Abstract

   Photoreceptor cells are highly susceptible to oxidative-stress-induced
   damage due to their high metabolic rate. Oxidative stress plays a key
   role in driving pathological events in several different ocular
   diseases, which lead to retinal degeneration and ultimately blindness.
   A growing number of studies have been performed to understand
   downstream events caused by ROS induced oxidative stress in
   photoreceptor cells; however, the underlying mechanisms of ROS toxicity
   are not fully understood. To shed light on ROS induced downstream
   pathological events, we employed a tandem mass tag (TMT)
   labelling-based quantitative mass-spectrometric approach to determine
   proteome changes in 661W photoreceptor cells following oxidative stress
   induction via the application of different concentrations of H[2]O[2]
   at different time points. Overall, 5920 proteins were identified and
   quantified, and 450 differentially expressed proteins (DEPs) were
   identified, which were altered in a dose and time dependent manner in
   all treatment groups compared to the control group. These proteins were
   involved in several biological pathways, including spliceosome and
   ribosome response, activated glutathione metabolism, decreased
   ECM-receptor interaction, oxidative phosphorylation, abnormally
   regulated lysosome, apoptosis, and ribosome biogenesis. Our results
   highlighted ECM receptor interaction, oxidative phosphorylation and
   spliceosome pathways as the major targets of oxidative stress that
   might mediate vascular dysfunction and cellular senescence.

   Keywords: proteomics, photoreceptor, oxidative stress, retina, ECM
   receptor interaction signalling pathway, oxidative phosphorylation,
   cellular senescence

1. Introduction

   Oxidative stress is defined as impaired balance between pro-oxidants
   and antioxidants, resulting in higher levels of reactive oxygen species
   (ROS) such as hydrogen peroxide (H[2]O[2]), nitric oxide (NO),
   superoxide (O[2]^−) and highly reactive hydroxyl radicals (HO). It
   plays a key role in the initiation and progression of several retinal
   degenerative diseases such as diabetic retinopathy (DR), glaucoma and
   age-related macular degeneration (AMD) [[46]1,[47]2]. Oxidative stress
   and accompanying neurodegeneration are also involved in several
   neurodegenerative disorders including Alzheimer’s disease (AD),
   Parkinson’s disease (PD) and Huntington’s disease (HD) [[48]3,[49]4].
   Photoreceptor cells account for about 60% of all cells in the retina;
   these neurons are also the most metabolically active neurons in the
   central nervous system, containing around 75% of all retinal
   mitochondria which are a primary source of ROS generation
   [[50]5,[51]6]. This feature, along with continuous exposure of the eye
   to environmental stimuli such as varying light intensities, UV
   radiation, smoke, air pollution and oxygen, makes these cells
   vulnerable to oxidative stress insults. Cumulative evidence suggests
   that oxidative stress contributes to DNA, lipid and protein damage and,
   consequently, cell death. However, the underlying mechanisms of these
   pathological events following oxidative stress in photoreceptor cells
   remain ill-defined [[52]7,[53]8].

   The accumulation of β-amyloid (Aβ) in the brain, known as the amyloid
   plaque, is a key pathological hallmark of AD that is also reported in
   the retina of glaucoma, DR and AMD subjects, suggesting overlapping
   mechanisms between AD and these conditions. Interestingly, Aβ has been
   suggested as a trigger of oxidative stress in these diseases
   [[54]9,[55]10]. In this regard, a previous proteomics study by our
   group showed increased oxidative stress in retinal photoreceptor cells
   upon treatment with Aβ, and investigated the underlying mechanisms of
   Aβ toxicity on photoreceptor cells [[56]11]. However, increased
   oxidative stress has also been observed in the absence of Aβ, and it
   has also been suggested that it may promote the aggregation and
   accumulation of the peptide [[57]12,[58]13,[59]14,[60]15,[61]16].
   Therefore, exploring the oxidative stress induced changes independent
   of Aβ accumulation can provide unique insights into
   oxidative-stress-derived pathological events.

   Herein, we studied the proteome changes in photoreceptor cells in
   oxidative stress conditions induced by different concentrations of
   H[2]O[2] in different times of exposure, using tandem mass tag (TMT)
   quantitative proteomics. We also compared the identified differentially
   expressed proteins (DEPs) with previously reported DEPs in
   photoreceptor cells upon treatment with Aβ, to reveal the common and
   distinct altered pathways.

2. Materials and Methods

2.1. Cell Culture and Treatments

   Mouse photoreceptor-derived 661W cells were kindly provided by Prof.
   Al-Ubaidi, University of Oklahoma. The cells were cultured in DMEM
   medium containing 10% (v/v) fetal bovine serum (FBS), 1% (v/v)
   penicillin/streptomycin (Thermo Fisher, San Jose, CA, USA). Cultures
   were maintained at 37 °C under a humidified atmosphere of 5% CO[2], as
   previously described [[62]11,[63]17]. The cells were exposed to two
   different concentrations of H[2]O[2] (Sigma, St Louis, MO, USA),10 µM
   and 50 µM, and then samples were separately collected at different
   times of exposure; after 0.5 h in treatments with 10 µM H[2]O[2], and
   after 6 h and 12 h in treatments with 50 µM H[2]O[2]. Therefore, there
   were four distinct sampling points including control (CTRL),
   concentration 1 (C1) (10 µM, 0.5 h), concentration 2 (C2) (50 µM, 6 h),
   and concentration 3 (C3) (50 µM, 12 h). Three biological replicates
   were prepared for each specific condition. Cells were stored at −20 ℃
   until protein extraction.

2.2. Protein Sample Preparation

   Protein samples were prepared as described previously [[64]11].
   Briefly, cells were solubilized in cold lysis buffer (20 mM HEPES, pH
   7.4, 150 mM NaCl, 1% Triton x-100, 1 mM EDTA) containing protease
   inhibitor under sonication. Lysates were centrifuged for 15 min at
   13,000 rpm at 4 ℃ and supernatant was collected, reduced with
   dithiothreitol (20 mM) for 30 min, and alkylated with 50 mM
   iodoacetamide for 30 min in the dark at room temperature. An additional
   40 mM DTT was added for 15 min to quench the alkylation reaction.
   Protein precipitation and contaminant removal were performed using
   chloroform–methanol extraction [[65]18]. Samples were air-dried for
   evaporation of the remaining methanol and then resuspended in 650 μL
   HEPES buffer (pH 8.8). Total protein quantities were measured using a
   BCA Protein Assay kit (Pierce, Rockford, IL, USA). An aliquot of 150 μg
   of proteins was digested with trypsin (protein: enzyme 50:1) at 37 °C
   overnight. Digestion was stopped by placing samples on ice and
   adjusting to a final concentration of 1 mM trifluoroacetic acid,
   followed by measurement of the peptide concentration using a Micro BCA
   assay kit (Pierce, Rockford, IL, USA).

2.3. TMT Labelling, LC-MS/MS Analysis and Peptide to Spectrum Matching

   An aliquot of 60 μg peptides from each sample was labelled with 0.25 mg
   of TMT reagent (Thermo Fisher, San Jose, CA, USA) at room temperature
   for one hour with occasional vortexing. To quench the reaction, 8 μL of
   5% fresh hydroxylamine was added to each sample, and then it was
   incubated at room temperature for 15 min with additional vortexing. TMT
   labelling was performed using three separate 10-plex labelling kits, as
   part of a larger scale experiment, so only the first four channels of
   labelling were used in each set. For the first set of TMT labels,
   designated R1, the labels were used as follows: control R1 = 126 C,
   (C1) (10 µM, 0.5 h) R1 = 127 N, (C2) (50 µM, 6 h) R1 = 127 C, and (C3)
   (50 µM, 12 h) R1 = 128 N. For the second set of TMT labels, designated
   R2, the labels were used as follows: control R2 = 126 C, (C1) (10 µM,
   0.5 h) R2 = 127 N, (C2) (50 µM, 6 h) R2 = 127 C, and (C3) (50 µM, 12 h)
   R2 = 128 N. For the third set of TMT labels, designated R3, the labels
   were used as follows: control R3 = 126 C, (C1) (10 µM, 0.5 h) R3 = 127
   N, (C2) (50 µM, 6 h) R3 = 127 C, and (C3) (50 µM, 12 h) R3 = 128 N.

   Following the labelling reaction, peptide samples were combined and
   evaporated to dryness in a vacuum centrifuge. The TMT labelled peptide
   mixture was redissolved in 1% formic acid and a Sep-Pak C18 cartridge
   (Waters, Milford, MA, USA) was employed to de-salt the mixture. TMT
   labelled peptides were then separated using a high pH reversed-phase
   peptide fractionation kit (Thermo Fisher, San Jose, CA, USA), with a
   total of 17 peptide containing fractions collected. These were
   evaporated to dryness, and peptides were redissolved in 1% formic acid
   and desalted for a second time using SDB-RPS (3M-Empore) Stage Tips. A
   nano flow liquid chromatography—tandem mass spectrometry system
   consisting of an EASY-nLC1000 nanoflow HPLC system (Thermo Scientific,
   San Jose, CA, USA) coupled to a Q Exactive Orbitrap mass spectrometer
   (Thermo Scientific, San Jose, CA, USA) was employed for the acquisition
   of mass spectrometric data for the identification and quantification of
   peptides.

   Peptide to spectrum matching was performed using Sequest HT and Mascot
   algorithms in Proteome Discoverer V2.1 (Thermo Scientific, San Jose,
   CA, USA). Spectra were searched against the SwissProt Mus Musculus
   protein sequence database ([66]http://www.ebi.ac.uk/swissprot/ (16,953
   sequences, accessed on 16 February 2019)), with parameters as described
   previously [[67]11,[68]19,[69]20]. Parent ion MS tolerance was
   specified as ±10 ppm, while fragment ion MS/MS tolerance was specified
   as 0.02 Da. Carbamidomethylation of cysteine was considered to be a
   fixed modification. A range of variable modifications were permitted,
   including Deamidation (N, Q), Acetylation (Protein N-Terminus), Glu- >
   pyro-Glu (N-term E), Gln- > pyro-Glu (N-term Q), Oxidation (M), and TMT
   10-plex labelling (N-term, K). The discrimination of peptide to
   spectrum matches, which involved calculating statistics including
   posterior error probabilities and Q-values, was achieved using the
   Percolator algorithm. Further filtering was performed to retain only
   peptides and proteins with FDR less than 1%. A summary of all identify
   proteins is included in [70]Supplementary Table S1, while complete
   details of all identify proteins are included in [71]Supplementary
   Table S2, including Sequest HT and Mascot Scores, peptide sequences and
   modifications, and details of which high pH reversed-phase peptide
   fractions the labelled peptides were identified in.

2.4. Analysis of Differentially Expressed Proteins

   TMTPrepPro [[72]21] was used for further quantitative analysis of
   identified proteins. The control reference (label 126) was used for the
   calculation of protein ratios, which were extracted from individual
   experiments and combined across all the runs involved. Differentially
   expressed proteins (DEPs) were identified across all conditions using
   two-way analysis of variance (ANOVA). Additional comparisons were
   performed using pairwise analysis of each treatment relative to control
   using the Student’s t-test. Abundance ratios of DEPS identified in
   ANOVA comparisons were log transformed prior to clustering. In
   comparisons performed in a pairwise manner, the relative quantitation
   of protein after peroxide treatment of cells was performed using the
   ratio of the detected TMT label in each of the treatments compared to
   the matched control. Geometric means of the respective ratios were used
   to calculate overall fold changes for individual proteins.

   Proteins were considered to be DEPs if they met the following criteria:
   fold change was greater than 1.20 (increased) or less than 0.833
   (decreased), with p-value less than 0.05 [[73]22,[74]23]. Finally, the
   identified DEPs were subjected to further bioinformatics analyses,
   including Gene Ontology (GO) and pathway enrichment analysis using the
   online tool String [[75]24,[76]25].

3. Results

3.1. Proteome Profiling of Photoreceptor Cells and Significantly Regulated
Proteins by H[2]O[2] Treatments

   A total of 5920 proteins was identified from 661W photoreceptor cells
   in this study ([77]Supplementary Table S1). These proteins were
   quantified with multiple peptides at an initial protein FDR less than
   1%. There were 410 proteins identified as differentially abundant under
   different conditions based on the ANOVA analysis (which met the
   requirements of both p-value < 0.05 and the absolute value of fold
   change > 1.2). Hierarchical clustering analysis was applied to those
   differentially expressed proteins and showed similar expression
   patterns of proteins increased or decreased in abundance between 6 h
   and 12 h of 50 µM H[2]O[2] treatments (C2 and C3, respectively)
   ([78]Figure 1). However, fewer proteins were changed in abundance after
   0.5 h of 10 µM H[2]O[2] treatment (C1), indicating that the cellular
   response under this relatively mild condition was quite distinct to
   that observed for the other two H[2]O[2] treatments.

Figure 1.

   [79]Figure 1
   [80]Open in a new tab

   Heatmap (hierarchical clustering) of the log-transformed ratios of
   differentially expressed proteins (differences between all experimental
   conditions via analysis of variance (ANOVA)) after H[2]O[2] treatments,
   row clustering only. Column colours indicate treatment types.

   The pairwise analysis between the mildest H[2]O[2] treatment (10 µM,
   0.5 h) versus control yielded 22 increased (p-value < 0.05 and fold
   change > 1.2) and 34 decreased proteins (p-value < 0.05 and fold change
   < −1.2), respectively ([81]Figure 2a and [82]Supplementary Table S3).
   In comparison, a greater portion of proteins was differentially
   expressed when cells were treated with a higher concentration of
   H[2]O[2] (50 µM). There were 90 proteins increased, and 86 proteins
   decreased in abundance, after cells were exposed to 50 µM H[2]O[2] for
   6 h ([83]Figure 2b). More proteins responded to H[2]O[2] when cells
   were treated for a longer duration (12 h) at this concentration (50
   µM), with 175 and 181 proteins being increased and decreased in
   abundance, respectively ([84]Figure 2c).

Figure 2.

   [85]Figure 2
   [86]Open in a new tab

   (a–c) Volcano plots demonstrating the dual thresholds for
   differentially regulated proteins in photoreceptor cells under three
   different H[2]O[2] treatments: C1 (10 µM, 0.5 h), C2 (50 µM, 6 h), and
   C3 (50 µM, 12 h). Each data point represents a single quantified
   protein. The x-axis represents log fold change in abundance (H[2]O[2]
   treatment vs. control). Vertical blue lines indicate ratios 1.2 and
   0.833. The −log (p-value) is plotted on the y-axis. Proteins above the
   red horizontal line indicate significance lower than 0.05. Proteins
   within the upper and outer quadrants meet both the fold change and
   p-value cut-off, and are therefore considered to be differentially
   abundant. (d) Top 10 differential proteins in cells under three
   different H[2]O[2] treatments, sorted by fold change values. Proteins
   increased in abundance by H[2]O[2] treatment are highlighted in orange,
   with colour intensity corresponding to fold change values, while
   proteins decreased in abundance by H[2]O[2] treatment are highlighted
   in blue, with colour intensity corresponding to fold change values.

   The top 10 differentially expressed proteins according to fold change
   values in photoreceptor cells under three specific H[2]O[2] treatments
   were further analysed ([87]Figure 2d). The fold changes of the
   differentially abundant proteins ranged from 2.4 to −2.04 under C1 (10
   µM, 0.5 h), 2.49 to −4.34 under C2 (50 µM, 6 h), and 5.58 to −4.48
   under C3 (50 µM, 12 h). This indicates that impacts of H[2]O[2]-induced
   oxidative stress on photoreceptor cells increased when they were
   exposed to a higher concentration of H[2]O[2] and treated for a longer
   time. Among these, four proteins were differentially abundant in all
   three conditions, including Ccn1, C4b, Pja1, and Pdcd4. Ccn1, a 40-kDa
   heparin-binding protein rich in cysteine residues and known as a
   connective tissue growth factor, was increased in abundance in C1 (10
   µM, 0.5 h) where a lower concentration of H[2]O[2] and a shorter time
   were applied to cells.

   However, the abundance of Ccn1 significantly decreased when cells were
   treated with 50 µM H[2]O[2] for either 6 or 12 h. C4 binding protein
   (C4b), which is essential for the propagation of the classical
   complement pathway, showed increased expression in cells under all
   these three treatment conditions. Pja1 protein is one of the
   RING-finger E3 ligases that are instrumental in the regulation of
   inflammatory cascades, apoptosis, and cancer. A reduced abundance of
   this protein was found in cells after all three H[2]O[2] treatments.
   Similarly, the abundance of programmed cell death protein 4 (Pdcd4)
   also decreased in all conditions of oxidative stress. Pdcd4 showed the
   greatest decrease in abundance in response to all H[2]O[2] treatments
   (average fold change of −3.62), while 40S ribosomal protein S30 (FAU)
   showed the greatest increase in abundance (5.58-fold change), in C3.

   A total of 450 proteins were differentially abundant in response to the
   three treatments when compared with the controls using pairwise
   analysis. Similar yet different changes in protein abundance were found
   among three treatments with changes to applied concentrations of
   H[2]O[2] and exposure time ([88]Figure 3a,b). Changes in protein
   abundance generally increased with the more concentrated H[2]O[2] and
   treatment time. The common differentially abundant proteins between any
   two treatments were 26 proteins between C1 and C2, 20 proteins between
   C1 and C3, and 109 proteins between C2 and C3, respectively. The
   expression patterns of most common proteins in two treatments were the
   same with either increase or decrease in abundance ([89]Figure 3c,d). A
   total of 17 proteins were commonly identified as differentially
   expressed in all three H[2]O[2] treatments, with 11 of these 17
   proteins decreased in abundance, while five were increased in
   abundance. Ccn1 was the only protein with decreased abundance in C1 (10
   µM, 0.5 h), but significantly increased abundance in C2 (50 µM, 6 h)
   and C3 (50 µM, 12 h), as mentioned above. All of these 17 proteins
   showed very similar expression patterns between C2 and C3 when a higher
   concentration and longer treatment time were applied to photoreceptor
   cells.

Figure 3.

   [90]Figure 3
   [91]Open in a new tab

   (a) Upset plot showing the overlap of proteins with either increased or
   decreased abundance in three different treatments. Blue and orange
   colours separately represent the decreased and increased abundance of
   proteins in photoreceptor cells under H[2]O[2] treatment. (b)
   Distribution of differentially abundant proteins relative to control in
   C1, C2 and C3. (c) Seventeen common altered proteins and their
   expression changes in C1, C2, and C3. (d) Heatmap indicating the
   expression pattern of 450 proteins with significant changes in
   abundance in photoreceptors under specific treatments. Orange and blue
   colours represent an increase in abundance (t-test, FC > 1.2, p-value <
   0.05) and decrease in abundance (t-test, FC < −1.2, p-value < 0.05),
   respectively.

3.1.1. GO Analysis of Regulated Proteins

   Gene ontology analysis indicated the response of proteins that were
   classified to a specific cellular component, biological process, and
   molecular function ([92]Figure 4a). More than 25 proteins belonging to
   the ribonucleoprotein complex responded with increased abundance to the
   H[2]O[2] treatment after 12 h. However, more than 25 proteins involved
   in regulating various aspects of the extracellular matrix, including
   eight collagen proteins, were decreased in abundance under the same
   conditions ([93]Figure 4a,b). The number of responsive proteins in
   different groups mostly increased with the increasing concentration of
   H[2]O[2] treatment, such as components from ribosome, lysosome, and
   actin cytoskeleton ([94]Figure 4a).

Figure 4.

   [95]Figure 4
   [96]Open in a new tab

   (a) GO analysis showing the number of proteins increased in abundance
   (orange colour) and decreased in abundance (blue colour), classified
   into the specific cellular component, biological pathway, and molecular
   function in photoreceptor cells under H[2]O[2] treatments. Different
   shapes of markers indicate specific treatment (square—C1, triangle—C2,
   round—C3). (b) Significantly decreased collagen proteins in C3 (50 µM,
   12 h) by H[2]O[2] (fold change < −1.2, p-value < 0.05) and their
   interactions.

3.1.2. Pathway Classification of Differentially Abundant Proteins

   The network of interacting proteins that were differentially expressed
   among all groups based on ANOVA was generated using String. KEGG
   pathway analysis was used to display the altered pathways, including
   ribosome, spliceosome, lysosome, ECM-receptor interaction, Alzheimer’s
   disease, apoptosis, ribosome biogenesis in eukaryotes, glutathione
   metabolism, and oxidative phosphorylation ([97]Figure 5a,b). As
   expected, C3 (50 µM, 12 h) treatment produced the maximum number of
   molecules regulated in these KEGG pathways. Most ribosome and
   spliceosome proteins were increased in abundance. In addition, three
   glutathione S-transferases—Gstm1, Gstm2 and Gstm5—were found with
   increased abundance under C3, indicating activated glutathione
   metabolism. In addition, proteins involved in extracellular matrix
   (ECM)-receptor interaction were decreased in abundance. Oxidative
   phosphorylation was disrupted, with seven out of eight proteins in this
   pathway decreased in abundance by longer and more concentrated H[2]O[2]
   treatment. Amyloid precursor protein (APP), belonging to the
   Alzheimer’s pathway, was decreased in abundance in cells with increased
   H[2]O[2] treatments, and had fold changes of −1.14, −1.35, and −1.63 in
   C1, C2, and C3, respectively (all p-value < 0.05). Another protein in
   this pathway, presenilin 1 (Psen1), was also decreased in abundance
   under C2. Intriguingly, membrane metallo-endopeptidase (MME), which is
   considered one of the important β-amyloid (Aβ)-degrading enzymes
   related to prevention of Alzheimer’s disease (AD) pathology [[98]26],
   was identified with increased abundance in all H[2]O[2] treatment
   groups. This may occur as part of an induced compensatory response
   against H[2]O[2] toxicity, which in turn decreased levels of target
   proteins including β-amyloid [[99]27]. Additionally, the increased
   abnormal regulation of proteins in KEGG pathways related to lysosomes,
   apoptosis, and ribosome biogenesis was identified with more severe
   H[2]O[2] treatment as well.

Figure 5.

   [100]Figure 5
   [101]Open in a new tab

   (a) Network of protein interactions and regulated KEGG pathways by
   H[2]O[2] treatment. (b) The change in abundance of proteins
   demonstrating the response of specific KEGG pathways, including
   increased spliceosome and ribosome response, activated glutathione
   metabolism, decreased ECM-receptor interaction and oxidative
   phosphorylation, abnormally regulated lysosome, apoptosis, and ribosome
   biogenesis.

3.1.3. Analysis of Proteins Found to Be Differentially Abundant in 661W
Photoreceptor Cells by Both H[2]O[2] and Aβ Treatments

   The results of this study were compared with those of our previously
   published study, in which the same cell line was treated with different
   concentrations of Aβ peptides. A total of 4636 out of 5837 proteins
   identified in photoreceptor cells under Aβ treatments were also found
   in this study. Among those, 61 proteins were changed in abundance in
   661W photoreceptor cells by both H[2]O[2] and Aβ treatments with either
   different concentrations or treatment time-points ([102]Supplementary
   Figure S1). Similar yet distinct changes in abundance and proteome
   complexity were identified between these two studies on photoreceptor
   cells, which separately mimicked the impact of Alzheimer’s disease and
   oxidative stress on the eye. For example, the Aβ degrading enzyme and
   membrane protein Mme was increased in abundance in both studies when
   cells were treated with either H[2]O[2] induced oxidative stress or
   longer Aβ treatment (12 h), which was believed to cause more severe
   oxidative stress-mediated toxic effects. However, cathepsin D (Ctsd),
   which is a lysosomal acid protease and plays a role in APP processing,
   the clearance of Aβ, and intracellular protein breakdown, was decreased
   in abundance in cells under the more severe Aβ (25 µm, 12 h) and
   H[2]O[2] (50 µM, 12 h) treatments. Further GO analysis classified 34
   out of these 61 proteins into specific groups, including different
   biological processes, molecular functions, or cellular components
   ([103]Supplementary Figure S2). This analysis identified six and nine
   components in the biological process of ribosome biogenesis and RNA
   processing, as well as 11 proteins with RNA binding molecular
   functions. In addition, the cellular component analysis indicated that
   five, six, and eight differentially abundant proteins belonged to the
   actin cytoskeleton, ribosomal subunit, and lysosomal networks,
   respectively.

4. Discussion

   Photoreceptor cells, including rods and cones, are retinal
   neuroepithelial cells that are crucial for visual phototransduction.
   Oxidative stress is now well established as a major player in
   photoreceptor cell degeneration; however, the downstream changes are
   not well defined [[104]28,[105]29]. We performed this study to
   investigate downstream events at the protein level in a retinal cone
   cell line (661W) following exposure to different concentrations of an
   oxidative stress inducer at different time points.

   Results of the present study showed that H[2]O[2] exposure altered the
   proteome complexity of 661W photoreceptor cells in a time and
   concentration dependent manner. Overall, 450 DEPs were identified in
   all treatment groups and several significantly altered biological
   pathways were identified, including ribosome, spliceosome, ECM-receptor
   interaction, oxidative phosphorylation, and glutathione metabolism.
   Moreover, 17 of these 450 DEPs were common between all three treatment
   groups, which were mainly involved in protein synthesis.

   The alteration of ribosomal proteins in retinal disorders has been
   reported previously [[106]30,[107]31]. In C1 treatment, a small number
   of ribosomal proteins were decreased in abundance while a greater
   proportion of proteins was increased in abundance when exposed to
   higher concentration of H[2]O[2] for a longer time. In agreement with
   our results, the down-regulation of several ribosomal genes was
   observed in a rat model of chronic glaucoma with early nerve injury,
   while more ribosomal genes were found to be up-regulated with advanced
   nerve injury [[108]32]. The increased expression of ribosomal genes has
   also been observed in retinal pigment epithelium (RPE)-choroid-sclera
   samples of AMD patients and aqueous humour (AH) of primary open-angle
   glaucoma (POAG) patients [[109]29]. Ribosomes are very important in
   protein synthesis, so a decrease in the abundance of some ribosomal
   proteins at lower concentrations of H[2]O[2] is possibly due to initial
   induced stress, while exposure to higher concentrations for longer time
   resulted in a considerable number of proteins increasing in abundance,
   which might be due to the increased production of proteins involved in
   spliceosome, glutathione metabolism and apoptosis, or might be a
   compensatory response for cell survival [[110]33]. In addition, a
   decrease in the abundance of Pdcd4 (which inhibits translation
   initiation) in all treatment groups also correlates with increased
   protein production [[111]34].

   The alteration of RNA splicing has been previously reported in human
   samples of POAG and AMD by our group and others
   [[112]35,[113]36,[114]37]. As shown in [115]Figure 5b, treatment with
   10 µM H[2]O[2] for 0.5 h resulted in a decreased abundance of several
   spliceosome proteins which were then shifted towards an increase in
   abundance following exposure to higher concentrations of H[2]O[2] (50
   µM, 6 h), and which increased even more upon a longer time of exposure
   (50 µM, 12 h). This change in abundance of the spliceosome proteins,
   particularly U2af2 and Prpf19, might be a compensatory response
   reflecting the initiation of global spliceosome activity in the early
   stage of cellular senescence, which is triggered by H[2]O[2] induced
   oxidative stress [[116]38,[117]39].

   The ECM consists of collagens, proteoglycans and glycoproteins that
   support the maintenance of cell structure and provide a scaffold
   essential for maintaining the organization of vascular endothelial
   cells into blood vessels. Interactions between ECM and cells are
   crucial for cell adhesion, migration, differentiation, proliferation,
   and apoptosis [[118]40,[119]41]. We observed a significant decrease in
   the abundance of collagen and laminin proteins, including col5a1/2,
   col6a1/2, col4a2/6, col1a1 col3a1, lamb2, and lamc1, especially in C3
   treatment. The down-regulation of genes involved in collagen synthesis
   and extracellular matrix organization has been reported in both human
   samples and cell models of keratoconus (KTCN), another eye disorder
   [[120]42,[121]43]. Laminin and collagen are involved in vasculature and
   blood vessel development [[122]44,[123]45]. The outer retina,
   containing photoreceptor cells, is known to be the most metabolically
   active tissue in the body, and the choroidal and retinal vasculature is
   developed to meet the specialized retina metabolic requirements. The
   retina has two vascular systems, including a central retinal artery in
   the inner retina and the choriocapillaris in the retinal pigment
   epithelium and outer retina [[124]46]. Choroidal vascular dysfunction
   and oxidative stress are two well-established factors in AMD
   pathogenesis [[125]47,[126]48], but little is currently known about the
   interplay between these two pathophysiological factors. Herein, our
   results showed that H[2]O[2]-induced oxidative stress resulted in a
   decreased abundance of collagen and laminin proteins. It is of interest
   that the decreased expression of ECM components, particularly collagen
   genes including COL1A1, COL1A2, and COL3A1, has also been reported in
   retinal pigment epithelial cells upon cellular senescence
   [[127]49,[128]50]. Moreover, the down-regulation of collagen genes such
   as COL3A1, COL4A1, COL4A2, and COL5A1 has been reported following
   cellular senescence induction in hepatic stellate cells and COL16A1 in
   human dermal fibroblast cells [[129]51,[130]52].

   Mitochondrial dysfunction, and consequently impaired oxidative
   phosphorylation, is another well-known feature of cellular senescence
   [[131]53] and has also been observed in this study. Oxidative
   phosphorylation is an important metabolic pathway and a major source of
   energy that occurs in the mitochondria. Oxidative phosphorylation
   impairment due to mitochondrial dysfunction has been widely reported in
   many eye diseases such as DR, glaucoma and AMD
   [[132]54,[133]55,[134]56,[135]57]. Therefore, mitochondrial dysfunction
   has become a well-established part of these disorders that drive
   disease progression. Photoreceptor cells require large amounts of ATP,
   and therefore healthy mitochondria are crucial for cellular survival
   and function, and impaired mitochondrial function leads to
   photoreceptor degeneration [[136]58,[137]59].

   Finally, another biochemical module that was enriched was glutathione
   metabolism, which remained relatively unchanged in C1 (10 µM, 6 h) but
   was significantly increased in abundance in C2 (50 µM, 6 h) and further
   increased with increasing time of exposure in C3 (50 µM, 12 h).
   Glutathione is the most abundant antioxidant molecule within the cell,
   and counteracts the effects of ROS and H[2]O[2] to prevent cellular
   damage. Although glutathione has been reported to be decreased in DR,
   AMD and glaucoma [[138]60], it was increased in abundance in our study,
   which has also been reported in early-stage AMD patients. This is
   possibly due to an initial compensatory response to neutralize the
   increased H[2]O[2]-induced oxidative stress and prevent cell senescence
   [[139]61].

   Interestingly, comparing these data with our previous findings in the
   same photoreceptor cell line upon treatment with Aβ [[140]11] yielded
   61 DEPs that were common between all treated groups in the two studies
   ([141]Supplementary Figure S1). These proteins were involved in
   ribosome biogenesis, ribosomal subunits, lysosome, actin cytoskeleton,
   and RNA binding and processing ([142]Supplementary Figure S2). In
   addition to these pathways, oxidative phosphorylation was also enriched
   in both studies. Intriguingly, these alterations were observed in
   decreased levels of β-amyloid, and therefore our results support the
   key role of oxidative stress in neurodegenerative diseases that may
   precede β-amyloid deposition [[143]62].

5. Conclusions

   The results of the present study highlight the biological pathways
   affected by oxidative stress in photoreceptor cells, and provide a set
   of differentially abundant proteins which are a valuable resource for
   future animal and human cell studies focused on oxidative stress
   response. In addition, our findings suggest the impairment of
   ECM-receptor interaction, oxidative phosphorylation, and spliceosome as
   the major targets of H[2]O[2]-induced oxidative stress, which might
   contribute to vascular dysfunction and photoreceptor cell degeneration.

6. Limitations

   This study has some inherent limitations due to the fact that it is a
   bottom-up proteomics analysis which involves first breaking proteins
   down into peptides. We acknowledge that this means this manuscript
   contains little or no information regarding different proteoforms,
   because that information is mostly lost when protease digestion is
   employed in the first stage of sample preparation. There is also little
   information in this manuscript on protein species, because we did not
   attempt to address any particular posttranslational modifications;
   rather, the study was focused on individual gene products.

   Proteomes are, by their very nature, complex systems. Describing the
   components of a complex system in terms of what is there and how much
   of it is present under different conditions is an essential first step
   in understanding how a complex system works. Comparative quantitative
   proteomics studies such as the one presented here are essential
   building blocks in understanding the overall proteome complexity
   involved in different cells and how they respond to external stress.
   Despite these caveats, identifying and quantifying proteins changed in
   abundance in photoreceptor cells’ responses to hydrogen peroxide
   demonstrates the power of proteomic analysis in the discovery of
   biomarkers of oxidative stress, and provides an important foundation
   for further investigation.

Supplementary Materials

   The following supporting information can be downloaded at:
   [144]https://www.mdpi.com/article/10.3390/proteomes11020012/s1.
   Supplementary Table S1—Details of all 5920 proteins identified from
   661W photoreceptor cells in this study; Supplementary Table S2—Complete
   details of identified proteins identified in this study including
   peptide sequences, and Sequest HT and Mascot Scores; Supplementary
   Table S3—Differentially abundant proteins identified in all three
   treatments; Supplementary Figure S1—Heat map table of differentially
   expressed proteins in 661W cells treated by H[2]O[2] and Aβ when
   compared to the control; Supplementary Figure S2—Network analysis of 61
   commonly regulated proteins by both H[2]O[2] and Aß in 661
   photoreceptor cells.
   [145]Click here for additional data file.^ (88MB, zip)

Author Contributions

   Conceptualization, M.M., S.L.G. and P.A.H.; methodology, L.D., V.G.
   (Vivek Gupta), M.A. and J.A.P.; formal analysis, V.G. (Veer Gupta),
   Y.Y., Y.W., K.P., N.C. and L.D.; investigation, L.D., M.A., K.P., Y.W.
   and N.C.; resources, M.M., V.G. (Vivek Gupta) and P.A.H.; data
   curation, L.D., M.M. and P.A.H.; writing—original draft preparation,
   L.D., M.M. and P.A.H.; writing—review and editing, L.D., V.G. (Vivek
   Gupta), M.A., N.C., K.P., Y.W., V.G. (Veer Gupta), Y.Y., J.A.P.,
   S.L.G., M.M. and P.A.H.; supervision, M.M., S.L.G., V.G. (Vivek Gupta)
   and P.A.H.; project administration, M.M., S.L.G., V.G. (Vivek Gupta)
   and P.A.H.; funding acquisition, M.M., S.L.G., V.G. (Vivek Gupta) and
   P.A.H. All authors have read and agreed to the published version of the
   manuscript.

Institutional Review Board Statement

   Not applicable.

Informed Consent Statement

   Not applicable.

Data Availability Statement

   Mass spectrometric data are available via ProteomeXchange with the
   identifier PXD037329.

Conflicts of Interest

   M.A. was employed by ProGene Technologies. The remaining authors
   declare that the research was conducted in the absence of any
   commercial or financial relationships that could be construed as a
   potential conflict of interest.

Funding Statement

   We acknowledge support from the Ophthalmic Research Institute of
   Australia, National Health and Medical Research Council (NHMRC) and
   Hillcrest Foundation and Macquarie University (MQRDG). The mass
   spectrometry analysis in this study was conducted at the Australian
   Proteome Analysis Facility supported by the Australian Government’s
   National Collaborative Research Infrastructure Scheme (NCRIS).

Footnotes

   Disclaimer/Publisher’s Note: The statements, opinions and data
   contained in all publications are solely those of the individual
   author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI
   and/or the editor(s) disclaim responsibility for any injury to people
   or property resulting from any ideas, methods, instructions or products
   referred to in the content.

References