Urinary Prostaglandin E₂ is a biomarker of early adaptive hyperfiltration in solitary functioning kidney

Abstract

Introduction

Hyperfiltration is a major contributor to progression of chronic kidney disease (CKD) in diabetes, obesity and in individuals with solitary functioning kidney (SFK). We have proposed hyperfiltration-induced injury as a continuum of overlapping glomerular changes caused by increased biomechanical forces namely, fluid flow shear stress (FFSS) and tensile stress. We have shown that FFSS is elevated in animals with SFK and, it upregulates prostaglandin E₂ (PGE₂), cyclooxygenase-2 and PGE₂ receptor EP2 in cultured podocytes and in uninephrectomized mice. We conceptualized urinary PGE₂ as a biomarker of early effects of hyperfiltration-induced injury preceding microalbuminuria in individuals with SFK. We studied children with SFK to validate our hypothesis.

Methods

Urine samples from children with SFK and controls were analyzed for PGE₂, albumin (glomerular injury biomarker) and epidermal growth factor (EGF, tubular injury biomarker). Age, gender, and Z-scores for height, weight, BMI, and blood pressure were obtained.

Results

Children with SFK were comparable to controls except for lower BMI Z-scores. The median values were elevated in SFK compared to control for urine PGE₂ [9.1 (n=57) vs. 5.7 (n=72), p=0.009] ng/mgCr and albumin [7.6 (n=40) vs. 7.0 (n=41), p=0.085] μg/mgCr, but not for EGF [20098 (n=44) vs. 18637 (n=44), p=0.746] pg/mgCr. Significant increase in urinary PGE₂ (p=0.024) and albumin (p=0.019) but not EGF (p=0.412) was observed using additional regression modeling. These three urinary analytes were independent of each other.

Conclusion

Increased urinary PGE₂ from elevated SNGFR and consequently increased FFSS during early stage of CKD precedes overt microalbuminuria and is a biomarker for early hyperfiltration-induced injury in individuals with SFK.

INTRODUCTION

Living kidney donors (LKD) specially, young donors (<35 years) face a greater risk of end-stage renal disease (ESRD) compared to healthy non-donors [1–6]. In contrast to conventional view, an estimated 20%−40% of children born with a solitary functioning kidney (SFK) are known to develop ESRD as young adults [7–9]. In both these populations i.e., children with congenital SFK and adults with acquired SFK, progression to ESRD is attributed to hyperfiltration-mediated injury [10,11]. Over time, adaptive hyperfiltration in the remaining functional nephrons turns maladaptive, and results in loss of glomerular barrier function, proteinuria, and progression to chronic kidney disease (CKD)/ ESRD. Classical indicators of renal dysfunction such as hypertension, microalbuminuria, proteinuria, and CKD (eGFR<60 ml/min/1.73m²) in individuals with SFK start to appear only after ~15 years [4–9]. Therefore, identifying a biomarker of early hyperfiltration-mediated injury prior to the development of overt albuminuria would be beneficial for both children and adults.

Biomechanical forces have been recently highlighted as mediators of the effects of hyperfiltration on glomerular structure and function [10–14]. Podocytes are especially vulnerable to hyperfiltration-associated increase in biomechanical forces [11]. Prior to overt CKD, hemodynamic adaptation in remaining functional nephrons leads to higher single nephron glomerular filtration rate (SNGFR) with increased ultrafiltrate flow causing fluid flow shear stress (FFSS) to podocytes. Additionally, tensile stress on podocyte foot processes increases due to radial stretch of the capillary wall consequent to increased glomerular capillary pressure [11–13]. We have reported a 1.5- to 2-fold increase in FFSS, which is largely due to increased SNGFR following unilateral nephrectomy in rodents [15]. We have proposed a model to study the significance of FFSS and tensile stress in hyperfiltration-mediated hemodynamic changes. In this model, FFSS and tensile stress cause podocyte injury sequentially and synergistically in a continuum as GFR declines from 100% to 0% (Figure 1).

Figure 1.

Hyperfiltration-mediated hemodynamic changes may be conceptualized as a continuum. In most kidney diseases, a gradual loss of functional nephrons over time leads to hyperfiltration in remaining glomeruli. Adaptive changes initially appear in the form of increased single nephron glomerular filtration rate (SNGFR) from changes in renal blood flow and ultrafiltration coefficient (Kf). Over time, progressive loss of functional nephrons results in glomerular hypertension due to increased glomerular capillary pressure (P_GC). We propose that the cyclooxygenase 2 enzyme - prostaglandin E₂ - prostanoid receptor EP2 (COX2-PGE₂-EP2) axis and components of the renin-angiotensin-aldosterone system (RAAS) are relevant in the early and late stages of the hyperfiltration continuum, respectively, (adapted with permission from Reference [10]).

FFSS applied to cultured podocytes in vitro as well as increased FFSS following unilateral nephrectomy of mice resulted in increased production of Prostaglandin E₂ (PGE₂), upregulation of COX2 and induction of PGE₂ receptor EP2 [16–18]. FFSS caused increased glomerular albumin permeability in vitro, an effect that was mimicked by PGE₂ and inhibited by indomethacin [18].

Urinary albumin and epidermal growth factor (EGF) are known biomarkers of glomerular and tubular injury, respectively. Incremental rise in urinary albumin within the reference range with gradual rise to microalbuminuria of >30 μg/mgCr and subsequently to overt proteinuria is considered a benchmark indicator of glomerular dysfunction in CKD. Early glomerular changes, prior to albuminuria, are indicated by increased glomerular albumin permeability in vitro in isolated glomeruli in rat models of puromycin-induced nephrosis (PAN), radiation nephropathy, and type 2 diabetes (obese Zucker rats), but similar studies are not possible in humans [19–21]. Distinct tubular changes are indicated by EGF levels which is highly expressed in the thick ascending limb of Henle and distal tubules [22–24]. A decrease in urine EGF has been validated as an excellent biomarker for progression of CKD of different etiologies in adults and children [24–30].

As mentioned, increased amounts of urinary PGE₂ and albumin are known in CKD, but the timeline of these changes is not known. The slow progression of CKD in hyperfiltration-mediated injury in individuals with SFK provides an opportunity to assess the early events (prior to the onset of overt albuminuria) by establishing a temporal distinction in the observed increase in urinary PGE₂ and albumin. We postulated that early glomerular changes in SFK due to hyperfiltration-associated increase in FFSS result in elevated levels of urinary PGE₂.

MATERIAL AND METHODS

Study subjects

Children (1–18 years) who received clinical care for SFK (unilateral renal agenesis or unilateral multicystic dysplastic kidney) or for monosymptomatic enuresis at CMH were eligible for this study. Children on dialysis, with kidney transplant, with diabetes, with voiding dysfunction, or on ACEIs (angiotensin converting enzyme inhibitors), ARBs (angiotensin receptor blockers) or NSAIDs (non-steroidal anti-inflammatory drugs) medications were excluded. Age, gender, height, weight and blood pressure (BP) data and scavenged urine sample were collected during the clinic visit. BP measurements were performed on automated oscillometric device (Dinamap™), and elevated BP was confirmed by manual auscultation. Systolic (SBP) and diastolic (DBP) BP Z-scores were calculated based on the 2017 Clinical Practice Guideline for Screening and Management of High Blood Pressure in Children and Adolescents [31]. Z-scores were calculated to correct for height, weight, body mass index (BMI) and BP for age and gender.

Urine sample collection

Spot urine samples were divided into 5–10 mL aliquots and refrigerated (4°C) immediately for urine sediment examination in a blinded manner. Urine samples that met the inclusion/exclusion criteria were scavenged for PGE₂, albumin, EGF and creatinine assays. Specimens were centrifuged (5000 rpm/min for 10 min, 4°C). Clear supernatants were transferred to new tubes and stored immediately at −80°C.

Urine Prostaglandin E₂ measurement by liquid chromatography-tandem mass spectrometry (LC-MS/MS)

Previously frozen urine samples were analyzed for PGE₂ metabolite 11-α-hydroxy-9,15-dioxo-2,3,4,5-tetranor-prostane-1,20-dioic acid (PGE-M) at the Eicosanoid Core Laboratory, Vanderbilt University Medical Center, Nashville TN. Briefly, urine (1mL) was acidified to pH 3 with hydrochloric acid and treated with methyloxime hydrochloride to convert analytes to the O-methyloxime derivative, passed through a C-18 Sep-Pak (Waters Corp. Milford, MA USA) and eluted with ethyl acetate [32]. A deuterated internal standard, [²H₆]-O-methyloxime PGE-M, was added for PGE-M quantification. Samples were dried (37°C) under dry nitrogen and reconstituted in 75μL mobile phase A for LC-MS/MS analysis. LC was performed on a 2.0 × 50 mm, 1.7μm particle Acquity BEH C18 column (Waters Corporation, Milford, MA, USA) using a Waters Acquity UPLC. Mobile phase A was 95:4.9:0.1 (v/v/v) 5 mM ammonium acetate:acetonitrile:acetic acid, and mobile phase B was 10.0:89.9:0.1 (v/v/v) 5 mM ammonium acetate:acetonitrile:acetic acid. Fractions separated using a gradient of 85–5% of mobile phase A at (375μl/min, 14min) were analyzed using a SCIEX 6500+ QTrap mass spectrometer. One representative quality control (QC) urine sample was analyzed with each batch of 18 samples. The QC urine samples were derived from multiple 24-hour urine collections that were pooled, thoroughly mixed, and separated into 1 mL aliquots before freezing at −80°C. Metabolite quantities were normalized by creatinine in respective sample and expressed as ng/mg creatinine.

Urine albumin and creatinine measurement

Urine albumin and creatinine were quantitated using the VITROS Chemistry analyzer (Ortho Clinical Diagnostics, Raritan, NJ). Albumin involves formation of an antibody/antigen complex that increases solution turbidity proportionately which is measured spectrophotometrically (340 nm) and used for determining albumin concentration from a calibration curve.

Urine creatinine was measured enzymatically using an assay compliant with the NKF-KDOQI recommendations. Briefly, creatinine is sequentially converted to creatine (by hydrolysis), to sarcosine by creatine amidinohydrolase, then to glycine, formaldehyde, and hydrogen peroxide by sarcosine oxidase. Finally, a peroxidase-catalyzed oxidation of leuco dye produces a colored product proportional to creatinine concentration.

Urine epidermal growth factor measurement

The Human EGF Quantikine ELISA kit (DEG00; R&D Systems, Minneapolis, MN) was used to measure EGF concentration following the manufacturer’s instructions. Three pre-measured quality control samples with high, medium, or low concentrations of urine EGF were included in each plate to determine inter-plate variations. All samples, quality controls, and standards were tested in duplicates.

Statistical analyses

The raw data were examined for normality, and then log-transformed for additional analysis. The data were presented as median with interquartile range. Wilcoxon-Mann-Whitney tests for unadjusted differences between the two groups were carried out as the data were not normally distributed. Log₂-transformation of values for PGE₂, albumin, and EGF reduced the skewness and the data were suitably normal for analysis. Differences in distribution between children in the control and SFK groups were examined using boxplots, followed by fitting linear regression models to test for differences between the two groups after adjustment for age, gender, weight z-score, height z-score, DBP z-score, and SBP z-score. Associations among three urine analytes, and between each analyte and age, gender, weight z-score, height z-score, DBP z-score, and SBP z-score, were determined by Spearman correlations. Correlations were computed separately for the control and SFK groups to allow for comparisons. Statistical analyses were carried out using SAS 9.4.

RESULTS

Demographics of the patient population

Urine PGE₂, albumin and EGF were measured in scavenged urine samples following an outpatient visit. Urine PGE₂ were available from 74 (51 boys and 23 girls) children in the control group and 60 (40 boys and 20 girls, χ² p=0.781) children with SFK. The SFK group included 35 (58%) children born with unilateral renal agenesis and 25 (42%) with multicystic dysplastic kidney. Urine albumin was available from 41 control children (29 boys and 12 girls) and 40 children (27 boys and 13 girls, χ² p=0.753) with SFK. The SFK group included 21 (53%) children born with unilateral renal agenesis and 19 (47%) with multicystic dysplastic kidney. Urine EGF was available from 44 control children (30 boys and 14 girls) and 44 children (29 boys and 15 girls, χ² p=0.821) with SFK. The SFK group included 23 (52%) children born with unilateral renal agenesis and 21 (48%) with multicystic dysplastic kidney. Table 1 shows the median and interquartile range for age, weight, height, BMI, systolic BP, diastolic BP, and their Z-scores in control and the SFK groups in whom PGE₂, albumin and EGF were measured. Children with SFK had significantly lower BMI and weight Z-scores compared to control children.

Table 1.

The table shows the median (inter-quartile range) of age, weight, height, body mass index (BMI), systolic BP (SBP), diastolic BP (DBP), and their Z-scores in control children and children with solitary functioning kidney (SFK) in whom urine prostaglandin E₂ (PGE₂), albumin and epidermal growth factor (EGF) were measured. Wilcoxon-Mann-Whitney test was used for comparison of the two groups. The weight, BMI and BMI Z-Score were lower in children with SFK compared to the control children.

		PGE2				Albumin				EGF
	Control (n=72)	SFK (n=57)	pvalue		Control (n=41)	SFK (n=40)	pvalue		Control (n=44)	SFK (n=44)	pvalue
Age (years)	10.8 (9.1, 12.4)	8.5 (5.2, 14.2)	0.180		10.5 (8.4, 12.6)	9.3 (5.2, 13.4)	0.247		10.5 (8.4, 12.6)	9.3 (5.2, 13.4)	0.095
Weight (kg)	42.0 (30.5, 57.5)	31.1 (20.6, 50.5)	0.018		42.2 (29.6, 55.1)	30.2 (20.6, 50.5)	0.028		42.2 (29.6, 55.1)	30.2 (20.6, 50.5)	0.006
Height (cm)	144.7 (133.4, 155.4)	134.8 (113.7, 159.3)	0.204		142.2 (131.3, 155.4)	136.5 (113.7, 157.3)	0.303		142.2 (131.3, 155.4)	136.5 (113.7, 157.3)	0.105
BMI (kg/m2)	20.0 (16.4, 25)	17.1 (15.9, 20.5)	0.005		20.2 (16.7, 26.2)	16.9 (15.6, 20.3)	0.003		20.2 (16.7, 26.2)	16.9 (15.6, 20.3)	0.001
SBP (mmHg)	108.0 (102.0, 117.0)	108.5 (102.8, 117.2)	0.827		108.0 (102.0, 115.0)	107.5 (103.0, 116.2)	0.780		108 (102, 115)	107.5 (103, 116.2)	0.462
DBP (mmHg)	63.0 (58.0, 67.0)	64.0 (56.8, 70.0)	0.692		64.0 (59.0, 67.0)	62.5 (56.0, 71.0)	0.469		64 (59, 67)	62.5 (56, 71)	0.475
Weight Z-Score	0.8 (−0.1, 2.0)	0.7 (−0.1, 1.1)	0.226		0.8 (0.0, 2.1)	0.7 (−0.1, 1.1)	0.077		0.8 (0, 2.1)	0.7 (−0.1, 1.1)	0.055
Height Z-Score	0.3 (−0.4, 1.2)	0.6 (−0.1, 1.5)	0.181		0.3 (−0.6, 1.1)	0.6 (−0.2, 1.3)	0.380		0.3 (−0.6, 1.1)	0.6 (−0.2, 1.3)	0.411
BMI Z-Score	1.0 (−0.1, 2.0)	0.2 (−0.3, 1.1)	0.012		1 (0.4, 2.1)	0.2 (−0.4, 1.0)	0.002		1 (0.4, 2.1)	0.2 (−0.4, 1)	0.001
SBP Z-score	0.4 (−0.1, 1.1)	0.7 (0.0, 1.2)	0.370		0.5 (0.0, 1.1)	0.7 (0, 1.3)	0.734		0.5 (0, 1.1)	0.7 (0, 1.3)	0.643
DBP Z-score	0.1 (−0.3, 0.5)	0.4 (−0.1, 1.0)	0.046		0.2 (−0.1, 0.7)	0.4 (−0.1, 1.0)	0.564		0.2 (−0.1, 0.7)	0.4 (−0.1, 1)	0.337

Urinary Prostaglandin E₂, Albumin and Epidermal Growth Factor

The median and inter-quartile ranges for urine PGE₂, albumin and EGF in control and children with SFK are shown in Table 2. The distribution of data for log transformed urine PGE₂, albumin and EGF are shown in Figure 2. Urinary PGE₂ and albumin were clearly elevated in children with SFK compared to controls (Figures 2(A) and (B)). Geometric means of 6.3 vs. 9.0 on urine PGE₂ and of 7.2 vs. 10.7 on urine albumin were observed in control and children with SFK, respectively. The Wilcoxon-Mann-Whitney tests provided additional evidence for differences on these urinary analytes, despite being unadjusted for patient characteristics on which the two groups differed (Table 2). In contrast, distribution of EGF in control and children with SFK were comparable (Table 2; Figure 2(C)).

Table 2.

Median and inter-quartile range (IQR) for urine prostaglandin E₂ (PGE₂), albumin, and epidermal growth factor (EGF) in control children and children with solitary functioning kidney (SFK) are shown in the table. Wilcoxon-Mann-Whitney test was used for comparison of the two groups.

	Control		SFK
	Median (IQR)	N	Median (IQR)	n	p-value
Urine PGE2 (ng/mgCr)	5.7 (4.0, 8.8)	72	9.1 (4.4, 16.7)	57	0.009
Urine Albumin (μg/mgCr)	7.0 (4.0, 10.3)	41	7.6 (4.7, 24.0)	40	0.085
Urine EGF (pg/mgCr)	18637 (15298, 25622)	44	20098 (13238, 30263)	44	0.746

Figure 2.

Distribution of log₂Urine PGE₂) on the left (A), log₂Urine Albumin in the middle (B) and log₂Urine EGF in control children and children with solitary functioning kidney (SFK).

Additional analysis provided strong evidence for differences in urine PGE₂ and urine albumin when age, gender, weight z-score, height z-score, DBP z-score, and SBP z-score were controlled for regression modeling (Table 3). According to model estimates, log₂Urine PGE₂ and log₂Urine albumin levels are, respectively, 0.41 and 0.55 standard deviations higher in children with SFK than in control (p = 0.024 and 0.019, respectively). By comparison, log₂Urine EGF levels were an estimated 0.19 standard deviations higher in children with SFK (p = 0.412).

Table 3.

Results from regression modeling using SAS 9.4. Controlling for model covariates, children with solitary functioning kidney (SFK) have substantially higher levels of urine PGE₂ and urine albumin than control patients; there was no evidence of decreased EGF in children with SFK.

Outcome	Explanatory variable	Beta	95% limits for Beta	p-value
Log2Urine PGE2	SFK (vs. control)	0.41	0.06, 0.77	0.024
	Female (vs. male)	0.27	−0.10, 0.63	0.155
	Age (years)	−0.06	−0.10, −0.02	0.008
	Weight z-score	0.07	−0.11, 0.25	0.446
	Height z-score	−0.21	−0.40, −0.01	0.039
	DBP z-score	0.04	−0.23, 0.31	0.749
	SBP z-score	−0.12	−0.34, 0.11	0.299
Log2Urine Albumin	SFK (vs. control)	0.55	0.09, 1.00	0.019
	Female (vs. male)	0.03	−0.43, 0.50	0.885
	Age (years)	0.00	−0.05, 0.05	0.872
	Weight z-score	0.17	−0.05, 0.39	0.132
	Height z-score	−0.29	−0.52, −0.06	0.015
	DBP z-score	0.32	−0.02, 0.67	0.066
	SBP z-score	−0.13	−0.43, 0.17	0.385
Log2EGF	SFK (vs. control)	0.19	−0.27, 0.66	0.412
	Female (vs. male)	−0.02	−0.49, 0.46	0.934
	Age (years)	−0.02	−0.08, 0.03	0.349
	Weight z-score	0.04	−0.20, 0.27	0.754
	Height z-score	0.10	−0.14, 0.34	0.402
	DBP z-score	0.16	−0.20, 0.52	0.373
	SBP z-score	−0.08	−0.40, 0.24	0.622

Spearman correlations between the three urinary analytes and patient characteristic variables (age, weight z-score, height z-score, BMI z-score, SBP z-score, and DBP z-score) were mostly small, and all except one were <0.30 in magnitude (Table 4). The correlation between height z-score and albumin in the SFK group was −0.41 (p = 0.009). Other correlations exceeding 0.25 in magnitude were as follows: between PGE₂ an age in the SFK group (r_s = −0.28, p = 0.034), between albumin and DBP z-score in the control group (r_s = 0.29, p = 0.069), and between EGF and height z-score in the SFK group (r_s = 0.26, p = 0.088).

Table 4.

Spearman correlations between urine Prostaglandin E₂ (PGE₂), urine albumin and urine epidermal growth factor (EGF), and age, weight Z-score, height Z-score, body mass index (BMI) Z-score, systolic BP (SBP) Z-score and diastolic BP (DBP) Z-score in control children and children with solitary functioning kidney (SFK) are shown in the table. The correlation between the three urinary analytes and patient characteristic variables (age, weight z-score, height z-score, BMI z-score, SBP z-score, and DBP z-score) were relatively small in magnitude.

	Urine PGE2 (Control)	Urine PGE2 (Control)	Urine PGE2 (SFK)	Urine PGE2 (SFK)
	Correlation	p-value	Correlation	p-value
AGE	−0.20	0.085	−0.28	0.034
WEIGHT Z-SCORE	−0.03	0.780	0.06	0.653
HEIGHT Z-SCORE	−0.19	0.102	−0.11	0.395
BMI Z-SCORE	0.06	0.630	0.08	0.549
SBP Z-SCORE	0.16	0.181	−0.08	0.533
DBP Z-SCORE	0.20	0.098	0.07	0.624
	Urine Albumin (Control)	Urine Albumin (Control)	Urine Albumin (SFK)	Urine Albumin (SFK)
	Correlation	p-value	Correlation	p-value
AGE	0.09	0.561	−0.21	0.188
WEIGHT Z-SCORE	0.02	0.918	−0.20	0.208
HEIGHT Z-SCORE	−0.20	0.221	−0.41	0.009
BMI Z-SCORE	0.07	0.653	−0.08	0.630
SBP Z-SCORE	0.20	0.214	0.11	0.506
DBP Z-SCORE	0.29	0.069	0.21	0.184
	Urine EGF (Control)	Urine EGF (Control)	Urine EGF (SFK)	Urine EGF (SFK)
	Correlation	p-value	Correlation	p-value
AGE	−0.09	0.563	0.02	0.881
WEIGHT Z-SCORE	−0.02	0.911	0.16	0.285
HEIGHT Z-SCORE	0.13	0.413	0.26	0.088
BMI Z-SCORE	−0.03	0.823	0.08	0.614
SBP Z-SCORE	0.25	0.108	−0.23	0.129
DBP Z-SCORE	−0.02	0.920	−0.09	0.540

Analysis of the relationship between urinary prostaglandin E₂, albumin and epidermal growth factor showed little evidence of correlation among the three urinary analytes (Table 5). Thus, the three urinary analytes are largely independent of each other.

Table 5.

Spearman correlations between urinary Prostaglandin E₂ (PGE₂), albumin, and epidermal growth factor (EGF) in control children and children with solitary functioning kidney (SFK) is shown in the table. There was no significant correlation among the three urinary analytes suggesting that they are largely independent of each other.

Control children	Urine Albumin (n=41)	Urine EGF (n=44)
Urine PGE2 (n=72)	rs = 0.06 (p = 0.722)	rs = −0.11 (p = 0.499)
Urine Albumin	-	rs = 0.01 (p = 0.933)
Children with SFK	Urine Albumin (n=40)	Urine EGF (n=44)
Urine PGE2 (n=57)	rs = −0.17 (p = 0.302)	rs = −0.02 (p = 0.905)
Urine Albumin	-	rs = −0.07 (p = 0.658)
All children	Urine Albumin (n=81)	Urine EGF(n=88)
Urine PGE2 (n=129)	rs = −0.01 (p = 0.951)	rs = −0.04 (p = 0.730)
Urine Albumin	-	rs = −0.02 (p = 0.847)

DISCUSSION

Results summarized above show increased urinary PGE₂ and albumin, but not EGF, in children with SFK compared to the control group which would suggest glomerular and not tubular injury. Results also show that urinary PGE₂, albumin and EGF are differentially altered and largely independent of each other in children with SFK. These results support our postulate on the significance of PGE₂ in early adaptive hyperfiltration-associated rise in FFSS prior to onset of clinical albuminuria (Figure 1). Thus, urinary PGE₂ is a potential biomarker for early stages of adaptive hyperfiltration (Figure 1).

The present study was undertaken to determine whether urine from children with SFK can be a potential source of biomarkers for determining early glomerular changes caused by hyperfiltration. To this end, we analyzed levels of urinary PGE₂, albumin, and EGF in a cohort of children with congenital SFK. Age, height and blood pressure in SFK and control groups were comparable except for lower weight and BMI in children with SFK (Table 1). Current standard of care does not include blood draw for evaluating kidney function in children with enuresis (control group) and is optional for children with SFK if urinalysis, growth and BP are normal. Thus, we did not have eGFR measurements for these children. Gender and prevalence of unilateral renal agenesis and multicystic dysplastic kidney in our cohort was comparable to the study by Urisarri et al [33] who identified renal agenesis and multicystic dysplastic kidney in 58% and 42% respectively, with 64% males, following ultrasound screening in 32,900 consecutive neonates.

In all forms of CKD including SFK, initial adaptive changes appear in the form of increased SNGFR from changes in renal blood flow and ultrafiltration coefficient (Kf) followed by increased glomerular capillary pressure. Increased SNGFR is an immediate consequence of reduced functional nephron number and a major determinant of increased FFSS. In adult kidney donors, Lenihan et al [34] found that the mean single kidney GFR increased from 47 (pre-donation) to 64 (early post-donation) to 66 (late post-donation) ml/min/1.73m². Similarly, the Assessing Long Term Outcomes of Living Kidney Donors (ALTOLD) study reported significant increase in GFR at a rate of 1.47 ± 5.02 ml/min/year in donors between 6 and 36months post-donation compared to a decrease of 0.36 ± 7.55 ml/min/year in controls [35]. We reported a 1.5- to 2-fold increase in FFSS from increased SNGFR following unilateral nephrectomy in rodent models of SFK [15]. These observations from human and animal studies suggest increased SNGFR and resulting increase in FFSS in individuals with SFK.

We had hypothesized an increase in urinary PGE₂ in individuals with SFK during early stages of glomerular injury caused by FFSS. This was based on our previous observations that application of FFSS to podocytes in vitro results in increased generation of PGE₂, upregulation of COX-2, induction of PGE₂ receptor EP2, and derangement of the actin cytoskeleton [16–18]. Likewise, we demonstrated increased expression of COX-2 and EP2 in podocytes in situ, and increased urine albumin excretion in unilaterally nephrectomized sv129 mice [18]. Recently, we found that experimental FFSS induces recruitment of the mediators of EP2 signaling namely, the Akt-GSK3β-β-catenin, ERK1/2, and p38MAPK signaling pathways, but not the cAMP-PKA pathway in podocytes [36]. Others have also reported increased urinary PGE₂ levels and renal COX-2 expression in rodent models of hyperfiltration caused by 5/6 nephrectomy, high dietary protein, and diabetes [37–39]. As postulated, we did find elevated urinary PGE₂ in children with SFK even after controlling for the covariates in the regression model (Tables 2 and 3). The results were not different after correcting urinary PGE₂ for BMI and weight (data not shown). Similarly, urinary PGE₂ is elevated in humans with hyperfiltration with moderate CKD or diabetes [40–42]. In these past studies in humans, it was not possible to differentiate the temporal change as elevated urinary PGE₂ and proteinuria were both present, and non-steroidal therapy in these models would decrease both these parameters [43–46]. We can now report that albuminuria is preceded by elevation in urinary PGE₂ in hyperfiltration-mediated injury. These observations from cell culture, animal and human studies support the finding of increased urinary PGE₂ in individuals with SFK and, also validates our concept that COX2-PGE₂-EP2 axis is upregulated in early stages of adaptive hyperfiltration (Figure 1).

Progressive increase in urinary albumin is an established indicator of gradual loss of glomerular filtration barrier function. As mentioned, we have proposed that persistent hyperfiltration-induced FFSS exerted on podocytes alters glomerular filtration barrier function. We have previously shown that FFSS or exogenous PGE₂ without FFSS can increase glomerular albumin permeability of isolated rat glomeruli that is blocked by indomethacin [18]. Thus, hyperfiltration (FFSS)-mediated glomerular injury in individuals with SFK would likely increase urinary albumin. The results outlined above show urinary albumin to be higher in SFK children compared to controls. In the ALTOLD study, urine albumin trended to increase over the 3-year follow-up period in donors but not in control subjects [47]. Urinary albumin levels may also be influenced by other factors such as obesity that is associated with microalbuminuria, proteinuria and obesity-related glomerulopathy [48]. However, in the present study we found that despite lower BMI, children with SFK had higher urinary albumin levels compared to the control group (Table 1 and 2) and were more likely to exceed the normal range cutoff of 30 μg/mgCr (18% vs. 5%; Fisher exact test p = 0.088). The difference in albumin levels was confirmed using the regression modeling adjusted for age, gender, height z-score, weight z-score, and BP z-scores (Table 3). Although urinary albumin was elevated in SFK children, 83% of values (33/40) were still within the normal reference range (<30 μg/mgCr), reflecting early stage of hyperfiltration-induced changes in the glomerular filtration barrier function. Large studies in adults including the Framingham Offspring Study, Nord-Trondelag Health (HUNT) study, LIFE study, HOPE study and PREVEND study have consistently shown increased risk for cardiovascular events and mortality with increasing urine albumin within the normal range- far below the current threshold of microalbuminuria (30 μg/mgCr) [49–53]. Thus, higher urine albumin excretion in children with SFK compared to the control group provides evidence for early effects of hyperfiltration (FFSS)-mediated injury to the glomerular filtration barrier. The increase in urine PGE₂ precedes the development of overt microalbuminuria has been difficult to discern as both are elevated in humans with hyperfiltration with moderate CKD or diabetes [43–46].

Urinary Epidermal growth factor (EGF) is considered a strong biomarker for evaluating tubular injury independent of circulating EGF. It belongs to the EGF-family of proteins and plays an important role in cell proliferation and attenuation of apoptosis. EGF transcript, highly expressed in the kidney, is enriched in the thick ascending limb of Henle and distal tubules compared to glomeruli, papillary tips, and renal pelvis [22–24]. EGF expression is decreased in tubular cells of the cortex and medulla in CKD, and it correlates inversely with interstitial fibrosis and tubular atrophy [24]. A decrease in urinary EGF is associated with the progression of CKD in adults and children with diabetes [24,25], IgA nephropathy [24,26], nephrotic syndrome [24], autosomal dominant polycystic kidney disease [27], Alport syndrome [28] and CKD [24,29,54].

Children with decreased renal function show significantly lowered urinary EGF while it remains unchanged in children with kidney disease without loss of renal function [29]. These observations parallel the findings in adults where urine EGF decreases along with eGFR [29,30,54]. In a study of children with ureteropelvic junction obstruction, low urinary EGF improved following surgical intervention, suggesting that EGF is involved in the pathogenesis of tubulointerstitial damage in congenital obstructive nephropathy [30]. Thus, comparable levels of urinary EGF in the SFK and control groups suggest an absence of significant tubular injury in the SFK group in the present study. Virtually no correlation between urine albumin and EGF in control or children with SFK (Table 5) was detected. Three large cohort studies (C-PROBE, NEPTUNE, and PKU-IgAN) also reported a weak to moderate negative correlation between urinary EGF and albuminuria (r = −0.14, −0.17, and −0.31, respectively) [24]. Thus, as discussed above, albuminuria and urine EGF reflect distinct pathophysiologic mechanisms associated with glomerular and tubulointerstitial injury, respectively.

In summary, elevated PGE₂ and albumin excretion without decreased EGF in the urine from children with SFK supports our postulate that effects of hyperfiltration represent a continuum of overlapping changes in the glomerular filtration barrier. Present findings are in line with our proposal that the COX2-PGE₂-EP2 axis and components of the renin-angiotensin-aldosterone system play important roles in the early and late stages of the hyperfiltration continuum, respectively (Figure 1). However, a limitation of the study is that the proposed hyperfiltration-mediated injury is largely based on experimental in vitro and animal models with only indirect support from human data. These cross-sectional data will need to be validated in a prospective follow-up studies to determine sequential changes in serum creatinine and urinary analytes in children and adults with SFK. Present observations provide heretofore unavailable evidence for a temporal distinction between the observed increase in urinary PGE₂ and albumin in human disease and animal models. Results also suggest urinary PGE₂ preceding overt albuminuria as a potential biomarker for initial effects of adaptive hyperfiltration and strengthens our long-term approach to study FFSS for identifying novel targets for treating hyperfiltration-mediated injury.

HIGHLIGHTS.

• Urinary PGE₂ and albumin rise in CKD but which increases first is not known.

• Children with solitary kidney (SFK) develop CKD due to hyperfiltration.

• Urinary albumin but not EGF is elevated in SFK indicating glomerular injury.

• Increase in urinary PGE₂ precedes clinical albuminuria in children with SFK.

• Urinary PGE₂ is a biomarker of hyperfiltration-mediated early glomerular injury.