Renin-angiotensin-aldosterone system inhibition exacerbates anemia in sickle cell disease

Key Points

• •RAAS inhibition is associated with worsening anemia and may increase red blood cell transfusion requirements in SCD.
• •Erythroid progenitor function is suppressed by losartan and can be rescued with erythropoietin.

Visual Abstract

Abstract

Renin-angiotensin-aldosterone system inhibitors (RAASi) are suggested for treating albuminuria in patients with sickle cell disease (SCD). RAASi may exacerbate anemia in the general population and those with diabetes through unclear mechanisms. The impact of RAASi on anemia in chronic hemolytic disorders, such as SCD, is unknown. In a cross-sectional analysis of 658 Walk-PHaSST participants, RAASi use was independently associated with lower hemoglobin concentrations, adjusting for age, sex, SCD genotype, estimated glomerular filtration rate, and erythroid-stimulating agent and hydroxyurea use (β, −0.46 ± 0.21; P = .032). In 2 longitudinal cohorts (University of Illinois Chicago, n = 24; multicenter losartan clinical trial, n = 32), RAASi therapy led to a reduction in hemoglobin concentrations compared with pretreatment values by −0.44 ± 0.14 g/dL (P = .006) and −0.53 ± 0.17 g/dL (P = .005), respectively. SCD mice treated with losartan demonstrated lower hemoglobin concentrations after 6 and 14 weeks (P < .001) and lower absolute reticulocyte counts by 14 weeks (P = .03) vs control mice without changes in circulating erythropoietin, interleukin-12p70 (IL-12p70), IL-3, or insulin-like growth factor 1 levels. Bone marrow cells from losartan-treated SCD mice had lower colony-forming units (P ≤ .09) with rescue of erythroid colony formation after exogenous erythropoietin supplementation (P = .02). Bone marrow histopathology demonstrated reduced erythroid relative to myeloid ratios in losartan-treated vs untreated SCD mice. Hemoglobin levels should be closely monitored when using RAASi in this high-risk population.

End-stage renal failure (ESRF) is associated with major reduction in life span for people with sickle cell disease (SCD), but specific measures to obviate ESRF in SCD are unproven. American Society of Hematology guidelines cautiously recommend the use of renin-angiotensin-aldosterone system inhibitors (RAASi) when patients with SCD have albuminuria, based on evidence for benefit in other diseases that cause ESRF. Eskandari and colleagues report the impact of RAASi on hemoglobin concentrations in patients with SCD, finding that, across 3 studies, a mean 0.4- to 0.9-g/dL decrease occurs, mirroring reductions seen in non-SCD patients and also in SCD mice. Their data provide pause for clinicians caring for patients with SCD who are experiencing symptomatic anemia, but they also support future trials of these drugs as well as erythropoietin-stimulating agents.

Introduction

Sickle cell disease (SCD) is characterized by chronic hemolytic anemia with an average red blood cell (RBC) life span of only 30 to 60 days.^{1, 2, 3} Chronic kidney disease is a common complication, affecting ∼40% of adults with SCD.^4,5 Renin-angiotensin-aldosterone system inhibitors (RAASi) are suggested by the American Society of Hematology and National Institutes of Health guidelines for treating albuminuria, a biomarker that defines early stages of kidney disease, in this patient population.^6,7 Consequently, RAASi are prescribed to patients with SCD, often in the setting of preexisting anemia.^8,9

In some non-SCD populations, RAASi worsen anemia, an effect attributed to either reduced erythropoietin production or by inhibiting angiotensin II signaling.^10,11 Angiotensin II functions as a growth factor for erythroid progenitors¹² and augments erythropoietin-mediated erythropoiesis.¹³ Whether RAASi exacerbate anemia in SCD, a disorder marked by high baseline erythropoietic demand, and the potential mechanisms for the reduced erythropoiesis remain unclear.

Here, we investigated the relationship between RAASi use and anemia in SCD using complementary clinical and experimental approaches. We analyzed this in a large cross-sectional and 2 longitudinal cohorts of patients with SCD and in transgenic SCD mice to test the hypothesis that RAASi worsen anemia through angiotensin II–dependent effects on erythroid progenitor function.

Study design

Clinical data

The studies were approved by the institutional review boards of the participating institutions, and patients provided written informed consent in accordance with the Declaration of Helsinki. Cross-sectional analyses were performed using baseline data from the multicenter Walk-PHaSST cohort (ClinicalTrials.gov identifier: NCT00492531)¹⁴ and longitudinal analyses in patients enrolled in a longitudinal registry at the University of Illinois Chicago (UIC)⁵ and from the multicenter, phase 2 losartan clinical trial (ClinicalTrials.gov identifier: NCT01479439).¹⁵ Patients from the Walk-PHaSST cohort with RAASi (n = 58) and hemoglobin data (n = 658) were included. Patients from UIC with at least 12 months of RAASi use with laboratory and clinical data available 12 months before and on treatment (n = 24) and who were not on chronic RBC transfusions were included. All 32 evaluable participants from the losartan clinical trial were analyzed, and eligibility criteria are provided in supplemental Table 1 (available on the Blood website).

Hemoglobin concentrations, erythropoietin levels, and reticulocyte values were compared by RAASi use in Walk-PHaSST participants by analysis of variance (ANOVA), adjusting for age, sex, SCD genotype, estimated glomerular filtration rate (eGFR), and erythroid-stimulating agent (ESA) and hydroxyurea use. Hemoglobin concentrations and reticulocyte values were compared in patients before RAASi use (UIC: closest value within the preceding 3 months; losartan clinical trial: average of screening and visit 1 measurements) vs on RAASi therapy (UIC: after 12 ± 3 months of therapy; losartan clinical trial: at 8 weeks and 6 months of therapy) using the paired t test or the Wilcoxon signed-rank test. In the UIC cohort, individual slopes for hemoglobin concentration were calculated by linear regression over time and compared pre-RAASi vs during RAASi treatment. A control group of UIC patients not on RAASi (n = 24) matched by age, sex, SCD genotype, hemoglobin concentration, and albuminuria was identified to compare hemoglobin slopes (supplemental Table 2). In the losartan clinical trial, all available hemoglobin values were analyzed using linear mixed-effects models. Time was defined as weeks from baseline and modeled as a continuous fixed effect. A patient-specific random intercept was included to account for within-patient correlation.

Mouse studies

All animal procedures were conducted under institutional animal care and use committee–approved protocols at UIC. Transgenic SCD mice (Townes model; The Jackson Laboratory, Bar Harbor, ME) were treated with losartan (0.2 mg/mL; TCI, Portland, OR) administered in drinking water beginning at the age of 12 weeks. Water consumption was monitored throughout the treatment period and did not differ between losartan-treated and control groups. Blood was collected retro-orbitally at baseline and after 6 and 14 weeks of treatment to compare hemoglobin concentrations and absolute reticulocyte counts with those of untreated SCD mice (6 males and 6 females per group). Plasma concentrations of erythropoietin (R&D System, Minneapolis, MN), interleukin-12p70 (IL-12p70; R&D Systems), IL-3 (Invitrogen, Waltham, MA), and insulin-like growth factor 1 (IGF-1; R&D Systems) were measured at 14 weeks of treatment by enzyme-linked immunosorbent assay. Comparisons between losartan and control groups were performed using ANOVA, adjusting for sex.

In a separate set of experiments, we harvested bone marrow cells from the tibias and right femurs of losartan-treated and untreated SCD mice after 4 weeks of exposure (2 males and 2 females per group). Cells were plated in methylcellulose-based media (StemCell Technologies, Vancouver, BC, Canada) for colony-forming unit (CFU) assays, and colonies were enumerated after 14 days of incubation. To assess the rescue of erythroid progenitor function, bone marrow cells from losartan-treated mice were cultured with erythropoietin (10 ng/mL; R&D Systems, Minneapolis, MN)¹⁶ or angiotensin II (5μM or 10μM; Sigma Aldrich, St. Louis, MO).¹⁷ Erythroid (CFU-erythroid [CFU-E] + burst-forming unit-erythroid [BFU-E]) and granulocyte/macrophage (CFU-GM) colonies were quantified as previously described.¹⁸ Left femurs were harvested after 4 weeks of treatment, fixed in formaldehyde, decalcified, and processed for hematoxylin and eosin staining for qualitative histologic assessment by a hematopathologist blinded to treatment arms. Comparisons between groups were performed using ANOVA.

Results and discussion

Clinical data

Baseline characteristics of the SCD cohorts are provided in supplemental Table 3. In the Walk-PHaSST cohort, patients with SCD on RAASi were older and had a lower eGFR than those not on RAASi (Table 1). Patients with SCD on RAASi had significantly lower hemoglobin concentrations than those not on RAASi after adjusting for age, sex, hydroxyurea use, ESA use, SCD genotype, and eGFR (β, −0.46 ± 0.21; P = .032). We did not observe significant differences in absolute reticulocyte counts, reticulocyte percentages, serum erythropoietin concentrations, or serum erythropoietin-to-hematocrit concentrations by RAASi use (P ≥ .5; Table 1).

Table 1.

Clinical and laboratory characteristics by RAASi use in the Walk-PHaSST cohort

	No RAASi (n = 600)	RAASi (n = 58)	P value
Age, y	35 (25-46)	42 (33-52)	.002
Male-to-female ratio, %	47:53	53:47	.3
Hb SS/β0-thalassemia, %	454 (76)	49 (84)	.1
Hydroxyurea use, %	220 (37)	26 (45)	.2
ESA use, %	16 (3)	2 (3)	.7
eGFR, mL/min per 1.73 m2	134 (106-152)	112 (71-139)	.0001
Hemoglobin, g/dL	9.3 (8.0-10.8)	8.4 (7.3-10.0)	.001
Reticulocyte, %	7.6 (4.1-11.7)	7.6 (4.9-11.6)	.5
Absolute reticulocyte count, ×109/L	218 (143-318)	212 (137-320)	.7
Erythropoietin level, mU/mL	62 (38-105)	59 (44-110)	.7
Erythropoietin-to-hematocrit ratio	3.1 (1.5-25.0)	3.4 (2.1-24.0)	.9

Median and interquartile range (IQR) values are provided.

Hb SS, hemoglobin SS.

Consistent with these findings, longitudinal analyses in both the UIC cohort and losartan clinical trial demonstrated a decline in hemoglobin concentrations with RAASi therapy. In the UIC cohort, the hemoglobin concentration declined by a mean of −0.44 ± 0.14 g/dL after 12 months of RAASi therapy (Figure 1A). There was an accelerated slope of hemoglobin decline with RAASi when using all hemoglobin values in the 12 months preceding vs on RAASi treatment (β, −0.31; 95% confidence interval, −0.55 to −0.06; P = .016) and when compared to hemoglobin slopes in similar patients with SCD not treated with RAASi (β, −0.66; 95% confidence interval, −1.23 to −0.08; P = .027). New RBC transfusion requirements were observed in 7 of the 24 (29%) patients while on RAASi therapy (Figure 1B). We did not observe a difference in absolute reticulocyte counts or reticulocyte percentages with RAASi therapy (Figure 1C). In the losartan clinical trial, hemoglobin values declined by a mean of −0.51 ± 0.10 g/dL and −0.53 ± 0.17 g/dL after 8 weeks and 6 months of therapy, respectively (Figure 1D). Reticulocyte percentage was unchanged after 8 weeks but increased by 2.4% ± 0.7% after 6 months of losartan treatment (Figure 1E). A significant decline in hemoglobin concentration over time was observed in the losartan clinical trial using all hemoglobin values in the mixed-effects model (P = .01; supplemental Figure 1).

Figure 1.

Effects of RAASi on anemia in patients with SCD. (A-C) In the UIC longitudinal cohort, hemoglobin concentration declined from a median of 8.5 g/dL (interquartile range [IQR], 7.9-9.5) before RAASi to 8.1 g/dL (IQR, 7.4-9.2) on RAASi therapy (A), new RBC transfusion requirements were observed in 7 of the 24 patients (B), with no significant changes in reticulocyte percentages (before RAASi: 8.8% [IQR, 5.0%-13.4%]; on RAASi: 8.2% [IQR, 4.9%-12.7%]) (C). (D-E) In the multicenter losartan clinical trial, a similar pattern for a decline in hemoglobin concentration (baseline: 8.6 g/dL [IQR, 7.8-9.8]; 8 weeks: 8.1 g/dL [IQR, 7.4-8.9]; and 6 months: 8.3 g/dL [IQR, 7.7-9.5]) was observed (D), whereas no significant change in reticulocyte percentage at 8 weeks but an increase at 6 months (baseline: 9.3% [IQR, 6.2%-13.1%]; 8 weeks: 10.2% [IQR, 7.0%-17.6%]; and 6 months: 12.2% [IQR, 7.5%-16.2%]) of losartan therapy was observed (E). Comparisons in the UIC cohort are based on 1 year of RAASi therapy. ns, not significant.

Mouse studies

Transgenic SCD mice were treated with losartan for 14 weeks. Losartan reduced hemoglobin levels after 6 and 14 weeks of therapy (P < .001; Figure 2A; supplemental Table 4). Absolute reticulocyte counts were significantly lower after 14 weeks (P = .03; corresponding to age of 26 weeks; Figure 2B) in the losartan vs control mice, whereas circulating erythropoietin and erythropoietin-to-hematocrit concentrations were relatively similar between the 2 groups (Figure 2C).

Figure 2.

Losartan effects on systemic erythropoiesis and hematopoietic progenitor function in transgenic SCD mice. (A-C) Changes in hemoglobin concentration (A) and absolute reticulocyte counts (B) after 6 and 14 weeks of treatment (corresponding to the age of 18 and 26 weeks, respectively) and EPO/HCT after 14 weeks (C) in losartan-treated vs untreated SCD mice. (E-F) Plasma levels of IL-12p70 (D), IL-3 (E), and IGF-1 (F) after 14 weeks (corresponding to the age of 26 weeks) in losartan-treated vs untreated mice. (G-H) CFU assays performed after 4 weeks of losartan treatment measured CFU-GM (G) and erythroid progenitors (BFU-E & CFU-E) (H) in losartan-treated vs untreated mice. Rescue of progenitors by exogenous EPO or angiotensin II is shown. (I) Bone marrow histopathology demonstrates balanced myeloid-to-erythroid progenitor ratios in control mice and relative erythroid hypoplasia in losartan-treated mice (hematoxylin and eosin stain; original magnification ×100). The black arrows indicate erythroid progenitors. ∗P < .05 and ∗∗P < .001 for differences between groups. Ang II, angiotensin II; BFU-E, burst-forming unit-erythroid; CFU-E, colony-forming unit-erythroid; EPO/HCT, erythropoietin-to-hematocrit ratio.

Because angiotensin II may influence erythropoiesis indirectly through systemic mediators,¹⁹ plasma IL-12p70, IL-3, and IGF-1 levels were measured after 14 weeks of losartan treatment. IL-3 and IGF-1 promote erythroid progenitor proliferation and differentiation,^{20, 21, 22, 23} whereas IL-12p70 is a pleiotropic cytokine with immunoregulatory effects²⁴ that positively affects hematopoietic progenitors.²⁵ None of these mediators were significantly different between losartan-treated vs control mice (Figure 2D-F; supplemental Table 4). CFU assays performed after 4 weeks of losartan treatment demonstrated a significant reduction in CFU-GM colonies and a trend for lower erythroid progenitor colonies (BFU-E + CFU-E; P = .09) vs control SCD mice (Figure 2G-H; supplemental Table 5). Exogenous erythropoietin partially rescued erythroid colony formation (P = .02), whereas exogenous angiotensin II led to nonsignificant increases in erythroid colonies; CFU-GM colonies were not rescued by either treatment (Figure 2G-H).

Qualitative bone marrow histopathology revealed preserved trilineage hematopoiesis in losartan-treated and control SCD mice (Figure 2I). In control mice, marrow cellularity ranged from 95% to 100%, with balanced myeloid and erythroid maturation and myeloid-to-erythroid ratios of 3:1 to 4:1. In contrast, losartan-treated mice demonstrated a relative shift toward myeloid predominance and relative erythroid hypoplasia, reflected by increased myeloid-to-erythroid ratios ranging from 6:1 to 8:1. Myeloid maturation remained intact, megakaryocytes were adequate in number, and no morphologic dysplasia was identified in any specimen.

Collectively, these clinical and experimental data demonstrate that RAAS inhibition exacerbates anemia, and the effects may be through reduced erythropoietic progenitor function. In 3 different clinical cohorts, we observed a reduction in hemoglobin concentration from 0.4 to 0.9 g/dL. Although this degree of anemia may seem mild, a further decline in hemoglobin concentration in the setting of a chronic hemolytic disorder such as SCD may not be as well tolerated. In a large meta-analysis of patients with SCD, a hemoglobin concentration that was 0.4 g/dL lower was observed in those with vs without cerebrovascular disease and 0.6 g/dL lower in those with vs without albuminuria.²⁶ Our study is limited by the relatively small longitudinal cohort sample sizes and RBC transfusion data without clinical context available in only 1 of the 3 cohorts. We corrected for eGFR in the Walk-PHaSST cohort and used paired analyses in our longitudinal analyses, but other factors, such as worsening renal function, may have contributed to hemoglobin decline. Although we did not observe significant changes in erythropoietin levels with RAASi in SCD mice or in the Walk-PHaSST cohort, erythropoietin supplementation improved erythroid colony formation, suggesting inadequate erythropoietin production with RAASi therapy. Larger studies are needed to evaluate the effects of different classes of RAASi on anemia, along with the impact of concurrent SCD treatments currently available (ESA and hydroxyurea) or in development. Nevertheless, our findings emphasize the need for close hemoglobin monitoring when RAASi are implemented in patients with SCD. Strategies, such as erythropoietin supplementation to overcome the negative effects of RAASi on erythroid progenitors, may help balance the renoprotective effects of RAASi and should be explored in future studies.

Conflict-of-interest disclosure: V.R.G. has served as a consultant for Emmaus, Global Blood Therapeutics/Pfizer, Vertex, and Modus Therapeutics and received research funding from Incyte and Novartis. S.L.S. has served on advisory boards or as a consultant for Agios, Beam Therapeutics, Forma/Novo Nordisk, Novartis, Global Blood Therapeutics/Pfizer, Merck, Chiesi, and Fulcrum. The remaining authors declare no competing financial interests.