Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Aug 1.
Published in final edited form as: Br J Haematol. 2021 Jul 5;194(3):617–625. doi: 10.1111/bjh.17663

Early initiation of hydroxyurea (hydroxycarbamide) using individualised, pharmacokinetics-guided dosing can produce sustained and nearly pancellular expression of fetal haemoglobin in children with sickle cell anaemia

Charles T Quinn 1,2,3, Omar Niss 1,2, Min Dong 2,4, Amanda Pfeiffer 1, Jennifer Korpik 1,3, Mary Reynaud 1,3,5, Holly Bonar 1,5, Theodosia A Kalfa 1,2,3, Luke R Smart 1,2,6, Punam Malik 1,2, Russell E Ware 1,2, Alexander A Vinks 2,4, Patrick T McGann 1,2
PMCID: PMC8319147  NIHMSID: NIHMS1713684  PMID: 34227124

Abstract

Hydroxyurea (hydroxycarbamide) is an effective treatment for sickle cell anaemia (SCA), but clinical responses depend primarily upon the degree of fetal haemoglobin (HbF) induction and the heterogeneity of HbF expression across erythrocytes. The number and characteristics of HbF-containing cells (F-cells) are not assessed by traditional HbF measurements. Conventional hydroxyurea dosing (e.g., fixed doses or low starting doses with stepwise escalation), produces a moderate heterocellular HbF induction, but haemolysis and clinical complications continue. Robust, pancellular HbF induction is needed to minimise or fully inhibit polymerisation of sickle haemoglobin. We treated children with hydroxyurea using an individualised, pharmacokinetics-guided regimen starting at predicted maximum tolerated dose (MTD). We observed sustained HbF induction (mean >30%) for up to 6 years, which was not dependent on genetic determinants of HbF expression. Nearly 70% of participants had ≥80% F-cells (near-pancellular), and almost half had ≥90% F-cells (pancellular). Mean HbF/F-cell content was approximately 12 pg. Earlier age of initiation and better medication adherence were associated with high F-cell responses. In summary, early initiation of hydroxyurea using pharmacokinetics-guided starting doses at predicted MTD can achieve sustained near-pancellular or pancellular HbF expression and should be considered an achievable goal for children with SCA treated with hydroxyurea at optimal doses.

Keywords: sickle cell anaemia, hydroxycarbamide, hydroxyurea, pharmacokinetics, fetal haemoglobin, flow cytometry, F-cell

Introduction

Sickle cell anaemia (SCA) is a common and life-threatening inherited disorder of haemoglobin (Hb). Without early initiation of disease-modifying therapy and lifelong treatment, SCA causes significant morbidity beginning in childhood and early mortality. The polymerisation of sickle Hb (HbS) upon deoxygenation is the primary pathophysiologic event in SCA that damages red blood cells (RBCs), producing inflexible, adhesive, dehydrated, and sickled forms. This ongoing RBC damage causes chronic, uncompensated haemolytic anaemia, vaso-occlusion, numerous acute and chronic clinical complications, progressive organ damage, and premature death.

Fetal haemoglobin (HbF) is a potent and the best-characterised inhibitor of HbS polymerisation, and robust induction of HbF production has been and remains a “holy grail” of SCA therapy (Saunthararajah, 2019; Telen et al, 2019; Eaton & Bunn, 2017). Hydroxyurea, the most widely available and well-established pharmacologic therapy for SCA, exerts its greatest therapeutic effects through the induction of HbF. With traditional dosing (using fixed doses or low starting doses with gradual, stepwise escalation), hydroxyurea therapy produces HbF levels of 10–25%, on average, which is sufficient to reduce the frequency of many overt complications of SCA. At these HbF levels, however, patients may still be symptomatic with ongoing haemolysis and progressive organ damage.

Traditional Hb separation and quantitation methods, such as high-performance liquid chromatography (HPLC) and Hb electrophoresis, quantify the percentage of HbF (%HbF), but that value is an average of the HbF content across all RBCs, including cells that contain HbF and cells with no detectible HbF. Variation in the cell-specific compartmentalization of HbF across RBCs is not assessed; this important information that quantifies the fraction of RBCs protected from HbS polymerisation is lost. In contrast, flow cytometric RBC analysis can determine the fraction of HbF-containing RBCs (F-cells) in the entire RBC population, enumerate HbF-containing reticulocytes (F-retics), and estimate the content of HbF per F-cell (F/F-cell).

In untreated SCA, HbF expression is variable and distributed unequally across RBCs (heterocellular expression with 10–50% F-cells) (Dover et al, 1978; Urio et al, 2016; Marcus et al, 1997; Steinberg et al, 1997). Traditional hydroxyurea dosing strategies raise HbF with usually heterocellular induction (20–90% F-cells) (Dai et al, 2017; Maier-Redelsperger et al, 1998; Steinberg et al, 1997). In contrast, compound heterozygosity for HbS and gene-deletion hereditary persistence of HbF [HbS/(δβ)0-HPFH] results in HbF levels of 30–40%, but with a pancellular distribution of HbF across RBCs (100% F-cells) (Wood et al, 1975; Marcus et al, 1997). Individuals with HbS/(δβ)0-HPFH have no or rare clinical manifestations of SCA, demonstrating the importance of the pancellular distribution of HbF (Ngo et al, 2012).

Several recent publications have suggested that curative therapies should aim to achieve HbF>30%, F-cells>70%, and >4–10 pg F/F-cell (Hebert et al, 2020; Buchanan, 2014; Steinberg et al, 2014). We have demonstrated that early initiation of hydroxyurea using individualised, pharmacokinetics (PK)-guided dosing results in robust and sustained HbF levels beyond 30–40% for most adherent patients (Dong & McGann, 2021; McGann et al, 2019). Here, we describe the utility of flow cytometric F-cell analysis and demonstrate sustained and near-pancellular or pancellular distribution of HbF in this cohort. We show that hydroxyurea, when given in this manner, can achieve or exceed the goals defined for “curative therapies” for SCA (Hebert et al, 2020; Buchanan, 2014; Steinberg et al, 2014).

Methods

The TREAT Study

This is a secondary, laboratory-focused analysis of The Therapeutic Response Evaluation and Adherence Trial (TREAT, NCT 02286154) (McGann et al, 2019; Dong et al, 2016). A detailed description of the primary study design, analysis, and reporting is published separately (McGann et al, 2019). Briefly, given the known inter-patient variability in the PK, pharmacodynamics, and dosing of hydroxyurea, TREAT was designed as a single-centre, non-randomised, open-label study designed to prospectively validate a novel individualised PK-guided hydroxyurea dosing strategy (Dong et al, 2016) with the primary endpoints of optimising dose and reducing time to maximum tolerated dose (MTD) (McGann et al, 2019). Enrolled participants received a single 20 mg/kg oral dose of liquid hydroxyurea, followed by sparse PK sampling with 3 samples collected at 3 time points over 3 hours. Hydroxyurea concentrations were incorporated into a population PK model to generate an optimal, patient-specific starting dose at the predicted MTD (see Supplementary Methods for details) (McGann et al, 2019). Following initiation of hydroxyurea at this PK-guided dose, dose escalation occurred per standard guidelines to target mild myelosuppression [absolute neutrophil count (ANC) 1–3.0×109/L]. In addition to the primary endpoint of dosing efficiency, TREAT includes serial monitoring of haematologic parameters and organ function in ongoing long-term follow-up studies. Standard clinical laboratory testing included serial complete blood counts (CBC) and Hb fractionation by capillary zone electrophoresis (CZE; Capillarys 2, Sebia, France). Metrics of adherence to hydroxyurea were also prospectively assessed (Supplementary Methods). All Emergency Department (ED) visits and hospitalizations were recorded during the study period and classified. An ED visit that resulted in a hospitalisation was classified as hospitalization. ED visit and hospitalisation rates were calculated using each participant’s follow-up time in the study in years.

There were 51 participants in TREAT, and we studied here all participants who had F-cell analysis performed at least once (N=48). Informed consent or assent was obtained from all participants or guardians. The study was conducted according to guidelines of the Declaration of Helsinki and the International Conference on Harmonisation.

F-cell Analysis

F-cell analysis was performed using an adaptation of a previously-described multiparametric flow cytometry method (Davis & Davis, 2004) Briefly, whole blood samples were fixed, permeabilised, and labeled with antibodies to HbF (PE-conjugated), CD235a (Glycophorin A; FITC-conjugated), and CD71 (transferrin receptor; BV421-conjugated). Flow cytometric analysis was performed using the BD FACSLyric Clinical System (BD Biosciences, San Jose, CA, USA). Patient samples were accompanied by two separate control samples: a healthy adult and a manufactured high HbF control. The total number of F-cells and F-retics (CD71+ F-cells) were enumerated. Mean concentration of F/F-cell for the entire F-cell population was calculated, but not individual cell F/F-cell values, so the variance of patients’ F/F-cell estimates is not known (Supplementary Methods). F-cell analyses were performed approximately yearly. For participants who enrolled on TREAT (beginning December 2014) before the availability of the F-cell analysis assay (beginning January 2019), no pre-hydroxyurea F-cell data were available and only traditional HbF measurements were obtained until January 2019. Here we define near-pancellular expression as total F-cells of 80–90%, which is greater than a suggested target for “curative” therapies for SCA (70% F-cells). We define pancellular as total F-cells >90%.

Genetic Testing

All patients had globin gene analysis to confirm their SCA genotype, determine α-thalassemia status, and exclude (δβ)0-HPFH. HBB, HBA1 and HBA2 were directly sequenced. Copy number variation analysis of the α-globin and β-globin gene clusters was performed by multiplex ligation-dependent probe amplification (MRC Holland, Amsterdam, The Netherlands). We genotyped three quantitative trait loci (QTL) associated with HbF levels in peripheral blood: the XmnI polymorphism in the HBG2 promoter (rs7482144); the erythroid specific enhancer of BCL11A (rs11886868, rs1427407, rs4671393); and the HBS1L-MYB intergenic region (rs28384513, rs9399137) (Cardoso et al, 2014; Chaouch et al, 2016; Lettre et al, 2008)

Statistical Analysis

Summary statistics, correlations, group comparisons, linear regression, and linear mixed modeling (for missing values) were performed with SPSS v27.0 (IBM Corp., Somers, NY). P-values <0.05 from two-sided tests was considered statistically significant. No study-wide correction was made for multiple comparisons. Figures were generated with Prism 9.0 (GraphPad Software, Inc., La Jolla, CA).

Results

Participants

We studied 48 patients with SCA (Table). None had (δβ)0-HPFH. There was an expected distribution of the common α-globin genotypes. Median age at start of hydroxyurea was 11.8 months (mean 40.2; min-max 6.4 – 234.6), of whom 52% (25/48) started before 1 year of age, 75% (36/48) before 3 years, and 81% (39/48) by 5 years. Median hydroxyurea dose was 27.6 mg/kg (mean 27.1; 17.8 – 38.6). Median duration of hydroxyurea therapy was 3.9 years (mean 3.8; 1.7 – 6.0), providing 183.6 patient-years of observation.

Table.

Characteristics of Participants.

Characteristic Median (mean; min – max) or Number (%)
Age
 Start of hydroxyurea (months) 11.8 (40.2; 6.4 – 234.6)
 Last follow-up (years) 6.0 (7.2; 2.3 – 21.9)
Sex
 Male 27 (56)
 Female 21 (44)
β-globin genotype
 βS/βS 46 (96)
 βS/β0 2 (4)
α-globin genotype
 αα/αα 30 (63)
 αα/–α 12 (25)
 −α/–α 4 (8)
 αα/ααα 2 (4)
Open in a new tab

Sustained haematologic benefits

Figure 1 shows baseline and follow-up blood counts. Hydroxyurea produced sustained increases in Hb, haematocrit (Hct) and mean corpuscular volume (MCV) and sustained reduction of absolute reticulocyte count (ARC), red cell distribution width (RDW), ANC, and platelet count. Mean Hb concentrations were ≥9.5 g/dL at all post-baseline timepoints. Even though this group includes many children who started hydroxyurea in the first year of life, a time when the Hb concentration is still high but falling, hydroxyurea therapy did not simply prevent its decline; rather, treatment increased Hb and Hct and decreased ARC. These laboratory benefits were sustained for up to 6 years with no evidence of waning effect over time.

Figure 1. Sustained haematologic benefits of individualised, PK-guided dosing of hydroxyurea.

Figure 1.

Open in a new tab

Yearly complete blood count measurements are shown from study timepoints beginning with the first available F-cell analysis. Timepoint B is the baseline, pre-hydroxyurea value (when available). Timepoints 1, 2 and 3 are the first and subsequent yearly post-baseline values measured on hydroxyurea therapy. Panel A shows Hb concentration. Panel B shows absolute reticulocyte count (ARC). Panel C shows MCV. Panel D shows red cell distribution width (RDW). Panel E shows absolute neutrophil count (ANC). Panel F shows platelet count. For all panels, individual participant’s measurements are shown as grey diamonds connected by grey lines across timepoints. The mean for all participants at each timepoint is shown as a black circle connected by a black line across timepoints. The light gray rectangle depicts the interquartile range (IQR) for the group at that timepoint.

Sustained induction of HbF

Before starting hydroxyurea, the mean baseline HbF was 24.8% (1.6–44.6). Baseline HbF was inversely correlated with age at start of hydroxyurea (r = −0.67, P < 0.001). Despite the high baseline HbF values of this group of mainly young children, hydroxyurea therapy produced a robust induction of HbF (Figure 2, Panel A). Hydroxyurea did more than prevent the expected decline in HbF with age; instead, it increased HbF values at all ages. Figure 3 illustrates this effect in one study participant. Robust HbF induction was maintained beyond 5 years of age for most adherent study participants, an age at which the switch from fetal to adult Hb should be complete. Mean HbF concentrations were ≥31.5% at all post-baseline timepoints (Figure 2, Panel A). HbF concentration is shown by starting age of hydroxyurea across study timepoints in Supplementary Figure 1.

Figure 2. Sustained HbF and F-cell responses.

Figure 2.

Open in a new tab

Yearly traditional HbF measurements and paired F-cell parameters are shown from study timepoints beginning with the first available F-cell analysis. Data are shown from study timepoints beginning with the first available F-cell analysis. Timepoint B is the baseline, pre-hydroxyurea value (when available). Timepoints 1, 2 and 3 are the first and subsequent yearly post-baseline values measured on hydroxyurea therapy. Panel A shows traditional HbF measurements (electrophoresis). Panel B shows the total F-cell fraction (percentage of all RBCs that contain detectable HbF). Panel C shows the F-retic fraction [percentage of all immature (CD71-expressing) reticulocytes that contain detectable HbF]. Panel D shows the mean F/F-cell estimates (mean HbF content per F-cell across the total F-cell population). For all panels, individual participant’s measurements are shown as grey diamonds connected by grey lines across timepoints. The mean for all participants at each timepoint is shown as a black circle connected by a black line across timepoints. The light gray rectangle depicts the interquartile range (IQR) for the group at that timepoint.

Figure 3. Example TREAT participant.

Figure 3.

Open in a new tab

Panel A shows the expected decline of HbF in the first 6 months of life. Hydroxyurea was started at 7 months of age, producing a robust and sustained increase in HbF despite the expected decline without treatment. Traditional HbF measurements (electrophoresis) are shown by the points and connecting line. Paired total F-cell measurements are indicated for 3 points. Panel B shows essentially pancellular expression of HbF (99% total F-cells) at the time of the traditional HbF measurement at 21 months (HbF 66%). Panel C highlights only the immature (CD71 expressing) reticulocyte fraction. This demonstrates that nearly 80% of reticulocytes are F-cells (F-retics), indicating that the main explanation for the very high peripheral F-cell number is increased production of F-cells rather than peripheral survival advantage and secondary enrichment of F-cells. For Panels B-C, fluorescence intensity of conjugated antibodies against HbF and CD71 are shown on the x- and y-axes, respectively.

Near-pancellular and pancellular expression of HbF

Serial F-cell analyses were performed approximately yearly, up to 3 per participant, for a total of 107 F-cell analyses (Figure 2, Panels B-D). F-cell results at baseline (before hydroxyurea) were measured for only 5 participants because this assay became available at our institution after TREAT enrollment began. Median total F-cells at baseline was 38.9%. With hydroxyurea therapy, the median total F-cells at all study timepoints was >80% (Figure 2, Panel B). At the first post-baseline timepoint, 69% of participants had at least near-pancellular HbF (F-cells ≥80%), and 46% had pancellular HbF (F-cells ≥90%). Similar high proportions of F-cells were seen across all post-baseline timepoints, indicating that sustained, pancellular HbF expression was achieved in nearly half of participants with hydroxyurea therapy (Figure 2, Panel B; Supplementary Table).

Clinical features associated with near-pancellular and pancellular expression of HbF

Participants whose F-cell fraction was <80% (heterocellular HbF expression) on treatment were older when they initiated hydroxyurea (mean 6.7 vs 1.8 years of age, P<0.001) and at the time of first F-cell analysis (5.6 vs. 3.9 years, P<0.001) compared to those who achieved an F-cell fraction ≥80% (at least near-pancellular expression). Total F-cell fraction is shown by starting age of hydroxyurea across study timepoints in Supplementary Figure 1. Age of initiation of hydroxyurea was linearly correlated with treatment F-cell fraction (r2 = 0.26, P=0.0002; Supplementary Figure 2). Participants who initiated hydroxyurea before age 3 years were more likely to have near-pancellular or pancelluar HbF expression than those who started later (Supplementary Figure 2).

Participants with total F-cells <80% had lower mean monthly number of hydroxyurea refills (0.80 vs 0.92, P=0.041; Supplementary Figure 3) and lower median self-reported measures of hydroxyurea adherence (4.5 vs 5, P=0.011; Supplementary Figure 3). No significant differences were found for sex, SCA genotype, or current hydroxyurea dose. The relatively high dose (MTD) of all participants (mean 27.1 mg/kg) likely explains the lack of relationship between hydroxyurea dose and F-cell response.

Increased production of F-reticulocytes

Both the mean and median F-reticulocyte percentages were approximately 50% across all post-baseline timepoints (Figure 2, Panel C), increased from approximately 25% in baseline samples. This indicates that a large determinant of the high total F-cell number with hydroxyurea therapy is the substantially increased production of HbF-expressing RBCs with further secondary enrichment by survival advantage of F-cells (rather than secondary enrichment as the predominant mechanism). F-reticulocyte fraction is shown by starting age of hydroxyurea across study timepoints in Supplementary Figure 1.

Increased F/F-cell content

We estimated the mean F/F-cell for each participant (Figure 2, Panel D). Prior studies indicate that an F/F-cell of 10 pg in an individual RBC can completely prevent its sickling, though others postulate that 4 pg per RBC may be functionally sufficient (Steinberg et al, 2010; Hebert et al, 2020) At baseline, mean F/F-cell was 7.4 pg. Only 1 participant had a mean F/F-cell at baseline ≥10 pg (10.2 pg). At the three timepoints on hydroxyurea therapy, mean F/F-cell values were 12.1, 12.5, and 14.1 pg. The percentage of participants with a mean F/F-cell ≥10 pg was 70.1 (34/48), 72.5 (29/40), and 92.9 (13/14) across these timepoints. F/F-cell content is shown by starting age of hydroxyurea across study timepoints in Supplementary Figure 1. A mean F/F-cell value of 10 pg does not indicate that all F-cells have 10 pg HbF, because there is an unmeasured variance around that mean F/F-cell value for each patient. Nevertheless, these data indicate that most participants have F-cells with a highly protective amount of F/F-cell.

Laboratory correlates high F-cell and F/F-cell responses

There was a curvilinear relationship between HbF and total F-cells (Figure 4, Panel A), similar to prior studies (Marcus et al, 1997; Steinberg et al, 1997). For all paired measurements of HbF and total F-cells, when HbF was ≥30% then nearly all (95%) F-cell measurements were ≥80%. When HbF was ≥35%, all post-baseline F-cell measurements were ≥80%. This indicates that a HbF measurement of at least 30% is a reasonable proxy for a marked increase in F-cells in peripheral blood (to at least 80% of all RBCs), consistent with prior studies in untreated SCA patients (Marcus et al, 1997), while HbF ≥35% gives a very high certainty of an F-cell response ≥80% in hydroxyurea-treated patients.

Figure 4. HbF and F-cell responses.

Figure 4.

Open in a new tab

Baseline (before hydroxyurea treatment) values are shown in red; all values obtained during hydroxyurea therapy are shown in black. Individual participants contribute more than one value (yearly measurements). Panel A shows the relationship between traditional HbF measurements (e.g., Hb electrophoresis or HPLC) and total F-cell number (flow cytometry). A traditional HbF measurement of at least 30% (first dashed vertical line) is a reasonable proxy for a marked increase in F-cells to at least 80% of all RBCs (horizontal dashed line). HbF ≥35% (second vertical dashed line) provides high certainty for at least near-pancellular expression (F-cells ≥80%). Panel B shows the relationship between total F-cell number and mean HbF content per F-cell (F/F-cell). An 80% F-cell response (first vertical dashed line) is associated with a mean F/F-cell ≥10 pg (horizontal dashed line) in 86.7% of paired measurements. When F-cells were ≥93% (second vertical dashed line), then mean F/F-cell was ≥10 pg in all paired samples. Panel C shows the relationship between Hb F/F-cell and absolute reticulocyte count (ARC). Normalization of the ARC (<150 × 109/L; horizontal dashed line) was achieved in most participants (85%) when the mean F/F-cell was ≥14.5 pg (vertical dashed line). For all panels, the box and whiskers depict IQR and Tukey hinges, and outliers are shown as points. In the box, the horizontal line is the median, and the “+” indicates the mean.

If we consider a mean F/F-cell of 10 pg as a minimum goal for a treatment response, then an 80% F-cell response is associated with a mean F/F-cell ≥10 pg in 89% of paired measurements (Figure 4, Panel B). Once F-cells were ≥93%, then all mean F/F-cell estimates were ≥10 pg. Here we show that at least 70% of hydroxyurea-treated young patients can achieve at least this high degree of F/F-cell content (Figure 2, Panel D). Normalization of the absolute reticulocyte count was achieved in most participants (91%) when the mean F/F-cell was ≥14.5 pg (Figure 4, Panel C).

Genetic determinants of F-cell responses

Baseline (pre-hydroxyurea) HbF levels were associated with the QTL polymorphisms (Supplementary Figure 4). On hydroxyurea therapy, HbF level, total F-cell number, and F/F-cell content increased across all QTL genotypes. There was an indication of higher HbF level, total F-cell number, and F/F-cell content in participants who had high HbF alleles at these QTLs (Supplementary Figure 4). Individuals with higher α-globin gene copy number tended to have higher HbF (baseline and treatment) as well as total F-cell and F/F-cell responses (Supplementary Figure 5). Nevertheless, hydroxyurea treatment produced good treatment responses in all genotype groups, indicating that all patients can benefit regardless of baseline genetic determinants.

Clinical complications

There were 25 hospitalisations (in 13 participants) and 12 emergency department (ED) visits (in 8 participants) for vaso-occlusive complications. The reasons for hospitalisation were acute chest syndrome (N=14), acute painful event (N=10), and acute splenic sequestration (N=1). All ED visits were for acute painful events. The mean rate of ED visits was 0.07 events/year (median 0), and the mean rate of hospitalizations was 0.15 events/year (median 0), giving a mean total event rate (ED visits + hospitalisations) of 0.23/year (median 0). Nearly 65% (31/48) had a total event rate of 0. The total event rate was significantly inversely correlated with HbF (r = −0.433, P = 0.002), total F-cells (r = −0.391, P = 0.006), and F/F-cell (r = −0.294, P=0.042). The mean total event rate was significantly lower for participants who had total F-cells ≥80% compared to <80% (0.33 vs 1.73, P=0.004). The mean total event rate was numerically lower for mean F/F-cell ≥10 pg compared to <10 pg (0.68 vs 1.0, P=0.52) but not statistically significant. Considering a higher cutoff, the mean total event rate was significantly lower for mean F/F-cell ≥12 pg compared to <12 pg (0.23 vs 1.23, P=0.027).

Adverse Events

There were no serious adverse events related to hydroxyurea in the clinical trial (McGann et al, 2019). Here, all hospitalisations and ED visits were caused by the underlying SCA. The only treatment-emergent adverse events were expected and included transient cytopenia (platelets < 50×109/L: 9 events in 6 participants; ANC < 1.0 ×109/L : 25 events in 19 participants) and temporary interruption of hydroxyurea therapy (16 interruptions in 13 participants). Cytopenia was associated with viral upper respiratory infections, otitis media, or acute chest syndrome in 20% of events (7/34). No participants permanently discontinued therapy during the study period.

Discussion

Here we show that PK-guided dosing of hydroxyurea with initiation at predicted MTD can achieve sustained haematologic improvements with near-pancellular and even pancellular HbF expression in children with SCA. This dosing strategy was well-tolerated with few clinical or laboratory adverse effects (McGann et al, 2019). We observed mean values of %HbF >30% in the TREAT cohort for up to 6 years. Despite initiation of hydroxyurea at a very young age, when Hb concentration and HbF values are still relatively high, these values increased further and did not show the otherwise expected decline in HbF due to developmental Hb switching. Because of the inherent limitations of traditional %HbF measurements, which do not reflect critical differences in the cell-specific compartmentalization of HbF across RBCs, flow cytometric F-cell analysis was used for individualised monitoring of hydroxyurea therapy. We demonstrated robust increases in total F-cells, F-reticulocytes, and F/F-cell content to levels greater than reported with traditional hydroxyurea dosing. Several participants had very high mean F/F-cell content consistent with nearly complete protection of RBCs against sickling. Moreover, the high degree of HbF induction in all participants was not conditional upon genetic determinants of HbF expression. Very early initiation of hydroxyurea, before the developmental transcriptional silencing of the γ-globin genes is complete, may underlie the potential for sustained pancellular HbF expression. Whatever the mechanism, marked and sustained HbF induction should be considered an achievable goal for children with SCA when hydroxyurea is initiated at PK-guided, individualised MTD at an early age.

Some have proposed that curative therapies should aim to achieve total F-cells >70% and minimum F/F-cell values ranging from >4 to >10 pg (Hebert et al, 2020; Buchanan, 2014; Steinberg et al, 2014). We suggest that an F-cell fraction of 80% and a mean F/F-cell of 10 pg should be the minimum goals for hydroxyurea therapy, especially when treatment is initiated in early childhood. An 80% F-cell fraction is associated with a mean F/F-cell ≥10 pg in nearly 90% of TREAT participants. We did not quantify F/F-cell in individual RBCs, as has been recently published (Hebert et al, 2020). Rather, for any given mean F/F-cell value calculated here, there is an unmeasured distribution of F/F-cell values from individual F-cells around that mean. So, a mean F/F-cell value of 10 pg does not indicate that all F-cells have 10 pg HbF and are fully protected against HbS polymerization. Our data suggest that a higher mean F/F-cell is needed, by shifting the entire F/F-cell distribution higher, to provide sufficient F/F-cell content to minimise clinical complications (F/F-cell ≥12 pg) and prevent haemolysis (F/F-cell ≥14.5 pg).

After initiation of therapy, optimal hydroxyurea dosing requires ongoing, rules-based adjustments for weight gain and maintenance of the ANC in the low-normal range (1 – 3 ×109/L). Especially for growing children, increases in hydroxyurea dose may be needed several times per year. Dose increases, especially when patients gain weight, should be considered even when the ANC is in a normal range (2.0 – 3.0 ×109/L), because additional HbF induction is often possible without significant myelosuppression. In TREAT, hydroxyurea was very well tolerated with no episodes of neutropenia-associated infection (McGann et al, 2019). Recent data from the NOHARM trial also demonstrate the safety and clinical efficacy of dose escalation and optimization (John et al, 2020). In addition to careful attention to optimal hydroxyurea dose selection, ongoing interventions to improve and support medication adherence are key.

This study has limitations. First, our results come from a comprehensive sickle cell treatment centre in a high-resource medical environment. We have a dedicated multidisciplinary team that includes 7 physicians with a clinical and research focus on SCA, nurse practitioners, nurses, social workers, and psychologists. Importantly, the psychologists are integrated in our team to provide individualised and ongoing support for adherence to hydroxyurea and other medications. These features may limit generalisability, but our results should be achievable in any high-resource environment. Second, the laboratory methods we used, including sparse PK sampling (and population modeling) and F-cell analysis may not be universally available, but these assays can be performed in central laboratories to support any hydroxyurea-treated patient. If F-cell analysis is not available, a HbF measurement of ≥30% is a reasonable proxy for F-cells ≥80%, and HbF ≥35% provides high certainty for at least near-pancellular expression. Third, we report a treatment duration of up to 6 years. HbF responses may wane over a longer period, so monitoring of TREAT participants is ongoing. Finally, the small sample of ~50 patients may not be representative of most with SCA. A separate, multicentre clinical trial, the Hydroxyurea Optimization Through Precision Study is currently underway with 116 planned participants in the US to validate and extend the findings here (Meier et al, 2020).

In summary, individualised, PK-guided dosing of hydroxyurea with initiation at predicted MTD can achieve sustained near-pancellular and even pancellular HbF expression in children with SCA. Such marked HbF induction is not conditional upon genetic determinants of HbF expression and should be considered an achievable goal for children with SCA treated with optimised hydroxyurea dosing at MTD. Initiation of hydroxyurea in early childhood, especially before 3 years of age, increases the likelihood of sustained, pancellular HbF expression. We suggest that a total F-cell fraction of 80% and a mean F/F-cell of 10 pg should be the minimum goals for any HbF-inducing therapy, whether pharmacological or genetic.

Supplementary Material

fS1-S5,tS1

Acknowledgements

This research was supported by a National Heart, Lung, and Blood Institute Patient-Oriented Career Development Award K23HL128885 (P.T.M.). We thank Lynette Fenchel and Kelly Clapp for assistance with patient care.

Footnotes

Conflicts of Interest

The authors declare no competing interests.

Clinical trial registration: NCT02286154 (clinicaltrials.gov)

References

  1. Buchanan GR (2014) “Packaging” of fetal hemoglobin in sickle cell anemia. Blood, 123, 464–465. [DOI] [PubMed] [Google Scholar]
  2. Cardoso GL, Diniz IG, Silva A.N.L.M. da, Cunha DA, Junior J.S. da S., Uchôa CTC, Santos S.E.B. dos, Trindade SMS, Cardoso M. do S. de O. & Guerreiro JF (2014) DNA polymorphisms at BCL11A, HBS1L-MYB and Xmn1-HBG2 site loci associated with fetal hemoglobin levels in sickle cell anemia patients from Northern Brazil. Blood Cells, Molecules, and Diseases, 53, 176–179. [DOI] [PubMed] [Google Scholar]
  3. Chaouch L, Moumni I, Ouragini H, Darragi I, Kalai M, Chaouachi D, Boudrigua I, Hafsia R & Abbes S (2016) rs11886868 and rs4671393 of BCL11A associated with HbF level variation and modulate clinical events among sickle cell anemia patients. Hematology, 21, 425–429. [DOI] [PubMed] [Google Scholar]
  4. Dai Y, Sangerman J, Nouraie M, Faller AD, Oneal P, Rock A, Owoyemi O, Niu X, Nekhai S, Maharaj D, Cui S, Taylor R, Steinberg M & Perrine S (2017) Effects of hydroxyurea on F‐cells in sickle cell disease and potential impact of a second fetal globin inducer. American Journal of Hematology, 92, E10–E11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Davis BH & Davis KT (2004) Enumeration of Fetal Red Blood Cells, F Cells, and F Reticulocytes in Human Blood. Current Protocols in Cytometry, 28, 6.17.1–6.17.16. [DOI] [PubMed] [Google Scholar]
  6. Dong M & McGann PT (2021) Changing the Clinical Paradigm of Hydroxyurea Treatment for Sickle Cell Anemia Through Precision Medicine. Clinical Pharmacology & Therapeutics, 109, 73–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Dong M, McGann PT, Mizuno T, Ware RE & Vinks AA (2016) Development of a pharmacokinetic‐guided dose individualization strategy for hydroxyurea treatment in children with sickle cell anaemia. British Journal of Clinical Pharmacology, 81, 742–752. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Dover GJ, Boyer SH, Charache S & Heintzelman K (1978) Individual Variation in the Production and Survival of F Cells in Sickle-Cell Disease. The New England Journal of Medicine, 299, 1428–1435. [DOI] [PubMed] [Google Scholar]
  9. Eaton WA & Bunn HF (2017) Treating sickle cell disease by targeting HbS polymerization. Blood, 129, 2719–2726. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Hebert N, Rakotoson MG, Bodivit G, Audureau E, Bencheikh L, Kiger L, Oubaya N, Pakdaman S, Sakka M, Liberto GD, Chadebech P, Vingert B, Pirenne F, Galactéros F, Cambot M & Bartolucci P (2020) Individual red blood cell fetal hemoglobin quantification allows to determine protective thresholds in sickle cell disease. American Journal of Hematology, 95, 1235–1245. [DOI] [PubMed] [Google Scholar]
  11. John CC, Opoka RO, Latham TS, Hume HA, Nabaggala C, Kasirye P, Ndugwa CM, Lane A & Ware RE (2020) Hydroxyurea Dose Escalation for Sickle Cell Anemia in Sub-Saharan Africa. New England Journal of Medicine, 382, 2524–2533. [DOI] [PubMed] [Google Scholar]
  12. Lettre G, Sankaran VG, Bezerra MAC, Araújo AS, Uda M, Sanna S, Cao A, Schlessinger D, Costa FF, Hirschhorn JN & Orkin SH (2008) DNA polymorphisms at the BCL11A, HBS1L-MYB, and beta-globin loci associate with fetal hemoglobin levels and pain crises in sickle cell disease. Proceedings of the National Academy of Sciences of the United States of America, 105, 11869–11874. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Maier-Redelsperger M, Montalembert M. de, Flahault A, Neonato MG, Ducrocq R, Masson MP, Girot R & Elion J (1998) Fetal hemoglobin and F-cell responses to long-term hydroxyurea treatment in young sickle cell patients. The French Study Group on Sickle Cell Disease. Blood, 91, 4472–9. [PubMed] [Google Scholar]
  14. Marcus SJ, Kinney TR, Schultz WH, O’Branski EE & Ware RE (1997) Quantitative analysis of erythrocytes containing fetal hemoglobin (F cells) in children with sickle cell disease. American Journal of Hematology, 54, 40–46. [DOI] [PubMed] [Google Scholar]
  15. McGann PT, Niss O, Dong M, Marahatta A, Howard TA, Mizuno T, Lane A, Kalfa TA, Malik P, Quinn CT, Ware RE & Vinks AA (2019) Robust clinical and laboratory response to hydroxyurea using pharmacokinetically guided dosing for young children with sickle cell anemia. American Journal of Hematology, 94, 871–879. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Meier ER, Creary SE, Heeney MM, Dong M, Appiah-Kubi AO, Nelson SC, Niss O, Piccone C, Quarmyne M-O, Quinn CT, Saving KL, Scott JP, Talati R, Latham TS, Pfeiffer A, Shook LM, Vinks AA, Lane A & McGann PT (2020) Hydroxyurea Optimization through Precision Study (HOPS): study protocol for a randomized, multicenter trial in children with sickle cell anemia. Trials, 21, 983. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Ngo DA, Aygun B, Akinsheye I, Hankins JS, Bhan I, Luo HY, Steinberg MH & Chui DHK (2012) Fetal haemoglobin levels and haematological characteristics of compound heterozygotes for haemoglobin S and deletional hereditary persistence of fetal haemoglobin. British Journal of Haematology, 156, 259–264. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Saunthararajah Y (2019) Targeting sickle cell disease root-cause pathophysiology with small molecules. Haematologica, 104, 1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Steinberg MH, Chui DHK, Dover GJ, Sebastiani P & Alsultan A (2014) Fetal hemoglobin in sickle cell anemia: a glass half full? Blood, 123, 481–485. [DOI] [PubMed] [Google Scholar]
  20. Steinberg MH, Lu Z-H, Barton FB, Terrin ML, Charache S & Dover GJ (1997) Fetal Hemoglobin in Sickle Cell Anemia: Determinants of Response to Hydroxyurea. Blood, 89, 1078–1088. [PubMed] [Google Scholar]
  21. Steinberg MH, McCarthy WF, Castro O, Ballas SK, Armstrong FD, Smith W, Ataga K, Swerdlow P, Kutlar A, DeCastro L, Waclawiw MA & Follow-Up, I. of the M.S. of H. in S.C.A. and M.P. (2010) The risks and benefits of long‐term use of hydroxyurea in sickle cell anemia: A 17.5 year follow‐up. American Journal of Hematology, 85, 403–408. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Telen MJ, Malik P & Vercellotti GM (2019) Therapeutic strategies for sickle cell disease: towards a multi-agent approach. Nature Reviews Drug Discovery, 18, 139–158. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Urio F, Lyimo M, Mtatiro SN, Cox SE, Mmbando B. p. & Makani J (2016) High prevalence of individuals with low concentration of fetal hemoglobin in F‐cells in sickle cell anemia in Tanzania. American Journal of Hematology, 91, E323–E324. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Wood WG, Stamatoyannopoulos G, Lim G & Nute PE (1975) F-cells in the adult: normal values and levels in individuals with hereditary and acquired elevations of Hb F. Blood, 46, 671–682. [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

fS1-S5,tS1
NIHMS1713684-supplement-fS1-S5_tS1.docx (1.5MB, docx)

RESOURCES