Introduction
AKI requiring dialysis (AKI-D) is associated with high mortality in patients who were hospitalized. Of those that survive, a fraction recovers kidney function, but most progress to CKD (12–3). Current assessment of kidney recovery primarily relies on urine output and measurement of serum creatinine to estimate GFR; however, these parameters are inconsistent and limited in predicting kidney recovery (4).
Previous studies have examined the potential of decreased levels of select kidney injury biomarkers (56–7) or inflammatory biomarkers (8,9) for the prediction of kidney recovery from AKI, including urine hepatocyte growth factor, urine neutrophil gelatinase-associated lipocalin (uNGAL), and plasma NGAL (pNGAL). Studies also found that urine insulin-like growth factor-binding protein 7, and tissue inhibitor of metalloproteinase 2 provided more accurate prediction of kidney recovery than other investigated urine AKI biomarkers (10,11). Recent studies showed that post-AKI proteinuria was associated with kidney disease progression (12) and preadmission proteinuria was an independent risk factor for nonrecovery for AKI-D (13). The Translational Research Investigating Biomarker Endpoints for AKI study found that high postoperative plasma levels of vascular endothelial growth factor A isoform (VEGF) and placental growth factor were independently associated with reduced risk for AKI, prolonged AKI, and mortality, whereas high levels of the antiangiogenic marker soluble VEGF receptor 1 were associated with increased risk for these outcomes in patients after cardiac surgery (14). These studies suggest that serum/plasma/urine AKI or angiogenic biomarkers with use of appropriate mathematical models could provide prognostic information for the prediction of kidney function recovery. However, the performance of these biomarkers/models has not been well established or validated.
Using the slow off-rate modified aptamers scan (SOMAscan) proteomics platform, we recently identified fibroblast growth factor-23 (FGF23), tissue plasminogen activator (tPA), neutrophil collagenase (matrix metalloproteinase-8), soluble urokinase plasminogen activator receptor, and IL-6 as potential mortality-associated biomarkers in patients with AKI-D (15). In this study, we aimed to identify biomarkers for prediction of kidney function recovery by analysis of serum samples obtained on day 8 of randomized RRT from 72 patients with AKI-D enrolled in the Veteran's Affairs/National Institutes of Health Acute Renal Failure Trial Network (ATN) study. Although most previous studies concentrated on kidney recovery within 60 or 90 days (5,16), this study focused on kidney recovery within 28 days. Our study not only confirmed that lower serum levels of AKI biomarkers were associated with AKI recovery but also identified higher levels of survival-related proteins as novel biomarkers of kidney recovery.
Methods
Study Design
The ATN study was a prospective, multicenter clinical trial to evaluate strategies of intensive versus conventional RRT in patients who were critically ill with AKI-D; 1124 patients were enrolled and randomly assigned to intensive or conventional RRT in 27 Veterans Affairs (VA) and 12 academic medical centers across the United States. Outcomes included 60-day mortality, recovery of kidney function, and intensive care unit and hospital lengths of stay. Details of the study protocol including inclusion and exclusion criteria have been previously published (17,18). Patients enrolled in the ATN study were critically ill adults (aged ≥18 years) who had AKI clinically consistent with acute tubular necrosis and failure of one or more nonrenal organ(s) (defined as a nonrenal sequential organ failure assessment [SOFA] score of ≥2) or sepsis. Patient consent for serum sample collection and at least one sample were obtained from 827 of the 1124 subjects who participated in the ATN study. A total of 819 patients provided samples on day 1 and 573 patients on day 8, with 565 patients contributing samples on both day 1 and day 8 of randomized RRT. Among the 626 patients who survived to day 28, 343 patients were dialysis independent on day 28, whereas the remaining 283 patients remained on dialysis. Selection of the samples used for this study was random, and the decision was made by the ATN Study coordinating center/biorepository, with constraints regarding survival status that were imposed. This post-hoc proteomic biomarker study of kidney recovery included day 8 serum samples from 72 randomly selected patients who either survived independent of dialysis (n=38) or survived dependent on dialysis (n=34) by day 28. We did not analyze urine samples because urine was not stored in the ATN study. Power analysis indicated that differentially expressed proteins could be identified with a power of >0.8. In this study, kidney function recovery was defined as alive and free of dialysis on day 28. Clinical data available through the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) data repository via a crosswalk file were then linked to the deidentified samples analyzed in this study. This post-hoc analysis was approved by the Salem Veterans Affairs Medical Center and Food and Drug Administration (FDA) Institutional Review Boards.
SOMAscan Proteomic Profiling
Quantitative proteome profiling of AKI samples was performed by SOMAscan assay (version 1.3k) developed by SomaLogic Inc. (Boulder, CO) as described previously (15,19). The assay quantified 1305 low, middle, and high abundance proteins using single-stranded DNA slow off-rate modified aptamers (SOMAmers), with five calibration samples and two quality control samples for each assay. Briefly, serum samples (50 µl) were incubated with preimmobilized SOMAmers, which were subjected to a series of washes to remove nonspecific bindings. SOMAmers specifically bound to their cognate proteins were released and hybridized to custom DNA microarrays. The microarrays were scanned using an Agilent C scanner (G2505C, Agilent Technologies, Palo Alto, CA). The raw data were processed as described previously (15). Mainly, hybridization control normalization was applied to relative fluorescence units (RFUs) to remove variation introduced during the hybridization and scanning processes, followed by median signal normalization to eliminate intrarun bias, and finally calibration to account for inter-run differences.
Olink Assay of Inflammatory Proteins
The Olink inflammation panel consisting of 92 proteins was measured by Olink Proteomics (Boston, MA) using the Proximity Extension Assay technology, with consumption of 1 µl of each serum sample, as previously described (20). Quality control was performed by adding four internal controls into all samples and running external controls in every assay plate. Assay results were reported in arbitrary, relative units as Normalized Protein eXpression (NPX) on a log2 scale. More information regarding the Olink platform, NPX, assay validation data, and the full list of proteins in the panel is available at Olink’s website (www.olink.com). The correlation between SOMAscan and Olink data was assessed using multivariate correlation analysis in JMP (version 12.1.0, SAS Institute Inc., Cary, NC) to calculate the correlation coefficient and P value.
Identification of Differential Proteins
Day 8 samples were stratified on the basis of dialysis status on day 28. Welch’s t test was performed for log-transformed SOMAscan RFUs to find significantly changed proteins between the dialysis-dependent and independent groups. A fold change of ≥1.2 and P<0.05 were set as significant changes as the criteria were evaluated and validated (15). The t test P values were also adjusted for this multiplex assay to calculate false discovery rate using the Benjamini and Hochberg method (21). For protein expression mean, SD, and fold changes, RFUs were used for SOMAscan data, whereas log-scale NPX was converted to a linear scale NPX (i.e., non–log transformed, linear NPX) for Olink data.
Ingenuity Pathway Analyses
Ingenuity Pathway Analysis (IPA) software was used for gene ontology, pathways, and core analysis comparison of proteins with significant fold changes in patients who discontinued dialysis versus those who continued dialysis on day 28. A P value <0.05 was considered significant.
Statistical Analysis of Proteins for Kidney Recovery
The top proteins with altered abundance (either higher or lower serum levels) were of interest in this study. Because the distributions of these proteins were relatively skewed, patients were classified into tertiles for each of these proteins, and the associations of categorized proteins with the kidney recovery response were further evaluated and the effects of elevated protein levels were quantified. Continuous variables were expressed as mean with SD or median (25th, 75th percentile) and categorical variables as frequencies and percentages. The differences between patients with and without kidney recovery were evaluated using t test or Mann–Whitney nonparametric test for continuous variables and chi-squared test (or Fisher’s exact test) for categorical variables. Tertile-specific recovery rate (% of recovery) was calculated for each protein, and chi-squared test or Fisher’s exact test was performed to test the difference in recovery rate among tertiles. The association of serum levels of each protein with the probability of kidney recovery was evaluated in the univariate logistic regression initially, and then in the multivariable logistic regression with adjustment for age, sex, congestive heart failure, cardiovascular SOFA (CV-SOFA) score, diabetes, intensive dialysis treatment, cause of AKI, baseline serum creatinine, day-8 urine volume, and estimated 60-day mortality risk. The lower tertile level for every protein was considered as the reference group. Reported P values were not adjusted for multiple comparisons. All statistical analyses were performed using SAS software version 9.4 (SAS Institute, Cary, NC).
Results
Patient Characteristics
We analyzed a total of 72 randomly selected samples obtained on day 8 of randomized RRT as part of the ATN study. Table 1 shows the characteristics of the patients stratified by the status of kidney function recovery by day 28 (i.e., dialysis on or off). Kidney function recovered (i.e., survived without dialysis) in 38 (53%) patients, whereas the remaining 34 (47%) patients survived with dialysis by day 28. Thus, recovery rate of kidney function in this subset of patients that we studied was similar to that in the overall ATN study cohort in which 343 of 626 patients (55%) surviving to day 28 were dialysis independent on day 28. The number (percentage) of patients with or without recovered kidney function by day 28 was also similar in diabetes, CV-SOFA score ≥2, intensive dialysis treatment arm, and cause of AKI, with no statistically significant differences (P>0.05). Higher baseline SOFA scores (both overall and CV component) were associated with mortality; however, they were not associated with recovery of kidney function among survivors. The median baseline serum creatinine levels and estimated 60-day mortality risk scores were the same between the two groups. In addition, the median urine volume on day 8 was not significantly different (P=0.06) between the kidney-recovered group (81.5, 60–478 ml/day) and the nonrecovered group (59.5, 10–207 ml/day). These urine volumes were far less than that expected for kidney function recovery (>500 ml/day).
Table 1. -
Characteristics of day 8 patient cohorts stratified by kidney recovery status by day 28
| Characteristic |
Overall (n=72) |
Not Recovered (n=34) |
Recovered (n=38) |
P Valuea |
| Age, yr |
60.1±15.1 |
62.8±15.3 |
57.6±14.8 |
0.15 |
| Sex, F |
27 (37.5%) |
10 (29.4%) |
17 (44.7%) |
0.18 |
| Congestive heart failure |
11 (15.7%) |
3 (9.1%) |
8 (21.6%) |
0.20 |
| Diabetes mellitus |
23 (32.4%) |
10 (29.4%) |
13 (35.1%) |
0.61 |
| CV-SOFA ≥2 |
42 (58.3%) |
21 (61.8%) |
21 (55.3%) |
0.58 |
| Intensive treatment arm |
43 (59.7%) |
19 (55.9%) |
24 (63.2%) |
0.53 |
|
Cause of AKI
|
|
|
|
|
|  Ischemic |
51 (70.8%) |
25 (73.5%) |
26 (68.4%) |
0.63 |
|  Nephrotoxic |
18 (25.0%) |
10 (29.4%) |
8 (21.1%) |
0.41 |
|  Sepsis |
40 (55.6%) |
17 (50.0%) |
23 (60.5%) |
0.37 |
|  Multifactorial |
39 (54.2%) |
19 (55.9%) |
20 (52.6%) |
0.78 |
| Baseline serum creatinine |
1.3 (1.05, 1.8) |
1.3 (1.0, 1.7) |
1.3 (1.1, 1.8) |
0.67 |
| Urine volume on day 8, ml/day |
81.5 (31, 255) |
59.50 (10, 207) |
81.5 (60, 478) |
0.06 |
| Estimated 60-day mortality risk |
0.55 (0.35, 0.81) |
0.55 (0.33, 0.81) |
0.55 (0.36, 0.81) |
0.77 |
| Serum calcium, mg/dl |
7.8±1.0 |
7.8±1.0 |
7.8±1.0 |
0.93 |
| Serum PO4, mg/dl |
5.4±1.9 |
5.5±2.1 |
5.4±1.8 |
0.77 |
Continuous variables are expressed as mean±SD or median (25th, 75th percentile), categorical variables as %; CV-SOFA, cardiovascular sequential organ failure assessment; F, female; PO4, phosphate.
aP values comparing patients who recovered kidney function versus those who did not recover were calculated using t test for continuous variables expressed as mean±SD, Mann–Whitney nonparametric test for those expressed as median (25th, 75th percentile), and chi-square test (or Fisher’s exact test if the count is <5) for categorical variables. Categorized causes of AKI in the Acute Renal Failure Trial Network (ATN) study include ischemic, nephrotoxic, and sepsis, and multifactorial if more than one of those three.
SOMAScan Proteomic Profile of Day 8 AKI-D Serum Samples
A volcano plot of SOMAscan analysis of 1305 serum proteins (Figure 1) demonstrates fold changes (FC; FC cutoff=1.2) and statistical significance (P<0.05) of those proteins that were higher (red dots) or lower (green dots) in day 8 serum samples from patients who discontinued dialysis compared with those who remained on dialysis by day 28. Although most serum proteins did not show a significant statistical difference (gray dots) between the two groups, 119 proteins showed statistically significant differences in serum levels, 53 of which were higher and 66 of which were lower in the group of patients who discontinued dialysis (Figure 1, Supplemental Table 1). Serum levels of mortality-associated proteins such as FGF23 (FC=−2.69, P=0.01) and IL-6 (FC=–2.44, P=0.009), AKI biomarkers NGAL (FC=–1.62, P=0.002) and chitinase-3-like protein 1 (YKL-40, FC=–1.44, P=0.04), and cardiac biomarker creatine kinase-MB (CK-MB, FC=–9.94, P=0.004) were lower on day 8 in the patients that recovered by day 28. In contrast, several growth and survival-related proteins were higher on day 8 in the patients that discontinued dialysis by day 28, such as tyrosine-protein kinase Fyn (FC=2.28, P=0.01) and BTK (FC=1.63, P=0.04), protein Wnt-7a (FC=1.83, P=0.003), Myc proto-oncogene protein (c-Myc, FC=1.62, P=0.0004), ghrelin (FC=1.42, P=0.005), platelet-derived growth factor C (PDGF-C, FC=1.40, P=0.005), survivin (FC=1.39, P=0.02), epidermal growth factor (EGF, FC=1.32, P=0.004), and neuregulin-1 (FC=1.31, P=0.03). The SOMAscan measurements of all 1305 proteins are presented in Supplemental Table 2.
;)
Figure 1.: Changes in serum protein levels in samples obtained on day 8 from patients who recovered kidney function, compared with patients who did not recover by day 28. The volcano plot is used for visualizing the relationship between fold changes and statistical significance. The vertical lines correspond to 1.2-fold changes in up and down expression, whereas the horizontal line represents a P value of 0.05. The gray dots represent proteins with no statistically significant differences (fold change <1.2 and P≥0.05); the red dots represent upregulated proteins with fold change ≥1.2 and P<0.05; the green dots represent downregulated proteins with fold change ≥1.2 and P<0.05; and the yellow dots represent proteins with fold change <1.2 and P<0.05. FYN, tyrosine-protein kinase Fyn; CXCL11, C-X-C motif chemokine 11 (I-TAC); CXCL2/CXCL3, C-X-C motif chemokine 2/C-X-C motif chemokine 3 (Gro-beta/gamma); CD86, T-lymphocyte activation antigen CD86 (activation B7-2 antigen); WNT7A, protein Wnt-7a; BTK, tyrosine-protein kinase BTK; c-Myc, Myc proto-oncogene protein; TIMP-3, metalloproteinase inhibitor 3; CCL5, C-C motif chemokine 5 (T-cell-specific protein RANTES); PDGF-C, platelet-derived growth factor C; CA2, carbonic anhydrase 2; IL-9, interleukin-9; EGF, epidermal growth factor; YKL-40, chitinase-3-like protein 1; CXCL16, C-X-C motif chemokine 16; NGAL, neutrophil gelatinase-associated lipocalin; IL1RL1, interleukin-1 receptor-like 1; IL-6, interleukin-6; CA3, carbonic anhydrase 3; FGF23, fibroblast growth factor 23; CK-MB, creatine kinase-MB.
IPA Analysis of Biologic Function Changes Related to Tissue Repair
IPA core analysis of proteins with significant changes in abundance revealed significant alterations in diseases and biologic functions on day 8 (Figure 2, Supplemental Table 3). The analysis indicated increased functions in cell-to-cell signaling and interaction, cellular growth and proliferation, cellular development, embryonic development, tissue morphology, tissue development, organismal development, molecular transport, and post-translational modification. The analysis also showed decreased function in cell death, connective tissue disorders, organismal injury and abnormalities, and inflammatory response and disease. The results suggest molecular pathways related to these biologic functions were already activated or suppressed on day 8, which potentially affected kidney recovery by day 28.
Figure 2.: Significantly altered biologic functions or diseases revealed from Ingenuity Pathway Analysis. The image blocks are sized by −log (P value) and colored by z score. The smaller the P value, the larger the block; and the larger the absolute z score, the deeper the color. Orange indicates a positive z score and functional increase. Blue indicates a negative z score and functional decrease.
Olink Verification of Top-Changed Proteins
Of the top 25 proteins with significant changes identified from the SOMAscan assay (Table 2), CXCL11, IL-6, and FGF23 were also included on the Olink inflammation panel (Supplemental Table 4). Consistent results were obtained for all three proteins between the SOMAscan and Olink assays. Strong positive correlations between the two assays were observed for CXCL11 and IL-6 with correlation coefficients of 0.86 (P<0.0001) and 0.89 (P<0.0001), respectively. SOMAscan and Olink intensities of FGF23 showed positive but moderate correlation with r=0.47 and P<0.0001 (Figure 3).
Table 2. -
Kidney recovery rates by tertiles of serum proteins
| Protein |
Tertile 1a (Low Levels) |
Tertile 2a (Intermediate Levels) |
Tertile 3a (High Levels) |
Chi-Square P Value |
|
Proteins with higher serum levels in the recovery group
|
|  FYN |
11 (45.8) |
11 (45.8) |
16 (66.7) |
0.25 |
|  CXCL11 |
6 (25.0) |
15 (62.5) |
17 (70.8) |
0.003 |
|  CXCL2/CXCL3 |
8 (33.3) |
13 (54.2) |
17 (70.8) |
0.03 |
|  CD86 |
9 (37.5) |
11 (45.8) |
18 (75.0) |
0.02 |
|  WNT7A |
10 (41.7) |
9 (37.5) |
19 (79.2) |
0.006 |
|  BTK |
9 (37.5) |
11 (45.8) |
18 (75.0) |
0.02 |
|  c-Myc |
9 (37.5) |
8 (33.3) |
21 (87.5) |
0.0002 |
|  TIMP-3 |
9 (37.5) |
10 (41.7) |
19 (79.2) |
0.006 |
|  CCL5 |
9 (37.5) |
12 (50.0) |
17 (70.8) |
0.07 |
|  Ghrelin |
9 (37.5) |
10 (41.7) |
19 (79.2) |
0.006 |
|  PDGF-C |
9 (37.5) |
13 (54.2) |
16 (66.7) |
0.13 |
|  Survivin |
9 (37.5) |
12 (50.0) |
17 (70.8) |
0.07 |
|  CA2 |
10 (41.7) |
10 (41.7) |
18 (75.0) |
0.03 |
|  IL-9 |
7 (29.2) |
16 (66.7) |
15 (62.5) |
0.02 |
|  EGF |
9 (37.5) |
11 (45.8) |
18 (75.0) |
0.02 |
|  Neuregulin-1 |
9 (37.5) |
11 (45.8) |
18 (75.0) |
0.02 |
|
Proteins with lower serum levels in the recovery group
|
|  YKL-40 |
15 (62.5) |
14 (58.3) |
9 (37.5) |
0.18 |
|  CXCL16, soluble |
18 (75.0) |
13 (54.2) |
7 (29.2) |
0.006 |
|  NGAL |
17 (70.8) |
11 (45.8) |
10 (41.7) |
0.09 |
|  IL1RL1 |
17 (70.8) |
12 (50.0) |
9 (37.5) |
0.07 |
|  Stanniocalcin-1 |
19 (79.2) |
12 (50.0) |
7 (29.2) |
0.002 |
|  IL-6 |
16 (66.7) |
14 (58.3) |
8 (33.3) |
0.06 |
|  CA3 |
18 (75.0) |
11 (45.8) |
9 (37.5) |
0.02 |
|  FGF23 |
17 (70.8) |
11 (45.8) |
10 (41.7) |
0.09 |
|  CK-MB |
15 (62.5) |
15 (62.5) |
8 (33.3) |
0.07 |
FYN, tyrosine-protein kinase Fyn; CXCL11, C-X-C motif chemokine 11 (I-TAC); CXCL2/CXCL3, C-X-C motif chemokine 2/C-X-C motif chemokine 3 (Gro-beta/gamma); CD86, T-lymphocyte activation antigen CD86 (activation B7–2 antigen); WNT7A, protein Wnt-7a; BTK, tyrosine-protein kinase BTK; c-Myc, Myc proto-oncogene protein; TIMP-3, metalloproteinase inhibitor 3; CCL5, C-C motif chemokine 5 (T-cell-specific protein RANTES); PDGF-C, platelet-derived growth factor C; CA2, carbonic anhydrase 2; IL-9, interleukin-9; EGF, epidermal growth factor; YKL-40, chitinase-3-like protein 1; CXCL16, C-X-C motif chemokine 16; NGAL, neutrophil gelatinase-associated lipocalin; IL1RL1, interleukin-1 receptor-like 1; IL-6, interleukin-6; CA3, carbonic anhydrase 3; FGF23, fibroblast growth factor 23; CK-MB, creatine kinase-MB.
an=24 patients in each tertile for all of the proteins. Data are presented in the number of patients with recovered kidney function and corresponding percentage in parenthesis in the tertile of each protein.
Figure 3.: Correlation between the slow off-rate modified aptamers scan (SOMAscan) and Olink measured serum levels for each of CXCL11, IL-6, and fibroblast growth factor 23 (FGF23). The ellipse is for confidence coefficient of 95%, within which 95% of the samples (points) were enclosed. Shown is also the correlation fit line. The SOMAscan and Olink values were log transformed.
Association of Significantly Changed Proteins with Kidney Function Recovery
To establish a direct association between the serum levels of proteins that were significantly different in the two groups and the end points of recovered or nonrecovered kidney function, we stratified their serum levels by tertiles of top-changed proteins (i.e., Fyn, CXCL11, CXCL2/CXCL3, CD86, Wnt-7a, BTK, c-Myc, TIMP-3, CCL5, ghrelin, PDGF-C, survivin, CA2, IL-9, EGF, neuregulin-1, YKL-40, soluble CXCL16, NGAL, IL1RL1, stanniocalcin-1, IL-6, CA3, FGF23, and CK-MB) in day-8 samples. Higher serum levels for each of CXCL11, CXCL2/CXCL3, CD86, Wnt-7a, BTK, c-Myc, TIMP-3, ghrelin, CA2, IL-9, EGF, and neuregulin-1 on day 8 were associated with increased recovery rate of kidney function by day 28, with a statistical significance P<0.05 as tested by the chi-square analyses (Table 2). Most significantly, 88% of the patients in the highest tertile of c-Myc levels recovered kidney function, whereas only 38% of the patients in the lowest tertile recovered kidney function. In contrast, higher serum levels of soluble CXCL16, stanniocalcin-1, and CA3 on day 8 were significantly (P<0.05) associated with decreased recovery rate of kidney function by day 28. Similarly, higher serum levels of the AKI biomarkers YKL-40, NGAL, IL-6, and FGF23, and cardiac biomarker CK-MB were associated with decreased rate of kidney recovery, albeit with less statistical significance (Table 2).
To further explore the relationships of serum levels of proteins with kidney recovery, logistic regression analysis was performed. Patients with upper serum levels for each of CXCL11, CXCL2/CXCL3, CD86, Wnt-7a, BTK, c-Myc, TIMP-3, CCL5, ghrelin, PDGF-C, survivin, CA2, IL-9, EGF, and neuregulin-1 had greater odds of kidney recovery compared with the patients with lower levels of each corresponding protein, and these associations remained significant (P<0.05) after multivariable adjustment for age, sex, congestive heart failure, diabetes mellitus, CV-SOFA scores, intensive dialysis treatment, cause of AKI, baseline serum creatinine, day-8 urine volume, and estimated 60-day mortality risk (Table 3). However, logistic regression analysis of all these confounding factors per se did not show any of them were significantly associated with kidney recovery (Supplemental Table 5). The association of higher serum levels of FYN with higher chance of kidney recovery was less significant (P>0.05) (Supplemental Table 6). Higher c-Myc levels were significantly associated with kidney recovery, with an approximately 56-fold increase in the chance of recovery compared with lower c-Myc levels (odds ratio=55.79, 95% CI, 5.53 to 562.99, P=0.0007). In contrast, for the proteins with lower serum levels on day 8, patients with the upper serum levels for each of soluble CXCL16, IL1RL1, stanniocalcin-1, IL-6, and FGF23 were unlikely to show kidney recovery by day 28 compared with those with the lower corresponding protein levels, and these associations remained significant (P<0.05) after multivariable adjustment (Table 3). The associations of the higher serum levels with lower chance of kidney recovery for proteins YKL-40, NGAL, CA3, and CK-MB were less significant after multivariable adjustment (Supplemental Table 6).
Table 3. -
Odds ratios and 95% confidence intervals of day 8 serum proteins in tertiles significantly associated with kidney function recovery by day 28 as analyzed by logistic regression
| Protein |
Tertilea |
Median (25th, 75th percentile)b |
Univariate Analysis |
Multivariable Analysisc |
| Odds Ratio (95% Confidence Interval) |
P Value |
Odds Ratio (95% Confidence Interval) |
P Value |
| CXCL11 |
1st |
2532 (2060–3467) |
1 |
|
1 |
|
|
2nd |
8658 (7016–10,109) |
5.00 (1.45 to 17.27) |
0.01 |
11.51 (2.22 to 59.63) |
0.004 |
|
3rd |
18,586 (14,894–37,590) |
7.29 (2.03 to 26.10) |
0.002 |
11.31 (2.10 to 60.78) |
0.005 |
| CXCL2/CXCL3 |
1st |
459 (406–512) |
1 |
|
1 |
|
|
2nd |
695 (643–766) |
2.36 (0.74 to 7.60) |
0.15 |
1.93 (0.40 to 9.32) |
0.41 |
|
3rd |
1198 (952–1386) |
4.86 (1.43 to 16.50) |
0.01 |
8.27 (1.77 to 38.67) |
0.007 |
| CD86 |
1st |
831 (762–898) |
1 |
|
1 |
|
|
2nd |
1105 (1065–1127) |
1.41 (0.45 to 4.46) |
0.56 |
1.45 (0.37 to 5.78) |
0.60 |
|
3rd |
1539 (1365–1748) |
5.00 (1.45 to 17.27) |
0.01 |
6.74 (1.47 to 30.80) |
0.01 |
| WNT7A |
1st |
270 (246–294) |
1 |
|
1 |
|
|
2nd |
376 (342–412) |
0.84 (0.26 to 2.68) |
0.77 |
1.16 (0.25 to 5.39) |
0.85 |
|
3rd |
556 (500–609) |
5.32 (1.49 to 19.06) |
0.01 |
13.10 (2.40 to 71.52) |
0.003 |
| BTK |
1st |
758 (673–837) |
1 |
|
1 |
|
|
2nd |
1005 (958–1106) |
1.41 (0.45 to 4.46) |
0.56 |
0.92 (0.20 to 4.20) |
0.92 |
|
3rd |
1544 (1233–1838) |
5.00 (1.45 to 17.27) |
0.01 |
5.48 (1.19 to 25.21) |
0.03 |
| c-Myc |
1st |
638 (556–683) |
1 |
|
1 |
|
|
2nd |
863 (771–915) |
0.83 (0.26 to 2.72) |
0.76 |
2.45 (0.42 to 14.40) |
0.32 |
|
3rd |
1247 (1155–1604) |
11.67 (2.70 to 50.49) |
0.001 |
55.79 (5.53 to 562.99) |
0.0007 |
| TIMP-3 |
1st |
1046 (680–1381) |
1 |
|
1 |
|
|
2nd |
3017 (2377–3486) |
1.19 (0.37 to 3.79) |
0.77 |
1.15 (0.21 to 6.32) |
0.87 |
|
3rd |
7306 (5328–9175) |
6.33 (1.75 to 22.91) |
0.005 |
36.98 (4.31 to 317.19) |
0.001 |
| CCL5 |
1st |
3697 (2197–4689) |
1 |
|
1 |
|
|
2nd |
12,782 (10,444–16,950) |
1.67 (0.53 to 5.27) |
0.38 |
2.38 (0.56 to 10.04) |
0.24 |
|
3rd |
38,218 (28,916–49,428) |
4.05 (1.21 to 13.54) |
0.02 |
8.64 (1.64 to 45.42) |
0.01 |
| Ghrelin |
1st |
1519 (1401–1707) |
1 |
|
1 |
|
|
2nd |
2120 (2001–2216) |
1.19 (0.37 to 3.79) |
0.77 |
2.00 (0.47 to 8.50) |
0.35 |
|
3rd |
2852 (2506–3311) |
6.33 (1.75 to 22.91) |
0.005 |
11.70 (2.21 to 61.77) |
0.004 |
| PDGF-C |
1st |
484 (411–529) |
1 |
|
1 |
|
|
2nd |
648 (631–723) |
1.97 (0.62 to 6.24) |
0.25 |
3.24 (0.73 to 14.44) |
0.12 |
|
3rd |
1001 (896–1669) |
3.33 (1.02 to 10.90) |
0.05 |
5.12 (1.12 to 23.41) |
0.04 |
| Survivin |
1st |
1113 (1006–1220) |
1 |
|
1 |
|
|
2nd |
1515 (1342–1578) |
1.67 (0.53 to 5.27) |
0.38 |
1.33 (0.28 to 6.39) |
0.72 |
|
3rd |
2002 (1855–2286) |
4.05 (1.21 to 13.54) |
0.02 |
10.71 (1.94 to 59.27) |
0.007 |
| CA2 |
1st |
260 (247–271) |
1 |
|
1 |
|
|
2nd |
380 (318–431) |
1.00 (0.32 to 3.15) |
1.00 |
0.94 (0.24 to 3.75) |
0.93 |
|
3rd |
599 (559–664) |
4.20 (1.23 to 14.37) |
0.02 |
5.11 (1.14 to 22.98) |
0.03 |
| IL-9 |
1st |
388 (371–415) |
1 |
|
1 |
|
|
2nd |
571 (515–634) |
4.86 (1.43 to 16.50) |
0.01 |
14.59 (2.30 to 92.51) |
0.004 |
|
3rd |
798 (691–979) |
4.05 (1.21 to 13.54) |
0.02 |
10.38 (1.83 to 58.72) |
0.008 |
| EGF |
1st |
747 (655–866) |
1 |
|
1 |
|
|
2nd |
1114 (1035–1346) |
1.41 (0.45 to 4.46) |
0.56 |
1.46 (0.30 to 7.11) |
0.64 |
|
3rd |
1899 (1577–2356) |
5.00 (1.45 to 17.27) |
0.01 |
12.06 (2.20 to 66.12) |
0.004 |
| Neuregulin-1 |
1st |
253 (223–282) |
1 |
|
1 |
|
|
2nd |
360 (333–408) |
1.41 (0.45 to 4.46) |
0.56 |
3.43 (0.77 to 15.34) |
0.11 |
|
3rd |
552 (488–805) |
5.00 (1.45 to 17.27) |
0.01 |
7.92 (1.61 to 38.90) |
0.01 |
| CXCL16, soluble |
1st |
12,884 (10,813–15,447) |
1 |
|
1 |
|
|
2nd |
27,219 (21,844–29,434) |
0.39 (0.12 to 1.34) |
0.14 |
0.38 (0.09 to 1.69) |
0.20 |
|
3rd |
40,953 (35,772–48,512) |
0.14 (0.04 to 0.49) |
0.002 |
0.09 (0.02 to 0.44) |
0.003 |
| IL1RL1 |
1st |
10,425 (7549–12,137) |
1 |
|
1 |
|
|
2nd |
30,420 (21,322–38,962) |
0.41 (0.13 to 1.35) |
0.14 |
0.41 (0.09 to 1.81) |
0.24 |
|
3rd |
66,622 (52,872–95,325) |
0.25 (0.07 to 0.83) |
0.02 |
0.11 (0.02 to 0.66) |
0.02 |
| Stanniocalcin-1 |
1st |
2594 (1831–3082) |
1 |
|
1 |
|
|
2nd |
4166 (3851–5157) |
0.26 (0.07 to 0.94) |
0.04 |
0.13 (0.02 to 0.77) |
0.02 |
|
3rd |
7557 (6282–16,587) |
0.11 (0.03 to 0.41) |
0.001 |
0.04 (0.01 to 0.28) |
0.0009 |
| IL-6 |
1st |
891 (764–972) |
1 |
|
1 |
|
|
2nd |
1312 (1213–1498) |
0.70 (0.22 to 2.27) |
0.55 |
0.59 (0.14 to 2.49) |
0.47 |
|
3rd |
3024 (2214–4157) |
0.25 (0.08 to 0.83) |
0.02 |
0.22 (0.05 to 0.99) |
0.05 |
| FGF23 |
1st |
1107 (1022–1171) |
1 |
|
1 |
|
|
2nd |
1693 (1452–1971) |
0.35 (0.11 to 1.15) |
0.08 |
0.32 (0.07 to 1.49) |
0.15 |
|
3rd |
5067 (2968–9366) |
0.29 (0.09 to 0.97) |
0.05 |
0.16 (0.03 to 0.79) |
0.03 |
CXCL11, C-X-C motif chemokine 11 (I-TAC); CXCL2/CXCL3, C-X-C motif chemokine 2/C-X-C motif chemokine 3 (Gro-beta/gamma); CD86, T-lymphocyte activation antigen CD86 (activation B7-2 antigen); WNT7A, protein Wnt-7a; BTK, tyrosine-protein kinase BTK; c-Myc, Myc proto-oncogene protein; TIMP-3, metalloproteinase inhibitor 3; CCL5, C-C motif chemokine 5 (T-cell-specific protein RANTES); PDGF-C, platelet-derived growth factor C; CA2, carbonic anhydrase 2; IL-9, interleukin-9; EGF, epidermal growth factor; CXCL16, C-X-C motif chemokine 16; IL1RL1, interleukin-1 receptor-like 1; IL-6, interleukin-6; FGF23, fibroblast growth factor 23.
an=24 patients in each tertile for all of the proteins.
bSOMAscan relative fluorescence units.
cAge, sex, congestive heart failure, diabetes mellitus, cardiovascular sequential organ failure assessment ≥2, intensive dialysis treatment, cause of AKI, baseline serum creatinine, day 8 urine volume, and estimated 60-day mortality risk were adjusted in the multivariable logistic regression analysis. The low level of each protein (tertile 1) was considered as the reference.
Discussion
Using routinely available clinical data including baseline eGFR, preadmission hemoglobin level, chronic liver disease, and age as the predictors, Lee et al. recently developed a predictive model for 90-day kidney recovery after AKI-D. However, the model's modest discrimination limits its clinical utility, and further studies are required to develop better models (16). This study is the first to use SOMAscan to discover proteomic biomarkers of kidney recovery from patients with AKI-D to develop such models. The VA/National Institutes of Health ATN study enrolled the patients with AKI and started randomized RRT on day 1. Serum samples were collected on day 1 and/or day 8. A recent study for the ATN cohort demonstrated that approximately half of the patients died during the first 8 days in the hyperacute and acute phases of AKI (22). Most patients who survived to day 8 continued to survive, and some recovered kidney function and discontinued dialysis. Using SOMAscan proteomic assays, we recently identified mortality-associated biomarkers, including FGF23, tPA, matrix metalloproteinase-8, and soluble urokinase plasminogen activator receptor in 100 serum samples collected on day 1. We also found that high serum levels of FGF23, tPA, and IL-6 were associated with mortality in 107 samples collected on day 8 (15). Because patients on day 1 were at hyperacute and acute phases of AKI, day-1 samples were more suitable for analysis of biomarker of mortality. Most patients with AKI-D who survived through day 8 had some degree of kidney recovery; thus we focused on the identification of potential biomarkers in day-8 samples to predict kidney recovery by day 28 in this study. In this study, kidney function recovery was defined for the patients with AKI who survived and discontinued dialysis by day 28.
Of the proteins identified from the SOMAscan assay that showed the greatest change, CXCL11, IL-6, and FGF23 were also verified by the Olink assay. Although a strong correlation between the two assays was observed for CXCL11 and IL-6 (r>0.85), FGF23 showed moderate correlation (r=0.47). In our previous analysis, the correlation between SOMAscan and ELISA for FGF23 was also moderate (r=0.61, P<0.0001) (15). SOMAscan, Olink, and ELISA assays are all affinity based and measure apparent quantitative changes as discussed previously (23); however, the affinity reagents of each assay might recognize a different region/epitope of a protein. Blood sample processing and storage could also differentially affect individual proteins under different preanalytical conditions (23). However, under standard operating procedures, the effects could be uniform for the samples from the same study cohort. A study showed that affinity-based proteomic approaches did not reveal any systematic effect of storage period on biobanked samples stored over a period of 13 to 17 years (24). In contrast, clinical factors such as disease status (e.g., AKI) and treatment might change the stability and structures of individual proteins, which in turn affect protein quantification. All these factors could potentially result in discordance between different assays.
Previous studies demonstrated that reduced levels of kidney injury biomarkers could serve as potential biomarkers for predicting kidney recovery. Plasma and urine levels of NGAL were decreased in patients with AKI who had recovering kidney function (56–7). In this study, we found that lower serum levels of AKI biomarkers NGAL and YKL-40 were associated, to a certain degree, with increased kidney recovery in patients with AKI-D, albeit with less statistical significance. Importantly, it was found that the association of lower serum levels of AKI mortality biomarker IL-6 with increased kidney recovery was statistically significant (P=0.05). The results suggest that AKI biomarkers have the potential to be predictive biomarkers of kidney recovery. However, these biomarkers need to be verified in larger clinical studies.
Inflammatory response is associated with phases of acute kidney injury and repair (25,26). Reduced levels of the inflammatory biomarker IL-18 were found to be indicative of kidney recovery (27). In this study, we found that lower serum levels of inflammation-related proteins (FGF23, soluble CXCL16, IL1RL1, and IL-6) on day 8 were associated with increased kidney recovery by day 28 and demonstrated that lower levels of these proteins were more significantly associated with kidney recovery than AKI biomarkers YKL-40 and NGAL. We also found that higher serum levels of CXCL11, CXCL2/CXCL3, CD86, CCL5, and IL-9 were significantly associated with kidney recovery. Chemokine CXCL11, CXCL2/CXCL3, and CCL5 target T lymphocytes, endothelial cells, monocytes, and macrophages during wound healing through their receptors on these cells (28). CXCL11 is a ligand of CXCR3, which has been shown to play roles in protecting the kidney from ischemia reperfusion injury by recruiting CXCR3+ natural killer T cells (29,30) and by promoting re-epithelialization in wound tissue repair (31). CXCL2 and CXCL3 are ligands of CXCR2, which also play important roles in wound healing (32). It has been shown that CXCL2/3-driven macrophage-myofibroblast crosstalk promotes intestinal repair (33). Studies have shown that CD86 is required for a robust regulatory T-cell response during the recovery phase for lung tissue repair after influenza A virus clearance (34). Furthermore, IL-9 is required to promote tissue repair in the recovery phase of inflammatory lung (35) and inhibit early podocyte injury and progressive glomerulosclerosis (36). These findings suggest reduced inflammation and increased chemokines/cytokines responsible for recruiting tissue repair immune cells or protecting kidney cells from injury may be useful predictive biomarkers of kidney recovery, underlining potential mechanism(s) of renal regeneration and repair.
This study indicates that a large fraction of the proteins associated with kidney recovery in patients with AKI-D are related to cellular growth, survival, or proliferation. Higher serum levels of Wnt-7a and c-Myc on day 8 were associated with kidney recovery by day 28. Protein Wnt-7a is a ligand in the canonical Wnt/beta-catenin signaling pathway, which has been shown to be involved in renal tubular protection after AKI (37) and repair and regeneration after kidney injury (38). Wnt/β-catenin signaling regulates the expression of the key driver of angiogenesis, VEGF (3940–41). The recent Translational Research Investigating Biomarker Endpoints for AKI study found that high postoperative plasma levels of VEGF A isoform and placental growth factor were associated with reduced risk for AKI and mortality (14). The present finding of high levels of Wnt-7a could coincide with VEGF effects on angiogenesis and endothelial remodeling, thus biomarkers associated with vascular regeneration could be potential biomarkers of kidney recovery (42). c-Myc is a transcription factor that activates the expression of growth-related genes (43). It is also well established that c-Myc is the major transcription factor promoting VEGFA gene expression (44), illustrating its role in cell growth and angiogenesis.
Coincidentally, higher serum levels of BTK, ghrelin, PDGF-C, survivin, EGF, and neuregulin-1 were associated with kidney recovery by day 28. BTK is a tyrosine-protein kinase and plays roles in B-cell development. PDGF-C, EGF, and neuregulin-1 are growth factors, and both EGF and neuregulin-1 belong to the epidermal growth factor family. Delayed recovery from AKI was found in mice with specific deletion of the EGF receptor in renal proximal tubule epithelial cells (45). Ghrelin is expressed in kidney and other organ tissues (46); however, ghrelin improves kidney function in mice with ischemic acute kidney failure (47) and reduces kidney tissue damage in rats (48), probably due to its anti-inflammatory and antioxidant effects (49). Increased expression of survivin in kidney epithelial cells was induced by AKI, and kidney recovery was markedly delayed in mice with renal proximal tubule–specific deletion of survivin (50), suggesting survivin plays a direct role in kidney recovery. Taken together, low levels of inflammation-associated proteins (e.g., FGF23, CXCL6), high levels of chemokines/cytokines for inflammatory tissue repair, and high levels of Wnt-7a, BTK, c-Myc, ghrelin, PDGF-C, survivin, EGF, and neuregulin-1 are likely to play concerted roles in the coordination and regulation of angiogenesis, endothelial and epithelial remodeling, kidney cell regeneration in the transition from inflammation to kidney recovery, and thus may be used as potential biomarkers for the prediction of kidney recovery.
This study had several limitations. First, the availability of randomized day-8 serum samples with secondary data analyses from the clinical trial study population was limited. This restricted the development of a predictive model. However, we addressed these shortcomings by using multiple statistical methods to attenuate potential biases. Second, both SOMAscan and Olink assays measured relative protein level changes, but not absolute protein concentrations in serum, and not all of the proteins discovered in the SOMAscan assay could be verified using Olink assays as many of them were not included on the Olink panel. Third, this is a biomarker discovery study for kidney recovery and there is a lack of an independent cohort of patients for validation of these novel biomarker candidates. The results from this study are informative for further verification in a larger cohort of samples. Finally, we were unable to follow the trajectory of protein changes over the course of kidney recovery due to lack of samples collected at later time points. Further studies specifically designed for this purpose may be able to address this issue. However, the novel biomarkers of kidney recovery discovered in our study warrant further independent validation in patients with AKI-D.
In conclusion, this study used a high-throughput proteomic technology to identify potential novel biomarkers of kidney recovery and defined potential biologic pathways associated with kidney remodeling, repair, and regeneration after AKI-D. Biomarkers involved in the progressive biologic processes of reducing inflammation, recruiting tissue repair immune cells, releasing growth factors, activating epithelial and endothelial remodeling, and increasing kidney cell growth and regeneration may be important targets for future studies to evaluate kidney function recovery.
Disclosures
D. Choudhury reports receiving research funding from Abbvie, AstraZeneca, Bayer, Boehringer Ingelheim, CorMedix, Duke Clinical Research Institute, GlaxoSmithKline (GSK), Hope, Janssen Pharma, and Minolock; reports receiving honoraria from the Virginia Tech Carilion School of Medicine; reports being a scientific advisor or member of Salem Research Institute VA nonprofit research corporation advisory board member; and reports having other interests/relationships as Board member of Salem Research Institute, and Salem VA Medical Center. D. Portilla reports being a scientific advisor or membership of Kidney360. J. Daniels reports receiving postdoctoral support through the Oak Ridge Institute for Science and Education from the National Center for Toxicological Research/FDA. J. Ma reports being a scientific advisor or member of the Editorial Board of American Journal of Physiology – Renal Physiology, Data and Safety Monitoring Board Member for the Clinical Trials network, the National Institute on Drug Abuse, and Statistician on the Editorial Board for Frontiers in Molecular Psychiatry. L.-R. Yu reports being a scientific advisor or membership as Member of the Editorial Board of Journal of Proteomics. P. Palevsky reports having consultancy agreements with Janssen Research & Development; reports being a scientific advisor or membership as Member of the Editorial Board of the Journal of Intensive Care Medicine, President and Member of the Scientific Advisory Board of the National Kidney Foundation, Member of the Quality, Safety and Accountability Committee of the Renal Physicians Association, Chair of the Medical Review Board for the Quality Insights Renal Network 4, and Section Editor, Renal Failure of UpToDate. R. Beger reports being a scientific advisor or member as Editor for Metabolomics Society and Scientific Reports and Coordinating Committee Member of the metabolomics Quality Assurance and quality Control Consortium (mQACC). All remaining authors have nothing to disclose. The ATN Study was conducted by the ATN Investigators and supported by the Cooperative Studies program of the Department of VA Office of Research and Development as CSP 530 and by the NIDDK under interagency agreement Y1-DK-3508-01.
Funding
This study was supported by the National Center for Toxicological Research, US Food and Drug Administration, Jefferson, Arkansas (Protocol E07571.11, to L.-R. Yu); the National Institutes of Health grants DK75976 and R01DK 12262401A1 (to D. Portilla).
Acknowledgments
This article reflects the views of the authors and does not necessarily reflect those of the US FDA. Any mention of commercial products is for clarification only and is not intended as approval, endorsement, or recommendation. This manuscript was not prepared in collaboration with the ATN Study Investigators and does not necessarily reflect the opinions or views of the ATN Study, VA, or NIDDK.
Author Contributions
R. Beger, D. Portilla, and L.-R. Yu conceptualized the study; J. Daniels and L.-R. Yu were responsible for data curation; Z. Cao, J. Ma, and L.-R. Yu were responsible for formal analysis; D. Portilla, J. Sun, and L.-R. Yu, were responsible for funding acquisition; P. Palevsky, D. Portilla, and L.-R. Yu were responsible for the investigation; J. Daniels and L.-R. Yu were responsible for the methodology; D. Choudhury, P. Palevsky, and D. Portilla were responsible for the resources; R. Beger provided supervision; Z. Cao and L.-R. Yu were responsible for the validation and for visualization; L.-R. Yu wrote the original draft; R. Beger, Z. Cao, D. Choudhury, J. Daniels, J. Ma, P. Palevsky, L. Pence, D. Portilla, L. Schnackenberg, and J. Sun reviewed and edited the manuscript.
Supplemental Material
This article contains the following supplemental material online at http://kidney360.asnjournals.org/lookup/suppl/doi:10.10.34067/KID.0002642021/-/DCSupplemental.
Supplemental Table 1. Proteins with significant changes in serum levels (1.2-fold and P<0.05) on day 8 as measured by SOMAscan assays in the patients who recovered (R) kidney function compared with those who did not recover (NR) by day 28.
Supplemental Table 2. All of the 1305 proteins measured by SOMAscan assays for day 8 serum samples from the patients who recovered (R) kidney function compared with those who did not recover (NR) by day 28.
Supplemental Table 3. Significantly altered biological functions or diseases revealed from Ingenuity Pathway Analysis.
Supplemental Table 4. All of the 92 proteins measured by the Olink assay for day 8 serum samples from the patients who recovered (R) kidney function compared with those who did not recover (NR) by day 28.
Supplemental Table 5. Logistic regression of kidney recovery on the baseline covariates.
Supplemental Table 6. Odds ratios and 95% confidence intervals of day 8 serum proteins in tertiles not significantly associated with kidney function recovery by day 28 as analyzed by logistic regression.
References
1. Wald R, McArthur E, Adhikari NK, Bagshaw SM, Burns KE, Garg AX, Harel Z, Kitchlu A, Mazer CD, Nash DM, Scales DC, Silver SA, Ray JG, Friedrich JO: Changing incidence and outcomes following dialysis-requiring acute kidney injury among critically ill adults: A population-based cohort study. Am J Kidney Dis 65: 870–877, 2015
https://doi.org/10.1053/j.ajkd.2014.10.017
2. Uchino S, Kellum JA, Bellomo R, Doig GS, Morimatsu H, Morgera S, Schetz M, Tan I, Bouman C, Macedo E, Gibney N, Tolwani A, Ronco C; Beginning and Ending Supportive Therapy for the Kidney (BEST Kidney) Investigators: Acute renal failure in critically ill patients: A multinational, multicenter study. JAMA 294: 813–818, 2005
https://doi.org/10.1001/jama.294.7.813
3. Bagshaw SM, Laupland KB, Doig CJ, Mortis G, Fick GH, Mucenski M, Godinez-Luna T, Svenson LW, Rosenal T: Prognosis for long-term survival and renal recovery in critically ill patients with severe acute renal failure: A population-based study. Crit Care 9: R700–R709, 2005
https://doi.org/10.1186/cc3879
4. Gaião SM, Paiva JAOC: Biomarkers of renal recovery after acute kidney injury. Rev Bras Ter Intensiva 29: 373–381, 2017
https://doi.org/10.5935/0103-507X.20170051
5. Srisawat N, Wen X, Lee M, Kong L, Elder M, Carter M, Unruh M, Finkel K, Vijayan A, Ramkumar M, Paganini E, Singbartl K, Palevsky PM, Kellum JA: Urinary biomarkers and renal recovery in critically ill patients with renal support. Clin J Am Soc Nephrol 6: 1815–1823, 2011
https://doi.org/10.2215/CJN.11261210
6. Srisawat N, Murugan R, Lee M, Kong L, Carter M, Angus DC, Kellum JA; Genetic and Inflammatory Markers of Sepsis (GenIMS) Study Investigators: Plasma neutrophil gelatinase-associated lipocalin predicts recovery from acute kidney injury following community-acquired pneumonia. Kidney Int 80: 545–552, 2011
https://doi.org/10.1038/ki.2011.160
7. Dewitte A, Joannès-Boyau O, Sidobre C, Fleureau C, Bats ML, Derache P, Leuillet S, Ripoche J, Combe C, Ouattara A: Kinetic eGFR and novel AKI biomarkers to predict renal recovery. Clin J Am Soc Nephrol 10: 1900–1910, 2015
https://doi.org/10.2215/CJN.12651214
8. Pike F, Murugan R, Keener C, Palevsky PM, Vijayan A, Unruh M, Finkel K, Wen X, Kellum JA; Biological Markers for Recovery of Kidney (BioMaRK) Study Investigators: Biomarker enhanced risk prediction for adverse outcomes in critically ill patients receiving RRT. Clin J Am Soc Nephrol 10: 1332–1339, 2015
https://doi.org/10.2215/CJN.09911014
9. Murugan R, Wen X, Shah N, Lee M, Kong L, Pike F, Keener C, Unruh M, Finkel K, Vijayan A, Palevsky PM, Paganini E, Carter M, Elder M, Kellum JA; Biological Markers for Recovery of Kidney (BioMaRK) Study Investigators: Plasma inflammatory and apoptosis markers are associated with dialysis dependence and death among critically ill patients receiving renal replacement therapy. Nephrol Dial Transplant 29: 1854–1864, 2014
https://doi.org/10.1093/ndt/gfu051
10. Aregger F, Uehlinger DE, Witowski J, Brunisholz RA, Hunziker P, Frey FJ, Jörres A: Identification of IGFBP-7 by urinary proteomics as a novel prognostic marker in early acute kidney injury. Kidney Int 85: 909–919, 2014
https://doi.org/10.1038/ki.2013.363
11. Meersch M, Schmidt C, Van Aken H, Martens S, Rossaint J, Singbartl K, Görlich D, Kellum JA, Zarbock A: Urinary TIMP-2 and IGFBP7 as early biomarkers of acute kidney injury and renal recovery following cardiac surgery. PLoS One 9: e93460, 2014
https://doi.org/10.1371/journal.pone.0093460
12. Hsu CY, Chinchilli VM, Coca S, Devarajan P, Ghahramani N, Go AS, Hsu RK, Ikizler TA, Kaufman J, Liu KD, Parikh CR, Reeves WB, Wurfel M, Zappitelli M, Kimmel PL, Siew ED; ASSESS-AKI Investigators: Post-acute kidney injury proteinuria and subsequent kidney disease progression: The Assessment, Serial Evaluation, and Subsequent Sequelae in Acute Kidney Injury (ASSESS-AKI) Study. JAMA Intern Med 180: 402–410, 2020
https://doi.org/10.1001/jamainternmed.2019.6390
13. Lee BJ, Go AS, Parikh R, Leong TK, Tan TC, Walia S, Hsu RK, Liu KD, Hsu CY: Pre-admission proteinuria impacts risk of non-recovery after dialysis-requiring acute kidney injury. Kidney Int 93: 968–976, 2018
https://doi.org/10.1016/j.kint.2017.10.017
14. Mansour SG, Zhang WR, Moledina DG, Coca SG, Jia Y, Thiessen-Philbrook H, McArthur E, Inoue K, Koyner JL, Shlipak MG, Wilson FP, Garg AX, Ishibe S, Parikh CR; TRIBE-AKI Consortium: The association of angiogenesis markers with acute kidney injury and mortality after cardiac surgery. Am J Kidney Dis 74: 36–46, 2019
https://doi.org/10.1053/j.ajkd.2019.01.028
15. Yu LR, Sun J, Daniels JR, Cao Z, Schnackenberg L, Choudhury D, Palevsky PM, Ma JZ, Beger RD, Portilla D: Aptamer-based proteomics identifies mortality-associated serum biomarkers in dialysis-dependent AKI patients. Kidney Int Rep 3: 1202–1213, 2018
https://doi.org/10.1016/j.ekir.2018.04.012
16. Lee BJ, Hsu CY, Parikh R, McCulloch CE, Tan TC, Liu KD, Hsu RK, Pravoverov L, Zheng S, Go AS: Predicting renal recovery after dialysis-requiring acute kidney injury. Kidney Int Rep 4: 571–581, 2019
https://doi.org/10.1016/j.ekir.2019.01.015
17. Palevsky PM, O’Connor T, Zhang JH, Star RA, Smith MW: Design of the VA/NIH Acute Renal Failure Trial Network (ATN) Study: Intensive versus conventional renal support in acute renal failure. Clin Trials 2: 423–435, 2005
https://doi.org/10.1191/1740774505cn116oa
18. Palevsky PM, Zhang JH, O’Connor TZ, Chertow GM, Crowley ST, Choudhury D, Finkel K, Kellum JA, Paganini E, Schein RM, Smith MW, Swanson KM, Thompson BT, Vijayan A, Watnick S, Star RA, Peduzzi P; VA/NIH Acute Renal Failure Trial Network: Intensity of renal support in critically ill patients with acute kidney injury. N Engl J Med 359: 7–20, 2008
https://doi.org/10.1056/NEJMoa0802639
19. Gold L, Ayers D, Bertino J, Bock C, Bock A, Brody EN, Carter J, Dalby AB, Eaton BE, Fitzwater T, Flather D, Forbes A, Foreman T, Fowler C, Gawande B, Goss M, Gunn M, Gupta S, Halladay D, Heil J, Heilig J, Hicke B, Husar G, Janjic N, Jarvis T, Jennings S, Katilius E, Keeney TR, Kim N, Koch TH, Kraemer S, Kroiss L, Le N, Levine D, Lindsey W, Lollo B, Mayfield W, Mehan M, Mehler R, Nelson SK, Nelson M, Nieuwlandt D, Nikrad M, Ochsner U, Ostroff RM, Otis M, Parker T, Pietrasiewicz S, Resnicow DI, Rohloff J, Sanders G, Sattin S, Schneider D, Singer B, Stanton M, Sterkel A, Stewart A, Stratford S, Vaught JD, Vrkljan M, Walker JJ, Watrobka M, Waugh S, Weiss A, Wilcox SK, Wolfson A, Wolk SK, Zhang C, Zichi D: Aptamer-based multiplexed proteomic technology for biomarker discovery. PLoS One 5: e15004, 2010
https://doi.org/10.1371/journal.pone.0015004
20. Assarsson E, Lundberg M, Holmquist G, Björkesten J, Thorsen SB, Ekman D, Eriksson A, Rennel Dickens E, Ohlsson S, Edfeldt G, Andersson AC, Lindstedt P, Stenvang J, Gullberg M, Fredriksson S: Homogenous 96-plex PEA immunoassay exhibiting high sensitivity, specificity, and excellent scalability. PLoS One 9: e95192, 2014
https://doi.org/10.1371/journal.pone.0095192
21. Benjamini Y, Hochberg Y: Controlling the false discovery rate: A practical and powerful approach to multiple testing. J R Stat Soc Series B Stat Methodol 57: 289–300, 1995
https://doi.org/10.1111/j.2517-6161.1995.tb02031.x
22. Zhang JH, Palevsky PM, Chertow GM, Hartigan J, O’Connor TZ, Guarino P, Zhou B: Piecewise analysis of patient survival after onset of AKI. Clin J Am Soc Nephrol 8: 1679–1684, 2013
https://doi.org/10.2215/CJN.07250712
23. Daniels JR, Cao Z, Maisha M, Schnackenberg LK, Sun J, Pence L, Schmitt TC, Kamlage B, Rogstad S, Beger RD, Yu LR: Stability of the human plasma proteome to pre-analytical variability as assessed by an aptamer-based approach. J Proteome Res 18: 3661–3670, 2019
https://doi.org/10.1021/acs.jproteome.9b00320
24. Hebels DG, Georgiadis P, Keun HC, Athersuch TJ, Vineis P, Vermeulen R, Portengen L, Bergdahl IA, Hallmans G, Palli D, Bendinelli B, Krogh V, Tumino R, Sacerdote C, Panico S, Kleinjans JC, de Kok TM, Smith MT, Kyrtopoulos SA; EnviroGenomarkers Project Consortium: Performance in omics analyses of blood samples in long-term storage: Opportunities for the exploitation of existing biobanks in environmental health research. Environ Health Perspect 121: 480–487, 2013
https://doi.org/10.1289/ehp.1205657
25. Okusa MD, Chertow GM, Portilla D; Acute Kidney Injury Advisory Group of the American Society of Nephrology: The nexus of acute kidney injury, chronic kidney disease, and World Kidney Day 2009. Clin J Am Soc Nephrol 4: 520–522, 2009
https://doi.org/10.2215/CJN.06711208
26. Mulay SR, Holderied A, Kumar SV, Anders HJ: Targeting inflammation in so-called acute kidney injury. Semin Nephrol 36: 17–30, 2016
https://doi.org/10.1016/j.semnephrol.2016.01.006
27. Hall IE, Yarlagadda SG, Coca SG, Wang Z, Doshi M, Devarajan P, Han WK, Marcus RJ, Parikh CR: IL-18 and urinary NGAL predict dialysis and graft recovery after kidney transplantation. J Am Soc Nephrol 21: 189–197, 2010
https://doi.org/10.1681/ASN.2009030264
28. Balaji S, Watson CL, Ranjan R, King A, Bollyky PL, Keswani SG: Chemokine involvement in fetal and adult wound healing. Adv Wound Care (New Rochelle) 4: 660–672, 2015
https://doi.org/10.1089/wound.2014.0564
29. Yang SH, Lee JP, Jang HR, Cha RH, Han SS, Jeon US, Kim DK, Song J, Lee DS, Kim YS: Sulfatide-reactive natural killer T cells abrogate ischemia-reperfusion injury. J Am Soc Nephrol 22: 1305–1314, 2011
https://doi.org/10.1681/ASN.2010080815
30. Zhang C, Zheng L, Li L, Wang L, Li L, Huang S, Gu C, Zhang L, Yang C, Zhu T, Rong R: Rapamycin protects kidney against ischemia reperfusion injury through recruitment of NKT cells. J Transl Med 12: 224, 2014
https://doi.org/10.1186/s12967-014-0224-z
31. Yates CC, Whaley D, Y-Chen A, Kulesekaran P, Hebda PA, Wells A: ELR-negative CXC chemokine CXCL11 (IP-9/I-TAC) facilitates dermal and epidermal maturation during wound repair. Am J Pathol 173: 643–652, 2008
https://doi.org/10.2353/ajpath.2008.070990
32. Zaja-Milatovic S, Richmond A: CXC chemokines and their receptors: A case for a significant biological role in cutaneous wound healing. Histol Histopathol 23: 1399–1407, 2008
33. Esser-von Bieren J, Volpe B, Sutherland DB, Bürgi J, Verbeek JS, Marsland BJ, Urban JF Jr, Harris NL: Immune antibodies and helminth products drive CXCR2-dependent macrophage- myofibroblast crosstalk to promote intestinal repair. PLoS Pathog 11: e1004778, 2015
https://doi.org/10.1371/journal.ppat.1004778
34. Moser EK, Hufford MM, Braciale TJ: Late engagement of CD86 after influenza virus clearance promotes recovery in a FoxP3+ regulatory T cell dependent manner. PLoS Pathog 10: e1004315, 2014
https://doi.org/10.1371/journal.ppat.1004315
35. Turner JE, Morrison PJ, Wilhelm C, Wilson M, Ahlfors H, Renauld JC, Panzer U, Helmby H, Stockinger B: IL-9-mediated survival of type 2 innate lymphoid cells promotes damage control in helminth-induced lung inflammation. J Exp Med 210: 2951–2965, 2013
https://doi.org/10.1084/jem.20130071
36. Xiong T, Attar M, Gnirck AC, Wunderlich M, Becker M, Rickassel C, Puelles VG, Meyer-Schwesinger C, Wiech T, Nies JF, Divivier M, Fuchs T, Schulze Zur Wiesch J, Taipaleenmäki H, Hoxha E, Wirtz S, Huber TB, Panzer U, Turner JE: Interleukin-9 protects from early podocyte injury and progressive glomerulosclerosis in Adriamycin-induced nephropathy. Kidney Int 98: 615–629, 2020
https://doi.org/10.1016/j.kint.2020.04.036
37. Zhou D, Li Y, Lin L, Zhou L, Igarashi P, Liu Y: Tubule-specific ablation of endogenous β-catenin aggravates acute kidney injury in mice. Kidney Int 82: 537–547, 2012
https://doi.org/10.1038/ki.2012.173
38. Lin SL, Li B, Rao S, Yeo EJ, Hudson TE, Nowlin BT, Pei H, Chen L, Zheng JJ, Carroll TJ, Pollard JW, McMahon AP, Lang RA, Duffield JS: Macrophage Wnt7b is critical for kidney repair and regeneration. Proc Natl Acad Sci U S A 107: 4194–4199, 2010
https://doi.org/10.1073/pnas.0912228107
39. Olsen JJ, Pohl SO, Deshmukh A, Visweswaran M, Ward NC, Arfuso F, Agostino M, Dharmarajan A: The role of Wnt signalling in angiogenesis. Clin Biochem Rev 38: 131–142, 2017
40. Lévy L, Neuveut C, Renard CA, Charneau P, Branchereau S, Gauthier F, Van Nhieu JT, Cherqui D, Petit-Bertron AF, Mathieu D, Buendia MA: Transcriptional activation of interleukin-8 by beta-catenin-Tcf4. J Biol Chem 277: 42386–42393, 2002
https://doi.org/10.1074/jbc.M207418200
41. Easwaran V, Lee SH, Inge L, Guo L, Goldbeck C, Garrett E, Wiesmann M, Garcia PD, Fuller JH, Chan V, Randazzo F, Gundel R, Warren RS, Escobedo J, Aukerman SL, Taylor RN, Fantl WJ: beta-Catenin regulates vascular endothelial growth factor expression in colon cancer. Cancer Res 63: 3145–3153, 2003
42. Bouchard J, Mehta RL: Angiogenesis markers and recovery from acute kidney injury: A piece of the puzzle? Am J Kidney Dis 74: 12–14, 2019
https://doi.org/10.1053/j.ajkd.2019.03.421
43. Pelengaris S, Khan M, Evan G: c-MYC: More than just a matter of life and death. Nat Rev Cancer 2: 764–776, 2002
https://doi.org/10.1038/nrc904
44. Shi Y, Xu X, Zhang Q, Fu G, Mo Z, Wang GS, Kishi S, Yang XL: tRNA synthetase counteracts c-Myc to develop functional vasculature. eLife 3: e02349, 2014
https://doi.org/10.7554/eLife.02349
45. Chen J, Chen JK, Harris RC: Deletion of the epidermal growth factor receptor in renal proximal tubule epithelial cells delays recovery from acute kidney injury. Kidney Int 82: 45–52, 2012
https://doi.org/10.1038/ki.2012.43
46. Mori K, Yoshimoto A, Takaya K, Hosoda K, Ariyasu H, Yahata K, Mukoyama M, Sugawara A, Hosoda H, Kojima M, Kangawa K, Nakao K: Kidney produces a novel acylated peptide, ghrelin. FEBS Lett 486: 213–216, 2000
https://doi.org/10.1016/S0014-5793(00)02308-5
47. Takeda R, Nishimatsu H, Suzuki E, Satonaka H, Nagata D, Oba S, Sata M, Takahashi M, Yamamoto Y, Terauchi Y, Kadowaki T, Kangawa K, Kitamura T, Nagai R, Hirata Y: Ghrelin improves renal function in mice with ischemic acute renal failure. J Am Soc Nephrol 17: 113–121, 2006
https://doi.org/10.1681/ASN.2004080626
48. Çimen S, Taşdemir C, Vardı N, Ateş B, Taşdemir S, Özaydoğdu Çimen A: Protective effects of ghrelin on kidney tissue in rats with partial ureteral obstruction. Turk J Med Sci 49: 696–702, 2019
https://doi.org/10.3906/sag-1802-17
49. Baatar D, Patel K, Taub DD: The effects of ghrelin on inflammation and the immune system. Mol Cell Endocrinol 340: 44–58, 2011
https://doi.org/10.1016/j.mce.2011.04.019
50. Chen J, Chen JK, Conway EM, Harris RC: Survivin mediates renal proximal tubule recovery from AKI. J Am Soc Nephrol 24: 2023–2033, 2013
https://doi.org/10.1681/ASN.2013010076
