Figures and data
Rod photoreceptor-specific knockout of glutaminase (GLS) displays rapid retinal degeneration and increased markers of cell death
(A) Western blot analysis showing decreased GLS protein levels in the retina of WT and cKO mice at post-natal day 14 (P14). Quantitation of Western blot results for N=3-4 animals per group. (B) Representative images for immunofluorescence of P14 mouse retinas (N=3 animals per group) stained for GLS (red), cone opsin (green) and nuclei (DAPI, blue) in WT and cKO mice. White arrows indicate remaining GLS expression in cone photoreceptors. Left and middle panel scale bars are 40 µm. Right image scale bar is 20 µm. (C) OCT images detailing outer retinal changes in cKO mice over time compared to WT. Retinal structures are comparable to WT mice at P14 but rapidly thin with age. (D) Total retinal thickness, (E) ONL thickness and (F) IS/OS thickness as determined by OCT in WT and cKO mice over time. N=4-9 eyes per group. (G) Representative hematoxylin and eosin stained retinal sections from rod photoreceptor-specific Gls conditional knockout (cKO) mice compared to wild-type (WT) mice at P14, P21 and P42. ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer. Scale bar is 40 µm. H) ONL cell counts as a percent of WT retinas at P14, P21 and P42. N=5-10 eyes per group. (I) Representative images of WT and cKO retinas stained to detect TUNEL-positive cells (green) at P21. Scale bar is 40 µm. N=3-5 animals per group. (J) Quantitation of percent TUNEL-positive cells at P21 showing an increase in TUNEL-positive cells in cKO animals. N=3-5 animals per group. (K) qRT-PCR of genes related to cell death pathways including apoptosis (Apop.), necroptosis (Nec.), autophagy (Aut.) and ferroptosis (Ferrop) in WT and cKO mice at P14. N=6 animals per group. Statistical differences in (A), (D), (E), (F), (H), (J) and (K) are based on an unpaired two-tailed Student’s t-test where *P <0.05, **P<0.01 and ***P<0.001. Data are presented as mean ± standard error of the mean. OCT: optical coherence tomography, OS: outer segment, IS: inner segment, ONL: outer nuclear layer, INL: inner nuclear layer, GCL: ganglion cell layer, TUNEL: terminal deoxynucleotidyl transferase dUTP nick and labeling.
Loss of Gls in rod photoreceptors impairs retinal function
(A) Representative scotopic and photopic electroretinography (ERG) tracings for the rod photoreceptor-specific Gls conditional knockout mice (cKO) compared wild-type (WT) mice at P21 and P42. (B) ERG scotopic a-wave and (C) b-wave amplitudes in cKO mice compared to WT mice at P21 and P42. A flash intensity of 32 cd*s/m2 was utilized. N=3-6 animals per group. (D) ERG photopic b-wave amplitudes in cKO mice compared to WT mice at P21 and P42. A flash intensity of 100 cd*s/m2 was used. N=3-6 animals per group. (E) Representative images from staining of the cone-specific marker peanut agglutinin (PNA, green) and nuclei (DAPI, blue) in retinal sections from cKO mice compared to WT mice at P14, P21 and P42. Scale bars are 40 µm. Statistical differences in (B), (C) and (D) are based on an unpaired two-tailed Student’s t-test where *P <0.05, **P<0.01 and ***P<0.001. Data are presented as mean ± standard error of the mean. ONL, outer nuclear layer; IS, inner segments; OS, outer segments. N=3 per group.
Rod photoreceptors require GLS for maintenance and maturation
Glsfl/fl mice carrying a tamoxifen (TAM)-inducible Cre-recombinase (IND-cKO) under the control of Pde6g (Glsfl/fl;Pde6g-CreERT2) compared to mice expressing only the inducible Cre-recombinase (Glswt/wt;Pde6g-CreERT2; WT). Both IND-cKO and WT were administered tamoxifen for 5 days starting at P22. (A) Quantitation of Western blot results showing decreased GLS protein levels in the retina of IND-cKO animals at 10 days after tamoxifen induction compared to WT mice. HSP90 was used as a loading control. N=3-4 animals per group. (B) Representative GLS immunofluorescence (red) in IND-cKO mice compared to the wild-type (WT) mouse 10 days after tamoxifen induction. N=3 animals per group. Scale bar is 40 µm. (C) OCT images detailing outer retinal changes in WT and IND-cKO animals over time. Total retinal thickness (D), outer nuclear layer (ONL) thickness (E) and inner segment/outer segment (IS/OS) thickness (F) as determined by OCT for 38 days post tamoxifen. N= 5-10 eyes per group. (G) Representative hematoxylin and eosin-stained retinal sections from IND-cKO mice compared to WT mice at 10 and 38 days after tamoxifen induction. N=3 animals per group. Scale bar is 40 µm. ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer. Quantitation of scotopic (H) and photopic (I) ERG a- and b-waves in WT and IND-cKO retina 10 days post TAM. N>12 eyes per genotype. Statistical differences in (A), (D), (E), (F), (H), and (I) are based on an unpaired two-tailed Student’s t-test where *P<0.05, **P<0.01 and ***P<0.001. Data are presented as mean ± standard error of the mean.
GLS cKO mice maintain levels of nucleotide, glycolytic and TCA cycle metabolites and mitochondrial function but demonstrate altered redox balance
(A) Schematic summarizing the biosynthetic and bioenergetic roles of glutamine. (B) Relative abundance of key intermediates in nucleotide metabolism in the retina of WT and cKO mice at P14 as determined by targeted metabolomics. N=6-7 animals per group. Relative abundance is the ion intensity normalized to the WT. (C) The NADP+/NADPH ratio, as determined by bioluminescence assay, is significantly increased in the cKO as compared to WT retina at P14. N=5-6 animals per group. (D) Relative abundance of GSSG in the retina of WT and cKO mice at P14, prior to PR degeneration, as determined by targeted metabolomics. N=6-7 animals per group. (E) qRT-PCR of genes related to redox homeostasis are significantly altered in cKO compared to WT mice. N=6 animals per group. (F) Relative abundance of metabolites in glycolysis and the TCA cycle in WT and cKO retina at P14. N=6-7 animals per group. (G) Mitochondrial stress test carried out on isolated WT and cKO retina at P14 using the BaroFuse. The baseline was established by perifusing the tissue for 90 min and then oligomycin, FCCP, and KCN were injected into the perifusate sequentially as indicated. (H) Comparison of the effects of oligomycin and FCCP on OCR in P14 WT and cKO retina. N=6-8 animals per group. (I) Western blot analysis and quantitation of the mitochondrial electron transport chain complexes show no differences between WT and cKO retina. N=3-4 animals per group. Fold change is in relation to WT. Statistical differences in (B-F), (H) and (I) are based on an unpaired two-tailed Student’s t-test where *P<0.05. Data are presented as mean ± standard error of the mean. R5P: ribose 5-phosphate, IMP: inosine monophosphate, UMP: uridine monophosphate, GSSG: glutathione disulfide, F6P: fructose 6-phosphate, DHAP: dihydroxyacetone phosphate, 2PG: 2-phospho-D-glycerate, PEP: phosphoenolpyruvate, Pyr: pyruvate, Lac: lactate, Cit: citrate, α-KG: alpha-ketoglutarate, Suc: succinate, Mal: malate, Oligo: oligomycin, FCCP: carbonyl cyanide p-trifluoromethoxyphenylhydrazone, KCN: potassium cyanide, CI-NDUFB8: complex 1, NADH:ubiquinone oxidoreductase subunit B8, CII-SDHB: complex 2, succinate dehydrogenase complex iron sulfur subunit B, CIII-UQCRC2: complex 3, ubiquinol-cytochrome c reductase core protein 2, CV-ATP5A: complex 5, ATP synthase F1 subunit alpha, HSP90: heat shock protein 90.
Loss of GLS in rod photoreceptors has significant effects on TCA cycle metabolism with only partial rescue upon α-KG supplementation
(A) Schematic summarizing 13C6-glucose labeling in glycolytic and TCA cycle intermediates. (B) Fractional labeling of glycolytic and TCA cycle metabolites in the retina following intraperitoneal injection of 13C6-glucose in WT and cKO mice at P14. N=5-6 animals per group. (C) Schematic summarizing 13C5-Gln labelling in the TCA cycle. (D) Fractional labeling of TCA cycle metabolites in the retina following intraperitoneal injection of 13C5-Gln in WT and cKO mice at P14. N=6-11 animals per group. (E) ONL thickness in cKO mice at P22 as assessed by OCT following α-KG supplementation (10 mg/mL) or vehicle (water) in the drinking water from P4-P22. N=5 animals per group. Statistical differences in (B), (D) and (E) are based on an unpaired two-tailed Student’s t-test where *P <0.05, **P<0.01 and ***P<0.001. Data are presented as mean ± standard error of the mean. F6P: fructose 6-phosphate, FBP: fructose 1,6-bisphosphate, DHAP: dihydroxyacetone phosphate, 2PG: 2-phosphoglycolate, PEP: phosphoenolpyruvate, Pyr: pyruvate, Lac: lactate, Gln: glutamine, Glu: glutamate, α-KG: alpha-ketoglutarate, Suc: succinate, Fum: fumarate, Mal: malate, OAA: oxaloacetate, Cit: citrate, Asp: aspartate.
GLS cKO retina demonstrates decreased nonessential amino acids, ISR activation, and decreased global protein synthesis
(A) Amino acids significantly altered in the Gls cKO mouse retina at P14. Relative abundance is the ion intensity relative to WT retina. N=6-7 animals per group. (B) Ratio of glutamine to glutamate in WT and cKO retina. N=6-7 animals per group. (C) Western blot of ISR proteins phospho-eIF1αS51 (p-eIF2α), total eIF2α and ATF4 in WT and cKO mice. N=3-4 animals per group. (D) Quantitation of Western blot in Panel C. (E) Western blot of protein puromycinylation in the WT and cKO mouse retina at P14 harvested 30 min after systemic puromycin administration. (F) Quantitation of puromycin incorporation in WT and cKO retina. N=4-5 animals per group. (G) ONL thickness at P21 in cKO mice as assessed by OCT following intraperitoneal injection of ISRIB (2.5 mg/kg) or vehicle (50% PEG 400, 43.4% saline, 6.6% DMSO) from P5-P21. N=3-6 animals per group. (H) ONL thickness at P21 in cKO mice as assessed by OCT following intraperitoneal injection of Asn (200 mg/kg) or vehicle (PBS) from P5-P21. N=3-5 animals per group. Statistical differences in (A), (B), (D), (F), (G), and (H) are based on an unpaired two-tailed Student’s t-test where *P <0.05, **P<0.01 and ***P<0.001. Data are presented as mean ± standard error of the mean. Gln: glutamine, Glu: glutamate, Asp: aspartate, ONL: outer nuclear layer.
GLS is the predominant isoform in the mouse retina and the rod photoreceptor-specific Gls conditional knockout mouse does not demonstrate compensatory upregulation of the Gls2 isoform
(A) qRT-PCR analysis of Gls and Gls2 in P14 mouse retina. Relative expression was analyzed by comparative threshold cycle (2^-ΔΔCT) method. Expression values were represented as fold change over Gls after normalization with β-actin. N=9 animals per group. (B) Representative GLS immunofluorescence in P21 WT animals. Scale bar is 40 µm. (C) Western blot analysis and quantitation of GLS in fractionated wild-type (WT) mouse retinas. TIM23 was used as a mitochondrial fraction marker and HSP90 as a cytosolic fraction marker. WR whole retina extract; C, cytosolic fraction; M, mitochondrial fraction. N=3 animals. (D) qRT-PCR analysis of Gls2 mRNA expression in the retina at P14 demonstrates no change between the rod photoreceptor-specific Gls conditional knockout mouse (cKO) and the wild-type (WT) mouse where both Gls alleles are present. Data are normalized to β-actin. N=5-6 animals per group. (E) Representative GLS2 immunofluorescence (red) images in cKO mice at P21 shows no increase in the expression of GLS2 as compared WT mice. Scale bar is 40 µm. Statistical differences in (A), (B) and (D) are based on an unpaired two-tailed Student’s t-test where **P<0.01 and ***P<0.001. Data are presented as mean ± standard error of the mean.
Gls conditional knockout mouse demonstrates similar degeneration under cyclic- and dark-reared conditions
(A) Representative OCT images from the rod photoreceptor-specific Gls conditional knockout (cKO) mouse compared to wild-type (WT) mice at P21 and P28 housed with 12-hour light/12-hour dark cycles (cyclic-reared) or dark-reared. The white vertical bar represents the thickness of the outer nuclear layer (ONL). (B) Quantitation of ONL thickness as determined by OCT in cKO mice compared to WT mice at P21 and P28 under the different housing conditions. N=4-6 animals per group. Statistical differences in (B) are based on a One-way ANOVA with Tukey’s post hoc test for multiple comparisons where *** P<0.001. Data are presented as mean ± standard error of the mean.
Histology confirms increased Müller glial cell activation in rod photoreceptor-specific Gls conditional knockout mice
Representative images from GFAP staining indicate increased Müller glial cell activation in cKO compared to WT mice at P21 and P42. Scale bar is 40 µm.
Rod photoreceptor-specific Gls conditional knockout mice have shorter outer segments
(A) Representative immunofluorescence images of rhodopsin (RHO) staining (green) in rod photoreceptor-specific Gls conditional knockout (cKO) mice compared to wild-type (WT) mice at P21. N=3 animals per group. Scale bar is 40 µm. (B) Representative transmission electron microscopy (TEM) images of retinal sections from cKO mice compared to WT mice at P21. Images taken at 250X magnification and scale bar is 5 µm. (C) TEM images of retinal sections from cKO mice compared to WT mice at P21 taken at 500X magnification and scale bar is 3 µm. (D) TEM images of retinal sections from cKO and WT mice at P21 taken at 8000X and 6000X magnification, respectively, to show ribbon synapses. Intact ribbon synapses in WT rods are shown in yellow boxes. A structurally intact ribbon synapse in cKO rods is denoted by the yellow arrow. Scale bars are 500 nm.
Rod photoreceptor-specific Gls knockout does not alter synaptic connectivity between photoreceptors and second-order neurons
(A) Representative retinal sections stained with wheat germ agglutinin (WGA) to label photoreceptor synaptic membranes and non-synaptic membranes in the rod photoreceptor-specific Gls conditional knockout (cKO) mice compared to wild-type (WT) mice at P14. OPL, outer plexiform layer; IPL, inner plexiform layer. N=3 animals per group. Scale bar is 40 µm. (B) Representative retinal sections stained with an antibody against Bassoon to label ribbon synapses of rods and cones in the OPL in cKO and WT mice at P14. N=3 animals per group. Scale bar is 40 µm. (C) Representative retinal sections from WT and cKO animals at P14, P21 and P42 are immunostained with antibodies against major cell-type specific markers (red): calretinin (amacrine cells), BRN3A (ganglion cells), CHX10 (bipolar cells). ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer. N=3 animals per group. Scale bars are 40 µm. (D) Inner retinal area as a percent of WT retinas at P14, P21 and P42. N=5-10 eyes per group. Statistical differences in (D) are based on an unpaired two-tailed Student’s t-test where **P<0.01. Data are presented as mean ± standard error of the mean.
Loss of Gls in rod photoreceptors alters the expression of retinal metabolism-related genes
qRT-PCR analysis of genes related to (A) glycolysis and pyruvate metabolism as well as (B) the TCA cycle in rod photoreceptor-specific Gls conditional knockout mice (cKO) compared to wild-type (WT) mice retina at P14. Expression values are represented as fold change over WT after normalization with β-actin. N=6 animals per group. Statistical differences in (A) and (B) are based on an unpaired two-tailed Student’s t-test where *P<0.05, **P<0.01 and ***P<0.001. Data are presented as mean ± standard error of the mean.
Loss of Gls in rod photoreceptors does not alter the expression of metabolism-related genes in eyecups
qRT-PCR analysis of genes related to (A) glycolysis and pyruvate metabolism as well as (B) the TCA cycle in rod photoreceptor-specific Gls conditional knockout mice (cKO) eyecups compared to wild-type (WT) mice eyecups at P14. Expression values are represented as fold change over WT after normalization with β-actin. N=3-4 eyecups per group. Statistical differences in (A) and (B) are based on an unpaired two-tailed Student’s t-test where **P<0.01. Data are presented as mean ± standard error of the mean.
Mass isotopologue distribution of 13C6-glucose in glycolysis and TCA cycle intermediates in rod photoreceptor-specific Gls conditional knockout (cKO) mice compared to wild-type (WT) mice
Fractional labeling of glycolytic and TCA cycle intermediates with uniformly-labeled, 13C6-glucose in retina from rod photoreceptor-specific Gls cKO mice compared to WT mice at P14. Statistical differences are based on an unpaired two-tailed Student’s t-test where *P<0.05 and **P<0.01. Data are presented as mean ± standard error of the mean.
Mass isotopologue distribution of 13C5-Gln in TCA cycle intermediates in rod photoreceptor-specific Gls conditional knockout mice (cKO) compared to wild-type (WT) mice
Fractional labeling of TCA cycle intermediates with uniformly-labeled, 13C5-Gln in retina from rod photoreceptor-specific Gls cKO mice compared to WT mice at P14. Statistical differences are based on an unpaired two-tailed Student’s t-test where *P<0.05, **P<0.01 and ***P<0.001. Data are presented as mean ± standard error of the mean.
Amino acid levels are altered in the IND-cKO retina
Relative abundance of glutamine (Gln), glutamate (Glu), aspartate (Asp) and asparagine (Asn) in retinal tissue from WT and IND-cKO 10 days post tamoxifen (TAM), which is prior to PR degeneration. N=5 animals per group. Statistical differences are based on an unpaired Student’s t-test where *P<0.05. Data are presented as mean ± standard error of the mean.
ISRIB does not impact retinal anatomy in WT mice and prolongs PR survival in the cKO retina
(A) ISRIB (2.5 mg/kg) or vehicle (50% PEG 400, 43.4% saline and 6.6% DMSO) was delivered intraperitoneally in WT mice starting at P5. Total retinal (TRT), outer nuclear layer (ONL) and the inner segment/outer segment thickness were measured at P21 by OCT. N=3 animals per group. (B) ONL thickness at P28 in cKO mice as assessed by OCT following intraperitoneal injection of ISRIB (2.5 mg/kg) or vehicle (50% PEG 400, 43.4% saline and 6.6% DMSO) from P5-P28. N=4-5 animals per group. Statistical differences in (A) and (B) are based on an unpaired two-tail Student’s t-test where *P <0.05 and **P<0.01. Data are presented as mean ± standard error of the mean.
Histology confirms ISRIB and Asn prolong photoreceptor survival in the GLS cKO retina
Representative hematoxylin and eosin stained retinal sections from rod photoreceptor-specific Gls conditional knockout (cKO) mice at P21 following intraperitoneal injection with A) ISRIB (2.5 mg/kg) or vehicle (50% PEG 400, 43.4% saline, 6.6% DMSO) or B) Asn (200 mg/kg) or vehicle (PBS) from P5-P21. ONL, outer nuclear layer; INL, inner nuclear layer. Asn, asparagine. Scale bar is 50 µm.
Primers utilized in qRT-PCR experiments to measure gene expression
