Histological Abnormalities Associated With DGF
Particular Banff scores were comparable in baseline donor biopsies (T 0) across the studied groups based on graft function development except for the rate of tubular atrophy (ct) that was slightly higher in the DGF group compared to the PGF group (P=0.07). The incidence of vascular nephrosclerosis, degree of interstitial fibrosis and glomerular sclerosis, and Remuzzi score were not different between the DGF and PGF groups (Table 3). Histological evaluation of preimplantation (T 1) and postimplantation (T 2) biopsies supported the above results after comparison between the studied groups (data not shown).
Molecular Changes Associated With DGF
To detect subtle changes associated with donor brain death and I/RI, we compared kidney allograft biopsy gene expression profiles of selected candidate genes at T 0, T 1, and T 2 from patients with DGF and PGF. The renal gene expression profile at T 0 was similar for DGF and PGF grafts. However, a higher Netrin-1 (NTN1) (16) expression level in donor kidneys that subsequently develop PGF was observed (lnRQ=0.52 [−0.54 to 1.14] in PGF vs. lnRQ=−0.47 [−7.23 to 0.28] in DGF, P=0.02; n-fold RQ PGF vs. DGF=3.77) (Fig. 1A). In T 1 biopsies of PGF, alpha-1-microglobulin (AMBP1), an adrenomedullin-binding protein with the potential to attenuate organ injury and organ inflammatory response (17–20), showed significantly higher expression levels compared with expression levels observed in patients with DGF (lnRQ=0.05 [−0.67 to 0.37] in PGF vs. lnRQ=−0.76 [−1.19 to −0.09] in DGF, P=0.02; n-fold RQ PGF vs. DGF=2.25). In T 2 biopsies, the higher NTN1 expression levels also separated organs with PGF from those with DGF (lnRQ=−0.22 [−1.44 to 0.32] in PGF vs. lnRQ=−1.77 [−9.2 to 0.38] in DGF, P=0.04; n-fold RQ PGF vs. DGF=4.72) (Fig. 1C). When performing analysis on the complete data set (including samples from all three biopsies), a significant down-regulation of NTN1, AMBP1, and caspase 9 (CASP9) was observed in DGF samples (P=0.05, one-sided test after Benjamini-Hochberg adjustment for multiple testing).
Immunohistochemical Evaluation for Netrin-1 Associated With DGF
The role of NTN1 in DGF was further confirmed at the protein level in T 0 and T 2 biopsy specimens. Moreover, immunohistochemical (IHC) evaluation showed higher intensity and higher numbers of positive tubular epithelial cells (mainly in the proximal tubules) in staining with anti–Netrin-1 antibody in cases with PGF compared to DGF grafts (Figure S1, SDC, at http://links.lww.com/TP/A884).
The Role of Tubular Damage in DGF
Since kidney proximal tubules are most vulnerable to I/RI (21) and tubular damage was associated with DGF in our study, the hypothesis that the tubular atrophy score (ct) combined with the Netrin-1 expression was predictive of DGF was tested. Intragraft gene expression of Netrin-1, known from univariate analyses to be significantly lower in DGF along with ct score (that was borderline significant, P=0.07) in T 0 biopsies were selected as predictor variables for training a logistic regression model parameterized on the donor biopsy data. This model achieved an area under the receiver operating curve (AUC, or c-statistic of 0.89) indicating a significant separation between DGF and PGF (Fig. 2, dashed line). To estimate how well the model generalizes, we repeatedly fit and tested the model using a leave-one-out (35-fold) cross-validation procedure. The average AUC (Fig. 2, solid line) resulting from this validation protocol was 0.81, confirming good discrimination power of this pair of predictors with respect to DGF.
In addition, we performed Spearman correlation analysis and demonstrated weak but negative correlation between the level of Netrin-1 and the duration of DGF (ρ=−0.284, P=0.004), which support the strength of Netrin-1 as a relevant biomarker of DGF.
Principal Component Analysis
Principal component analysis (PCA) is a mathematical procedure that reduces high dimensionality to two dimensions. Briefly, samples with similar gene expression profiles appear close to each other in the PCA diagram, while dissimilar samples are mutually more distant. Principal component analysis was calculated on the entire data set (n=105 cases), Significant overlap between DGF and PGF samples as well as between the respective biopsy groups (three sequential biopsies, i.e., T 0, T 1, and T 2) was observed (Figure S2, SDC, at http://links.lww.com/TP/A884), suggesting that it was not possible to discriminate between the respective groups through PCA. Visual inspection of the PCA diagram also suggests that sample heterogeneity within biopsy groups grows over time, that is, T 2 biopsies were more distantly spaced than T 1 biopsies, which in turn are more spread than T 0 biopsies. We verified this trend on the original (pre-PCA) data by computing all pairwise Euclidean distances (dissimilarities) between samples (real-valued vectors of expression) within each of the three biopsy groups. On the basis of a nonparametric Wilcoxon test, the mean of these distances (i.e., the heterogeneity) was significantly (P<0.001) smaller in T 0 group than in T 1 group, and significantly (P<0.001) smaller in the T 1 group than in the T 2 group. This result reflects growing sample heterogeneity from T 0 through T 1 and then to T 2 biopsies. While donor kidneys are quite homogenous at the time of organ procurement, I/RI leads to increased transcriptome diversity between allografts suggesting different organ vulnerability to ischemia and reperfusion.
This study established that donor kidney tubular cell atrophy and Netrin-1 gene and protein expression levels were associated with DGF of renal allografts. Moreover, this study was the first to describe in detail molecular changes occurring in three consecutive biopsies taken during organ procurement and transplantation as a means of identifying molecular events associated with either ischemia or reperfusion injury. Despite the absence of changes in histological abnormalities during I/RI, we found altered expression levels of genes associated with kidney graft injury. While cold storage did not lead to significant changes in the level of expression of the evaluated gene transcripts, reperfusion was associated with enhanced activation of innate and adaptive immune responses and apoptotic programs. These data demonstrated that reperfusion was associated with more enormous changes to the transcriptome in comparison with ischemia as visually shown by PCA where gene expression heterogeneity increased over time, with the highest heterogeneity observed in postimplantation biopsies. The possible explanation for the molecular silence during the ischemia is suppression of metabolism as the basic physiological mechanism during the hibernation. It is well known that acute kidney injury is characterized by a dissociation of pathology and physiology (22, 23). Our data provided quantitative measurements of inflammation and immune activation after reperfusion and support data suggesting that molecular profiling goes beyond histopathology by characterizing a discrete set of transcripts activated by I/RI (24). Molecular events found in postimplantation (T 2) biopsies reflect acute tissue injury and repair otherwise not detectable microscopically. Proinflammatory responses are likely to be the consequence of reperfusion as observed in T 2 biopsies and may augment tissue injury in excess of that produced by ischemia alone (25).
During the reperfusion, there were several transcripts up-regulated. The most significant ones were FOS, TNF, and HSPA1A. The FOS proteins have been implicated as regulators of cell proliferation, differentiation, and transformation and also involved in apoptosis. Tumor necrosis factor (TNF) is a cytokine mainly involved in the regulation of a wide spectrum of biological processes including cell proliferation, differentiation, apoptosis, lipid metabolism, and coagulation. Heat shock 70 kDa protein 1 (HSP) is a protein that in humans is encoded by the HSPA1A gene. The protein stabilizes existing proteins against aggregation and mediates the folding of newly translated proteins in the cytosol and in organelles. All three genes have been shown to play some role in autophagy, which is an essential, conserved lysosomal degradation pathway that controls the quality of the cytoplasm by eliminating protein aggregates and damaged organelles. Moreover, autophagy likely plays an important role in normal proximal tubule function and recovery from acute ischemic kidney injury (26, 27).
The primary aim of this study was to identify potential risk factors associated with DGF. However, no associations between donor, recipient, organ procurement, surgery conditions, and DGF development were found. However, tubular atrophy scores were higher among donor kidneys that developed DGF. Besides severe tubular atrophy, the lower messenger RNA expression of Netrin-1 was observed in both T 0 and T 2 biopsies in patients who suffered from DGF. This indicated the importance of tubular quality in the context of posttransplant outcomes. It was shown that preexisting renal tubular atrophy in donor kidney reflected tubular cell loss after chronic renal insult (28). I/RI triggers apoptosis that predominantly affects tubular epithelial cells resulting in acute tubular necrosis. Apoptosis is not only an important cause of a cell death immediately after I/RI but also contributes to recovery from I/RI by providing a balance between excessive renal cell proliferation associated with the process of replacing irreversibly injured tubular epithelial cells and restoring tubular integrity (29). In the event of apoptotic cell death exceeding mitotic replacement, renal cell loss results in tubular atrophy (30). The effector apoptotic pathway is constantly suppressed by survival factors. Their deficiency triggers apoptosis while their availability facilitates recovery from ischemic injury (31). Netrin-1 facilitates recovery from ischemic injury by stimulating the proliferation of surviving renal tubular cells and inhibition of apoptosis (32). Normally, Netrin-1 is not expressed (or is present at low levels) in healthy tubular epithelial cells (16) but is expressed in tubular cells during recovery from ischemia. In addition, Netrin-1 expression levels are significantly elevated in urine of patients with various forms of acute kidney injury including ischemic injury (16, 33, 34). The beneficial effects of Netrin-1 on the suppression of tissue injury were first described in mice (34, 35). One of the most interesting findings described in our report was higher expression of Netrin-1 in PGF compared to DGF grafts in T 0 biopsies as well as in T 2 biopsies. Renal tubular cell apoptosis induced by acute I/RI, especially in combination with preexisting chronic tubular atrophy, was associated with a higher incidence of DGF (36, 37). The worse short-term outcomes after deceased versus live donor transplant kidneys are due to impaired tubular counterbalance of oxidative stress at the time of implantation (38). Despite the inconsistent conclusions regarding the effects of the donor kidney biopsy’s predictive value (39), a large study examining expanded criteria donors showed that kidney allocation based on histological evaluation of donor biopsies (40), eventually combining with donor and recipient risk factors, helped improve graft survival (41). Here we suggest that Netrin-1 gene expression might help to better understand vulnerability to I/RI, especially in expanded criteria donor kidneys with preexisting chronic changes that are uniquely susceptible to I/RI because of older donor age, comorbidities, and terminal effects of brain death on kidneys with limited functional reserve. Moreover, recently it was shown that Netrin-1 stimulates HSP, a gene associated with autophagy and thus contributes to the protection of hypoxia-induced apoptosis via the inhibition of mitochondrial dysfunction (42). Therefore, it is likely that higher autophagy-related gene transcripts may be associated with the reduction of DGF.
There are some limitations that need to be acknowledged and addressed regarding the present study. In our study, we adhered to statistical rigor when identifying differentially expressed genes. The genes considered to be differentially expressed had to fulfill not only statistical significance but also biological significance when minimal cutoff for differential expression was two-fold. We were unable to discriminate between DGF and PGF grafts after target gene analysis as shown by PCA. This observation suggested limitations in the target genes selected, suggesting that other yet-to-be-defined molecules likely contributed to the pathogenesis of DGF. The sample size might appear as relatively small. However, the value of this study includes the molecular analysis of 92 target genes in 105 samples of repeated three consecutive biopsies taken during organ procurement and transplantation in 35 human donor kidneys. Furthermore, we were not able to discriminate a role of recipient blood leukocytes during reperfusion on analyzed genes expression.
In conclusion, analysis of donor kidneys identified poor tubular cell quality and low Netrin-1 expression levels to be associated with DGF. Netrin-1 is a survival factor that plays a protective role in recovery from I/RI. In the future, targeted therapeutic intervention might help to decrease the incidence of DGF in organs susceptible to I/RI. Despite minimal histopathological abnormalities occurring during reperfusion injury, reperfusion was associated with enhanced activation of innate and adaptive immune responses and apoptotic programs, contrary to relative molecular silence of ischemia. Taking together, the combination of both conventional histology and molecular pathology techniques during transplantation surgery may represent a more accurate tool to discriminate grafts at risk for DGF.
MATERIALS AND METHODS
Study Design and Patients
Allograft biopsy specimens from 38 deceased donor kidneys that were consecutively allocated to transplant recipients at our center between October 2009 and June 2010 were used in this study. Transplant recipients participating in concurrent immunosuppressive drug studies and patients with a bleeding history were not included. The study was performed in accordance with guidelines established by the Declaration of Helsinki, and the study protocol was approved by the Ethics Committee of the Institute for Clinical and Experimental Medicine. Informed consent was obtained from all patients (IRB approval number F2002/08).
Definition of DGF
Delayed graft function was defined as the need for dialysis during the first week after transplantation (except dialysis for hyperkalemia or volume overload in the first 24 hr) (2, 43).
To investigate changes in the intrarenal transcription profile during I/RI, three sequential biopsies were performed during organ procurement and transplantation, that is, at the time the kidney was procured (baseline donor, T 0), preimplantation biopsy (T 1), and postimplantation “30-min” (T 2) biopsy:
- 1. The “T 0” biopsy is the standard intraoperative, subcapsular donor wedge biopsy obtained at the time of kidney procurement right after perfusion with 4°C histidine-tryptophan ketoglutarate preservation solution.
- 2. The “T 1” biopsy was taken during the transplantation surgery at the back table at the end of cold ischemia time just before reperfusion using a sterile 14-gauge biopsy needle, reflecting molecular and structural changes during cold ischemia.
- 3. The “T 2” biopsy was performed 30 min after release of the vascular clamps allowing for the detection of reperfusion changes in the allograft. The core of the biopsy specimen cortical zone was placed immediately in RNAlater stabilization reagent (Qiagen, Hilden, Germany) and stored at 4°C overnight and then cooled to −80°C until RNA was extracted.
Histological Evaluation of Kidney Biopsies
Histological evaluation of donor biopsies at baseline was carried out according to the 2005 Banff working classification criteria (Table 3) (12). The median Remuzzi score (13) used to quantify donor kidney quality was 4 of 12. We did not observe any changes in histological abnormalities or histopathological signs of inflammation in preimplantation or postimplantation biopsies (data not shown).
Molecular Evaluation of Kidney Biopsies
To investigate changes in the intrarenal transcription profile during ischemia phase of I/RI, we compared the expression profile of the preimplantation biopsy (T 1) with the donor biopsy (T 0). In a similar fashion, by comparing the postimplantation (T 2) and preimplantation biopsies (T 1), the reperfusion period was characterized (Fig. 3).
Gene Probe Set Selection
The expression profile of 92 candidate genes was used to evaluate injury and repair signals associated with organ procurement and reperfusion. Studied genes included genes associated with apoptosis, innate and adaptive immunity, and oxidative stress signaling. We identified target genes potentially associated with graft function development on the basis of available published work describing complementary DNA microarray and reverse transcriptase quantitative polymerase chain reaction analyses (Table S2, SDC, at http://links.lww.com/TP/A884). Housekeeping genes were selected based on expression stability (Figure S3, SDC, at http://links.lww.com/TP/A884). Additional supporting information on real-time reverse transcriptase polymerase chain reaction analysis, expression data analysis, immunohistochemistry, and immunosuppressive treatment of recipients may be found in the online version of this article (Materials and Methods, SDC, at http://links.lww.com/TP/A884).
For demographic and the gene expression data, nonparametric testing was performed, and data were expressed as median and interquartile range values. Continuous clinical variables were compared using the Mann-Whitney U test, and categorical variables were compared using the Fisher exact test. Significance was established by two-sided P values <0.05 unless otherwise specified. The Benjamini-Hochberg adjustment was used when appropriate. Only genes that fulfilled the selection criteria (combination of both statistical and biological significance criteria; P<0.05, at least two-fold gene expression difference between studied groups, respectively) were considered as significantly differentially expressed. Additional supporting information on multivariable logistic regression, leave-one-out cross-validation procedure, receiver operating curve analysis, functional interpretation of genes, and principal component analysis (PCA) may be found in the online version of this article (Supplemental Digital Content, Materials and Methods).
The authors thank the transplantation coordinators, nurses, and patients for their cooperation and help. Special thanks goes to Romana Polackova for expert technical assistance and Vera Polaskova for her help with data collection.
1. Ojo AO, Wolfe RA, Held PJ, et al. Delayed graft function: risk factors and implications for renal allograft survival. Transplantation 1997; 63: 968.
2. Halloran PF, Hunsicker LG. Delayed graft function: state of the art, November 10–11, 2000. Summit meeting, Scottsdale, Arizona, USA. Am J Transplant 2001; 1: 115.
3. El-Husseini A, Sabry A, Zahran A, et al. Can donor implantation renal biopsy predict long-term renal allograft outcome? Am J Nephrol 2007; 27: 144.
4. Mas VR, Archer KJ, Yanek K, et al. Gene expression
patterns in deceased donor kidneys developing delayed graft function after kidney transplantation. Transplantation 2008; 85: 626.
5. Yoshida T, Kurella M, Beato F, et al. Monitoring changes in gene expression
in renal ischemia-reperfusion in the rat. Kidney Int 2002; 61: 1646.
6. Supavekin S, Zhang W, Kucherlapati R, et al. Differential gene expression
following early renal ischemia/reperfusion. Kidney Int 2003; 63: 1714.
7. Hauser P, Schwarz C, Mitterbauer C, et al. Genome-wide gene-expression patterns of donor kidney biopsies distinguish primary allograft function. Lab Invest 2004; 84: 353.
8. Mueller TF, Reeve J, Jhangri GS, et al. The transcriptome of the implant biopsy identifies donor kidneys at increased risk of delayed graft function. Am J Transplant 2008; 8: 78.
9. Hoffmann SC, Kampen RL, Amur S, et al. Molecular and immunohistochemical characterization of the onset and resolution of human renal allograft ischemia-reperfusion injury. Transplantation 2002; 74: 916.
10. Avihingsanon Y, Ma N, Pavlakis M, et al. On the intraoperative molecular status of renal allografts after vascular reperfusion and clinical outcomes. J Am Soc Nephrol 2005; 16: 1542.
11. Perco P, Kainz A, Wilflingseder J, et al. Histogenomics: association of gene expression
patterns with histological parameters in kidney biopsies. Transplantation 2009; 87: 290.
12. Solez K, Colvin RB, Racusen LC, et al. Banff 07 classification of renal allograft pathology: updates and future directions. Am J Transplant 2008; 8: 753.
13. Remuzzi G, Grinyo J, Ruggenenti P, et al. Early experience with dual kidney transplantation in adults using expanded donor criteria. Double Kidney Transplant Group (DKG). J Am Soc Nephrol 1999; 10: 2591.
14. Basile DP, Fredrich K, Weihrauch D, et al. Angiostatin and matrix metalloprotease expression following ischemic acute renal failure. Am J Physiol Renal Physiol 2004; 286: 893.
15. Viappiani S, Sariahmetoglu M, Schulz R. The role of matrix metalloproteinase inhibitors in ischemia-reperfusion injury in the liver. Curr Pharm Des 2006; 12: 2923.
16. Urbschat A, Obermuller N, Haferkamp A. Biomarkers of kidney injury. Biomarkers 2011; 16 (Suppl 1): S22.
17. Shah KG, Rajan D, Jacob A, et al. Attenuation of renal ischemia and reperfusion injury by human adrenomedullin and its binding protein. J Surg Res 2010; 163: 110.
18. Yang J, Wu R, Qiang X, et al. Human adrenomedullin and its binding protein attenuate organ injury and reduce mortality after hepatic ischemia-reperfusion. Ann Surg 2009; 249: 310.
19. Chaung WW, Wu R, Ji Y, et al. Peripheral administration of human adrenomedullin and its binding protein attenuates stroke-induced apoptosis and brain injury in rats. Mol Med 2011; 17: 1075.
20. Carrizo GJ, Wu R, Cui X, et al. Adrenomedullin and adrenomedullin-binding protein-1 downregulate inflammatory cytokines and attenuate tissue injury after gut ischemia-reperfusion. Surgery 2007; 141: 245.
21. Molitoris BA, Wagner MC. Surface membrane polarity of proximal tubular cells: alterations as a basis for malfunction. Kidney Int 1996; 49: 1592.
22. Schlondorff DO. Overview of factors contributing to the pathophysiology of progressive renal disease. Kidney Int 2008; 74: 860.
23. Rosen S, Stillman IE. Acute tubular necrosis is a syndrome of physiologic and pathologic dissociation. J Am Soc Nephrol 2008; 19: 871.
24. Mueller TF, Solez K, Mas V. Assessment of kidney organ quality and prediction of outcome at time of transplantation. Semin Immunopathol 2011; 33: 185.
25. Kurian SM, Flechner SM, Kaouk J, et al. Laparoscopic donor nephrectomy gene expression
profiling reveals upregulation of stress and ischemia associated genes compared to control kidneys. Transplantation 2005; 80: 1067.
26. Kimura T, Takabatake Y, Takahashi A, et al. Autophagy
protects the proximal tubule from degeneration and acute ischemic injury. J Am Soc Nephrol 2011; 22: 902.
27. Singh KK, Yanagawa B, Quan A, et al. Autophagy
gene fingerprint in human ischemia and reperfusion. Thorac Cardiovasc Surg 2013 [Epub ahead of print].
28. Lieberthal W, Levine JS. Mechanisms of apoptosis and its potential role in renal tubular epithelial cell injury. Am J Physiol 1996; 271: 477.
29. Bonventre JV. Dedifferentiation and proliferation of surviving epithelial cells in acute renal failure. J Am Soc Nephrol 2003; 14 (Suppl 1): S55.
30. Mullen TD, Jenkins RW, Obeid LM, et al. Apoptosis in the kidney. In: Reed JC, Green DR, eds. Apoptosis: Physiology and Pathology. Cambridge, England: Cambridge University Press, 2011: 242.
31. Raff MC. Social controls on cell survival and cell death. Nature 1992; 356: 397.
32. Wang W, Reeves WB, Pays L, et al. Netrin-1
overexpression protects kidney from ischemia reperfusion injury by suppressing apoptosis. Am J Pathol 2009; 175: 1010.
33. Reeves WB, Kwon O, Ramesh G. Netrin-1
and kidney injury. II. Netrin-1
is an early biomarker of acute kidney injury. Am J Physiol Renal Physiol 2008; 294: 731.
34. Wang W, Reeves WB, Ramesh G. Netrin-1
and kidney injury. I. Netrin-1
protects against ischemia-reperfusion injury of the kidney. Am J Physiol Renal Physiol 2008; 294: 739.
35. Wu TW, Li WW, Li H. Netrin-1
attenuates ischemic stroke–induced apoptosis. Neuroscience 2008; 156: 475.
36. Oberbauer R, Rohrmoser M, Regele H, et al. Apoptosis of tubular epithelial cells in donor kidney biopsies predicts early renal allograft function. J Am Soc Nephrol 1999; 10: 2006.
37. Ueda N, Shah SV. Tubular cell damage in acute renal failure—apoptosis, necrosis, or both. Nephrol Dial Transplant 2000; 15: 318.
38. Kainz A, Mitterbauer C, Hauser P, et al. Alterations in gene expression
in cadaveric vs. live donor kidneys suggest impaired tubular counterbalance of oxidative stress at implantation. Am J Transplant 2004; 4: 1595.
39. Sung RS, Christensen LL, Leichtman AB, et al. Determinants of discard of expanded criteria donor kidneys: impact of biopsy and machine perfusion. Am J Transplant 2008; 8: 783.
40. Remuzzi G, Cravedi P, Perna A, et al. Long-term outcome of renal transplantation
from older donors. N Engl J Med 2006; 354: 343.
41. Anglicheau D, Loupy A, Lefaucheur C, et al. A simple clinico-histopathological composite scoring system is highly predictive of graft outcomes in marginal donors. Am J Transplant 2008; 8: 2325.
42. Son TW, Yun SP, Yong MS, et al. Netrin-1
protects hypoxia-induced mitochondrial apoptosis through HSP27 expression via DCC- and integrin α6
-dependent Akt, GSK-3β, and HSF-1 in mesenchymal stem cells. Cell Death Dis 2013; 4: 563.
43. Yarlagadda SG, Coca SG, Garg AX, et al. Marked variation in the definition and diagnosis of delayed graft function: a systematic review. Nephrol Dial Transplant 2008; 23: 2995.
44. Urbanova M, Brabcova I, Girmanova E, et al. Differential regulation of the nuclear factor-kappaB pathway by rabbit antithymocyte globulins in kidney transplantation. Transplantation 2012; 93: 589.
45. Sanchez-Espiridion B, Sanchez-Aguilera A, Montalban C, et al. A TaqMan lowdensity array to predict outcome in advanced Hodgkin’s lymphoma using paraffin-embedded samples. Clin Cancer Res 2009; 15: 1367.
46. Livak KJ, Schmittgen TD. Analysis of relative gene expression
data using real-time quantitative PCR and the 2(−Delta Delta CT) method. Methods 2001; 25: 402.
47. Collette S, Collette L, Budiharto T, et al. Predictors of the risk of fibrosis at 10 years after breast conserving therapy for early breast cancer: a study based on the EORTC Trial 22881-10882 ‘boost versus no boost’. Eur J Cancer 2008; 44: 2587.
48. Holec M, Klema J, Zelezny F, et al. Comparative evaluation of set-level techniques in predictive classification of gene expression
samples. BMC Bioinformatics 2012; 13 (Suppl 10): S15.
49. Goeman JJ, Buhlmann P. Analyzing gene expression
data in terms of gene sets: methodological issues. Bioinformatics 2007; 23: 980.
50. Goeman JJ VdGS, van Houwelingen HC. Testing against a high dimensional alternative. J R Statist Soc B 2006; 68: 477.
Renal transplantation; Gene expression; Netrin-1; TaqMan Low Density Array; Donor biopsy; Ischemia/reperfusion injury; Autophagy
Supplemental Digital Content
© 2014 by Lippincott Williams & Wilkins