The value of single-molecule real-time technology in the diagnosis of rare thalassemia variants and analysis of phenotype–genotype correlation

Thalassemia is a single gene disease, which is difficult to cure but more straightforward to diagnose and be prevented clinically. Its gene mutation types are diverse and complex. As of 2021, the LOVD (https://databases.lovd.nl/shared/genes) database had more than 2000 thalassemia and abnormal hemoglobin-related variant sites, and most of the sites have not been studied by conventional genetic testing methods, especially the large deletion variant type. At present, common clinical testing techniques for thalassemia genes include Gap-PCR, reverse dot hybridization, PCR-flow fluorescence hybridization, gene chip, MLPA, Sanger sequencing, and next generation sequencing. The conventional screening mode can only detect variants in known gene loci, which is far from sufficient for the detection of other variant loci, leading to missed diagnoses and misdiagnoses. There is therefore an urgent need to use more accurate and effective diagnostic techniques to screen thalassemia patients in clinical practice. In recent years, there have been reports of missed detections of thalassemia using conventional genetic testing methods. The SMRT technology can detect the thalassemia gene without interrupting the DNA, and can directly read the full-length gene sequence. The DNA does not need to be amplified by PCR during sequencing, which facilitates individual sequencing of each DNA molecule, and it has very long read lengths (a read length up to 30–100 kb), high accuracy (QV30 > 99.8%), no GC preference, and single-molecule resolution characteristics [7]. SMRT technology can facilitate the simultaneous detection of α-thalassemia and β-thalassemia in 1 μL of whole blood or 10–15 mL of amniotic fluid sample. It can also detect hotspots and rare variant sites and their arrangements with high accuracy, including comprehensive coverage of 2062 variant sites related to thalassemia, and detection of 18 α-globin gene deletion variants, four α-globin genes triplicate and two β-globin gene deletion. It can detect 96 samples at a time with high efficiency and high accuracy.

Xu et al. [5] first used the SMRT to sequence full-length thalassemia-related genes (HBA1/2 and HBB) to obtain complete variant information of two alleles that were difficult to obtain by conventional genetic testing techniques. Twelve hospitals in southern China assessed a comprehensive analysis of thalassemia alleles (CATSA) for identifying both α and β thalassemia genetic carrier status by third-generation sequencing (TGS). Compared with standard thalassemia variant PCR panel testing, TCS can detected 33 more positive variants, and found that the traditional PCR detection technology had 1 false negative and 8 false positive result [6]. The present study used the SMRT and conventional technologies to test the thalassemia gene in the thalassemia screening positive population in this area. The results showed that the percentage of thalassemia gene was high and the genotype was complex, rare variant types of thalassemia and the phenotypes were diverse. Among the 434 cases, 49 variant types were detected, of which 19 were detected by conventional technology and 47 were detected by SMRT technology. Compared with conventional technology, SMRT technology detected 28 more variant types. The positive detection of SMRT was 9.91% higher than that of conventional technology, and SMRT technology increased the detection of thalassemia genes. At present, the detection range of the reagents we used only included 2062 variant sites related to thalassemia on the HBA1/2 and HBB genes. HS-40 deletion occurs upstream of the α-globin gene cluster, and HBG1-HBG2 deletion occurs upstream of the HBB gene cluster. The SMRT method developed in this study focused on detection of variants in HBA1, HBA2, and HBB genes, which consisted the vast majority of thalassemia variants. With expanded primer pairs, the SMRT technology can definitely detect HS-40 and HBG1-HBG2 deletions. However, the sequencing cost will increase with more primer pairs [7]. So, it was the limit of the design of SMRT method in this study but not SMRT technology itself.

This study found 14 cases of rare deletions or triplicate α-globin genes. Among them, the —^SEA/−α^2.4 and −α^3.7/HS-40 deletion patients all manifested with HbH disease [8, 9]. Carriers of ααα^anti3.7 and ααα^anti-4.2 had normal phenotypes, but HbF was significantly increased by 12.3% and 16.2%, respectively, and HbA2 was reduced. When compounded with β-thalassemia, it can manifest as intermediate β-thalassemia due to the aggravation of the imbalance between the α and β chains, and HbF is also significantly increased [10, 11]. In the present study, among the thalassemia carriers whose detection results were −α^3.7/αα by conventional methods, two of them were found to be HKαα/αα using SMRT technology, and the misdiagnosis rate was as high as 4.17% (2/48). HKαα/αα patients presented with silent α-thalassemia, and HKαα/–^SEA patients presented with mild α-thalassemia, which is consistent with past reports [12]. Although the HBG1-HBG2 deletion combined with c.126_129delCTTT/WT had two allelic variants in the HBB gene, HBG1-HBG2 was functionally closed in adulthood and did not affect the expression of β globin, so it was clinically mild β-thalassemia. SMRT method showed that the genotype of sample D141966 was −α^3.7/αα, while by conventional Gap-PCR it was −α^3.7/−α^3.7. Validation by MLPA confirmed D141966 had heterozygous −α^3.7 deletion. To investigate the basis of this discordance, we analyzed the SNV/indels in the αα allele identified by SMRT method and found there were three SNPs in the 3′-terminal of HBA2, which caused dropout of the αα allele in conventional Gap-PCR method that designed primer in this region.

This study found 16 rare HBA gene variants. Among them, c.34A>C, c.51G>C, c.84G>T, and c.19G>T were located in the HBA1 gene. c.34A>C and c.51G>C showed normal hematology, and abnormal hemoglobin was detected [13, 14]. Carriers with HBA1:c.84G>T, HBA1:c.19G>T, and HBA1:c.55G>C genotypes had a normal blood phenotype. When they were compounded with other deletion types, they could be mild or silent [15,16,17]. Carriers of these gene variants all showed abnormal hemoglobin, and no HbH phenotype was found in the compound Southeast Asian deletion. The —^SEA/HBA2:c.2delT, —^SEA/HBA2:c.2T>C, and —^SEA/HBA2:c.52G>T are located in the more functional HBA2 gene, causing α chain synthesis to be affected, showing that the non-deletion HbH disease was more serious than the deletion of HbH disease [18,19,20]. HBA2: c.91G>C has a normal phenotype, the main manifestation is abnormal hemoglobin, and HbA2 is reduced [21]. Qadah et al. [22] reported that the HBA2: c.−59C>T variant caused a significant reduction in the transcription level of HBA2 by 53.7%. Our study reported, for the first time, HBA2:c.−59C>T and HBA2:c.91G>C compound heterozygous cases. Hemoglobin electrophoresis detected abnormal hemoglobin peaks at 3.784 min and 4.349 min, and the routine blood phenotype was normal, due to the abnormal hemoglobin peak time being very close to HbA2, so it could be easily misdiagnosed as a significant increase in HbA2. There are related reports of HBA2:c.256G>C [23], but no related reports of HBA2:c.256G>A. The phenotype of this case was normal, the main manifestation was abnormal hemoglobin, and HbA2 was reduced.

Among the rare variants in the HBB gene, carriers with c.−100G>A, c.−136C>G, c.315+5G>C, c.380T>G, and c.−81A>C were manifested as silent or mild β-thalassemia. The normal hematological phenotype of some cases is consistent with related reports [24,25,26,27]. Carrier with c.91A>G was manifested as mild β thalassemia, which is consistent with related reports [28]. Carriers with c.170G>A, c.431A>G, c.232C>T, c.341T>A, and c.431A>G s had normal hematological phenotypes. The content of HbA2 and HbF was within the reference range, and abnormal hemoglobin was detected [29,30,31,32]. In the first report of c.341T>A/c.315+5G>C case, abnormal hemoglobin accounted for 93.5%, and HbA was not detected. Carrier with c.431A>G had the peak time of abnormal hemoglobin and HbA2 overlapped, and the content of each component could not be detected correctly. The blood routine examination of c.−248A>G carrier was normal, mainly manifested as a decrease in HbA2 content [33].The phenotype of c.316-45G>C, c.316-179A>C, and c.315+308delA combined with other types of α-thalassemia may be as silent or mild α-thalassemia.