Coexpression of factor VIII and factor von Willebrand variants in a woman with heavy menstrual bleeding

Abstract

 

Heavy menstrual bleeding is one of the most common causes of consultation in haematology. We present the clinical case of a 20-year-old woman referred by her gynaecologist due to heavy menstrual bleeding since menarche, complicated by iron deficiency anaemia. Haemostasis work-up was initially suggestive of a von Willebrand disease type 1. Genetic analyses by whole exome sequencing lead to a fortuitous discovery of haemophilia by identifying a heterozygous missense mutation in F8, exon 8 c.1127T>G:p.Val376Gly, previously reported in a patient with mild haemophilia A. The bleeding phenotype worsened by concomitant low von Willebrand factor (VWF) due to VWF variants influencing VWF levels. Our case highlights how whole exome sequencing can help to correct an erroneous diagnosis and identify polymorphisms that eventually contribute to the overall haemostatic balance.

Introduction

Haemophilia in women can be due to an inherited heterozygous variant on a single X chromosome. Lyonization, the inactivation of one of the X chromosomes early in embryogenesis, is in principle random but can result in low plasma factor VIII (FVIII) in haemophilia A carriers resulting in woman with mild, moderate or severe haemophilia. Compared to boys, the age at haemophilia diagnosis is delayed in girls ^[1]. The late diagnosis is particularly deleterious for appropriate anticipation of menarche. Indeed, heavy menstrual bleeding (HMB) is one of the most frequent bleeding experiences in haemophilia A carriers, resulting in a poor health-related quality of life ^[1]. A major challenge in the diagnosis and management of haemophilia A carriers is the relatively weak association between the baseline factor level and the bleeding phenotype. It is likely that other features beyond the baseline factor, such as the FVIII/FIX genotype or the coinheritance of another bleeding-associated variant participate to the bleeding severity in haemophilia A carriers. Here, we present the rare clinical case of a woman with mild haemophilia A with a bleeding phenotype worsened by concomitant low von Willebrand factor (VWF) due to VWF variants influencing VWF levels. Our case highlights how whole exome sequencing can help to correct an erroneous diagnosis and identify polymorphisms that eventually contribute to the overall haemostatic balance.

Methods

Informed consent was obtained from our patient. A 20-year-old woman was referred by her gynaecologist due to HMB since menarche, complicated by iron deficiency anaemia. Her periods were regular and last for more than 10 days with a sensation of flooding. Her pictorial blood assessment with the Janssen score was of 790 points (normal range <185) with bleeding through sanitary protection in 2 h, large blood clots, soaking through bedclothes and change during the night. She also reported frequent gum bleeding lasting for more than 10 min, abnormal bleeding occurring after teeth extraction and recurrent nosebleeds requiring medical consultation. The bleeding score evaluated by the International Society of Thrombosis and Haemostasis (ISTH) bleeding assessment tool (BAT) was of eight points (normal range for women less than six points). The family history revealed that her maternal grandfather was possibly haemophiliac, her mother was asymptomatic. Pelvic ultrasonogram was normal. The laboratory work-up is summarized in Table 1.

Table 1 - Biological data

Hemostasis work-up	Results	Normal reference range
PT (%)	77	>70
aPTT (s)	40.8	26–37
Fibrinogen (Clauss) (g/l)	1.6	2–4
Fibrinogen antigen (g/l)	1.8	2–4
Thrombin time (s)	24.2	22–28
Reptilase time (s)	20.2	18–22
Factor VIII (%)	28	50–100
Factor IX (%)	73.8	70–130
Factor XI (%)	72	>50
Factor XII (%)	77	>70%
Factor XIII (%)	105.3	75–130
VWF : GPIbR (%)	37	50–150
VWF : Ag (%)	39	50–150
VWF : CB (%)	37	50–150
VWF : GPIbR/VWF : Ag	0.9	NA
VWF : CB/VWF : Ag	0.9	NA
Factor VIII/VWF : Ag	0.7	NA
Platelet aggregometry
Arachidonic acid 1 mmol/l (%)	85	>84
ADP 2 μmol/l (%)	82	>30
Adrenaline 10 μmol/l (%)	83	>83
Collagen 2 μmol/l (%)	87	>85
Ristocetin 0.5 mg/ml (%)	4	<12
TRAP 10 μmol/l (%)	88	>84
U46619 1 μmol/l (%)	87	>82

Ag, antigen; aPTT, activated partial thromboplastin time; CB, collagen binding; GPIbR, recombinant GPIb; NA, not applicable; PT, prothrombin time; TRAP, thrombin receptor activating peptide; VWF, von Willebrand factor.

Results

Results were initially suggestive of a von Willebrand disease (VWD) with normal FVIII : C/VWF : Ag ratio. The patient's blood group was O. Multiple testing showed persistent low FVIII and VWF activities values, ranging from 25 to 36% and 37 to 46%, respectively. The PFA with ADP was 136 s (range 66–129) and with EPI was 169 s (range 86–189). Fibrinogen activity (Clauss) and antigen were at the lower range of reference, probably due to consumption in the setting of chronic bleeding as suggested by decreased levels of ferritin despite oral substitution. Successive fibrinogen measurement was in the normal range. A significant and sustained increase in plasma FVIII and VWF activity was observed following DDAVP administration (1 h after injection, FVIII 173% and VWF 185%, ratio 0.9; 4 h after injections FVIII 124% and VWF 155%, ratio 0.8).

As indicated in Table 2, genetic analyses by whole exome sequencing lead to a fortuitous discovery of haemophilia by identifying a heterozygous missense mutation in F8, exon 8 c.1127T>G:p.Val376Gly, previously reported in a patient with mild haemophilia A ^[2]. Another heterozygous missense mutation in F8 was also identified, the exon 14 c.3780C>G:p.Asp1260Glu polymorphism which despite previous reports associating it with haemophilia A is now considered benign. In addition, we detected five variants in VWF, three reported as polymorphisms according to the EAHAD Coagulation Factor Variant Databases (https://dbs.eahad.org) and two rare missense mutations in exon21 c.2771G>A:p.Arg924Gln and exon 37 c.6532G>T:p.Ala2178Ser. Finally, four variants were detected in F5, all considered benign by computation predictions (SIFT and PolyPhen-2a) and a well-known variant in F13A1 (exon 2 c.103G>T:p.Val34Leu) ^[3]. Of note, no variants were detected in fibrinogen genes.

Table 2 - Genetics data

Gene	Sequence	Status	Reference SNP ID	Current ClinVar interpretationa	Population frequency totalb
F8	exon8 c.1127T>G:p.Val376Gly	Heterozygous	–	–	–
F8	exon14 c.3780C>G:p.Asp1260Glu	Heterozygous	rs1800291	Benign	0.1958
VWF	exon13 c.1415A>G:p.His484Arg	Heterozygous	rs1800378	Benign/Likely benign	0.6106
VWF	exon18 c.2365A>G:p.Thr789Ala	Heterozygous	rs1063856	Benign/Likely benign	0.3224
VWF	exon21 c.2771G>A:p.Arg924Gln	Heterozygous	rs33978901	Conflicting interpretations of pathogenicity; VUS(3), Benign(4), Likely benign(1)	0.01067
VWF	exon28 c.4141A>G:p.Thr1381Ala	Heterozygous	rs216311	Benign/Likely benign	0.6862
VWF	exon37:c.6532G>T:p.Ala2178Ser	Heterozygous	rs34230288	Conflicting interpretations of pathogenicity; VUS(1), Likely benign(1)	0.01148
F5	exon13:c.2773A>G:p.Lys925Glu	Heterozygous	rs6032	Benign/Likely benign	0.2735
F5	exon13:c.2594A>G:p.His865Arg	Heterozygous	rs4525	Benign/Likely benign	0.2735
F5	exon13:c.2573A>G:p.Lys858Arg	Heterozygous	rs4524	Conflicting interpretations of pathogenicity; VUS (1), Benign(2), Likely benign(1)	0.2759
F5	exon16:c.5290A>G:p.Met1764Val	Heterozygous	rs6030	Conflicting interpretations of pathogenicity; VUS(1); Benign(2), Likely benign(1)	0.3371
F13A1	exon2:c.103G>T:p.Val35Leu	Heterozygous	rs5985	Conflicting interpretations of pathogenicity; VUS(1), Benign(3)	0.2062

In bold variants considered pathogenic or likely pathogenic in this study.

^aCurrent ClinVar Interpretation dated 04/01/2023; Variant of Unknown Significance (VUS), numeric value in brackets following conflicting interpretations i.e. VUS(1) corresponds to the count of submitted interpretations.

^bAllele Frequency according to gnomAD v2.1.1.

Initially, HMB were unsuccessfully treated by a cyclic combined oral contraceptive (COC) (ethynyl estradiol 20 μg–levonorgestrel 100 μg), then by a progestin-only regimen (desogestrel 75 μg) and finally by another COC (ethynyl estradiol 30 μg–levonorgestrel 150 μg). All these treatments did not significantly improve the bleeding symptoms. The introduction of tranexamic acid (1 g three times per day) did not provide any improvement neither. Subcutaneous injections of desmopressin (0.3 μg/kg) at the first 2 days of menses resulted in a partial but transient response on menses intensity. Lastly, the introduction of an intrauterine device releasing levonorgestel (14 μg/jour) with concomitant transdermal oestrogen (oestradiol 100 μg/24 h twice a week) finally allowed a significant decrease of the severity of menstrual blood loss. Gum bleeding was successfully treated with local application of tranexamic acid.

Discussion

A wide range in clotting factor levels is observed in haemophilia A carriers. The phenotypic expression of low FVIII in haemophilia A carriers is partially explained by lyonization. In a recent study, including 73 haemophilia A carriers, a significant association was observed between a nonrandom chromosome X inactivation pattern, the FVIII or FIX levels and the bleeding phenotype assessed by the ISTH-BAT ^[4]. However, other genetic modifiers can affect the coagulation factors levels and more importantly the bleeding phenotype. In the large Global Emerging Hemostasis Panel Study, no difference was identified between bleeding scores or factor levels when comparing haemophilia A carriers with null mutations and those with nonnull mutations ^[5]. Substantial variability in bleeding phenotype has even been described among haemophilia A carriers cases with identical causal variants. Coinheritance of variants in VWF in haemophilia A carriers, as reported in this case report, could be considered as a rare cause of bleeding variability in this patient population. In fact, when inherited in cis the two identified heterozygous missense mutations in VWF, exon 21 c.2771G>A:p.Arg924Gln and exon 37 c.6532G>T:p.Ala2178Ser, have been linked to a type 2B-like VWD phenotype in a prior case ^[6]. Functional testing in the previous study confirmed the causative effect of these heterozygous variants when in cis and identified that they can promote conformational transitions in the VWF molecule. Nevertheless, in the case presented here no family members were available for testing and it remains unknown if the two VWF variants are inherited in cis thus limiting any significant comparisons between the cases being made.

The pathogenicity of exon21 c.2771G>A:p.Arg924Gln alone is contentious and highlighted by the conflicting interpretations on ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/) and the EAHAD Coagulation Factor Variant Database. The Arg924 amino acid is well conserved and has been implicated in the interaction between VWF and FVIII. The substitution of arginine to glutamine will result in a change of polarity and charge which in modelling, undertaken elsewhere, has shown this will cause a partial change in the proteins structure ^[6]. This is however not reflected in the in-silico predictions of PolyPhen-2a and SIFT which show the variant respectively as benign tolerated. In 2010, 1115 healthy controls and 148 index cases from the MCMDM-1VWD study were genotyped for c.2771G>A. Results showed that the allele bearing c.2771A, identified in six index cases and 35 controls, was associated with lower VWF and FVIII levels, particularly in combination with blood group O⁷. It most probable that the inheritance alone of this variant is insufficient for a VWD diagnosis but that the variant is in linkage disequilibrium with another allele causing the reported effect or that a second VWF mutation is required. Historic cases reporting Arg924Gln as the only identifiable causal variant in VWD patients may therefore have missed the true cause, in part due to a lack of sensitivity and selectivity of the genetic testing used at the time ^[7].

Reports of the exon37:c.6532G>T:p.Ala2178Ser variant in VWF on both ClinVar and the EAHAD Coagulation Factor Variant Database indicate only conflicting interpretations of pathogenicity with no supporting evidence. Again, the amino acid (Ala217) is well conserved and the substitution to serine also results in a change in polarity. The missense mutation is predicted in modelling by Sacco et al. to cause a partial conformational change to the proteins structure. Nonetheless the variant is predicted by PolyPhen-2a and SIFT to be benign and tolerated. For sufficient independent interpretation of this variant, supplementary functional and/or segregation data is required.

The low incidence of haemophilia A and VWD in the general population means that the combined deficiency is an uncommon condition. Very few cases of genetically identified combined haemophilia A and type 1 vWD have been reported. Daidone et al.^[8] described an Italian family in which the VWF allele significantly exacerbates the haemorrhagic complications in patients with mild haemophilia A but with only mild effects on haemophilia A carriers. A cross-sectional study from 28 haemophilia comprehensive treatment centres in the United States between 1988 and 2002 identified eight patients with concomitant haemophilia and VWD, including three haemophilia A carriers women with mild bleeding phenotype ^[9]. The impact of other variants identified in our case is purely speculative. The F5 missense variants rs6032, rs4524 and rs4525, reported in several studies as being part of the so-called G allele where adenine is replaced by guanine at certain nucleotide positions in the F5 gene, could be associated with a decreased risk of thrombosis. This protective effect could be mediated by reduced FVIII levels and reduced activated protein C resistance ^[10]. The FXIIIA1 rs5985 is associated with altered blood clot structure which could influence haemorrhagic events.

In conclusion, we report a rare case of coexpression of FVIII and VWF variants in a woman with a bleeding phenotype. Based on recent definitions from the ISTH, this patient should be classified as having mild haemophilia A, associated with VWF variants not clearly pathogenic but most likely influencing VWF levels. In this case the serendipitous identification of a pathogenic FVIII variant will aid the patients genetic counselling. Haematologists should consider the possibility of overlapping diagnoses. It is likely that the increasing use of next-generation sequencing will lead to the identification of many more combined haemostatic defects and genetic modifiers allowing a more individualized diagnosis and treatment.

Acknowledgements

Author contributions: A.Ca., A.Co. and M.N.-A. contributed to the acquisition of data. A.Ca. wrote the article. M.Y., P.F. and A.Ca. were in care of the patient. M.N.-A. and A.Co. performed genetic analysis. All authors performed critical review of the article. All authors approved the final version of the article.

Conflicts of interest

There are no conflicts of interest.