Evaluation of alpha hemoglobin stabilizing protein (AHSP) as a genetic modifier in patients with β thalassemia

Although β thalassemia results from reduced synthesis of the β globin chains of adult hemoglobin (α₂β₂), the pathophysiology is predominantly determined by excess free α globin chains which precipitate and cause oxidative damage in developing red cells (causing dyserythropoiesis) and in mature red cells (causing hemolysis).^1,2  Up to a certain threshold, the erythron can cope with and compensate for excess α globin chains by known and unknown mechanisms. For example, individuals with the β thalassemia trait have somewhat abnormal red cells but no significant anemia.¹  By contrast, most patients who are compound heterozygotes or homozygotes for β thalassemia have severe forms of anemia (thalassemia intermedia or major).¹  However, within this group of patients who have a significant degree of anemia there is still considerable variability in hematologic and clinical severity and this is thought to be determined primarily by variations in the amounts of free α chains.^1,3  In general, patients who inherit mild forms of thalassemia (β⁺⁺ and β⁺) tend to be less severely affected than those with β⁰ thalassemia.^4,5  Similarly, patients who inherit 2 α genes⁶  (-α/-α and --/αα) or 3 α genes⁷  (-α/αα) are often less severely affected than those with 4 (αα/αα), 5 (αα/ααα), or more α genes.⁸  Finally, those patients who coinherit any increased propensity to produce fetal (γ) globin chains in adult life⁹  (eg, the Xmn I+ allele) also tend to be less severely affected since γ chains can combine with excess α globin to produce fetal hemoglobin (Hb F; α₂γ₂), adding to the pool of functional hemoglobin and reducing unpaired α globin chains. However, even taking all of these factors into consideration it still appears that there are additional, as-yet-unidentified genetic determinants of the clinical outcome of patients with β thalassemia (reviewed in Weatherall³ ).

Recently, it has been shown that a highly expressed protein called alpha hemoglobin stabilizing protein (AHSP) can act as a chaperone for free α chains and prevent their precipitation.^10,11  Mice lacking AHSP have abnormal red cell production and lifespan, thought to be caused by a relatively small excess of unchaperoned α globin chains.¹⁰  Furthermore, the phenotype of mice with β thalassemia intermedia is exacerbated by concomitant loss of AHSP.¹²  Based on these observations it has been proposed that alleles altering the levels or function of AHSP might account for some of the clinical variability observed in patients with β thalassemia.^10,13  Here, we have fully characterized the AHSP gene in 120 Thai patients with clinically mild, moderate, and severe forms of Hb E-β thalassemia (a subtype of β thalassemia) to determine whether there are common alleles of AHSP in this population that modify the clinical presentation of β thalassemia. The influence of distinct alleles of AHSP were directly compared with the previously determined genetic determinants of severity, including the interacting β thalassemia alleles, the presence of α thalassemia, and the coinheritance of the high γ expression (Xmn I+) allele.

We recruited and obtained informed consent from 120 patients with Hb E-β thalassemia who have mild, moderate, or severe clinical phenotypes based on criteria previously defined by the Thalassemia International Federation¹⁴  (TIF). These criteria are onset of anemia, baseline levels of hemoglobin, symptoms of anemia, degree of splenomegaly and/or requirement to undergo splenectomy, transfusion requirement and/or treatment regimen (occasional transfusion, hypertransfusion with iron chelation or stem cell transplantation), and linear growth and height development during the follow-up period. In this study the average follow-up was 7.6 ± 0.79 years. This study was approved by a local ethical committee at Siriraj Hospital, Thailand.

The molecular basis of β thalassemia was determined using reverse dot-blot (RDB) hybridization,¹⁵  which simultaneously identifies 17 previously characterized β thalassemia mutations common in Thailand. DNA from patients with negative RDB results was subsequently analyzed by direct genomic sequencing of the β globin genes (2.4 kb). In addition, gap-polymerase chain reaction (PCR) analyses were performed to identify 3 previously characterized β globin deletions found in Thailand (3.5 kb, 619 bp, and 105 bp).^1,16,17  Coinheritance of either deletional or nondeletional forms of α thalassemia and the Xmn I polymorphism were analyzed by standard protocols^1,18  in all individuals. Two sets of PCR primers were used to amplify a 1.4-kb fragment of AHSP gene from genomic DNA. They are as follows: AHSP-F, 5′-tgcacagagagattcacgcacc-3′ (12179-12200) and AHSP-R, 5′-gcactggtctttattgaggtgtcag-3′ (13048-13072), producing an 895-bp coding sequence at annealing temperature 58°C; and AHSP-PF, 5′-cacgcttacaggtggcttatctg-3′ (11698-11720) and AHSP-PR, 5′-gtcacgattttcccaggttgg-3′ (12271-12251), for 574 bp of promoter and upstream region encompassing a polythimidine tract ((T)_n) repeat at annealing temperature 55°C (coordinates according to the GenBank accession number AC106 730.2 are shown in parentheses). PCR products were sequenced using fluorescent-labeled dideoxyterminators (Applied Biosystems, Foster City, CA). The promoter region containing (T)_n repeat PCRs were subcloned using pGEM-T Vector plasmid (Promega, Madison, WI) and DH5-α-competent cells (Invitrogen, Paisley, UK) before sequencing. Mapping analysis of AHSP loci on chromosome 16p was performed in genomic DNA from 30 patients (10 with mild and 20 with moderate and severe phenotypes) who have identical globin genotypes (β^41/42/β^E, αα/αα, Xmn I +/-). In each case, DNA was digested with HindIII and BglII and hybridized with AHSP genomic probe (895-bp PCR product).

We successfully identified the β globin mutations in all patients studied; 115 had β⁰ or severe β⁺ alleles and 5 had mild β⁺ alleles (Table 1). Most individuals with Hb E-mild β⁺ thalassemia had a mild phenotype (n = 4) therefore they were excluded from further study. A summary of the β globin mutations, hematologic parameters, and clinical features in the 3 different clinical categories are presented in Table 1 and Table 2.

As expected, despite being compound heterozygotes for β⁰ (or severe β⁺) and β^E mutations, approximately 50% of mildly affected patients coinherited either α thalassemia (-α or α^Tα) or were homozygous for the Xmn I polymorphism (+/+) (Table 3). None of these determinants were found in the severe group and only one patient (β^CD17/β^E, Xmn I +/-) who fell into the moderate group coinherited an α thalassemia allele (α^CSα) (Table 3).

To determine whether AHSP might play a role in modifying the clinical severity of patients with β thalassemia, we sequenced the AHSP and its promoter region in 50 individuals with identical β globin genotypes (β^41/42/β^E, β^CD17/β^E, and β⁶⁵⁴/β^E), normal α globin genotypes (αα/αα), and Xmn I polymorphism (+/- or -/-) including representatives from each of the different severity groups (17 mild, 13 moderate, and 20 severe). No nucleotide mutation that might obviously alter structure or function of AHSP was identified. However, we did find several single nucleotide polymorphisms (SNPs; Figure 1). These polymorphic sites have been noted previously when we sequenced a number of control samples (n = 20) from different ethnic origins to establish the genomic sequence of the AHSP gene (data not shown). These SNPs form distinct haplotypes with different numbers of (T)_n repeats at the promoter (haplotypes I and II; Figure 1). By using associations at coordinate 12 390 (I with G and II with A), we could identify these 2 haplotypes using a mismatch PCR restriction fragment length polymorphism (RFLP) for a Mae III site (supplemental information is available through the Blood website; see the Supplemental Document link at the top of the online article). We analyzed the haplotype in all patients (mild, moderate, and severe) but found no statistically significant association between a particular haplotype and disease severity (χ² test).

Interestingly, the allele frequency of haplotype II is higher in the patient groups compared with a group of nonthalassemic controls obtained in Bangkok, which represents the general Thai population. This might reflect different ethnic origins of these 2 groups. It has been proposed that Hb E has a single origin in Southeast Asia in the area adjoining Thailand, Cambodia, and Laos (the so-called “Hb E Triangle”).¹⁹  Hence, we analyzed 30 healthy nonthalassemic individuals (normal Hb A-A₂) from Surin, a province located in the Hb E-Triangle where 50% of population are carriers of Hb E, to determine whether the AHSP haplotype II might occur at a high frequency in this region. This population has AHSP allele frequencies (0.43) and haplotype combinations that are more similar to our patient groups (Table 3), suggesting that both Hb E and AHSP haplotype II may occur independently at high frequencies in this subpopulation. This finding highlights the importance of obtaining properly selected control groups for genetic-association studies in Southeast Asians, since the population admixture might give rise to a false-positive association if the origin of the subgroup analyzed is not clearly identified.

Finally, we excluded the possibility of any large deletions, duplications, or rearrangements, which may result in any differences in AHSP expression in these cases by comprehensive gene mapping analysis (data not shown).

The amounts of free α globin chains appear to be a major factor contributing to differences in hematologic and clinical severity in patients with β thalassemia. This study confirmed that coinheritance of α thalassemia alleles and increased γ globin expression (associated with the Xmn I+ polymorphism) ameliorate clinical severity.²⁰  AHSP is a good candidate for modifying β thalassemia since it selectively binds and stabilizes free α chains¹¹  and is highly expressed in hemoglobin-synthesizing erythroid precursors.¹⁰  The ability of reduced AHSP gene dosage to exacerbate the β thalassemia phenotype was demonstrated recently in mice.¹²  Moreover, Galanello et al²¹  reported that severity of β thalassemia is concordant with reduced AHSP mRNA expression in 2 unrelated Sardinian families. The current study demonstrates that the AHSP gene does not modulate clinical severity in Thai patients with Hb E-β thalassemia. However, since we were unable to examine AHSP protein or mRNA levels, it remains possible that reduced AHSP expression by genetic or epigenetic factors unlinked to its gene impact the severity of β thalassemia in the patients described here. In order to fully define the role of AHSP as a modifier of β thalassemia, it will be important to further examine its gene inheritance and expression in various patient populations.