Cholelithiasis is the most common gastrointestinal disease to require hospitalization worldwide [1]. Most gallstones are primarily comprised of cholesterol; however, in certain patient risk groups, bile pigment stones are extremely common and represent a significant cause of morbidity. These stones are associated with increased serum unconjugated bilirubin levels, which occur in a number of conditions including hemolytic anemia. In other patient groups, ineffective erythropoiesis, ileal diseases, or gastrointestinal resections producing increased colonic spillage of bile salts, predispose to bile pigment stones [2].

Patients with sickle cell disease (SCD), particularly those with sickle cell anemia (homozygosity for HBB glu6val), are at risk for bile pigment cholelithiasis due to the association of this disease with hemolysis which produces an unconjugated hyperbilirubinemia [3]. Cholelithiasis occurs in up to 70% of these patients, resulting from the precipitation of polymerized calcium bilirubinate within the gallbladder lumen when the ion product of calcium and unconjugated bilirubin exceeds its solubility. Up to 50% of these patients undergo cholecystectomy at some point during their lifetimes, suggesting increased associated morbidity, creating a need for hospitalization, and potential mortality from infectious or other complications related to this process(1).

Cholelithiasis risk, for both cholesterol and bile pigment stones, is thought to be partially mediated by genes impacting bilirubin metabolism [4], [5], [6]. Genome-wide association studies (GWAS) of three large Caucasian cohorts have identified single nucleotide polymorphisms (SNPs) in UGT1A1 that are responsible for 18% of clinical variability of serum bilirubin [7]. Sequencing of the UGT1A1 gene in patients with Gilbert’s syndrome, associated with unconjugated hyperbilirubinemia, revealed that an important modulator of enzyme activity is the length of TA repeats in the promoter region; the greater the number of repeats being associated with the highest serum bilirubin levels and cholelithiasis risk [8], [9].

Genome-wide studies examining hyperbilirubinemia and cholelithiasis risk have not been performed in persons of African descent with sickle cell anemia. However, sequencing of UGT1A1 has been undertaken in SCD cohorts in Guadeloupe, Portugal, Jamaica, India and the United States; genotyping of 324 SCD children from the Cooperative Study of Sickle Cell Disease (CSSCD) revealed that the common (TA)6 and (TA)7 alleles were present in 75% [10]–[15]. The (TA)7/(TA)7 genotype was associated with increased serum bilirubin levels and cholelithiasis risk compared with the (TA)6/(TA)7 and (TA)6/(TA)6 genotypes across cohorts with suggestions that those with a (TA)8 allele are at highest risk, suggesting a similar pattern of inheritance as observed in Caucasians [10]–[15]. Treatment with hydroxyurea, known to decrease the hemolytic rate in sickle cell anemia patients, was unable to decrease bilirubin levels to normal levels in those with the (TA)7/(TA)7 genotype suggesting that this genetic defect may represent a new pharmacologic target for these patients [14].

One of the reasons for the focus on UGT1A1 in genetic studies relates to its known importance in bilirubin metabolism [8], [9]. Upon release into the plasma, heme is metabolized to bilirubin via the actions of heme oxygenase-1. Bilirubin is poorly water soluble and in order for excretion to occur, disruption of its hydrogen bonds via a process termed glucuronidation, or conjugation, is essential. Glucuronidation of bilirubin is mediated by a family of enzymes termed the uridine-diphosphoglucuronate glucuronosyltransferases (UGT). Of these, UGT1A1 is the most important regulator of this process clinically and is one of the isoforms of the UGT1A gene complex comprised of 9 transcriptional units. Unconjugated or indirect bilirubin is responsible for all of the toxic effects of hyperbilirubinemia, supporting the clinical importance of this class of enzymes. Additionally, as alluded to previously, genetic studies of other patient populations characterized by hyperbilirubinemia, Gilbert’s syndrome and cystic fibrosis, have demonstrated an association between genetic variants within dinucleotide repeats in the promoter of UGT1A1 and bile pigment stone formation [7].

We conducted a GWAS of serum total bilirubin in the largest number of African American sickle cell anemia patients studied to date to determine if genetic variants played a role in the increased risk of cholelithiasis observed in this population. Our analysis shows that UGT1A1 is the major regulator of bilirubin levels in African Americans with sickle cell anemia, and that genetically mediated differences in bilirubin conjugation play an important role in cholelithiasis risk in these patients.

Intravascular hemolysis in sickle cell anemia patients produces increased circulating heme levels and has been associated with clinical complications including cholestatic jaundice and cholelithiasis [10], [23]–[25], [26]–[28] Unconjugated hyperbilirubinemia is a risk factor for the development of bile pigment stones in sickle cell disease and other hemolytic anemias [12], [29]–[41]. Typically, bilirubin is formed from metabolism of heme, which is derived principally from erythrocytic hemoglobin. Heme is oxidatively metabolized by heme oxygenase-1 into biliverdin, which is subsequently reduced enzymatically by biliverdin reductase to bilirubin. Upon formation, bilirubin is water insoluble, due to the presence of internal hydrogen bonding that creates a contorted molecule structure, inhibiting solubility. This form of bilirubin, termed unconjugated or indirect bilirubin, is responsible for all of the toxic effects of bilirubin observed clinically. To permit excretion, bilirubin complexes with albumin and is transported via the bloodstream to the liver, where it undergoes glucuronidation by the UGT family of enzymes; UGT1A1 is the only one with clinical relevance in humans [42]. UGT1A is a gene complex composed of 9 transcriptional units encoding an isoform of the UGT gene [30], [31]. The most common genetic cause of impaired bilirubin glucuronidation occurs in Gilbert’s syndrome, which follows an autosomal recessive inheritance pattern and most commonly affects Caucasians of European descent [43], [44]. Clinically, these patients have a mild indirect hyperbilirubinemia, which worsens during periods of stress or febrile illnesses [45]. Genetic studies of these patients have identified a dinucleotide repeat polymorphism (TA)_5–8 in the TATA box of the UGT1A1 gene promoter that is associated with reduced UGT expression and produces hyperbilirubinemia [30]–[31], [37], [46]–[47]. This polymorphism has also been observed in small cohorts of sickle cell anemia patients suggesting a common pathogenic link between ethnically divergent etiologies of indirect hyperbilirubinemia [5], [10]–[15], [33]. In the current study, genome-wide genetic screening of a large cohort of sickle cell anemia patients not only confirmed these findings but also supported their biological relevance.

In addition to defects in bilirubin metabolism, genetic alterations in the glucuronidation pathway have been linked to abnormalities in the hepatic metabolism of certain medications, pre-disposition to cardiovascular disease and certain types of malignancy suggesting a clinical level of importance beyond cholelithiasis [35], [42], [48]–[53]. How this relates to the pathogenesis of sickle cell anemia is not entirely clear clinically, but may be worthy of future study.

In genetic studies of primarily Caucasian cohorts, cholelithiasis and serum bilirubin levels are heritable traits. Family-based studies show that the UGT locus accounted for a significant proportion of the variation observed in both of these variables [18], [54]–[56]. Our study is the first to use GWAS to examine an African American population with sickle cell anemia, a condition where hemolysis predisposes patients to hyperbilirubinemia, to determine the heritability of serum bilirubin levels. Fifteen SNPs in UGT1A1, UGT1A3, UGT1A4, UGT1A5, UGT1A6, UGT1A7, UGT1A8 and UGT1A9 reached genome-wide significance for association with total bilirubin levels in the CSSCD cohort; 13 of these were also associated with the presence or history of cholelithiasis. These findings were confirmed in 4 other cohorts totaling 3,269 patients, unusual for a rare disease such as sickle cell anemia, and representing a greater number of patients than all of the previous studies of this population combined.

SNPs we identified in the current study have been linked to hyperbilirubinemia in other populations. SNP rs887829, the most significant SNP associated with bilirubin levels identified in the CSSCD, Duke and Walk-PHaSST cohorts, was previously observed in associated with total bilirubin levels in a GWAS of 4,300 Sardinians [57]. This study also reported a significant association with bilirubin and the SLCO1B3 locus; however, we were unable to replicate these results in our study due to unavailability or low MAF of the markers in our array. In a study performed by Cheng et al. examining genetic variants associated with serum bilirubin in 619 healthy African Americans revealed that the top SNP was rs887829 (p = 1.77×10⁻²²) [58]. This suggests that this variant is not unique to African Americans with sickle cell disease. It has been theorized that this SNP may confer protection against malaria which may explain its penetrance amongst African populations. SNP rs887829 is located in the promoter region of UGT1A1, 221 bp upstream from the (TA)_n repeat sequence. Horsfall et al. found a strong association was found between rs887829 and the (TA)₇ and (TA)₈ repeat sequence suggesting that this may be a marker for the (TA)7/(TA)7 or 8 genotypes [34]. This proximity to the promoter element and strong level of association suggest that rs887829 is a marker for the (TA)_n repeats and that additional sequencing of this gene would be redundant [58]. SNP rs6742078, identified in the current study, was associated with total bilirubin in a meta-analysis previously performed by Johnson et al., including the Framingham Heart Study, Rotterdam Study and Age, Gene, Environment and Susceptibility-Reykjavik study cohorts [33]. In addition, SNPs rs3755319, rs7586110, and rs6759892 were also found to be significantly associated with total bilirubin in a cohort of 4,300 Sardinians [57]. SNP rs887829 was found to be associated with cholelithiasis in a study published by Buch et al., in which a cohort of 2,606 German cholelithiasis patients was compared with 1,121 South American controls (OR = 1.73, p value .003) [11].

There was no association between LDH, hemoglobin concentration or reticulocyte count, markers of hemolysis, and the SNPs identified in the present study. A potential explanation for this is the weak association between bilirubin levels and hemolytic rate. While the "gold standard" for hemolysis is the red cell lifespan, this is rarely performed and surrogate blood biomarkers such as reticulocyte count, LDH, bilirubin and haptoglobin levels are used clinically to estimate the degree of hemolysis. None of these measures are specific for hemolysis. An elevated serum bilirubin level in sickle cell anemia reflects multiple pathophysiologic processes that include liver disease in addition to hemolysis and therefore may lack the phenotypic specificity required for genetic association studies of hemolysis. Another possible reason for this finding is that the genes associated with bilirubin levels are all involved in bilirubin catabolism and not production suggesting that they are not reflective of the processes leading to bilirubin formation.

In summary, SNPs in UGT1A1 are most tightly associated with bilirubin levels in African Americans with sickle cell anemia. This study in conjunction with those conducted in other populations at risk for unconjugated hyperbilirubinemia and other cohorts of sickle cell anemia patients support the concept that genetically mediated differences in bilirubin conjugation play an important role in the cholelithiasis risk. It is possible that targeting this pathway pharmacologically may offer new therapeutic options for these patients.