Spectrum of pathogenic alterations identified after two decades of HBB gene sequencing for molecular diagnosis of beta-thalassaemias and haemoglobinopathies

Inherited haemoglobin disorders, including reduced globin expression (β-thalassaemia) and structural haemoglobin variants (haemoglobinopathies) are common genetic blood disorders worldwide. Screening is performed using a full blood count, haemoglobin analysis of haemoglobin A2 (HbA2) and haemoglobin F (HbF) by capillary electrophoresis or high performance liquid chromatography, and gel electrophoresis. Microcytic anaemia with an elevated level of HbA2 or HbF suggests β-thalassaemia or a haemoglobinopathy, which are caused by pathogenic alternations in the beta globin (HBB) gene. Located on the short arm of chromosome 11 (11p15.5), the HBB gene consists of three exons encoding 147 amino acids. The majority (95%) of β-thalassaemias and haemoglobinopathies are due to small nucleotide substitutions, deletions or insertions, and only a minority of β-thalassaemias are due to large gene deletions. Currently there are at least 800 HBB pathogenic alterations reported in curated databases including HbVar, IthaGenes and HGMD. Different ethnic groups and populations from different geographical locations have distinct mutational spectra of HBB variants.^, Many molecular laboratories adopt targeted mutational analysis using reverse dot blot or primer specific amplification assays to detect commonly occurring variants in the populations from which their patients originate. While being cost effective and time efficient, targeted testing may miss rare and novel variants. Mutation scanning and sequence analysis have an analytical sensitivity of 99% for mutation detection. Capillary sequencing is the best current method for interrogating all possible small genetic variants in the HBB gene.

Located in Singapore, a city with a multi-ethnic South East Asian population with a substantial proportion of international immigrants, our laboratory has been using capillary sequencing to genotype the HBB gene (NM_000518.5) in patients with clinical suspicion of β-thalassaemia and haemoglobinopathies since 1998. Genomic DNA is extracted from whole blood and polymerase chain reaction (PCR) is performed using a set of primers flanking the nucleotides c.-177 and c.∗459 to generate a PCR product of 2060 nucleotides, which is then run on gel electrophoresis. Subsequently, the purified PCR product is sequenced with ABI BigDye Terminator v1.1 and v3.1 chemistries (Thermo Fisher Scientific, Australia) using multiple forward and reverse sequencing primers followed by capillary electrophoresis on ABI Genetic Analysers 310 and 3130XL (Thermo Fisher Scientific). All primers used have been checked for specificity and possible mismatches using PrimerBLAST and SNPCheck V3 (https://genetools.org/SNPCheck/credits.htm), respectively. Sequence analysis is performed for the entire coding region and clinically relevant non-coding regions of the HBB gene including the proximal promoter, 5ʹ untranslated region (UTR), partial introns 1 and 2, 3ʹ UTR and the polyadenylation signal site. Notably, the 619-bp deletion (HBB:c.316-149_∗342delinsAAGTAGA), which is known to be frequently detected in patients of Indian ethnicity, has been identified by detecting the 1.4kb PCR product on gel electrophoresis with the deletional breakpoint ascertained by capillary sequencing.

Of a total of 825 individual patients who have had their HBB gene sequenced since 1998, 25 (3.0%) had pathogenic variants in homozygous state and 39 (4.7%) had two different heterozygous pathogenic variants. Another 560 patients (67.9%) were carriers of one pathogenic variant. Our capillary sequencing assay did not identify any pathogenic variants in 201 patients (24.4%). This could be due to the patients having other rare large deletional variants causing their clinical symptoms, or more likely it is due to a low pre-test probability as the patients' blood samples were sent for genetic testing as part of asymptomatic prenatal screening for thalassaemias.

A total 48 different pathogenic alterations were found in our patients (Fig. 1). The top ten most frequently identified pathogenic alterations were: c.79G>A (Hb E; 20.6%), c.126_129delCTTT (Codons 41/42 –TTCT; 15.8%), c.316-197C>T (IVS-II-654 C>T; 14.4%), c.92+5G>C (IVS-I-5 G>A; 10.5%), c.-78A>G (-28 A>G; 4.8%), c.52A>T (Codon 17 A>T; 3.9%), c.316-238C>T (IVS-II-613 C>T; 3.9%), c.20A>T (Hb S; 3.6%), c.92+1G>T (IVS-I-1 G>T; 3.2%) and c.316-149_∗342delinsAAGTAGA (619 bp deletion; 3.1%). Together, these contributed to 83.9% of all the pathogenic alterations identified. These results are consistent with a previous study by Ng et al. which found that the c.126_129delCTTT, c.316-197C>T, c.92+5G>C, c.-78A>G and c.52A>T variants were among the most common mutations in β-thalassaemia patients in the Singapore population. A further 38 pathogenic variants accounted for the remaining 16.2% of HBB pathogenic alterations identified in our patients (Fig. 1). The broad mutational spectrum of HBB pathogenic alterations identified may be due to an increase in immigrants and/or international patients, as other studies have shown that migration can broaden the range of haemoglobinopathy mutations. If the β-Globin StripAssay SEA (ViennaLab Diagnostics GmbH, Austria) targeted panel had been used (which claims to detect >90% of β-globin defects found in South East Asia), only 18 of the 48 identified pathogenic alterations would have been detected (Fig. 1), translating into an expected analytical sensitivity of 80.3% in comparison to our current sequencing assay. This demonstrates the value of the capillary sequencing-based approach for molecular diagnosis of β-thalassaemias and haemoglobinopathies by detecting many rare pathogenic variants in addition to those common pathogenic variants.

However, with the whole gene analysis approach by capillary sequencing, we are occasionally challenged with interpretation of the pathogenicity of rare and novel variants. We detected six heterozygous variants of uncertain significance (VUS) in the HBB gene (Table 1). They are located in the 5ʹUTR or intron 2 of the gene, which are known to harbour clinically significant variants. These variants are not present in the HbVar, IthaGenes and HGMD databases. Three variants, c.-84G>A, c.316-233T>C and c.316-100T>A, are not present in dbSNP, ClinVar and population databases such as gnomAD, and appear to be novel variants. The MCV, MCH and HbA2 of the individuals with the c.-84G>A and c.316-100T>A variants do not appear very abnormal; unfortunately no data from family members was available to help interpret the pathogenicity of the variants. The individual with the c.-51T>C variant also was a carrier of the South East Asian deletion (--SEA/αα) and had a HBB:c.-136C>T pathogenic variant. Capillary sequencing was unable to discriminate whether the c.-51T>C and c.-136C>T variants are in cis on the same allele or in trans on two different alleles of the HBB gene in this individual. The c.316-45G>C variant was detected in seven individuals. This variant has been described as benign, likely benign and of uncertain significance in ClinVar. The two individuals with this variant with available haemoglobin analysis did not have elevated HbA2. Thus, this variant is likely benign.

The full gene sequencing approach improves the molecular diagnosis of β-thalassaemia and haemoglobinopathies by increasing the total number of pathogenic variants identified. Although the occurrence of each individual rare pathogenic variant is low, together they make up to more than 10% of the total identifiable variants in our test population, which could have been potentially missed if a targeted approach detecting only ethnic or population specific variants was applied. The identification of all pathogenic variants in the patients is important for genetic counselling for couples regarding the risk to their offspring.