Detection of hemoglobin H disease by long molecule sequencing.

BACKGROUND: Hemoglobin H (Hb H) disease is a moderate-to-severe form of α-thalassemia (α-thal), and parts of patients may require intermittent transfusion therapy, especially during intercurrent illness. However, rare Hb H diseases remain undetected using routine methods being outside of the testing scope. In this study, we present an approach to detecting Hb H disease by long molecule sequencing (LMS).
METHODS: A total of 206 known genotype samples were collected and carried to blind detected by LMS on the PacBio Sequel platform. Circular consensus sequencing reads were aligned to the hg19 reference genome using Free-Bayes finished LMS. LMS accuracy would be compared with routine methods, including Gap-PCR and PCR-Reverse dot blot hybridization (PCR-RDB).
RESULTS: The assay could detect carriers of both deletion and point mutations. It had an overall accuracy of 100% when compared with routine methods. In addition, LMS detected six mutations based on routine methods and corrected three case results. Hb H diseases were identified using LMS, whether a common or rare genotype, a deletion or non-deletion genotype. However, two cases of Hb H disease were misdiagnosed using routine methods.
CONCLUSIONS: Long molecule sequencing can be suggested as a rapid and reliable assay to detect probable carriers of hemoglobinopathies. LMS accurately identified the common and rare genotypes of Hb H disease.

INTRODUCTION

Hemoglobinopathy is one of the most common inherited disorders in southern China, including thalassemia and hemoglobin (Hb) variant. , , α‐globin gene cluster on chromosome 16 mainly has two expressed HBA2 and HBA1 genes, and mutations cause α‐thalassemia (α‐thal) in the α‐globin gene cluster. Deletion or dysfunction of one gene results in thalassemia silent, and two genes result in thalassemia trait; compound deletion of two genes and deletion or point mutation of one gene cause α‐thal intermedia hemoglobin H disease (Hb H disease), while loss of all four genes causes Hb Bart's hydrops fetalis. There are two types of Hb H disease: deletional Hb H (‐‐/‐α) and non‐deletional Hb H (‐‐/αTα). In the southern China, most α‐thals are caused by deletions, and the rest are caused by point mutation. Therefore, there is more deletional Hb H than non‐deletion Hb H disease. However, a kit with Chinese FDA approval for the deletional thalassemia test in clinical laboratories is only used to detect ‐‐SEA, ‐‐THAI, ‐α3.7, and ‐α4.2 in the Chinese population. , Predictably, there will be misdiagnosed cases in our routine clinical works if the routine genetic analysis is used. In the high‐prevalence area of thalassemia, it is thus important to have a rapid and reliable detection method for the deletional α‐thal s test, which avoids misdiagnosing of Hb H disease or Hb Bart's hydrops fetalis.

Third‐generation sequencing (TGS), as a current new molecule technique, was described previously to detect the thalassemia and Hb variants on the PacBio Sequel II platform. , Long molecule sequencing (LMS) is the main characteristic of TGS. In this study, we applied the LMS to test 206 suspected thalassemia samples, compared with the routine genetic analysis. Our results suggested that LMS, the same as routine genetic analysis, could detect deletional mutations and point mutations; moreover, it also detected rare mutations to diagnose Hb H disease. So, it can be an effective and reliable approach for thalassemia screening of suspected individuals with rare mutations or outside of the detection scope.

MATERIALS AND METHODS

Sample

A total of 206 subjects with suspected thalassemia and Hb variants were collected from thalassemia screening in the People's Hospital of Guangxi Zhuang Autonomous Region. Their age ranged from 9‐month to 65 years. The study was approved by the People's Hospital of Guangxi Zhuang Autonomous Region ethics committee. All the individuals provided informed written consent. Whole blood samples were collected in tubes with K2‐EDTA.

Complete blood count and hemoglobin analysis

Currently, complete blood count (CBC) and Hb analysis are routinely used as screening tests for thalassemia. Red cell indices were determined with an automated hematology analyzer (Sysmex XT 1800i; Sysmex Corporation). Subjects with a mean corpuscular volume (MCV) of <82 fl or a mean corpuscular (MCH) of <27 pg were suspected of thalassemia. Hb analysis was performed by capillary electrophoresis (CE) (Capillary 2 Flex Piercing; Sebia). When HbA2 was>3.5%, β‐thalassemia but not excluded α‐thal was considered, whereas when HbA2 was<2.5%, α‐thal was suspected (This is the thalassemia screening process recommended by our government policy, but we have to explain that HbA2 value of 2.5%–3.5% cannot rule out the presence of α‐thal.). In addition, abnormal peaks of CE are also regarded as positive screening results. All the suspected samples were performed by routine genetic analysis and LMS, respectively.

Gap‐PCR and PCR–Reverse dot blot hybridization (PCR–RDB) (routine genetic analysis/routine methods, RGA)

Gap‐PCR and PCR–RDB as routine genetic analysis methods are widely used in the clinical laboratories in China. Gap‐PCR was applied to detect the common deletion‐type α‐thal in the Chinese population, including ‐‐SEA, ‐‐THAI, ‐α3.7, ‐α4.2. PCR–RDB was conducted to detect non‐deletion α‐globin mutations Hb Constant Spring (Hb CS, HBA2: c.427 T>C, p.Ter143GlnextTer31), Hb Westmead (Hb WS, HBA2: c.369C>G, p.His123Gln), Hb Quong Sze (Hb QS, HBA2: c.377 T>C, p.Leu126Pro). PCR‐RDB was also carried out to detect the 17 common β‐thalassemia mutations: ‐32 (HBB: c.‐82C>A), −30 (HBB: c.‐80 T>C), −29 (HBB: c.‐79A>G), −28 (HBB: c.‐78A>G), CDs14‐15 (HBB: c.45_46insG), CD17 (HBB: c.52A>T, p.Lys18Ter), CD26 (HBB: c.79G>A, p.Glu27Lys), CDs27‐28(HBB: c.84_85insC), CD31 (HBB: c.94delC), CDs41‐42 (HBB: c.124_127delTTCT), CD43(HBB: c.130G>T, p.Glu44Ter), CDs71‐72 (HBB: c.216_217insA), IVS‐I‐1 (HBB: c.92 + 1G > T), IVS‐1‐5 (HBB: c.92 + 5G > C), IVS‐II‐654 (HBB: c.316–197C > T), Int (HBB: c.2 T > G), and CAP (HBB: c.‐50A>C). Genomic DNA was extracted from 100 μl whole blood samples using the DNA extraction kit (YanengBio). Gap‐PCR (Yishengtang Bio) and PCR–RDB (Yaneng Bio) were done following the kit's instructions.

Long molecule sequencing and data analysis

As previous studies described, , genomic DNA samples prepared with the DNA blood mini kit (QIAGEN) were performed multiplex long PCR using an assay Kit (Berry Genomics) after assessing DNA quality. Each PCR product was ligated with a unique identifying barcoded adaptor following purification and end‐repair. Single‐molecule Real‐time (SMRT) bell libraries were prepared using the Sequel Binding and Internal Ctrl Kit 3.0 (Pacific Biosciences). Then the DNA‐polymerase complexes were loaded on SMRT cells, and sequencing was performed on the PacBio Sequel platform. In order to obtain aligned subreads, raw reads were processed by the Lima application and the circular consensus sequencing (CCS) software application. The filtered CCS reads were aligned to the human reference genome (hg19), and NGMLR (version 0.2.6) or BLASR and variants called using FreeBayes (version 1.2.0) was selected according to different purposes. Alignments of variant and wild‐type molecules were presented by Integrative Genomics Viewer (IGV).

Confirmation of rare mutations detected by LMS

Rare α‐thal deletions were confirmed by Gap‐PCR using a rare deletional thalassemia kit (Yaneng Bio). Point mutations of the α‐globin gene and β‐globin gene were confirmed by Sanger sequencing as we previously described. ,

RESULTS

Gap‐PCR and PCR–RDB

Of the 206 suspected cases, 140 subjects were diagnosed with thalassemia and Hb variants using routine Gap‐PCR and PCR–RDB methods. Of these thalassemias and Hb variants, the ‐‐SEA/αα mutation (34.3%) was the most common α‐thal genotype, followed by ‐α3.7/αα (12.1%) and αCSα/αα (5.7%); the most common β‐thalassemia mutation was βCDs41‐42/βN (15.7%), followed by βCD17/βN (3.6%) and β−28/βN (2.9%) (Table 1). Besides, 9 cases were diagnosed with compound α and β‐thalassemia, including eight genotypes (Table 1).

TABLE 1

The α‐globin and β‐globin gene mutation spectrum in this study, excluding Hb H disease

Mutation type	Number	Detection by RGA	Detection by LMS	Concordance of two methods
α‐thalassemia/Hb variant
‐‐SEA/αα	48	48	48	Y*
‐α3.7/αα	17	17	17	Y
αCSα/αα (Hb CS)	8	8	8	Y
‐α4.2/αα	4	4	4	Y
αWSα/αα (Hb WS)	2	2	2	Y
αQSα/αα (Hb QS)	1	1	1	Y
‐α3.7/αWSα	1	1	1	Y
‐α4.2/αCSα	1	1	1	Y
‐‐SEA/‐‐SEA	1	1	0	N
β‐thalassemia/Hb variant
βCDs41‐42/βN	22	22	22	Y*
βCD17/βN	5	5	5	Y
β−28/βN	4	4	4	Y
βCDs71‐72/βN	3	3	3	Y
βIVS‐II‐654 /βN	1	1	1	Y
βCDs27/28/βN	1	1	1	Y
βCD26/βN (Hb E)	1	1	1	Y
βIVS‐I‐1 /βN	1	1	1	Y
Compound α and β‐thalassemia
αCSα/αα, βCDs41‐42/βN	2	2	2	Y
‐α3.7/αα, β−28/βN	1	1	1	Y
‐α4.2/αα, βCDs41‐42/βN	1	1	1	Y*
‐‐SEA/αα, βCDs71‐72/βN	1	1	1	Y
αQSα/αα, β−28/βN	1	1	1	Y
αWSα/αα, βCDs41‐42/βN	1	1	1	Y
αWSα/αα, βCDs71‐72/βN	1	1	1	Y
‐α3.7/αα, βIVS‐II‐654 /βN	1	1	1	Y*

Note: Y*: The detection results of the two methods are consistent, but the LMS has additional detections.

Abbreviations: LMS, long molecule sequencing; N, No; RGA, routine genetic analysis; Y, Yes.

In this study, 10 cases were detected with Hb H disease. Among them, 6 cases were diagnosed with deletional Hb disease (‐‐SEA/‐α3.7 and ‐‐SEA/‐α4.2), while 3 cases were non‐deletional Hb H disease (‐‐SEA/αCSα and ‐‐SEA/αWSα). A case of compound Hb H and β‐thalassemia intermedia was accidentally detected. As Table 2 illustrates, 8 out of 10 Hb H diseases showed anemia, and the remaining two cases of ‐‐SEA/αWSα had no anemia but decreased MCV and MCH values.

TABLE 2

Hematological phenotypes and genotypes of Hb H disease in this study

Mutation type	Sex	Age (year)	Hb (g/L)	MCV (fl)	MCH (pg)	Hb A (%)	Hb A2 (%)	Hb F (%)	Hb H (%)	Hb Bart's (%)	Detection by RGA	Detection by LMS
‐‐SEA/‐α3.7	M	45	98	59.4	16.9	97.8	1.1	0	1.1	0	Y	Y
‐‐SEA/‐α3.7	F	31	89	57.3	17.4	98.0	1.3	0.2	0.5	0	Y	Y
‐‐SEA/‐α3.7	F	5	89	53.8	15.5	96.0	1.5	0.3	1.7	0.5	Y	Y
‐‐SEA/‐α3.7	F	35	86	65.6	18.6	96.5	0.9	0	2.4	0.2	Y	Y
‐‐SEA/‐α4.2	F	25	89	52.5	16.9	98.7	1.3	0	0	0	Y	Y
‐‐SEA/‐α4.2	F	9/12	88	45.0	15.7	96.1	1.6	2.3	0	0	Y	Y
‐‐SEA/αCSα	M	27	82	74.2	20.1	87.6	1.1	0	8.5	0	Y	Y
‐‐SEA/αWSα	M	23	122	74.8	23.1	97.8	2.2	0	0	0	Y	Y
‐‐SEA/αWSα	M	26	145	63.8	20.4	97.5	2.5	0	0	0	Y	Y
‐‐SEA/‐α3.7 βCDs41‐42/β−28	M	30	105	55.6	25.1	32.7	10.9	56.4	0	0	Y	Y
‐‐SEA/‐α27.6	F	1	82	53.7	15.3	98.3	1.7	0	0	0	N	Y
‐‐SEA/Fusion gene	M	27	108	75.5	21.2	67.5	0.5	0	31.0	1.0	N	Y

Abbreviations: LMS:long molecule sequencing; RGA: routine genetic analysis;

LMS analysis

After the raw data was processed, we obtained high‐quality sequencing data. LMS identified 81 carriers of α‐thal or α‐chain Hb variants, 38 carriers of β‐thalassemia, and 11 carriers of compound α‐thal (α‐chain Hb variants) and β‐thalassemia. All of the mutations detected by routine genetic analysis could also be detected by LMS (Table 1). LMS additionally identified four rare Hb variants, including Hb Hekinan II (HBA1:c.84G > T), Hb G‐Siriraj (HBB: c.22G > A), Hb Binyang (HBA2:c.40G > T), and Hb Q‐Thailand (HBA1:c.223G > C), which were not detected by routine genetic analysis. A case genotype was Hong Kong αα by LMS, but the routine genetic analysis showed a 3.7 deletion heterozygote (Table 3).

TABLE 3

Different results detected by LMS and RGA in this study

Number	Detection by RGA	Detection by LMS	Comment
1	‐‐SEA/αα	‐‐SEA/fusion gene	Omission diagnosis
2	‐‐SEA/‐‐SEA	‐‐SEA/‐α27.6	Omission diagnosis
3	‐α4.2/αα, βCDs41‐42/βN	‐α4.2‐Q‐Thailand/αα, βCDs41‐42/βN	Omission diagnosis
4	‐α3.7/αα, βIVS‐II‐654 /βN	HKαα/αα, βIVS‐II‐654 /βN	Mistake diagnosis
5	αα/αα, βN /βN	αα/αα, βCD7 /βN (Hb G‐Siriraj)	Omission diagnosis
6	αα/αα	αα/αCD13α (Hb Binyang)	Omission diagnosis
7	‐‐SEA/αα, βN /βN	‐‐SEA/αα, β−50/βN	Omission diagnosis
8	αα/αα	αα/αCD30α	Omission diagnosis
9	αα/αα, βCDs41‐42/βN	αα/αCD27α(HbHekinanII), βCDs41‐42/βN	Omission diagnosis

Abbreviations: LMS, long molecule sequencing; RGA, routine genetic analysis.

In addition, LMS detected 12 cases with Hb H disease (Figures 1 and 2), two more cases than routine genetic analysis. In other words, 2 cases with Hb H disease were misdiagnosed using the routine genetic analysis method (Figure 3). One case was detected as a SEA heterozygote (‐‐SEA/αα) by routine genetic analysis, but compound SEA heterozygote and ψα1–α2 fusion gene (‐‐SEA/Fusion gene) by LMS (Table 3). The other case was identified as a SEA homozygote (‐‐SEA/‐‐SEA) by routine genetic analysis, but LMS presented compound SEA deletion and α27.6 deletion (‐‐SEA/‐α27.6) (Table 3).

FIGURE 1

Integrative Genomics Viewer (IGV) plots of common deletional Hb H disease, including ‐‐SEA/‐α3.7 (A) and ‐‐SEA/‐α4.2 (B). IGV showed the ‐‐SEA deletion in the yellow area, the ‐α3.7 deletion in the blue area (A), and the –α4.2 deletion in the blue area (B).

FIGURE 2

Integrative Genomics Viewer (IGV) plots of common non‐deletional Hb H disease, including ‐‐SEA/αCSα (A) and ‐‐SEA/αWSα (B). IGV of (A) showed the ‐‐SEA deletion in the yellow area and one allele with αCS mutation in the blue area. (B) revealed the ‐‐SEA deletion in the yellow area and one allele with αWS mutation in the blue area.

FIGURE 3

Integrative Genomics Viewer (IGV) plots of rare deletional Hb H disease, including ‐‐SEA/‐α27.6 (A) and ‐‐SEA/Fusion gene (B). (A) Displayed the ‐‐SEA deletion in the yellow area and the ‐α27.6 deletion in the blue area. (B) Presented the ‐‐SEA deletion in the yellow area and the other allele with the ψα1–α2 fusion gene.

DISCUSSION

In the present study, LMS, as well as the routine genetic analysis, detected 140 carriers with deletion or point mutations. In addition, nine mutations were detected, which were not detected by routine genetic analysis. The thalassemia routine kits are only used to identify the 24 common thalassemias and variants in the Chinese population, so misdiagnosis frequently occurs. However, LMS performs long fragment amplification, including entire HBA1/2 and HBB genes, and then mutations can be easily viewed with the IGV tool after finished bioinformatics analysis. Therefore, since the complete gene regions are analyzed, the assay can detect mutations outside the scope or difficult to detect using routine genetic analysis.

The deletional Hb H (‐‐/‐α) disease is caused by compound heterozygote of α0‐ and α+‐thalassemia (‐α/) alleles. The common genotype of deletional Hb H diseases are ‐‐SEA/‐α3.7 and ‐‐SEA/‐α4.2, and a few are ‐‐THAI/‐α3.7 and ‐‐THAI/‐α4.2 in southern China. , Our study showed that LMS identified 6 cases of common deletional Hb H disease and a case of Hb H combined with β‐thalassemia intermedia as routine methods. The difference was that 1 case was identified as SEA homozygote (‐‐SEA/‐‐SEA, Hb Bart's hydrops fetalis) by routine genetic analysis, while LMS was identified as (‐‐SEA/‐α27.6, Hb H disease) (Figure 3A). The ‐α27.6 is a rare deletional mutation type and was first reported in a Chinese family, outside the routine method's test scope. In fact, there were some occasional reports of similarly rare large deletion of α‐chain. , , Since the normal control primers of the routine genetic analysis are designed on the HBA2 gene, it will be detected as a SEA homozygote if combined with a SEA heterozygote as long as the large fragment deletional sample involves this region. However, we considered Hb H disease combined with the comprehensive screening results. Based on the results described earlier, we recommend that LMS operate as a reliable and rapid approach for detecting these similar rare Hb H diseases as ‐‐SEA/‐α27.6.

As shown in Table 3, although most Hb H diseases were deletional types, there are also non‐deletional types. The common non‐deletional Hb H diseases were identified either by the routine genetic analysis or LMS, such as ‐‐SEA/αCSα and ‐‐SEA/αWSα (Figure 2 A,B). However, we need to clarify that the phenotype of ‐‐SEA/αWSα is similar to SEA heterozygotes, which is not classified as Hb H disease, but the Hb WS is a mutation that is routinely detected in China. Surprisingly, LMS detected a case of Hb H disease with compound SEA heterozygote and ψα1–α2 fusion gene, but it was displayed as a SEA heterozygote using routine genetic analysis (Figure 3B). The fusion gene is a point mutation type involving seven alleles (HBA2:c.*64 T>C, HBA2:c.*68A>C, HBA2:c.*71G>A, HBA2:c.*74C>A, HBA2:c.*82G>A, HBA2:c.*92A>G, and HBA2:c.*98 T>C), and compound SEA heterozygote and ψα1–α2 fusion gene is a rare type of Hb H disease. , This suggested that detecting non‐deletional Hb H disease using LMS had high accuracy.

Hemoglobin H disease is heterogeneous, and the clinical manifestations vary widely from mild asymptomatic anemia to a severely anemic state. An accurate diagnosis of thalassemia genotype is important for couples who have pregnancies affected with Hb H disease or Hb Bart's hydrops fetail syndrome or anemia patients who need to be treated. Currently, diagnosis of Hb H disease needs several tests by routine methods, which are required for non‐deletion and deletion mutations in separate tests. In addition, the test mutations of the routine kits are limited. If a rare or novel large fragment deletion is suspected, multiplex ligation‐dependent probe amplification (MLPA) screening is first required, followed by Gap‐PCR amplification, and then Sanger sequencing determines the break sites. However, as presented in this study, LMS provided a rapid and more effective method for identifying the common Hb H genotype and enabled the detection of the rare Hb H genotype. It can avoid the time‐consuming and laborious of MLPA method and possible misdiagnosis. So, for anemia patients with iron deficiency anemia excluded, or couples of childbearing age with suspicious thalassemia screening results, LMS can be considered a suitable method when the routine genetic analysis cannot be detected (Figure 4).

FIGURE 4

Work flow of Hb H disease diagnosis by routine genetic analysis and LMS.

AUTHOR CONTRIBUTIONS

All the authors have accepted responsibility for the entire content of this manuscript and approved its submission.

FUNDING INFORMATION

This study was supported by the Health Department Research Fund of Guangxi (Z20200076).

CONFLICT OF INTEREST

Authors state no conflict of interest.