Hydroxyurea improves nitric oxide bioavailability in humanized sickle cell mice

Abstract

Despite advancements in disease management, sickle cell nephropathy, a major contributor to mortality and morbidity in patients, has limited therapeutic options. Previous studies indicate hydroxyurea, a commonly prescribed therapy for sickle cell disease (SCD), can reduce renal injury in SCD but the mechanisms are uncertain. Because SCD is associated with reduced nitric oxide (NO) bioavailability, we hypothesized that hydroxyurea treatment would improve NO bioavailability in the humanized sickle cell mouse. Humanized male 12-wk-old sickle (HbSS) and genetic control (HbAA) mice were treated with hydroxyurea or regular tap water for 2 wk before renal and systemic NO bioavailability as well as renal injury were assessed. Untreated HbSS mice exhibited increased proteinuria, elevated plasma endothelin-1 (ET-1), and reduced urine concentrating ability compared with HbAA mice. Hydroxyurea reduced proteinuria and plasma ET-1 levels in HbSS mice. Untreated HbSS mice had reduced plasma nitrite and elevated plasma arginase concentrations compared with HbAA mice. Hydroxyurea treatment augmented plasma nitrite and attenuated plasma arginase in HbSS mice. Renal vessels isolated from HbSS mice also had elevated nitric oxide synthase 3 (NOS3) and arginase 2 expression compared with untreated HbAA mice. Hydroxyurea treatment did not alter renal vascular NOS3, however, renal vascular arginase 2 expression was significantly reduced. These data support the hypothesis that hydroxyurea treatment augments renal and systemic NO bioavailability by reducing arginase activity as a potential mechanism for the improvement on renal injury seen in SCD mice.

INTRODUCTION

Initially described in 1910, sickle cell disease (SCD) is an autosomal recessive blood disorder distinguished by hemoglobin polymerization. Ultimately, this results in erythrocytes with a distorted sickle shape (1, 2), which promotes red cell destabilization, chronic hemolytic anemia, vaso-occlusive events, ischemia-reperfusion injury, nitric oxide (NO) deficiency, and systemic inflammation (3–5). As expected, each of these processes can have a profound effect on disease progression and lead to multisystem, end-organ damage, namely, pulmonary arterial hypertension, stroke, and nephropathy (4, 6).

Patients with SCD suffer from dysregulated NO signaling as seen by depressed plasma NO concentrations and impaired NO-dependent blood flow (7, 8). NO is a free-radical signaling molecule biosynthesized from the amino acid l-arginine by a family of enzymes known as nitric oxide synthases (NOS) via induction of the soluble guanylyl cyclase (sGC) pathway to result in vasodilation and other anti-inflammatory and antioxidant effects (9, 10). Chronic intravascular hemolysis due to decreased erythrocyte stability in SCD promotes the release of extraerythrocytic hemoglobin and free heme, which both sequester plasma NO leading to NO depletion (11). In addition, arginase released from erythrocytes during intravascular hemolysis competes with NOS for the substrate arginine, which further exacerbates NO depletion (12). Patients with SCD have reduced arginine bioavailability, which promotes reactive oxygen species (ROS) generation by NOS uncoupling and NO scavenging by ROS molecules, such as peroxynitrite and superoxide (12, 13). As NO is necessary for the maintenance of vascular tone, NO insufficiency can lead to systemic vasculopathy and endothelial dysfunction in SCD (10, 14). Although cumbersome on a chronic basis, continuous exogenous NO administration can be an effective therapy for improving clinical outcomes associated with acute chest syndrome (ACS) and pulmonary arterial hypertension (15, 16). To date, there have been few studies, that we know of, investigating exogenous NO administration therapy on SCN.

Hydroxyurea has become the cornerstone of therapy for patients suffering from SCD since its debut in the mid-90s as the first drug to effectively reduce the rate of vaso-occlusive events (17). Hydroxyurea is an antineoplastic agent believed to function primarily through fetal hemoglobin induction to reduce hemoglobin polymerization and erythrocyte sickling (18, 19). Hydroxyurea is also used as a cancer therapeutic due to its ability to inhibit ribonucleotide reductase, an enzyme required for the conversion of ribonucleotides into deoxyribonucleotides in DNA synthesis (18). In patients with SCD, chronic hydroxyurea therapy has demonstrated long-term clinical efficacy through preservation of splenic function, diminished end-organ damage, and reduced mortality and morbidity (18, 20, 21). In addition, short-term hydroxyurea therapy reduces renal injury indicated by reduced albuminuria in both patients and animal models of SCD (22, 23). The induction of fetal hemoglobin, a potent inhibitor of deoxy sickle hemoglobin polymerization, remains regarded as one of the prominent mechanisms by which hydroxyurea exerts its therapeutic effects (24, 25). Studies have shown that this process is mediated, in part, by the activation of the soluble guanylyl cyclase (sGC) pathway in erythroid cells similar to NO (26). Hydroxyurea-induced γ-globin gene expression in erythroid progenitor cells is suppressed by blocking sGC pathway activation (26). Moreover, hydroxyurea therapy has been shown to augment guanosine 3′,5′-cyclic monophosphate (cGMP), a secondary messenger in the sGC pathway, and fetal hemoglobin in patients with SCD (27). Although the many clinical benefits of hydroxyurea are often attributed to its ability to induce fetal hemoglobin, increased NO bioavailability has emerged as a possible mechanism for its therapeutic effects. Oral hydroxyurea therapy has been shown to augment both intravascular and intraerythrocytic NO production in patients with SCD in as early as 2 h postadministration with a concomitant reduction in plasma vascular adhesion molecule-1 expression (27, 28). This mechanism, however, has not been examined in SCN.

Prior studies suggest that augmented NO production may serve as a potential therapeutic mechanism of hydroxyurea treatment in sickle cell nephropathy (SCN) (28–30). However, details of the mechanism for improving NO bioavailability in SCD are very limited. Therefore, the objective of this study was to determine if improved systemic and renal NO bioavailability is involved in the mechanism of short-term hydroxyurea treatment in a mouse model of SCD.

METHODS

Animals

Humanized sickle cell mice with endogenous murine hemoglobin genes replaced with human hemoglobin genes were used for our studies. These humanized knock-in sickle cell mice developed by Townes et al. have the notation: B6; 129-Hba^tm1(HBA)Tow Hbb^tm2(HBG1,HBB*^)Tow/Hbb^{tm3(HBG1,HBB)Tow}/J (31). Homozygosity at the β-globin locus for the mutant sickle allele or the control allele defined sickle cell mice (HbSS) and genetic control mice (HbAA), respectively. All experimental animals were 14-wk-old male mice of both genotypes and maintained in a standard 12-h light/dark cycle. Mice were euthanized between 9:00 AM and 12:00 PM for all plasma and tissue collections. Food and water were provided to the experimental mice ad libitum (7917 Irradiated NIH-31 Mouse/Rat diet, Envigo, Indianapolis, IN). Mice were placed in standard metabolic cages for 3 days for 24-h urine collection. After a 24-h acclimation period, urine was collected each day for the following 2 days. An average of two 24-h urine collections was calculated for the daily excretion rate. The University of Alabama at Birmingham Institutional Animal Care and Use Committee approved all experimental protocols and animal care. Experiments followed the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines and the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The number of animals in each experiment varied because of the variable output from our colony.

Drug Treatment Protocols

Experimental mice were treated with hydroxyurea (50 mg/kg/day; Sigma-Aldrich Lot No. MKBR1926V; St. Louis, MO) in the drinking water daily for 2 consecutive wk. Control mice were given drinking water without hydroxyurea. Daily water intake was measured and used to adjust concentrations appropriately, as previously described (32). Mice were treated from the age of 12 wk old to 14 wk old, at which time measures of nephropathy and NO bioavailability were assessed.

Plasma and Urine Analyses

Blood was collected in ethylenediaminetetraacetic acid (EDTA)-coated tubes from experimental mice via cardiac puncture, aliquoted, and frozen in liquid nitrogen. Plasma was obtained from the whole blood sample. Plasma endothelin-1 (ET-1) concentrations were measured by immunoassay as per the manufacturer’s instructions (QET00B, R&D Systems, Minneapolis, MN). Plasma nitrite and nitrate, metabolites of nitric oxide, were measured by high-performance liquid chromatography (ENO-30, Eicom, Kyoto, Japan) as previously described (33). Briefly, plasma mixed with 100% methanol (1:1) was centrifuged for 10 min at 8,000 rpm. The supernatant was separated on a column. Subsequently, nitrate was reduced to nitrite which reacts with Griess reagent to produce diazo compounds measured by absorbance at 540 nm and quantified using a nitrite and nitrate standard curve. A colorimetric assay kit was used to measure plasma lactate dehydrogenase in plasma samples, as per the manufacturer’s instructions (LDH Assay kit, Abcam, Cat. No. ab102526, Cambridge, MA). Daily 24-h urine collections were collected under mineral oil, aliquoted, and frozen in liquid nitrogen.

Protein Extraction and Western Blot Analysis

Renal vessels were isolated from fresh kidneys as previously described (34). Vessels were placed in homogenization buffer containing 50 mM Tris·HCl, pH 7.4, 0.1 mM ethylenediaminetetraacetic acid, 0.1 mM ethyleneglycol-bis(β-aminoethyl ether)-N,N′-tetra-acetic acid, 0.1% vol/vol β-mercaptoethanol, 10% vol/vol glycerol, 1 mM phenylmethylsulfonyl fluoride, 1 μM pepstatin A, 1 µM leupeptin, and 1 µM phosphatase inhibitor (Phosphatase Inhibitor Cocktail A, Santa Cruz). Tissue was homogenized using a glass homogenizer. Homogenate was sonicated and gently rocked for 30 min at 4°C. The resulting mixture was then centrifuged at 5,000 g at 4°C. Protein concentration was determined for tissue lysates using the Bradford assay. Equal amounts of tissue lysate from some experimental groups were run on a 10% sodium dodecyl-sulfate polyacrylamide gel at 120 V for 1.5 h. Proteins were transferred to a polyvinylidene difluoride (PVDF) membrane in a Tris-glycine buffer for 1.5 h at 90 V. Blots were blocked for 1 h at room temperature (RT) with a 1:1 mixture of a Tris-buffered saline (TBS) solution with a TBS blocking buffer [Odyssey blocking Buffer (TBS), Li-Cor, Lincoln, Nebraska]. Primary antibodies [NOS3, BD Transductions, Cat. No. 610297, 1:1,000; pNOS3 S1177, 1:200, Cell Signaling, Cat. No. 9571; Arginase-2, 1:1,000, Arg-2 321–335 (Sigma-Aldrich, Cat. No. A6232, St. Louis, MO)] in a 1:1 TBS with Tween-20 (TBST) with TBS blocking buffer with gentle rocking overnight at 4°C. Blots were then washed with TBST (3 × 5 min) before incubating in secondary antibody (IRDye 800CW goat anti-rabbit, 1:500; IRDye 700 goat anti-mouse, 1:10,000) diluted in a 1:1 mixture of TBST with TBS blocking buffer for 1 h at RT. Blots were washed again with TBST (3 × 5 min) before being developed with near-infrared fluorescent imaging (Odyssey CLx Imager, Li-Cor, Lincoln, Nebraska).

Arginase Activity

Arginase activity was measured in plasma and renal vessels using a colorimetric assay as per the manufacturer’s instructions (Arginase Activity Assay kit, Sigma-Aldrich, St. Louis, MO). For renal vessels, equal amounts of protein were loaded into the assay for arginase activity measurement. Plasma samples were depleted of endogenous urea by centrifuging samples twice for 30 min at 10,000 g in 3 K filters before assay measurement (Amicon Ultra Centrifugal filters, MilliporeSigma, Burlington, MA).

Fetal Hemoglobin

Whole blood samples were centrifuged at 3,000 g for 5 min at 4°C. Plasma was removed and the remaining red blood cells were washed three times with cold PBS. Cells were spun at 3,000 g for 5 min at 4°C after each wash. Cells were then snap frozen in liquid nitrogen and allowed to thaw at room temperature. This process was performed twice to ensure lysis of the red blood cells. The resulting mixture was vortexed briefly and spun at 5,000 g for 10 min at 4°C to remove cellular debris. The lysate was removed and used for analysis. Fetal hemoglobin was measured in red blood cell lysate using an ELISA [ELISA kit for Fetal Hemoglobin (HbF), MyBioSource, Cat. No. MBS2700434, San Diego, CA), per the manufacturer’s instructions.

Plasma Amino Acid Analyses

l-Citrulline and l-arginine were quantified by HPLC-MS with a Shimdazu 20 series HPLC in line with a Sciex API 4000 triple quadrupole mass spectrometer. Analytes in samples were compared against authentic mixed amino acid EZFaast standard (Phenomenex, Torrence, CA). A standard curve ranging from 1 µM to 200 µM over seven points was employed. Five microliters of standard or sample were added to 600 µL of 50% methanol and incubated for 10 min at room temperature. Samples were centrifuged for 10 min at 15,000 g. Next, 400 µL of the supernatant was transferred to an HPLC vial for analysis. Samples were chromatographically separated by hydrophilic interaction chromatography (HILIC) on an Intrada Amino Acid 3 µM 75 × 3-mm column (Imtakt) at 35°C with a flow rate of 0.6 mL/min. Mobile phases were 100 mM ammonium acetate for A and acetonitrile/water/formic acid (95/4.7/0.3 vol/vol/vol) for B. Gradient schedule and mass spectroscopy parameters were adapted from Liu et al. (35). Additional transitions were added for l-arginine (175.1 > 70.1), l-ornithine (133.1 > 70), and l-citrulline (176 > 113) to the MS method. Data acquisition was carried out using Analyst v. 1.6.2 and data analysis with MulitQuant v. 3.0.2 (Sciex). Standard curves were linear with 1/x² weighting.

Luminex Soluble Factor Analysis

We also analyzed plasma from HbAA and HbSS groups treated with hydroxyurea for a series of soluble inflammatory factors using a Luminex MAGPIX System [Cat. No. MCVD1MAG-77K-05 MILLIPLEX Mouse CVD Panel 1 with sE-selectin, ICAM-1, sP-selectin, plasminogen activator inhibitor-1 (PAI-1), and matrix metallopeptidase-9 (proMMP-9)] according to the manufacturer’s instructions. For soluble factor analysis by Luminex, samples were thawed on ice and plasma was analyzed according to the manufacturer’s instructions.

Statistical Analyses

GraphPad Prism 7.0 software was used to perform all statistical analyses. Two-way ANOVA (genotype ± drug treatment) with Tukey’s post hoc correction for multiple comparisons were used for all appropriate analyses as indicated in the figure legends. Statistical analysis was performed using all treatment groups and genotypes. Outliers were evaluated using Grubbs’ or ROUT. All data are presented as means ± SE with P < 0.05 being considered statistically significant.

RESULTS

Physical characteristics of all experimental animals were assessed after 2 wk of treatment with either hydroxyurea or regular drinking water (Table 1). Untreated HbSS mice exhibited significant splenomegaly as seen by increased spleen-to-body weight ratio. Hydroxyurea treatment yielded no significant effects on body weight, spleen-to-body weight ratio, or kidney size in either HbSS or HbAA mice (Table 1).

Table 1.

Physical characteristics of 2 wk treated HbAA and HbSS mice

	HbAA		HbSS
	Control	Hydroxyurea	Control	Hydroxyurea
n	8	6	7	9
Body wt, g	27.3 ± 0.9	27.9 ± 0.7	27.4 ± 0.8	26.9 ± 0.5
Right kidney/body wt, mg/g	5.9 ± 0.2	5.2 ± 0.1	6.3 ± 0.5	6.4 ± 0.3
Left kidney/body wt, mg/g	5.9 ± 0.2	5.8 ± 0.2	6.2 ± 0.6	6.4 ± 0.3
Spleen-to-body wt ratio (%)	0.5 ± 0.1	0.4 ± 0.1	6.4 ± 0.4*	5.7 ± 0.3*

Data are presented as means ± SE; n, number of animals. *P < 0.05 vs. similarly treated HbAA group. HbAA, genetic control mice; HbSS, sickle cell disease mice.

We confirmed that HbSS mice exhibited evidence of renal damage by significant proteinuria compared with HbAA mice (Fig. 1A) (23, 36, 37). As expected, untreated HbSS mice displayed evidence for tubular injury as seen by defects in urinary concentration ability (Fig. 1B) (37). HbSS mice also had significantly elevated plasma ET-1 concentrations, which has been previously reported in this model (Fig. 1C) (37). Short-term treatment with hydroxyurea in HbSS mice significantly reduced proteinuria and plasma ET-1 compared with HbSS mice (Fig. 1A and C). There was no significant improvement in urine concentrating ability with hydroxyurea treatment (Fig. 1B).

Figure 1.

Hydroxyurea reduces renal injury in sickle cell disease mice. Total urinary protein excretion (A) (P_treatment = 0.011, P_genotype = 0.006, P_interaction = 0.014), urine osmolality (B) (P_treatment = 0.09, P_genotype < 0.001, P_interaction > 0.1), plasma endothelin-1 (ET-1) (C) were assessed in HbSS mice after 2 wk of treatment (P_treatment = 0.058, P_genotype < 0.001, P_interaction = 0.034). Data were analyzed by two-way ANOVA with Tukey’s post hoc comparison of individual means. HbSS, sickle cell mice.

We further investigated the effects of short-term hydroxyurea treatment on NO bioavailability in mice. Plasma nitrite is a measure of systemic NO bioavailability (28, 38). Untreated HbSS mice had significantly lower plasma nitrite concentrations compared with HbAA mice (Fig. 2A). This was significantly increased with hydroxyurea treatment (Fig. 2A). There was no significant difference in plasma nitrate concentration between groups (Fig. 2B). As arginase competes with NOS for the substrate l-arginine (39), we measured plasma arginase activity in HbAA and HbSS mice. Untreated HbSS mice had significantly elevated plasma arginase activity compared with HbAA mice (Fig. 2C). The lower plasma arginase activity observed in HbSS mice treated with hydroxyurea was statistically significant (Fig. 2C). In addition, plasma arginine concentration was significantly reduced in untreated HbSS mice compared with HbAA mice (Fig. 2D). However, hydroxyurea treatment did not improve plasma arginine concentration in HbSS mice (Fig. 2D). Citrulline, an amino acid generated by the enzymatic formation of NO, was also assessed as an indirect measure of systemic NO production (40). There was no significant difference in plasma citrulline concentration in HbSS and HbAA mice regardless of treatment (Fig. 2E).

Figure 2.

Hydroxyurea improves nitric oxide (NO) bioavailability in sickle cell disease mice. Plasma nitrite concentration (A) (P_treatment > 0.1, P_genotype > 0.1, P_interaction = 0.017), nitrate concentration (B) (P_treatment > 0.1, P_genotype > 0.1, P_interaction > 0.1), plasma arginase activity (C) (P_treatment > 0.1, P_genotype < 0.001, P_interaction = 0.019), plasma arginine concentration (D) (P_treatment > 0.1, P_genotype < 0.001, P_interaction > 0.1), and plasma citrulline concentration (E) (P_treatment > 0.1, P_genotype > 0.1, P_interaction > 0.1) were measured in 2-wk treated HbAA and HbSS mice. Two-way ANOVA with Tukey’s post hoc correction for multiple comparisons was used for HbAA and HbSS controls and hydroxyurea-treated mice. HbAA, genetic control mice; HbSS, sickle cell mice.

Next, we wanted to investigate whether changes in the arginine/NO pathway within the renal vascular system could be a potential mechanism for improved renal injury in hydroxyurea-treated HbSS mice. In renal vessels isolated from untreated HbSS mice, we observed significantly elevated arginase-2 protein abundance (Fig. 3, A and B) compared with HbAA mice. Though NOS3 abundance was elevated in the renal vessels of untreated HbSS mice, it was not statistically significant (Fig. 3, A and C). However, a significant genotype effect was observed in renal vascular NOS3 abundance in HbSS mice compared with HbAA mice such that the increase in NOS3 was significant in the hydroxyurea-treated HbSS mice. In addition, renal vascular arginase activity was significantly increased in HbSS mice compared with HbAA mice (Fig. 3D). This effect was ameliorated with hydroxyurea treatment in HbSS mice. Taken together, these findings demonstrate that short-term hydroxyurea treatment improves renal NO, most likely, by reducing both arginase protein abundance and arginase activity in renal vessels.

Figure 3.

Hydroxyurea improves nitric oxide (NO) bioavailability via reduced arginase activity in renal vessels of sickle cell mice. (A) representative Western blot images of Arg-2 and nitric oxide synthase 3 (NOS3) protein abundance in the renal vessels of 2-wk-treated HbAA and HbSS mice. Quantification of Arg-2 abundance (B) (P_treatment < 0.001, P_genotype < 0.001, P_interaction < 0.01), NOS3 abundance (C) (P_treatment > 0.1, P_genotype < 0.001, P_interaction > 0.1), and renal vascular arginase activity (D) (P_treatment < 0.01, P_genotype < 0.01, P_interaction = 0.01) in renal vessels from 2-wk treated HbAA and HbSS mice. Two-way ANOVA with Tukey’s post hoc correction for multiple comparisons was used for HbAA and HbSS controls and hydroxyurea-treated mice. HbAA, genetic control mice; HbSS, sickle cell mice.

We examined several hemolysis markers to determine whether this could be a contributing factor in reducing NO bioavailability in HbSS mice. Hemoglobin and hematocrit, both indirect markers of hemolysis, were decreased in untreated HbSS mice compared with HbAA mice (Fig. 4, A and B). These hematological parameters were restored with hydroxyurea treatment in HbSS mice (Fig. 4, A and B). We then measured plasma lactate dehydrogenase (LDH), which is an enzyme that converts pyruvate to lactate and is a surrogate marker of hemolysis in SCD patients (41). HbSS mice had significantly elevated plasma LDH activity compared with HbAA mice (Fig. 4C). Plasma LDH activity was significantly reduced by hydroxyurea treatment in HbSS mice (Fig. 4C) compared with controls.

Figure 4.

Hydroxyurea treatment improves anemia status via reduced hemolysis in sickle cell mice. Hemoglobin (A) (P_interaction = 0.02, P_treatment > 0.1, P_genotype < 0.001), hematocrit (B) (P_interaction < 0.01, P_treatment < 0.01, P_genotype < 0.001), plasma lactate dehydrogenase activity (C) (P_interaction > 0.1, P_treatment = 0.03, P_genotype < 0.01), and erythrocyte fetal hemoglobin (D) (P_interaction < 0.001, P_treatment < 0.001, P_genotype < 0.001) were measured. Two-way ANOVA with Tukey’s post hoc correction for multiple comparisons was used for HbAA and HbSS controls and hydroxyurea-treated mice. HbAA, genetic control mice; HbSS, sickle cell mice.

One of the primary benefits of hydroxyurea treatment in patients with SCD is reported to be induction of fetal hemoglobin, which can reduce the frequency of sickling and the frequency of vaso-occlusive crises, as well as improve mortality rates (24, 42–44). However, short-term hydroxyurea did not increase erythrocyte fetal hemoglobin production in HbSS mice (Fig. 4D). It is interesting to note that our 2-wk treatment regimen significantly increased fetal hemoglobin in HbAA control mice (Fig. 4D).

As hydroxyurea has been shown to have anti-inflammatory properties in patients with SCD, we assessed five inflammatory cytokines in plasma from HbAA and HbSS mice: sE selectin, sP selectin, soluble intracellular adhesion molecule-1 (sICAM-1), plasminogen activator inhibitor-1 (PAI-1), and matrix metallopeptidase-9 (proMMP-9). There were no significant differences in sE selectin, sP selectin, sICAM-1, and proMMP-9 plasma concentrations between groups, regardless of genotype or treatment (Fig. 5, A–C and E). However, plasma PAI-1 was significantly elevated in HbSS vehicle-treated mice compared with genetic controls and was subsequently reduced after hydroxyurea treatment in HbSS mice (Fig. 5D).

Figure 5.

Hydroxyurea treatment specifically reduces plasminogen activator inhibitor-1 (PAI-1) in sickle cell mice. Multiplex assay results for soluble E-selectin (A) (P_interaction > 0.1, P_treatment > 0.1, P_genotype > 0.1), soluble intracellular adhesion molecule-1 (ICAM-1) (B) (P_interaction > 0.1, P_treatment > 0.1, P_genotype > 0.1), soluble P-selectin (C) (P_interaction > 0.1, P_treatment > 0.1, P_genotype > 0.1), plasma PAI-1 (D) (P_interaction < 0.001, P_treatment < 0.001, P_genotype < 0.001), and plasma matrix metallopeptidase-9 (proMMP-9) (E) (P_interaction > 0.1, P_treatment > 0.1, P_genotype > 0.1). Two-way ANOVA with Tukey’s post hoc correction for multiple comparisons was used for HbAA and HbSS controls and hydroxyurea-treated mice. HbAA, genetic control mice; HbSS, sickle cell mice.

DISCUSSION

Our results provide evidence that hydroxyurea treatment improves systemic NO bioavailability, which can be partly attributed to reduced hemolysis along with lower renal vascular arginase activity as well as reduced PAI-1 levels. Although it is generally believed that loss of NO bioavailability is due to scavenging by reactive oxygen species (45), we provide further evidence that increased arginase activity and PAI-1 levels, which are attenuated with hydroxyurea treatment, could account for the loss of NO bioavailability.

Previous studies from our laboratory have extensively characterized the development of renal injury in the Towne’s mouse model of sickle cell disease (23, 36, 37). These mice develop proteinuria, albuminuria, glomerulomegaly, focal segmental glomerulosclerosis, tubular damage, and interstitial inflammation by 12 wk of age. Furthermore, we also have several publications showing that ET_A blockade (Ambrisentan) reduced these measures of renal injury, and more recently, hydroxyurea alone can attenuate renal injury as shown by histological analysis (23, 37). The current study sought to determine if improved NO bioavailability is a contributing mechanism to the therapeutic effects of hydroxyurea we observed in our previous study.

Chronic systemic NO depletion contributes to the extensive vasculopathy and end-organ damage seen in patients with SCD (10, 38). Our studies suggest prevalent systemic defects in NO signaling in SCD mice, indicated by suppressed plasma nitrite concentrations. Global dysregulation of NO signaling has already been implicated in the development of other comorbidities associated with SCD, such as pulmonary hypertension and endothelial dysfunction (46). Nonetheless, ours is one of the few studies to implicate decreased NO bioactivity in the pathogenesis of kidney disease in a humanized mouse model of SCD. It is clear that hemolysis can reduce NO signaling through heme scavenging mechanisms (10). However, whether this contributes to reductions in NO bioavailability in SCD has largely been driven by association of hemolysis with vascular disorders such as pulmonary hypertension (47). Although many investigators presume that hemolysis and oxidative stress reduces NO bioavailability in SCD (10), this has not been fully accepted because direct information on vascular NO production and signaling are lacking. In contrast to our results, Almeida et al. (48) recently reported that both humans with SCD as well as the Berkeley (BERK) mouse model of SCD have elevated blood nitrite and cGMP. Some of the recruited patients, however, were taking hydroxycarbamide during the study, which was confirmed by elevated HbF levels. Moreover, the BERK SCD mouse strains are notoriously difficult to manage without careful phenotyping, so compensatory mechanisms can easily arise to improve survival without monitoring reticulocyte count. At this point, however, we do not know whether differences in methods or mouse colonies could explain the contrasting findings between Almeida’s work and the current study.

Another novel observation we report here for the Townes humanized sickle cell mouse model is elevated plasma arginase activity and renal vascular arginase activity that can function to reduce substrate availability for NOS-dependent NO generation (38). In steady state, patients with SCD also experience decreased plasma concentrations of the NO precursor l-arginine and elevated arginase activity (8, 49). These phenomena can exacerbate NO depletion in SCD by promoting NOS uncoupling accompanied by subsequent ROS generation and NO consumption (50). Elevated plasma and vascular arginase activity with concomitant vasculopathy have been well established in earlier developed mouse models of SCD; however, until now, these pathways remained unexplored in the Townes mouse model (51). Similar to findings in other studies, our data show elevated plasma arginase activity in HbSS mice; however, our novel discovery of elevated arginase-2 activity in the renal vasculature has not been previously investigated (51, 52).

In our current study, we provide evidence that short-term treatment with hydroxyurea is sufficient to blunt arginase activity and reduce renal injury in SCD. Previous work has established fetal hemoglobin induction as the primary therapeutic mechanism by which hydroxyurea exerts its pleiotropic effects in both human and animal models (29, 44, 53). Recently, enhanced NO signaling has been proposed as a potential therapeutic pathway involved with hydroxyurea treatment in SCN (54). Inhaled NO therapy has been shown to provide rapid and significant improvement in pulmonary arterial pressure, systemic vascular resistance, and blood oxygenation in patients with SCD suffering from acute chest syndrome (15, 55, 56). Another study in pediatric patients with SCD reported reductions in opioid usage and pain severity during vaso-occlusive crises with inhaled NO therapy (54). Despite the many studies on NO therapy and its beneficial effects, few have investigated NO therapy in the setting of SCN. Both acute and chronic hydroxyurea administration reduce renal injury in SCD (22, 23). In addition, hydroxyurea treatment has also been shown to induce NO production in as little as 2 h postadministration in patients with SCD (28). Although some have proposed the benefits of hydroxyurea could be by serving as an NO donor (28), this has not gained widespread acceptance. We recently reported that limited hydroxyurea treatment (2 wk) reduced glomerular injury, renal cortical inflammation, and plasma ET-1 concentration in HbSS mice (23). Although hydroxyurea can increase fetal hemoglobin levels, we observed reduced arginase and increased NO bioavailability without elevating fetal hemoglobin. Our current study provides an alternative mechanism for oral, short-term hydroxyurea administration by reducing arginase activity as a means of augmenting systemic NO bioavailability and perhaps reducing risk of end-organ damage such as proteinuria. Our results also show reduced hemolysis, indicated by reduced plasma LDH concomitant with improved NO bioavailability. These data suggest that improved red cell stability occurs with short-term hydroxyurea therapy, which may be caused by the NO augmentation. Interestingly, improved systemic NO bioavailability was accompanied by variable, but significantly lower plasma arginase activity, consistent with the concept that hydroxyurea may improve substrate availability for NOS although it cannot be ruled out that hydroxyurea serves as a direct NO donor in circulation. Previous studies have identified the in vivo chemical oxidation of hydroxyurea as a potential NO-synthesizing mechanism in patients with SCD (28, 57, 58). Despite studies investigating the NO-donating capabilities of hydroxyurea, the complete mechanism has yet to be elucidated.

Arginase has two major isoforms expressed throughout the body: arginase-1 and arginase 2. Unlike arginase-1, which is a cytosolic, hepatic enzyme primarily responsible for total plasma arginase activity in the body, arginase 2 is a mitochondrial enzyme expressed predominately in the kidney (59). Arginase catalyzes the conversion of arginine, the substrate for NOS, to l-ornithine and urea (60). This reaction reduces bioavailability of the NOS substrate, which leads to NOS uncoupling, increased ROS, and reduced NO production (61, 62) (Fig. 6). Elevated plasma arginase activity concomitant with NO insufficiency has been well established in patients with SCD (12, 63, 64). In this study, we report elevated circulating arginase activity in untreated HbSS mice. This finding is substantiated by the observed concomitant decrease in plasma arginine concentration in these mice as well. Interestingly, there were no significant changes in plasma citrulline concentration regardless of treatment and genotype. Treatment with hydroxyurea did not restore plasma arginase activity or plasma arginine concentration to control levels. Taken together, these data suggest that reduced systemic arginase activity is not the primary mechanism responsible for the renal therapeutic effects of short-term hydroxyurea therapy in SCD.

Figure 6.

Impact of arginase on nitric oxide (NO) bioavailability in sickle cell disease (SCD). Deoxygenated, sickle-shaped red blood cells (RBCs) have compromised cell membrane integrity and often lyse in circulation. When lysed, these cells release free, mutant HbS hemoglobin and arginase into the microenvironment. Circulating arginase competes with nitric oxide synthase (NOS) for the arginine substrate, thereby depleting arginine reserves. In addition, free HbS sequesters generated NO. These processes both contribute to the reduced NO bioavailability observed in SCD patients. HbS, hemoglobin S; HbA, hemoglobin A.

Prior studies have also reported augmented vascular arginase activity in mouse models of SCD (51). In isolated renal vessels, our data show that hydroxyurea improves renal vascular arginase-2 expression and activity in SCD mice consistent with improved NO-generating ability. A study by Iyamu et al. (65) reported reduced erythrocyte arginase activity with an accompanied elevation in plasma NOS activity in patients with SCD on hydroxyurea therapy. Interestingly, patients on short-term hydroxyurea therapy (<3 mo) showed no correlation between fetal hemoglobin and arginase activity, despite having increased NOS activity (65). These findings, in addition to the data provided in this study, support the notion that therapeutic benefits associated with hydroxyurea therapy may not be entirely dependent on fetal hemoglobin induction. Further investigation is necessary to determine the exact mechanism behind hydroxyurea therapy and reduced arginase activity.

PAI-1, a profibrotic factor that promotes cell adhesion and fibrinolysis, is known to be elevated in patients with SCD during vaso-occlusive crises and in steady state (66–71). Conditions endemic to the sickle cell milieu, such as hypoxia and oxidative stress, are thought to be the cause of increased PAI-1 (72, 73). Interestingly, a recent study also determined that PAI-1 directly inhibits NOS3 activity and NO biosynthesis (74). In the current study, we observed elevated PAI-1 in plasma of HbSS mice that was attenuated with hydroxyurea treatment. Our untreated HbSS mice also displayed reduced plasma nitrite concentrations that was significantly improved with hydroxyurea treatment. These data are consistent with PAI-1 contributing reduced NO biosynthesis in SCD. These data also suggest that suppressed PAI-1 activity may be a contributing mechanism to the therapeutic benefits of hydroxyurea.

In our prior studies, we observed improvements in renal outcomes with the 2-wk hydroxyurea treatment (23). The current study demonstrates that this improvement occurs in the absence of fetal hemoglobin changes in the blood. Long-term hydroxyurea treatment has been demonstrated to improve clinical outcomes in patients with SCD as shown by our current findings (17–19). Although many of its therapeutic effects have been attributed to its fetal hemoglobin-inducing properties, studies have shown clinical benefits in patients with SCD independent of fetal hemoglobin (17, 65). Improved NO bioavailability has been established as a potential therapeutic mechanism for hydroxyurea (27, 28). In addition, NO administration and augmentation therapy have proven clinically beneficial in both patients and animal models of SCD (16, 75). Despite this evidence, the role of hydroxyurea and NO bioavailability in SCN has yet to be fully elucidated. Our study demonstrates that short-term hydroxyurea therapy reduces vascular arginase and proteinuria in the kidney. Both direct arginine therapy and arginase inhibition have been shown to improve endothelial dysfunction, hemolysis, and oxidative stress in preclinical models of SCD (51, 52). Therefore, it seems plausible that similar therapeutic mechanisms could be responsible for the reduction in renal injury we observed in this and previous studies (23). Although there does appear to be some subtle changes in NOS3 expression in the renal vasculature that could account for improved NO production, the dramatic effects on arginase would suggest that this is an underappreciated mechanism and beneficial effect of hydroxyurea therapy in SCD. Collectively, our data reveal new insight into the short-term therapeutic mechanisms of hydroxyurea and provides tangible targets for the development of future SCD therapies.

PERSPECTIVES AND SIGNIFICANCE

Although improving NO bioavailability has been used as a possible therapy for other disorders, it has never been directly investigated in the context of an animal model of SCN. The anti-inflammatory effects of NO, as well as its ability to promote vasodilation and enhance blood flow, makes it a prime mechanistic target to investigate in SCD. In our current study, we highlight the NO-enhancing capabilities of hydroxyurea as a potential mechanistic pathway in the treatment of SCN independent of the presumed effects on fetal hemoglobin induction. These findings may initiate studies examining the use of hydroxyurea to treat other diseases in which enhanced NO bioavailability may be beneficial. Our results may also provide a viable mechanistic pathway to be targeted by novel therapeutics in SCD by reducing arginase activity. Undoubtedly, more studies are required to fully elucidate the therapeutic role of improved NO bioavailability in SCN. However, this study may serve as a catalyst for the research community to begin delving deeper into this mystery.

GRANTS

This study is supported by the National Heart, Lung, and Blood Institute Grants U01 HL117684 (to D.M.P. and J.S.P.), K99 HL 144817 (to M.K.), F31 HL149235 (to L.S.D.), and F31 HL151264 (to P.A.M.); a Porter Fellowship from the American Physiological Society (to C.T.); the National Institute of General Medical Sciences Grant T32 GM 109780 (to L.S.D.); the National Institute of Diabetes and Digestive and Kidney Diseases Grant F31 DK111067 (to R.S.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

C.T., J.S.P., and D.M.P. conceived and designed research; C.T., M.K., R.S., P.A.M., and L.S.D. performed experiments; C.T., R.S., P.A.M., and L.S.D. analyzed data; C.T., M.K., R.S., P.A.M., L.S.D., J.S.P., and D.M.P. interpreted results of experiments; C.T. and P.A.M. prepared figures; C.T. drafted manuscript; C.T., M.K., R.S., P.A.M., L.S.D., J.S.P., and D.M.P. edited and revised manuscript; C.T., M.K., R.S., P.A.M., L.S.D., and D.M.P. approved final version of manuscript.