Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Jul 1.
Published in final edited form as: Exp Eye Res. 2019 Mar 16;184:135–145. doi: 10.1016/j.exer.2019.03.007

Broad Spectrum Metabolomics for Detection of Abnormal Metabolic Pathways in a Mouse Model for Retinitis Pigmentosa

Ellen R Weiss a,b,c, Shoji Osawa a, Yubin Xiong a, Suraj Dhungana d,g, James Carlson d,h, Susan McRitchie d,f, Timothy R Fennell d,e
PMCID: PMC6570542  NIHMSID: NIHMS1528009  PMID: 30885711

Abstract

Retinitis pigmentosa (RP) is a degenerative disease of the retina that affects approximately 1 million people worldwide. There are multiple genetic causes of this disease, for which, at present, there are no effective therapeutic strategies. In the present report, we utilized broad spectrum metabolomics to identify perturbations in the metabolism of the rd10 mouse, a genetic model for RP that contains a mutation in Pde6β. These data provide novel insights into mechanisms that are potentially critical for retinal degeneration. C57BL/6J and rd10 mice were raised in cyclic light followed by either light or dark adaptation at postnatal day (P) 18, an early stage in the degeneration process. Mice raised entirely in the dark until P18 were also evaluated. After euthanasia, retinas were removed and extracted for analysis by ultra-performance liquid chromatography-time of flight-mass spectrometry (UPLC-QTOF-MS). Compared to wild type mice, rd10 mice raised in cyclic light or in complete darkness demonstrate significant alterations in retinal pyrimidine and purine nucleotide metabolism, potentially disrupting deoxynucleotide pools necessary for mitochondrial DNA replication. Other metabolites that demonstrate significant increases are the Coenzyme A intermediate, 4’-phosphopantothenate, and acylcarnitines. The changes in these metabolites, identified for the first time in a model of RP, are highly likely to disrupt normal energy metabolism. High levels of nitrosoproline were also detected in rd10 retinas relative to those from wild type mice. These results suggest that nitrosative stress may be involved in retinal degeneration in this mouse model.

Keywords: retinal degeneration, UPLC-TOF-MS, metabolomics, deoxynucleotide, nitric oxide

1. Introduction

Retinal degeneration is caused by a variety of genetic and environmental insults that result in partial or total blindness. In the majority of cases, therapeutic intervention is minimally effective or nonexistent (Bramall et al., 2010). As of January 2019, approximately 307 genes and loci have been identified that contribute to various retinal degeneration phenotypes (RetNet, https://sph.uth.edu/retnet/home.htm). For one of these diseases, retinitis pigmentosa (RP), it is estimated that 1 in 4,000 individuals worldwide is afflicted with dominant, recessive or X-linked forms that result in the loss of rods, followed by cones at later stages (Hartong et al., 2006). Although it is widely accepted that oxidative stress plays a role in RP, due to the large consumption of oxygen and ATP in the retina (Okawa et al., 2008), very little is known regarding the metabolic perturbations that lead to oxidative stress from these diverse genetic insults and how these pathways result in retinal degeneration.

Broad spectrum metabolomics is an untargeted method that has been successful in identifying novel biomarkers for diseases, such as cardiovascular disease (Griffin et al., 2011), diabetes (Sas et al., 2015), Parkinson’s (Caudle et al., 2010), and Alzheimer’s. This analytical method was also able to identify metabolic changes in glaucoma and therapeutic interventions for this disease in mouse models, rabbits and patients (Agudo-Barriuso et al., 2013; Barbosa-Breda et al., 2017; Williamson et al., 2018), as well as detection of biomarkers in the plasma of patients with macular degeneration (Osborn et al., 2013). However, no studies of RP using broad spectrum metabolomics have been reported. Therefore, the range of metabolic changes that occur during retinal degeneration has never been studied directly in the retina.

In the present study, we acquired broad spectrum metabolomics data from retinas of rd10 (a model for recessive RP) and wild type C57Bl/6J mice to identify potential metabolites and pathways involved in disease progression at postnatal day (P) 18, which is early in degeneration. The rod photoreceptor cells in the rd10 mouse have a missense mutation (Arg560Cys) in the gene for PDE6β (Pde6b), one of the subunits of the cGMP-phosphodiesterase, PDE6, where many human mutations are located (Chang et al., 2002; Han et al., 2013). Retinas in these mice demonstrate reduced levels of PDE6β and PDE6γ as well as mistargeting to the inner segment (Chang et al., 2007; Wang et al., 2018). The reduced levels of PDE6β result in higher cGMP levels in photoreceptor cells, which maintain a greater number of cGMP-gated cation channels in an open state. The pathways that cause disease are presently unknown. Using broad spectrum metabolomics, our study represents the first direct analysis of retinal tissue to identify pathways that are potentially involved in early disease progression in RP. Our novel results demonstrate significant elevations in deoxyribonucleotides and related metabolites in rd10 mice that may affect mitochondrial DNA synthesis or repair of genomic DNA. In addition, an intermediate metabolite involved in Coenzyme A generation, as well as a number of acyl carnitines, were also elevated, potentially disrupting energy production in the retina. We also detected high levels of nitrosoproline, which suggests a role for nitrosative stress in the rd10 mouse model.

2. Methods

2.1. Mice

All animal protocols were approved by the University of North Carolina Institutional Animal Care and Use Committee (protocol #18–114). C57BL/6J and rd10 mice (The Jackson Laboratory, Bar Harbor, ME) were raised under varying light conditions in standard housing, temperature and access to food and water. Some mice were raised under normal light/dark cycle at ambient light levels of approximately 80 lux. At 18 days after birth (P18), which is after terminal differentiation of the retina (Chang et al., 2007; Gargini et al., 2007), mice were exposed to 250 lux for 3 h (light-adapted), in order to normalize the light exposure for all these mice, followed by euthanasia using cervical dislocation. The eyes were enucleated and the retinas dissected in PBS in the light under a microscope, followed by rapid freezing by dropping them individually into a container filled with liquid nitrogen. Each set of replicates (4 retinas) were pooled in tubes containing liquid nitrogen, which was allowed to evaporate in a −80 °C freezer. Other mice raised in cyclic light were adapted overnight at P17 in complete darkness (dark-adapted) before euthanasia at P18. Using infrared goggles and an infrared microscope, these retinas were processed identically to the light-adapted samples except that all procedures were performed in the dark. Mice raised in total darkness were euthanized at P18 and dissected in the dark using infrared goggles and an infrared microscope as described above. Animals were randomized with regard to gender and cage number. Although C57Bl/6J mice do not synthesize melatonin, to reduce the possibility of variance due to light cycles, mice were euthanized at approximately noon to 2 μm. All samples were stored at −80 °C until processed for metabolomic analysis.

2.2. Broad Spectrum Metabolomics Analysis

All analytical raw data files and processed data files have been uploaded to www.metabolomicsworkbench.com. Four retinas from 2 mice (~15–25 mg wet weight) were treated as a single sample. Each sample was homogenized in buffer containing 50:50 acetonitrile:water at a ratio of 1mg:25 μl tissue:buffer. A minimum of 5 samples were prepared for each condition. An aliquot of tissue homogenate from samples with sufficient mass was pooled to created quality control (QC) samples. The study samples and QC sample were prepared using the same method. A 320 μl aliquot of homogenate (12.8 mg tissue equivalent) was transferred to a tube and centrifuged. The supernatant was dried and re-suspended in 100 μl 95:5 H2O:methanol for UPLC-TOF-MS analysis.

Samples were analyzed on a SYNAPT G2/G2Si quadrupole time-of-flight mass spectrometer coupled to an Acquity UPLC (Waters Corporation, MA) for broad-spectrum metabolomics analysis. Metabolites were separated on a Waters Acquity HSS T3 column (2.1 × 100mm, 1.8 μm particle size) operating at 50 °C using a reversed-phase chromatographic method. A gradient mobile phase consisting of water with 0.1% formic acid (A) and methanol with 0.1% formic acid (B) was used as previously described (Dunn et al., 2011). The LC gradient used for separation included: 0.0–1.0 min 1% B, 1.0–16.0 min 1–99% B, 16.0–20.0 min 99% B, 20.0–20.5 min 1% B, and 20.5–22.0 min 1% B. The MS data were acquired over 50–1000 m/z in ESI positive and negative ion modes. Leucine enkephalin was used as the lock mass. A lock mass scan was collected every 45 s and averaged over 3 scans to perform lock mass correction. Source and desolvation temperatures were set at 110 °C and 450 °C, respectively.

All MS data analyses (alignment, normalization, and peak picking) were performed using Progenesis QI (Waters). Multivariate analysis (principal component analysis [PCA] and orthogonal projection to latent structures discriminant analysis [OPLS-DA]) of the normalized metabolomics data were performed using SIMCA (Sartorius Stedim Biotech, Umeå, Sweden) with mean centering and unit variance (UV) scaling to reduce the dimensionality and to visualize the study groups (Eriksson et al., 2013; Trygg et al., 2007). PCA and OPLS-DA are pattern recognition methods that are commonly used to analyze high dimensional multicollinear data such as metabolomics data (Eriksson et al., 2013; Trygg et al., 2007). PCA is an unsupervised analysis (the outcome is not used in the analysis) and the dimensionality is reduced by projecting the data onto a new coordinate system based on the principal components where the first component (t[1 ]) is the vector that maximizes the variance in the data, and the second component (t[2]) is a vector that is orthogonal to the first component and accounts for the next highest variance (Eriksson et al., 2013). PCA allows for visualizing any clustering in the data and identifying outliers (Eriksson et al., 2013). The quality of the broad-spectrum data was assessed by examining the PCA score plot to ensure that the QC pools were clustered, which is a quality control method widely used in broad spectrum metabolomics studies (Masson et al., 2011; Sangster et al., 2006). OPLS-DA is a supervised multivariate analysis method which uses the outcome and a rotation so that the first component (t[1]) is the predictive component that maximizes the between group variation and other systematic variation (within group variance) is included in the orthogonal components (to[1]) (Eriksson et al., 2013). Loading plots and variable influence on projection (VIP) plots from the OPLS-DA analysis were inspected and peaks that had a VIP ≥ 1 with a jackknife confidence interval that did not include 0 were determined to be important for differentiating study groups. All models used a 7-fold cross-validation to assess the predictive variation of the model (Q2). Descriptive statistics and hypothesis testing were conducted using SAS 9.4 (SAS Institute Inc., Cary, NC).

Putative identification of the group differentiating metabolites identified from OPLS-DA (VIP ≥ 1 with a jackknife confidence interval that did not include 0) was made using a database search against the NIH Eastern Regional Comprehensive Metabolomics Resource Core’s in-house exact mass retention time library of standards (~900 compounds) and the Human Metabolome Database. Peaks that could not be matched to the library were classified as unknown peaks. Pathway enrichment analyses using library-matched metabolites that distinguished study groups were performed using Metaboanalyst (http://www.metaboanalyst.ca) (Xia and Wishart, 2016). Fold changes, p-values and VIP are reported for peaks with a putative identification in the Supplemental Tables.

3. Results

Wild type and rd10 mice were analyzed at P18, which is relatively early in disease progression, since only 11% of the nuclei are reported to exhibit TUNEL staining in light-raised rd10 mice at that age (Arango-Gonzalez et al., 2014; Dong et al., 2017; Samardzija et al., 2012). For our experiments, mice were raised either in complete darkness (RD-D) or under cyclic light (RCL). The mice raised in cyclic light were either dark-adapted overnight (RCL-D) at P17 or exposed to light (250 lux) at P18 for 3 h (RCL-L) before euthanasia and removal of the retinas as described in the Methods. Alternatively, mice were raised in absolute darkness and sacrificed at P18 for metabolomic analysis. The principal component analysis (PCA) scores plot for the negative mode data is shown in Fig. 1. The interpretation of the PCA score plot for the positive mode data (not shown) is consistent with the negative mode data. The quality control (QC) pools (black pentagon) are clustered, indicating stability of the platform throughout the run. Wild type mice raised in cyclic light (green circle, light-adapted, WT-RCL-L; and grey diamond, dark-adapted, WT-RCL-D) or raised in the dark (brown hexagon, WT-RD-D) were primarily clustered on the left side of the PCA plot. In contrast, both sets of rd10 mice raised in cyclic light clustered on the right side of the PCA plot (red triangle, light-adapted, rd10-RCL-L, and blue square, dark-adapted, rd10-RCL-D). The rd10 mice raised entirely in the dark clustered on the left side (green star, rd10-RD-D) without overlapping the wild type mice, indicating that the differences in metabolic profiles of the rd10 mice raised in cyclic light compared to rd10 mice raised in the dark are greater than the wild type mice raised in different light conditions. Therefore, it appears that metabolism in the rd10 retinas is significantly more sensitive to light exposure during rearing than in the wild type mice.

Fig. 1.

Fig. 1

Open in a new tab

Scores Plot of Principal Component Analysis (PCA) of P18 retinas from wild type (WT) and rd10 mice including QC samples from UPLC-TOF-MS in the negative mode (R2X=0.79, Q2=0.57).

For each light condition, 5 to 6 samples, each containing 4 retinas, were evaluated by UPLC-TOF-MS. Untargeted metabolomics data were collected in both positive and negative modes. Only the negative mode data are shown but results were similar in the positive mode. Following normalization, the data were mean centered and scaled by unit variance. Quality control (QC) pools (black hexagon) cluster, indicating stability of the platform throughout the run. Retinas from cyclic light-raised WT mice (green circles and grey diamonds) were not clearly differentiated from dark-reared WT mice (brown hexagons). In contrast, retinas from rd10 mice raised in cyclic light (RCL; blue squares, red triangles) were differentiated from the mice raised in the dark (RD; green stars) along the t[1] axis.

Pairwise comparisons of the metabolic profiles within each genotype raised under different light conditions were performed using orthogonal projection to latent structures discriminant analysis (OPLS-DA) (Fig. 2). The wild type mice and rd10 mice raised in different light conditions could be differentiated in these analyses. Both positive and negative mode UPLC-TOF-MS data were analyzed, and because the results are similar, only the negative mode data are shown. Forty-three peaks could be library-matched to metabolites differentiating wild type mice raised in cyclic light or raised in the dark and both dark-adapted (Fig.2A). Sixty-two peaks could be library-matched to metabolites differentiating wild type mice raised in cyclic light and light-adapted from those raised in cyclic light and dark-adapted (Fig. 2B). The compounds that differentiate wild type mice under different light conditions are shown in Supplemental Tables S1 and S2 (negative mode). Similarly, rd10 mice raised in different light conditions could be differentiated using OPLS-DA. One hundred peaks could be library-matched to metabolites differentiating rd10 mice raised in cyclic light from those that were reared in the dark (Fig. 2C) and fifty-three peaks could be library-matched to metabolites that distinguished rd10 mice raised in cyclic light, then either exposed to light for 3 h or kept in the dark overnight (Fig. 2D). One rd10-RCL-L sample was an outlier in the PCA (Fig. 1), and was removed from the rd10-RCL-L vs rd10-RCL-D analysis. Metabolites important to differentiating the rd10 mice under different light conditions are shown in Supplemental Tables S3 and S4 (negative mode). Therefore, differences in light exposure during rearing and adaptation both appear to influence metabolism in wild type and rd10 mice.

Fig. 2.

Fig. 2

Open in a new tab

Orthogonal Projection to Latent Structures-Discriminant Analysis (OPLS-DA) of wild type (WT) and rd10 mice.

Pairwise comparisons of WT and rd10 retinas in mice reared under the same light conditions revealed distinct metabolo mic profiles for the 2 genotypes. Therefore, WT and rd10 retinas respond differently to light rearing and light adaptation. A. Separation of WT-RCL-D (grey diamonds) and WT-RD-D (brown hexagons) mouse retinas [R2X = 0.36; R2Y= 0.91; Q2= −0.03]. B. Separation of WT-RCL-L (green circles) and WT-RCL-D (grey diamonds) mouse retinas [R2X = 0.89; R2Y= 1.0; Q2=0.68]. C. Separation of rd10-RCL-D (blue squares) and rd10-RD-D type (green stars) mouse retinas [R2X = 0.72; R2Y= 1.0; Q2=0.94]. D. Separation of rd10-RCL-L (red triangles) and rd10-RCL-D (blue squares) mouse retinas [R2X=0.91, R2Y=1.0, Q2Y=0.79]

Comparison between rd10 and wild type mice is most important for discovering changes in metabolites that contribute to the disease process. Notably, differences between rd10 and wild type mice raised in different light conditions and adapted in either the dark or the light prior to euthanasia were also distinguished by OPLS-DA (Figure 3). Sixty-five metabolites were important for differentiating mice raised in the cyclic light and light adapted (RCL-L) (Fig. 3A) and are shown in Supplemental Table S7 (negative mode). rd10 and wild type mice raised in cyclic light and dark-adapted (RCL-D) could be distinguished by one-hundred-one metabolites (Fig. 3B) and mice raised entirely in dark (RD-D) could be distinguished by forty-eight library-matched compounds (Fig. 3C). Metabolites important to the differentiations are shown in Supplemental Tables S6 and S5, respectively (negative mode). Therefore, the data strongly suggest that metabolic differences exist between these two genotypes under all rearing and adaptation conditions.

Fig. 3.

Fig. 3

Open in a new tab

OPLS-DA of rd10 and wild type (WT) mice raised in different light conditions. A. Separation of rd10 retinas (red triangle, left) and WT retinas (bright green circles, right) raised under RCL-L conditions [R2X = 0.65; R2Y= 1.0; Q2=0.94]. B. Separation of rd10 retinas (blue squares, left) and WT retinas (grey diamonds, right) raised under RCL-D conditions [R2X = 0.70; R2Y= 1.0; Q2=0.96]. C. Separation of rd10-RD-D retinas (green stars, left) and WT-RD-D retinas (brown hexagons, right) [R2X = 0.86; R2Y= 1.0; Q2=0.84].

Metabolite data identified as important to differentiating the study groups in both negative and positive mode UPLC-TOF-MS were used for pathway enrichment analysis to obtain a comprehensive view of potentially perturbed pathways. Pathways that were most significantly altered under each of the three light conditions (RCL-L, RCL-D and RD-D) for rd10 compared to wild type mice are shown in Fig. 4. Pyrimidine and purine metabolism pathways are significantly perturbed under all three light conditions. Based on these results, we analyzed the fold differences for the pyrimidine and purine metabolites that discriminate the two genotypes (VIP ≥ 1) under the most extreme conditions, which were raised in cyclic light and light-adapted (RCL-L) compared to mice raised in the dark (RD-D). Four compounds in the pyrimidine pathway are dramatically elevated in rd10 mouse retinas compared to retinas from wild type mice when raised in cyclic light and light-adapted (Fig. 5A; Table 1, Column 7, bold). These compounds include deoxycytidine (74.2-fold), thymidylic acid (dTMP; 34.3-fold), and thymidine (4.7-fold). In addition, 2-deoxy ribose 1-phosphate (87.9-fold), a breakdown product of thymidine, was also highly elevated. In contrast, dUMP displays a moderately lower ratio (0.8-fold). Comparing dark-reared rd10 and wild type mice, we also observed differences in four compounds but pairwise ratios were smaller than in light-reared mice (deoxycytidine, 18.7-fold; dTMP, 7.3-fold; thymidine, 1.8-fold; 2-deoxy ribose 1-phosphate, 23.0-fold; Fig. 5B; Table 1, Column 5, bold). This is consistent with observations based on morphology from several laboratories, including our own, that degeneration proceeds more slowly in dark-reared rd10 mice than in rd10 mice raised in cyclic light (Chang et al., 2007; Cronin et al., 2012; Dong et al., 2017; Pang et al., 2008). Interestingly, light and dark adaptation did not make a large difference in the number of metabolites that were altered between rd10 and wild type retinas when animals were raised in cyclic light, suggesting that adaptation (3 h light or overnight dark) plays a very small role in altering pyrimidine metabolism compared to rearing conditions (Table 1; Columns 6 vs. 7, bold). These deoxyribonucleotide-related pyrimidine metabolites are required for DNA synthesis, but cells in the retina are no longer mitotic. Therefore, the results suggest disruption of these pathways may affect mitochondrial DNA synthesis and repair of damaged genomic DNA. Other pyrimidines did not show large differences in pairwise comparisons, although they are still statistically significant (Table 1, below the double line).

Fig. 4.

Fig. 4

Open in a new tab

Pathway analysis of compounds found to be statistically different between rd10 and wildtype mice using the computer program, Metaboanalyst. A. raised in cyclic light and light-adapted (RCL-L); B. raised in cyclic light and dark-adapted (RCL-D); C. raised in the dark (RD-D). Data from both positive and negative mode MS were used for pathway analysis.

Fig. 5.

Fig. 5

Open in a new tab

Comparative analysis of pyrimidine metabolites in rd10 vs. wild type (WT) mice. A selected region of the pyrimidine pathway from the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (http://www.genome.jp/kegg/) demonstrating significant changes in pyrimidine levels in rd10 mice reared in cyclic light and light-adapted (A) and reared in the dark (B) compared to WT mice. Compounds in bold and red lettering represent metabolites found to have significant fold differences between the 2 genotypes with a VIP ≥ 1. The values in parentheses represent fold differences for rd10/WT from Table 1. #1, thymidine phosphorylase; #2, thymidine kinase.

Table 1.

Pyrimidine Metabolites Identified in Pairwise Comparisons

Within Genotype Between Genotypes
Genotvoe: RCL, RD or RCL/RD ratio Wild type:RCL/RD Wild type:RCL rd10:RCL/RD rd10:RCL rd10/Wild type:RD rd10/Wild type:RCL rd10/Wild type:RCL
Adaptation Dark L vs. D Dark L vs. D Dark Dark Light
Fold Change (VIP) Fold Change (VIP) Fold Change (VIP) Fold Change (VIP) Fold Change (VIP) Fold Change (VIP) Fold Change (VIP)
Column Number: 1 2 3 4 5 6 7
Deoxycytidine X X 6.09 (1.06) X 18.73 (1.75) 83.24 (1.11) 74.17 (1.26)
‡dUMP 0.88 (1.61) X 0.68 (1.21) X X 0.85 (1.01) 0.83 (0.96)
‡ Thymidylic Acid (dTMP) 0.45 (1.11) 3.1 (1.18) 6.46 (1.24) X 7.33 (1.61) 105.82 (1.25) 34.34 (1.22)
* Thymidine X X 3.14 (1.27) X 1.83 (1.74) 4.97 (1.28) 4.67 (1.41)
‡ 2-deoxyribose 1-phosphate X X 9.04 (1.25) X 23.04 (1.42) 140.52 (1.22) 87.85 (1.23)
‡ Uridine 5’ diphosphate (UDP) 1.01 (1.04) X 1.25 (1.18) 1.2 (1.12) X 1.22 (1.03) 1.28 (1.09)
‡ P1, P4-Bis (5-uridyl) tetraphosphate X X X 1.77 (1.39) X X 2.57 (1.35)
‡ Orotidylic acid X 2 (1.13) 2.06 (1.12) 2.06 (1.15) X 1.72 (1.09) 1.64 (0.96)
* Uridine X X X X 0.9 (1.36) 0.87 (0.96) 0.82 (1.3)
* L-glutamine X X 1.63 (1.14) X X 2.52 (1.21) X
‡ Uridylic Acid (UMP) 1.08 (1.07) X 1.52 (1.02) X X 1.44 (0.96) X
Open in a new tab

Compounds in bold lettering are shown in Fig. 5. Numbers represent fold change. Numbers in parentheses are the VIP values for the comparison.

*

Compounds detected in positive mode.

‡

Compounds detected in negative mode. X, no detectible fold change

In the purine pathway (Fig. 6A; Table 2, Column 7, bold), we identified nine compounds that were elevated in rd10 retinas compared to retinas from wild type mice raised in cyclic light and light-adapted. The compounds with the greatest changes were deoxyadenosine monophosphate (dAMP; 9.4-fold), deoxyadenosine (3.1 fold), deoxyinosine (189.3-fold), and uric acid (13.0-fold). Deoxyinosine (a deoxyribonucleoside) is known to be an intermediate in the purine degradation pathway to uric acid, although the consequence of the accumulation of deoxyinosine is not clear. Uric acid has been identified as a marker for oxidative stress that is either protective or destructive depending on the environmental conditions and tissue (Glantzounis et al., 2005). Other compounds that displayed smaller increases were cAMP (2.1-fold), AMP (2.9-fold) adenine (2.9-fold) and adenosine (1.7-fold). Similar to our observations of the pyrimidine pathway, there was a reduction in the number of compounds and the fold differences that were significant in comparisons of rd10 and wild type mice raised in the dark (Fig. 6B) with the exception of deoxyinosine, which remained high in both cyclic light- and dark-reared rd10 mice (Fig. 6B; Table 2, Columns 7 and 5, bold; 189-fold and 187-fold, respectively). Interestingly, in both pyrimidine and purine pathways, a number of significant changes were also observed in a pairwise comparison between rd10 mice raised in cyclic light vs. raised in the dark (Tables 1 and 2; Column 3), whereas few changes were observed in comparisons of wild type or rd10 mice raised in cyclic light and light- or dark-adapted (Tables 1 and 2; Columns 2 and 4, respectively). With the exception of deoxyinosine (Table 2; Column 1), wild type mice did not show large differences under different rearing or adaptation conditions.

Fig 6.

Fig 6

Open in a new tab

Comparative analysis of purine metabolites in rd10 vs. wild type (WT) mice. A selected region of the purine pathway from the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (http://www.genome.jp/kegg/) demonstrated increases in purines in rd10 mice reared in cyclic light and light-adapted (A) and reared in the dark (B) compared to WT mice. Compounds in bold and red lettering represent metabolites whose fold difference was significant with a VIP ≥ 1. The values in parentheses represent fold differences for rd10/WT from Table 2. #1, adenylosuccinate lyase.

Table 2.

Purine Metabolites Detected in Pairwise Comparisons

Within Genotype Between Genotypes
Genotvoe: RCL, RD or RCL/RD ratio Wild type:RCL/RD Wild type:RCL rd10:RCL/RD rd10:RCL rd10/Wild type:RD rd10/Wild type:RCL rd10/Wild type:RCL
Adaptation Dark L vs. D Dark L vs. D Dark Dark Light
Fold Change (VIP) Fold Change (VIP) Fold Change (VIP) Fold Change (VIP) Fold Change (VIP) Fold Change (VIP) Fold Change (VIP)
Column Number: 1 2 3 4 5 6 7
*3’−5’ -cyclic AMP X X 1.65 (1.11) X X X 2.11 (1.23)
Adenosine monophosphate X X X X X X 2.86 (1)
*Adenosine X X 1.56 (1.07) X X 1.78 (1.05) 1.73 (1.21)
*Adenine X X 2.12 (1.06) X X 2.65 (1.09) 2.86 (1.1)
*Deoxyadenosine X X X X 1.7 (1.58) 3.82 (1.08) 3.06 (1.31)
‡Deoxyadenosine monophosphate (dAMP) X 0.51 (1.28) 3.82 (1.07) X 2.33 (1.52) 6.6 (1.13) 9.38 (1.23)
*Deoxyinosine 9 (1.87) X 3.23 (1.26) X 186.83 (1.75) 67.07 (1.31) 189.25 (1.33)
*Hypoxanthine X X 1.5 (1.13) X X 1.95 (1.2) 2.2 (1.36)
‡Uric Acid X X 4.72 (1.17) X 2.48 (1.4) 14.45 (1.28) 13.03 (1.38)
‡ ADP-ribose X X 2.2 (1.23) X X 5.42 (1.27) 4.26 (1.4)
* dITP X X 2.46 (1.24) X 0.64 (1.29) 2.08 (1.09) 1.4 (1.01)
* Guanosine X X 0.83 (1.21) 1.1 (1.27) 0.92 (1.45) 0.76 (1.25) 0.82 (1.21)
‡ P1, P4-Bis(5’-xanthaosyl) tetraphosphate X X 1.36 (1.23) 1.21 (1.17) X 1.45 (1.13) 1.46 (1.17)
‡ 2 (formamido)-NI- (5’phosphoribosyl) acetamide 1.32 (1.13) X X X 4.15 (1.5) 8.49 (1.2) X
‡ Guanine X X X X X 0.82 (1.03) X
Open in a new tab

Compounds in bold lettering are shown in Fig. 6. Numbers represent fold change. Numbers in parentheses are the VIP values for the comparison.

*

Compounds detected in positive mode.

‡

Compounds detected in negative mode. X, no detectible fold change

Another purine, ADP-ribose, which is a breakdown product of NAD and central to NAD metabolism in mitochondria (Dolle et al., 2013), is also modestly elevated in pairwise comparisons between rd10 and wild type mouse retinas raised in cyclic light (Table 2; Columns 6 and 7; 5.4-fold and 4.3-fold respectively). It is also elevated in rd10 retinas from mice raised in cyclic light compared to dark-reared mice (2.2-fold; Table 2, Column 3). In contrast, no changes were apparent in wild type mice. ADP-ribose is also reported to be involved in second messenger signaling and posttranslational modification of calcium channels (Dolle et al., 2013; Perraud et al., 2005). Taken together, these data demonstrate that pyrimidine and purine metabolism in rd10 mouse retinas is more sensitive to light rearing conditions compared to wild type retinas.

Other pathways potentially related to energy metabolism also appear to be significantly affected. For example, 4’-phosphopantothenate, which is involved in pantothenate and Coenzyme A (CoA) biosynthesis pathways (http://www.genome.jp/kegg/), was elevated 39-fold and 63-fold in comparisons of retinas from rd10 and wild type mouse raised in cyclic light and either light- or dark-adapted, respectively (Table 3; Columns 7 and 6, respectively). There was also a 21-fold higher level of 4’-phosphopantothenate in rd10 mice compared to wild type mice raised in the dark (Table 3; Column 5). In contrast, levels were elevated only 3-fold in rd10 mice raised in the light compared to those raised in the dark (Table 3; Column 3). The accumulation of 4’-phosphopantothenate may indicate reduced production of CoA. Since CoA is a carrier for acyl residues, and participates extensively in energy metabolism, energy production may be significantly disrupted in rd10 mice (Ferrier, 2014).

Table 3.

Other Metabolites Detected in Pairwise Comparisons

Within Genotype Between Genotypes
Genotvoe: RCL, RD or RCL/RD ratio Wild type:RCL/RD Wild type:RCL rd10:RCL/RD rd10:RCL rd10/Wild type:RD rd10/Wild type:RCL rd10/Wild type:RCL
Adaptation Dark L vs. D Dark L vs. D Dark Dark Light
Fold Change (VIP) Fold Change (VIP) Fold Change (VIP) Fold Change (VIP) Fold Change (VIP) Fold Change (VIP) Fold Change (VIP)
Column Number: 1 2 3 4 5 6 7
4’-Phosphopantothenate X X 3.05 (1.0) X 20.97 (1.42) 63.38 (1.26) 38.9 (1.28)
3-Nitrotyrosine X 1.68 (1.16) 2.07 (1.08) 1.78 (1.07) 0.6 (1.42) X X
N-Nitrosoproline X X 2.76 (0.98) X 83.62 (1.78) 67.12 (1.12) 215.65 (1.24)
Open in a new tab

Numbers represent fold change. Numbers in parentheses are the VIP values for the comparison. All compounds were detected in the negative mode. X, no detectible fold change

Long chain fatty acid metabolism may also be affected in rd10 mice, based on modest changes in the levels of large numbers of carnitine compounds (Table 4). Carnitines function in fatty acid metabolism, including the transfer of long chain fatty acids across the inner mitochondrial membrane for production of ATP via fatty acid oxidation. Fold differences range from 0.4 to 4.1 in pairwise comparisons between rd10 and wild type mouse retinas raised in cyclic light (Table 4; Columns 6 and 7). Notably, these carnitines are also perturbed in rd10 mice in a pairwise comparison between light- and dark-reared animals but not in wild type animals (Table 4; Columns 3 and 1, respectively). These perturbations may be indicative of more oxidative stress in light-reared rd10 mice (Ferrier, 2014). In contrast, a comparison of retinas from rd10 and wild type mice raised in the dark show almost no differences in carnitine ratios (Table 4; Column 5). Therefore, it appears that increases in carnitine levels are sensitive to light-rearing in the rd10 mice.

Table 4.

Carnitine Compounds Detected in Pairwise Comparisons

Within Genotype Between Genotypes
Genotvoe: RCL, RD or RCL/RD ratio Wild type:RCL/RD Wild type:RCL rd10:RCL/RD rd10:RCL rd10/Wild type:RD rd10/Wild type:RCL rd10/Wild type:RCL
Adaptation Dark L vs. D Dark L vs. D Dark Dark Light
Fold Change (VIP) Fold Change (VIP) Fold Change (VIP) Fold Change (VIP) Fold Change (VIP) Fold Change (VIP) Fold Change (VIP)
Column Number: 1 2 3 4 5 6 7
2-Hydroxyhexadecanoylcamitine X X 2.31 (1.27) X X 2.39 (1.17) 2.5 (1.21)
2-Hydroxymyristoylcarnitine X X X X 1.26 (1.14) X X
11z-Octadecenylcarnitine 1.42 (1.83) X 2.28 (1.02) X X X X
12-hydroxy-12
octadecanoylcarnitine		 X X 1.96 (1.14) X X X 1.55 (0.99)
3-Hydroxy-9Z-
octadecenoylcarnitine		 X X 1.54 (1.11) X X 1.82 (1.15) X
3-Hydroxyhexadecanoylcarnitine X X 2.31 (1.27) X X 2.39 (1.51) 2.5 (1.21)
3-hydroxy 5–8- tetradecadiencarnitine X X 0.54 (1.2) X X 0.43 (1.3) X
Butyrylcarnitine X X 1.35 (1.09) X X 1.35 (0.99) X
Cervonycarnitine X X 3.88 (1.11) X X 3.0 (1.02) 1.61 (0.99)
L-acetylcarnitine X X 1.4 (0.99) X X X X
Linoelaidyl carnitine X X 2.86 (1.18) X X 3.37 (1.17) 2.52 (1.19)
Linoleyl carnitine 1.52 (1.77) X 4.83 (1.06) X 1.27 (1.24) 4.05 (0.97) 3.11 (1.15)
L-Palmitoylcarnitine X X 2.04 (1.23) X X 1.76 (1.13) 1.74 (1.16)
Propionylcarnitine X 1.08 (1.01) X X X 1.26 (1.04) 1.31 (0.99)
Stearoylcarnitine X X 2.62 (1.32) 0.66 (1.99) 1.18 (1.03) 2.69 (1.32) 2.03 (1.27)
Valerylcarnitine X X 1.56 (1.07) X X 1.53 (0.99) X
Open in a new tab

Numbers represent fold change. Numbers in parentheses are the VIP values for the comparison. All compounds were detected in the positive mode. X, no detectible fold change

We also observed altered ratios of other compounds not clearly associated with specific metabolic pathways (Table 3). One of these compounds, nitrosoproline, is extremely elevated in rd10 compared to wild type mice (215-fold RCL-L, column 7; 67-fold RCL-D, column 6; 83.6-fold, RD-D, column 5). Nitrosoproline was originally found in the urine of individuals who smoke cigarettes, but is no longer thought to be carcinogenic (Arimoto-Kobayashi et al., 2002). Rather, it has been correlated with the presence of nitric oxide (NO) in tissues (L’hirondel, 2002). Importantly, a comparison of changes between different light conditions within the same genotype are either much lower (e.g. 2.8-fold in rd10, RCL vs. RD; Table 3; Column 3) or no statistical difference was found in our dataset (wild type and rd10 retinas, RCL-L vs. RCL-D, Columns 1,2 and 4 in Table 3). High levels of nitric oxide (NO) are indicative of nitrosative stress (Thomas et al., 2006) that, along with oxidative stress, is known to inactivate enzymes, modify DNA to form 8-oxo and 8-nitroguanine, as well as generate lipid hydroperoxides and other modified lipids (Aicardo et al., 2016; Liu et al., 2002). In contrast to nitrosoproline, nitrotyrosine exhibited fewer differences between rd10 and wild type retinas. This is likely due to the fact that nitrotyrosine is found primarily in proteins and also there are differences in the mechanism that generates this modified amino acid. However, its presence is also an indicator of NO in the retina (Kaur and Halliwell, 1994).

4. Discussion

This is the first time that changes in metabolism have been evaluated in an RP model using broad spectrum metabolomics. Broad spectrum metabolomics has the advantage of being able to identify entirely unexpected metabolic changes in diseased compared to normal tissues, although we did not identify some metabolites previously known to distinguish between genotypes and between dark and light-raised wild type mice using targeted metabolomics, such as cGMP (Du et al., 2016). However, our studies have led to the identification of novel metabolites in retinas of the rd10 compared to wild type mice that have not been found using a targeted approach. These changes were observed relatively early (P18) in the degeneration process (Arango-Gonzalez et al., 2014; Dong et al., 2017; Samardzija et al., 2012). The rd10 mouse is a useful model for human RP because mutations in the catalytic domain of PDE6β also cause RP in humans (Hartong et al., 2006). By performing a pairwise comparison between wild type and rd10 mouse retinas under a variety of light conditions, we were also able to identify disease-related changes exacerbated by light. Strikingly, in the PCA analysis (Fig. 1), rd10 mice reared in cyclic light are well differentiated from rd10 mice reared in the dark, whereas wild type mice raised in the dark are clustered but not well differentiated from wild type mice reared in cyclic light. These data suggest greater perturbations in rd10 mice than wild type mice with light rearing conditions. However, OPLS-DA analysis revealed that both rd10 and wild type mice do have distinct metabolic profiles when reared in light compared to in the dark, as well as mice that are light- or dark-adapted (Fig. 2). These results are consistent with changes in energy utilization and metabolism in retinas from dark- and light-exposed wild type mice reported previously (Du et al., 2016; Linton et al., 2010). In addition, retinas from rd10 mice have metabolic profiles that are distinct from wild type mice under all light conditions (Fig. 3).

Mitochondrial DNA codes for 37 (human) genes important for mitochondrial function. Thirteen of them code for proteins involved in oxidative phosphorylation. Mitochondria are dynamic organelles that undergo fission and fusion in normal and diseased cells (Mishra and Chan, 2016; Westermann, 2010). This has been shown by electron microscopy in mouse photoreceptor inner segments looking at the influence of aging on mitochondrial dynamics (Kam and Jeffery, 2015). Mitochondrial DNA replication and repair are processes that occurs during mitochondrial dynamics. Therefore, deoxyribonucleotides and related metabolites involved in pyrimidine and purine pathways are critical for mitochondrial survival and function (Ricchetti, 2018). In both pyrimidine and purine pathways, we observed significantly altered ratios of metabolites in comparisons between rd10 and wild type mice when reared in either light or dark conditions (Figs. 5 and 6). However, we also observed that the number of metabolites demonstrating significant differences were smaller and the fold changes were lower in dark-reared than in light-reared mice. These results are consistent with the reduced rate of degeneration demonstrated in dark-reared rd10 mice (Chang et al., 2007; Dong et al., 2017). These data also strengthen the hypothesis that significant changes in the levels of some of these metabolites are involved in fundamental processes of retinal degeneration since these changes are present even in dark-reared rd10 retinas. In fact, Berkowitz et al. (2018) reported that oxidative stress is still present in dark-reared rd10 mice at P23. Some of our results suggest an imbalance in the deoxynucleotide pools that could negatively affect mitochondrial DNA synthesis (Mathews and Song, 2007), thereby compromising the survival of the photoreceptors. On the other hand, we cannot rule out the possibility that changes in ribonucleotide-related purines negatively influence the generation of various types of RNAs, since moderate increases in AMP and adenosine in rd10 compared to wild type mice were also observed.

Previously published work on mutations in enzymes that affect mitochondrial deoxynucleotide pools have been shown to cause serious illnesses in early childhood, such as mitochondrial depletion syndrome, neurogastrointestinal encephalopathy (MNGIE), myopathy and neurotoxicity, further strengthening our hypothesis that mitochondrial dysfunction results from the rd10 mutation (Di Meo et al., 2015; Donti et al., 2016; El-Hattab and Scaglia, 2013; Garone et al., 2011; Jurecka et al., 2015). For example, in the pyrimidine pathway, reduction in the activity of thymidine phosphorylase (Fig. 5, #1) results in elevated levels of thymidine and deoxyuridine (Suomalainen and Isohanni, 2010) and causes mitochondrial neurogastrointestinal encephalopathy (MNGIE) (Di Meo et al., 2015). One clinical case of loss of vision has been reported (Garone et al., 2011). Defects in thymidine kinase (TK) (Fig. 5, #2), which synthesizes dTMP from thymidine, is associated with myopathic conditions in early childhood (El-Hattab and Scaglia, 2013). Both cytosolic (TK1) and mitochondrial (TK2) forms of these enzymes are critical for mitochondrial survival.

Less is known regarding the clinical effects of purine metabolism disruption. However, a deficiency in adenylosuccinate lyase (Fig. 6, #1), which converts adenylosuccinate to AMP, causes a metabolic disorder resulting in neurotoxicity (Donti et al., 2016; Jurecka et al., 2015). Since we observed the opposite effect - an increase in AMP levels in rd10 retinas - the consequence of this increase is not clear. In our studies of the rd10 mice, we speculate that enzymes involved in pyrimidine and purine synthesis may be denatured or inactivated in the presence of high levels of oxidative/nitrosative stress (see below), causing an imbalance in the levels of these metabolites. Another possibility is inactivation of these enzymes by calpain digestion, which has been demonstrated in rd10 retinas as early as P18 (Arango-Gonzalez et al., 2014). These potential mechanisms form the basis for some of our future studies.

The purine, ADP-ribose, also exhibited a modest increase in rd10 compared to wild type mice. This compound is a product of NAD degradation in mitochondria, as well as participating in signaling pathways in the cytoplasm. ADP-ribose has been reported to be released from mitochondria, where it activates the transient receptor potential cation channel 2 (TRPM-2) to allow Ca2+ influx from the extracellular environment in response to oxidative and nitrosative stress (Dolle et al., 2013; Perraud et al., 2005). Expression of TRPM-2 is upregulated in both rd1 and rds mouse models of RP (Wong et al., 1994), providing a potential mechanism for Ca2+ influx into the cell. Ca2+ influx is thought to be necessary for retinal degeneration in rd1 mice, based on reduced degeneration with Ca2+ channel blockers, although the effectiveness of the different blockers is controversial (Frasson et al., 1999; Takano et al., 2004). However, deletion of the cyclic nucleotide-gated (CNG) ion channels on the rd10 background does slow degeneration, further suggesting that an influx of Ca2+ may play a role in degeneration of retinas in mice with mutations in PDEb (Wang et al., 2018). In addition, administration of cGMP analogs does appear to reduce the rate of degeneration in rd10 mice (Vighi et al., 2018). It has also been proposed that that elevated cGMP in rd10 mice may stimulate PKG to increase calpain activity. While intriguing, these hypotheses are not yet fully developed (Iribarne and Masai, 2017). On the other hand, inhibition of PKG does reduce degeneration in the rd1 mouse (Paquet-Durand et al., 2009).

One surprising observation was the dramatic increase in nitrosoproline (one of the nitrosamines) in retinas from rd10 mice compared to those from wild type mice raised in cyclic light (215-fold) or in the dark (83-fold). Nitrosamines are found in multiple tissues at nanomolar levels, including brain, heart, liver kidney and lung, and have been detected in plasma and urine (Rassaf et al., 2002). Nitrosoproline is generated from proline and NO via the dinitrogen trioxide (N2O3) pathway (Bryan et al., 2004; Thomas et al., 2006). The presence of such a high level of nitrosoproline in rd10 mouse retinas suggests that NO levels are very high at P18. NO can be generated via 3 nitric oxide synthase (NOS) enzymes: eNOS, iNOS and nNOS. nNOS is reported to be expressed primarily in amacrine cells and possibly bipolar cells in normal retinas (Vielma et al., 2012; Wei et al., 2012). NO may also come from microglia that are activated during inflammation through upregulation of iNOS (Liu et al., 2002). Since NO can easily diffuse small distances within tissue, the place of synthesis is not necessarily indicative of its functional location. In the case of our studies, the site of NO synthesis is unknown. For example, Donovan et al. (2001) demonstrated a role for nNOS in light-induced photoreceptor apoptosis. Therefore, it is possible that one or more of the NOS enzymes is upregulated under conditions of degenerative stress. Alternatively, NO can be generated from nitrite (NO2−) in the bloodstream (Aicardo et al., 2016). The role of nitrosoproline is presently unknown, but it is possible that proline acts as a scavenger for NO, thereby attempting to reduce nitrosative stress. These data indicate that in addition to oxidative stress, which is high even in normal retinas, nitrosative stress is likely to play a role in degeneration. Importantly, NO binds the superoxide anion, O2−, which comes from mitochondria and produces a peroxynitrite anion, ONOO−, a highly reactive compound that is harmful for multiple cell functions via modification of protein, DNA and lipids (Aicardo et al., 2016). Interestingly, NO is also able to stimulate cGMP synthesis through soluble guanylate cyclases in neurons, which could lead to upregulation of PKG (Paquet-Durand et al., 2009).

Since most cells in the mouse retina are photoreceptor cells, which make up 65.6% of the cells in the mouse retina (Macosko et al., 2015), it is reasonable to speculate that many of these metabolic changes occur in rods. Alternatively, Müller glia, which surround each photoreceptor, play a role in virtually all aspects of photoreceptor function and in remodeling of neuronal connections during degeneration. These cells could be targets of metabolic changes even though their processes occupy less than 10% of the tissue volume (Reichenbach and Bringmann, 2013). A recent publication reported that high levels of proline are transported from the bloodstream into retinal pigment epithelial (RPE) cells for the generation of citrate. Therefore, we cannot rule out the possibility that nitrosoproline found in our samples comes from the RPE (Chao et al., 2017) since, as described above, NO can diffuse small distances within tissue. The ability to characterize overall metabolic pathways that are altered during disease progression in mouse models of retinitis pigmentosa will be of critical importance in design of drugs for treatment of this genetic disease.

Supplementary Material

1
2
3
4
5
6
7

Highlights.

  • First use of broad spectrum metabolomics to identify metabolic perturbations in a mouse model for retinitis pigmentosa.

  • Significant differences in metabolite profiles were identified between wild type and rd10 mice raised in cyclic light or in the dark.

  • Pathways found to be most affected involved purine and pyrimidine deoxynucleotides.

  • Elevated levels of nitrosoproline suggest NO plays an important role in the degeneration program in rd10 mice.

  • These studies provide novel insights into retinal degeneration pathways that may lead to novel drug development.

5. Acknowledgments

The metabolomics analysis was funded by the NIH Common Fund (U24 DK097193, SCJ Sumner, PI) and a supplement from the NIH Common fund to NEI R01EY12224 (ERW, PI). The authors would like to thank Zachery Acuff and Danny Vu for their contributions to this study.

Abbreviations

NO

nitric oxide

NOS

nitric oxide synthase

N2O3

dinitrogen trioxide

NO2−

nitrite

ONOO−

peroxynitrite anion

OPLS-DA

orthogonal projection to latent structures-discriminant analysis

PCA

principal component analysis

PDE

phosphodiesterase

QC

quality control

RP

retinitis pigmentosa

UPLC-TOF-MS

Ultrahigh performance liquid chromatography-time-of-flight-mass spectrometry

Footnotes

Conflict of Interest: All authors declare no conflicts of interest.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

6. References

  1. Agudo-Barriuso M, Lahoz A, Nadal-Nicolas FM, Sobrado-Calvo P, Piquer-Gil M, Diaz-Llopis M, Vidal-Sanz M, Mullor JL, 2013. Metabolomic changes in the rat retina after optic nerve crush. Invest. Ophthalmol. Vis. Sci 54, 4249–4259. [DOI] [PubMed] [Google Scholar]
  2. Aicardo A, Martinez DM, Campolo N, Bartesaghi S, Radi R, 2016. Biochemistry of Nitric Oxide and Peroxynitrite: Sources, Targets and Biological Implications, in: Gelpi RJ, Boveris A, Poderoso JJ (Eds.), Biochemistry of Oxidative Stress. Springer International Publishing, Cham, Switzerland, pp. 49–77. [Google Scholar]
  3. Arango-Gonzalez B, Trifunovic D, Sahaboglu A, Kranz K, Michalakis S, Farinelli P, Koch 5, Koch F, Cottet S, Janssen-Bienhold U, Dedek K, Biel M, Zrenner E, Euler T, Ekstrom P, Ueffing M, Paquet-Durand F, 2014. Identification of a common non-apoptotic cell death mechanism in hereditary retinal degeneration. PLoS One 9, e112142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Arimoto-Kobayashi S, Ando Y, Horai Y, Okamoto K, Hayatsu H, Lowe JE, Green MH, 2002. Mutation, DNA strand cleavage and nitric oxide formation caused by N-nitrosoproline with sunlight: a possible mechanism of UVA carcinogenicity. Carcinogenesis 23, 1537–1540. [DOI] [PubMed] [Google Scholar]
  5. Barbosa-Breda J, Himmelreich U, Ghesquiere B, Rocha-Sousa A, Stalmans I, 2017. Clinical Metabolomics and Glaucoma. Ophthalmic research. [DOI] [PubMed] [Google Scholar]
  6. Berkowitz BA, Podolsky RH, Berri AM, Dernay K, Graffice E, Shafie-Khorassani F, Roberts R, 2018. Dark Rearing Does Not Prevent Rod Oxidative Stress In Vivo in Pde6b(rd10) Mice. Invest Ophthalmol & Vis Sci 59, 1659–1665. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bramall AN, Wright AF, Jacobson SG, McInnes RR, 2010. The genomic, biochemical, and cellular responses of the retina in inherited photoreceptor degenerations and prospects for the treatment of these disorders. Annu. Rev. Neurosci 33, 441–472. [DOI] [PubMed] [Google Scholar]
  8. Bryan NS, Rassaf T, Maloney RE, Rodriguez CM, Saijo F, Rodriguez JR, Feelisch M, 2004. Cellular targets and mechanisms of nitros(yl)ation: an insight into their nature and kinetics in vivo. Proc Natl Acad Sci U S A 101, 4308–4313. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Caudle WM, Bammler TK, Lin Y, Pan S, Zhang J, 2010. Using ‘omics’ to define pathogenesis and biomarkers of Parkinson’s disease. Expert Rev Neurother 10, 925–942. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Chang B, Hawes NL, Hurd RE, Davisson MT, Nusinowitz S, Heckenlively JR, 2002. Retinal degeneration mutants in the mouse. Vision Res 42, 517–525. [DOI] [PubMed] [Google Scholar]
  11. Chang B, Hawes NL, Pardue MT, German AM, Hurd RE, Davisson MT, Nusinowitz S, Rengarajan K, Boyd AP, Sidney SS, Phillips MJ, Stewart RE, Chaudhury R, Nickerson JM, Heckenlively JR, Boatright JH, 2007. Two mouse retinal degenerations caused by missense mutations in the p subunit of rod cGMP phosphodiesterase gene. Vision Res. 47, 624–633. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Chao JR, Knight K, Engel AL, Jankowski C, Wang Y, Manson MA, Gu H, Djukovic D, Raftery D, Hurley JB, Du J, 2017. Human retinal pigment epithelial cells prefer proline as a nutrient and transport metabolic intermediates to the retinal side. J Biol Chem 292, 12895–12905. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Cronin T, Lyubarsky A, Bennett J, 2012. Dark-rearing the rd10 mouse: implications for therapy. Adv Exp Med Biol 723, 129–136. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Di Meo I, Lamperti C, Tiranti V, 2015. Mitochondrial diseases caused by toxic compound accumulation: from etiopathology to therapeutic approaches. EMBO molecular medicine 7, 1257–1266. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Dolle C, Rack JG, Ziegler M, 2013. NAD and ADP-ribose metabolism in mitochondria. FEBS J 280, 3530–3541. [DOI] [PubMed] [Google Scholar]
  16. Dong E, Bachleda A, Xiong Y, Osawa S, Weiss ER, 2017. Reduced phosphoCREB in Müller glia during retinal degeneration in rd10 mice. Mol. Vis 23, 90–102. [PMC free article] [PubMed] [Google Scholar]
  17. Donovan M, Carmody RJ, Cotter TG, 2001. Light-induced photoreceptor apoptosis in vivo requires neuronal nitric-oxide synthase and guanylate cyclase activity and is caspase-3-independent. J Biol Chem 276, 23000–23008. [DOI] [PubMed] [Google Scholar]
  18. Donti TR, Cappuccio G, Hubert L, Neira J, Atwal PS, Miller MJ, Cardon AL, Sutton VR, Porter BE, Baumer FM, Wangler MF, Sun Q, Emrick LT, Elsea SH, 2016. Diagnosis of adenylosuccinate lyase deficiency by metabolomic profiling in plasma reveals a phenotypic spectrum. Molecular genetics and metabolism reports 8, 61–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Du J, Rountree A, Cleghorn WM, Contreras L, Lindsay KJ, Sadilek M, Gu H, Djukovic D, Raftery D, Satrustegui J, Kanow M, Chan L, Tsang SH, Sweet IR, Hurley JB, 2016. Phototransduction influences metabolic flux and nucleotide metabolism in mouse retina. J Biol Chem 291,4698–4710. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Dunn WB, Broadhurst D, Begley P, Zelena E, Francis-McIntyre S, Anderson N, Brown M, Knowles JD, Halsall A, Haselden JN, Nicholls AW, Wilson ID, Kell DB, Goodacre R, 2011. Procedures for large-scale metabolic profiling of serum and plasma using gas chromatography and liquid chromatography coupled to mass spectrometry. Nat Protoc 6, 1060–1083. [DOI] [PubMed] [Google Scholar]
  21. El-Hattab AW, Scaglia F, 2013. Mitochondrial DNA depletion syndromes: review and updates of genetic basis, manifestations, and therapeutic options. Neurotherapeutics 10, 186–198. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Eriksson L, Byrne T, Johansson E, Trygg J, Vikstrom C, 2013. Multi- and Megavariate Data Analysis: Basic Principles and Applications, 3rd ed. Umetric Academy, Umea, Sweden. [Google Scholar]
  23. Ferrier DR, 2014. Biochemistry, 6th ed. Wolters Kluwer/Lippincot Williams & Wilkins, Philadelphia. [Google Scholar]
  24. Frasson M, Sahel JA, Fabre M, Simonutti M, Dreyfus H, Picaud S, 1999. Retinitis pigmentosa: rod photoreceptor rescue by a calcium-channel blocker in the rd mouse. Nature medicine 5, 1183–1187. [DOI] [PubMed] [Google Scholar]
  25. Gargini C, Terzibasi E, Mazzoni F, Strettoi E, 2007. Retinal organization in the retinal degeneration 10 (rd10) mutant mouse: a morphological and ERG study. J. Comp. Neurol 500, 222–238. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Garone C, Tadesse S, Hirano M, 2011. Clinical and genetic spectrum of mitochondrial neurogastrointestinal encephalomyopathy. Brain 134, 3326–3332. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Glantzounis GK, Tsimoyiannis EC, Kappas AM, Galaris DA, 2005. Uric acid and oxidative stress. Current pharmaceutical design 11, 4145–4151. [DOI] [PubMed] [Google Scholar]
  28. Griffin JL, Atherton H, Shockcor J, Atzori L, 2011. Metabolomics as a tool for cardiac research. Nature reviews. Cardiology 8, 630–643. [DOI] [PubMed] [Google Scholar]
  29. Han J, Dinculescu A, Dai X, Du W, Smith WC, Pang J, 2013. Review: the history and role of naturally occurring mouse models with Pde6b mutations. Mol Vis 19, 2579–2589. [PMC free article] [PubMed] [Google Scholar]
  30. Hartong DT, Berson EL, Dryja TP, 2006. Retinitis pigmentosa. Lancet 368, 1795–1809. [DOI] [PubMed] [Google Scholar]
  31. Iribarne M, Masai I, 2017. Neurotoxicity of cGMP in the vertebrate retina: from the initial research on rd mutant mice to zebrafish genetic approaches. Journal of neurogenetics 31, 88–101. [DOI] [PubMed] [Google Scholar]
  32. Jurecka A, Zikanova M, Kmoch S, Tylki-Szymanska A, 2015. Adenylosuccinate lyase deficiency. Journal of inherited metabolic disease 38, 231–242. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Kam JH, Jeffery G, 2015. To unite or divide: mitochondrial dynamics in the murine outer retina that preceded age related photoreceptor loss. Oncotarget 6, 26690–26701. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Kaur H, Halliwell B, 1994. Evidence for nitric oxide-mediated oxidative damage in chronic inflammation. Nitrotyrosine in serum and synovial fluid from rheumatoid patients. FEBS Lett 350, 9–12. [DOI] [PubMed] [Google Scholar]
  35. L’hirondel J, 2002. The Case Against Nitrate: a Critical Examination, Nitrate and Man: Toxic, Harmless Or Beneficial? CABI, Oxfordshire, UK, pp. 35–68. [Google Scholar]
  36. Linton JD, Holzhausen LC, Babai N, Song H, Miyagishima KJ, Stearns GW, Lindsay K, Wei J, Chertov AO, Peters TA, Caffe R, Pluk H, Seeliger MW, Tanimoto N, Fong K, Bolton L, Kuok DL, Sweet IR, Bartoletti TM, Radu RA, Travis GH, Zagotta WN, Townes-Anderson E, Parker E, Van der Zee CE, Sampath AP, Sokolov M, Thoreson WB, Hurley JB, 2010. Flow of energy in the outer retina in darkness and in light. Proc. Natl. Acad. Sci. U.S.A 107, 8599–8604. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Liu B, Gao HM, Wang JY, Jeohn GH, Cooper CL, Hong JS, 2002. Role of nitric oxide in inflammation-mediated neurodegeneration. Ann N Y Acad Sci 962, 318–331. [DOI] [PubMed] [Google Scholar]
  38. Macosko EZ, Basu A, Satija R, Nemesh J, Shekhar K, Goldman M, Tirosh I, Bialas AR, Kamitaki N, Martersteck EM, Trombetta JJ, Weitz DA, Sanes JR, Shalek AK, Regev A, McCarroll SA, 2015. Highly Parallel Genome-wide Expression Profiling of Individual Cells Using Nanoliter Droplets. Cell 161, 1202–1214. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Masson P, Spagou K, Nicholson JK, Want EJ, 2011. Technical and biological variation in UPLC-MS-based untargeted metabolic profiling of liver extracts: application in an experimental toxicity study on galactosamine. Anal Chem 83, 1116–1123. [DOI] [PubMed] [Google Scholar]
  40. Mathews CK, Song S, 2007. Maintaining precursor pools for mitochondrial DNA replication. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 21,2294–2303. [DOI] [PubMed] [Google Scholar]
  41. Mishra P, Chan DC, 2016. Metabolic regulation of mitochondrial dynamics. The Journal of cell biology 212, 379–387. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Okawa H, Sampath AP, Laughlin SB, Fain GL, 2008. ATP consumption by mammalian rod photoreceptors in darkness and in light. Curr. Biol 18, 1917–1921. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Osborn MP, Park Y, Parks MB, Burgess LG, Uppal K, Lee K, Jones DP, Brantley MA Jr., 2013. Metabolome-wide association study of neovascular age-related macular degeneration. PLoS One 8, e72737. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Pang JJ, Boye SL, Kumar A, Dinculescu A, Deng W, Li J, Li Q, Rani A, Foster TC, Chang B, Hawes NL, Boatright JH, Hauswirth WW, 2008. AAV-mediated gene therapy for retinal degeneration in the rd10 mouse containing a recessive PDEbeta mutation. Invest. Ophthalmol. Vis. Sci 49, 4278–4283. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Paquet-Durand F, Hauck SM, van Veen T, Ueffing M, Ekstrom P, 2009. PKG activity causes photoreceptor cell death in two retinitis pigmentosa models. J Neurochem 108, 796–810. [DOI] [PubMed] [Google Scholar]
  46. Perraud AL, Takanishi CL, Shen B, Kang S, Smith MK, Schmitz C, Knowles HM, Ferraris D, Li W, Zhang J, Stoddard BL, Scharenberg AM, 2005. Accumulation of free ADP-ribose from mitochondria mediates oxidative stress-induced gating of TRPM2 cation channels. J Biol Chem 280, 6138–6148. [DOI] [PubMed] [Google Scholar]
  47. Rassaf T, Bryan NS, Kelm M, Feelisch M, 2002. Concomitant presence of N-nitroso and S-nitroso proteins in human plasma. Free radical biology & medicine 33, 1590–1596. [DOI] [PubMed] [Google Scholar]
  48. Reichenbach A, Bringmann A, 2013. New functions of Müller cells. Glia 61,651–678. [DOI] [PubMed] [Google Scholar]
  49. Ricchetti M, 2018. Replication stress in mitochondria. Mutation research 808, 93–102. [DOI] [PubMed] [Google Scholar]
  50. Samardzija M, Wariwoda H, Imsand C, Huber P, Heynen SR, Gubler A, Grimm C, 2012. Activation of survival pathways in the degenerating retina of rd10 mice. Exp. Eye Res. 99, 17–26. [DOI] [PubMed] [Google Scholar]
  51. Sangster T, Major H, Plumb R, Wilson AJ, Wilson ID, 2006. A pragmatic and readily implemented quality control strategy for HPLC-MS and GC-MS-based metabonomic analysis. The Analyst 131, 1075–1078. [DOI] [PubMed] [Google Scholar]
  52. Sas KM, Karnovsky A, Michailidis G, Pennathur S, 2015. Metabolomics and diabetes: analytical and computational approaches. Diabetes 64, 718–732. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Suomalainen A, Isohanni P, 2010. Mitochondrial DNA depletion syndromes--many genes, common mechanisms. Neuromuscular disorders : NMD 20, 429–437. [DOI] [PubMed] [Google Scholar]
  54. Takano Y, Ohguro H, Dezawa M, Ishikawa H, Yamazaki H, Ohguro I, Mamiya K, Metoki T, Ishikawa F, Nakazawa M, 2004. Study of drug effects of calcium channel blockers on retinal degeneration of rd mouse. Biochem Biophys Res Commun 313, 1015–1022. [DOI] [PubMed] [Google Scholar]
  55. Thomas DD, Ridnour L, Donzelli S, Espey MG, Mancardi D, Isenberg JS, Feelisch M, Roberts DD, Wink DA, 2006. The chemistry of protein modifications elicited by nitric oxide and related nitrogen oxides, in: Dalle-Donne I, Scaloni A, Butterfield DA (Eds.), Redox Proteomics: From Protein Modifications to Cellular Dysfunction and Diseases, 1 ed, pp. 25–50. [Google Scholar]
  56. Trygg J, Holmes E, Lundstedt T, 2007. Chemometrics in metabonomics. J Proteome Res 6, 469–479. [DOI] [PubMed] [Google Scholar]
  57. Vielma AH, Retamal MA, Schmachtenberg O, 2012. Nitric oxide signaling in the retina: what have we learned in two decades? Brain Res 1430, 112–125. [DOI] [PubMed] [Google Scholar]
  58. Vighi E, Trifunovic D, Veiga-Crespo P, Rentsch A, Hoffmann D, Sahaboglu A, Strasser T, Kulkarni M, Bertolotti E, van den Heuvel A, Peters T, Reijerkerk A, Euler T, Ueffing M, Schwede F, Genieser HG, Gaillard P, Marigo V, Ekstrom P, Paquet-Durand F, 2018. Combination of cGMP analogue and drug delivery system provides functional protection in hereditary retinal degeneration. Proc Natl Acad Sci U S A 115, E2997–e3006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Wang T, Reingruber J, Woodruff ML, Majumder A, Camarena A, Artemyev NO, Fain GL, Chen J, 2018. The PDE6 mutation in the rd10 retinal degeneration mouse model causes protein mislocalization and instability and promotes cell death through increased ion influx. J Biol Chem 293, 15332–15346. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Wei T, Schubert T, Paquet-Durand F, Tanimoto N, Chang L, Koeppen K, Ott T, Griesbeck O, Seeliger MW, Euler T, Wissinger B, 2012. Light-driven calcium signals in mouse cone photoreceptors. J Neurosci 32, 6981–6994. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Westermann B, 2010. Mitochondrial fusion and fission in cell life and death. Nature reviews. Molecular cell biology 11, 872–884. [DOI] [PubMed] [Google Scholar]
  62. Williamson BK, Hawkey NM, Blake DA, Frenkel JW, McDaniel KP, Davis JK, Satija C, Beazer A, Dhungana S, Carlson J, McRitchie S, Ayyala RS, 2018. The Effects of Glaucoma Drainage Devices on Oxygen Tension, Glycolytic Metabolites, and Metabolomics Profile of Aqueous Humor in the Rabbit. Transl Vis Sci Technol 7, 14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Wong P, Borst DE, Farber D, Danciger JS, Tenniswood M, Chader GJ, van Veen T, 1994. Increased TRPM-2/clusterin mRNA levels during the time of retinal degeneration in mouse models of retinitis pigmentosa. Biochem Cell Biol 72, 439–446. [DOI] [PubMed] [Google Scholar]
  64. Xia J, Wishart DS, 2016. Using MetaboAnalyst 3.0 for Comprehensive Metabolomics Data Analysis. Current protocols in bioinformatics 55, 14.10.11–14.10.91. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1528009-supplement-1.xlsx (9.8KB, xlsx)
2
NIHMS1528009-supplement-2.xlsx (9.9KB, xlsx)
3
NIHMS1528009-supplement-3.xlsx (14.1KB, xlsx)
4
NIHMS1528009-supplement-4.xlsx (9.9KB, xlsx)
5
NIHMS1528009-supplement-5.xlsx (10.4KB, xlsx)
6
NIHMS1528009-supplement-6.xlsx (13.6KB, xlsx)
7
NIHMS1528009-supplement-7.xlsx (12.3KB, xlsx)

RESOURCES