Mutation detection and inhibitor risk in Iranian patients with Hemophilia A: Six novel mutations

Abstract This investigation facilitates a better understanding of inhibitor development, the critical treatment morbidity in HA patients. Furthermore, six novel mutations are reported, which would expand the mutation spectrum of the F8 gene.


| INTRODUCTION
Hemophilia A is known as a bleeding abnormality accompanied with recessive X-linked inheritance. More than 3000 various mutations have been found in the F8 gene. Intron 22 and Intron 1 inversions (Inv22 and Inv1) are the most frequent molecular variations observed in severe HA patients. The present study aimed to identify the F8 gene mutations in Iranian patients with HA and examine their correlation with the development of inhibitors. This research was conducted on 30 Iranian HA patients, including 28 severe and two mild patients. IS-PCR (inverse-shifting PCR) was used to assay the inv22 and inv1 and to screen the other mutations, In Inv22, and Inv1-negative patients, direct sequencing of whole 26 exons was performed by Sanger sequencing method. Finally, multiplex PCR was carried out to detect probable large deletions. Of all the detected mutations, 73.3% mutations (22 of 30) were classified as high-risk mutations, including Inv22 in 12/30 patients (40%), frameshift mutations in 6/30 (23.3%), substitution nonsense in 2/30 (6.7%), and large deletions in 2/30 (6.7%). The analysis revealed six different mutations in the F8 gene, which had not been previously reported based on different HA databases. Inhibitors were observed in 32% of patients with severe HA. No significant difference was found between the mutation risk groups and inhibitor development. We demonstrated that the distribution of the F8 gene mutations and the prevalence of F8 inhibitors in Iranian patients were consistent with those in different populations concerned in previous studies. The lack of correlation between mutation risk and inhibitor incidence can be attributed to the high prevalence of inhibitors in patients with splice site and missense mutations, as documented in some previous studies.
Hemophilia is considered as one of the most serious hemorrhagic disorders and is categorized into hemophilia A (HA), hemophilia B (HB), and hemophilia C (HC). There are three different coagulation factor deficiencies in these categories: factor VIII (FVIII) in HA, factor IX (FIX) in HB, and factor XI (FXI) in HC. 1 HA (OMIM 306700) is defined as a bleeding abnormality with the X-linked recessive inheritance that predominately affects males, with one in 5000 at birth. According to the level of coagulation FVIII, HA is classified into severe <1 IU/dL, moderate 1-5 IU/dL, and mild 5-40 IU/dL. 2 There are several bleeding sites in this disease that vary from life-threatening to non-life-threatening. The fatal hemorrhages generally include organ bleeding, intracranial bleeding, gastrointestinal and genitourinary bleeding, and pharyngeal bleeding. Mucosal bleeding, bleeding into the joints (hemarthrosis), and muscles are classified as non-life-threatening type. Hemarthrosis is one of the usual features of HA. 3 FVIII, also known as antihemophilic factor (AHF), is one of the coagulation factors involved in the coagulation cascade. Activated FVIII (FVIIIa) that functions in the intrinsic tenase complex as the FIXa cofactor increases the catalytic efficiency of FIXa. Consequently, with activating FX, the pathway would convert prothrombin (FII) to thrombin (FIIa) and eventually fibrin to fibrinogen. 4 Different mutations, including large rearrangements, intragenic deletions or insertions, and point mutations in the F8 gene, have been found as the primary cause of the disease. 5 Intron 22 (int22) and Intron 1 (int1) inversions (invs) are proved as the most common mutations in severe HA with the incidence rates of 45% and 5%, respectively. 6 Int22 contains a CPG island that functions as a bidirectional promoter for genes known as factor VIII-associated A (F8A) and B (F8B). Two copies of F8A, also known as int22h-1, are also located in the telomeric side of the gene, including int22h-2 (proximal copy) and int22h-3 (distal copy). There is also an identical copy of int1, known as int1-2, upstream of the F8 gene. 7,8 Homologous recombination between int1-1 and int1-2, int22h-1 with one of its two copies, either int22h-2 or int22h-3, results in inv1 and inv22, respectively. 9 Inv1 produces two-hybrid transcripts, including the promoter and a part of exon 1 of the F8 gene along with alternative exons of the VBP1 gene, and exons of BRCC3 gene, with alternative exons 2-26 of the F8 gene. 10 Inv22, which divides the F8 gene into two distinct regions (exons 1-22 and exons 23-26), produces defective coagulation FVIII due to the lack of a secretory polypeptide chain. 11 Although inv22 type1 is more common than type 2, no clinical discrepancy is observed between the two groups of patients. To date, more than 3000 different mutations have been reported in F8 Gene Variant Databases such as several large deletions or insertions, point mutations including nonsense variations, missense mutations, and mRNA splicing site errors. Moreover, frameshift mutations caused by deletions or insertions may result in severe HA. 12 Coagulation tests fail to identify carrier parents and their affected child in the prenatal diagnosis process. Hence, molecular detection has become a pivotal approach in prenatal diagnosis and also the genetic counseling of patients with a history of this disease. 13,14 Additionally, it is well-known that specific F8 mutations contribute to the development of FVIII inhibitors; hence, the detection of these mutations is essential to manage HA treatment. 15 This study aimed to analyze the F8 gene in Iranian HA patients and examine the correlation between the detected mutations and the inhibitor presence. The identified mutations were categorized into high-risk (namely Inv22, Inv1, large deletions, nonsense, and frameshift mutations) and low-risk (namely missense variants, in-frame deletion/insertions, and splice site mutations) genotypes, consistent with Rodin's study suggesting an association between the F8 genotype and inhibitor development. 16

| Patients
This study was conducted on 30 HA patients referred to our Genetic laboratory of Iranian Hemophilia and Thrombophilia Association (MAHTA). Ethics approvals were gained from all of the patients. The patients' age range varied from three months to 60 years. The clinical severity of the patients had been previously measured using standard 1-stage clotting assay. A majority of patients had severe symptoms (28 severe and two mild patients). They manifested symptoms of the disease at an early age. Many of the severe patients suffered from hemorrhage, spontaneous mucosal bleeding, and bruising. Regarding the mild patients, symptoms were experienced at an older age. The FVIII inhibitor titers in all the patients were quantified using the Nijmegen modification of the Bethesda assay 17 (Table 1).

| Sample collection and DNA extraction
After obtaining informed consent from all the participants following the Declaration of Helsinki, peripheral blood samples were taken from the patients using EDTA tubes, and genomic DNA was extracted utilizing the standard salting-out method. 18

IS-PCR
The patients were first screened for the detection of Inv22 (type I and type II) and Inv1 rearrangements by the use of inverse-shifting PCR (IS-PCR) technique. We used the IS-PCR protocol of Rossetti et al 9 To summarize, genomic DNA was digested by BCL1 enzyme, and digested DNA fragments were separated. Then, DNA fragments were ligated with T4 DNA ligase and purified. Finally, PCR amplification was | 2977 NASIRNEJAD SOLA Et AL.
performed, and the products were run on a 2% agarose gel. The diagnosis was made based on different fragment lengths of the IS-PCR products for the F8 gene.

| Sequencing the whole exons by Sanger sequencing method
The F8 gene was directly sequenced for all 26 exons using the Sanger sequencing method. PCR was performed, and the products were then sequenced on ABI 3500 Genetic analyzer. The miscellaneous lines of in silico computational analysis (MutationTaster, CADD, PolyPhen, VarSome, SIFT, etc) were used to predict the pathogenicity of variations. These variants have also been investigated in HA databases (COSMIC, EAHAD, HGMD) to check previous reports.

| Multiplex PCR
In two patients, when PCR was performed, some exons had not been amplified to be directly sequenced, suggesting a deletion affecting these exons. To determine the breakpoint regions, multiplex PCR was done for the associated exons, as described by Fernandez-Lopez et al. 19 T A B L E 1 Clinical

| RESULTS
This study was performed on 30 Iranian patients with HA from unrelated families. The analysis of these patients revealed various types of mutations in the F8 gene. Across the whole study group, the rate of mutation detection success was 100%. IS-PCR revealed that 12 out of 30 samples (40%) were positive for Inv22 type1 with severe HA. Inv22 type 2 and inv1 were not found in any of the severe or mild patients. Eighteen patients with no Inv22 mutation were then analyzed for other F8 gene variations, and disease-causing variants were detected in 16 patients (53.3%).
The mutations included six frameshifts, five missense substitutions, two nonsense substitutions, two substitutions in the splicing region, and one in-frame deletion in distinct areas of the F8 gene.
The detected c.2013_2015del mutation in the F8 gene has not been previously reported for its pathogenicity. Based on the ACMG guidelines, this variant has been classified as a variant of uncertain significance (VUS); hence, direct sequencing was performed in some family members with or without the disease to confirm the mutation as a causing variant of the disease. The same mutation, c.2013_2015del (p.Phe672del), on the F8 gene, was detected in a heterozygous state in the proband's mother and his affected maternal grandfather hemizygously.
According to multiplex PCR results, large deletions in the exons 8-9 and exon 15 were identified in two other patients. Of all causative mutations identified in this study, 73.3% (22 of 30) were classified as high risk, including 12 invs, six frameshifts, two nonsense mutations, and two large deletions. Table 2 shows a list of all mutations.
Regarding the clinical severity of the 30 patients, 28 cases (93.3%) were severe, and two patients (6.6%) were mild ones.  The frequency of inhibitor positive patients was found to be 30% (9 out of 30). In patients with severe HA, large deletions, nonsense mutations, and splicing errors showed the highest incidence of inhibitors. The lowest inhibitor frequency in severe HA was found in patients with missense and frameshift mutations. Table 3 presents the details of the data.
In the high-risk mutation group, inhibitor positivity was found in 7 out of 22 patients (32%). In the low-risk mutation group, inhibitor development was detected in 2 out of 8 patients (25%). The association between the mutation risk group and inhibitor development was not statistically significant (P > .05). In the patients with Inv22, the frequency of inhibitor positivity was 33.3% (4 out of 12). The association between inv mutations and inhibitor development was not statistically significant (P > .05), as shown in Table 4.

| DISCUSSION
The current study was conducted to investigate the F8 gene mutations in 30 Iranian patients with HA. This research revealed different mutations in the F8 gene, of which six have not been previously reported in the existing mutation register databases. Previous studies have reported inv22 and inv1 as the most common mutations in severe HA. Rossetti et al observed the presence of inv22 in 45%-50% of severe patients. 9 Moreover, in similar reports in Spain, Italy, Argentina, India, Egypt, Mexico, and Iran, inv22 is interpreted as the most frequent cause of severe HA. [20][21][22] It should be noted that we reported the highest prevalence of inv22 mutation (40%), and this strongly supports previous studies. In most reports, the frequency of inv1 mutation has been stated to be 0%-5% of patients with severe HA. 23,24 In our survey, none of the patients showed inv1 mutation. As expected, our findings confirm that invs are the most prevalent mutations in patients with severe HA; therefore, they should be examined as the first diagnostic step. Missense errors are the most common mutations with a frequency of 70%-80% among mild/moderate HA patients. 25 We found missense mutations in both patients with mild HA (S29/S30). However, our findings are based on a limited number of mild patients. This research focused on patients with a severe phenotype, whereas it might be necessary to include more mild/moderate patients as well. It is probably because patients with severe HA are more likely to be medically concerned due to their more serious bleeding phenotype.
The FVIII mutation database consists of more than 3000 unique mutations in hemophiliac patients (EAHAD-DB at dbs.eahad.org). We sequenced 18 patients with negative results for inv mutations and recognized six novel mutations not reported in any F8 gene variant databases. Mutations c.95delC (p.S32Yfs*45), c.3943delA (p.R1315Gfs*20), and c.5509A>T (p.K1837*) are considered to be pathogenic as they introduce premature stop codons, resulting in protein truncation or loss altogether through mRNA nonsense-mediated decay. Besides, we found two mutations in mRNA splicing region, c.1010-2A>G and c.1444-1G>T, in exons 8 and 10, respectively. The detected canonical splice site variants in the F8 gene have not been previously reported for their pathogenicity. On the other hand, null variants, including canonical splice site errors affecting the F8 gene, are a known mechanism of the disease, and multiple lines of in silico computational analysis (MutationTaster, CADD, etc) support the deleterious effect of the variants on the gene or gene product.
According to the data reported worldwide, severe HA patients have a higher prevalence of inhibitors than mild/moderate deficiency, as observed in 25%-30% of severe patients and approximately 3%-13% of patients with mild/moderate HA. 26 In a study, the incidence of inhibitor development in 635 Iranian patients with severe HA was reported to be 22.8%. 27 In another review study, 100 out of 355 patients (28%) from six main hemophilia care centers in Iran were positive for the FVIII inhibitor. 28 Our study showed that 32% of patients with severe HA had factor VIII inhibitors. This finding was almost similar to those reported in previous studies.
In many investigations, the relationship between F8 genotypes and inhibitor development varies noticeably; however, most reviews confirm that the inhibitor incidence is correlated with invs, large deletions, and nonsense mutations in severe HA. 29 A survey in Germany showed a higher incidence of inhibitor in patients with large deletions, nonsense, and inv22 mutations, compared to patients with small deletions or insertions, missense, and splicing errors. 30 In a large cohort of US patients with severe HA, the frequency rates of inhibitors were 57.1%, 35.7%, and 26.8% in large deletions, splice sites, and inv22, respectively. 31 In another study, 1104 cases with HA from three different countries, including Iran, France, and the Netherlands, were analyzed. In this study, in contrast to previous studies, in patients with severe HA, splicing errors had the highest frequency of inhibitors, and the lowest inhibitor prevalence was observed in patients with missense mutations. 32 Generally, in most studies, the inhibitor frequency ranges in severe HA patients are 21%-27%, 25%-40%, 10%-16%, and 5%-10% for inv22, nonsense mutations, frameshift mutations, and missense variations, respectively. For other genotypes, the results are more different, ranging from 22% to 67% and 17% to 50% for large deletions, and splice site changes, respectively. 31 This study showed the equal inhibitor incidence in patients with large deletions, nonsense mutations, and splice site errors, which were higher than those in cases with inv22. Contrary to most of the previous studies, we observed a higher range of inhibitor frequency in patients with missense mutations and splice site changes. Hence, concerning the classification of mutations into low-and high-risk groups, the incidence of inhibitor development in low-risk mutations was close to those in the high-risk group. In a study in Turkey, the examination of 270 patients with HA revealed a significant correlation between high-risk genotypes and inhibitor formation. In our study, however, no significant difference was found between the mutation risk group and inhibitor development. 33 The lack of correlation in the present study can be attributed to the high prevalence of inhibitors in patients with splice errors, as documented in one previous study. Furthermore, missense variations are associated with an overall lower risk of inhibitor development; however, different missense mutations are correlated with specific risks for inhibitor formation. For example, the likelihood of inhibitor formation in association with the F8 missense mutations is significantly higher if the amino acid substitution is related to another physicochemical class than the original residue. Another possible reason for this is that we cannot rule out that positive family history of inhibitors, ethnicity, and other genetic factors, including certain polymorphisms in immune-modulatory genes, might have influenced the risk of inhibitor development. Interestingly, F8 genotypes are believed to be equally distributed across different ethnic populations; however, this may not be true for other genetic risk factors. 34 One of the limitations in this research was that the surveys included a small number of patients; nevertheless, the assessment of the inhibitor risk caused by genetic factors would require the collection of data on much larger numbers of patients using a population-based approach. Despite this, we can still claim a close agreement with the findings of previous publications in many aspects, including the prevalence of mutations, overall inhibitor incidence, and its distribution in each of the mutations, and this confirms that our findings are appreciable. We recommend that further studies such as meta-analysis studies can yield more precise estimates of the inhibitor risk for different types of F8 mutations.

| CONCLUSION
A high mutation detection rate was achieved using the broad molecular techniques performed in this study, which included six novel mutations. The results provided additional support for the distribution rate of F8 gene mutations. A high range of inhibitor frequency was observed in patients with missense and splice site mutations. The prevalence of F8 inhibitors was close to those observed in previous reviews. However, the relationship between F8 genotypes and inhibitor development was not significant.