Accepted Manuscript Retinal proteome changes following experimental branch retinal vein occlusion and intervention with Ranibizumab Lasse Jørgensen Cehofski, Anders Kruse, Martin Bøgsted, Sigriður Olga Magnusdottir, Allan Stensballe, Bent Honoré, Henrik Vorum PII:
S0014-4835(16)30244-5
DOI:
10.1016/j.exer.2016.09.002
Reference:
YEXER 7018
To appear in:
Experimental Eye Research
Received Date: 8 January 2016 Revised Date:
7 August 2016
Accepted Date: 8 September 2016
Please cite this article as: Cehofski, L.J., Kruse, A., Bøgsted, M., Magnusdottir, S.O., Stensballe, A., Honoré, B., Vorum, H., Retinal proteome changes following experimental branch retinal vein occlusion and intervention with Ranibizumab, Experimental Eye Research (2016), doi: 10.1016/ j.exer.2016.09.002. 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 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.
ACCEPTED MANUSCRIPT
Branch retinal vein occlusion and anti-VEGF treatment Laser induced BRVO in both eyes
Right eye anti-VEGF
Dissection
Left eye 0.9% NaCl
SC
RI PT
Argon laser
0.200
0.200
0.150
0.150
0.050
0.050
Cytosol
AC C
1
EP
TE D
Retinal branch vein
2
Integrin-β1
Extracellular matrix
M AN U
0.10
0.10
0
Experimental BRVO and inhibition of VEGF-A
3
Adhesion Migration Proliferation
In the present study branch retinal vein occlusion (BRVO) was induced in the inferior retina in both eyes of the pigs. After 24 hours vascular endothelial growth factor A (VEGF-A) was inhibited in the right eyes through an injection with Ranibizumab whilst the left eyes were treated with 0.9% sodium chloride. Retinas were excised after 3 days. In retinas treated with Ranibizumab protreomic analyses revealed a downregulation of integrin β-1 suggesting that inhibition of VEGF-A downregulated adhesion, migration and proliferation processes associated with BRVO.
ACCEPTED MANUSCRIPT
RI PT
Retinal Proteome Changes Following Experimental Branch Retinal Vein Occlusion and Intervention with Ranibizumab
SC
Lasse Jørgensen Cehofski1,2,3, Anders Kruse1, Martin Bøgsted3,4, Sigriður Olga Magnusdottir2, Allan Stensballe5, Bent Honoré3,6 and Henrik Vorum1,3
Department of Ophthalmology, Aalborg University Hospital, Aalborg, Denmark
2
Biomedical Research Laboratory, Aalborg University Hospital, Aalborg, Denmark
3
Department of Clinical Medicine, Aalborg University, Aalborg, Denmark
4
Department of Haematology, Aalborg University Hospital, Aalborg, Denmark
5
Department of Health Science and Technology, Aalborg University, Denmark
6
Department of Biomedicine, Aarhus University, Aarhus, Denmark
AC C
EP
Correspondance to: Lasse Jørgensen Cehofski, Department of Ophthalmology, Aalborg University Hospital, Hobrovej 18-22, 9100 Aalborg, Denmark. Phone +45 53558878 Fax +45 97662581 E-mail [email protected]
TE D
M AN U
1
1
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
Abstract Animal models of experimental branch retinal vein occlusion (BRVO) provide a unique opportunity to study protein changes directly in retinal tissue. Results from these experimental models suggest that experimental BRVO is associated with an upregulation of extracellular matrix remodeling and adhesion signaling processes. To study whether these processes could be blocked by inhibition of VEGF-A, a porcine model of experimental BRVO was combined with proteomic analyses. In six Danish Landrace pigs experimental BRVO was induced with argon laser in both eyes. After 24 hours an injection of 0.05 mL Ranibizumab was given in the right eyes of the animals while left eyes received an injection of 0.05 mL 9 mg/mL sodium chloride water. Retinas were dissected three days after BRVO and the retinal samples were analyzed with label-free quantification as well as tandem mass tag based proteomics. In retinas treated with Ranibizumab five proteins exhibited statistically significant changes in content with both proteomic techniques. These five proteins, which were all decreased in content, included integrin β-1, peroxisomal 3-ketoacyl-CoA thiolase, OCIA domain-containing protein 1, calnexin and 40S ribosomal protein S5. As anti-integrin therapies are under development for inhibition of angiogenesis in retinal diseases it is interesting that inhibition of VEGFA in itself resulted in a small decrease in the content of integrin β-1. The decreased content of integrin β-1 indicates that extracellular matrix remodeling and adhesion processes associated with BRVO are at least partly reversed through inhibition of VEGF-A.
AC C
EP
TE D
1. Introduction Studies on protein changes following branch retinal vein occlusion (BRVO) have traditionally been conducted on aqueous humor and the vitreous body samples from patients with macular edema secondary to BRVO. These analyses have led to the identification of important proteins that are involved in the formation of macular edema (Campochiaro et al., 2009; Noma et al., 2011; Noma et al., 2012; Noma et al., 2014). Animal models of experimental BRVO allow for protein studies to be conducted directly on retinal tissue where major pathological changes are thought to take place (Cehofski et al., 2015a; Cehofski et al., 2014). We recently published a proteomic study on large-scale protein changes in porcine retinas 15 days after exposure to experimental BRVO (Fig 1A) and proposed the hypothesis that experimental BRVO was associated with an upregluation of proteins involved in extra cellular matrix (ECM) remodeling and focal adhesion processes (Cehofski et al., 2015b). Proteins involved in ECM remodeling and focal adhesion signaling included laminin subunit β-2, laminin subunit γ-1, lipocalin-7, nidogen-2, osteopontin, integrin β-1, isoform 2 of α-actinin-1, talin-2 and filamin C (Cehofski et al., 2015b). As these proteins have not previously been associated with BRVO, the present study was established to examine if the content of the proteins would change when subjected to an intervention with Ranibizumab. Ranibizumab is a humanized monoclonal antibody fragment that binds VEGF-A isoforms such as VEGF110, VEGF121 and VEGF165 (Gaudreault et al., 2005). The drug is widely used for the treatment of macular edema secondary to BRVO (Campochiaro et al., 2010). However, the effects of Ranibizumab on the large-scale retinal protein expression remain largely unstudied. By using two different proteomic techniques this study aimed at identifying retinal proteins and biological processes that had not been associated with sensitivity to Ranibizumab previously. 2. Materials and methods 2.1. Animal preparation The study was approved by the Danish Animal Experiments Inspectorate, permission no. 2013-15-293400775.
2
ACCEPTED MANUSCRIPT Six Danish Landrace pigs of approximately 30-40 kg were used in the present study. The animals were anesthetized by using tiletamine-zolazepam (Zoletil), a mixture of 2 dissociative anesthetics (ketamine 6.25 mg/mL, and tiletamine 6.25 mg/mL), a benzodiazepine (Zolazepam 6.25 mg/mL), a synthetic opioid (butorphanol, 1.25 mg/mL), and xylazine (6.5 mg/mL). This mixture was administered as an intramuscular injection at 1 mL/10 kg
RI PT
2.2. Experimental branch retinal vein occlusion Experimental BRVO was induced as earlier described (Fig. 1A) (Cehofski et al., 2015a; Cehofski et al., 2015b). The eyes were anesthetized with Oxybuprocaine Hydro 0.4% (Bausch & Lomb) and Tetracaine 1 % (Bausch & Lomb) followed by dilatation with Tropicamide 0.5% (Mydriacyl; Bausch & Lomb) and Phenylephrine 10% (Metaoxidrin; Bausch & Lomb). Systane Ultra eye drops (Polyethylene Glycol 400, Propylene Glycol; Alcon Copenhagen Denmark) were used to lubricate the eyes.
M AN U
SC
Experimental BRVO was induced in both eyes of the pigs (Fig. 1B). A standard argon laser (532 nm) given by indirect ophthalmoscopy was used to induce the occlusion by creating a patch of laser burns around an inferior branch vein to create a narrowing of the vein followed by laser application directly on the vein until stagnation of the venous blood flow was observed. Stagnation of the venous blood flow was generally observed after 30-40 laser applications with a power of 400 mW and a duration of 550 ms. Peripheral flameshaped hemorrhages developed within 10 minutes after application of the laser burns.
TE D
After 24 hours the animals were given an intervention with Ranibizumab (Fig. 1B). The animals were anesthetized as described above followed by topical anesthesia with Oxybuprocaine Hydro 0.4% (Bausch & Lomb) and Tetracaine 1 % (Bausch & Lomb). The eyelids including cilia were disinfected with swabs containing 82% ethanol and 0.5% chlorhexidine (Kirudan) followed by disinfection with iodine. Under sterile conditions the right eyes received an injection of 0.05 mL Ranibizumab 10 mg/mL (Novartis) whilst all left eyes were given an injection of 0.05 mL sodium chloride 9 mg/mL (B. Braun).
AC C
EP
After 3 days the eyes were enucleated and dissected (Fig. 1B). Prior to dissection the eyes were rinsed in saline water to remove blood residues. The eyes were opened by cutting 2-3 mm posteriorly to the corneal limbus followed by removal of the anterior compartments and the vitreous body. Vitreous remnants adhering to the retina were gently aspired with a needle into a 5 mL syringe to ensure that vitreous contamination was avoided. After removal of the vitreous a sector shaped piece of the inferior neurosensory retina containing the occlusion was peeled from the retinal pigment epithelium and stored in Eppendorf tubes at -80°C until proteomic analysis was conducted. All animals were euthanized immediately after enucleation. 2.3. Sample preparation for mass spectrometry Essentially, sample preparation was performed as earlier described (Bennike et al., 2015; Cehofski et al., 2015b). The samples were thawed on ice and a volume of 200 µL cold PBS was added. The samples were then transferred to impact resistant bead beading tubes containing 0.9-2.0 mm steel beads. The retinal tissue was homogenized by placing the bead beading tubes on a Vortex mixer for 30 seconds followed by bead beading in a Bullet Blender (Next Advance Inc.) for four minutes at speed level 8. The samples were cooled during bead beating. To ensure that all tissue was washed off the bead beading balls another 100 µL cold PBS were added and the samples were placed on a Vortex mixer for 30 seconds. The protein concentration was determined by measuring the absorption at 280 nm (Implen Nanophotometer). A total of 10 µg of lysate from each sample was separated by reducing SDS PAGE (12%, Biorad) and visualized using Krypton fluorescent stain to assess the quality of retinal lysates.
3
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
2.4. Label-free quantification Essentially, label-free liquid chromatograpy tandem mass spectrometry was performed as earlier described (Bennike et al., 2015; Cehofski et al., 2015b). An aliquot of 100 µg retinal homogenate and 100 µL lysis buffer of 10% sodium deoxycholate (NaDOC) in mM triethylammonium bicarbonate (TEAB); pH 8.0, was transferred to 0.5 mL low bind Eppendorf tubes and heat inactivated for 5 min at 95 ºC. Another 250 µL lysis buffer of 10% sodium deoxycholate (NaDOC) in mM triethylammonium bicarbonate (TEAB); pH 8.0, were added. The samples were then reduced with 1.2 µL TCEP followed by homogenization on a Vortex mixer and incubation for 30 min at 60 ºC. The lysates were transferred to YM-10kDa spinfilters (Millipore) and centrifuged at 14,000g for 15 min. Centrifugation was repeated until all liquid was spun out of the filter. The samples were alkylated by adding chloroacetamide (final concentration 50 mM) in 0.5% NaDOC with 50 mM TEAB, homogenized on a Vortex mixer followed by incubation in the dark for 20 min. The samples were centrifuged for 30 min at 14,000g until a residual volume of 50 µL was reached. Buffer exchange was performed by adding 400 µL digestion buffer containing 0.5% NaDOC in 50 mM TEAB; pH 8.0 followed by centrifugation. A volume of 50 µL 0.5% digestion buffer containing 1 µg trypsin was added followed by homogenization on a Vortex mixer. The samples were put into a plastic bag with wet tissue to prevent the samples from drying out followed by incubation overnight at 37 ºC.
TE D
After incubation with trypsin the spin filters containing the retinal samples were transferred to a new collection tube which was used to collect the peptides. The samples were centrifuged at 14,000g for 15 min. A volume of 200 µL of 50mM TEAB was added and centrifugation at 14,000g for 15 min was repeated. NaDOC was removed from the samples by acid extraction with ethyl acetate acidified by adding concentrated formic acid to lower pH to below 2. The samples were mixed on a Vortex mixer and centrifuged at 14,000g for 2 min for phase separation to be obtained. The upper phase was aspirated and discarded. Ethyl acetate extraction was repeated and the entire upper phase was aspirated and discarded. The samples were dried in a vacuum centrifuge and stored at -80ºC.
EP
Before mass spectrometry the samples were resuspended in 20 µL loading buffer consisting of 2% acetonitrile and 0.1% formic acid. The peptide samples were supplemented with 100 fmol MS Qual/Quant Mix (Sigma-Aldrich) as internal standard and analyzed by nanoUPLC tandem mass spectrometry. Prior to nanoUPLC tandem mass spectrometry the peptide concentration was determined by absorption at 280nm by uv-vis-spectroscopy (IMPLEN NanoPhotometer P-300). A volume of 1.5 µL was applied for measurement using lid factor x 50 and blank calibration using dissolution solvent (2% acetonitrile in 0.1% formic acid).
AC C
A sample volume of 5 µl was injected onto a Dionex RSLC nanoUPLC system that was connected to a Quadrupole Orbitrap (Q Exactive Plus) mass spectrometer equipped with a NanoSpray Flex ion source (Thermo Scientific, Bremen, Germany). The flow settings were 8 µL per min for the sample loading onto a trapping column (Acclaim PepMap100 C18 5µm, Thermo Scientific). The nanoflow was set to 300 nL per min for the peptide separation on the analytical column, which was a 50 cm Acclaim Pepmap RSLC, 75 um ID connected with nanoviper fittings. The nano-electrospray was done using a Picotip ‘Silicatip’ emitter from New Objectives. The LC buffers were buffer A (0.1% FA) and buffer B (99.9% acetonitrile, 0.1% formic acid). The applied gradient was from 10% to 30% buffer B over 230 min. The mass spectrometer was operated in data-dependent acquisition mode. A full MS scan in the mass range of 350 to 1400 m/z was acquired at a resolution of 70,000 with an AGC target of 1 x 106 and maximum fill time set to 100 ms. In each cycle, the mass spectrometer would trigger up to 12 MS/MS acquisitions on abundant peptide precursors ions. The MS/MS scans were acquired with a dynamic mass range at a resolution of 17,500 and with an AGC target of 5 x 105 and max fill time of 50 ms. The precursor ions were
4
ACCEPTED MANUSCRIPT isolated using a quadrupole isolation window of 2.0 m/z and fragmented in the HCD trap with a normalized collision energy set to 30. The under-fill ratio was set to 1.0 % with the intensity threshold at 1.0 x 105. Apex triggering was 3 to 10 s with charge and isotopes exclusion on. Dynamic exclusion was set to 30 s. All samples were analyzed in three technical replicates. MS data were submitted to the ProteomXchange Consortium via the PRIDE partner repository with the dataset identifier PXD003044 (Supplementary material – access to MS data).
M AN U
SC
RI PT
Mass spectrometry data were searched using the open-source software MaxQuant (version 1.5.0.31) against a pig protein database optimized from Uniprot containing Sus Scofa proteins and the closest related human protein gene identified based on BLAST (Hesselager et al., 2016). The label-free quantification (LFQ) algorithm was activated in MaxQuant. Standard settings included a peptide and protein false discovery rate of 1%. P-values were calculated by a two-tailed, heteroscedastic t-test. Raw data from the MaxQuant search (Supplementary material – database search output) were submitted to Perseus (version 1.5.2.8) for filtration. Proteins were filtered by unique peptides + razor peptides > 2 followed by removal of proteins identified from a peptide that was found to be part of a protein derived from the reversed part of the decoy database. LFQ values were log2 transformed and triplicates were averaged by mean. Statistical analysis by paired t-test that was conducted for all proteins that were identified in at least four out of five pigs.
TE D
2.5. Isobaric label quantification using mass tag For analysis using isobaric labeling a TMT10plex Mass Tag Labeling kit from Thermo Scientific was used. One hundred µg from each of the 10 samples were transferred into a new tube and 100 mM TEAB was added until a final volume of 100 µL was reached. A volume of 5 µL 200 mM TCEP was added and the samples incubated at 55°C for 1 hour. A volume of 5 µL 375mM iodoacetamide was added to the samples and incubated for 30 minutes at room temperature protected from light and 600 µL of pre-chilled 20% acetone were added and precipitation proceeded overnight. The samples were centrifuged at 8000g for 10 minutes at 4°C. The acetone was decanted and the pellet was allowed to dry for 2-3 minutes. The protein pellets were resuspended with 100 µL of 100 mM TEAB and 2.5 µg trypsin were added to the samples for digestion overnight at 37°C.
AC C
EP
Prior to peptide labeling a volume of 41 µL of anhydrous acetonitrile was added to the 0.8 mg vials containing the TMT label reagents followed by occasional vortexing for 5 minutes and brief centrifugation. Forty-one µL of the TMT label reagent was added to each 100 µL sample followed by incubation for 1 hour at room temperature. A volume of 8 µL of 5% hydroxylamine was added to each sample followed by incubation for 15 minutes. The 10 samples were combined at equal amounts in a new microcentrifuge tube and stored at -80°C. Peptide samples were thawed and purified with a Pierce C18 Spin Column (Thermo Scientific). The spin column was placed in a receiver tube and 200 µL activation solution (50% methanol) were added to wet resin followed by centrifugation at 1500g for 1 minute, the flow-through was discarded and the step was repeated. Subsequently, 200 µL of equilibration solution (0,5% TFA in 5% acetonitrile) were added and the sample was centrifuged at 1500g, the flow-through was discarded and the step was repeated. The sample was loaded on top of the resin bed and the column was placed in a received tube and centrifuged at 1500g for 1 minute. The flow-through was recovered and centrifugation was repeated. The column was placed in a new receiver tube and 200 µL wash solution (0.5%TFA and 5% acetonitrile) were added and the column was centrifuged for 1 minute at 1500g. Wash and centrifugation were repeated. The column was placed in a new
5
ACCEPTED MANUSCRIPT receiver tube and 20 µL elution buffer (70% acetonitrile) were added followed by centrifugation at 1500g for 1 minute, the step was repeated and the sample was dried in a vacuum evaporator.
M AN U
SC
RI PT
Peptides were fractionated using a Pierce High pH Reversed-Phase Peptide Fractionation Kit (Themo Scientific). The spin column was loaded with 300 µL of acetonitrile and centrifuged for 2 minutes at 5000g. This step was repeated. The spin column was washed twice with 0.1% TFA solution. Elution solutions consisting of increasing concentrations of acetonitrile in 0.1% triethylamine were prepared according to instructions provided by the manufacturer. The digested samples were dissolved in 300 µL 0.1% TFA. The spin column was placed into a 2.0 mL sample tube and 300 µL of the sample solution were loaded onto the column that was centrifuged for 2 minutes at 3000g and the eluate was retained as “flow-through” fraction. The spin column was placed in a new 2.0 mL sample tube and 300 µL of water were loaded onto the column that was centrifuged at 3000g for 2 minutes followed by an additional column wash with 300 µL 5 % acetonitrile and 0.1% TEA. The spin column was placed in a new 2.0 mL sample tube and loaded with elution solution and centrifuged at 3000g for 2 minutes and the fraction was collected. This step was repeated with all 8 elution solutions and the collected fractions were dried using vacuum centrifugation. Samples were resuspended in 0.1% formic acid before LC-MS.
AC C
EP
TE D
The labeled samples were analyzed by LC-MS by injecting 5 µL sample volumes into a Dionex RSCL nanoUPLC system connected to an Orbitrap Fusion mass spectrometer equipped with an Easy Spray ion source (Thermo Scientific). The flow settings were 30 µL per min for the sample loading onto a trapping column (5 mm x 300 µm, C18 PepMap100, 5 µm, 100Å, Thermo Scientific). The nanoflow was set to 300 µL per min for separation of peptides on an analystical column (500 mm x 75 µm PepMap RSCL, C18, 2 µm, 100 Å, Thermo Scientific). The buffers were buffer A (0.1% FA) and buffer B (80% acetonitrile, 20% water, 0.1% FA). The applied gradient was performed over either 125 min or 225 min using a gradient of buffer B from 6% to 90%. The Fusion was operated in the TMT SPS MS3 mode using full Orbitrap scans in the mass range 380-1500 m/z acquired at a resolution of 120,000 with an AGC target of 2 x 105 and maximum injection time of 50 ms. In each cycle the mass spectrometer would trigger MS2 acquisitions using the linear ion trap in the mass range 400-1200 m/z with a CID collision energy at 35%, an AGC target of 1 x 104 and a maximum injection time of 50 ms. The precursor ions were isolated using the quadrupole set with an isolation window of 0.7 m/z. Up to 10 reporter ions isolated with a window of 2 m/z were detected in MS3 using synchronous precursor selection in the Orbitrap in the mass range 120-500 m/z with a HCD collision energy at 65% acquired at a resolution of 60,000 with an AGC target of 1 x 105 and a maximum injection time of 120 ms. Dynamic exclusion was set to 30 s or to 70 s. In total 4 technical replicates, each fragmented in 8 samples were performed generating in total 32 data files. The raw data files were further processed in Proteome Discoverer 2.1. The Sequest HT search engine was used against the SwissProt Homo sapiens database (TaxID=9606_and_subtaxonomies; version 2016-04-13) and the SwissProt Sus scrofa database (TaxID=9823_and_subtaxonomies; version 2016-04-13) with a precursor mass tolerance of 10 ppm and a fragment mass tolerance of 0.6 Da. Dynamic modifications were allowed for the TMT labels as well as oxidation of Met, deamidation of Asn and Gln and acetylation of the N-terminal. Static modification was carbamidomethyl on Cys. A false discovery rate of 1% was used and at least 2 unique peptides were required for identification. Quantification was required in each of the 10 samples and each channel was normalized based on total peptide amount.
6
ACCEPTED MANUSCRIPT 2.6. Statistics LFQ values and TMT abundances were log2 transformed prior to statistical analysis. For both datasets a paired t-test was conducted between Ranibizumab injected eyes and NaCl injected eyes in STATA 13.1. The paired t-test was performed for all proteins that were identified in at least four out of five pigs in the nonlabeled analysis and in all pigs in the labeled analysis. Proteins that were significantly changed in both datasets were considered significantly regulated.
SC
RI PT
3. Results 3.1. Experimental branch retinal vein occlusion Based on the development of stagnation of venous blood and peripheral haemorrhages upstream of the site of occlusion, BRVO was successfully induced in six right eyes and six left eyes. Fundoscopic images to document BRVO were taken with a KOWA GENESIS-D handheld retinal camera within the first 30 minutes after BRVO had been induced (Fig. 2). Upon enucleation one animal had developed a retinal detachment in one eye and was excluded from the study. Thus, proteomic analysis was conducted on five retinas from right eyes treated with Ranibizumab and five retinas from left eyes treated with saline water.
M AN U
3.2. Label-free quantification and TMT quantification The present study analyzed large-scale protein changes following BRVO and intervention with Ranibizumab. BRVO was induced in both eyes of the animals and after 24 hours an intervention with Ranibizumab was given in right eyes and saline water was injected into all left eyes. Proteins in the retinal samples were quantified with label-free quantification as well as TMT quantification. Mass spectrometry data were searched against the databases. A paired t-test was used as statistical analysis of both datasets. Proteins that were statistically significantly changed in content following label-free quantification as well as TMT quantification were considered significantly regulated.
EP
TE D
With label-free quantification 2837 proteins were identified in at least 4 animals. In eyes treated with Ranibizumab a total of 237 proteins exhibited a statistically significant change in content based on label-free quantification (Fig. 3) (Supplementary table 1). In the quantification based on labeling 3209 proteins were identified. With TMT quantification we identified 83 proteins that exhibited a statistically significant change in content (Fig. 3) (Supplementary table 2). Five proteins had a statistically significant change in content in both datasets (Fig. 3). These proteins included integrin β-1, peroxisomal 3-ketoacyl-CoA thiolase, OCIA domain-containing protein 1, calnexin and 40S ribosomal protein S5 (Table 1). These five proteins all exhibited a significant decrease in content.
AC C
Platelet derived growth factor receptor-β (PDGFRB) was found be significantly downregulated in the dataset that was generated through label-free quantification (fold change = 0.83; p = 0.041). PDGFRB was not identified with TMT. Findings that were exclusively identified with TMT included high increases in immunoglobulin gamma-1 chain C region (fold change = 7.37; p = 0.0000613) and immunoglobulin kappa chain C region (fold change = 6.64; p = 0.0000696) in the right eyes that received anti-VEGF intervention. With TMT a decline in the content of retinal dehydrogenase 1 was found in retinas treated with Ranibizumab (fold change = 0.34; p = 0.00761). With label-free quantification retinal dehydrogenase 1 was only identified in two animals (fold changes 0.19 and 0.34).
7
ACCEPTED MANUSCRIPT 4. Discussion 4.1. Experimental BRVO In our previous study we analyzed retinal porcine protein changes 15 days after experimental BRVO (Fig. 1A) (Cehofski et al., 2015b). In this previous study validating fluorescein angiography showed successful occlusion after 9 days. Therefore, no re-canalization was to be expected in the present study in which the retinas were collected after 3 days.
SC
RI PT
4.2. Integrin β-1 and OCIA-domain-containing protein 1 Using two fundamentally different proteomic techniques we found that inhibition of VEGF-A with Ranibizumab resulted in a small significant downregulation of integrin β-1. This finding is of interest as our previous experiment (Fig. 1A) (Cehofski et al., 2015b) identified increased levels of integrin β-1 in retinas exposed to experimental BRVO. As integrin β-1 is a key protein in ECM-signaling processes it is likely that BRVO is associated with ECM remodeling processes that are inhibited when VEGF-A signaling is blocked by Ranibizumab (Fig. 4A and Fig. 4B).
M AN U
It is important to note that the retinal samples in our previous study (Fig. 1A) (Cehofski et al., 2015b) and the present study (Fig. 1B) were collected after 15 days and 3 days respectively. In the present study (Fig. 1B) the retinas were collected 3 days after BRVO as there is growing evidence in favor of early inhibition of VEGF and dissection within the acute stage of BRVO. Firstly, our previous study (Fig. 1A) (Cehofski et al., 2015b) found that angiogenesis was not among the biological processes that were altered 15 days after BRVO. Secondly, a study of gene expression following experimental retinal vein occlusion in rat retinas has demonstrated that VEGF expression was upregulated within day 1 of retinal vein occlusion and had normalized 3 days after BRVO (Rehak et al., 2009).
TE D
The anti-integrin Luminate (Allegro Ophthalmics) binds to several integrins including αvβ3, αvβ5, α3β1, α5β1 (Karageozian, 2015). Considering the fact that Luminate is now in phase 2 clinical trials for the treatment of wet age-related macular degeneration and diabetic macular edema (Karageozian, 2015) it is interesting that inhibition of VEGF-A in itself resulted in a small decrease in the content of integrin β-1.
AC C
EP
Interactions between VEGF-A and integrin β-1 have been described by Yamamoto et al. (Yamamoto et al., 2015) in a study on murine retinas which established integrin β-1 as a promoter of endothelial sprouting, vascular ECM deposition and organization of the vessel wall. Oommen et al. (Oommen et al., 2011) demonstrated that VEGF-A can bind to a specific binding site on integrin α9β1 to promote endothelial cell adhesion and migration. Studying human dermal microvascular endothelial cells (HMVEC) Vlahakis et al. (Vlahakis et al., 2007) found that endothelial cell migration induced by either VEGF-A165 or VEGF-A121 was inhibited by α9β1 blocking antibodies. Furthermore, VEGF-A-induced angiogenesis in chicken embryo chorioallantoic assay was inhibited when an integrin α9β1 antibody was added. Ovarian cancer immunoreactive antigen domain containing 1 (OCIA-domain-containing protein 1) has been shown to downregulate cell attachment to integrin β-1 in ovarian cancer cell lines. Thus, the decreased levels of integrin β-1 and OCIA-domain-containing protein 1 following anti-VEGF treatment may potentially affect recruitment of cells to retinal areas affected by the occlusion (Wang et al., 2010). 4.3. Peroxisomal 3-ketoacyl-CoA thiolase, calnexin and 40S ribosomal protein S5 Peroxisomal 3-ketoacyl-CoA thiolase is involved in peroxisomal β-oxidation of fatty acids (Hu et al., 2005), but its role in retinal diseases and interactions with VEGF-A remain largely unstudied. Calnexin is a lectin
8
ACCEPTED MANUSCRIPT
SC
RI PT
chaperone protein that has a key function in the protein quality control of the endoplasmic reticulum (Wang et al., 2015) where it controls that newly synthesized glycoproteins are properly assembled and folded (Rosenbaum et al., 2006; Wang et al., 2015). Rosenbaum and co-workers (Rosenbaum et al., 2006) demonstrated that calnexin has multiple functions in the retina in Drosophila where it regulates cytosolic Ca2+ and promotes rhodopsin (Rh1) maturation. Compromised regulation of cytosolic Ca2+ and rhodopsin maturation resulted in retinal degeneration in flies with mutations in the calnexin gene. Interactions between calnexin and VEGF have been elucidated by Demeure et al. (Demeure et al., 2016) who found that antiVEGF treatment with intraperitoneal Bevacizumab injections increased calnexin levels in glioblastoma xenografts in rats that received weekly injections for three weeks. Thus, it is possible that the endoplasmic reticulum reacts differently on prolonged anti-VEGF treatment. The role of calnexin in retinal diseases and potential interactions with VEGF remain to be established, but given the decrease in 40S ribosomal protein S5, the lower content of calnexin in retinas treated with Ranibizumab may represent changes in glycoprotein trafficking that occur following VEGF-A inhibition.
M AN U
4.4. Platelet derived growth factor receptor-β In retinas treated with Ranibizumab a small decrease in the content of PDGFRB was identified with labelfree quantification. The finding is of interest as drugs targeted at platelet-derived growth factor homodimer BB (PDGF-BB) are under development for the treatment of wet age-related macular degeneration (Tolentino et al., 2015). For example, Fovista (Ophthotech) binds to platelet-derived growth factor homodimer BB to prevent it from interacting with its receptor PDGFRB (Tolentino et al., 2015). PDGFRB was not identified with TMT and additional studies are needed to examine if intravitreal anti-VEGF injections decrease the level of the retinal content of PDGFRB.
TE D
4.5. Immunoglobulin gamma-1 chain C region and immunoglobulin kappa chain C region In the TMT experiment substantial increases in the content of immunoglobulin gamma-1 chain C region and immunoglobulin kappa chain C region were identified. As Ranibizumab is an IgG1 kappa isotype monoclonal antibody Fab fragment (Meyer and Holz, 2011) it is likely that immunoglobulin gamma-1 chain C region and immunoglobulin kappa chain C region are components of Ranibizumab. The finding indicates that Ranibizumab successfully reached the retina after intravitreal injection.
AC C
EP
4.6. Retinal dehydrogenase 1 The TMT experiment showed that the content of retinal dehydrogenase 1 was significantly decreased in retinas treated with Ranibizumab. In the experiment based on label-free proteomics retinal dehydrogenase 1 was detected in two animals (fold changes 0.19 and 0.34). Retinal dehydrogenase 1 irreversibly oxidizes retinaldehyde to retinoic acid, the most potent metabolite of vitamin A (Bchini et al., 2013). Thus, our study raises the question if anti-VEGF treatment may affect the production of retinoic acid in the retina. 5. Conclusion In retinas that received Ranibizumab intervention after BRVO we identified decreased contents of integrin β1, peroxisomal 3-ketoacyl-CoA thiolase, OCIA domain-containing protein 1, calnexin and 40S ribosomal protein S5. The downregulation of integrin β-1 suggests that ECM remodeling and focal adhesion processes associated with BRVO are at least partly reversed when an intervention with Ranibizumab is given. With anti-integrin therapies under development for the treatment of diabetic macular edema and wet age-related macular degeration it is interesting that inhibition of VEGF-A in itself resulted in a small decrease in the content of integrin β-1. With label-free quantification a small downregulation of PDGFRB was identified in
9
ACCEPTED MANUSCRIPT eyes that received Ranibizumab treatment. Further studies are required validate if intravitreal anti-VEGF treatment decreases retinal PDGFRB levels.
RI PT
Fig. 1. (A) Experimental set-up from previous study (Cehofski et al., 2015b). In this previous work BRVO was induced in the right eyes of the animals by applying argon green laser directly onto a vein in the inferior retina. In the left eyes that served as controls a similar area of laser burns was created in an area devoid of vessels. The retinas were dissected after 15 days followed by proteomic analysis. (B) In the present intervention study BRVO was induced in both eyes of the animals. After 24 hours VEGF-A was inhibited in all right eyes with an intravitreal injection of Ranibizumab while all left eyes received an intravitreal injection of 0.9% sodium chloride. The retinas were dissected after 3 days.
SC
Fig. 2. Image of laser-induced experimental BRVO taken with a KOWA GENESIS handheld retinal camera approximately 20 minutes after BRVO was induced. (A) Dilation of branch vein and stagnation of venous blood following laser induced occlusion ( : Area of laser treatment, : Peripheral haemorrhages, : Dilation of vein). (B) Peripheral haemorrhages upstream to the occlusion ( : Peripheral haemorrhages).
M AN U
Fig. 3. Summary of results. The retinal proteins were quantified with label-free quantification as well as tandem mass tag quantification. With label-free quantification we identified 237 significantly changed proteins while 77 significantly changed proteins were identified with TMT. There was an overlap of five significantly changed proteins between the two datasets. These proteins included integrin β-1, peroxisomal 3ketoacyl-CoA thiolase, OCIA domain-containing protein 1, calnexin and 40S ribosomal protein S5.
TE D
Fig. 4. Hypothesis about protein changes following experimental BRVO. (A) We earlier identified an upregulation of integrin β-1 following experimental BRVO. (B) Inhibition of VEGF-A resulted in a downregulation of integrin β-1 suggesting that ECM remodeling, adhesion, migration and proliferation processes associated with experimental BRVO are downregulated due to intervention with Ranibizumab.
EP
Financial support This work was supported by the Svend Andersen Foundation, the Bagger-Sørensen Foundation, the Obel Family Foundation, the Herta Christensen Foundation, the North Denmark Region (2013-0076797) and Heinrich Kopps Foundation.
AC C
Disclosures Ranibizumab was provided complimentarily by Novartis Ophthalmics, Denmark. The authors have no financial interests to declare. Contributions Lasse Cehofski, Anders Kruse, Bent Honoré and Henrik Vorum conceived the idea of studying the retinal proteome following experimental BRVO and intervention with Ranibizumab. All authors participated in the writing of the manuscript. Lasse Cehofski performed experimental, laboratory work and data analysis. Anders Kruse performed laser treatments. Sigriður Olga Magnusdottir anesthetized the animals. Martin Bøgsted contributed to statistical analyses. Allan Stensballe conducted sample preparation for mass spectrometry, mass spectrometry analysis and data analysis. Bent Honoré participated in study design, data analysis, the writing of the manuscript and conducted the TMT experiments. The study was supervised by Henrik Vorum who participated in study design and data analysis.
10
ACCEPTED MANUSCRIPT References
AC C
EP
TE D
M AN U
SC
RI PT
Bchini, R., Vasiliou, V., Branlant, G., Talfournier, F., Rahuel-Clermont, S., 2013. Retinoic acid biosynthesis catalyzed by retinal dehydrogenases relies on a rate-limiting conformational transition associated with substrate recognition. Chem. Biol. Interact. 202, 78-84. Bennike, T.B., Carlsen, T.G., Ellingsen, T., Bonderup, O.K., Glerup, H., Bogsted, M., Christiansen, G., Birkelund, S., Stensballe, A., Andersen, V., 2015. Neutrophil Extracellular Traps in Ulcerative Colitis: A Proteome Analysis of Intestinal Biopsies. Inflamm. Bowel. Dis. 21, 2052-2067. Campochiaro, P.A., Choy, D.F., Do, D.V., Hafiz, G., Shah, S.M., Nguyen, Q.D., Rubio, R., Arron, J.R., 2009. Monitoring ocular drug therapy by analysis of aqueous samples. Ophthalmology 116, 2158-2164. Campochiaro, P.A., Heier, J.S., Feiner, L., Gray, S., Saroj, N., Rundle, A.C., Murahashi, W.Y., Rubio, R.G., 2010. Ranibizumab for macular edema following branch retinal vein occlusion: six-month primary end point results of a phase III study. Ophthalmology 117, 1102-1112.e1101. Cehofski, L.J., Kruse, A., Kjaergaard B, Stensballe A, Honore, B., Vorum, H., 2015a. Dye Free Porcine Model of Experimental Branch Retinal Vein Occlusion – a Suitable Approach for Retinal Proteomics. J. Ophthalmol., article ID 839137. Cehofski, L.J., Kruse, A., Kjaergaard, B., Stensballe, A., Honore, B., Vorum, H., 2015b. Proteins involved in focal adhesion signaling pathways are differentially regulated in experimental branch retinal vein occlusion. Exp. Eye Res. 138, 87-95. Cehofski, L.J., Mandal, N., Honore, B., Vorum, H., 2014. Analytical platforms in vitreoretinal proteomics. Bioanalysis 6, 3051-3066. Demeure, K., Fack, F., Duriez, E., Tiemann, K., Bernard, A., Golebiewska, A., Bougnaud, S., Bjerkvig, R., Domon, B., Niclou, S.P., 2016. Targeted Proteomics to Assess the Response to Anti-Angiogenic Treatment in Human Glioblastoma (GBM). Mol. Cell. Proteomics 15, 481-492. Gaudreault, J., Fei, D., Rusit, J., Suboc, P., Shiu, V., 2005. Preclinical pharmacokinetics of Ranibizumab (rhuFabV2) after a single intravitreal administration. Invest. Ophthalmol. Vis. Sci. 46, 726-733. Hesselager, M.O., Codrea, M.C., Sun, Z., Deutsch, E.W., Bennike, T.B., Stensballe, A., Bundgaard, L., Moritz, R.L., Bendixen, E., 2016. The Pig PeptideAtlas: A resource for systems biology in animal production and biomedicine. Proteomics 16, 634-644. Hu, T., Foxworthy, P., Siesky, A., Ficorilli, J.V., Gao, H., Li, S., Christe, M., Ryan, T., Cao, G., Eacho, P., Michael, M.D., Michael, L.F., 2005. Hepatic peroxisomal fatty acid beta-oxidation is regulated by liver X receptor alpha. Endocrinology 146, 53805387. Karageozian, V., 2015. Investigational drug Luminate targets integrin receptors. Ophthalmol. Times, 949/940-8130. Meyer, C.H., Holz, F.G., 2011. Preclinical aspects of anti-VEGF agents for the treatment of wet AMD: ranibizumab and bevacizumab. Eye (Lond) 25, 661-672. Noma, H., Funatsu, H., Mimura, T., Eguchi, S., Hori, S., 2011. Soluble vascular endothelial growth factor receptor-2 and inflammatory factors in macular edema with branch retinal vein occlusion. Am. J. Ophthalmol. 152, 669-677.e661. Noma, H., Funatsu, H., Mimura, T., Tatsugawa, M., Shimada, K., Eguchi, S., 2012. Vitreous inflammatory factors and serous macular detachment in branch retinal vein occlusion. Retina 32, 86-91. Noma, H., Mimura, T., Yasuda, K., Shimura, M., 2014. Role of soluble vascular endothelial growth factor receptors-1 and -2, their ligands, and other factors in branch retinal vein occlusion with macular edema. Invest. Ophthalmol. Vis. Sci. 55, 3878-3885. Oommen, S., Gupta, S.K., Vlahakis, N.E., 2011. Vascular endothelial growth factor A (VEGF-A) induces endothelial and cancer cell migration through direct binding to integrin {alpha}9{beta}1: identification of a specific {alpha}9{beta}1 binding site. J. Biol. Chem. 286, 1083-1092. Rehak, M., Hollborn, M., Iandiev, I., Pannicke, T., Karl, A., Wurm, A., Kohen, L., Reichenbach, A., Wiedemann, P., Bringmann, A., 2009. Retinal gene expression and Muller cell responses after branch retinal vein occlusion in the rat. Invest. Ophthalmol. Vis. Sci. 50, 2359-2367. Rosenbaum, E.E., Hardie, R.C., Colley, N.J., 2006. Calnexin is essential for rhodopsin maturation, Ca2+ regulation, and photoreceptor cell survival. Neuron 49, 229-241. Tolentino, M.J., Dennrick, A., John, E., Tolentino, M.S., 2015. Drugs in Phase II clinical trials for the treatment of age-related macular degeneration. Expert Opin. Investig. Drugs 24, 183-199. Vlahakis, N.E., Young, B.A., Atakilit, A., Hawkridge, A.E., Issaka, R.B., Boudreau, N., Sheppard, D., 2007. Integrin alpha9beta1 directly binds to vascular endothelial growth factor (VEGF)-A and contributes to VEGF-A-induced angiogenesis. J. Biol. Chem. 282, 15187-15196. Wang, C., Michener, C.M., Belinson, J.L., Vaziri, S., Ganapathi, R., Sengupta, S., 2010. Role of the 18:1 lysophosphatidic acidovarian cancer immunoreactive antigen domain containing 1 (OCIAD1)-integrin axis in generating late-stage ovarian cancer. Mol. Cancer Ther. 9, 1709-1718. Wang, Q., Groenendyk, J., Michalak, M., 2015. Glycoprotein Quality Control and Endoplasmic Reticulum Stress. Molecules 20, 13689-13704. Yamamoto, H., Ehling, M., Kato, K., Kanai, K., van Lessen, M., Frye, M., Zeuschner, D., Nakayama, M., Vestweber, D., Adams, R.H., 2015. Integrin beta1 controls VE-cadherin localization and blood vessel stability. Nat. Commun. 6, 6429.
11
ACCEPTED MANUSCRIPT Table 1 Protein exibiting a significant change in content in both proteomic methods Protein
Fold change LFQ p-value LFQ
Fold change TMT p-value TMT
Gene name
0.78
0.00798
0.92
0.00728
ITGB1
0.81
0.00744
0.91
0.0277
ACAA1
OCIA domain-containing protein 1
0.81
0.0460
0.90
0.0357
OCIAD1
40S ribosomal protein S5
0.77
0.0403
0.89
0.00299
RPS5
Calnexin
0.80
0.00468
0.93
0.0495
CANX
AC C
EP
TE D
M AN U
SC
RI PT
Integrin β-1 Peroxisomal 3-ketoacyl-CoA thiolase
ACCEPTED MANUSCRIPT
BRVO vs. control - previous study Right retina - laser induced BRVO
Dissection
Left retina - lasered, no BRVO Argon laser
A
0
1
2
3
4
5
Right retina - laser induced BRVO
TE D
Intervention study - present study
Left retina - laser induced BRVO
Figure 1
15
Left eye 0.9% NaCl
Days
Dissection
2
3
4
0.200
0.200
0.150
0.150
0.10
0.10
AC C B
1
Right eye anti-VEGF
10
Argon laser
EP
Argon laser
0
M AN U
SC
RI PT
Argon laser
0.050
0.050
5
10
15
Days
A
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
B
Figure 2
ACCEPTED MANUSCRIPT
RI PT
Summary of results Tandem mass tag quantification 83 significantly changed proteins
M AN U
SC
Label-free quantification 237 significantly changed proteins
5
83
EP
TE D
237
AC C
Significantly changed proteins identified in both datasets - Integrin β-1 - Peroxisomal 3-ketoacyl-CoA thiolase - OCIA domain-containing protein 1 - Calnexin - 40S ribosomal protein S5 Figure 3
ACCEPTED MANUSCRIPT
Experimental BRVO and inhibition of VEGF-A
RI PT
Experimental BRVO
Integrin-β1
M AN U
SC
Integrin-β1
Extracellular matrix
TE D
Extracellular matrix
Cytosol
EP
Cytosol
Adhesion Migration Proliferation
AC C
Adhesion Migration Proliferation
A Figure 4
B
ACCEPTED MANUSCRIPT The content of integrin β-1 was decreased following treatment with Ranibizumab. Calnexin and 40S ribosomal protein S5 decreased in content when VEGF-A was blocked. OCIA domain-containing protein 1 was downregulated following inhibition of VEGF-A.
AC C
EP
TE D
M AN U
SC
RI PT
A role of integrin β-1 in branch retinal vein occlusion is proposed.
S0014-4835(16)30244-5
DOI:
10.1016/j.exer.2016.09.002
Reference:
YEXER 7018
To appear in:
Experimental Eye Research
Received Date: 8 January 2016 Revised Date:
7 August 2016
Accepted Date: 8 September 2016
Please cite this article as: Cehofski, L.J., Kruse, A., Bøgsted, M., Magnusdottir, S.O., Stensballe, A., Honoré, B., Vorum, H., Retinal proteome changes following experimental branch retinal vein occlusion and intervention with Ranibizumab, Experimental Eye Research (2016), doi: 10.1016/ j.exer.2016.09.002. 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 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.
ACCEPTED MANUSCRIPT
Branch retinal vein occlusion and anti-VEGF treatment Laser induced BRVO in both eyes
Right eye anti-VEGF
Dissection
Left eye 0.9% NaCl
SC
RI PT
Argon laser
0.200
0.200
0.150
0.150
0.050
0.050
Cytosol
AC C
1
EP
TE D
Retinal branch vein
2
Integrin-β1
Extracellular matrix
M AN U
0.10
0.10
0
Experimental BRVO and inhibition of VEGF-A
3
Adhesion Migration Proliferation
In the present study branch retinal vein occlusion (BRVO) was induced in the inferior retina in both eyes of the pigs. After 24 hours vascular endothelial growth factor A (VEGF-A) was inhibited in the right eyes through an injection with Ranibizumab whilst the left eyes were treated with 0.9% sodium chloride. Retinas were excised after 3 days. In retinas treated with Ranibizumab protreomic analyses revealed a downregulation of integrin β-1 suggesting that inhibition of VEGF-A downregulated adhesion, migration and proliferation processes associated with BRVO.
ACCEPTED MANUSCRIPT
RI PT
Retinal Proteome Changes Following Experimental Branch Retinal Vein Occlusion and Intervention with Ranibizumab
SC
Lasse Jørgensen Cehofski1,2,3, Anders Kruse1, Martin Bøgsted3,4, Sigriður Olga Magnusdottir2, Allan Stensballe5, Bent Honoré3,6 and Henrik Vorum1,3
Department of Ophthalmology, Aalborg University Hospital, Aalborg, Denmark
2
Biomedical Research Laboratory, Aalborg University Hospital, Aalborg, Denmark
3
Department of Clinical Medicine, Aalborg University, Aalborg, Denmark
4
Department of Haematology, Aalborg University Hospital, Aalborg, Denmark
5
Department of Health Science and Technology, Aalborg University, Denmark
6
Department of Biomedicine, Aarhus University, Aarhus, Denmark
AC C
EP
Correspondance to: Lasse Jørgensen Cehofski, Department of Ophthalmology, Aalborg University Hospital, Hobrovej 18-22, 9100 Aalborg, Denmark. Phone +45 53558878 Fax +45 97662581 E-mail [email protected]
TE D
M AN U
1
1
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
Abstract Animal models of experimental branch retinal vein occlusion (BRVO) provide a unique opportunity to study protein changes directly in retinal tissue. Results from these experimental models suggest that experimental BRVO is associated with an upregulation of extracellular matrix remodeling and adhesion signaling processes. To study whether these processes could be blocked by inhibition of VEGF-A, a porcine model of experimental BRVO was combined with proteomic analyses. In six Danish Landrace pigs experimental BRVO was induced with argon laser in both eyes. After 24 hours an injection of 0.05 mL Ranibizumab was given in the right eyes of the animals while left eyes received an injection of 0.05 mL 9 mg/mL sodium chloride water. Retinas were dissected three days after BRVO and the retinal samples were analyzed with label-free quantification as well as tandem mass tag based proteomics. In retinas treated with Ranibizumab five proteins exhibited statistically significant changes in content with both proteomic techniques. These five proteins, which were all decreased in content, included integrin β-1, peroxisomal 3-ketoacyl-CoA thiolase, OCIA domain-containing protein 1, calnexin and 40S ribosomal protein S5. As anti-integrin therapies are under development for inhibition of angiogenesis in retinal diseases it is interesting that inhibition of VEGFA in itself resulted in a small decrease in the content of integrin β-1. The decreased content of integrin β-1 indicates that extracellular matrix remodeling and adhesion processes associated with BRVO are at least partly reversed through inhibition of VEGF-A.
AC C
EP
TE D
1. Introduction Studies on protein changes following branch retinal vein occlusion (BRVO) have traditionally been conducted on aqueous humor and the vitreous body samples from patients with macular edema secondary to BRVO. These analyses have led to the identification of important proteins that are involved in the formation of macular edema (Campochiaro et al., 2009; Noma et al., 2011; Noma et al., 2012; Noma et al., 2014). Animal models of experimental BRVO allow for protein studies to be conducted directly on retinal tissue where major pathological changes are thought to take place (Cehofski et al., 2015a; Cehofski et al., 2014). We recently published a proteomic study on large-scale protein changes in porcine retinas 15 days after exposure to experimental BRVO (Fig 1A) and proposed the hypothesis that experimental BRVO was associated with an upregluation of proteins involved in extra cellular matrix (ECM) remodeling and focal adhesion processes (Cehofski et al., 2015b). Proteins involved in ECM remodeling and focal adhesion signaling included laminin subunit β-2, laminin subunit γ-1, lipocalin-7, nidogen-2, osteopontin, integrin β-1, isoform 2 of α-actinin-1, talin-2 and filamin C (Cehofski et al., 2015b). As these proteins have not previously been associated with BRVO, the present study was established to examine if the content of the proteins would change when subjected to an intervention with Ranibizumab. Ranibizumab is a humanized monoclonal antibody fragment that binds VEGF-A isoforms such as VEGF110, VEGF121 and VEGF165 (Gaudreault et al., 2005). The drug is widely used for the treatment of macular edema secondary to BRVO (Campochiaro et al., 2010). However, the effects of Ranibizumab on the large-scale retinal protein expression remain largely unstudied. By using two different proteomic techniques this study aimed at identifying retinal proteins and biological processes that had not been associated with sensitivity to Ranibizumab previously. 2. Materials and methods 2.1. Animal preparation The study was approved by the Danish Animal Experiments Inspectorate, permission no. 2013-15-293400775.
2
ACCEPTED MANUSCRIPT Six Danish Landrace pigs of approximately 30-40 kg were used in the present study. The animals were anesthetized by using tiletamine-zolazepam (Zoletil), a mixture of 2 dissociative anesthetics (ketamine 6.25 mg/mL, and tiletamine 6.25 mg/mL), a benzodiazepine (Zolazepam 6.25 mg/mL), a synthetic opioid (butorphanol, 1.25 mg/mL), and xylazine (6.5 mg/mL). This mixture was administered as an intramuscular injection at 1 mL/10 kg
RI PT
2.2. Experimental branch retinal vein occlusion Experimental BRVO was induced as earlier described (Fig. 1A) (Cehofski et al., 2015a; Cehofski et al., 2015b). The eyes were anesthetized with Oxybuprocaine Hydro 0.4% (Bausch & Lomb) and Tetracaine 1 % (Bausch & Lomb) followed by dilatation with Tropicamide 0.5% (Mydriacyl; Bausch & Lomb) and Phenylephrine 10% (Metaoxidrin; Bausch & Lomb). Systane Ultra eye drops (Polyethylene Glycol 400, Propylene Glycol; Alcon Copenhagen Denmark) were used to lubricate the eyes.
M AN U
SC
Experimental BRVO was induced in both eyes of the pigs (Fig. 1B). A standard argon laser (532 nm) given by indirect ophthalmoscopy was used to induce the occlusion by creating a patch of laser burns around an inferior branch vein to create a narrowing of the vein followed by laser application directly on the vein until stagnation of the venous blood flow was observed. Stagnation of the venous blood flow was generally observed after 30-40 laser applications with a power of 400 mW and a duration of 550 ms. Peripheral flameshaped hemorrhages developed within 10 minutes after application of the laser burns.
TE D
After 24 hours the animals were given an intervention with Ranibizumab (Fig. 1B). The animals were anesthetized as described above followed by topical anesthesia with Oxybuprocaine Hydro 0.4% (Bausch & Lomb) and Tetracaine 1 % (Bausch & Lomb). The eyelids including cilia were disinfected with swabs containing 82% ethanol and 0.5% chlorhexidine (Kirudan) followed by disinfection with iodine. Under sterile conditions the right eyes received an injection of 0.05 mL Ranibizumab 10 mg/mL (Novartis) whilst all left eyes were given an injection of 0.05 mL sodium chloride 9 mg/mL (B. Braun).
AC C
EP
After 3 days the eyes were enucleated and dissected (Fig. 1B). Prior to dissection the eyes were rinsed in saline water to remove blood residues. The eyes were opened by cutting 2-3 mm posteriorly to the corneal limbus followed by removal of the anterior compartments and the vitreous body. Vitreous remnants adhering to the retina were gently aspired with a needle into a 5 mL syringe to ensure that vitreous contamination was avoided. After removal of the vitreous a sector shaped piece of the inferior neurosensory retina containing the occlusion was peeled from the retinal pigment epithelium and stored in Eppendorf tubes at -80°C until proteomic analysis was conducted. All animals were euthanized immediately after enucleation. 2.3. Sample preparation for mass spectrometry Essentially, sample preparation was performed as earlier described (Bennike et al., 2015; Cehofski et al., 2015b). The samples were thawed on ice and a volume of 200 µL cold PBS was added. The samples were then transferred to impact resistant bead beading tubes containing 0.9-2.0 mm steel beads. The retinal tissue was homogenized by placing the bead beading tubes on a Vortex mixer for 30 seconds followed by bead beading in a Bullet Blender (Next Advance Inc.) for four minutes at speed level 8. The samples were cooled during bead beating. To ensure that all tissue was washed off the bead beading balls another 100 µL cold PBS were added and the samples were placed on a Vortex mixer for 30 seconds. The protein concentration was determined by measuring the absorption at 280 nm (Implen Nanophotometer). A total of 10 µg of lysate from each sample was separated by reducing SDS PAGE (12%, Biorad) and visualized using Krypton fluorescent stain to assess the quality of retinal lysates.
3
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
2.4. Label-free quantification Essentially, label-free liquid chromatograpy tandem mass spectrometry was performed as earlier described (Bennike et al., 2015; Cehofski et al., 2015b). An aliquot of 100 µg retinal homogenate and 100 µL lysis buffer of 10% sodium deoxycholate (NaDOC) in mM triethylammonium bicarbonate (TEAB); pH 8.0, was transferred to 0.5 mL low bind Eppendorf tubes and heat inactivated for 5 min at 95 ºC. Another 250 µL lysis buffer of 10% sodium deoxycholate (NaDOC) in mM triethylammonium bicarbonate (TEAB); pH 8.0, were added. The samples were then reduced with 1.2 µL TCEP followed by homogenization on a Vortex mixer and incubation for 30 min at 60 ºC. The lysates were transferred to YM-10kDa spinfilters (Millipore) and centrifuged at 14,000g for 15 min. Centrifugation was repeated until all liquid was spun out of the filter. The samples were alkylated by adding chloroacetamide (final concentration 50 mM) in 0.5% NaDOC with 50 mM TEAB, homogenized on a Vortex mixer followed by incubation in the dark for 20 min. The samples were centrifuged for 30 min at 14,000g until a residual volume of 50 µL was reached. Buffer exchange was performed by adding 400 µL digestion buffer containing 0.5% NaDOC in 50 mM TEAB; pH 8.0 followed by centrifugation. A volume of 50 µL 0.5% digestion buffer containing 1 µg trypsin was added followed by homogenization on a Vortex mixer. The samples were put into a plastic bag with wet tissue to prevent the samples from drying out followed by incubation overnight at 37 ºC.
TE D
After incubation with trypsin the spin filters containing the retinal samples were transferred to a new collection tube which was used to collect the peptides. The samples were centrifuged at 14,000g for 15 min. A volume of 200 µL of 50mM TEAB was added and centrifugation at 14,000g for 15 min was repeated. NaDOC was removed from the samples by acid extraction with ethyl acetate acidified by adding concentrated formic acid to lower pH to below 2. The samples were mixed on a Vortex mixer and centrifuged at 14,000g for 2 min for phase separation to be obtained. The upper phase was aspirated and discarded. Ethyl acetate extraction was repeated and the entire upper phase was aspirated and discarded. The samples were dried in a vacuum centrifuge and stored at -80ºC.
EP
Before mass spectrometry the samples were resuspended in 20 µL loading buffer consisting of 2% acetonitrile and 0.1% formic acid. The peptide samples were supplemented with 100 fmol MS Qual/Quant Mix (Sigma-Aldrich) as internal standard and analyzed by nanoUPLC tandem mass spectrometry. Prior to nanoUPLC tandem mass spectrometry the peptide concentration was determined by absorption at 280nm by uv-vis-spectroscopy (IMPLEN NanoPhotometer P-300). A volume of 1.5 µL was applied for measurement using lid factor x 50 and blank calibration using dissolution solvent (2% acetonitrile in 0.1% formic acid).
AC C
A sample volume of 5 µl was injected onto a Dionex RSLC nanoUPLC system that was connected to a Quadrupole Orbitrap (Q Exactive Plus) mass spectrometer equipped with a NanoSpray Flex ion source (Thermo Scientific, Bremen, Germany). The flow settings were 8 µL per min for the sample loading onto a trapping column (Acclaim PepMap100 C18 5µm, Thermo Scientific). The nanoflow was set to 300 nL per min for the peptide separation on the analytical column, which was a 50 cm Acclaim Pepmap RSLC, 75 um ID connected with nanoviper fittings. The nano-electrospray was done using a Picotip ‘Silicatip’ emitter from New Objectives. The LC buffers were buffer A (0.1% FA) and buffer B (99.9% acetonitrile, 0.1% formic acid). The applied gradient was from 10% to 30% buffer B over 230 min. The mass spectrometer was operated in data-dependent acquisition mode. A full MS scan in the mass range of 350 to 1400 m/z was acquired at a resolution of 70,000 with an AGC target of 1 x 106 and maximum fill time set to 100 ms. In each cycle, the mass spectrometer would trigger up to 12 MS/MS acquisitions on abundant peptide precursors ions. The MS/MS scans were acquired with a dynamic mass range at a resolution of 17,500 and with an AGC target of 5 x 105 and max fill time of 50 ms. The precursor ions were
4
ACCEPTED MANUSCRIPT isolated using a quadrupole isolation window of 2.0 m/z and fragmented in the HCD trap with a normalized collision energy set to 30. The under-fill ratio was set to 1.0 % with the intensity threshold at 1.0 x 105. Apex triggering was 3 to 10 s with charge and isotopes exclusion on. Dynamic exclusion was set to 30 s. All samples were analyzed in three technical replicates. MS data were submitted to the ProteomXchange Consortium via the PRIDE partner repository with the dataset identifier PXD003044 (Supplementary material – access to MS data).
M AN U
SC
RI PT
Mass spectrometry data were searched using the open-source software MaxQuant (version 1.5.0.31) against a pig protein database optimized from Uniprot containing Sus Scofa proteins and the closest related human protein gene identified based on BLAST (Hesselager et al., 2016). The label-free quantification (LFQ) algorithm was activated in MaxQuant. Standard settings included a peptide and protein false discovery rate of 1%. P-values were calculated by a two-tailed, heteroscedastic t-test. Raw data from the MaxQuant search (Supplementary material – database search output) were submitted to Perseus (version 1.5.2.8) for filtration. Proteins were filtered by unique peptides + razor peptides > 2 followed by removal of proteins identified from a peptide that was found to be part of a protein derived from the reversed part of the decoy database. LFQ values were log2 transformed and triplicates were averaged by mean. Statistical analysis by paired t-test that was conducted for all proteins that were identified in at least four out of five pigs.
TE D
2.5. Isobaric label quantification using mass tag For analysis using isobaric labeling a TMT10plex Mass Tag Labeling kit from Thermo Scientific was used. One hundred µg from each of the 10 samples were transferred into a new tube and 100 mM TEAB was added until a final volume of 100 µL was reached. A volume of 5 µL 200 mM TCEP was added and the samples incubated at 55°C for 1 hour. A volume of 5 µL 375mM iodoacetamide was added to the samples and incubated for 30 minutes at room temperature protected from light and 600 µL of pre-chilled 20% acetone were added and precipitation proceeded overnight. The samples were centrifuged at 8000g for 10 minutes at 4°C. The acetone was decanted and the pellet was allowed to dry for 2-3 minutes. The protein pellets were resuspended with 100 µL of 100 mM TEAB and 2.5 µg trypsin were added to the samples for digestion overnight at 37°C.
AC C
EP
Prior to peptide labeling a volume of 41 µL of anhydrous acetonitrile was added to the 0.8 mg vials containing the TMT label reagents followed by occasional vortexing for 5 minutes and brief centrifugation. Forty-one µL of the TMT label reagent was added to each 100 µL sample followed by incubation for 1 hour at room temperature. A volume of 8 µL of 5% hydroxylamine was added to each sample followed by incubation for 15 minutes. The 10 samples were combined at equal amounts in a new microcentrifuge tube and stored at -80°C. Peptide samples were thawed and purified with a Pierce C18 Spin Column (Thermo Scientific). The spin column was placed in a receiver tube and 200 µL activation solution (50% methanol) were added to wet resin followed by centrifugation at 1500g for 1 minute, the flow-through was discarded and the step was repeated. Subsequently, 200 µL of equilibration solution (0,5% TFA in 5% acetonitrile) were added and the sample was centrifuged at 1500g, the flow-through was discarded and the step was repeated. The sample was loaded on top of the resin bed and the column was placed in a received tube and centrifuged at 1500g for 1 minute. The flow-through was recovered and centrifugation was repeated. The column was placed in a new receiver tube and 200 µL wash solution (0.5%TFA and 5% acetonitrile) were added and the column was centrifuged for 1 minute at 1500g. Wash and centrifugation were repeated. The column was placed in a new
5
ACCEPTED MANUSCRIPT receiver tube and 20 µL elution buffer (70% acetonitrile) were added followed by centrifugation at 1500g for 1 minute, the step was repeated and the sample was dried in a vacuum evaporator.
M AN U
SC
RI PT
Peptides were fractionated using a Pierce High pH Reversed-Phase Peptide Fractionation Kit (Themo Scientific). The spin column was loaded with 300 µL of acetonitrile and centrifuged for 2 minutes at 5000g. This step was repeated. The spin column was washed twice with 0.1% TFA solution. Elution solutions consisting of increasing concentrations of acetonitrile in 0.1% triethylamine were prepared according to instructions provided by the manufacturer. The digested samples were dissolved in 300 µL 0.1% TFA. The spin column was placed into a 2.0 mL sample tube and 300 µL of the sample solution were loaded onto the column that was centrifuged for 2 minutes at 3000g and the eluate was retained as “flow-through” fraction. The spin column was placed in a new 2.0 mL sample tube and 300 µL of water were loaded onto the column that was centrifuged at 3000g for 2 minutes followed by an additional column wash with 300 µL 5 % acetonitrile and 0.1% TEA. The spin column was placed in a new 2.0 mL sample tube and loaded with elution solution and centrifuged at 3000g for 2 minutes and the fraction was collected. This step was repeated with all 8 elution solutions and the collected fractions were dried using vacuum centrifugation. Samples were resuspended in 0.1% formic acid before LC-MS.
AC C
EP
TE D
The labeled samples were analyzed by LC-MS by injecting 5 µL sample volumes into a Dionex RSCL nanoUPLC system connected to an Orbitrap Fusion mass spectrometer equipped with an Easy Spray ion source (Thermo Scientific). The flow settings were 30 µL per min for the sample loading onto a trapping column (5 mm x 300 µm, C18 PepMap100, 5 µm, 100Å, Thermo Scientific). The nanoflow was set to 300 µL per min for separation of peptides on an analystical column (500 mm x 75 µm PepMap RSCL, C18, 2 µm, 100 Å, Thermo Scientific). The buffers were buffer A (0.1% FA) and buffer B (80% acetonitrile, 20% water, 0.1% FA). The applied gradient was performed over either 125 min or 225 min using a gradient of buffer B from 6% to 90%. The Fusion was operated in the TMT SPS MS3 mode using full Orbitrap scans in the mass range 380-1500 m/z acquired at a resolution of 120,000 with an AGC target of 2 x 105 and maximum injection time of 50 ms. In each cycle the mass spectrometer would trigger MS2 acquisitions using the linear ion trap in the mass range 400-1200 m/z with a CID collision energy at 35%, an AGC target of 1 x 104 and a maximum injection time of 50 ms. The precursor ions were isolated using the quadrupole set with an isolation window of 0.7 m/z. Up to 10 reporter ions isolated with a window of 2 m/z were detected in MS3 using synchronous precursor selection in the Orbitrap in the mass range 120-500 m/z with a HCD collision energy at 65% acquired at a resolution of 60,000 with an AGC target of 1 x 105 and a maximum injection time of 120 ms. Dynamic exclusion was set to 30 s or to 70 s. In total 4 technical replicates, each fragmented in 8 samples were performed generating in total 32 data files. The raw data files were further processed in Proteome Discoverer 2.1. The Sequest HT search engine was used against the SwissProt Homo sapiens database (TaxID=9606_and_subtaxonomies; version 2016-04-13) and the SwissProt Sus scrofa database (TaxID=9823_and_subtaxonomies; version 2016-04-13) with a precursor mass tolerance of 10 ppm and a fragment mass tolerance of 0.6 Da. Dynamic modifications were allowed for the TMT labels as well as oxidation of Met, deamidation of Asn and Gln and acetylation of the N-terminal. Static modification was carbamidomethyl on Cys. A false discovery rate of 1% was used and at least 2 unique peptides were required for identification. Quantification was required in each of the 10 samples and each channel was normalized based on total peptide amount.
6
ACCEPTED MANUSCRIPT 2.6. Statistics LFQ values and TMT abundances were log2 transformed prior to statistical analysis. For both datasets a paired t-test was conducted between Ranibizumab injected eyes and NaCl injected eyes in STATA 13.1. The paired t-test was performed for all proteins that were identified in at least four out of five pigs in the nonlabeled analysis and in all pigs in the labeled analysis. Proteins that were significantly changed in both datasets were considered significantly regulated.
SC
RI PT
3. Results 3.1. Experimental branch retinal vein occlusion Based on the development of stagnation of venous blood and peripheral haemorrhages upstream of the site of occlusion, BRVO was successfully induced in six right eyes and six left eyes. Fundoscopic images to document BRVO were taken with a KOWA GENESIS-D handheld retinal camera within the first 30 minutes after BRVO had been induced (Fig. 2). Upon enucleation one animal had developed a retinal detachment in one eye and was excluded from the study. Thus, proteomic analysis was conducted on five retinas from right eyes treated with Ranibizumab and five retinas from left eyes treated with saline water.
M AN U
3.2. Label-free quantification and TMT quantification The present study analyzed large-scale protein changes following BRVO and intervention with Ranibizumab. BRVO was induced in both eyes of the animals and after 24 hours an intervention with Ranibizumab was given in right eyes and saline water was injected into all left eyes. Proteins in the retinal samples were quantified with label-free quantification as well as TMT quantification. Mass spectrometry data were searched against the databases. A paired t-test was used as statistical analysis of both datasets. Proteins that were statistically significantly changed in content following label-free quantification as well as TMT quantification were considered significantly regulated.
EP
TE D
With label-free quantification 2837 proteins were identified in at least 4 animals. In eyes treated with Ranibizumab a total of 237 proteins exhibited a statistically significant change in content based on label-free quantification (Fig. 3) (Supplementary table 1). In the quantification based on labeling 3209 proteins were identified. With TMT quantification we identified 83 proteins that exhibited a statistically significant change in content (Fig. 3) (Supplementary table 2). Five proteins had a statistically significant change in content in both datasets (Fig. 3). These proteins included integrin β-1, peroxisomal 3-ketoacyl-CoA thiolase, OCIA domain-containing protein 1, calnexin and 40S ribosomal protein S5 (Table 1). These five proteins all exhibited a significant decrease in content.
AC C
Platelet derived growth factor receptor-β (PDGFRB) was found be significantly downregulated in the dataset that was generated through label-free quantification (fold change = 0.83; p = 0.041). PDGFRB was not identified with TMT. Findings that were exclusively identified with TMT included high increases in immunoglobulin gamma-1 chain C region (fold change = 7.37; p = 0.0000613) and immunoglobulin kappa chain C region (fold change = 6.64; p = 0.0000696) in the right eyes that received anti-VEGF intervention. With TMT a decline in the content of retinal dehydrogenase 1 was found in retinas treated with Ranibizumab (fold change = 0.34; p = 0.00761). With label-free quantification retinal dehydrogenase 1 was only identified in two animals (fold changes 0.19 and 0.34).
7
ACCEPTED MANUSCRIPT 4. Discussion 4.1. Experimental BRVO In our previous study we analyzed retinal porcine protein changes 15 days after experimental BRVO (Fig. 1A) (Cehofski et al., 2015b). In this previous study validating fluorescein angiography showed successful occlusion after 9 days. Therefore, no re-canalization was to be expected in the present study in which the retinas were collected after 3 days.
SC
RI PT
4.2. Integrin β-1 and OCIA-domain-containing protein 1 Using two fundamentally different proteomic techniques we found that inhibition of VEGF-A with Ranibizumab resulted in a small significant downregulation of integrin β-1. This finding is of interest as our previous experiment (Fig. 1A) (Cehofski et al., 2015b) identified increased levels of integrin β-1 in retinas exposed to experimental BRVO. As integrin β-1 is a key protein in ECM-signaling processes it is likely that BRVO is associated with ECM remodeling processes that are inhibited when VEGF-A signaling is blocked by Ranibizumab (Fig. 4A and Fig. 4B).
M AN U
It is important to note that the retinal samples in our previous study (Fig. 1A) (Cehofski et al., 2015b) and the present study (Fig. 1B) were collected after 15 days and 3 days respectively. In the present study (Fig. 1B) the retinas were collected 3 days after BRVO as there is growing evidence in favor of early inhibition of VEGF and dissection within the acute stage of BRVO. Firstly, our previous study (Fig. 1A) (Cehofski et al., 2015b) found that angiogenesis was not among the biological processes that were altered 15 days after BRVO. Secondly, a study of gene expression following experimental retinal vein occlusion in rat retinas has demonstrated that VEGF expression was upregulated within day 1 of retinal vein occlusion and had normalized 3 days after BRVO (Rehak et al., 2009).
TE D
The anti-integrin Luminate (Allegro Ophthalmics) binds to several integrins including αvβ3, αvβ5, α3β1, α5β1 (Karageozian, 2015). Considering the fact that Luminate is now in phase 2 clinical trials for the treatment of wet age-related macular degeneration and diabetic macular edema (Karageozian, 2015) it is interesting that inhibition of VEGF-A in itself resulted in a small decrease in the content of integrin β-1.
AC C
EP
Interactions between VEGF-A and integrin β-1 have been described by Yamamoto et al. (Yamamoto et al., 2015) in a study on murine retinas which established integrin β-1 as a promoter of endothelial sprouting, vascular ECM deposition and organization of the vessel wall. Oommen et al. (Oommen et al., 2011) demonstrated that VEGF-A can bind to a specific binding site on integrin α9β1 to promote endothelial cell adhesion and migration. Studying human dermal microvascular endothelial cells (HMVEC) Vlahakis et al. (Vlahakis et al., 2007) found that endothelial cell migration induced by either VEGF-A165 or VEGF-A121 was inhibited by α9β1 blocking antibodies. Furthermore, VEGF-A-induced angiogenesis in chicken embryo chorioallantoic assay was inhibited when an integrin α9β1 antibody was added. Ovarian cancer immunoreactive antigen domain containing 1 (OCIA-domain-containing protein 1) has been shown to downregulate cell attachment to integrin β-1 in ovarian cancer cell lines. Thus, the decreased levels of integrin β-1 and OCIA-domain-containing protein 1 following anti-VEGF treatment may potentially affect recruitment of cells to retinal areas affected by the occlusion (Wang et al., 2010). 4.3. Peroxisomal 3-ketoacyl-CoA thiolase, calnexin and 40S ribosomal protein S5 Peroxisomal 3-ketoacyl-CoA thiolase is involved in peroxisomal β-oxidation of fatty acids (Hu et al., 2005), but its role in retinal diseases and interactions with VEGF-A remain largely unstudied. Calnexin is a lectin
8
ACCEPTED MANUSCRIPT
SC
RI PT
chaperone protein that has a key function in the protein quality control of the endoplasmic reticulum (Wang et al., 2015) where it controls that newly synthesized glycoproteins are properly assembled and folded (Rosenbaum et al., 2006; Wang et al., 2015). Rosenbaum and co-workers (Rosenbaum et al., 2006) demonstrated that calnexin has multiple functions in the retina in Drosophila where it regulates cytosolic Ca2+ and promotes rhodopsin (Rh1) maturation. Compromised regulation of cytosolic Ca2+ and rhodopsin maturation resulted in retinal degeneration in flies with mutations in the calnexin gene. Interactions between calnexin and VEGF have been elucidated by Demeure et al. (Demeure et al., 2016) who found that antiVEGF treatment with intraperitoneal Bevacizumab injections increased calnexin levels in glioblastoma xenografts in rats that received weekly injections for three weeks. Thus, it is possible that the endoplasmic reticulum reacts differently on prolonged anti-VEGF treatment. The role of calnexin in retinal diseases and potential interactions with VEGF remain to be established, but given the decrease in 40S ribosomal protein S5, the lower content of calnexin in retinas treated with Ranibizumab may represent changes in glycoprotein trafficking that occur following VEGF-A inhibition.
M AN U
4.4. Platelet derived growth factor receptor-β In retinas treated with Ranibizumab a small decrease in the content of PDGFRB was identified with labelfree quantification. The finding is of interest as drugs targeted at platelet-derived growth factor homodimer BB (PDGF-BB) are under development for the treatment of wet age-related macular degeneration (Tolentino et al., 2015). For example, Fovista (Ophthotech) binds to platelet-derived growth factor homodimer BB to prevent it from interacting with its receptor PDGFRB (Tolentino et al., 2015). PDGFRB was not identified with TMT and additional studies are needed to examine if intravitreal anti-VEGF injections decrease the level of the retinal content of PDGFRB.
TE D
4.5. Immunoglobulin gamma-1 chain C region and immunoglobulin kappa chain C region In the TMT experiment substantial increases in the content of immunoglobulin gamma-1 chain C region and immunoglobulin kappa chain C region were identified. As Ranibizumab is an IgG1 kappa isotype monoclonal antibody Fab fragment (Meyer and Holz, 2011) it is likely that immunoglobulin gamma-1 chain C region and immunoglobulin kappa chain C region are components of Ranibizumab. The finding indicates that Ranibizumab successfully reached the retina after intravitreal injection.
AC C
EP
4.6. Retinal dehydrogenase 1 The TMT experiment showed that the content of retinal dehydrogenase 1 was significantly decreased in retinas treated with Ranibizumab. In the experiment based on label-free proteomics retinal dehydrogenase 1 was detected in two animals (fold changes 0.19 and 0.34). Retinal dehydrogenase 1 irreversibly oxidizes retinaldehyde to retinoic acid, the most potent metabolite of vitamin A (Bchini et al., 2013). Thus, our study raises the question if anti-VEGF treatment may affect the production of retinoic acid in the retina. 5. Conclusion In retinas that received Ranibizumab intervention after BRVO we identified decreased contents of integrin β1, peroxisomal 3-ketoacyl-CoA thiolase, OCIA domain-containing protein 1, calnexin and 40S ribosomal protein S5. The downregulation of integrin β-1 suggests that ECM remodeling and focal adhesion processes associated with BRVO are at least partly reversed when an intervention with Ranibizumab is given. With anti-integrin therapies under development for the treatment of diabetic macular edema and wet age-related macular degeration it is interesting that inhibition of VEGF-A in itself resulted in a small decrease in the content of integrin β-1. With label-free quantification a small downregulation of PDGFRB was identified in
9
ACCEPTED MANUSCRIPT eyes that received Ranibizumab treatment. Further studies are required validate if intravitreal anti-VEGF treatment decreases retinal PDGFRB levels.
RI PT
Fig. 1. (A) Experimental set-up from previous study (Cehofski et al., 2015b). In this previous work BRVO was induced in the right eyes of the animals by applying argon green laser directly onto a vein in the inferior retina. In the left eyes that served as controls a similar area of laser burns was created in an area devoid of vessels. The retinas were dissected after 15 days followed by proteomic analysis. (B) In the present intervention study BRVO was induced in both eyes of the animals. After 24 hours VEGF-A was inhibited in all right eyes with an intravitreal injection of Ranibizumab while all left eyes received an intravitreal injection of 0.9% sodium chloride. The retinas were dissected after 3 days.
SC
Fig. 2. Image of laser-induced experimental BRVO taken with a KOWA GENESIS handheld retinal camera approximately 20 minutes after BRVO was induced. (A) Dilation of branch vein and stagnation of venous blood following laser induced occlusion ( : Area of laser treatment, : Peripheral haemorrhages, : Dilation of vein). (B) Peripheral haemorrhages upstream to the occlusion ( : Peripheral haemorrhages).
M AN U
Fig. 3. Summary of results. The retinal proteins were quantified with label-free quantification as well as tandem mass tag quantification. With label-free quantification we identified 237 significantly changed proteins while 77 significantly changed proteins were identified with TMT. There was an overlap of five significantly changed proteins between the two datasets. These proteins included integrin β-1, peroxisomal 3ketoacyl-CoA thiolase, OCIA domain-containing protein 1, calnexin and 40S ribosomal protein S5.
TE D
Fig. 4. Hypothesis about protein changes following experimental BRVO. (A) We earlier identified an upregulation of integrin β-1 following experimental BRVO. (B) Inhibition of VEGF-A resulted in a downregulation of integrin β-1 suggesting that ECM remodeling, adhesion, migration and proliferation processes associated with experimental BRVO are downregulated due to intervention with Ranibizumab.
EP
Financial support This work was supported by the Svend Andersen Foundation, the Bagger-Sørensen Foundation, the Obel Family Foundation, the Herta Christensen Foundation, the North Denmark Region (2013-0076797) and Heinrich Kopps Foundation.
AC C
Disclosures Ranibizumab was provided complimentarily by Novartis Ophthalmics, Denmark. The authors have no financial interests to declare. Contributions Lasse Cehofski, Anders Kruse, Bent Honoré and Henrik Vorum conceived the idea of studying the retinal proteome following experimental BRVO and intervention with Ranibizumab. All authors participated in the writing of the manuscript. Lasse Cehofski performed experimental, laboratory work and data analysis. Anders Kruse performed laser treatments. Sigriður Olga Magnusdottir anesthetized the animals. Martin Bøgsted contributed to statistical analyses. Allan Stensballe conducted sample preparation for mass spectrometry, mass spectrometry analysis and data analysis. Bent Honoré participated in study design, data analysis, the writing of the manuscript and conducted the TMT experiments. The study was supervised by Henrik Vorum who participated in study design and data analysis.
10
ACCEPTED MANUSCRIPT References
AC C
EP
TE D
M AN U
SC
RI PT
Bchini, R., Vasiliou, V., Branlant, G., Talfournier, F., Rahuel-Clermont, S., 2013. Retinoic acid biosynthesis catalyzed by retinal dehydrogenases relies on a rate-limiting conformational transition associated with substrate recognition. Chem. Biol. Interact. 202, 78-84. Bennike, T.B., Carlsen, T.G., Ellingsen, T., Bonderup, O.K., Glerup, H., Bogsted, M., Christiansen, G., Birkelund, S., Stensballe, A., Andersen, V., 2015. Neutrophil Extracellular Traps in Ulcerative Colitis: A Proteome Analysis of Intestinal Biopsies. Inflamm. Bowel. Dis. 21, 2052-2067. Campochiaro, P.A., Choy, D.F., Do, D.V., Hafiz, G., Shah, S.M., Nguyen, Q.D., Rubio, R., Arron, J.R., 2009. Monitoring ocular drug therapy by analysis of aqueous samples. Ophthalmology 116, 2158-2164. Campochiaro, P.A., Heier, J.S., Feiner, L., Gray, S., Saroj, N., Rundle, A.C., Murahashi, W.Y., Rubio, R.G., 2010. Ranibizumab for macular edema following branch retinal vein occlusion: six-month primary end point results of a phase III study. Ophthalmology 117, 1102-1112.e1101. Cehofski, L.J., Kruse, A., Kjaergaard B, Stensballe A, Honore, B., Vorum, H., 2015a. Dye Free Porcine Model of Experimental Branch Retinal Vein Occlusion – a Suitable Approach for Retinal Proteomics. J. Ophthalmol., article ID 839137. Cehofski, L.J., Kruse, A., Kjaergaard, B., Stensballe, A., Honore, B., Vorum, H., 2015b. Proteins involved in focal adhesion signaling pathways are differentially regulated in experimental branch retinal vein occlusion. Exp. Eye Res. 138, 87-95. Cehofski, L.J., Mandal, N., Honore, B., Vorum, H., 2014. Analytical platforms in vitreoretinal proteomics. Bioanalysis 6, 3051-3066. Demeure, K., Fack, F., Duriez, E., Tiemann, K., Bernard, A., Golebiewska, A., Bougnaud, S., Bjerkvig, R., Domon, B., Niclou, S.P., 2016. Targeted Proteomics to Assess the Response to Anti-Angiogenic Treatment in Human Glioblastoma (GBM). Mol. Cell. Proteomics 15, 481-492. Gaudreault, J., Fei, D., Rusit, J., Suboc, P., Shiu, V., 2005. Preclinical pharmacokinetics of Ranibizumab (rhuFabV2) after a single intravitreal administration. Invest. Ophthalmol. Vis. Sci. 46, 726-733. Hesselager, M.O., Codrea, M.C., Sun, Z., Deutsch, E.W., Bennike, T.B., Stensballe, A., Bundgaard, L., Moritz, R.L., Bendixen, E., 2016. The Pig PeptideAtlas: A resource for systems biology in animal production and biomedicine. Proteomics 16, 634-644. Hu, T., Foxworthy, P., Siesky, A., Ficorilli, J.V., Gao, H., Li, S., Christe, M., Ryan, T., Cao, G., Eacho, P., Michael, M.D., Michael, L.F., 2005. Hepatic peroxisomal fatty acid beta-oxidation is regulated by liver X receptor alpha. Endocrinology 146, 53805387. Karageozian, V., 2015. Investigational drug Luminate targets integrin receptors. Ophthalmol. Times, 949/940-8130. Meyer, C.H., Holz, F.G., 2011. Preclinical aspects of anti-VEGF agents for the treatment of wet AMD: ranibizumab and bevacizumab. Eye (Lond) 25, 661-672. Noma, H., Funatsu, H., Mimura, T., Eguchi, S., Hori, S., 2011. Soluble vascular endothelial growth factor receptor-2 and inflammatory factors in macular edema with branch retinal vein occlusion. Am. J. Ophthalmol. 152, 669-677.e661. Noma, H., Funatsu, H., Mimura, T., Tatsugawa, M., Shimada, K., Eguchi, S., 2012. Vitreous inflammatory factors and serous macular detachment in branch retinal vein occlusion. Retina 32, 86-91. Noma, H., Mimura, T., Yasuda, K., Shimura, M., 2014. Role of soluble vascular endothelial growth factor receptors-1 and -2, their ligands, and other factors in branch retinal vein occlusion with macular edema. Invest. Ophthalmol. Vis. Sci. 55, 3878-3885. Oommen, S., Gupta, S.K., Vlahakis, N.E., 2011. Vascular endothelial growth factor A (VEGF-A) induces endothelial and cancer cell migration through direct binding to integrin {alpha}9{beta}1: identification of a specific {alpha}9{beta}1 binding site. J. Biol. Chem. 286, 1083-1092. Rehak, M., Hollborn, M., Iandiev, I., Pannicke, T., Karl, A., Wurm, A., Kohen, L., Reichenbach, A., Wiedemann, P., Bringmann, A., 2009. Retinal gene expression and Muller cell responses after branch retinal vein occlusion in the rat. Invest. Ophthalmol. Vis. Sci. 50, 2359-2367. Rosenbaum, E.E., Hardie, R.C., Colley, N.J., 2006. Calnexin is essential for rhodopsin maturation, Ca2+ regulation, and photoreceptor cell survival. Neuron 49, 229-241. Tolentino, M.J., Dennrick, A., John, E., Tolentino, M.S., 2015. Drugs in Phase II clinical trials for the treatment of age-related macular degeneration. Expert Opin. Investig. Drugs 24, 183-199. Vlahakis, N.E., Young, B.A., Atakilit, A., Hawkridge, A.E., Issaka, R.B., Boudreau, N., Sheppard, D., 2007. Integrin alpha9beta1 directly binds to vascular endothelial growth factor (VEGF)-A and contributes to VEGF-A-induced angiogenesis. J. Biol. Chem. 282, 15187-15196. Wang, C., Michener, C.M., Belinson, J.L., Vaziri, S., Ganapathi, R., Sengupta, S., 2010. Role of the 18:1 lysophosphatidic acidovarian cancer immunoreactive antigen domain containing 1 (OCIAD1)-integrin axis in generating late-stage ovarian cancer. Mol. Cancer Ther. 9, 1709-1718. Wang, Q., Groenendyk, J., Michalak, M., 2015. Glycoprotein Quality Control and Endoplasmic Reticulum Stress. Molecules 20, 13689-13704. Yamamoto, H., Ehling, M., Kato, K., Kanai, K., van Lessen, M., Frye, M., Zeuschner, D., Nakayama, M., Vestweber, D., Adams, R.H., 2015. Integrin beta1 controls VE-cadherin localization and blood vessel stability. Nat. Commun. 6, 6429.
11
ACCEPTED MANUSCRIPT Table 1 Protein exibiting a significant change in content in both proteomic methods Protein
Fold change LFQ p-value LFQ
Fold change TMT p-value TMT
Gene name
0.78
0.00798
0.92
0.00728
ITGB1
0.81
0.00744
0.91
0.0277
ACAA1
OCIA domain-containing protein 1
0.81
0.0460
0.90
0.0357
OCIAD1
40S ribosomal protein S5
0.77
0.0403
0.89
0.00299
RPS5
Calnexin
0.80
0.00468
0.93
0.0495
CANX
AC C
EP
TE D
M AN U
SC
RI PT
Integrin β-1 Peroxisomal 3-ketoacyl-CoA thiolase
ACCEPTED MANUSCRIPT
BRVO vs. control - previous study Right retina - laser induced BRVO
Dissection
Left retina - lasered, no BRVO Argon laser
A
0
1
2
3
4
5
Right retina - laser induced BRVO
TE D
Intervention study - present study
Left retina - laser induced BRVO
Figure 1
15
Left eye 0.9% NaCl
Days
Dissection
2
3
4
0.200
0.200
0.150
0.150
0.10
0.10
AC C B
1
Right eye anti-VEGF
10
Argon laser
EP
Argon laser
0
M AN U
SC
RI PT
Argon laser
0.050
0.050
5
10
15
Days
A
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
B
Figure 2
ACCEPTED MANUSCRIPT
RI PT
Summary of results Tandem mass tag quantification 83 significantly changed proteins
M AN U
SC
Label-free quantification 237 significantly changed proteins
5
83
EP
TE D
237
AC C
Significantly changed proteins identified in both datasets - Integrin β-1 - Peroxisomal 3-ketoacyl-CoA thiolase - OCIA domain-containing protein 1 - Calnexin - 40S ribosomal protein S5 Figure 3
ACCEPTED MANUSCRIPT
Experimental BRVO and inhibition of VEGF-A
RI PT
Experimental BRVO
Integrin-β1
M AN U
SC
Integrin-β1
Extracellular matrix
TE D
Extracellular matrix
Cytosol
EP
Cytosol
Adhesion Migration Proliferation
AC C
Adhesion Migration Proliferation
A Figure 4
B
ACCEPTED MANUSCRIPT The content of integrin β-1 was decreased following treatment with Ranibizumab. Calnexin and 40S ribosomal protein S5 decreased in content when VEGF-A was blocked. OCIA domain-containing protein 1 was downregulated following inhibition of VEGF-A.
AC C
EP
TE D
M AN U
SC
RI PT
A role of integrin β-1 in branch retinal vein occlusion is proposed.
