2-D hydrophilic interaction liquid chromatography-RP separation in urinary proteomics--minimizing variability through improved downstream workflow compatibility.
Summary (9 min read)
1 Introduction
- Solid organ transplantation is a unique treatment option for organ failure where the failing organ function is replaced by organs obtained from either a living or deceased donor.
- Most of the transplants are performed between genetically non-identical individuals, where the immune response of the recipient against the foreign graft is one of the principal obstacles to a successful transplantation.
- Acute rejection (AR), which is subcategory, predominately appears the first 3 months posttransplant, but can also emerge after several years.
- The protein concentration in urine of healthy subjects is low (less than 100 mg/L) compared to other body fluids.
- Urinary proteomics seems very promising in the search for biomarkers and is a rapid growing field [1].
1.1 Kidney transplantation
- 1.1 Kidney transplantation in general and the status in Norway Renal transplantation is the ultimate renal replacement therapy (RRT) for most patients with end-stage kidney disease [3].
- The last few years there has been a slight increase in number of transplantations in Norway and in 2009 a total of 292 renal transplants were performed at Rikshospitalet, which was a new all-time high [4].
- The mean age of the recipients from living donors were 46.9 years (range 1-78) while for those receiving from deceased donors the mean age was 57 years (range 14-80).
- The observed two-year patient survival was 84 % for patients transplanted in the period of 2000-2004 while the five-year survival was approximately 70 % for the same group [5].
- The basis immunosuppressive protocol at the hospital has since 2007 been quadruple treatment.
Cause
- Acute rejection is a serious and relative frequent complication after renal transplantation affecting long-term graft outcome.
- The allograft rejection is caused by several elements of the immune system including antibody, complement, T-cells and other cell types [6].
- Mechanisms believed to be responsible are thoroughly reviewed by Cornell et al. [7], see Figure 1 for cells and mediators involved.
- T-cell-mediated acute rejection is characterized by accumulation of mononuclear cells (mostly T-cells) in the interstitium, accompanied by inflammation of tubules and sometimes arteries.
- The pathology has however a wide spectrum and could also include a component of acute cellular rejection.
Diagnosis
- Examination of immunological activity by histological analysis in renal biopsies is currently the gold standard for diagnosis of acute rejection episodes.
- This is carried out on suspicion of acute rejection, often made on basis of clinical symptoms of impaired renal function (elevated plasma creatinine levels).
- One of the challenges is that CsA and tacrolimus can give the same symptoms, but then as a result of high dosage.
- The biopsies are classified according to the Banff criteria, which is a standardization of renal allograft biopsy interpretation based on international consensus.
- Antibody-mediated rejection type is identified by positive C4d staining in addition to other criteria [8].
Effect on outcome
- The event of AR in renal transplants increases the risk of developing chronic allograft nephropathy and is also associated with reduced long-term survival [10-14].
- Several factors including the timing and severity of the acute rejection episode and the post rejection recovery of renal function affects the chronic allograft injury [15-17].
- Antibodymediated rejections generally has worse prognosis and demands a different form of therapy than the usual T-cell-mediated rejection [8].
1.2 Proteomics
- The proteome can be described as the protein complement of the expressed genome, including protein modifications occurring during and after translation [18].
- Proteomics is the study of protein properties like expression levels, post-translational modifications, interactions etc. on a large scale to obtain a view of disease processes, cellular processes and networks at the protein level [19].
- Detection of proteins using mass spectrometry (MS) can either be done by a top-down approach where intact proteins are analyzed or by a bottom-up approach where proteins are digested into smaller peptides prior to analysis.
- A bottom-up approach has been applied; the principal workflow is presented in Figure 2.
- The main steps, presented in the following sections, are sample preparation to isolate the proteins of interest from the matrix followed by digestion of the proteins into peptides using a specific protease with known digestion pattern.
Protein identification Quantification
- Since the protein concentration in urine is relatively low, effective protein enrichment is advantageous in the sample preparation.
- 5 kDa cut-off centrifugation and protein precipitation using ethanol and trichloroacetic acid (TCA) have been tested.
- Thus, depletion of proteins has become a standard approach for in-depth analysis of the proteome.
- This results in unfolding of the proteins to make the cleavage sites more accessible to trypsin, yielding a more efficient digestion.
Separation by liquid chromatography
- Separation of proteins has in proteomics routinely been done using two-dimensional gel electrophoresis (2-DE) followed by in-gel digestion prior to MS [2,38].
- The challenge with this approach is the massively increased sample complexity due to all the peptides originating from a single protein after digestion.
- The practical achievable peak capacity will however be limited by the orthogonality of the system, which means that if the two dimension of separation are not completely orthogonal , the achievable peak capacity is lower than theoretically expected.
- Both ion exchange and electrostatics are weak compared to other HILIC phases and the main influence of retention is partitioning between the mobile phase and the adsorbed water layer for the ZIC-HILIC column.
- Combination of HILIC and RP has shown to give a higher orthogonality and peak capacity compared with alternatives like SCX-RP and size exclusion chromatography (SEC)-RP [49].
Ionization and MS detection of peptides
- A requirement for peptide detection in a mass spectrometer is that the molecule is ionized before entering the mass analyzer.
- MALDI-TOF-MS was not used for this work and will thus not be described further.
- The process is essentially the same as with regular ESI but because of the low flow rate, droplet formation occurs more readily requiring only applied voltage to generate spray.
- Different types of mass analyzers were used in this work including ion trap, TOF, single quadrupole and linear ion trap-Orbitrap .
- Analysis and interpretation of these ions (in addition to several other ions produced by fragmentation) are then used to elucidate the amino acid sequence of the peptide.
The Orbitrap mass analyzer
- Several different types of mass analyzers were used in this work; most of them are established and have been used routinely for years.
- Recently this was combined with a linear ion trap combining the mass spectrometric features of the ion trap with the high resolution and mass accuracy of the Orbitrap which resulted in the hybrid instrument named LTQ Orbitrap.
- Another challenge with 18O-labeling is back exchange to 16O when labeled samples are mixed with unlabeled samples before LC-MS analysis, a reaction which is likely as long as trypsin is present [71].
- Corresponding mass values are then scored in a way that allows for identification of the peptides and the proteins that best matches the peptide composition in the sample.
- The database is highly annotated including detailed information regarding protein structure, functions etc. and is updated at a regular basis.
2 Aim of the study
- The current gold standard for diagnosis of suspected acute rejection episodes in kidney transplants is done by histological examination of renal core biopsies.
- Analysis of urine is particularly useful as biomarker matrix since it contains both proteins originating from plasma as well as locally in the kidney.
- Downscale analysis to nanoscale separation (nanoLC-MS/MS) to increase sensitivity.
- Analyze urine samples from kidney transplants experiencing acute rejections to identify associated proteins.
3 Results and discussion
- The methodology in bottom-up proteomics is complex, time demanding, labor intensive and there are several possible pitfalls.
- In this thesis the focus has been on developing a urinary proteomics method to be able to find differentially expressed proteins associated with acute rejection episodes in kidney transplants.
- The first three papers have been focused around the sample preparation, tryptic digestion and the chromatographic separation.
- In Paper IV, a quantification method was modified and implemented in the workflow before the complete method was utilized in the patient study (Paper V).
- In addition, much effort has been put on developing a more time efficient methodology than current standard protocols.
3.1 Sample preparation and separation in urinary proteomics
- A bottom-up proteomics experiment is a complex multi-step procedure typically including sample preparation, depletion and multidimensional separation followed by MS-detection.
- In addition, the chemicals used in each step are not always compatible with the next step making extra sample handling necessary.
- Simplification and streamlining was one of the main principles laid to ground in the method development in order to decrease variability and increase repeatability and time efficiency.
- For the study in Paper V, urine was collected as part of an at that time ongoing study at Oslo University Hospital (n=20) [83].
- Patient samples used in Paper I and IV was collected from anonymous kidney transplant patients in a stable phase post-transplant.
Choice of method
- For a successful urinary proteome analysis, isolation and purification of the proteins is necessary.
- In Paper I, several sample preparation approaches were tested.
- Criteria for 21 evaluation were high protein recovery, possibilities for enrichment and effective desalting of the sample.
- Precipitation using 10 % TCA had no effect in urine and only low protein recoveries (<20 %) were obtained.
- The use of centrifugal filtration was also shown to be an effective desalting step, which is important for the further analysis of urine.
Optimization
- In order to improve the recovery and thus covering a larger part of the proteome, different solutions with increasing volumes (600 µL – 2400 µL) were added to the remaining volume over the 5 kDa-filter of the device.
- Figure 10 shows that for all solvents tested the recovery increased with increasing volumes up to approximately 1800 µL.
- The highest recovery was obtained using 8 M urea, but this was considered to be unsuitable due to the downstream incompatibility with HSA depletion and the requirement of an extra step to remove excess urea.
- High recoveries were achieved with 10 mM TrisHCl/150 mM NaCl 22 (pH 7.4), and more important, this solution was downstream compatible with the HSA depletion step that made it a better choice than urea.
HSA depletion
- Hence, the dynamic range is reduced and depletion of only HSA has been reported to be sufficient to be able to identify low abundant proteins in urine [28].
- A combination of this and the risk of information loss after depletion (see chapter 1.2.1) lead to the choice of depleting only HSA.
- 23 3.1.3 Chromatographic separation of the peptides A proteolytically digested protein sample usually yields highly complex peptide mixtures where the separation power offered by standard RP columns is far from sufficient to obtain quality data from the MS analysis.
- In order to improve this, the introduction of multiple chromatographic separations is often done to achieve a higher separation power and increased amount of information obtained.
- The protein recoveries were improved from 42 % to 76 % and 0.1 % formic acid was thus chosen for the further work.
First-dimension separation: ZIC-HILIC
- In the development phase both 80 % MeCN and 95 % MeCN were investigated as starting conditions for the gradient elution and also sample solvent for the respective setups.
- Chromatograms separating a cyt c digest in both gradients are displayed in Figure 12, which shows a significant difference not only in peak height of the peptides, but also in total number of peaks detectable.
- This may be due to decreased solubility of the peptides in the 95 % MeCN mobile phase, which particularly affects the hydrophilic peptides.
- As a consequence only the most hydrophobic peptides may be solubilized resulting in lower peak heights and fewer peaks.
- 25 Hence, it was decided that 80 % MeCN was used as gradient starting mobile phase and as sample solvent.
Second-dimension separation: Reversed Phase
- Reversed phase (RP) chromatography was used as the separation technique in the second dimension when multidimensional separation was applied.
- Furthermore are mounting of columns and changing of other parts of the flow-line critical operations 26 where small details like for example an inadequate tightened coupling can lead to large changes of the chromatography.
- The fraction number from the ZIC-HILIC separation was plotted against the retention time of the peaks in the 2nd dimension (RP) shown in Figure 14.
- In order to evaluate the variability of the current method, 6 replicates of a pooled urine sample from 3 renal transplant recipients were analyzed.
- Both variations in protein recovery and retention times in the first dimension affect the signal variability in the last step in addition to sources directly related to that step (e.g. electrospray ionization).
3.2 Tryptic digestion & protein identification
- 2.1 Optimization of digestion conditions using immobilized trypsin beads Tryptic digestion of proteins has traditionally been carried out in-solution [85,86], which also was the case in Paper I.
- As a strategy to reduce the total time frame of the workflow, enzymatic digestion using immobilized trypsin was tested as a replacement for in-solution digestion (Paper II).
- Similar intensities were found for many peptides, however, both higher and lower intensities of several peptides were observed after digestion in urine compared to in buffered solution.
- The system consisted of pH gradient SAX chromatography of native proteins in the first dimension which then were fractionated and stored on trap columns (C4-C4) for subsequent on-column reduction and alkylation.
- In the method development, different LC-MS/(MS) equipment were utilized.
Paper LC MS ID Peptides Proteins
- Peptides/proteins is probably lower than compared with results obtained using ion trap MS/MS.
- This is related to the high mass accuracy of the Orbitrap, which reduces the number of possible peptide hits from a certain m/z-value considerably.
- In Paper V, the identified proteins were also validated by searching against the reversed database in order to eliminate false positive identifications.
- This is probably strongly correlated to the use of PMF, which has considerable limitations in complex protein samples.
- If MS/MS had been used for identification, the list of identified proteins would probably be larger.
3.3 Accelerated quantification in urinary proteomics utilizing 18O-
- As described in section 1.2.4; 18O-labeling was chosen as the preferred quantification strategy for the current work.
- Focus of the experiments was not to study each reaction in detail but a more practical approach, optimizing a method best fit for the application.
- To investigate this further, an experiment assessing both pH and reaction time was carried out to identify the conditions where complete labeling could be achieved in the shortest possible time.
- The ratio of the lysine terminated peptides increased at a much slower rate than arginine terminated peptides.
- The increased digestion efficiency observed after implementation of immobilized trypsin beads (section 3.2.1) combined with the fact that both the digestion and labeling steps were carried out using trypsin beads, lead to the idea of a closer integration of the two steps.
3.4 Differential expressed proteins following acute rejection in renal
- To improve current methods, a specific and more sensitive biomarker that could be obtained non-invasively and detect initiating rejection episodes at an earlier time would have been of great value.
- In order to find proteins associated with AR, urine samples from the day a biopsy was taken to investigate if there was a true acute rejection episode were compared with the first urine sample available after transplantation by the proteomic 43 method described and developed in this thesis.
- In particular for the highly down-regulated proteins where the isotope pattern from the unlabeled peptide peak would 44 interfere with the low signals from the labeled peptide giving uncertain quantification results.
- 48 All patients in the rejection group were regulated above the predefined threshold in at least one of the proteins groups of which no significant regulation was observed in the control group.
- SERPINF1 did not show any clear regulation pattern while CD44 was increased during AR in 4 patients, but only in 1 of the patients was the up-regulation more than two-fold.
3.5 Future perspectives
- The developed method has shown proof of concept in biomarker discovery of the present setup by identifying several urinary proteins associated with acute rejection episodes in kidney transplants in this pilot study.
- Another technique which could be used is Multiple Reaction Monitoring (MRM) of unique peptide products from the proteins of interest.
- Both of these techniques can be carried out without the extensive sample preparation and without the need 52 of multidimensional chromatography, reducing both the workload and time used for each sample.
- In order to remove doubt over possible sub-clinical rejections not discovered in the norejection group, the control group should preferably consist of kidney transplanted that are confirmed non-rejectors.
- Such information could be very valuable from a clinical point of view.
4 Concluding remarks
- In the presented work a proteomic method has been developed and optimized in order to analyze urine from kidney transplanted patients.
- This could make the proteins useful in a clinical setting enabling earlier recognition of acute rejection episodes in a non-invasive manner.
- The result was a method with a minimal amount of sample handling between each step to eliminate possible sources of variability.
- Several approaches to tryptic digestion of proteins were also tested, introducing immobilized trypsin and different technical solutions.
- A complete on-line method including all steps in a proteomic workflow was also evaluated.
Introduction
- Patients whom experience an acute rejection (AR) after renal transplantation have reduced long-term graft survival and an increased risk of developing chronic allograft nephropathy (1- 5).
- It suffers from sampling heterogeneity and correlates poorly with treatment response and prognosis.
- The authors performed a small prospective pilot study in order to try to identify urinary proteins associated with AR episodes in the early phase following kidney transplantation.
- In shotgun proteomics proteins are enzymatically digested into peptides which are separated by liquid chromatography coupled to a mass spectrometer.
- Analysis of individual samples gives information on inter-individual variation.
Study design and samples
- The authors used urine samples from 6 renal transplant patients with biopsy proven acute rejection (BPAR) and from 6 renal transplant patients with stable graft function, matched for age, immunosuppression and time after transplantation.
- All urine samples were collected prospectively as part of an at that time ongoing study at Oslo University Hospital (n=20) (24).
- On average urine samples were available from 4.7±2.7 days after transplantation and the patients were followed for 8-10 weeks.
- All patients’ received induction with i.v. basiliximab on day 0 and 4, cyclosporine A (CsA), mycophenolate mofetil 1 g BID, steroids, sulfacotrimoxacole and proton pump inhibitor.
- Acute rejections were suspected in patients based on an increased plasma creatinine of 20%, without other plausible causes, and were verified with a renal core biopsy according to the Banff 97 criteria (6).
Urine sample preparation
- Further sample preparation of urine was performed as previously described (25).
- 5 mL of stored urine was centrifuged at 9000 × g for 10 minutes and applied to Vivaspin 5 kDa cut-off centrifugal filter (Vivascience Sartorius Group, Stonehouse, UK) for desalting and up-concentration of urinary proteins, followed by washing and reconstitution (1200 µL) using 10 mM TrisHCl/150 mM NaCl (pH 7.4), also known as In brief.
- Total protein concentrations in each sample was measured using Bradford’s method (26) and the samples from each patient was normalized with respect to total protein content.
- The key parameters were as follows: Subsequently, the samples were subjected to 18O/16O-labeling using the same beads, but with a different buffer (pH 6.0) at 37 °C for 3 hours under shaking (1200 rpm).
- The AR samples were labeled with 18O and mixed with both unlabeled baseline samples and unlabeled stable samples (7-11 days prior to rejection) in the AR-group.
2D LC-MS/MS
- Two-dimensional LC-MS/MS was used for separation and detection of the tryptic digested peptide mixture.
- Hydrophilic Interaction Liquid Chromatography was used as the first dimension of separation and was done exactly as described previously (25, 27).
- Fractions were collected every minute, in total 30 fractions per sample.
- Subsequently, the elution strength was increased to 100 %.
- Data dependent MS/MS with wide band activation carried out on the highest m/z value for a maximum of one spectrum in the linear ion-trap, also known as Scan event 2.
Identification and selection of proteins
- The acquired mass spectrometric data were analyzed and processed using Proteome Discoverer 1.2 software.
- Carboxymethyl (C) was set as constant modification while oxidation (M) and 18O (2) on the C-terminal were chosen as variable modifications.
- Grouping of proteins were enabled and only the top ranked peptide hits below the FDR threshold (< 0.05) were accepted.
- In the second search node precursor ions were detected for quantification.
- For all protein and protein groups: up-regulation was defined as a fold change of ≥1 (log 2) in protein level observed between baseline and AR.
Statistics
- For the evaluation of the demographic data and comparison of the groups, the Mann-Whitney U test was used.
- A P-value of <0.05 was considered statistical significant and all analyses were performed by Minitab version 16.1 (Minitab Inc., Coventry, UK).
Patient demographics
- Demographic data of the twelve patients, six with acute rejection and six controls, included in the present analysis are shown in Table 1.
- Three urine samples from each patient in the AR-group were analyzed; the first available urine sample after transplantation (5.0±3.6 days post transplant, baseline), one sample obtained in a clinically stable phase (7-11 days prior to BPAR) and one at the day of BPAR, obtained prior to biopsying.
- One protein, MEP1A, did not fit any of these groups and is presented separately.
- For the immune response protein group, up- regulation was observed in 4 out of 6 patients in the AR-group and in none of the control patients.
- Figure 2 shows the log 2 changes in protein levels for the specified protein groups between baseline and the time of BPAR in the AR-group.
Discussion
- The present analysis identifies several up-regulated urinary proteins in association with acute rejection episodes in the early post transplant phase after kidney transplantation.
- When comparing results from pooled samples taken from patients with acute rejection and samples taken from stable patients Sigdel et al. found several up- and down-regulated proteins.
- In addition, other proteins have been investigated using a more targeted approach (e.g. ELISA) but these were not confirmed by their investigation (10-19, 21).
- In addition individual samples were analyzed in the present study, not pooled urine.
- In conclusion, this study shows the applicability of shotgun proteomics in combination with quantification by 18O/16O-labeling in biomarker discovery in sequential urine samples.
Disclosure
- The study was entirely funded by internal budgets at the University of Oslo.
- The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.
Legend to figures
- Box plot showing fold change (log2) of immune proteins, growth factors and MEP1A from baseline to acute rejection in the AR-group compared with the control group.
- The center point (Clinically stable) is 7-11 days before BPAR, at stable serum creatinine levels.
Did you find this useful? Give us your feedback
Citations
194 citations
62 citations
59 citations
Cites methods from "2-D hydrophilic interaction liquid ..."
...[4] H....
[...]
...Additionally, the proteolytic peptides are subsequently separated with multidimensional (MD) LC–MS methods [2,4,5]....
[...]
58 citations
29 citations
References
194 citations
190 citations
188 citations
147 citations
136 citations
Related Papers (5)
Frequently Asked Questions (2)
Q2. What are the future works in "Searching for biomarkers of acute rejection in renal transplant recipients – development and optimization of a urinary proteomic approach" ?
Further prospective studies are therefore needed in larger populations, where biopsies also are performed in the control patients, in order to elucidate on the involvement of these proteins in acute rejection and their potential usability as diagnostic biomarkers. The use of urine and a trend towards an increase of proteins levels prior to deterioration of graft function potentially opens for early, specific and non-invasive detection of acute rejection episodes.