University of Groningen Donation of kidneys after brain death van Dullemen, Leon

(1)

Donation of kidneys after brain death

van Dullemen, Leon

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from

it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date:

2017

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

van Dullemen, L. (2017). Donation of kidneys after brain death: Protective proteins, profiles, and treatment

strategies. Rijksuniversiteit Groningen.

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

(2)

512316-L-bw-Dullemen 512316-L-bw-Dullemen 512316-L-bw-Dullemen 512316-L-bw-Dullemen Processed on: 9-8-2017 Processed on: 9-8-2017 Processed on: 9-8-2017

Processed on: 9-8-2017 PDF page: 221PDF page: 221PDF page: 221PDF page: 221

Deceased donor kidney

proteomic profiles correlate

with 12-month allograft

function after transplantation

Maria Kaisar

Leon F.A. van Dullemen+ M. Zeeshan Akhtar+ Marie-Laëtitia Thézénas Honglei Huang Nicholas A. Watkins Benedikt M. Kessler Rutger J. Ploeg

*Authors contributed equally

(3)

222

ABBREVIATIONS

BD: Brain Death CAT: Catalase

CIT: Cold Ischaemia Time CKD: Chronic Kidney Dysfunction DGF: Delayed Graft Function DBD: Donation after Brain Death DCD: Donation after Circulatory Death ECD: Extended Criteria Donors FDR: False Discovery Rate GO: Good Outcome Group GST: Glutathione S-Transferase HCA: Hierarchical Cluster Analysis HLA: Human Leukocyte Antigen IF: Immediate Function

IRI: Ischaemia Reperfusion Injury LD: Living Donor

MAPK1: Mitogen Activated Protein Kinase-1 NHS: National Health Service UK

NHSBT: NHS Blood and Transplant

PDGFRa: Platelet-Derived Growth Factor Receptor-a PRX3: Peroxiredoxin-3

QUOD: Quality in Organ Donation

SHCA: Supervised Hierarchical Cluster Analysis SO: Suboptimal Outcome Group

STAT1: Signal Transduced and Activator of Transcription-1 TGFb1: Transforming Growth Factor-b1

(4)

8

223

ABSTRACT

Donation after brain death (DBD) is the main source of deceased donor kidneys for transplantation. Complex and abnormal pathophysiological changes following brain death adversely impact the graft quality following transplantation. Diagnostic tools for donor organ quality as predictors of transplantation outcomes could improve donor organ utilisation and guide targeted interventions to repair potential grafts in the donor or during preservation. This study was performed to discover what pathways affect DBD kidney graft quality.

Procurement DBD kidney biopsies were obtained from the UK QUOD biobank for analysis and selected on the basis of post-transplant glomerular filtration rate. Renal function of both kidney recipients with the same donor were consulted for sample selection. Analysed biopsy samples were obtained from donors of whom both kidneys had developed delayed graft function (DGF) and a 3-month mean eGFR ≤29.8±7ml/min (suboptimal outcome). These samples were compared to biopsy samples from kidneys that functioned immediately and had at a 3-month mean eGFR ≥ 65.1±8ml/min (good outcome). Samples were analysed by label free proteomics (n=10) and candidate markers were validated in another set of donor kidney samples (n=28). There was no difference in acute or chronic kidney injury between the two analysed donor groups according to the AKIN and Remuzzi score. However, we could differentiate the donor kidneys that had a suboptimal function from those with a good function after transplantation on the basis of a proteomic profile. Pro-fibrotic and apoptotic markers correlated to suboptimal allograft function while increased antioxidant protein levels were associated with good functioning allografts.

(5)

224

INTRODUCTION

Organ transplantation is a life-saving treatment for patients with end stage organ disease. However, the persistent shortage of deceased donor organs meant that, whilst on the waiting list between April 2014 and March 2015, in the United Kingdom 429 patients died and 807 patients became too ill to receive a transplant (1). Due to our aging population with a higher incidence of comorbid conditions the demand for transplants may further increase as a consequence of increasing prevalence of diabetes, hypertension, and obesity. To reduce morbidity, mortality, and the time patients have to spend on the waiting list, the deceased donor pool has been expanded to include extended criteria donors with co-morbidities (ECD) (2-4). As a result the utilisation of organs from older donors has increased by 20% and currently donors over 60 years old represent almost 40% of the donation after brain death (DBD) donors in the UK (1).

Allografts transplanted from ECDs have suboptimal transplant outcomes compared to those obtained from standard criteria or living donors (LDs) (5,6). Early kidney dysfunction in the recipient manifests itself as primary non function (PNF) or delayed graft function (DGF) requiring dialysis treatment and may result in complications, prolonged hospitalisation, and increased risk of chronic allograft dysfunction (CAD) on the long-term(7,8). Despite the development of new immunosuppressive drugs to prevent acute rejection, a progressive decline of kidney function over time and the development of CAD leads to the loss of 50% of kidney transplants within a decade of transplantation, a clinical outcome that has been unchanged since 1989 (9,10).

Brain death in the donor induces a profound pro-inflammatory and pro-coagulatory systemic response that increases immunogenicity in the allografts (11-13). This hostile non-physiological environment makes grafts vulnerable to prolonged cold ischaemia time (CIT) and ischaemia-reperfusion injury (IRI) following engraftment, resulting in higher rates of DGF and suboptimal long-term kidney function and graft survival (14). To yield good transplant outcomes and decrease the waiting list it is important to select donor kidneys of good (enough) quality for transplantation, while preventing the unnecessary discard of organs.

To date, there remains a lack of objective criteria to assess donor organ quality helping the clinical decision at time of offering whether to accept or decline the donor kidney for a particular recipient. Current methods are limited to a few surrogate markers of kidney function during donor management together with other potential risk factors such as donor age or e.g. histological assessment using the Remuzzi scoring. This uncertainty may result in possibly unnecessary decline and the discard of donor kidneys.

To discover sensitive and specific markers, we have recently established the UK Quality in Organ Donation (QUOD) biobank to collect longitudinal blood, urine, and biopsy samples from deceased donors and provide a link to donor and recipient clinical and demographic data(15). In addition, we have created a platform to analyse the large number of well curated clinical samples using powerful analytical methods such as proteomics to identify new molecular profiles associated with clinical relevant phenotypes in donation and transplantation.

(6)

8

225 Proteomic profiling by mass spectrometry can rapidly identify significant changes to proteins in tissues and has been widely used in medical research including the field of transplantation with great promise (16,17).

The aim of this study was to evaluate the hypothesis that we can discriminate between donor kidneys that will have good- or suboptimal long-term outcome post transplantation by comparing the proteomic profile of donor kidney biopsies from both groups. Furthermore, we validated the main findings of our proteomic analysis in another set of donor kidney biopsy samples and show how activation of biological pathways of injury and cytoprotection in the donor affects kidney function in renal transplantation.

MATERIALS AND METHODS

Study samples

Kidney biopsy samples from deceased brain dead donors were obtained from the UK QUOD biobank following approval from the QUOD steering committee. Kidney biopsies were obtained

ex-situ from the upper pole of kidney cortex during preparation at the back table using a 23mm

needle biopsy gun. Each biopsy specimen was divided in two; one half was stored in RNAlater followed by subsequent storage in liquid nitrogen and the other half in formalin. The study was performed under the ethical approval of QUOD project 13/NW/0017

Sample selection

Donor kidney biopsy samples (n=38) were selected on the basis of recipient transplantation outcomes of both kidneys transplanted per donor. We interrogated National Health Service UK Blood and Transplant - Organ Donation and Transplantation (NHSBT ODT) database to identify pairs of recipients from the same donor with 3-month post-transplantation follow up clinical data. We created two subgroups on the basis of the eGFR in both recipients at 3 months post-transplant. In the suboptimal (SO) cohort at least one donor kidney from each donor pair had developed DGF and together the recipients had a 3-month mean eGFR £ 33.4±10 ml/min. In the good outcome cohort none of the grafts developed DGF and the mean eGFR at 3 months was ³ 62.5±10 ml/min in the recipients. The donor kidneys selected (one of the two per donor) for further analysis (n=19 per group) had the following kidney function at 3 months; the SO group all developed DGF and had a mean eGFR £ 29.8±7 ml/min, while in the GO group non of the grafts developed DGF and had a mean eGFR of ³ 65.1±8 ml/min.

DGF was defined as the need for dialysis within the first week post-transplantation after excluding urine obstruction and hypercalcaemia. The experimental protocol is shown in Figure 1. Baseline donor and recipient demographic characteristics were considered to create as homogeneous groups as possible, and there were no significant differences between risk factors such as age and CIT (Table 1). After the start of this study we acquired the 12-month eGFR follow-up values for the selected cohort of recipients from NHSBT ODT as shown in Table 1.

(7)

226

AKIN assessment

AKIN assessment was performed by calculating the fold change of terminal serum creatinine compared to the baseline serum creatinine at donor admission in the intensive care (18). Fold change <1.5-2 was defined as AKIN-1, 2-3 as AKIN-2, and >3 fold change as AKIN-3.

Histological scoring

Kidney paraffin sections (4µm) were de-waxed, rehydrated, and stained with 0.5% periodic-acid for 5 minutes, rinsed with distilled water and then placed in Schiffs reagent for 15 minutes followed by 5 minutes washing in tap-water. The slides were counter-stained with Mayer’s haematoxylin and washed in tap water for 5 minutes. Remuzzi scoring assessment was performed blindly by a pathologist.(19)

LC-MS/MS proteomic analysis

Kidney biopsy samples were lysed in 300 µL of RIPA buffer (150 mM NaCl, 1.0% NP-40, 0.5% sodium deoxycholate, 1% SDS, 50 mM Tris, pH 8.0) containing protease (Roche, USA) and phosphatase inhibitor cocktail (Sigma, UK). Homogenization was performed on a beads beater at 6500 rpm for 40 seconds.

We analysed 5 individual biopsy samples per subgroup by LFQ-MS/MS (Figure 1B). Fifteen micrograms of protein were used for each individual sample. Peptide tryptic digests were generated from the biopsy proteins and were analysed in duplicates by nano ultra-high performance liquid chromatography tandem mass spectrometry LC-MS/MS. Samples were reduced for 1 hour by addition of 200 mM dithiothreitol (DTT) followed by alkylation with 200 mM iodoacetamide (IAA) for 30 minutes. Trypsin digestion was performed overnight at 37 °C with gentle mixing using a 1:50 trypsin: protein ratio. Samples were acidified with 1% FA or TFA. Peptide digests were then desalted using Sep-Pak C18 cartridges (Waters) and dried by Speed Vac centrifugation. Pellets were resuspended in 30 µl of buffer A (98 % Milli-Q-H2O, 2 % acetonitrile, 0.1 % formic acid) prior to LC-MS/MS analysis. Peptides were analysed in duplicates by nano ultra-high performance liquid chromatography tandem mass spectrometry (nUHPLC-MS/MS) using a Dionex Ultimate 3000 UHPLC (C18 column with a 75 μm × 250 mm, 1.7 μm particle size; Thermo coupled to a Q Exactive tandem mass spectrometer; Thermo Scientific, Bremen, Germany). MS data were processed and identified proteins were quantified using an in house bioinformatics tool pipeline CPFP (20). Protein abundance, SING intensity, was calculated based on spectral counts of the total number of MS/MS spectra identified per protein as described (21).

Hierarchical clustering analysis

Unsupervised hierarchical clustering analysis was performed using the individual abundance values of all the identified proteins without a priori assumptions. Supervised analysis included all the proteins significantly changed among the experimental groups (p<0.05, T-test). Unsupervised and supervised hierarchical clustering analysis and data visualisation was performed using the PermutMatrix software (22). For a given protein, missing abundance data were imputed with the mean value of protein expression for all the runs in the same cohort.

(8)

8

227 Data was imputed when only 40% of individual samples had missing data. The proteins with missing values for more than 40% of samples per cohort for a specific protein were excluded from the analysis. Dissimilarity between the 10 donors was assessed by single linkage (Closest neighbour linkage) for columns and rows and the dendrogram was constructed using multi fragment heuristic as a tree seriation algorithm.

Validation

Samples containing 15 µg of protein were denatured at 95°C for 5 minutes in Laemmli buffer and loaded into 8–12% pre-cast SDS-PAGE gels (Bio-Rad, USA). Proteins separated by SDS-PAGE were transferred to hydrophobic PVDF membranes (Merck Millipore, USA) in transfer buffer (25 mM Tris, 192 mM glycine and 10% methanol) overnight. PVDF membranes were blocked for 1 hour in TBST buffer (25 mM Tris, pH 7.5, 0.15 M NaCl, 0.05% Tween 20) containing 5% milk. Membranes were incubated overnight at 4 °C with anti TRX1 (ab26320, 1:1000), PRX3 (ab73349, 1:1000), Catalase (ab16731, 1:1000), GST (ab53940, 1:3000)(Abcam, UK), STAT-1 (mab14901, 1:1000), PDGFRα (af307na, 1:800)(R&D Systems, US) antibodies. Rabbit monoclonal anti Beta-actin served as a loading control (ab8227, 1:5000) (Abcam, UK). Membranes were washed for 30 minutes with 5 changes of wash buffer and then incubated at room temperature for 1 hour in blocking buffer containing a 1:5000 dilution of Dye-800-conjugated anti-mouse, -rabbit, or -goat secondary antibody (Li-Cor, Nebraska, USA), and visualization was performed with Odyssey CLx (Li-Cor Nebraska, USA). Detected signal was quantified and normalized to the beta-actin signal on the same blot.

Statistical analyses

Statistical comparison of protein abundance changes observed between GO and SO was performed using the Mann Whitney-U test and accepting as significant baseline p<0.05. Candidate proteins were shortlisted based on a significant change between the two groups (SO vs GO) with a protein identified in at least three of the five kidney samples per group. The Chi squared or Fisher’s exact test were used to compare categorical variables.

RESULTS

Acute kidney injury network (AKIN) classification assessment and Remuzzi scoring of deceased donor biopsies

To assess if donors included in the study had developed acute kidney injury (AKI) during donor management we evaluated the donor AKIN classification (18). Table 1 shows there is no significant difference in AKIN stages between the two experimental cohorts with 90% of the donor kidneys with a suboptimal outcome (SO) and 95% with good outcome (GO) classified as AKIN 1 stage.

We also evaluated the procurement biopsies for chronic kidney disease (CKD) using Remuzzi scoring (19). Retrospectively, we assessed biopsies from all the 38 donors. We successfully scored biopsies from n=30 donors, as 8 donors had insufficient glomeruli for scoring.

(9)

Sixty-512316-L-bw-Dullemen 512316-L-bw-Dullemen 512316-L-bw-Dullemen 512316-L-bw-Dullemen Processed on: 9-8-2017 Processed on: 9-8-2017 Processed on: 9-8-2017

228

eight% of the donor kidneys with SO and 79% of donor kidneys with GO had a Remuzzi score between 0-3. Overall there was no difference in the AKIN classification or Remuzzi score of the donors included in the analysis (Table 1).

Proteomic signature differentiates DBD kidneys according to their transplantation outcomes

Initially, we analysed 10 individual kidney biopsy samples (Figure 1B) by label free quantitative proteomic analysis (n=5 per transplantation outcome) and found 1743 unique proteins with a false discovery rate (FDR) <1%. To explore the uniformity of data within each group and to investigate the degree of closeness in the two groups we performed hierarchical clustering analyses (HCA) (22). On the basis of similarity between individual samples, unsupervised HCA placed 4 out of 5 donor kidneys with the same outcome in adjacent positions (Figure 2A). Subsequently we shortlisted 214 (Table S1) proteins that were significantly altered between the two groups (p£0.05). Supervised hierarchical clustering analysis (sHCA) using the proteomic signature of these 214 proteins separated the biopsy samples into two clearly distinct groups according to transplantation outcomes (Figure 3A).

After analysing the data we subsequently acquired the 12-month eGFR values of the kidney recipients. Figure 3B shows the association between the sHCA derived dendrogram to the recipient kidney function observed at the 3- and 12-month follow up. This further supported the results of the hierarchical clustering to recipient kidney function (Figure 3B). The donor kidneys with the lowest and highest eGFR values at 3- and 12-month post transplantation were clustered furthest apart while the suboptimal S1 and S2 kidneys, which had a noticeable improvement at 12-month eGFR follow-up, were segregated towards the centre of the dendrogram (Figure 3B). For the good outcome kidneys, those that clustered furthest to the right (G2, G4, G3) had consistently good function at 3- and 12-months (Figure 3B). The association of the hierarchical clustering of the donor kidneys with kidney function in the recipients at 3- and 12-months indicates that the kidney proteomic signature may be useful to predict post-transplantation outcome.

Donor kidney proteins indicate activation of injury and repair pathways

The magnitude of the change in protein abundance and the statistical significance of those changes were plotted in a volcano plot (Figure 2B), revealing a greater proportion of the significant altered proteins were increased in the suboptimal group (right side).

Pathway analysis using Reactome and String open access bioinformatics of these 214 proteins indicated that they were enriched for proteins regulating cellular metabolic processes including ROS detoxification (n=107, with a FDR: 2 10E-3), cellular response to stress with (n=26 proteins; FDR:2 10E-3), and proteins related to cell surface receptor signalling pathways (n=40; FDR: 3 10E-3) (Figure 4). Among the predominant KEGG pathways were metabolic dysregulation (n=107 FDR 7.13 10E-6) and tight junction molecules (n=40 proteins; FDR; 310E-3).

Pro-fibrotic and apoptotic protein levels were increased in donor kidneys with SO

Proteomic data analysis revealed a significant increase of key mediators in the SO group that was associated with injury and cellular stress. Among these proteins were the signal

(10)

8

229 transduction proteins MAPK-1, TGF-β1, and STAT-1. The levels of both TGF-β1 (p=0.04) and STAT-1 (p=0.0003) increased by 7-fold in the kidneys in SO compared to GO kidneys, and the levels of MAPK-1 were increased by 2.1-fold (p=0.03) (Figure 4). In addition, there was also an increase of proteins in the SO compared to the GO group that affect the immune response, including a 3.1-fold increased abundance of H-FABP (p=0.007) and a 2.2-fold increase of basigin (p=0.0001).

As STAT-1 was identified in all SO biopsies analysed by proteomics and since it demonstrated the most significant change between the two experimental groups (7-fold, p=0.0003), we further validated this protein by Western blot. Western blot validation and densitometric quantification of ß-actin normalised intensities values was performed in an independent cohort of n=28 donor kidney biopsy samples and confirmed the significant increased expression of STAT-1 in the SO compared to GO kidneys (p=0.02) (Figure 5). To investigate whether there was evidence of increased occurrence of kidney fibrosis in the SO in comparison to GO, platelet derived growth factor receptor-a (PDGFRα) was chosen to be analysed by western blot on the validation kidney biopsy cohort mainly because of the synergetic role with TGF-β1 in the onset and propagation of fibrosis. Western blot analysis indeed demonstrated the significantly increased levels of PDGFRα in the donor kidneys with suboptimal transplantation outcomes when compared to good functioning kidneys (p=0.01) (Figure 5).

Antioxidant protein levels were increased in donor kidneys with GO

In contrast to the injury associated proteins with increased levels of expression in the SO kidneys, increased levels of key antioxidant and cytoprotective proteins correlated to kidneys with GO. Initial interrogation of the proteomic data showed a 3.4-fold increased level of thioredoxin reductase-2 (TRX2) (p=0.04). To further investigate the association of increased antioxidant protein expression of donor kidneys with good post-transplantation function, key antioxidant proteins of the thioredoxin group and glutathione metabolism system were analysed by immunoblotting. Thioredoxin-1 (TRX1), glutathione-S transferase (GST), Peroxiredoxin-3 (PRX3), and catalase were analysed by immunoblotting on the validation cohort of n=28 biopsy samples. Western blot validation confirmed that TRX1 (p=0.005), GST (p=0.02), and PRDX3 (p=0.04) had significantly increased levels in the donor kidneys with GO (Figure 6). In contrast to the rest of the antioxidant proteins, catalase was not significantly changed between the two (Figures 4, 6).

DISCUSSION

The lack of diagnostic tools to assess donor organ quality has led to high kidney discard rates in addition to unsatisfactory long-term kidney function post transplantation (23). In this study we show that based on donor kidney proteomic signatures we can differentiate at time of retrieval the donor kidneys that will develop suboptimal outcomes from those donor kidneys with good, long-term transplantation outcomes (Figure 2, 3). In contrast, the AKIN classification and histological evaluations failed to identify those kidneys at risk for a

(11)

230

suboptimal allograft function. The comparison of the proteomic signature between the GO and SO revealed increased levels of injury mediators in SO kidneys and enhanced levels of cytoprotective proteins in GO kidneys suggesting that the activation of either injury or repair pathways are determinants of kidney function post transplantation (Figure 7).

Our donor inclusion criteria were based on the kidney allograft outcomes to ensure that our observations most likely reflect donor organ quality. Although there was a small improvement in the mean eGFR values between 3- and 12-month follow up (29.8 ± 7 to 35.9 ± 6) for the suboptimal group, kidney recipients from this group were still classified as chronic kidney disease stage 3b as defined by MDRD (24). These kidneys still have an increased hazard to progress to stages 4 and 5 and eventually graft failure as recently (25,26). The sustained reduced kidney function of these suboptimal donors could be the result of an ongoing subclinical inflammation, which is know to inducing injury that could progressively lead to fibrosis (27,28). Our study indicates that unfavourable pathophysiological changes already start in the donor following brain death and could predispose the allograft to future injury and dysfunction. Pathway analysis of our proteomic data set indicates dynamic interactions of metabolic-, cellular stress-, inflammatory catabolic-, and cytoprotective pathways. These observations are consistent with our previous findings of a brain death rodent model using a proteomics and metabolomics integrative approach, showing that dysregulation of metabolic pathways lead to mitochondrial dysfunction and ROS generation, further promoting cellular injury in the kidney (29). The dynamic relationship between ROS and TGF-β has been extensively examined

in vitro and in vivo (30-32). Our analysis show significant increased levels of TGF-β1 in donor

kidneys with SO compared to GO (Figure 4). TGF-β mediates ROS production and suppresses antioxidant activity causing an imbalance in redox homeostasis that could further promote oxidative stress. This potent mediator has a key role in the onset and progression of fibrosis in multiple organ systems. Animal model studies have shown that blocking TGF-β can limit or prevent the progression of fibrosis (33). Considering the key role of TGF-β in the progression of kidney injury it was not a surprise that the increase of TGF-β1 was correlated with increased levels of the pro-fibrotic protein PDGFRα in the same donor cohort (Figures 5). PDGFRα is one of the two components of the PDGF receptors and although all PDGF factors play a pivotal role in healthy kidney development, PDGFRα can promote kidney fibrosis through a maladaptive process of wound healing (34-36). A downstream effect of the PDGF receptors is activation of the proinflammatory genes, one of which is 1.(37). PDGF receptors can induce STAT-1 phosphorylation in a JAKSTAT-1/2 dependent process (38-40). These observations are consistent with our results showing the parallel increase of PDGFRa and STAT-1 and their correlation with suboptimal functioned kidneys. As our study is not a mechanistic study, we can only hypothesize that the increased levels of STAT-1 might also indicate downstream signalling pathways activation of cytokine releasing and pro-apoptotic genes.

The increased expression of STAT-1 in parallel with MAPK-1 could indicate activation of JAK/ STAT and mitogen-activated protein kinase (MAPK)/p38 pathways, leading to inflammation and a graft more susceptible to the recipient’s immune system and the insult of IRI during the process of transplantation.

(12)

8

231 It is well established that brain death results in a proinflammatory state in the donor (41). In addition, haemodynamic instability in the donor could lead to organ hypoperfusion and cellular hypoxia, promoting oxidative stress and activation of pathways of apoptosis (42). Following organ retrieval, donor kidneys are subjected to a hypoxic period and subsequent reperfusion injury when oxygenated kidney perfusion is restored after transplantation.

As a counterbalance for the ROS production, antioxidant cellular mechanisms become activated to reinstate a healthy cellular environment (43-45). In this study we have shown that increased levels of antioxidants were associated with a good kidney function suggesting that the ability to neutralise reactive oxygen species renders these DBD kidneys less susceptible to ischaemic injury.

Two distinct cellular redox systems are the thioredoxin and glutathione systems. The thioredoxin system includes thioredoxin (TRX), thioredoxin reductase (TRXR), and thioredoxin peroxidases (PRX), while detoxification by glutathione is mediated by glutathione s-transferase (GST) (46). In the DBD donors we identified several components of the redox signalling pathways that act downstream of the generation of intracellular ROS (Figure 6). Increased levels of cytoplasmic TRX1, mitochondrial TRXR2, and PRX3 in the biopsies of GO kidneys suggest either effective detoxifying processes and/or a protective role associated with improved kidney function post transplantation. TRX1 is selectively expressed in the cortex of human kidneys and localised mainly in the proximal tubules and to lesser extent in distal tubules (47). In ischaemia reperfusion (I/R) animal models TRX1 and PRX3 are both increased immediately after reperfusion (48), and overexpression of TRX1 appears to be protective following IRI in the kidney (49). GST, which was also increased in the good functioning kidneys, has mitochondrial protective properties through a process of detoxification of products of oxidative stress or lipid peroxidation (50,51).

In summary, our study has shown that a discriminative proteomic profile can distinguish donor kidneys that will have suboptimal function after kidney transplantation. Subsequently, we verified a small panel of proteins and confirmed our initial hypothesis that the balance between injury and survival mechanisms is an important determinant of allograft function. As these results only reflect that of a small cohort, validation studies will be needed to determine the subclinical diagnostic utility of a proteomic signature that could provide an indication of the extent of injury or the degree of repair that donor kidneys have undergone prior to transplantation.

(13)

232

DISCLOSURE

The authors declare no conflicts of interest.

ACKNOWLEDGEMENTS

We thank the QUOD consortium for the clinical samples analysed in this study. We thank Dr Sergei Maslau for his support on the statistical analysis and Dr Roman Fischer for his expert input on mass spectrometry. We are in debt to Dr Astrid Klooster for performing the histological assessment. This work was supported by NHS Blood and Transplant Trust Fund TF031 to M.K., a John Fell Fund 133/075 and Wellcome Trust grant 097813/Z/11/Z to B.M.K. and COPE FP7 grant award to R.J.P.

(14)

8

233

TABLES AND FIGURES

Table 1. Donor and recipient demographic and clinical characteristics

Suboptimal outcome (n=19) Good outcome (n=19) p-value Donor characteristics Age (yr)* 56±13 54±11 0.5 Gender (%) Male 10 (53) 7 (37) 0.5 Race (%) White 18 (95) 16 (84) 0.6 Other 1 (5) 3 (16) Cause of death (%) 0.6 Intracranial haemorrhage 10 (53) 10 (53)

Hypoxic brain injury 4 (21) 2 (11)

Other 5 (26) 7 (36) AKIN classification (%) 0.5 1 17 (89) 18 (95) 2 1 (5) 1 (5) 3 1 (5) 0 Remuzzi scorea _0.5 0-3 13 (68) 15 (79) 4-7 0 1 (5) ≥8 1 (5) 0 Recipient characteristics Age (yr)* 50.1±13.5 48.4±13.6 0.5 Gender (%) Male 13 (68) 13 (68) 1 Race (%) 0.17 White 15 (79) 10 (53) Other 4 (21) 9 (47) HLA mismatches (%)b _0.22 1 3 (16) 1 (5) 2 5 (26) 7 (36) 3 11 (58) 7 (36) 4 - 2 (11) CIT (h) 1 0-12 6 (32) 6 (32) 12-26 13 (68) 13 (68)

Post transplantation kidney

function (eGFR ml/min)* <0.0001

3-month 29.8±7 65.1±8

12-month 35.9±6 73±18

* indicates mean +/- SD. CIT: Cold ischaemia time. Fisher exact test was used to calculate the P-values of categorical data and t-test of numerical values.

a_{Remuzzi scoring was performed in n=30 donor kidneys, n=8 donor kidney parafin biopsies had not} enough glomeruli for a reliable assessment.

(15)

234

Figure 1. Experimental workflow

A. Overview for the kidney sample selection. Analysed biopsy samples were selected on the basis of

transplantation outcomes, and two experimental cohorts (suboptimal and good) were created on the basis of 3-month post transplantation kidney function. Selected donors offered both kidneys as single transplants. Both allografts had similar outcomes post transplantation although analysed biopsies were obtained from only one kidney. Analysed kidneys with suboptimal transplantation outcomes had at 3-month follow up mean eGFR ≤ 29.8±7 ml/min and those with good post transplantation outcomes a 3-month follow up mean eGFR ≥ 65.1±8 ml/min. B. Experimental workflow for LC-MS/MS and Westernblot.

(1) Initially n=10 individual kidney biopsies (n=5 per transplantation outcome) were analysed by label free

quantitative mass spectrometry (LC-MS/MS). Biopsy samples homogenates were precipitated, proteins were enzymatically digested to tryptic peptides and analysed by mass spectrometry. Bioinformatic analysis resulted in the identification of 1743 proteins (FDR<1%). (2) Western blot validation analysis was performed in an independent cohort of DBD kidney biopsies n=28 (n=14 biopsy samples per transplantation outcome).

(16)

8

235 Figure 2. Unsupervised hierarchical clustering analysis (HCA).

A. HCA of 1743 proteins identified by LC-MS/MS analysis of n=10 individual donor kidney biopsy samples

with either suboptimal (S) outcome or good (G) outcome post transplantation. B. Volcano plot shows the distribution of the identified proteins by LC-MS/MS according to log fold change (x-axis) and log p-values (Student’s t-test analysis) (y-axis) between suboptimal- and good transplantation outcomes. Black dotes represent the proteins with p>0.05, red and green dotes represent proteins with ≥2 fold change and p≤0.05.

(17)

236

Figure 3. Hierarchical clustering analysis and association to transplantation outcomes at 3- and 12-month post transplantation.

A. Supervised hierarchical clustering analysis (HCA) segregated individual donor kidneys in two distinct

groups according to 3-month kidney function post transplantation. B. Association between the HCA derived dendrogram of the recipient kidney function recorded at 3- and 12-month post transplantation. GF: graft failure; MV: missing value. S1, S2, S3, S4, S5: Individual donor kidneys with suboptimal outcomes. G1, G2, G3, G4, G5: Individual donor kidneys with good outcomes.

(18)

8

237 Figure 4. Altered cellular metabolic dysregulation- and response to stress pathways between the two experimental groups.

A. 214 proteins were significantly changed between the two groups (p<0.05). STRING pathway analysis

indicated that the identified proteins were enriched for regulation of cellular metabolic processes n=107 proteins (FDR: 2 10E-3). B. SINQ intensity of pro-fibrotic, and apoptotic proteins TGF-β1, Mitogen activated protein kinase-1 (MAPK-1), Signal transduction and activator of translational-1 (STAT-1), and Basigin in donor kidneys in the suboptimal outcome group (p≤0.05). Thioredoxin reductase-2 (TRX2) (p=0.03) and

(19)

238

Catalase (CAT) (p>0.05) were enriched in the good outcome group.

Figure 5. STAT-1 and PDGFRα are increased in donor kidneys with SO.

Western blot analysis of an independent cohort of n= 28 biopsy samples for Signal transduction and activator of translational-1 (STAT-1) and Platelet-derived growth factor receptor-alfa (PDGFRα) (n=14 for the suboptimal and n=14 for the good outcome cohort). Normalised by ß-actin densidometric analysis shows significant increased levels of STAT-1 and PDGFRa in the suboptimal donor kidney biopsies (p≤0.05).

(20)

8

239

Figure 6. Antioxidant proteins are enriched in the donor kidneys with GO.

Western blot analysis on an independent cohort of n=28 DBD kidney biopsies (n=14 for the suboptimal and n=14 for the good outcome cohort). Normalised for ß-actin densidometric analysis shows significant increased levels of Thioredoxin-1 (TRX1), Glutathion S-transferase (GST), and Peroxiredoxin-3 (PRX3) in good outcome donor kidney biopises (p≤0.05). Catalase (CAT) was not changed.

(21)

240

Figure 7. Transplantation outcomes depend on a balance between injury and cytoprotection in the donor kidney.

On the left, red indicates increased levels of the identified proteins associated with PDGFRa signalling in donor kidneys that can promote subclinical injury to kidney allografts and suboptimal function 3- and 12-months post transplantation. On the right, green indicates increased levels of the identified proteins of thioredoxin- and glutathione systems that promote cytoprotection and ROS elimination in donor kidneys with good transplantation outcomes.

(22)

8

241

REFERENCES

1. NHS Blood and Transplant. Organ Donation and Transplantation Activity Report 2014/15. NHS Blood and Transplant (2015).

2. Tanriover B, Mohan S, Cohen DJ, Radhakrishnan J, Nickolas TL, Stone PW, et al. Kidneys at higher risk of discard: expanding the role of dual kidney transplantation. Am J Transplant. 2014 Feb;14(2):404–15.

3. Collins MG, Chang SH, Russ GR, McDonald SP. Outcomes of transplantation using kidneys from donors meeting expanded criteria in Australia and New Zealand, 1991 to 2005. Transplantation. 2009 Apr 27;87(8):1201–9. 4. Singh RP, Farney AC, Rogers J, Zuckerman J,

Reeves-Daniel A, Hartmann E, et al. Kidney transplantation from donation after cardiac death donors: lack of impact of delayed graft function on post-transplant outcomes. Clin Transplant. 2011 Mar;25(2):255–64. 5. Moers C, Leuvenink HGD, Ploeg RJ.

Non-heart beating organ donation: overview and future perspectives. Transpl Int. 2007 Jul;20(7):567–75.

6. Metzger RA, Delmonico FL, Feng S, Port FK, Wynn JJ, Merion RM. Expanded criteria donors for kidney transplantation. Am J Transplant. 2003;3 Suppl 4:114–25.

7. Ali T, Dimassi W, Elgamal H, Alabassi A, Aleid H, Altalhi M, et al. Outcomes of kidneys utilized from deceased donors with severe acute kidney injury. QJM. 2015 Oct;108(10):803–11.

8. Dempster NJ, Ceresa CDL, Aitken E, Kingsmore D. Outcomes following renal transplantation in older people: a retrospective cohort study. BMC Geriatr. 2013 Jul 24;13:79.

9. Lamb KE, Lodhi S, Meier-Kriesche H-U. Long-term renal allograft survival in the United States: a critical reappraisal. Am J Transplant. 2011 Mar;11(3):450–62.

10. Dharnidharka VR, Lamb KE, Zheng J, Schechtman KB, Meier-Kriesche H-U. Lack of significant improvements in long-term allograft survival in pediatric solid organ transplantation: A US national registry analysis. Pediatr Transplant. 2015 Aug;19(5):477–83.

11. Bos EM, Leuvenink HGD, van Goor H, Ploeg RJ. Kidney grafts from brain dead donors: Inferior quality or opportunity for improvement? Kidney International. 2007 Jul 25;72(7):797–805.

12. Westendorp WH, Leuvenink HG, Ploeg RJ. Brain death induced renal injury. Current Opinion in Organ Transplantation. 2011 Apr;16(2):151–6.

13. Schuurs TA, Morariu AM, Ottens PJ, ‘t Hart NA, Popma SH, Leuvenink HGD, et al. Time-Dependent Changes in Donor Brain Death Related Processes. Am J Transplant. 2006 Dec;6(12):2903–11.

14. Wu WK, Famure O, Li Y, Kim SJ. Delayed graft function and the risk of acute rejection in the modern era of kidney transplantation. Kidney International. 2015 Oct;88(4):851–8. 15. https://www.nds.ox.ac.uk/research/quod. 16. Sigdel TK, Lee S, Sarwal MM. Profiling

the proteome in renal transplantation. Proteomics Clin Appl. 2011 Jun;5(5-6): 269–80.

17. Sigdel TK, Salomonis N, Nicora CD, Ryu S, He J, Dinh V, et al. The identification of novel potential injury mechanisms and candidate biomarkers in renal allograft rejection by quantitative proteomics. Mol Cell Proteomics. 2014 Feb;13(2):621–31.

18. Ricci Z, Cruz DN, Ronco C. Classification and staging of acute kidney injury: beyond the RIFLE and AKIN criteria. Nat Rev Nephrol. 2011 Apr;7(4):201–8.

19. Remuzzi G, Grinyò J, Ruggenenti P, Beatini M, Cole EH, Milford EL, et al. Early experience with dual kidney transplantation in adults using expanded donor criteria. Double Kidney Transplant Group (DKG). J Am Soc Nephrol. 1999 Dec;10(12):2591–8.

(23)

242

20. Trudgian DC, Thomas B, McGowan SJ, Kessler BM, Salek M, Acuto O. CPFP: a central proteomics facilities pipeline. Bioinformatics. 2010 Apr 15;26(8):1131–2. 21. Trudgian DC, Ridlova G, Fischer R, Mackeen

MM, Ternette N, Acuto O, et al. Comparative evaluation of label-free SINQ normalized spectral index quantitation in the central proteomics facilities pipeline. Proteomics. 2011 Jul;11(14):2790–7.

22. Caraux G, Pinloche S. PermutMatrix: a graphical environment to arrange gene expression profiles in optimal linear order. Bioinformatics. 2005 Apr 1;21(7):1280–1. 23. Mittal S, Adamusiak A, Horsfield C,

Loukopoulos I, Karydis N, Kessaris N, et al. A Re-Evaluation of Discarded Deceased Donor Kidneys in the UK: Are Usable Organs Still Being Discarded? Transplantation. 2016 Oct 25.

24. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med. 1999 Mar 16;130(6):461–70.

25. Kasiske BL, Israni AK, Snyder JJ, Skeans MA, Patient Outcomes in Renal Transplantation (PORT) Investigators. The relationship between kidney function and long-term graft survival after kidney transplant. Am J Kidney Dis. 2011 Mar;57(3):466–75. 26. Clayton PA, Lim WH, Wong G, Chadban SJ.

Relationship between eGFR Decline and Hard Outcomes after Kidney Transplants. Journal of the American Society of Nephrology. 2016 Nov;27(11):3440–6.

27. O’Connell PJ, Zhang W, Menon MC, Yi Z, Schröppel B, Gallon L, et al. Biopsy transcriptome expression profiling to identify kidney transplants at risk of chronic injury: a multicentre, prospective study. Lancet. 2016 Sep 3;388(10048):983–93.

28. Naesens M, Salvatierra O, Benfield M, Ettenger RB, Dharnidharka V, Harmon W, et al. Subclinical inflammation and chronic renal allograft injury in a randomized trial on steroid avoidance in pediatric kidney transplantation. Am J Transplant. 2012 Oct;12(10):2730–43.

29. Akhtar MZ, Huang H, Kaisar M, Faro Lo ML, Rebolledo R, Morten K, et al. Using an Integrated -Omics Approach to Identify Key Cellular Processes That Are Disturbed in the Kidney After Brain Death. Am J Transplant. 2016 May;16(5):1421–40.

30. Liu R-M, Desai LP. Reciprocal regulation of TGF-β and reactive oxygen species: A perverse cycle for fibrosis. Redox Biol. 2015 Dec;6:565–77.

31. Djamali A, Vidyasagar A, Adulla M, Hullett D, Reese S. Nox-2 is a modulator of fibrogenesis in kidney allografts. Am J Transplant. 2009 Jan;9(1):74–82.

32. Jenkins RH, Martin J, Phillips AO, Bowen T, Fraser DJ. Transforming growth factor β1 represses proximal tubular cell microRNA-192 expression through decreased hepatocyte nuclear factor DNA binding. Biochem J. 2012 Apr 15;443(2):407–16.

33. Zhong X, Chung ACK, Chen H-Y, Meng X-M, Lan HY. Smad3-mediated upregulation of miR-21 promotes renal fibrosis. Journal of the American Society of Nephrology. 2011 Sep;22(9):1668–81.

34. Boor P, Ostendorf T, Floege J. PDGF and the progression of renal disease. Nephrol Dial Transplant. 2014 Feb;29 Suppl 1:i45–i54. 35. Eitner F, Ostendorf T, Kretzler M, Cohen

CD, Eriksson U, Gröne H-J, et al. PDGF-C expression in the developing and normal adult human kidney and in glomerular diseases. J Am Soc Nephrol. 2003 May;14(5):1145–53.

36. Floege J, Eitner F, Alpers CE. A new look at platelet-derived growth factor in renal disease. Journal of the American Society of Nephrology. 2008 Jan;19(1):12–23.

(24)

8

243

37. Yamamoto H, Crow M, Cheng L, Lakatta E, Kinsella J. PDGF receptor-to-nucleus signaling of p91 (STAT1 alpha) transcription factor in rat smooth muscle cells. Experimental Cell Research. 1996 Jan 10;222(1):125–30.

38. Vignais ML, Sadowski HB, Watling D, Rogers NC, Gilman M. Platelet-derived growth factor induces phosphorylation of multiple JAK family kinases and STAT proteins. Molecular and Cellular Biology. 1996 Apr;16(4):1759–69.

39. He C, Medley SC, Hu T, Hinsdale ME, Lupu F, Virmani R, et al. PDGFRβ signalling regulates local inflammation and synergizes with hypercholesterolaemia to promote atherosclerosis. Nat Commun. 2015 Jul 17;6:7770.

40. Paukku K, Valgeirsdóttir S, Saharinen P, Bergman M, Heldin CH, Silvennoinen O. Platelet-derived growth factor (PDGF)-induced activation of signal transducer and activator of transcription (Stat) 5 is mediated by PDGF beta-receptor and is not dependent on c-src, fyn, jak1 or jak2 kinases. Biochem J. 2000 Feb 1;345 Pt 3:759–66.

41. Nijboer WN, Schuurs TA, van der Hoeven JAB, Fekken S, Wiersema-Buist J, Leuvenink HGD, et al. Effect of Brain Death on Gene Expression and Tissue Activation in Human Donor Kidneys. Transplantation. 2004 Oct;78(7):978–86.

42. Stiegler P, Sereinigg M, Puntschart A, Bradatsch A, Seifert-Held T, Wiederstein-Grasser I, et al. Oxidative stress and apoptosis in a pig model of brain death and living donation. Journal of Translational Medicine; 2013 Oct 2;11(1):1–1.

43. Chouchani ET, Pell VR, Gaude E, Aksentijević D, Sundier SY, Robb EL, et al. Ischaemic accumulation of succinate controls reperfusion injury through mitochondrial ROS. Nature. 2014 Nov 20;515(7527):431–5. 44. Gorrini C, Harris IS, Mak TW. Modulation of

oxidative stress as an anticancer strategy. Nat Rev Drug Discov. 2013 Dec;12(12):931–47.

45. Eltzschig HK, Eckle T. Ischemia and reperfusion--from mechanism to translation. Nat Med. 2011 Nov 7;17(11):1391–401. 46. Hanschmann E-M, Godoy JR, Berndt C,

Hudemann C, Lillig CH. Thioredoxins, glutaredoxins, and peroxiredoxins--molecular mechanisms and health significance: from cofactors to antioxidants to redox signaling. Antioxid Redox Signal. 2013 Nov 1;19(13):1539–605.

47. Netto LES, Antunes F. The Roles of Peroxiredoxin and Thioredoxin in Hydrogen Peroxide Sensing and in Signal Transduction. Mol Cells. 2016 Jan;39(1):65–71.

48. Kasuno K, Nakamura H, Ono T, Muso E, Yodoi J. Protective roles of thioredoxin, a redox-regulating protein, in renal ischemia/ reperfusion injury. Kidney International. 2003 Oct;64(4):1273–82.

49. Advani A, Gilbert RE, Thai K, Gow RM, Langham RG, Cox AJ, et al. Expression, localization, and function of the thioredoxin system in diabetic nephropathy. Journal of the American Society of Nephrology. 2009 Apr;20(4):730–41.

50. Suvakov S, Damjanovic T, Stefanovic A, Pekmezovic T, Savic-Radojevic A, Pljesa-Ercegovac M, et al. Glutathione S-transferase A1, M1, P1 and T1 null or low-activity genotypes are associated with enhanced oxidative damage among haemodialysis patients. Nephrol Dial Transplant. 2013 Jan;28(1):202–12.

51. Walsh J, Jenkins RE, Wong M, Olayanju A, Powell H, Copple I, et al. Identification and quantification of the basal and inducible Nrf2-dependent proteomes in mouse liver: biochemical, pharmacological and toxicological implications. J Proteomics. 2014 Aug 28;108:171–87.

(25)

244

SUPPLEMENTARY TABLE

Table S1. Significantly changed proteins identified with LC-MS/MS Accession code Description Peptide

Sequences

p-value Ratio SO/GO

P35613-2 Isoform 2 of Basigin 3 2,089E-05 1,957E+00 P50990-2 Isoform 2 of T-complex protein 1

subunit theta 22 4,406E-05 2,599E+00

Q13885 Tubulin beta-2A chain 29 7,411E-04 2,863E+00 Q695T7 Sodium-dependent neutral amino acid

transporter B(0)AT1 3 1,183E-03 4,490E-01 Q14204 Cytoplasmic dynein 1 heavy chain 1 70 1,576E-03 2,198E+00 P11234-2 Isoform 2 of Ras-related protein Ral-B 4 2,200E-03 2,045E-01 Q9UHL4 Dipeptidyl peptidase 2 5 2,850E-03 5,566E-01 P63010-2 Isoform 2 of AP-2 complex subunit

beta 17 3,337E-03 2,281E+00

P05387 60S acidic ribosomal protein P2 6 3,573E-03 3,026E+00 P14854 Cytochrome c oxidase subunit 6B1 7 3,779E-03 2,008E+00 Q9HAV7 GrpE protein homolog 1, mitochondrial 3 3,960E-03 2,220E-01 P52209-2 Isoform 2 of 6-phosphogluconate

dehydrogenase, decarboxylating 14 4,148E-03 2,142E+00 P33176 Kinesin-1 heavy chain 8 4,369E-03 1,983E+00 Q01082 Spectrin beta chain, non-erythrocytic 1 93 4,623E-03 1,208E+00

P07339 Cathepsin D 17 4,773E-03 1,831E+00

Q15124-2 Isoform 2 of Phosphoglucomutase-like

protein 5 2 4,783E-03 2,941E-01

P05109 Protein S100-A8 4 4,840E-03 4,648E+00

P01860 Ig gamma-3 chain C region 12 4,858E-03 1,868E+00 P08237-3 Isoform 3 of ATP-dependent

6-phosphofructokinase, muscle type 10 5,034E-03 2,887E+00 Q9Y2R0 Cytochrome c oxidase assembly factor

3 homolog, mitochondrial 3 5,088E-03 2,502E+00 P51659 Peroxisomal multifunctional enzyme

type 2 18 5,219E-03 2,169E+00

Q13423 NAD(P) transhydrogenase,

mitochondrial 41 5,395E-03 1,513E+00

Q93077 Histone H2A type 1-C 8 5,784E-03 4,156E-01 P48047 ATP synthase subunit O 16 6,648E-03 6,082E-01 Q99832-3 Isoform 3 of T-complex protein 1

subunit eta 17 7,202E-03 2,008E+00

O14818-2 Isoform 2 of Proteasome subunit alpha

type-7 7 7,481E-03 2,327E+00

P07237 Protein disulfide-isomerase 30 7,859E-03 1,546E+00 P26641 Elongation factor 1-gamma 18 8,397E-03 1,643E+00 Q16563-2 Isoform 2 of Synaptophysin-like

protein 1 2 8,558E-03 5,067E-01

P05413 Fatty acid-binding protein, heart 9 8,677E-03 2,675E+00 P16615 Sarcoplasmic/endoplasmic reticulum

calcium ATPase 2 12 9,042E-03 2,614E+00 Q15181 Inorganic pyrophosphatase 8 1,048E-02 3,678E+00

(26)

8

245

Q96HC4 PDZ and LIM domain protein 5 12 1,097E-02 2,313E+00 P46781 40S ribosomal protein S9 11 1,107E-02 1,551E+00 Q96I99 Succinyl-CoA ligase [GDP-forming]

subunit beta, mitochondrial 35 1,148E-02 1,773E+00 O00151 PDZ and LIM domain protein 1 2 1,204E-02 3,681E-01 P07814 Bifunctional glutamate/proline--tRNA

ligase 26 1,208E-02 2,393E+00

P80294 Metallothionein-1H 4 1,322E-02 4,295E-01

Q16891-2 Isoform 2 of MICOS complex subunit

MIC60 33 1,379E-02 1,385E+00

P40227 T-complex protein 1 subunit zeta 15 1,381E-02 1,716E+00 P62195-2 Isoform 2 of 26S protease regulatory

subunit 8 5 1,386E-02 1,465E+00

P62851 40S ribosomal protein S25 7 1,645E-02 1,841E+00 Q7KZF4 Staphylococcal nuclease

domain-containing protein 1 14 1,699E-02 2,031E+00 P62258 14-3-3 protein epsilon 22 1,726E-02 1,577E+00 Q14764 Major vault protein 30 1,749E-02 2,726E+00 Q07075 Glutamyl aminopeptidase 10 1,850E-02 2,604E+00 P39023 60S ribosomal protein L3 10 1,894E-02 2,045E+00 P24311 Cytochrome c oxidase subunit 7B,

P31946-2 Isoform Short of 14-3-3 protein beta/

alpha 17 1,966E-02 2,814E+00

Q9UNF0-2 Isoform 2 of Protein kinase C and casein kinase substrate in neurons protein 2

5 1,997E-02 2,440E+00 Q8NFV4-4 Isoform 4 of Alpha/beta hydrolase

domain-containing protein 11 2 2,014E-02 2,060E+00 O94903 Proline synthase co-transcribed

bacterial homolog protein 3 2,064E-02 5,778E-01 P18859-2 Isoform 2 of ATP synthase-coupling

factor 6, mitochondrial 3 2,105E-02 2,257E+00

P37837 Transaldolase 17 2,120E-02 1,536E+00

P14927 Cytochrome b-c1 complex subunit 7 10 2,128E-02 6,689E-01 P07919 Cytochrome b-c1 complex subunit 6,

Q86XE5 4-hydroxy-2-oxoglutarate aldolase,

mitochondrial 14 2,431E-02 6,820E-01

Q16527 Cysteine and glycine-rich protein 2 9 2,484E-02 2,090E+00 P39019 40S ribosomal protein S19 8 2,515E-02 1,501E+00 P09497-2 Isoform Non-brain of Clathrin light

chain B 7 2,521E-02 2,074E+00

Q6P1X6-2 Isoform 2 of UPF0598 protein C8orf82 4 2,532E-02 4,529E-01 Q9UI09 NADH dehydrogenase [ubiquinone] 1

alpha subcomplex subunit 12 6 2,537E-02 2,188E+00

Q9HBL0 Tensin-1 21 2,559E-02 1,762E+00

P02766 Transthyretin 8 2,581E-02 1,906E-01

P22695 Cytochrome b-c1 complex subunit 2,

(27)

246

Q9NRX4-2 Isoform 2 of 14 kDa phosphohistiine

phosphatase 5 2,634E-02 7,431E-02

Q96HY6-2 Isoform 2 of DDRGK

domain-containing protein 1 3 2,704E-02 1,561E+00 P50502 Hsc70-interacting protein 9 2,791E-02 1,704E+00 P62857 40S ribosomal protein S28 3 2,832E-02 1,767E+00 Q9HB40-2 Isoform 2 of Retinoid-inducible serine

carboxypeptidase 4 3,023E-02 5,546E-01

P46777 60S ribosomal protein L5 9 3,030E-02 1,854E+00 P14868-2 Isoform 2 of Aspartate--tRNA ligase,

cytoplasmic 8 3,043E-02 2,294E+00

P62917 60S ribosomal protein L8 3 3,191E-02 1,849E+00 Q07955 Serine/arginine-rich splicing factor 1 7 3,198E-02 2,358E+00 Q14247-3 Isoform 3 of Src substrate cortactin 20 3,240E-02 2,132E+00 P27695 DNA-(apurinic or apyrimidinic site)

lyase 8 3,243E-02 1,616E+00

Q9NPJ3 Acyl-coenzyme A thioesterase 13 4 3,357E-02 3,780E-01 Q01813 ATP-dependent

6-phosphofructokinase, platelet type 9 3,449E-02 2,875E+00 P12277 Creatine kinase B-type 21 3,474E-02 2,084E+00 P28482 Mitogen-activated protein kinase 1 6 3,548E-02 2,113E+00 P62269 40S ribosomal protein S18 8 3,675E-02 1,765E+00 Q99436 Proteasome subunit beta type-7 7 3,677E-02 2,441E+00 Q15019-2 Isoform 2 of Septin-2 10 3,691E-02 1,819E+00 Q9UQ80 Proliferation-associated protein 2G4 11 3,729E-02 4,766E-01 P14866 Heterogeneous nuclear

ribonucleoprotein L 12 3,754E-02 1,660E+00 P52907 F-actin-capping protein subunit alpha-1 8 3,788E-02 7,236E-01 Q9NZ45 CDGSH iron-sulfur domain-containing

protein 1 3 3,792E-02 2,458E-01

O15143 Actin-related protein 2/3 complex

subunit 1B 5 3,853E-02 4,706E-01

Q01844-2 Isoform EWS-B of RNA-binding protein

EWS 4 3,866E-02 2,345E+00

P27824-2 Isoform 2 of Calnexin 23 3,932E-02 1,810E+00 Q16851 UTP--glucose-1-phosphate

uridylyltransferase 17 4,001E-02 2,281E+00 P62424 60S ribosomal protein L7a 4 4,116E-02 5,700E-01 P55786 Puromycin-sensitive aminopeptidase 17 4,203E-02 1,472E+00 Q92896-2 Isoform 2 of Golgi apparatus protein 1 4 4,248E-02 3,103E+00 Q07157-2 Isoform Short of Tight junction protein

ZO-1 8 4,285E-02 3,976E+00

P35221 Catenin alpha-1 24 4,334E-02 1,729E+00

P17987 T-complex protein 1 subunit alpha 17 4,420E-02 1,783E+00 O75306 NADH dehydrogenase [ubiquinone]

iron-sulfur protein 2, mitochondrial 11 4,436E-02 1,906E+00 P51580 Thiopurine S-methyltransferase 9 4,527E-02 1,980E+00 O43294 Transforming growth factor

beta-1-induced transcript 1 protein 6 4,531E-02 7,416E+00

(28)

8

247

Q16851-2 Isoform 2 of UTP--glucose-1-phosphate

uridylyltransferase 16 4,550E-02 2,246E+00 P04792 Heat shock protein beta-1 17 4,612E-02 1,570E+00 P21397 Amine oxidase [flavin-containing] A 16 4,943E-02 2,083E+00 O95428-4 Isoform 4 of Papilin 2 4,961E-02 4,529E-01 P68366-2 Isoform 2 of Tubulin alpha-4A chain 26 4,964E-02 7,579E-01

P27216-2 Isoform B of Annexin A13 4 - GO only

P05093 Steroid 17-alpha-hydroxylase/17,20

lyase 3 - 0,04856798

P53041 Serine/threonine-protein phosphatase

5 3 - SO only

P18827 Syndecan-1 3 - 4,859583333

O75340 Programmed cell death protein 6 4 - 0,763578402 Q07092-2 Isoform 2 of Collagen alpha-1(XVI)

chain 2 - 1,683993723

Q5T440 Putative transferase CAF17,

mitochondrial 2 - 0,540656907

Q8TCC7-2 Isoform 2 of Solute carrier family 22

member 8 3 - 3,332736309

Q9NRZ7-2 Isoform 2 of

1-acyl-sn-glycerol-3-phosphate acyltransferase gamma 3 - 0,428849685 O95782-2 Isoform B of AP-2 complex subunit

alpha-1 8 - 1,646076146

P08134 Rho-related GTP-binding protein RhoC 9 - 0,552189127 O14874-2 Isoform 2 of [3-methyl-2-oxobutanoate

dehydrogenase [lipoamide]] kinase, mitochondrial

3 - 0,892437755

P21964-2 Isoform Soluble of Catechol

O-methyltransferase 6 - 5,089639481

O14976-2 Isoform 2 of Cyclin-G-associated kinase 2 - 2,53432282

P10301 Ras-related protein R-Ras 5 - 0,671553122

P22234-2 Isoform 2 of Multifunctional protein

ADE2 5 - 0,811649485

P15531-2 Isoform 2 of Nucleoside diphosphate

kinase A 6 - 0,903039565

P62316 Small nuclear ribonucleoprotein Sm D2 3 - SO only P62714 Serine/threonine-protein phosphatase

2A catalytic subunit beta isoform 7 - 0,859516189

Q16543 Hsp90 co-chaperone Cdc37 8 - 0,953049229

P0CG39 POTE ankyrin domain family member J 9 - SO only P01903 HLA class II histocompatibility antigen,

DR alpha chain 4 - 1,172035059

P62854 40S ribosomal protein S26 2 - 1,181710379

Q53SF7-2 Isoform 2 of Cordon-bleu protein-like 1 5 - SO only P26885 Peptidyl-prolyl cis-trans isomerase

FKBP2 4 - 0,257124457

P28074 Proteasome subunit beta type-5 4 - 0,914455882 Q8N465 D-2-hydroxyglutarate dehydrogenase,

(29)

248

Q96EY1-2 Isoform 2 of DnaJ homolog subfamily

A member 3, mitochondrial 3 - 1,857178035

P35580-3 Isoform 3 of Myosin-10 87 - 1,860164679

Q7Z406-2 Isoform 2 of Myosin-14 17 - 0,866672198

P36222 Chitinase-3-like protein 1 3 - GO only

Q07812-2 Isoform Beta of Apoptosis regulator

BAX 2 - 1,036157802

Q12792-4 Isoform 4 of Twinfilin-1 4 - 1,076899321 P42224 Signal transducer and activator of

transcription 1-alpha/beta 7 - 6,940175953 Q9NX40 OCIA domain-containing protein 1 4 - SO only Q9NY33-4 Isoform 4 of Dipeptidyl peptidase 3 4 - SO only P43034 Platelet-activating factor

acetylhydrolase IB subunit alpha 5 - 0,455978906

Q9UGB7 Inositol oxygenase 4 - 0,235509138

P49327 Fatty acid synthase 7 - 2,827369888

Q14108 Lysosome membrane protein 2 4 - 0,629382918 Q9Y241 HIG1 domain family member 1A,

mitochondrial 3 - 0,984621736

P50552 Vasodilator-stimulated

phosphoprotein 5 - 2,44707808

Q15436 Protein transport protein Sec23A 7 - 1,859601755 Q15437 Protein transport protein Sec23B 5 - 1,633090024 Q9Y2W1 Thyroid hormone receptor-associated

protein 3 3 - SO only

Q86Y82 Syntaxin-12 2 - 1,033402848

Q7Z434-4 Isoform 4 of Mitochondrial

antiviral-signaling protein 2 - 0,915445321

Q969H8 UPF0556 protein C19orf10 2 - 0,929933212

Q53T59 HCLS1-binding protein 3 2 - SO only

P12268 Inosine-5’-monophosphate

dehydrogenase 2 2 - 0,977966988

Q04637-3 Isoform B of Eukaryotic translation

initiation factor 4 gamma 1 4 - 2,540892696 Q10589-2 Isoform 2 of Bone marrow stromal

antigen 2 2 - 0,438320087

O60826 Coiled-coil domain-containing protein

22 4 - SO only

Q9UHJ6 Sedoheptulokinase 2 - SO only

Q53H12 Acylglycerol kinase, mitochondrial 2 - 0,860362694 Q13177 Serine/threonine-protein kinase PAK 2 2 - 0,990568107

O75954 Tetraspanin-9 2 - SO only

Q99569-2 Isoform 2 of Plakophilin-4 3 - 0,942591071 Q9Y3Z3-2 Isoform 2 of Deoxynucleoside

triphosphate triphosphohydrolase SAMHD1

2 - 1,60239952

O14639-2 Isoform 2 of Actin-binding LIM protein

1 2 - 1,796940538

A3KMH1-3 Isoform 3 of von Willebrand factor A

domain-containing protein 8 6 - 0,601438849

(30)

8

249

P15924-3 Isoform DSPIa of Desmoplakin 7 - 0,6937321 P24666 Low molecular weight

phosphotyrosine protein phosphatase 2 - 0,282199128 P40616-2 Isoform 2 of ADP-ribosylation

factor-like protein 1 3 - 2,963466298

Q13435 Splicing factor 3B subunit 2 2 - 1,111253197 Q495M3 Proton-coupled amino acid

transporter 2 3 - 1,236696023

P20042 Eukaryotic translation initiation factor

2 subunit 2 2 - 1,572728607

O00203-3 Isoform 2 of AP-3 complex subunit

beta-1 3 - SO only

P26196 Probable ATP-dependent RNA helicase

DDX6 2 - SO only

P41250 Glycine--tRNA ligase 5 - 1,276476423

Q9NZB2-4 Isoform D of Constitutive coactivator

of PPAR-gamma-like protein 1 4 - 0,461871891

P0DJI8 Serum amyloid A-1 protein 7 - 2,105586249

P67812-2 Isoform 2 of Signal peptidase complex

catalytic subunit SEC11A 2 - SO only

O43681 ATPase ASNA1 3 - GO only

P36405 ADP-ribosylation factor-like protein 3 2 - 0,942880259

P42677 40S ribosomal protein S27 2 - 17,75541126

P47897-2 Isoform 2 of Glutamine--tRNA ligase 5 - 0,145503408 Q96EY8 Cob(I)yrinic acid a,c-diamide

adenosyltransferase, mitochondrial 2 - 1,478656331

O43813 LanC-like protein 1 2 - 0,905555556

Q15102 Platelet-activating factor

acetylhydrolase IB subunit gamma 2 - 1,298580442

Q15393 Splicing factor 3B subunit 3 5 - SO only

P53420 Collagen alpha-4(IV) chain 3 - 2,236086702 P06730-2 Isoform 2 of Eukaryotic translation

initiation factor 4E 3 - 0,938811189

Q9Y2V2 Calcium-regulated heat stable protein

1 2 - 1,218217562

Q00688 Peptidyl-prolyl cis-trans isomerase

FKBP3 2 - 6,190016639

P12004 Proliferating cell nuclear antigen 2 - 1,770178166 Q6UXV4 MICOS complex subunit MIC27 4 - 1,564849161 O00442-2 Isoform 2 of RNA 3’-terminal

phosphate cyclase 3 - 1,068251785

Q16531 DNA damage-binding protein 1 4 - 0,462925345 Q8NC56 LEM domain-containing protein 2 3 - 1,052026083 O43491-3 Isoform 3 of Band 4.1-like protein 2 8 - 1,016841436 Q9NTJ5-2 Isoform 2 of Phosphatidylinositide

phosphatase SAC1 2 - 3,2262309

P56192-2 Isoform 2 of Methionine--tRNA ligase,

cytoplasmic 2 - 1,513146168

P49914-2 Isoform 2 of 5-formyltetrahydrofolate

(31)

250

P34949-2 Isoform 2 of Mannose-6-phosphate

isomerase 2 - 1,079799919

Q86VB7-2 Isoform 2 of Scavenger receptor

cysteine-rich type 1 protein M130 3 - SO only O43615 Mitochondrial import inner membrane

translocase subunit TIM44 3 - 0,916211878

O75915 PRA1 family protein 3 2 - 0,878698068

Q9H078-2 Isoform 2 of Caseinolytic peptidase B

protein homolog 2 - 1,128007974

Q15637-2 Isoform 2 of Splicing factor 1 2 - 0,449719218 Q8N142 Adenylosuccinate synthetase isozyme

1 2 - 0,678067079

P10644-2 Isoform 2 of cAMP-dependent protein

kinase type I-alpha regulatory subunit 2 - SO only

Q03135 Caveolin-1 3 - 0,685502392

O43684-2 Isoform 2 of Mitotic checkpoint

protein BUB3 2 - 1,607132906

O14828-2 Isoform 2 of Secretory

carrier-associated membrane protein 3 2 - SO only Q9UGT4 Sushi domain-containing protein 2 3 - 0,86122449 Q13724-2 Isoform 2 of Mannosyl-oligosaccharide

glucosidase 2 - 1,113869188

Q16625-2 Isoform 2 of Occludin 2 - SO only

Q9Y2A7-2 Isoform 2 of Nck-associated protein 1 4 - 0,421976967

Q9UBV8 Peflin 4 - 3,91780174

P36507 Dual specificity mitogen-activated

protein kinase kinase 2 2 - 2,125073746

SO: Suboptimal outcome group; GO: good optimal outcome group.

(32)

8

(33)