Haptoglobin genotype predicts severe acute vaso-occlusive pain episodes in children with sickle cell anemia

Chronic hemolysis is a prevalent characteristic of SCA that intensifies during times of illness, such as infection, acute chest syndrome (ACS), and vaso-occlusive pain episodes.¹ Hemolysis causes the release of cell-free hemoglobin and free plasma heme² which can damage lipid, protein, and DNA through the generation of reactive oxygen species and inflammatory signaling pathways.³ The plasma protein haptoglobin (HP) binds to and clears cell-free hemoglobin from circulation thereby preventing its deleterious effects.^4,5

A common variant in the HP gene may offer a genetic risk factor for disease-related complications in SCA. HP has two predominant alleles in humans (1 and 2) leading to three possible genotypes: HP1–1, HP1–2, and HP2–2. The HP2–2 protein has shown a decreased efficiency and lower affinity for binding cell-free hemoglobin; thus, the ability to scavenge cell-free hemoglobin is genotype dependent.⁶ The HP2–2 protein also has a lower circulating concentration and poorer protection against oxidative damage from cell-free hemoglobin in vitro.^7–9 The role of HP genotype and its relation to disease severity has been studied in multiple disease states such as epilepsy,¹⁰ cardiovascular disease,^11,12 hypercholesteremia,¹³ diabetes,¹⁴ hemochromatosis,¹⁵ and sub-arachnoid hemorrhage,^16,17 with variable results. However, there is little data investigating the role of these genotypes in SCA. We hypothesized that children with SCA and the HP2–2 genotype will have an increased incidence of vaso-occlusive pain and ACS episodes, compared to those with the HP1–1 and HP1–2 genotypes. To test this hypothesis, we examined the impact of HP genotype on the development of pain and ACS episodes in a primary cohort of children with SCA. We then sought to validate our findings in a replication cohort of age-matched children with SCA.

The primary cohort included participants enrolled in the Sleep and Asthma Cohort (SAC), a prospective cohort study of 246 children and adolescents with SCA (HbSS, HbSβ°). Children aged 4 to 18 years were enrolled and followed between 2005 and 2011 as previously described.¹⁸ A total of 199 participants from the SAC study with DNA available for genotyping were included in the primary cohort. These participants had a mean age of 10.8 years (range 4.0–19.3 years), had DNA available, laboratory data collected at baseline, and were followed for at least one year during the study period (mean 5.0, range 1.1–6.7 years).

The replication cohort included participants enrolled in the Cooperative Study of Sickle Cell Disease (CSSCD), a multi-institutional prospective cohort study of 4,085 individuals with SCA (HbSS, HbSβ°) from 23 U.S. academic hospitals that followed participants from 1977 to 1995 in a standardized fashion.^19,20 A total of 458 participants from the CSSCD were included as they had GWAS data available, laboratory data collected at baseline, were between 4–19 years of age (mean age 11.1 years, range 4.0–19.3yrs), and were followed for at least one year during the study period (mean 6.3, range 2–8.2 years), in order to harmonize the data with the primary (SAC) cohort. Uniform definitions for acute pain and ACS were utilized in both the SAC and CSSCD cohorts.²¹ An acute pain episode was defined as a hospitalization for SCA-associated pain, excluding headaches, and requiring opioid treatment. ACS was defined as a new clinical or radiographic pulmonary infiltrate in the context of an acute illness characterized by respiratory symptoms with or without fever. Respiratory symptoms included cough, wheezing, rales, chest pain, decreased oxygen saturation (decrease >2% from baseline), use of accessory muscles of respiration, or increased respiratory rate. Pneumonia was included in this ACS definition. Both pain and ACS rates were calculated as events per year to account for variable follow-up duration. Table 1 shows descriptive statistics of both cohorts.

Table 1.

Descriptive statistics of the study populations from the primary cohort (SAC) and replication cohort (CSSCD)

	SAC N= 199	CSSCD N= 458	P Value*
Age, years, mean (SD)	10.8 (4.2)	11.1 (4.4)	0.419
Male, N (%)	105 (52.8)	230 (50.2)	0.549
Hemoglobin, g/dL, mean (SD)	8.1 (1.1)	8.2 (1.1)	0.292
White blood cell count, 109/, mean (SD)	12.1 (3.9)	12.2 (3.7)	0.788
Prospective time, years, median (IQR)	5.1 (4.2 – 5.9)	6.3 (5.7 – 6.5)	<0.001#
Rate of pain episodes per year, mean (SD)	0.9 (1.3)	0.8 (2.3)	<0.001‡
Rate of ACS episodes per year, mean (SD)	0.2 (0.3)	0.1 (0.3)	<0.001‡
Haptoglobin 1–1 genotype, N (%)	60 (30.2)	108 (23.6)
Haptoglobin 1–2 genotype, N (%)	91 (45.7)	237 (51.7)
Haptoglobin 2–2 genotype, N (%)	48 (24.1)	113 (24.7)

^*Overall comparison between CSSCD and SAC across all haptoglobin genotypes. T test for mean difference and Chi-square test for frequencies, unless otherwise noted.

^#Mann-Whitney U test

^‡Negative binomial regression

HP genotype was assessed by two methods: direct PCR and imputation, with a subset of subjects undergoing both methods in order to determine agreement between methods. For the SAC cohort, analysis of HP genotype was done by direct HP genotyping using an established real-time TaqMan PCR method.²² Next, in a subset of 187 SAC participants who had both DNA available for direct PCR genotyping and GWAS level genotyping, we validated a previously published imputation method¹³ in this US population of African Ancestry. Using GWAS data, HP genotype was determined based on imputation of HP structural features from surrounding SNPs.¹³ We observed excellent agreement between genotyping methods (Cohen’s Kappa 0.84, balanced accuracy 90%, 95% CI [85%; 93%]). After this validation, imputation was used to determine HP genotype in the replication CSSCS cohort (Supplemental figure 1 and Supplemental table 1).

The distribution of haptoglobin genotypes in both cohorts was similar to distribution previously reported among patients with SCA²³ and individuals of African ancestry.⁸ In the SAC cohort, 60 participants were HP1–1 (30%), 91 were HP1–2 (46%), and 48 were HP2–2 (24%). In the CSSCD cohort, 108 were HP1–1 (24%), 237 were HP1–2 (52%), and 113 were HP2–2 (25%) (Supplemental Figure 2). In multivariable negative binomial regression models, HP2–2 genotype was associated with an increased rate of severe vaso-occlusive pain when compared to the HP1–1 and HP1–2 genotypes in both the SAC cohort (incidence rate ratio [IRR] 1.64, 95% confidence interval [CI] 1.04–2.58, p= 0.03) and the CSSCD cohort (IRR 1.57, 95% CI 1.03–2.39, p= 0.04) (Figure 1). Older age and increased WBC count were also associated with a modest increase in pain rate in the SAC cohort, as has been demonstrated previously.^24,25 In the CSSCD cohort, higher WBC was associated with a significantly decreased rate of pain, but with a minimal effect size (IRR 0.95, 95% CI 0.90–0.99, p=0.04). In contrast, HP2–2 was not independently associated with an increased rate of ACS episodes in either cohort. Increased WBC was associated with an increased rate of ACS episodes in the SAC cohort (IRR 1.07, 1.02–1.13, p=0.01) while no covariates were associated with an increased rate of ACS episodes in the CSSCD cohort (Supplemental Table 2).

Figure 1.

Negative binomial regression model demonstrating increased risk of severe acute vaso-occlusive pain episodes in individuals with sickle cell anemia and the HP2–2 genotype compared to the HP1–1 and HP1–2 genotypes in both the primary and replication cohorts. Incidence rate ratio (IRR) with bars representing confidence intervals for each clinical covariate.

SCA is a chronic hemolytic disease with a range of severity of hemolysis both over time in an individual and between individuals. Kato et al have put forth a working hypothesis that the products of intravascular hemolysis damage the vascular system and contribute to the pathophysiology of SCA.²⁶ We have expanded on this concept to show that genetic variation in the ability to scavenge cell-free hemoglobin is also implicated in SCA-related morbidity. In children with SCA, the HP2–2 genotype is associated with a significantly increased rate of severe acute vaso-occlusive pain episodes in both our primary and replication cohorts, with an increase of 64% and 57% respectively compared to HP1–1 and HP1–2 genotypes. This association was shown in the primary cohort by direct molecular analysis of HP genotyping and confirmed in the replication cohort by imputation of the HP2–2 genotype from GWAS data. In addition, we validated the use of HP imputation for assessment of the impact of HP variants on clinical outcomes in a US population of African ancestry as demonstrated by the high correlation between PCR-based and imputed HP genotypes.

The effect of haptoglobin genotype on morbidity and mortality has been studied in other acute and chronic diseases, with variable results. Individuals with the HP2–2 genotype have been found to have increased risk for diabetic cardiovascular and renal disease,^11,27 cognitive decline among those with type 2 diabetes,²⁸ delayed brain injury after subarachnoid hemorrhage,¹⁶ and development of ARDS in sepsis,²⁹ while other studies have shown no apparent association between risk for iron accumulation in hereditary hemochromatosis,¹⁵ cardiovascular risks in individuals with elevated HbA1C,³⁰ or outcomes after subarachnoid hemorrhage.³¹ To our knowledge, this is the first study to examine how variations in the haptoglobin genotype may contribute to morbidity in children with SCA. Previous studies have shown that the HP2–2 protein has a decreased efficiency and lower binding affinity for cell-free hemoglobin.⁶ These data, taken in conjunction with the finding that heme infusion leads to endothelial damage and vaso-occlusion in transgenic SCA mice,³² highlight the importance of heme-mediated injury in the pathophysiology of SCA. Additionally, studies conducted in SCA mice demonstrate infusion of human haptoglobin prevents hemoglobin mediated vaso-occlusion and protects against heme-induced death.³² Haptoglobin infusions have been proposed as a therapeutic intervention for adults with SCA, however this intervention has yet to be explored in clinical trials.³³

The current study finds an increased risk for episodes of acute severe vaso-occlusive pain among children with SCA and the HP2–2 genotype, supporting an important role for hemoglobin scavengers in modulating heme-mediated injury in SCA. Although further studies will be needed to provide insight into biological mechanisms of these processes, therapeutic strategies to decrease cell-free hemoglobin may complement other treatment strategies such as hydroxyurea, or anti-selectin antibodies. These complementary therapeutic strategies may attenuate the incidence rates of severe acute pain episodes in children with SCA.

In summary, in two large cohorts of children with SCA, HP2–2 genotype was associated with an increased risk for the development of acute severe vaso-occlusive pain episodes. HP genotype may prove to be an important predictor of morbidity in children with SCA. Further research is needed to fully understand the role of the haptoglobin genotype and the oxidative effect of cell-free hemoglobin and free plasma heme in the pathophysiology of SCA.