Identification of novel mutations in patients with fibrinogen disorders and genotype/phenotype correlations

Abstract

Background

Congenital fibrinogen disorders are caused by variants occurring within the fibrinogen gene cluster. We describe ten subjects with disease-causative variants, adding information on such disorders.

Materials and methods

Ten subjects were referred to our Centre because of likely hypo/dysfibrinogenaemia. We evaluated the function and quantity of fibrinogen, using Clauss and immunoreactive assays, and performed genetic investigations by direct sequencing of alpha, beta and gamma chain-encoding genes. Mutations were analysed using SIFT and Polyphen-2 algorithms.

Results

We identified one afibrinogenaemic patient (alpha p.Arg178* homozygote) with bleeding/thrombotic events, three heterozygous patients with hypo/dysfibrinogenaemia (gamma p.Thr47ILeu combined with beta IVS7+1G>T; beta p.Cys95Ser; beta p.Arg196Cys) referred for bleeding or thrombotic episodes and six heterozygous subjects with hypofibrinogenaemia (alpha p.Glu41Lys; gamma p.Gly191Val; beta p.Gly288Ser; gamma p.His333Arg; gamma p.Asp342Glu and p.343–344 duplication; gamma p.Asp356Val), of whom four were symptomatic. Five novel missense changes and one novel duplication variant were found, all in hypofibrinogenaemic subjects: p.Glu41Lys (SIFT score 0, Polyphen-2 score 0.986) was identified in a woman with bleeding after major orthopaedic surgery; p.Gly191Val (SIFT score 0.02, Polyphen-2 score 1) in an asymptomatic woman; p.His333Arg (SIFT score 0, Polyphen-2 score 1) in a woman with a post-partum haemorrhage; and p.Asp342Glu (SIFT score 0.23, Polyphen-2 score 0.931); and an Asn-343 and Asp-344 duplication in a child who developed a haematoma following a fall.

Discussion

All but one of the novel mutations were in symptomatic subjects and are predicted to be deleterious. Our findings shed more light on genotype-phenotype relationships in congenital fibrinogen disorders.

Introduction

Fibrinogen is a complex protein of 340 kDa composed of three non-identical chains - alpha, beta, and gamma - that are encoded by three paralogous genes (fibrinogen alpha chain [FGA], fibrinogen beta chain [FGB], and fibrinogen gamma chain [FGG] genes), which are clustered in a 50 kb region on chromosome 4 (4q31.3–q32.1) (UCSC Genome browser; http://genome.ucsc.edu/; GRCh38/hg38 human genome assembly, December 2013). The fibrinogen gene cluster guides fibrinogen transcription, assembly, domain stability and secretion¹. Genetic variants within all three fibrinogen genes can cause quantitative or qualitative congenital disorders, which are rare inherited recessive or dominant traits^2,3. The quantitative disorders are classified as afibrinogenaemia or hypofibrinogenaemia, while the qualitative ones are termed dysfibrinogenaemia and hypo-dysfibrinogenaemia³. Congenital fibrinogen disorders are mostly due to FGA and FGG nucleotide variations. Overall, 141 variants have been reported to cause afibrinogenaemia or hypofibrinogenaemia, whereas 78 variants have been described so far in dysfibrinogenaemic patients (see HGMD^® [Human Gene Mutation Database] professional, http://www.biobase-international.com/product/hgmd, accessed 15^th May 2018, QIAGEN Bioinformatics, Redwood City, CA, USA).

A diagnosis of afibrinogenemia is suggested by the complete absence of functional and immunoreactive fibrinogen, while that of hypofibrinogenaemia is suggested by a combined reduction of functional and immunoreactive fibrinogen. The prevalence of afibrinogenaemia in the general population is 1:1,000,000¹. Dysfibrinogenaemia is suggested by reduced functional fibrinogen coupled with normal immunoreactive fibrinogen¹.

Patients with afibrinogenaemia may present with haemorrhagic events, which can range from mild to severe^1,3, while hypofibrinogaemic and hypo-dysfibrinogaemic subjects may bleed following trauma. However, fibrinogen disorders may also be a cause of thromboembolic complications or go undiagnosed¹.

Afibrinogenaemia and hypofibrinogenaemia are usually associated with the same genetic variants, which influence the quantity of fibrinogen^1,3.

Here we report on genotype-phenotype correlations in a cohort of ten subjects with documented quantitative and/or qualitative fibrinogen disorders. Mutational screening led to the identification of four novel missense variants and one novel duplication.

Materials and methods

Ethical standards

The study complied with the principles of the Declaration of Helsinki and was approved by the Ethics Committee of the I.R.C.C.S. “Casa Sollievo della Sofferenza”. Written informed consent to participation in the study was obtained from each subject.

Patients

From 2011 to 2016, ten subjects (9 females, 1 male) were referred to our Centre because of a likely diagnosis of congenital fibrinogen deficiency. All subjects are from southern Italy. Their clinical histories were collected using a specific questionnaire focusing on any bleeding or thrombotic episodes, as well as an obstetric history in women. Congenital fibrinogen disorders were classified as quantitative or qualitative according to whether the Clauss/antigen ratio was above or below 0.7, respectively⁴. Bleeding was evaluated using a bleeding assessment tool (BAT) developed by the Scientific and Standardization Committee Joint Working Group of the International Society on Thrombosis and Haemostasis⁵.

Plasma- and DNA-based investigations

Plasma- and DNA-based investigations were performed on samples obtained from peripheral whole blood. Blood specimens were collected into BD Vacutainer^® Citrate Tubes (Becton Dickinson, Franklin Lakes, NJ, USA) with 3.8% buffered sodium citrate. Plasma samples were prepared by centrifugation at 3,000×g for 10 min and stored at −80 °C until use. DNA was obtained as previously described⁶.

Coagulation tests

Plasma fibrinogen was assayed functionally by the Clauss clotting method, as previously described⁷. Both the reagents and the coagulometer (ACL TOP 300) were from Instrumentation Laboratory Haemostasis systems (Instrumentation Laboratory Company, Milan, Italy). The normal range of plasma fibrinogen concentrations was 1.5–4.0 g/L. Plasma fibrinogen antigen was evaluated by means of a NOR Partigen Fibrinogen kit (Dade Behring, Marburg, Germany), with normal reference values of 1.82–3.39 g/L.

Mutational screening

Polymerase chain reaction analysis was performed according to standard protocols⁸. All coding regions of fibrinogen chain genes and intron/exon boundaries were amplified using sense and antisense oligonucleotides designed on the basis of known sequences of fibrinogen gene loci (Genbank accession numbers M64982, M64983, and M10014).

The amplified fragments underwent direct cycle sequencing using a BigDyeTerminator v3.1 Cycle Sequencing kit and the ABI PRISM 3130 Genetic Analyzer Sequencer (PE Biosystems, Foster City, CA, USA).

In silico predictions

Multiple alignments of alpha and gamma chains protein sequences were generated by the computer programme MUSCLE (version 3.6)⁹ on the HomoloGene automated system (http://www.ncbi.nlm.nih.gov/homologene). The damaging effects of missense mutations were predicted using the web-based algorithms SIFT (Sorting Intolerant from Tolerant, http://sift.bii.a-star.edu.sg/)¹⁰ and Polyphen-2 (Polymorphism Phenotyping v2, http://genetics.bwh.harvard.edu/pph2/)¹¹. Residue changes were classified as probably damaging in the presence of a SIFT score ranging from 0.00 to 0.05 or a Polyphen-2 score ranging from 0 to 1, with a score of 1 indicating a probably damaging residue. Protein modelling was performed by means of the Swiss-PdBViewer application v4.1.

Results

The clinical and biological data of the ten investigated subjects are summarised in Table I. One was diagnosed with afibrinogenaemia, five with hypofibrinogenaemia, two with hypo-dysfibrinogenaemia and two with dysfibrinogenaemia. The symptomatic individuals had quantitative or qualitative disorders.

Table I.

Clinical, biological and genetic data of the patients.

Patient reference	Gender	Age (years)	Functional fibrinogen (g/L) Reference range 1.50–4.00 g/L	Fibrinogen antigen (g/L) Reference range 1.82–3.39 g/L	Clauss/antigen ratio	Phenotype	Genotype	Gene region	Nucleotide	Variant numbering (literature-based)	Variant numbering (HGVS nomenclature)	Clinical features	BAT score
P1	Female	37	Undetectable	Undetectable	Not applicable	Afibrinogenaemia	Homozygous	FGA exon 5	c.532C>T	Arg159Stop	p.Arg178Ter	Haemorrhagic and thrombotic disorders	4
P2	Female	23	1.25	1.71	0.73	Hypofibrinogenaemia	Heterozygous	FGA exon 2	c.121G>A	New	p.Glu41Lys	Bleeding in the knee after surgery	3
P3	Female	31	1.20	1.98	0.60	Dysfibrinogenaemia	Heterozygous	FGB exon 4	c.586C>T	Arg166Cys	p.Arg196Cys	Pregnancy loss	0
P4	Female	43	1.00	1.75	0.57	Hypo-dysfibrinogenaemia	Heterozygous	FGB intron 7	IVS7 +1G>T	IVS7 +1G>T	IVS7 +1G>T	Pregnancy-related thrombosis	0
							Heterozygous	FGG exon 3	c.140C>T	Thr21Ile	p.Thr47Ile
P5	Female	32	0.92	1.20	0.76	Hypofibrinogenaemia	Heterozygous	FGB exon 6	c.682G>A	p.Gly288Ser	p.Gly288Ser	Asymptomatic	0
P6	Male	40	0.35	1.4	0.25	Hypo-dysfibrinogenaemia	Heterozygous	FGB exon 2	c.284G>C	p.Cys95Ser	p.Cys95Ser	Asymptomatic	0
P7	Female	22	1.19	1.3	0.91	Hypofibrinogenaemia	Heterozygous	FGG exon 6	c.572G>T	New	p.Gly191Val	Asymptomatic	0
P8	Female	38	1.165	0.925	1.26	Hypofibrinogenaemia	Heterozygous	FGG exon 8	c.998A>G	New	p.His333Arg	Post-partum haemorrhage	4
P9	Female	50	0.88	2.5	0.35	Dysfibrinogenaemia	Heterozygous	FGG exon 8	c.1067A>T	Asp330Val	p.Asp356Val	Disseminated intravascular coagulation	1
P10	Female	7	1.12	0.48	2.33	Hypofibrinogenaemia	Heterozygous	FGG exon 8	c.1026C>A	New	p.Asp342Glu	Haematoma	1
									c.1027–1032 duplication	New	p.Asn343-Asp344 duplication

HGVS: Human Genome Variation Society; BAT: bleeding assessment tool.

Four novel missense changes and one novel duplication variant were identified: one in the alpha chain, the remaining in the gamma chain. We also identified seven genetic variants (5 missense, 1 frameshift and 1 nonsense) already described in the Human Gene Mutation Database (HGMD^®, QIAGEN Bioinformatics). The p.Cys95Ser missense change in the fibrinogen gamma chain is not reported in the above-mentioned database because it was only recently described¹².

Afibrinogenaemia

Afibrinogenaemia was diagnosed in subject P1, whose fibrinogen levels were undetectable in qualitative and quantitative assays (Table I). The patient experienced epistaxis, haematomas, and menorrhagia and was treated with prophylactic doses of fibrinogen. At the age of 34 she suffered from an intracerebral haemorrhage in the right temporal and parietal lobes complicated by thrombosis in the right transverse sinus and left internal jugular vein. The molecular investigation revealed that she had a homozygous C-to-T transition at codon 178 in the fibrinogen alpha chain, responsible for the p.Arg178* variant, which causes a truncated protein.

Dysfibrinogenaemia

Dysfibrinogenaemia was diagnosed in subjects P3 and P9. Both had normal levels of fibrinogen but abnormal function of the protein (Table I). P3 had had two miscarriages and P9 a haematocolpos after a vaginal delivery complicated by disseminated intravascular coagulation. Both the patients were heterozygous carriers for previously described variants. More in detail, P3 had a C-to-T transition within FGB, responsible for the p.Arg196Cys missense change, whereas P9 carried an A-to-T transversion within FGG, causing the p.Asp356Val missense change.

Hypo-dysfibrinogenemia

Hypo-dysfibrinogenemia was observed in subjects P4 and P6. Both had previously described genetic variants (Table I). P4 experienced a pregnancy-related venous thrombosis. Molecular investigation revealed that she was heterozygous for the IVS7+1G>T frameshift variant in FGB and for the FGG c.140C>T nucleotide variation, which is responsible for the p.Thr47Ile. Her mother was asymptomatic (Clauss: 0.69 g/L; antigen: 0.84 g/L) and carried the IVS7 +1G>T variant. No data were available for her father, whereas a paternal aunt carried the p.Thr47Ile with normal fibrinogen levels (Clauss: 2.27 g/L [normal values: 1.5–4.0 g/L]; antigen: 3.52 g/L [normal values 1.82–3.39 g/L]). P6 was an asymptomatic male and carried the p.Cys95Ser missense change in the fibrinogen beta chain.

Hypofibrinogenaemia

Five subjects were diagnosed with hypofibrinogenaemia. They showed low functional and antigen levels (Table I). Three of them were symptomatic, manifesting mild to severe bleeding, which occurred after a vaginal delivery, after knee surgery (for an intra-articular venous malformation), and after a fall. Overall, six variants were identified, of which five were missense changes and one was a duplication variant. To the best of our knowledge, five of these six variants were novel: one was in the alpha chain, one in the beta chain and the remaining were in the gamma chain.

Of note, one novel missense change and the novel duplication variant were observed in cis in patient P10.

Characterization of the novel missense variants

p.Glu41Lys

The p.Glu41Lys change was found in a hypofibrinogenaemic, symptomatic 23-year old woman (patient P2). Multiple alignments of alpha chain sequences showed that glutamate-41 is a highly conserved residue (Figure 1A). According to the prediction algorithms, the p.Glu41Lys missense change is likely to have a damaging effect (Table II). The glutamate-41 residue is located at the N-terminal end of the alpha chain close to the cysteine-47 residue, which is in a disulphide ring, acting to ensure stabilisation of the fibrinogen. At protein position 41, the p.Glu41Lys change leads to a substitution of a negatively charged residue (glutamate) by a positively charged one (lysine), which also has a longer side chain. Several mutations have been reported to affect the N-terminal end polymerization site of the fibrinogen alpha chain¹³. We hypothesise that differences between glutamic acid and lysine at a crucial protein site¹⁴, in terms of chemical properties, could affect the stabilisation and polymerisation of fibrinogen molecules.

Figure 1.

Multiple alignment of (A) FGA and (B, C) FGG protein sequences generated by MUSCLE version 3.6 (using option: -maxiters 2) from the HomoloGene automated system.

(A) Glu41 (according to H. sapiens FGA sequence numbering) is shown in the narrow box. (B) Gly191, (C) His333 and Asp342 (according to H. sapiens FGG sequence numbering) are shown in the narrow boxes. NP_000499.1: H. sapiens; XP_001136067.1: P. troglydytes; XP_001088855.2: M. mulatta; XP_532697.2: C. lupus; NP_001104518.1: M. musculus; NP_001008724.1: R. norvegicus; NP_001258840.1: G. gallus; NP_001181918.1: D. rerio; XP_002933535.2: X. tropicalis).

Table II.

Polyphen-2 and SIFT predictions.

Gene	Mutation	Polyphen-2			SIFT
		Score	Sensitivity	Specificity	Score	Median information*
FGA	p.Glu41Lys	0.986	0.74	0.96	0	2.70
FGG	p.Gly191Val	1	0	1	0.02	2.57
FGG	p.His333Arg	1	0.00	1.00	0	2.57
FGG	p.Asp342Glu	0.923	0.81	0.94	0.23	3.08

SIFT: sorting intolerant from tolerant algorithm;

^*warning about predictions when the median information value is greater than 3.25.

p.Gly191Val

Glycine-191 is a highly conserved residue across the species considered (Figure 1B). Thus, the p.Gly191Val could affect fibrinogen gamma chain structure. This hypothesis is justified by in silico predictions, which suggested a probably damaging effect due to the p.Gly191Val missense change (Table II). Glycine-191 initiates a β-sheet structure in the gamma chain (Figure 2A) and is generally a peculiar residue for the overall conformational features of proteins. Indeed, a Ramachandran plot displays glycine-191 in the so-called “disallowed regions” in the gamma chain (Figure 2B). Moreover, it has been shown previously that glycine-191 may be a key gamma chain residue because of its role in the folding pattern of the protein C-terminal globular domain¹⁵.

Figure 2.

The glycine-191 residues in the fibrinogen gamma chain.

(A) Model-template alignment using the Swiss Model Server. The glycine-191, in the circle, is in a β-sheet structure indicated with an arrow. (B) A representative Ramachandran plot shows glycine-191 (small square at the bottom right) outside the so-called “allowed regions”.

p.His333Arg

This mutation was identified in a hypofibrinogenaemic woman (P8) who experienced postpartum hemorrhage after her first pregnancy. Histidine-333 is a highly conserved residue in the multiple alignment of gamma chain protein sequences (Figure 1C). Based on computational algorithmic predictions, the change may have a structurally relevant effect (Table II). Histidine has peculiar chemical properties and for this reason, proteins barely tolerate substitution with other residues. Histidine at position 333 is located within the globular D domain of the gamma chain and close to the D:D interface¹⁶ (Figure 3).

Figure 3.

Fibrinogen gamma histidine-333 is close to the D:D fibrinogen interface, contributing to generate H-bonds in this area.

The image was generated with Swiss-PdbViewer v4.1 using the Protein Data Bank file 1FZF.PDB.

p.Asp342Glu and the duplication variant

The p.Asp342Glu and duplication variants are due to nucleotide changes within FGG exon 8. Multiple alignments revealed that aspartate-342 is a highly conserved amino acid (Figure 1C). The duplication variant involves AATGAC nucleotides, which code for Asn-343 and Asp-344 residues. The p.Asp342Glu is predicted to be a likely deleterious variant according to Polyphen-2, but not according to SIFT (Table II). Protein modelling by means of the Swiss-Pdbviewer application showed that the variant causes the disappearance of two H-bonds and a new interface generated contains a new H-bond (Figure 4). Moreover, it has been experimentally proven that Asp-342, together with Asp-344 and Asp-346, is involved in fibrinogen Ca²⁺ binding in a polymerisation site¹⁷. As a consequence, modified biochemical interactions occurring at this site could impair fibrinogen polymerisation. Trio investigation revealed that the proposita’s mother had both the variants, while the father had no variants in his FGA, FGB, or FGG sequences. Thus, the p.Asp342Glu and duplication variants presumably exert cis-acting effects.

Figure 4.

Molecular modelling of fibrinogen.

(A) Aspartate-342 forms two H-bonds. (B) Substitution of aspartate-342 with glutamic acid causes the rupture of the two H-bonds and generation of one new bond. This image was generated with Swiss-PdbViewer v4.1 using the Protein Data Bank file 3GHG.PDB.

Discussion

In this study of ten subjects with fibrinogen disorders, we confirmed the known heterogeneity in terms of clinical manifestations, which varied from bleeding to thrombotic events and pregnancy complications¹⁸.

We identified four novel missense mutations and one duplication variant, all of which were found in hypofibrinogenaemic women. The variants p.Glu41Lys, p.His333Arg and p.Asp342Glu (in cis with the duplication variant) were clinically relevant. They were found in hypofibrinogenaemic subjects with symptomatic bleeding, which ranged from mild (haematoma after a fall) to severe (post-partum haemorrhage) bleeding episodes. Hypofibrinogenaemic individuals are traditionally categorised as asymptomatic; however, specific challenges may determine the bleeding phenotype.

In silico predictions suggest that these novel missense changes could be deleterious and affect the biochemical and structural features of fibrinogen. In agreement with other authors, we suggest screening FGA exons 1–3 in individuals with quantitative fibrinogen disorders, when the most frequent mutations are not identified¹⁸.

The p.Gly191Val and p.His333Arg variants involve residues forming the gamma chain C-terminal end and they would be mutational amino-acid sites that cause hypofibrinogenaemia. The former may cause a polymerisation defect, affecting the folding pattern of the C-terminal globular domain of the fibrinogen gamma chain¹⁵. The latter has been previously described in hypofibrinogenaemic women with post-partum bleeding and could play a key role in fibrinogen assembly, as it interacts with the Gly-Pro-Arg binding cavity¹⁹.

Furthermore, Asp−342, −344 and −346, located in a D-domain in a C-terminal segment could have a role in binding to the platelet membrane, fibrin polymerisation and cross-linking of fibrin by factor XIII¹⁷. The previously reported p.Arg196Cys missense change, the so-called fibrinogen Longmont, modifies profibril organisation in the context of fibrinogen polymerisation and was previously described in heterozygous dysfibrinogenaemic patients²⁰ with menhorragia, post-partum haemorrhages and epistaxis, but no history of thrombosis²¹. At variance with the previously described individuals, the p.Arg196Cys heterozygous carrier here described did not have a history of bleeding episodes but had experienced miscarriages. Thus, the clinical symptoms associated with the p.Arg196Cys variant are quite heterogeneous.

In our case-series, carriership of FGB IVS7+1G>T was associated with pregnancy-related thrombosis. Phenotype comparisons cannot help in reaching conclusions on the possible pathgenicity of IVS7+1G>T, since the variant was previously described in homozygosis in an asymptomatic afibrinogenaemic 2-year old child²² and it would be premature to draw conclusions about this child’s phenotype given his, as yet, breif exposure to haemorrhagic or prothrombotic challenges.

As expected, the p.Gly288Ser variant, previously described in a homozygous afibrinogenaemic patient with a history of severe bleeding symptoms²³, when present in heterozygosis is associated with asymptomatic hypofibrinogenaemia (P5).

Again, as expected, the gamma chain p.Asp356Val and the beta p.Cys95Ser variants were not associated with clinically relevant events in our case series, as they were found in heterozygosis²⁴. The former, responsible for defective polymerisation in reptilase-treated plasma samples, is associated in homozygosis with ischaemic events²⁵.

In conclusion, the novel variants herein reported involve residues belonging to well-known fibrinogen motifs, which affect fibrinogen polymerisation and assembly. It is, therefore, conceivable that the resulting mutated fibrinogen molecules may not be assembled properly and may not be able to enter the secretory pathway. Our data also confirm the heterogeneity of phenotypes of fibrinogen disorders, which could be explained, at least in part, by some differences in clot structure and stability. Therefore, global tests, such as thromboelastometry and a thrombin generation assay, could help to clarify clinical variability in fibrinogen disorders.

Conclusions

This work contributes to expanding the mutational spectrum of congenital fibrinogen disorders. We identified five novel variants in hypofibrinogenaemic subjects, confirming that molecular investigations help to reach the correct diagnosis. In silico algorithms, based on multiple alignment data and on amino-acid chemico-physical properties, suggested a possible deleterious effect of these new variants on the fibrinogen structure. However, inferences from biochemical and functional studies could help to give a comprehensive characterisation of these novel variants.