male animals (Fig. 2A). We found that visual acuity was not significantly different between genotypes with age (Fig. 2A; p > 0.05). Additionally, aged males had similar numbers of RGCs between genotypes (Fig. 2B; p > 0.05).

Fig. 2.

Visual acuity and RGC counts in aged male mice. (A) Normalized visual acuity in aged male sGCα1^−/− mice is comparable to age-matched WT mice (p > 0.05). (B) RGC counts in aged male animals were not different in sGCα1^−/− mice compared to WT (p > 0.05). Data normalized to respective WT control and presented as means ± SEM. Animal numbers represented in bars.

sGC knockout reduces retinal glucose uptake in females with age

To identify retinal gene expression changes that may drive RGC dysfunction and degeneration in female mice, we carried out whole retina RNA sequencing. A complete set of most significantly up- and down-regulated genes can be found in Supplemental Table 1. In exploring the most significant changes between sGCα1^−/− and WT mice, we noted that glucose transporter GLUT1(Slc2a1) was significantly upregulated in sGCα1^−/− mice compared to WT (log₂FC = 1.55, p_adj=7.31 × 10^{− 7}; Supplementary Table 1). To explore whether sGC knockout affected glucose uptake, we carried out ex vivo retinal glucose uptake assays of the fluorescent glucose analog 6-NBDG as previously described⁴⁵. In young animals, no significant difference in total retinal fluorescence was observed between WT and sGCα1^−/− mice of either sex although in young females there was a trend for increased uptake of 6-NBDG in sGCα1^−/− mice (p = 0.1143; Fig. 3). However, with age, male sGCα1^−/− mice exhibited elevated levels of 6-NBDG uptake (+ 52%; Fig. 3; *p = 0.020) while female sGCα1^−/− mice showed the opposite trend (−25%; Fig. 3; **p = 0.004). The increase in glucose uptake in aged male sGCα1^−/− mice is reminiscent of the increase observed in young female mice. Overall these results indicate that with age, 6-NBDG uptake into the retina is decreased in female sGCα1^−/− mice.

Fig. 3.

Retinal uptake of the glucose analog 6-NBDG is reduced in female sGCα1^−/− animals with age. In young animals, no statistically significant difference was observed in either group (males = blue bars, females = pink bars). However, in aged male animals, 6-NBDG uptake was increased (*p = 0.020) while in female animals it was reduced (**p = 0.004). Uptake was not significantly changed in young male (p = 0.905) or female (p = 0.1143) animals. Animal numbers indicated in bars. Due to number of variables and methodology, sGCα1^−/− data normalized to it’s respective WT control for each group (i.e., male young, female young, male aged, and female aged) and presented as means ± SEM. Statistical analyses = Mann-Whitney t-test.

Female sGCα1^−/− mice have decreased retinal GLUT1 expression with age

Since our RNA sequencing results pointed towards a decrease in Glut1 transcription in knockout retina and decreased glucose uptake, we next determined whether this change was reflected at the protein level in male vs. female animals. We measured whole retinal protein levels of GLUT1 using Western blot analysis and visualized localization of GLUT1 protein in retinal layers using immunohistochemistry (Fig. 4). Densitometry analysis and representative protein bands for each group are shown (Fig. 4A). In young animals, both male and female sGCα1^−/− show elevated GLUT1 protein levels in the retina compared to age-matched WT controls; the increase in GLUT levels in females was more striking (+ 13.7% and + 93.7%, respectively; Fig. 4A, B, and C). Elevation in protein levels is seen clearly throughout the retina of young males (Fig. 4B) and females (Fig. 4C) but most strongly in the ganglion cell layer (GCL), inner plexiform layer (IPL), outer plexiform layer (OPL) and to some extent in the outer nuclear layer (ONL). With age, male sGCα1^−/− animals still exhibited higher GLUT1 protein levels than WT animals (+ 34.4%; Fig. 4A, C). Elevation in protein predominantly localized to the OPL, ONL and inner/outer segments of the photoreceptor layer (Fig. 4B, C). With age, the trend in female animals reversed; GLUT1 levels were reduced in sGCα1^−/− animals compared to WT (−39.6%; Fig. 4C), an effect observed throughout the retina.

Fig. 4.

Knockout of sGC reduces GLUT1 expression in females with age. Western blot densitometry and representative protein bands from (A) young male, aged male, young female, and aged female animals (males = blue bars, females = pink bars). In young animals, both male and female sGCα1^−/− exhibited higher GLUT1 protein levels with respect to their WT controls (*p = 0.020, ***p = 0.007). With age, GLUT1 expression remained increased in male sGCα1^−/− animals (**p = 0.01) but decreased in female sGCα1^−/− animals (***p = 0.007). Representative retinal immunohistochemical samples are shown for each group (B; RNFL = retinal nerve fiber flayer, GCL = ganglion cell layer, IPL = inner plexiform layer, INL = inner nuclear layer, OPL = outer plexiform layer, ONL = outer nuclear layer, IO/OS = inner segments/outer segments of photoreceptors). Animal numbers indicated in bars. Due to number of variables and methodology, sGCα1^−/− data normalized to it’s respective WT control for each group (i.e., male young, female young, male aged, and female aged) and presented as means ± SEM. Statistical analyses = Mann-Whitney t-test.

Single cell transcriptomics highlight retinal cell co-expression of GUCY1A1 and SLC2A1

To understand the implications of decreased GLUT expression in the retina we investigated which retinal cell types co-express GLUT1 (Slc2a1) and sGC (Gucy1a1) using an open access single cell RNA sequencing dataset⁵⁰ (Fig. 5). Co-expression of Slc2a1and Gucy1a1 genes was most robust in amacrine cells, bipolar cells, endothelial cells, ganglion cells, and Müller glia. These data show that both genes are widely expressed across multiple cell classes in the retina including neuronal, glial, and vascular cells. Because of their widespread (and in most cases positively correlated) expression, dysregulation in the sGC-cGMP signaling axis likely has far reaching effects on retinal function.

Fig. 5.

Retinal cell type specific expression of sGCα1 and Glut1. Single cell RNA sequencing data from the Mouse Retinal Cell Atlas⁵⁰ shows cell type specific expression of GUCY1A1 (sGCα1) and SLC2A1 (Glut1). Co-expression is common and most robust in amacrine cells, bipolar cells, endothelial cells, ganglion cells, and Müller glia. Points in each subplot represent a paired, scaled, mean expression value of mRNA transcripts within individual retinal cells. Data were pre-processed and normalized by the authors of the original study prior to our download and curation. Datapoint color corresponds to broad cell classes; green represents neuronal, magenta represents glial, orange represents vascular and mural cells, and cyan pigment epithelium. Spearman correlation coefficients and associated P-values are shown for each cell type.

sGC knockout leads to increased nitrosative stress at the GCL in female mice with age

To evaluate levels and localization of nitrosative stress in the retina, we carried out whole retinal Western blotting and immunohistochemistry in retinal sections using a highly specific antibody for 3-nitrotyrosine (3-NT; Fig. 6)⁵¹. In young males and females, whole retinal levels of 3-NT were not significantly changed between WT and sGCα1^−/− animals (Fig. 6A). This was reflected in retinal sections where levels were minimal through the retinas of both genotypes and both sexes (Fig. 6B). In aged males, levels of 3-NT were reduced in sGCα1^−/− mice compared to WT (−26%; Fig. 6C, F; *p = 0.029). However, in aged females, total retinal levels of 3-NT were increased (+ 18%; Fig. 6C; *p = 0.024). The increase in 3-NT was primarily localized to the ganglion cell layer  (GCL) (Fig. 6D). To investigate whether the increase in 3-NT was due to elevated NO levels, we looked at levels of NADPH diaphorase (NQO1), an in situ indicator of NO^52,53, in aged animals (Fig. 6E). We saw an elevation in NQO1 localised to the GCL layer in aged female sGCα1^−/− mice when compared to WT; we did not see any differences between aged male genotypes.

Fig. 6.

Nitrosative stress is elevated in female sGCα1^−/− mice with age at the retinal ganglion cell layer. Western blot protein analysis of total 3-NT protein levels in (A) young male and female WT and sGCα1^−/− mice. There was no significant change in 3-NT levels in young male (p = 0.943) or female (p = 0.100) animals. (B) Representative confocal micrographs. (C) Aged males exhibited decreased 3-NT levels (*p = 0.029), while females had the opposite trend (*p = p = 0.024). (D) Representative confocal micrographs. Increased 3-NT was predominantly localized to the GCL layer (C; GCL = ganglion cell layer, IPL = inner plexiform layer, INL = inner nuclear layer, OPL = outer plexiform layer, ONL = outer nuclear layer, and IS/OS = inner/outer segments of photoreceptors). (E) Representative confocal micrographs of NQO1 staining in aged male and female mice. Due to number of variables and methodology, sGCα1^−/− Western blot data normalized to it’s respective WT control for each group (i.e., male young, female young, male aged, and female aged) and presented as means +/- S.E.M; n = 5–10 animals per group as indicated in bars. Statistical test = Mann-Whitney nonparametric t-test.

Mitochondrial protein expression is dysregulated in female sGCα1^−/− mice with age

The presence of nitrosative stress in aged female retina is suggestive of elevated levels of reactive oxygen species and NO⁵⁴. Nitrosative stress is a primary cause of mitochondrial dysfunction in a number of neurodegenerative diseases⁵⁵. In our whole retinal RNA sequencing analyses the most significantly changed transcripts in aged female sGCα1^−/− mice compared to WT were mitochondrial-related genes (Table 1). Amongst them were: Tomm20l (Log₂FC = 3.42, p_adj=2.9 × 10^{− 13}), encoding TOMM20L, a mitochondrial membrane-associated protein involved in protein targeting to the mitochondrion; Mt-nd3 (Log₂FC = 3.38, p_adj=4.8 × 10^{− 8}) encoding NDUFS3, a component of complex I of the respiratory chain; Mt-nd4 (Log₂FC = 3.06, p_adj=1.06 × 10^{− 5}) encoding NDUFS4, another component of complex I, Mss51 (Log₂FC = 1.24, p_adj=1.35 × 10^{− 15}) encoding MSS51 a mitochondrial translational activator protein, and the mitochondrial regulator Pyroxd2 (PYROXD2; Log₂FC=−1.90, p_adj=2.75 × 10^{− 9}; Table 1).

Table 1.

Bulk RNA sequencing in aged, female WT and sGCα1^−/− mouse retina indicates dysregulation of mitochondrial genes. Bulk RNA sequencing in retinal samples from aged sGCα1^−/− (n = 5) and WT (n = 5) mice. Log₂ fold change (Log₂FC) data comparing transcripts in sGCα1^−/− vs. age-matched WT animals.

Gene	Function	Log2FC	Padj value
Tomm20l	Mitochondrial outer membrane protein	3.41	2.9 × 10− 13
Mt-Nd3	Mitochondrial NADH dehydrogenase	3.38	4.77 × 10− 8
Mt-Nd4	NADH-ubiquinone oxidoreductase	3.06	1.06 × 10− 5
Mss51	Mitochondrial Translational Activator	1.24	1.35 × 10− 5
Pyroxd2	Putative mitochondrial regulator	−1.90	2.75 × 10− 9

To determine whether transcriptional changes were reflected at the protein level, we carried out whole retinal Western blotting in young and aged mice of both sexes (Fig. 7). Our results indicate that in young mice of both sexes, protein levels largely remain unchanged between genotypes (Figs. 7A-C; p > 0.05). The only exception was COXIV in young males where protein levels in sGCα1^−/− mice were higher compared to WT (*p = 0.01). In aged animals however, there were decreases in protein levels between WT and sGCα1^−/− animals in the female group for TOMM20, NDUFS3, and COXIV (Figs. 7A-C; **p < 0.01, *p < 0.04). These results were in contrary to the increases in mRNA observed in bulk sequencing data (Table 1). Although changes in aged male mice were not observed for TOMM20 and NDUFS3, there was a more pronounced decrease in COXIV levels in sGCα1^−/− mice compared to WT (Fig. 7C; *p < 0.04).

Fig. 7.

Mitochondrial protein expression is dysregulated in female sGCα1^−/− mice with age. Western blot densitometry and representative protein bands showing retinal levels of (A) TOMM20, (B) NDUFS3, and (C) COXIV proteins in young and aged WT and sGCα1^−/− mice. Data presented as means ± SEM. Number of animals per group as indicated in bars where possible (n = 8 WT and 7 sGCα1^−/− in B). Statistical tests = Mann-Whitney nonparametric t-test; p*≤0.04, **p ≤ 0.01, ns = p > 0.05. (D) Mitochondrial DNA quantitative PCR analysis comparing transcript levels of the mitochondrial-specific gene Mt-nd1 to nuclear-encoded gene Hk2. In young animals, mitochondrial number remains unchanged between WT and sGCα1^−/− mice of either sex (p > 0.05). The same trend was observed in aged animals (p > 0.05). Due to number of variables and methodology, sGCα1^−/− data normalized to it’s respective WT control for each group (i.e., male young, female young, male aged, and female aged) and presented as means ± SEM. Number of animals per group as indicated in bars. Statistical tests = Mann-Whitney nonparametric t-test.

Mitochondrial biogenesis is unaffected by sGC knockout

Mitochondrial biogenesis is enhanced by NO through the action of cGMP⁵⁶. To determine whether protein changes in sGCα1^−/− mice were due to changes in mitochondrial copy number, we carried out mitochondrial DNA quantitative PCR analysis in retina and compared transcript levels of the mitochondrial-specific gene Mt-nd1 to nuclear-encoded gene Hk2 (Fig. 7D). Our results show that in young and aged animals, mitochondrial number did not change between genotypes in either sex (Fig. 7D). These results confirm that changes in mitochondrial protein levels in aged females are not due to changes in the number of mitochondria, rather, a downregulation of mitochondrial genes and proteins themselves.

Mitochondrial size increases with age in male, but not female, sGCα1^−/− mice

To assess whether changes in mitochondrial protein expression accompanied changes in the overall size of mitochondria, we used transmission electron microscopy in cross-sections of optic nerve tissue; representative electron micrographs are shown in Fig. 8. In female animals of both ages, mitochondrial size remained the same between genotypes as indicated by the minimal shifts in groups in histograms (Fig. 8C, D, G, and H). In male mice, young animals exhibited a slight shift towards a smaller mitochondrial size; male sGCα1^−/− mice had higher frequencies of mitochondria with areas of 60–80 µm² (Fig. 8A and B). Interestingly, unlike aged female mice, in male mice there was a broader range of mitochondrial sizes in both WT and sGCα1^−/− mice (Fig. 8E, F), with a shift towards a higher percentage of mitochondria larger than 100 µm² in sGCα1^−/− mice compared to WT. When taking a closer look at mitochondrial structure we noticed a trend for a larger intermembrane space in sGCα1^−/− mice compared to their WT counterparts across all groups (Fig. 8; red arrows). In WT animals, the inner and outer membranes of mitochondria across all groups were tightly associated. This was less apparent in sGCα1^−/− animals, where space between the membranes was more apparent. This phenomenon was noted but not quantified further.

Fig. 8.

Mitochondria are larger in male sGCα1^−/− mice compared to WT in young and aged animals. Representative transmission electron micrographs of optic nerve mitochondria in (A) young male, (C) young female, (E) aged male and (G) aged female animals. Red arrows indicate larger intermembrane space. Scale bars = 500 nm. Individual mitochondrial area measurements were calculated using a custom macro and percentage of total mitochondria at each size were plotted in frequency histograms (B-H).

To further explore these data, all mitochondrial area measurements from each experimental group, along with averages of each animal were plotted (Fig. 9A). In male animals of both ages, sGCα1^−/− mice exhibited changes in mitochondrial area; in young animals, sGCα1^−/− mice had slightly smaller mitochondria, whereas in aged animals, mitochondria were significantly larger. No significant difference was found between genotypes in females at either age. Next, we plotted the number of mitochondria measured, normalized to the total area analyzed (Fig. 9B). No significant difference between the number of mitochondria analyzed were observed across ages or genotypes, although a slight trend was observed in aged male animals towards fewer mitochondria in sGCα1^−/− mice compared to WT. To explore whether an increase in mitochondrial area correlated with mitochondrial number, we compared these measures for each animal in male (Fig. 9C) and female mice (Fig. 9D). We next tested if the differences between linear regression slopes were different for any groups. The difference between the slopes of aged male WT and aged male sGCα1^−/− mice was significant (P = 0.01). Across all other groups (including sex, age, and genotype), no significant differences in slopes were found.

Fig. 9.

Mitochondrial area changes more dramatically with age in male sGCα1^−/− mice compared to WT. (A) Mitochondrial area measurements in young and aged, male and female WT and sGCα1^−/− mice. Total individual mitochondrial measurements are shown for each group; mouse averages are indicated in white circles. Means ± SEM are indicated in black error bars. *p = 0.02, ****p < 0.0001; Mann-Whitney non-parametric t-test. Statistics carried out on all mitochondrial measurements. (B) Number of mitochondria in each sample normalized to total area analyzed, all comparisons non-significant (p > 0.05). (C) Correlation scatter plot indicating mitochondrial average area and number of mitochondria for young and aged male WT and sGCα1^−/− mice. (D) Correlation scatter plot indicating mitochondrial average area and number of mitochondria for young and aged male WT and sGCα1^−/− mice. Lines indicate linear regressions.

Mitochondrial function is impaired in retina and nerve of aged female sGCα1^−/− mice

To investigate mitochondrial physiology in aged male and female mice we measured oxygen consumption rate (OCR) in the presence of mitochondrial an uncoupler (BAM15) and an inhibitor of respiration, rotenone and antimycin A (Rot + AA) in retina and ON (Fig. 10). In aged male retinas, there was no difference between OCR at baseline, or respiration rates after addition of BAM15 and Rot + AA (Fig. 10A). Furthermore, when comparing the slope of OCR decline after inhibition of respiration by Rot + AA, there was a prompt decrease in OCR that occurred at the same rate in both genotypes (WT slope = −0.107 vs. sGCα1^−/− =−0.101). In aged male ON samples, basal OCR in WT was higher compared to sGCα1^−/− and the slope of OCR at inhibition by Rot + AA was steeper in WT vs. sGCα1^−/− mice (WT slope = −0.03 vs. sGCα1^−/− =−0.01), suggesting that WT animals were more reliant on oxidative phosphorylation for ATP production.

Fig. 10.

Retinal and optic nerve oxidative phosphorylation is impaired in aged female sGCα1^−/− mice. Comparison of oxygen consumption rate (OCR) between WT and sGCα1^−/− retina and optic nerve groups in (A) males and (B) females. Dotted lines indicate injection of the uncoupler BAM15 or Rotenone + Antimycin A (Rot + AA). Data normalized to total µg protein in each sample and plotted as means of 4–10 biological replicates ± SEM. Statistical test = 2-way ANOVA; *p < 0.0001 and #p ≤ 0.03.

In aged female retinas, sGCα1^−/− retinal OCR was significantly lower than WT; baseline OCR was on average 41% lower than WT (WT = 1.98 vs. sGCα1^−/− = 1.16 pmol/min/µg). In addition, the slope of OCR at inhibition by Rot + AA was steeper in WT vs. sGCα1^−/− mice (WT slope = −0.09 vs. sGCα1^−/− =−0.04), again suggesting little reliance on ATP production by oxidative phosphorylation in aged female sGCα1^−/− mice. In the case of aged female sGCα1^−/− mice, taking into consideration a lower basal OCR, these data are suggestive of mitochondrial dysfunction. We also noted that the effect on Rot + AA on aged female WT retina was somewhat less pronounced than in aged male WT retina. In ON tissue, similarly to males, female sGCα1^−/− ON samples had a lower basal OCR compared to WT and the slope of OCR at inhibition by Rot + AA was steeper in WT vs. sGCα1^−/− mice (WT slope = −0.03 vs. sGCα1^−/− =−0.02), again suggesting that WT animals were more reliant on oxidative phosphorylation for ATP production.

Discussion

The NO-sGC-cGMP signaling axis is best known for its role in blood vessel dilation, however, novel neuroprotective roles are evident^5,10–12,14. In our previous work, we showed that female sGCα1^−/− mice that lack the NO-binding domain of sGC, develop RGC degeneration with age^4,24,33,49. Degeneration was prevented by the preservation of cGMP pools by supplementation with tadalafil, a PDE5 inhibitor⁴. PDE5 inhibition also prevented RGC degeneration in the microbead model of elevated intraocular pressure (IOP), and in ex vivo retinal explants suggesting that cGMP-mediated neurodegeneration occurred independently of changes in IOP⁴. Current therapies for glaucoma focus on reducing IOP however, patients still progress to vision loss, therefore, neuroprotective strategies that act independently of IOP are warranted. Our findings in sGCα1^−/− mice suggest that targeting the cGMP pathway may fill this critical therapeutic need.

In previous work, male mice were excluded due to their development of hypertension which can be a confound in studies of glaucoma⁵⁷. Here we decided to explore how both sex and aging impact RGC degeneration in sGCα1^−/−mice. We first confirmed levels of sGC WT and mutant sGCα1^−/− protein and sensitivity of the proteins to NO by measuring retinal levels of cGMP in the presence and absence of NO (Supplemental Fig. 1). Since deletion of exon 6 in the sGCα1^−/− mice is not a null mutation, it was important to determine levels of cGMP in the retina of all animal groups and to confirm that deletion of the alpha-1 exon resulted in NO-insensitivity of the enzyme in retinal tissue. We found that in WT female retina, cGMP levels were slightly higher than in WT males (Supplemental Fig. 1 A; p = 0.04). This is intriguing and may suggest a higher dependence on cGMP in the female retina, higher levels of NO in female mice at baseline driving higher cGMP levels, or higher expression of the membrane-bound particulate guanylate cyclase in female animals, an enzyme that also produces cGMP in the retina; a sex difference in expression levels of particulate guanylate cyclase has been observed in human and mouse brain⁵⁸. Furthermore, our results show that NO addition leads to a much greater increase in the levels of NO in female WT mice as compared with males. This result, paired with a tendency for female WT animals to express higher levels of sGC protein, suggests a greater level of expression and activity of sGC in the retina of female WT animals. Since we observe greater pathology in the absence of the NO-sensitive enzyme in female sGCα1^−/− animals, these results may point to a greater dependence on the sGC enzyme in females vs. males.

We found that visual acuity deficits and RGC degeneration are not observed in male sGCα1^−/− mice with age and that ~ 20% RGC loss in female mice is coincident with an even bigger reduction in ATP-dependent axonal transport (Fig. 1). These initial findings suggest that dysfunction within the cGMP pathway differentially affects neuronal survival in a sex-dependent manner. This was intriguing, since neurodegenerations of the CNS are more predominant in females with age and in patients, a decrease in cGMP is correlated with cognitive decline in both Alzheimer’s disease and glaucoma^25–27. We next sought to interrogate the mechanisms by which a lack of sGC promoted RGC degeneration in aging female animals.

A reduction in ATP-dependent axonal transport in females with age is indicative of issues with cellular energy production. Dysfunctional ATP generation is common across all CNS neurodegenerative diseases and can occur at critical steps; transport of energetic resources such as glucose into cells can fail and downstream issues with mitochondrial respiration or glycolytic pathways can also occur^59–62. Our bulk RNA sequencing identified an increase in gene expression of the glucose transporter-1 (GLUT1; Slc2a1), in aged female sGCα1^−/− retina. GLUT1 is most highly expressed on vascular endothelial cells where it facilitates the movement of glucose from the blood to surrounding cells and tissue, however, it is also expressed on neurons and astrocytes^63–65. In the retina and CNS, astrocytes mediate basal glucose uptake from the vasculature through GLUT1^63,66. Although RGCs can uptake glucose directly via the expression of GLUT3, RGCs distant from the vasculature depend on astrocytic delivery of energetic resources, including astrocytic lactate made from glucose^67,68.

Since GLUT1-mediated uptake of glucose is necessary for retinal function, we next measured the uptake of glucose in live retina using a novel ex vivo method⁴⁵. Our results showed that only females exhibited a reduction in glucose uptake with age, whereas males showed the opposite trend (Fig. 3). We confirmed that reduced glucose uptake was correlated with an overall reduction of GLUT1 protein expression throughout the retina in aging female sGCα1^−/− retina (Fig. 4). In males, we observed an increase in GLUT1 levels spanning the retinal layers with age. These results were striking, since GLUT1 protein levels are also diminished in the brains of patients with CNS neurodegenerative diseases such as Huntington’s and Alzheimer’s disease^69–71. Furthermore, pharmacologic stimulation of sGC increased GLUT1 expression in rat brain⁷². Our results suggest then that functional cGMP signaling is important for GLUT1 expression in female retina.

To better understand the cellular co-expression of GLUT1and sGC in the retina we mined an existing open-access single-cell RNA sequencing dataset⁵⁰. Surprisingly, co-expression was most robust in amacrine cells, bipolar cells, endothelial cells, ganglion cells, and Müller glia (Fig. 5). Retinal immunohistochemistry has also highlighted the expression of sGC in these cells types in mice, rats, and humans^33,73,74. Widespread expression of sGC in the retina suggests a potential role of cGMP in more than just regulating vascular tone. Co-expression of sGC with GLUT1 may point to a coordinated role between the two proteins in the retina. Outside of the CNS, cGMP has been implicated in the regulation of GLUT1 expression and its translocation⁷⁵. When GLUT1 is produced in cells, it resides in the cytosol until it is needed at the cytoplasmic membrane⁷⁶. GLUT1 can also be trafficked to the mitochondrial membrane where it serves as a transporter for dehydroascorbic acid (DHA), the precursor for ascorbic acid⁷⁷. GLUT1 facilitates the movement of DHA into mitochondria where it is reduced to ascorbic acid, a potent antioxidant. It is possible that dysfunctional cGMP signalling may impact not only expression but mitochondrial localization of GLUT1, reducing its antioxidant capacity in female mice. Further investigation into the role of cGMP in GLUT1 antioxidant capacity is necessary. Our results suggest that age-, and sex-related changes in cGMP signaling decrease the expression of GLUT1 in the retina, which may have a direct influence on RGC survival through diminished glucose-mediated energy metabolism.

We hypothesized that a decrease in retinal glucose uptake in aged female sGCα1^−/− mice might lead to an increase in superoxide (O₂⁻) production due to reduced oxidative phosphorylation⁷⁶. Excessive NO (due to lack of binding by sGC in knockout animals) along with increased O₂⁻ can lead to the formation of potentially toxic levels of peroxynitrite (ONOO⁻), which can form rapidly in tissue⁷⁸. This reaction not only limits NO bioavailability for physiological signal transduction but drives nitrooxidative processes, including the nitrosylation of proteins⁷⁹. One of the footprints of peroxynitrite is the formation of protein 3-nitrotyrosine (3-NT), an oxidative posttranslational modification that leads to mitochondrial dysfunction and apoptosis⁷⁹. With age, only female sGCα1^−/− mice exhibited high levels of 3-NT localized to the RGC layer; in aged male sGCα1^−/− mice this was absent (Fig. 6). In fact, we detected a decrease in 3-NT proteins in aged male sGCα1^−/− mice compared to the increase observed in females. Although the reason for a decrease in 3-NT levels in males is not yet clear, it is possible that male animals do not exhibit the same production of O₂⁻ as females. Expression of O₂⁻ detoxification enzymes such as superoxide dismutase (SOD) can also be influenced by sex and levels of NO^80–83. These results suggest that mitochondrial function is affected by sGCα1 knockout in females, but not in males. We next sought to determine whether this was the case.

During oxidative phosphorylation in mitochondria, a small number of electrons leak from the electron transport chain and react with oxygen, producing O₂⁻⁸⁴. Although a small amount of O₂⁻ production is normal during homeostasis, with aging or when mitochondria are dysfunctional, the amount produced can increase significantly^85,86. To explore whether mitochondria were dysfunctional with age, we first investigated mitochondrial protein expression. Our bulk retinal RNA sequencing analysis in aged female animals highlighted changes in mitochondrial gene transcription (Table 1). Specifically, an up-regulation of TOMM20, NDUFS3, and COXIV at the transcript level, however, we show data that suggest that mitochondrial proteins TOMM20, NDUFS3, and COXIV are reduced in aged, female sGCα1^−/− mice compared to WT (Fig. 7). This could be a response to reduced protein levels by the cells, or an issue with translation of the mRNA into protein. Interestingly, high levels of NO can cause a ribotoxic stress response (RSR) in cells; NO can inhibit protein translation by inducing ribosome stalling and collision^87,88. When we explored our bulk RNA sequencing dataset for evidence of ribosomal gene changes (Supplementary Table 1), the putative ribosomal protein RPL13a was significantly upregulated (Log₂FC = 2.12; P_adj=0.0008). It would be interesting to investigate the role of dysfunctional sGC in ribosomal function and its role in neuronal survival with age as a potential downstream mechanism of neuronal degeneration.

In contrast to females, in male animals there was no significant change in protein levels except for a reduction in levels of COXIV. The enzyme COXIV is a cytochrome c oxidase contains two heme groups to which NO can bind⁸⁹. When NO levels increase, such as we hypothesize is the case in sGCα1^−/− mice (male or female) due to their lack of NO receptor (sGC), NO becomes a potent inhibitor of COXIV, which can in turn affect expression of the protein⁸⁹. To determine whether the changes in protein levels were due to a reduction in overall mitochondrial number we quantified mitochondria copy number. We saw no changes in the number of mitochondria between animal groups, suggesting that mitochondrial protein expression is reduced by sGC knockout in female retina.

Using electron microscopy, we found that with age with age, male sGCα1^−/− mice may exhibit an increase in mitochondrial area compared to WT, whereas females did not show this trend (Figs. 8 and 9). Since our sample sizes in this study are low, these analyses must be repeated to confirm any changes in mitochondrial size. In muscle and brain, mitochondrial area increases with age, although the physiological significance of this change is not well understood^86,90. Since female sGCα1^−/− mice exhibit RGC degeneration with age, the increase in mitochondrial area observed in male sGCα1^−/− mice may represent a beneficial change in size that may promote neuronal function and survival. Further studies to better examine mitochondrial morphology (for example using 3-dimensional EM to resolve intact mitochondrial volumes⁹¹, as well as mitochondrial functions that can affect morphology and physiology such as mitophagy and fusion/fission are warranted.

Since we saw a lack of nitrosative stress in male sGCα1^−/− retina, and mitochondrial protein expression was largely unchanged, we hypothesized that sGC knockout impacted mitochondrial metabolism in female, but not male retina. After measurements of OCR in aged animals, we observed a pronounced reduction in mitochondrial baseline OCR and inhibition by Rot + AA in female, but not male retina (Fig. 10). Interestingly results from male and female sGCα1^−/− ONs showed lower OCR than WT, suggesting sGCα1^−/− mitochondria in the ON are less reliant on oxidative phosphorylation for energy production.

Our physiological results, along with our reports of possible changes in mitochondrial protein expression and levels of oxidative stress suggest that there is mitochondrial dysfunction in female sGCα1^−/− mice with age that is not present in male animals at that time point. We also noted that when comparing WT retina between the sexes, female mice were less reliant on oxidative phosphorylation for ATP production than males (as indicated by the slope at inhibition by Rot + AA; Fig. 10). Sex differences in oxygen consumption and metabolism that occur with age in humans are apparent; in the brain the cerebral metabolic rate of oxygen consumption is higher in males with age than in females⁹² and sex hormones including estrogen can directly enhance mitochondrial function⁹³. In Alzheimer’s disease, metabolic changes are detected at early stages of the disease with impaired energy metabolism preceding cognitive impairment, indicating that mitochondrial dysfunction may be a formative factor in progression^94,95. Furthermore, Alzheimer’s patients exhibit abnormally low glucose metabolism in the brain that correlates with disease severity^96–98. Our results suggest that cGMP signaling may play an important role in GLUT-mediated metabolism, particularly in RGCs and that sex may play an important role in modulating this pathway in females.

When looking at the findings in this study we noticed an intriguing trend in results from retinal glucose uptake, GLUT1 protein expression, and nitrosative stress levels. Aged male sGCα1^−/− mice often had results that mimicked young female sGCα1^−/− mice. This led us to wonder whether males might be progressing at a slower rate with RGC degeneration than female mice. This is an important question that we plan to address in a current study with a longer timeline and increased number of animals. It is important to note that due to the large number of experimental outcomes and large amount of tissue that these outcomes necessitate, some of our experimental groups may be underpowered, as such we show all data and individual datapoints for each animal analysed. A post-hoc power analysis of all results is included in the supplementary datafile. Finally, the results in this manuscript come from a global sGCα1 knockout. In future studies, targeted knockout of sGC α1 using a Cre-LoxP conditional knockout in RGCs would be valuable in delineating the role of sGC in RGCs specifically.

Taken together our results suggest that NO-sGC-cGMP signalling is an important signalling axis for RGC metabolism and survival with increasing age in female mice. The biggest question that has arisen from our findings is why male RGCs are seemingly ‘protected’ from the effects of sGC depletion. Since male mice exhibit hypertension, we hypothesized that there would be an increase in degeneration compared to females; hypertension is associated with incidence and progression of glaucoma and is also a risk factor for hypertension-associated retinopathy⁹⁹. The primary discovery from this study is that female RGC survival with age is more reliant on NO-cGMP signalling than males, an interesting next step in this research will be to determine how sex hormones interplay with sGC to promote RGC survival with age. Estrogen is known to be neuroprotective in the CNS; the association of estrogen decline in menopause and its relationship to glaucoma onset is gaining traction in the literature as sex-based investigations are increasing^100–103. Future studies investigating the mechanism of the potential neuroprotective role of sGC in both sexes could have far-reaching benefits for therapeutic strategies aimed at preventing neurodegeneration in female-dominated CNS diseases.

Conclusions

In summary, our findings uncover a sex-specific vulnerability to RGC degeneration driven by disrupted cGMP signaling and altered retinal metabolism. Dysregulation of the cGMP signaling pathway has been implicated in the pathogenesis of several neurodegenerative diseases of the CNS, including Alzheimer’s disease^17–19, Huntington’s disease^20,21, and glaucoma^4,22–24. Interestingly, these diseases also predominantly occur in females. Our results therefore suggest that cGMP plays a neuroprotective role in the aging female CNS. Our study also highlights an important question that warrants further investigation; how do female sex hormones interplay with cGMP to promote survival of RGCs with age. These results establish a mechanistic link between cGMP signaling and retinal neurodegeneration with age in females and underscore the potential of targeting the sGC-cGMP signaling pathway as a therapeutic strategy for glaucoma and other CNS neurodegenerative diseases.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

LKW designed the research; LKW, OLB, KLC, LAS, SW, RK, and EA performed research; LKW, OB, GOBG, JMH and SM analyzed data; TSR and EDK consulted on experimental design and protocols; LKW wrote the manuscript and all authors reviewed and edited the manuscript.

Funding

Support provided by National Institutes of Health grant EY036002 to LKW, and an RPB Unrestricted grant to the VEI. Imaging and transmission electron microscopy supported through the Vanderbilt University Medical Center Cell Imaging Shared Resource core facility and NIH grants CA68485, DK20593, DK58404, DK59637, EY08126, S10OD034315, and R24OD037694. Additional funding to support imaging was supplied by the Vanderbilt Vision Research Center Core Grant 5P30EY008126-38.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. RNA sequencing data published in this manuscript is available online at https://doi.org/10.7910/DVN/WMN7EL.

Declarations

Competing interests

The authors declare no competing interests.