Liver histology of an afibrinogenemic patient with the Bβ-L353R mutation showing no evidence of hepatic endoplasmic reticulum storage disease (ERSD); comparative study in COS-1 cells of the intracellular processing of the Bβ-L353R fibrinogen vs. the ERSD-associated γ-G284R mutant


  • S. DUGA,

    1. Department of Biology and Genetics for Medical Sciences, University of Milan, Milan, Italy
    Search for more papers by this author

    1. Department of Medicine, Surgery and Dentistry A.O. San Paolo and IRCCS (Istituto di Ricovero e Cura a Carattere Scientifico) Maggiore Hospital, University of Milan, Milan, Italy
    Search for more papers by this author

    1. Department of Biology and Genetics for Medical Sciences, University of Milan, Milan, Italy
    Search for more papers by this author

    1. Department of Medicine, Surgery and Dentistry A.O. San Paolo and IRCCS (Istituto di Ricovero e Cura a Carattere Scientifico) Maggiore Hospital, University of Milan, Milan, Italy
    Search for more papers by this author

    1. Angelo Bianchi Bonomi Hemophilia and Thrombosis Center and Fondazione Luigi Villa, Department of Internal Medicine, University of Milan and IRCCS Maggiore Hospital, Milan, Italy
    Search for more papers by this author

    1. Department of Medicine, Surgery and Dentistry A.O. San Paolo and IRCCS (Istituto di Ricovero e Cura a Carattere Scientifico) Maggiore Hospital, University of Milan, Milan, Italy
    Search for more papers by this author
  • G. COGGI,

    1. Department of Medicine, Surgery and Dentistry A.O. San Paolo and IRCCS (Istituto di Ricovero e Cura a Carattere Scientifico) Maggiore Hospital, University of Milan, Milan, Italy
    Search for more papers by this author

    1. Department of Biology and Genetics for Medical Sciences, University of Milan, Milan, Italy
    Search for more papers by this author

    1. Department of Biology and Genetics for Medical Sciences, University of Milan, Milan, Italy
    Search for more papers by this author

Maria Luisa Tenchini, Department of Biology and Genetics for Medical Sciences via Viotti, 3/5–20133 Milano, Italy.
Tel.: + 39 02 5031 5840; fax: + 39 02 5031 5864; e-mail:


Summary. Background: Type I fibrinogen deficiencies (hypofibrinogenemia and afibrinogenemia) are rare congenital disorders characterized by low or unmeasurable plasma fibrinogen antigen levels. Their genetic bases are represented by mutations within the three fibrinogen genes. Among the 11 reported missense mutations, a few have been characterized by expression studies and found to have an impaired fibrinogen assembly and/or secretion. Histopathological analyses were previously reported in two hypofibrinogenemic cases with discernible hepatic disease, revealing that both underlying mutations (γ-Gly284Arg and γ-Arg375Trp) were associated with hepatic fibrinogen endoplasmic reticulum storage disease (ERSD). Objective: The objective of this study was to investigate the liver histology in an afibrinogenemic patient, homozygous for the Bβ-Leu353Arg mutation, and to study the intracellular processing of the mutant protein. Patients and methods: Liver histology was evaluated by light microscopy, electron microscopy and immunocytochemistry. Intracellular processing of mutant fibrinogen was analyzed by pulse–chase labeling and immunoprecipitation experiments. Messenger RNA levels were determined by real-time reverse transcription–polymerase chain reaction (RT-PCR). Results: The histopathological characterization of the liver showed no signs of fibrinogen accumulation, a difference from the previously reported findings in two hypofibrinogenemic kindreds with ERSD. To evaluate whether the Bβ-Leu353Arg mutation and the ERSD-associated γ-Gly284Arg mutation affected intracellular fibrinogen trafficking differently, both mutant proteins were expressed in COS-1 cells. Bβ-Leu353Arg led to a more severe secretion defect, but no differences that could explain phenotype–genotype correlation were found in the intracellular processing. Endoglycosidase-H analysis demonstrated a secretion block before translocation to the Golgi medial stacks. Real-time RT-PCR studies showed normal levels of the Bβ mRNA in the patient's liver. Conclusions: The results confirm that Bβ-Leu353Arg is associated with impaired fibrinogen secretion, but not with hepatic ERSD.


Fibrinogen is a 340-kDa glycoprotein mainly synthesized by hepatocytes and circulating in plasma as a hexamer made up by two sets of three homologous polypeptide chains (Aα, Bβ, and γ) [1,2]. The three-dimensional structure of hexameric fibrinogen is characterized by a central E domain (containing the N-terminus of each chain) and two lateral globular D domains (containing the C-terminus of the Bβ and γ chains). Central and lateral domains are linked by coiled-coil triple helix structures [3]. The three chains are encoded by three different genes (FGA, FGB and FGG, for Aα, Bβ and γ chains, respectively) clustered on chromosome 4q31.3-4q32.1 [4]. Fibrinogen assembly occurs in a stepwise manner: initially Aα–γ and Bβ–γ dimers are produced, subsequently a third chain is added, and finally the Aα–Bβ–γ half-molecules dimerize to form hexameric fibrinogen [5]. This process occurs rapidly in the endoplasmic reticulum (ER) with the aid of chaperone proteins such as binding protein, glucose-regulated protein 94, protein disulfide isomerase and calnexin [6,7]. Glycosylation begins in the ER at single asparagine residues, 364 and 52, in the Bβ and γ chains, respectively [8]. After maturation of N-linked oligosaccharides, hydroxylation, sulfation, and phosphorylation of specific side chains through the Golgi network [5], the mature protein is secreted into the circulation. Congenital fibrinogen defects can be divided into two subclasses on the basis of plasma concentration: type I deficiencies (hypofibrinogenemia and afibrinogenemia), with low or unmeasurable antigen levels, and type II deficiencies (dysfibrinogenemia and hypo-dysfibrinogenemia) with normal or altered antigen levels associated with reduced coagulant activity. While dysfibrinogenemias are generally autosomal dominant diseases, type I fibrinogen deficiencies are inherited as autosomal recessive traits. Most patients with mild to moderate hypofibrinogenemia are indeed heterozygotes for mutations that, in the homozygous or compound heterozygous state, would result in a severe deficiency. Patients affected by congenital afibrinogenemia (Mendelian Inheritance in Man #202400) or severe hypofibrinogenemia (Mendelian Inheritance in Man + 134820, *134830 and *134850) may experience bleeding manifestations ranging from mild to very severe [9,10]. Moreover, both afibrinogenemia and hypofibrinogenemia are sometimes associated with thromboembolic disease, and thrombotic complications occurring spontaneously or after infusion of fibrinogen-containing preparations have been reported in afibrinogenemic patients [for review see ref. 11].

The genetic bases of type I fibrinogen deficiencies are invariantly represented by mutations (mostly null alleles) within the three fibrinogen genes [12]. All 11 missense mutations so far described (Table 1) [13–22] are clustered in the Bβ- and γ-chain genes and, when characterized by protein expression studies, they were shown to impair either the assembly [19] or the secretion process [15–17]. Access to liver biopsies allowed histological and ultrastructural studies only in two hypofibrinogenemic patients, revealing that both underlying mutations, namely γ-Gly284Arg (Fibrinogen Brescia) and γ-Arg375Trp (Fibrinogen Aguadilla), were associated with hepatic ER storage disease (ERSD). Intrahepatic fibrinogen storage was observed by immunohistological techniques also in two German families with hypofibrinogenemia, but no molecular studies were performed [23,24].

Table 1.  Known missense mutations underlying type I fibrinogen deficiencies
ExonMutation designation*Nucleotide substitutionPheno-typeGeno-typeFg:Ag (mg dL−1)Fg:C (mg dL−1)Protein expressionHistological studyRef.
  1. *Mutation designation refers to the mature protein (without signal peptide); †numbering is according to GenBank accession numbers M64983 and M10014 for FGB and FGG, respectively; ‡fibrinogen level was measured by a gravimetric method; §mean level of six probands' relatives, all carriers of the mutation.

  2. Fg:Ag = fibrinogen antigen level (normal range 160–400 mg dL−1); Fg:C = fibrinogen functional level (normal range 160–400 mg dL−1); nr =not reported; hypo = hypofibrinogenemia; afib = afibrinogenemia; hetero = heterozygous; homo = homozygous; comp hetero = compound heterozygous.

FGB gene
6Arg255His6654G > Ahypohetero100 (fibrino-peptide)130nono13
7Asp316Tyr7044G > Thypohetero90100nono14
7Leu353Arg7156T > GaFibhomo0.13< 5yesthis paper15
8Gly400Asp7915G > AaFibhomo1.59< 5yesno15
8Gly414Ser7956G > AaFibcomp.
8Trp437Gly8025T > GaFibhomo0.062< 5yesno17
FGG gene
4Ala82Gly2525C > Ghypohetero8070/110nono18
6Cys153Arg4647A > Ghypohetero8781yesno19
8Gly284Arg7502G > Chypohetero10020this paperyes20
9Thr371Ile9402C > Thypohetero80§90§nono21
9Arg375Trp9413C > Thypoheteronr60noyes22

Here we describe the histological, ultrastructural, immunocytochemical and molecular characterization of an afibrinogenemic patient, whose genetic defect was previously identified as a homozygous Bβ-Leu353Arg missense mutation [15].

Patients, materials and methods


The clinical features and the molecular characterization of the analyzed Italian afibrinogenemic patient were previously reported [15]. He was found to carry a homozygous missense mutation (Leu353Arg) in FGB, resulting in a secretion defect. As a consequence of treatment with whole blood and fibrinogen concentrates at birth and afterwards, abnormally high levels of transaminases were first detected at 6 years of age and hepatitis C virus (HCV) infection was diagnosed a few years later, when specific serological tests became available. Transient jaundice and acute pancreatitis occurred at the age of 19 years when gallbladder stones were detected by abdominal ultrasonography. Medical therapy allowed a complete recovery from pancreatitis, however, mild jaundice episodes recurred during the next year and laparotomic cholecystectomy was planned. The surgical setting was considered adequate to perform a liver biopsy that would provide histological data useful for the clinical management of chronic liver disease. The patient signed an informed consent for both surgical procedures. Solvent-detergent-treated plasma (Octaplas®; Octapharma, Vienna, Austria) and fibrinogen concentrate (Fibrinogeno Tim 3®; Baxter, Vienna, Austria) were successfully administered to maintain the hemostasis during the peri- and postoperative periods.

Liver samples

The liver biopsy was divided into three fragments; the first was fixed in formalin and routinely processed for light microscopy examination; the second was routinely processed for ultrastructural studies; the third was snap frozen in liquid nitrogen and maintained at − 80 °C for immunohistochemical and molecular studies. Normal liver samples were obtained from three patients who were undergoing surgical resections of benign liver lesions and were used as calibrator samples in real-time reverse transcription–polymerase chain reaction (RT-PCR) experiments. A fourth sample was obtained by a wedge liver biopsy from a patient with chronic B hepatitis and used as anti-fibrinogen antibody immunoreactivity positive control.

Light microscopy

Sections (4 μm thick) were obtained from formalin-fixed paraffin-embedded liver tissue. Sections were routinely stained with hematoxylin & eosin, Masson trichrome, and PAS (periodic acid Schiff) with and without diastase digestion and Pearl's iron.

Electron microscopy

The sample was fixed in 2.5% glutaraldehyde in 0.13 mol L−1 phosphate buffer and routinely processed for ultrastructural study. Counterstained ultra-thin sections were examined with a Jeol JEM1010 electron microscope (Jeol, Tokyo, Japan).


Sections of 4–5 μm thickness from paraffin blocks were deparaffinized, pretreated for antigen retrieval in 0.01 mol L−1 citrate buffer (pH 6) twice for 6 min in a microwave oven, and incubated with the following antibodies (Dako Cytomation, Glostrup, DK): polyclonal anti-human fibrinogen (1 : 400 dilution); monoclonal anti-human hepatocyte antigen (hepatocyte paraffin, HePar 1, clone OCVH 1E5; 1 : 200 dilution); and monoclonal anti-cytokeratin 7 (clone OV-TL 12/30; 1 : 200 dilution). Cryosections from frozen samples were incubated with fluorescein-labeled anti-fibrinogen antibody diluted 1 : 30. All slides were incubated in an automated slide stainer (Biogenex, S. Ramon, CA, USA); the reaction products were visualized by indirect staining methods using En Vision (Dako).

Liver sections from the hepatitis B surface antigen-infected patient were added as anti-fibrinogen-antibody-positive controls [25]. Human hepatocyte (OCH1E5) and cytokeratin 7 (OVTL12/30) antigens were stained to test the immunoreactivity of liver tissue.

Expression plasmids

pRSV-Neo-Aα, pRSV-Neo-Bβ, and pRSV-Neo-γ plasmids [26,27], containing the full-length cDNAs coding for the three fibrinogen chains, were kindly provided by Dr C. M. Redman (New York Blood Center, New York, NY, USA). To produce the pRSV-Neo-γ-G284R plasmid expressing the Gly284Arg mutant γ chain, the pRSV-Neo-γ plasmid was used as the template for site-directed mutagenesis using the QuickChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA), according to the manufacturer's instructions. The resulting mutant plasmid was checked by sequencing. The pRSV-Neo-Bβ-L353R plasmid was produced as described [15].

Cell cultures, transfections and metabolic labeling

African green monkey kidney COS-1 cells were cultured according to standard procedures [15]. All plasmids were extracted using the EndoFree Plasmid Maxi Kit (Qiagen, Hilden, Germany). Semi-confluent COS-1 cells were cotransfected with equimolar amounts of pRSV-Neo-Aα, pRSV-Neo-Bβ, and pRSV-Neo-γ plasmids, to express the wild-type protein. Fibrinogens containing either the Leu353Arg Bβ chain or the Gly284Arg γ chain were expressed cotransfecting the relevant mutant plasmid (pRSV-Neo-Bβ-L353R or pRSV-Neo-γ-G284R) together with equimolar amounts of the two remaining wild-type expression vectors. As negative control, COS-1 cells were transfected with an unrelated plasmid (pUC18). Transfections and pulse–chase experiments were performed as described previously [17], with the exception that chase periods were 2, 4, 8, 12 and 16 h.

Immunoprecipitation and sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE)

Conditioned media and cell lysates were prepared as described [15]. Immunoprecipitations were performed using magnetic beads coated with protein G (Dynabeads Protein G; Dynal Biotech, Oslo, Norway) and polyclonal antibodies to human fibrinogen. Briefly, Dynabeads Protein G (14 μL per sample) were washed twice in phosphate-buffered saline containing 1.5% Triton X-100 and a protease inhibitor mixture (Complete; Roche, Basel, Switzerland) and incubated with polyclonal anti-human fibrinogen antibodies (80 μg per sample) for 40 min at room temperature under gentle shaking. Coated beads were then washed three times as described above and added to 1 mL of cell lysate or conditioned medium. Binding was performed at 4 °C for 1 h under gentle mixing and the protein–bead complexes were washed three times as described above, resuspended in 20 μL SDS–PAGE Laemmli loading buffer (without reducing reagent), and stored at − 80 °C until use. The immunoprecipitated proteins were released from the beads by boiling for 5 min. Samples were fractionated either by non-reducing 4% SDS–PAGE or, after the addition of a reducing agent (Bond-Breaker; Pierce Biotechnology, Rockford, IL, USA), by reducing 7.5% SDS–PAGE. Gels were dried under vacuum at 80 °C for 1 h, exposed overnight to a storage phosphor screen, and analyzed using a Typhoon 9200 phosphor imager (Amersham Pharmacia Biotech, Uppsala, Sweden).

Endoglycosidase-H treatment

Endoglycosidase H (Endo H) was obtained from New England BioLabs (Beverly, MA, USA). Five-microliter aliquots of immunoprecipitated proteins were incubated for 10 min at 100 °C in the presence of 1 × denaturing buffer and subsequently with 500 U of Endo H and 1 μL of 10 × G5 buffer for 2 h at 37 °C, according to the manufacturer's instructions. Samples were analyzed by 7.5% SDS–PAGE under reducing conditions and proteins were visualized as described above.

RNA extraction and cDNA synthesis

Total RNA was isolated from liver samples using the Trizol kit (Invitrogen, Carlsbad, CA, USA) following the manufacturer's instructions. The reverse transcription mixture (50 μL final volume) contained 200 ng per sample of total RNA, 1 × PCR buffer II, 5.5 mmol L−1 MgCl2, 1 mmol L−1 deoxynucleotide triphosphates, 2.5 μmol L−1 random hexamers, 20 U RNase inhibitor, and 62.5 U Moloney murine leukemia virus (M-MuLV) reverse transcriptase. Samples were incubated at 25 °C for 10 min, 48 °C for 30 min, and 95 °C for 5 min. All reagents were obtained from Applied Biosystems (Foster City, CA, USA).

Real-time RT-PCR

FGB and FGG mRNA levels were evaluated in liver samples by real-time RT-PCR based on TaqMan methodology, using the ABI PRISM 7700 Sequence Detection System (Applied Biosystems) as described elsewhere [28]. FGB and FGG expression was determined applying the ΔΔ cycle threshold (Ct) method [29]. A calibrator sample representing the unitary amount of the target of interest (1 × sample) was used. The other samples express n-fold mRNA of interest relative to the calibrator. One of the three normal liver samples was used as calibrator. To normalize the amount of total RNA added to each reaction mixture, two different internal controls, the mRNAs for β2-microglobulin (B2M) and β-actin (ACTB), were quantified.

Intron-spanning primers and TaqMan probes for B2M and ACTB mRNAs were designed using the primer express software (Applied Biosystems; sequences available on request). FGB and FGG mRNAs were quantified by using ready-to-use assays (Assay-on-Demand™ Gene Expression Products; Applied Biosystems): Hs00170586_m1 and Hs00241037_m1 for FGB and FGG mRNAs, respectively.

All PCRs were performed in duplicate using the Universal TaqMan 2 × PCR mastermix (Applied Biosystems) in a volume of 25 μL containing 5 μL cDNA. The thermal profile included 2 min at 50 °C, 10 min at 95 °C, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min.



Light microscopy  The liver presented a normal lobular structure with focal mild enlargement of portal tracts and focal presence of thin portal-to-portal, mostly incomplete, septa; minimal fibrosis of Disse spaces was focally observed; a mild lymphoid infiltrate was present in a few portal tracts, consistent with HCV infection, without evidence of interface hepatitis or intralobular necroinflammatory activity. Hepatocytes were normal in size and shape; no pigments or globular intracytoplasmic inclusions were observed with PAS and PAS-diastase staining; mild focal ceroid pigment was present in Kupffer's cells (Figs 1a,b).

Figure 1.

Light microscopy of the afibrinogenemic patient's liver. (a) PAS staining after diastase digestion, showing the lack of cytoplasmic inclusions in hepatocytes (magnification 400 ×); (b) Masson staining, displaying mild fibrosis of portal spaces, with thin portal-to-portal septa, and minimal flogistic infiltration, without evidence of interface hepatitis (magnification 200 ×); (c) immunostaining with polyclonal anti-human fibrinogen antibodies, showing no detectable immunostained fibrinogen in the cytoplasm of hepatocytes. The insert shows the positive control in a hepatitis B virus-infected liver (magnification 400 ×).

Electron microscopy  Only mild ultrastructural alterations were detected in the sample, i.e. focal cytoplasmic accumulation of lipofuscin granules and rare pleomorphic and giant mitochondria (Fig. 2). No dilatation of rough ER cisternae with inclusion bodies was observed.

Figure 2.

Transmission electron microscopy of the afibrinogenemic patient's liver. Mild ultrastructural changes are present in hepatocytes near a polymorphonuclear inflammatory cell (PMN); scale bar = 2 μm. In the enlargement of the boxed area numerous lipofuscin granules and a giant pleomorphic mitochondrium (Mt) are visible.

Immunohistochemistry  Hepatocytes were not immunoreactive using anti-fibrinogen antibody, neither with indirect technique on paraffin section (Fig. 1c) nor with direct immunofluorescent technique on cryosections (not shown); all positive controls revealed focal clear cytoplasmic immunoreactivity in each experiment. Anti-human hepatocyte antigen and anti-cytocheratin 7 antibodies diffusely stained hepatocyte cytoplasm in the sample under study and in positive controls.

Molecular studies

Pulse-chase analysis of Bβ-Leu353Arg and γ-Gly284Arg fibrinogens  To evaluate whether the Bβ-Leu353Arg mutation (carried by the patient analyzed here) and the γ-Gly284Arg mutation (previously reported in a hypofibrinogenemic patient associated with fibrinogen ERSD) [20] showed any difference in secretion and/or intracellular accumulation when expressed ex vivo in eukaryotic cells, pulse–chase experiments were performed in COS-1 cells (not expressing fibrinogen). To this purpose, COS-1 cells, transiently transfected with the pRSV-Neo-Bβ-L353R or with the pRSV-Neo-γ-G284R plasmid together with equimolar amounts of plasmids expressing the other two wild-type chains, were pulse-labeled for 2 h, and subsequently chased for 2–16 h. Expression of the mutant Leu353Arg Bβ chain did not lead to secretion of hexameric fibrinogen, consistent with previous experiments performed with shorter chase periods [15](Fig. 3). In media conditioned by cells expressing the γ-Gly284Arg chain, small amounts of hexameric fibrinogen were detected after 8 and 12 h of chase, suggesting that fibrinogen secretion was highly reduced but not completely abolished (Fig. 3). Concerning intracellular fibrinogen, no differences appeared between the two mutants: in both cases a radioactive band corresponding to hexameric fibrinogen was visible up to 16 h of chase.

Figure 3.

In vitro expression of Bβ-Leu353Arg, γ-Gly284Arg, and wild-type fibrinogens. Cells were labeled with [35S]methionine and [35S]cysteine for 2 h and chased with an excess of the corresponding cold residues for 2–16 h. Immunoprecipitated proteins from cell lysates and conditioned media of COS-1 cells cotransfected with pRSV-Neo-Bβ-L353R (Bβ-Leu353Arg) or pRSV-Neo-γ-G284R (γ-Gly284Arg) and equimolar amounts of plasmids expressing the other two wild-type chains; with the three wild-type plasmids (wild-type); or with the unrelated pUC18 plasmid (pUC18); were separated on 4% SDS–PAGE under non-reducing conditions, and visualized using a Typhoon 9200 phosphor imager. Arrowheads indicate the 340-kDa hexameric fibrinogen or its expected position within the gel.

Endoglycosidase analysis  The fibrinogen Bβ and γ chains are N-glycosylated at a single residue (Asn364 and Asn52, respectively) [8]. Therefore, a treatment with Endo H, which specifically cleaves high-mannose N-linked oligosaccharides (typical of proteins present in the ER and in the cis-Golgi compartment) but does not affect complex oligosaccharides (typical of proteins present in the trans-Golgi compartment), was used to evaluate the translocation from the ER to the Golgi of the Bβ-Leu353Arg or γ-Gly284Arg mutant fibrinogens. Immunoprecipitated proteins from lysates of cells expressing either the wild-type or each of the mutant fibrinogens were Endo-H treated and analyzed by SDS–PAGE under reducing conditions. Endo-H treatment increased the electrophoretic mobility of the bands corresponding to Bβ and γ chains both in wild-type (data not shown) and mutant samples (Fig. 4), indicating that they did not contain complex oligosaccharides and therefore might be located in the ER or in the cis-Golgi complex. As expected, the wild-type fibrinogen secreted into the medium was Endo-H resistant, as demonstrated by the lack of shifting of the bands corresponding to the fibrinogen Bβ and γ chains (Fig. 4).

Figure 4.

Endoglycosidase-H treatment of intracellular Bβ-Leu353Arg and γ-Gly284Arg fibrinogens. Immunoprecipitated proteins from lysates of pulse–chase labeled COS-1 cells were treated (+) or not (–) with Endo H, and separated by 7.5% SDS–PAGE, under reducing conditions. The glycosylated Bβ and γ chains were Endo-H sensitive at each analyzed chase-period. The slight change in the electrophoretic mobility of the Aα chain (not glycosylated) probably reflects a proteolytic cleavage as a result of contaminating proteases during the Endo-H treatment. As controls, treated and untreated wild-type fibrinogen immunoprecipitated from conditioned media was also loaded into the gels. Lys = lysates; CM = conditioned media; BIP = binding protein, coimmunoprecipitated by polyclonal anti-human fibrinogen antibodies; BβM and γM = fully glycosylated mature Bβ and γ chains; BβP and γP = precursor forms of Bβ and γ chains; BβD and γD = deglycosylated Bβ and γ chains; AαDEG = proteolyzed Aα chain.

FGB and FGG mRNA expression  The mean Ct values of two internal controls were 20.49 (range 19.7–22) and 21.5 (range 21–21.9) for B2M and ACTB, respectively, documenting the appropriate quality of RNA. FGB : FGG ratios were determined after separate measurement of FGB and FGG mRNA levels for each sample. The FGB : FGG ratios of three normal livers and of the sample of interest are reported in Table 2. No significant differences were detected between the normal liver control group and the afibrinogenemic patient's liver.

Table 2. FGB/FGG ratios in the analyzed afibrinogenemic patient's liver and in normal livers
 FGB/FGG ratio (B2M endogenous control)FGB/FGG ratio (ACTB endogenous control)
  1. B2M = β-2 microglobulin; ACTB = β-actin.

Normal liver 1 (calibrator sample)1.001.00
Normal liver 21.091.09
Normal liver 31.431.43
Mean (SD)1.17 (± 0.23)1.17 (± 0.23)
Case under study1.421.41


Our previous molecular study on the afibrinogenemic patient reported here demonstrated a lack of secretion of recombinant fibrinogen containing the mutant Leu353Arg Bβ chain. Owing to ethical constraints, which denied access to liver biopsies, no clues could be gathered on the possible in vivo intracellular accumulation of fibrinogen within hepatocytes [15]. Recently, the occasion of a cholecystectomy, combined with the necessity for a histological evaluation of the proband's liver because of his HCV infection, made a liver biopsy available. This represents the first case of a histological study of the liver in an afibrinogenemic patient, even though accurate liver histopathological examinations have been previously reported in hypofibrinogenemic patients [20,22]. Histology showed mild fibrosis with minimal chronic portal infiltrate, without interface or interlobular necroinflammatory activity, consistent with minimal lesion, HCV-related chronic hepatitis. The only additional feature was minimal fibrosis of Disse spaces, which is not typical of the early stages of chronic HCV infection. No cytoplasmic inclusions were detected after PAS staining and diastase digestion, which is different from what was described in the case of Fibrinogen Brescia [20] and from what is commonly observed in α1-anti-trypsin deficiency [30]. No signs of fibrinogen accumulation within hepatocytes could be documented by light microscopy, immunocytochemistry, or electron microscopy.

Hepatic retention of mutant fibrinogens within the ER was previously reported as a result of missense mutations (Gly284Arg or Arg375Trp) in the fibrinogen γ chain, leading to quantitative fibrinogen deficiency [20,22]. In particular, the γ-Gly284Arg mutation was found in an Italian family in which two heterozygous carriers (out of seven) had chronic liver disease. The remaining five carriers, however, had hypofibrinogenemia but no apparent symptoms of liver disease [20]. A similar variability in the pattern of liver disease symptoms was found for the γ-Arg375Trp mutation, which was associated with chronic liver disease in two heterozygous hypofibrinogenemic sisters but not in their father, who, although hypofibrinogenemic and heterozygous for the same mutation, had only mildly increased transaminase levels [22].

To compare the intracellular fate of different mutant fibrinogens, associated or not with ERSD, we therefore performed pulse–chase experiments in COS-1 cells by expressing either the Bβ-Leu353Arg or the γ-Gly284Arg [20] mutants, and prolonging the chase period up to 16 h. No major differences that could explain phenotype–genotype correlation were demonstrated in the intracellular processing of Bβ-Leu353Arg and γ-Gly284Arg mutants. Concerning secretion of the mutant proteins, analysis of the supernatants suggested that the Bβ-Leu353Arg mutation could lead to a more severe secretion defect than γ-Gly284Arg. In fact, fibrinogen containing the Bβ-Leu353Arg mutation was absent from the conditioned media, in agreement with previous expression data [15]. Only a very faint band was visible after 12 h of chase. On the contrary, small amounts of fibrinogen could be detected in the supernatants of cells expressing the γ-Gly284Arg mutation. This is interesting considering that the functional fibrinogen level of the proband studied by Brennan and colleagues [20] (0.2 mg mL−1) was significantly lower than the immunoreactive fibrinogen level (1.0 mg mL−1), suggesting the possible presence in plasma of dysfunctional molecules. However, the presence of circulating aberrant molecules was excluded using different techniques [20]. If tiny amounts of mutant γ-Gly284Arg fibrinogen are secreted, these molecules might be unstable in plasma.

Since Bβ- and γ-fibrinogen chains are known to be glycosylated, to provide insight into intracellular protein trafficking, immunoprecipitated proteins from lysates of COS-1 cells expressing either Bβ-Leu353Arg or γ-Gly284Arg mutants were treated with Endo H. Intracellular fibrinogen was found to be Endo-H sensitive at each analyzed time period, even though faint bands probably corresponding to the fully glycosylated mature Bβ and γ chains were visible on the gels, demonstrating that small amounts of fibrinogen were processed through the Golgi apparatus (Fig. 4). N-linked oligosaccharides from wild-type secreted fibrinogen had been processed to the complex type in the Golgi, as demonstrated by their Endo-H resistance. The marked reduction of Endo H-resistant forms for both mutant proteins is suggestive of a block in intracellular transport before translocation to the medial stacks of the Golgi apparatus. This is consistent with ER accumulation of fibrinogen containing the γ-Gly284Arg chain found in the proband carrier of this mutation [20]. The absence of a correlation between the amount of fibrinogen retained intracellularly and the differences in the phenotypes of patients (ERSD vs. normal liver) could be because aggregation and accumulation of fibrinogen in the ER are probably very slow processes that cannot be mimicked by in vitro experiments. The demonstration of the lack of intracellular fibrinogen accumulation in the afibrinogenemic patient's liver raised the question whether the Bβ-Leu353Arg mutation could alter in vivo the transcription rate or the half-life of FGB mRNA. To this purpose, real-time RT-PCR assays were used to evaluate the Bβ-chain transcript in the patient's liver. Normal levels of FGB mRNA were found, suggesting that the extremely reduced levels of secreted fibrinogen as well as the lack of hepatic fibrinogen accumulation were not the result of either reduced transcription of the mutant FGB or of decreased mRNA stability.

In conclusion, this paper reports the first histological examination of the liver in an afibrinogenemic patient. All the reported data suggest that the Bβ-Leu353Arg mutation causes afibrinogenemia by altering protein secretion without causing fibrinogen accumulation within the ER. This contrasts with the effect of other missense mutations found in the γ-chain gene of patients affected by hypofibrinogenemia. A possible explanation of these differences could be the presence of a concomitant defect in the ER degradation pathway in some patients, predisposing to ERSD, as demonstrated for ZZ-antitrypsin deficiency, in which only 10–15% of ZZ homozygotes develop liver injury [30]. The tendency of mutant fibrinogens to aggregate within the ER could also depend on the type of missense mutation leading to the secretion defect and, possibly, on the involved fibrinogen chain (only missense mutations in the γ chain have been so far associated with ERSD). Missense mutations causing hepatic fibrinogen accumulation change amino acids (Gly284 or Arg375) exposed to the protein surface, which are probably more prone to cause fibril aggregation. Conversely, missense mutations so far reported in the Bβ chain involve residues buried underneath the protein surface.

Moreover, availability of fresh liver biopsy enabled the experimental demonstration of normal mRNA levels in this organ, thus confirming that the Bβ-Leu353Arg mutation exerts its pathogenetic role at the protein level.


The financial support of Telethon – Italy (Grant no. GGP030261) is gratefully acknowledged. This work was supported by MURST (Ministero dell'Università e della Ricerca Scientifica e Tecnologica; Grant no. 2002061282), and FIRB (Fondo per gli Investimenti della Ricerca di Base; Grant no. RBAU01SPMM). We thank the proband (GE) and his family for their participation in this study. We wish to thank Dr C.M. Redman and Dr H. Xia (Lindsley F. Kimball Research Institute of the New York Blood Center, New York, USA) for their helpful suggestions on the interpretation of endoglycosidase H experiments.

All the authors participated in the conception and design of the present study, in the analysis and interpretation of data, and in revising the manuscript. S.D. and R.A. were responsible for the conception of the study, the interpretation of results, and for the writing of the manuscript; they also performed site-directed mutagenesis, expression experiments, and immunoprecipitation experiments; P.B. and M.M. were responsible for light microscopy, electron microscopy and immunocytochemistry experiments; E.S. was responsible for the clinical management of the patients; C.P. was responsible for real-time PCR experiments; G.C. and M.M. were involved in the study design, in the discussion of the results, and in the reviewing of the manuscript; MLT supervised the entire study.