Introduction

In the United States, hemoglobin SS (HbSS) disease is more prevalent in African Americans, occurring in 1/400 births of children of African origin [1]. Sickle cell disease (SCD) is a genetic disorder with a single base-pair mutation resulting in an amino acid change of Glu > Val in the β globin gene. Patients with HbSS disease are homozygous for the βS globin gene; they are characterized by severe anemia and morbidities involving multiple organ systems such as vaso-occlusive crises, acute chest syndrome, cerebrovascular events, retinopathy, and renal damage. Other hemoglobinopathies such as HbSC or HbSD disease are mixed heterozygote states with milder phenotypes than SS disease.

Sickle cell nephropathy (SCN) is a common complication that occurs in one-third of SCD patients [2–4]. Renal failure is a major cause of mortality in SCD [5, 6], developing in up to 18% of adult patients with SCD [7]. As SCN affects both the glomerular and tubular functions of the kidney, SCD patients have varied presentations [7]. Patients with glomerular dysfunction may present with proteinuria, nephrotic syndrome, hematuria, and renal failure. SCD patients with tubular dysfunction may present with urinary concentrating defects, impaired potassium secretion, renal tubular acidosis, and renal failure.

Microalbuminuria (MA) is an early marker of glomerular dysfunction and has a prevalence of up to 46% in children with SCN of between 10 and 18 years of age. It correlates with increasing age and lower hemoglobin levels [8–10]. Higher molecular weight proteinuria, evidenced by nephrotic syndrome is a predictor of progression to end-stage renal disease and correlates with early mortality [3, 11]. Tubular dysfunction is also very common in SCD. Recent studies suggest that tubular injury is independent of glomerular damage; Marsenic et al. reported increased urinary excretion of retinol binding protein as a marker of tubular injury without MA [12].

Currently, MA and creatinine are the only clinical markers of advanced renal injury. To find non-invasive urinary markers of early renal damage in SCD, we examined urinary levels of proteins involved in acute kidney failure and the initiation of fibrosis and tubular dysfunction, and correlated them with MA and the estimated glomerular filtration rate (GFR). We also aimed to study whether the urinary levels of these proteins correlated with the severity of SCD. If an association could be found, then we would propose that these urinary proteins could be used in future prospective studies to monitor the progression of sickle cell nephropathy and the effectiveness of its treatment.

Transforming growth factor β-1 (TGF-β1) is a ubiquitously expressed cytokine with roles in embryonic development, cellular homeostasis and extracellular matrix remodeling [13]. It causes renal fibrogenesis and nephron loss by various mechanisms, such as apoptosis of endothelial cells and podocytes; it also plays a role in epithelial to mesenchymal cell transition [14]. The TGFβ-1 pathway has been shown to be an important player in the development and progression of diabetic nephropathy [15, 16]. Recent studies have suggested that the TGF-β1 pathway may be activated by the renin–angiotensin aldosterone system (RAAS) [17]. The intrarenal RAAS system may be involved in the pathogenesis of sickle cell nephropathy, although there are no studies that directly address this. However, Falk et al. reported a reduction in proteinuria following treatment with an ACE inhibitor [18]. Single nucleotide polymorphisms in several genes of the TGF-β1/BMP superfamily have been shown to be associated with more severe renal disease and complications such as stroke, osteonecrosis, priapism, lower extremity ulcers, pulmonary hypertension, and in general as a measure of overall disease severity [19]. To date, to the best of our knowledge, there have been no published reports evaluating urinary excretion of TGFβ-1 in SCD.

Neutrophil gelatinase-associated lipocalin (NGAL) is expressed and secreted by immune cells, hepatocytes, and renal tubular cells in various pathological states. It is a member of the lipocalin family of proteins that has been extensively studied in acute kidney injury (AKI) [20, 21]. Mitsnefes et al. showed that serum NGAL correlates well with measured GFR, and may be a useful marker for monitoring and grading chronic kidney disease (CKD) even at GFR < 30 ml/min/1.73 m2 [22]. We assessed urinary NGAL levels in patients with SCD to determine its usefulness in assessing chronic renal injury in SCD.

Subjects and methods

The study was conducted at The Children’s Hospital at Montefiore/Albert Einstein College of Medicine. The Montefiore Medical Center Institutional Review Board approved the protocol for this study. Fifty-one children with SCD and 21 healthy controls between the ages of 2 and 21 years were enrolled from January 2009 to December 2009. We enrolled children with a diagnosis of sickle cell disease (HbSS, HbSC, HbSD or Sβ thalassemia) during their regular health maintenance visits to the outpatient hematology department. Control subjects were recruited and enrolled in the outpatient general pediatric clinics. The controls were primarily African Americans and Hispanics. Patients with a recent hospitalization within the last 30 days, history of blood transfusions within the last 90 days prior to enrollment, history of current pregnancy, neoplasm, hypertension, any known glomerular disease, patients who were on antihypertensive or steroid therapy for any reason were excluded. A total of 42 patients had HbSS, 8 HbSC, and 1 HbSD.

Clinical and laboratory data, obtained during routine medical care, was gathered by chart review of the visit at which the patients were enrolled. Estimated GFR was measured by the Schwartz formula [23]. Urinary microalbumin and creatinine were determined by turbidimetric analysis (Olympus 640). Microalbumin excretion was determined by the urinary microalbumin to urinary creatinine ratio. MA was defined as albumin/creatinine > 20 mg/g, as previously used in other studies [8]. Data analysis was performed using descriptive statistics and correlation analysis. Fresh urine samples were collected at room temperature within 2 h, centrifuged (10,000 rpm for 10 min) and the supernatant was stored at −80°C for analysis. Enzyme-linked immunosorbent assay (ELISA) kits were used to detect urinary excretion of activated TGF-β1 and NGAL (R&D Systems, Minneapolis, MN, USA). Values were normalized to measured urine creatinine. Mean arterial pulse index (MAPI) was calculated as the average MAP (calculated from systolic and diastolic values) for a given patient divided by the MAP value for the 95th percentile for age, gender, and height based on the Fourth Report of the National High Blood Pressure Education Program Working Group [24]. Self-reported patient ethnicity was confirmed from the medical record of each patient.

Statistical analysis was performed using STATA (Stata Corporation, College Station, TX, USA) and Microsoft Office Excel 2003 (Microsoft Corporation, Redmond, WA, USA). P < 0.05 was considered statistically significant.

Results

We analyzed the urine samples of 51 SCD subjects: 42 had HbSS, 8 had HbSC, and 1 had HbSD. Of the 42 HbSS subjects, 16 (38%) were receiving treatment with hydroxyurea (HU). We compared all the SCD subjects against 21 healthy controls (CTR). There were no statistically significant differences with regard to mean age, gender, and race between the SCD group and the CTR group. Demographic information for the SCD subjects and the CTR group are shown in Table 1.

Table 1 Clinical parameters
Full size table

All SCD patients had elevated eGFR (174.7 ± 40.3 ml/min/1.73 m2 in the HbSS group and 137.8 ± 40.0 ml/min/1.73 m2 in the HbSC group, p = 0.01); eGFR data on control subjects was not available. Further clinical and laboratory parameters characterizing the HbSS and HbSC/SD groups are shown in Table 2.

Table 2 Clinical parameters in subjects with sickle-cell disease (SCD)
Full size table

Urinary excretion of TGF-β1 was 26.4 ± 1.5 pg/mgCr in SCD subjects vs 15.0 ± 2.4 pg/mgCr in CTR, p < 0.00001). TGF-β1 was higher in both HbSS and HbSC subjects (HbSS vs CTR, p <0.00001, and HbSC vs CTR, p < 0.01), but there was no statistical difference in TGF-β1 levels between HbSS and HbSC patients (Fig. 1). We found a negative correlation between urinary TGF-β1 levels and age and a positive correlation between urinary creatinine and age in CTR (Fig. 2, Spearman’s rho =−0.8, and Spearman’s rho = 0.87 respectively, p < 0.00001), but this correlation was not observed in SCD subjects. SCD patients with hemoglobin less than 9 g/dl had higher urinary TGF-β1 than SCD patients with milder anemia (p  = 0.002). Urinary TGF-β1 was lower in HbSS patients on HU (23.61 ± 2.6 pg/mgCr) vs HbSS patients who were not on HU (27.69 ± 1.8 pg/mgCr), although this difference did not reach statistical significance (p = 0.055).

Fig. 1
figure 1

Box plot of urinary transforming growth factor (TGF) beta-1 levels in controls and subjects with sickle cell disease (SCD). ±SEM standard error of the mean. p = 0.0001 comparison amongst three groups (Kruskal–Wallis). p < 0.00001 comparison between controls (CTR) and HbSS (Mann–Whitney). p = 0.71 comparison between HbSS and HbSC/SD (Mann–Whitney). p < 0.01 comparison between CTR and HbSc/SD (Mann–Whitney)

Full size image
Fig. 2
figure 2

Negative correlation between urinary TGF-β1 levels (open squares) and age in CTR (Spearman’s rho =−0.8, p < 0.00001), and positive correlation between urinary creatinine (triangles) and age in CTR (Spearman’s rho = 0.87, p < 0.00001)

Full size image

MA (urine microalbumin >30 mg/gCr) was detected in 10% of subjects with SCD. All of the subjects with MA were older than 10 years. Urinary TGF-β1 was 33.9 ± 19.7 pg/mgCr in patients with MA vs 25.8 ± 9.5 pg/mgCr in patients with no MA (p = 0.49, Mann–Whitney test). There was no correlation between urinary TGF-β1 and levels of urinary microalbumin (Spearman’s rho = 0.2, p = 0.17) or eGFR (Spearman’s rho =−0.07, p = 0.64).

We found no statistical difference in urinary NGAL excretion between healthy controls and SCD subjects; 7.8±5.3 ng/mgCr in controls, 7.9±6.6 ng/mgCr in HbSS and 7.8±5.3 ng/mgCr (p = 0.85, Kruskal–Wallis). Mean arterial pressure index (MAPI) was 0.8 ± 0.08 in HbSS, 0.8 ± 0.07 in HbSC/SD, and 0.7 ± 0.09 CTR (p = 0.21). Hospitalizations per person-year in HbSS group was 1.4 per year vs 0.3 in the HbSC/SD group (p = 0.02).

Discussion

TGF-β1 is a known profibrogenic cytokine that plays an important role in the progression of CKD [13]. TGF-β1 has been shown to directly induce both increased synthesis and reduced degradation of matrix proteins, leading to a net accumulation of pathological matrix, thereby playing a role in the fibrotic process. In our study, we found elevated levels of urinary TGF-β1 in patients with SCD compared with controls. However, we did not observe statistical difference in urinary TGF-β1 levels and MA between HbSS and HbSC patients. One explanation for this finding may be the limited number of HbSC patients in our study. While there was a trend toward MA in SS patients more than in SC/SD patients, TGF-β1 levels were elevated in both groups. Although we observed that lower Hgb is associated with higher TGF-β1 urine levels, we were unable to demonstrate that less anemic patients (SC group) had lower urinary TGF-β1 levels. This could be explained by the relatively small SC sample size; or alternatively, it may be that the relationship between Hb and TGF-β1 is more complex and other factors may modulate it, i.e. SS patients were treated with hydroxyurea. Urinary TGF-β1 in SCD subjects did not correlate with age, eGFR or MA. Our finding of increased prevalence of MA in older children with SCD (all were older than 10 years) and urinary TGF-β1 excretion that did not correlate with age (p = 0.14), suggests that greater renal injury was present earlier than would be recognized based on MA testing alone. There are no systematic studies examining urine excretion of TGF-β1 over the course of childhood. However, in Japanese patients Okamoto et al., demonstrated that the serum TGF-β1 level of children (aged 0–14 years) was significantly higher than that of adults (over 15 years of age; p < 0.01) [25]. Interestingly in our study, we did note a strong negative correlation between urinary TGF-β1 (normalized to urine creatinine) and age, and a positive correlation between urinary creatinine and age in CTR subjects. This may suggest that in controls the higher ratio of TGF-β1/urine Cr is determined by lower urinary creatinine excretion.

Close follow-up of SCD patients is important to identify those who are at increased risk of developing renal dysfunction. For example, lower hemoglobin has been associated with a higher risk of proteinuria and renal dysfunction [26, 27]. In our SCD cohort, lower hemoglobin was correlated with higher urinary TGF-β1 levels, and it is likely that these are the patients most at risk of developing renal disease. The small sample size of HbSC patients makes it difficult to interpret our findings in the SC group. Long-term monitoring of this cohort is necessary to determine the clinical relevance of our findings. Alvarez et al. reported that blood transfusions protected against MA in patients with SCD [28]. As we excluded patients who had received blood transfusions in the 90 days prior to enrollment, which excludes patients receiving chronic transfusion therapy, we did not assess MA in that subset of patients. Hydroxyurea with ACE inhibitors may further reduce proteinuria in SCD subjects [29]. In our HbSS group, patients being treated with HU tended to have lower urinary TGF-β1 excretion compared with those not receiving treatment with HU; the difference was not statistically significant, which may be attributed to our small sample size. Serum TGF-β1 protein levels have been shown to be positively and significantly associated with the systolic and diastolic blood pressure in African Americans after controlling for age, gender, and body mass index [30]. However, in our population MAPI was not increased in the SCD group compared with CTR.

Urinary NGAL is a diagnostic biomarker that has been shown to predict the development of acute kidney injury (AKI) and has been associated with CKD [20–22]. In our population, we found no difference in urinary excretion of NGAL between SCD patients and CTR. This could be attributed to the fact that, based on our exclusion criteria, we excluded patients who were clinically less stable, likely patients experiencing frequent and/or severe vaso-occlusive crises and resultant episodes of AKI. In addition, none of our SCD subjects had CKD.

Reports in adults have shown that asymptomatic proteinuria can progress to overt proteinuria and ultimately to renal failure [1, 11, 31]. Renin and angiotensin II have been shown to directly stimulate the expression of renal TGF-β1 independent of renal hemodynamics [32–34]. Therefore, blockade of the RAAS is an important therapeutic strategy for prevention of the progression of renal disease; for example, angiotensin receptor blockers or angiotensin converting enzyme inhibitors (ACEI) have been shown to decrease albuminuria and preserve renal function in diabetic nephropathy [35, 36]. ACEI have been successfully used in SCD, suggesting that glomerular capillary hypertension may be a pathogenic factor in SCN [18]. In our cohort of SCD patients, we did show evidence of glomerular hyperfiltration, which may support the use of ACE inhibitors even before the advent of MA. It is unclear if such a treatment would correspondingly decrease urinary TGF-β1 excretion. Interestingly, blockade of TGF-β1 directly in experimental models has been shown to prevent the progression of CKD [37, 38], suggesting a role for new treatment strategies. The long-term effectiveness of these therapies in prevention of renal disease in SCD patients needs to be further studied [18, 39, 40].

One of the limitations of our study is that the urine microalbumin evaluation was not performed on urine collected in a first morning void to exclude elements of orthostatic proteinuria. Also, as we recruited CTR subjects during routine well child visits, we neither had laboratory values available to calculate GFR nor noted hemoglobin or urine microalbumin.

Although serum creatinine is currently the clinically available tool used to estimate renal function, more and more evidence points to the fact that creatinine clearance may not be an accurate assessment of renal function in general [41]. Creatinine clearance compared with inulin clearance is significantly increased in SCD individuals due to the oversecretion of creatinine by the proximal tubules, above the amount filtered by the glomerulus [39]. In SCD subjects, even those with normal GFR, the presence of albuminuria is indicative of a decreased ultrafiltration coefficient, suggesting an injury to the filtration barrier [42]. Serum cystatin C has been suggested as a marker of GFR in children and adults with SCD [27, 43, 44], but further long-term studies are required to assess the reliability of serum cystatin C as an early marker of renal dysfunction in SCD patients. Urinary TGF-β1 is elevated in SCD subjects and may serve both as a marker to detect patients at risk of developing impending renal injury and as a treatment target. Our finding needs to be evaluated in long-term prospective trials.