Significance of borderline HbA2 levels in β thalassemia carrier screening

Increased HbA2 levels are the characteristic feature of β-thalassemia carriers. A subset of carriers however do not show HbA2 levels in the typical carrier range (≥ 4.0%) but show borderline HbA2 levels. As a result, these carriers escape diagnosis and carry the risk of having β-thalassemia major offspring. Borderline HbA2 values may occur as a consequence of mild β-thalassemia mutations, co-inherited β-thalassemia and α- or δ- thalassemia or iron deficiency anemia. However, there is insufficient knowledge regarding the cause of borderline HbA2 levels in specific populations. This study aimed to identify the determinants of borderline HbA2 levels (which we have considered as HbA2 3.0–3.9%) in 205 individuals. Primary screening involved detecting the presence of iron deficiency anemia followed by molecular analysis of α, β and δ globin genes. Remarkably, 168 of 205 individuals were positive for a defect. 87% (149/168) of positive individuals were heterozygous for β thalassemia with (59/149) or without (90/149) the presence of co-existing IDA, α or δ gene defects. Notably, 20 of 149 β thalassemia carriers showed HbA2 < 3.5% and MCV > 80fL. 7 of these 20 carriers were married to carriers of hemoglobinopathies. Our findings describe the genetic basis of borderline HbA2 levels and emphasize the necessity of a molecular diagnosis in these individuals in the routine clinical setting.

Primary screening and molecular analysis. 10 8.9] was performed, when required, to rule out the presence of abnormal hemoglobins. Genomic DNA was isolated from peripheral blood leucocytes using the QIAamp Blood Mini Kit [Qiagen GmbH, Hilden, Germany]. β-thalassemia mutations were characterized by reverse dot-blot hybridization 13 or the amplification refractory mutation system [ARMS] 15 . Analysis of eight common deletional α thalassemia determinants [-α 3.7 , -α 4.2 , --SEA , --THAI , --FIL , --MED , -[α] 20.5 and [--SA ] was performed using multiplex polymerase chain reaction 16  Ethics approval. The study was approved by the National Institute of Immunohematology-Institutional Ethics Committee. All methods were performed in accordance with relevant guidelines and regulation. Consent to participate. Informed consent was obtained from all individual participants included in the study.

Results
HbA 2 and MCV levels of the study subjects. Of 205 individuals analysed in this study, 19 individuals had HbA 2 ranging between 3.0 and 3.2%, 59 individuals had HbA 2 ranging between 3.3 and 3.4% and 127 individuals had HbA 2 ranging between 3.5 and 3.9%. Molecular analysis of the β globin gene revealed a β-thalassemia allele in 149 of 205 [73%] individuals and a normal β genotype in the remaining 56 [27%] individuals. Figure 1. shows the frequency of β-thalassemia carriers identified at different HbA 2 levels investigated in our study.  www.nature.com/scientificreports/ β-thalassemia mutations. Eleven β-thalassemia mutations were identified in the 149 carriers with borderline HbA 2 levels (Table 1). β + IVS1-5 G>C was the most common mutation (51%), followed by the milder β ++ mutations such as Cap site +1 A>C, Poly A T>C and Poly A -AATAA, together accounting for 40% of carriers. β° mutations were identified in 9% of carriers with borderline HbA 2 levels.
In the HbA 2 3.0-3.2% range, 19 individuals were analysed and a β-thalassemia allele was identified in 15 individuals (79%). 12 of 15 β thalassemia heterozygotes showed presence of a β ++ thalassemia allele [Cap + 1 A > C or Poly A T > C with co-inheritance of either − α 3.7 /αα deletion, HbA2 Saurashtra or IDA] and 3 of 15 β thalassemia heterozygotes showed presence of the β + IVS1-5 G > C allele. β 0 thalassemia defects were not identified in this group. Four individuals in this HbA 2 group showed a normal β genotype of whom one was iron deficient and in the remaining three individuals, no defects could be identified.
MCV-based analysis. Since MCV and MCH values, together with HbA 2 levels, play a critical role in carrier diagnosis we analysed these parameters in the 149 β thalassemia heterozygotes identified in our study ( Table 2). 10 of 15 β thalassemia heterozygotes with HbA 2 between 3.0 and 3.2% showed MCV > 80 fL. Similarly, 10 of 33 β thalassemia heterozygotes with HbA 2 between 3.3 and 3.4% showed MCV > 80 fL. Thus, 20 of 149 [14.0%] of all β thalassemia heterozygotes would be missed from being detected if MCV < 80 fL and HbA 2 ≥ 3.5% were used as cut-off for carrier screening. These findings are of clinical significance because 5 of the 10 β thalassemia heterozygotes with HbA 2 between 3.0-3.2% and MCV > 80 fL and 2 of the 10 β thalassemia heterozygotes with HbA 2 between 3.3-3.4% and MCV > 80 fL were married to carriers of a hemoglobinopathy. A "missed" diagnosis of these 7 heterozygotes could have led to the birth of an affected child. Notably, of these 20 above mentioned β thalassemia heterozygotes with HbA 2 between 3.0 and 3.4% and MCV > 80, 10 heterozygotes also showed

MCV < 80 fL (n = 130). A molecular defect could be identified in 123 of 130 individuals with borderline
HbA 2 levels and MCV < 80fL. Heterozygosity for β-thalassemia accounted for 93.4% (115 of 123) of defects identified. In this MCV group, the β + IVS1-5 G > C mutation was identified in 60 of 115 (52%) carriers, β ++ mutations were identified in 45 of 115 (39%) carriers and β 0 mutations were identified in 10 of 115 (8.6%) carriers. Among the 115 β-thalassemia heterozygotes identified in this MCV group, 64 of 115 (56%) did not show presence of co-existing defects and 51 of 115 (44%) β-thalassemia heterozygotes showed concomitant IDA or co-inheritance of α or δ globin gene defects. A novel δ globin gene defect CD 46 G > T/HBD:c.140G > T was identified in a β-thalassemia heterozygote in this MCV group. In the eight individuals with a normal β genotype, α thalassemia was detected in two individuals, δ globin defects were identified in two individuals, IDA was noted in two individuals, one iron deficient sample showed the -α 4.2 /αα deletion and lastly one iron deficient sample showed the δ promoter defect −68 C > T. Genotypes and hematologic findings of these individuals are shown in Table 3.  In the remaining 11 individuals with a normal β genotype, three showed presence of α deletions and four showed a δ globin gene mutation, two individuals were iron deficient, one showed IDA; δ globin gene defect and one showed coinheritance of α thalassemia and a δ globin gene defect. Genotypes and hematologic findings of these individuals are shown in Table 3. In our study, heterozygosity for β + IVS1-5 G > C (52% and 47%) and heterozygosity for the β ++ thalassemia mutations (47% and 44%) were the most prevalent defects in individuals with MCV < 80fL and MCV > 80fL, respectively. To get a better understanding of the distribution of MCV and HbA 2 levels of these heterozygotes we plotted a scatter diagram of the MCV and HbA 2 levels of the carriers (Fig. 3). From the plot we observe that carriers of β + IVS1-5 G > C and β ++ defects show largely similar HbA 2 levels and MCV values and note the absence of IVS1-5 G > C β thalassemia carriers with MCV < 80 and HbA 2 < 3.3%.
To summarize our findings, we show that individuals with borderline HbA 2 levels are not rare in the Indian population with approximately 82% [168/205] of them harbouring a molecular or acquired defect. We found that at all HbA 2 levels investigated in this study, and irrespective of MCV (< 80 or > 80 fL), heterozygosity for β-thalassemia was the most common defect. The severe β + thalassemia allele, IVS1-5 G > C, was identified in Table 3. Hematological analysis of the individuals with MCV ≤ 80 fL and MCV ≥ 80 fL. www.nature.com/scientificreports/ heterozygotes with MCV < 80 fL and HbA 2 > 3.5% as well as in heterozygotes with MCV > 80 fL and HbA 2 < 3.5%. We found that the co-inheritance of α or δ globin gene defects was more common in individuals with borderline HbA 2 levels and MCV < 80fL (12.1% and 16.5%, respectively) as compared to that in individuals with borderline HbA 2 levels and MCV > 80fL (5.9% and 5.9%, respectively). 40% of individuals with borderline HbA 2 levels and MCV > 80fL did not show any defect. Since most of the defects identified were common to all three HbA 2 groups and both MCV groups we were unable to associate any defect exclusively to a particular HbA 2 or MCV group.
In the course of our study we identified 48 β-thalassemia heterozygotes who showed HbA 2 3.0-3.4%. Of these 48 heterozygotes, 20 also showed MCV > 80fL. These cases emphasize that a molecular work up of the β globin gene is the only way to achieve a confident diagnosis of individuals with borderline HbA 2 levels.

Discussion
Individuals affected by β-thalassemia major require regular blood transfusions and lifelong medical care to survive. Research in ameliorating the pathological effects and in the treatment of the disease are ongoing 18 , however, the newer treatment modalities are high-priced and often unaffordable to the general population in low and middle income countries. It cannot be stressed enough that this complex disease is preventable if simple and cost-effective measures such as carrier screening and genetic counselling are rigorously employed. The heterozygous form of β-thalassemia is associated with a mild persistent anemia and distinctly elevated levels of HbA 2 , which form the basis for screening programs world-over. Approximately 80-90 million people worldwide are carriers of β-thalassemia 19 with India alone harbouring 35-45 million carriers 14 . From this perspective, the implications of borderline HbA 2 levels in the diagnosis of β-thalassemia holds immense significance. In this study we have identified the factors responsible for borderline HbA 2 levels, have investigated the effects of confounding factors on HbA 2 and MCV levels and have interpreted the effects of these parameters in β-thalassemia carrier screening.
Heterozygosity for β-thalassemia was the most common cause of borderline HbA 2 levels in our study population. 40% β-thalassemia heterozygotes showed concomitant presence of IDA or α or δ globin gene defects while 60% heterozygotes showed absence of these confounding factors.
Presently, our lab and different laboratories worldwide consider HbA 2 > 3.2% as the second level of diagnosis of β thalassemia. However, during the time this study was conducted (2009-2013) our laboratory and other laboratories offering β thalassemia screening 4,21 used to consider HbA 2 ≥ 4.0% as cut-off for diagnosis of heterozygous β thalassemia which is why, in this study, we have considered borderline HbA 2 levels as 3.0-3.9%.
On stratifying the β-thalassemia heterozygotes on the basis of HbA 2 , we found that 32% had HbA 2 levels in the range of 3.0-3.4% and 68% had HbA 2 levels 3.5-3.9%. Several laboratories consider HbA 2 ≥ 3.5% as the cutoff for identification of β-thalassemia carriers during screening programmes and in the absence of molecular testing, had HbA 2 levels been the sole diagnostic determinant for β-thalassemia carriers, the 48 heterozygotes showing HbA 2 3.0-3.4% would be misdiagnosed as non-carriers [false negatives]. Interestingly, 15 of the 48 false negative heterozygotes were partners of HbS or β-thalassemia heterozygotes. 20 of these 48 false negative heterozygotes showed MCV > 80fL and 10 of these 20 individuals also showed MCH > 27 pg. These findings highlight how, despite using a combination of MCV and HbA 2 values (and sometimes, even MCH) for screening, β thalassemia heterozygotes could still be misdiagnosed as non-carriers. As documented by Giambona et al. 1 , it is important and highly relevant to detect all β-thalassemia carriers for a prevention screening program aimed at the identification of at-risk couples. Collectively, our findings emphasize the need to offer molecular screening of the β globin gene to partners of carriers of hemoglobinopathies, irrespective of their hematological   13 . In our study, 11 β-thalassemia mutations were associated with borderline HbA 2 levels. Five mutations identified in our borderline HbA 2 individuals overlapped with the spectrum of common β-thalassemia mutations in India (IVS1-5 G > C, IVS1-1 G > T, CD 8/9 + G, CD 41/42 -CTTT, CD 15 G > A and CD 30 G > C). We identified β ++ , β + and β 0 thalassemia mutations in our borderline HbA 2 β-thalassemia heterozygotes. A study by Rangan et al. 12 has previously also identified β ++ , β + or β 0 thalassemia defects in 8 of 25 [32%] individuals with HbA 2 3.0 -4.0%. Association of β 0 and β + thalassemia mutations in borderline HbA 2 individuals is also not uncommon in other populations [21][22][23] . Overall, β + IVS1-5 G > C mutation was the most common defect identified in borderline HbA 2 individuals in our study [51%], followed by the milder β ++ thalassemia mutations [40%] and the β 0 thalassemia mutations [9%].
We found that at lower HbA 2 levels (3.0-3.2%), the β ++ thalassemia alleles are more common, β 0 thalassemia alleles did not occur at HbA 2 < 3.2% but as HbA 2 levels increase (3.3-3.9%) the number of heterozygotes for β + IVS1-5 G > C increase. These findings can be explained by the fact that IVS1-5 G > C mutation is the most common β-thalassemia mutation in the Indian population with a prevalence of about 60% in Western India 24 . Our study for the first time highlights that this severe β + thalassemia allele that has been conventionally associated with elevated HbA 2 levels in past studies may also be the most common allele in borderline HbA 2 carriers in India. Indeed, heterozygosity for a β + thalassemia allele, IVS1-6 T > C, is reported as the most common cause of borderline HbA 2 levels in the Mediterranean region 1,25 .
Many factors influence HbA 2 levels besides the β-thalassemia alleles, such as α-thalassemia, δ-thalassemia and severe IDA 26 . The co-inheritance of α-or δ-thalassemia in β-thalassemia carriers has also been reported to lower/normalize HbA 2 levels 25,27,28 . To identify if these genotypes were associated with HbA 2 levels in our study population, we analysed their α and δ globin genotypes and evaluated the presence of iron deficiency anemia.
A high prevalence of associated α thalassemia is reported in Indian β-thalassemia carriers 29 . In our study, the co-inheritance of α thalassemia was noted in 10.7% β-thalassemia heterozygotes. A recent study from Thailand has reported that as many as 43.75% β thalassemia heterozygotes with HbA 2 3.5-3.9% also co-inherited α thalassemia 30 . Reports also show that co-inheritance of α-and β-thalassemia could normalize MCV and MCH levels leading to misdiagnosis 28,31 . We did not observe such an effect in our study: of the 16 α and β-thalassemia double heterozygotes identified in our study, 14 showed MCV < 80 fL and 14 showed MCH < 27 pg. Another group has also reported a significant improvement in MCV and MCH values due to β + or β° thalassemia mutations interacting with one or two α globin gene abnormalities [(− α/αα), (α T α/αα) or (− − /αα)] in the Thai population 32 . In our study, of the seven β + thalassemia heterozygotes with co-inherited α thalassemia, five showed MCV 61.5 ± 4.6 fL and MCH 17.62 ± 2.25 pg and two showed MCV 78.0 ± 0.0 fL and MCH 25.05 ± 5.16 pg. One β° thalassemia heterozygote with co-inherited α thalassemia showed MCV 76.1 fL and MCH 23.4 pg. Overall, it appears that the co-inheritance of α thalassemia did not normalize MCV or MCH values in borderline HbA 2 β-thalassemia carriers in our study.
The co-existence of δ thalassemia in cis or in trans leading to reduction of HbA 2 levels and a change in typical hematological phenotype of β-thalassemia trait is not uncommon 1,25,27,33 . Co-inheritance of δ thalassemia was noted in 14% β-thalassemia heterozygotes in our study. The defects identified included δ promoter defects −68 C/T and −68 T/T, HbA2 Yialousa, CD 83 G > A/HBD:c.251G > A/ Hb A2 Nishishinbashi, a novel δ globin gene defect CD 46 G > T/HBD: c.140G > T, and a δ globin structural variant; HbA2 Saurashtra 34 . MCV of the β and δ double heterozygotes was lower (Mean ± SD, 69 ± 7.9fL) than that of β-thalassemia carriers without any other defect (Mean ± SD, 73.53 ± 11.8fL) although the difference was not statistically significant (p = 0.13).
India is one of the countries with the highest prevalence of β-thalassemia and anemia and hence it is not uncommon to identify iron deficient β-thalassemia carriers. IDA is associated with a melange of red cell abnormalities coupled with decreases in HbA 2 that may at times lead to misdiagnosis of β-thalassemia carriers 35 . In borderline HbA 2 β thalassemia carriers, the significant iron depletion caused by IDA reduces the already poorly elevated HbA 2 fraction of hemoglobin into the normal range 36 . However, other studies dispute the decrease in HbA 2 levels in the presence of iron deficiency 37,38 . In our study, co-existence of IDA was noted in 10% β-thalassemia heterozygotes with HbA 2 between 3.0 and 3.9%. HbA 2 levels of β thalassemia heterozygotes with co-existing IDA was not different (Mean ± SD, 3.5 ± 0.2%) from that of β-thalassemia carriers without any other defect (Mean ± SD, 3.5 ± 0.2%). However, we noticed that the mean HbA 2 values of the iron deficient β ++ thalassemia heterozygotes were 3.2 ± 0.2% and that of iron deficient β + thalassemia heterozygotes at 3.6 ± 0.2% (p = 0.0459). The co-existence of IDA in β ++ thalassemia heterozygotes appears to decrease HbA 2 levels in our study population.
On assessing our study population on the basis of MCV, we found that 93% individuals with MCV < 80 fL and 76% individuals with MCV > 80 fL were β-thalassemia carriers. In both MCV groups, heterozygosity for β + IVS1-5 G > C and for the β ++ mutations were the most common etiologies associated with borderline HbA 2 levels. A previous study 1 has suggested that with the evaluation of both MCV and HbA 2 , it is possible to differentiate mild mutations from more severe β globin gene defects. Notably, in this study 1 the β 0 and β + alleles were associated with MCV < 80fL while the β ++ alleles were noted in both MCV groups. In our study, heterozygotes for all three types of β thalassaemia alleles (β 0 , β + and β ++ ) could be identified in both MCV groups. We also identified a subset of β + and β ++ thalassemia heterozygotes showing MCV > 80fL with HbA 2 levels < 3.5% (Fig. 3.)  www.nature.com/scientificreports/ population. These atypical carriers can be missed if molecular analyses of the β globin gene is not undertaken in a laboratory setting and may potentially add to the β thalassemic burden of the nation. In our study, associated α thalassemia [MCV < 80fL group 12% versus MCV > 80fL group 6%] and δ thalassemia [MCV < 80fL group 16% versus MCV > 80fL group 6%], were more prevalent in β thalassemia heterozygotes showing MCV < 80 fL and MCH < 27 pg. Our findings suggest that both co-inheritance of α and δ in borderline HbA 2 β thalassemia heterozygotes reduce MCV and MCH.
In this study, we also identified 26 of 205 individuals showing HbA 2 between 3.5 and 3.9% with a normal β genotype, of whom six individuals had MCV < 80 fL (false positives). Three of these six false positive individuals showed presence of IDA, α thalassemia and δ thalassemia while no defects could be identified in the remaining three individuals. Another Indian study 11 also reported borderline HbA 2 levels [3.5-3.9%] in 3.6% individuals with absence of a β-thalassemia defect. The probable etiologies in such individuals have been discussed by Giambona et al. 1 . Individuals with MCV < 80 fL are thought to harbour rare globin gene defects, sequence changes in locus control regions or enhancer regions of the β globin gene while individuals with MCV > 80 fL are speculated to be the result of increased δ globin gene expression, use of antiretroviral drugs, presence of co-morbidities such as hyperthyroidism or genetic disorders such as Pseudoxanthoma Elasticum or defects in genes regulating synthesis of specific protein factors.
Few studies, so far, have delineated the molecular basis of borderline HbA 2 levels in different populations as discussed in a recent review 10 . Results from these studies collectively reveal that the causes of borderline HbA 2 levels may include heterozygosity for β-thalassemia, presence of KLF1 gene mutations, α thalassemia, co-inheritance of β and δ thalassemia, co-inheritance of β and α thalassemia, α globin gene triplication or the presence of hereditary persistence of fetal hemoglobin. In our study, the most common causes of borderline HbA 2 levels were heterozygosity for β-thalassemia, co-inheritance of β and δ-thalassemia, co-inheritance of βand α-thalassemia and the co-existence of IDA in β-thalassemia carriers. A recent study by Hariharan et al. 39 in the Indian population has demonstrated the prevalence of KLF1 gene variations to be 7.6% in individuals with borderline HbA 2 levels [3.3-3.9%]. Although a limitation of the present study remains the non-assessment of KLF1 gene mutations and α gene triplications, we have for the first time delineated the molecular basis of normal/borderline HbA 2 levels in the Indian population. We have described the potential of missing β-thalassemia carriers when HbA 2 levels alone or in combination with MCV are used as stand-alone diagnostic determinants. We strongly recommend a comprehensive molecular work up of the β globin gene in individuals with borderline HbA 2 levels with special attention to cases with a hemoglobinopathy carrier partner. www.nature.com/scientificreports/ Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.