Genotype-phenotype correlation in combined deficiency of factor V and factor VIII

Combined deficiency of factor V and factor VIII (F5F8D) is an autosomal recessive disorder characterized by simultaneous reduction in the levels of factor V (FV) and factor VIII (FVIII) activity and antigen and mild-to-moderate bleeding symptoms.¹  The reported levels of FV and FVIII in F5F8D patients range from as low as 1% to as high as 46% of normal, but generally fall between 5% and 30%.¹  Positional cloning identified 2 genes that are associated with the disorder.^2,3  Mutations in LMAN1 account for approximately 70% of F5F8D families, while mutations in MCFD2 account for the remaining 30%.^4-6  LMAN1 is a type-1 transmembrane protein that cycles between the endoplasmic reticulum (ER) and the ER-Golgi intermediate compartment (ERGIC).^7,8  It contains a mannose-specific carbohydrate recognition domain on the ER lumenal side and ER exit and retrieval motifs on the cytoplasmic side.⁹  MCFD2 is an EF-hand domain protein that interacts with LMAN1 in a Ca²⁺-dependent manner.³  The LMAN1-MCFD2 protein complex functions as a cargo receptor that facilitates the transport of FV and FVIII from the ER to the Golgi apparatus.^3,10 

All the LMAN1 mutations reported to date are null mutations with the exception of a cysteine-to-arginine mutation that disrupts a disulfide bond that is required for its oligomerization and also destabilizes the protein.⁶  In contrast, both null mutations and missense mutations have been identified in MCFD2. All 4 MCFD2 missense mutations reported to date change highly conserved amino acid residues in the EF hand domains,^3,6,11  and 2 have been shown to abolish LMAN1 binding,³  indicating that LMAN1 and MCFD2 must function as a unit to transport FV and FVIII. Although MCFD2 appears to directly interact with FVIII,¹⁰  it is unclear whether LMAN1 binds directly to FV/FVIII or indirectly via MCFD2.

Here we report identification of mutations in MCFD2 and LMAN1 in an additional 11 F5F8D families. With the availability of these and other published data on a large number of patients with known mutations and the corresponding FV/FVIII levels, we also address the question of whether there is a phenotypic difference between patients with LMAN1 mutations and those with MCFD2 mutations, and whether FV in plasma and FV in platelets are similarly affected.

Table 1 lists the 15 affected individuals studied from 11 new families with affected individuals. Peripheral blood samples were obtained with written informed consent from probands and family members after diagnosis of F5F8D in accordance with the Declaration of Helsinki. FV and FVIII activities were measured at local clinical laboratories using one-stage assays based on prothrombin time (for FV) and activated partial thromboplastin time (for FVIII). For families B19 to B22 and B26 to B28, amplification by polymerase chain reaction (PCR) of exons and intron-exon junctions of the LMAN1 and MCFD2 genes and DNA sequencing were performed as reported previously.^2,3  For families B23 to B25, the presence of previously described mutations in Tunisian and Middle Eastern Jewish populations was confirmed by PCR amplification and restriction analysis as previously reported.¹² 

Cloning of wild-type MCFD2 and MCFD2^D129E into the pcDNA3.1-myc-his expression vector was described previously.³  The Stratagene mutagenesis II kit (La Jolla, CA) was used to introduce individual point mutations into the wild-type MCFD2 expression vector. The mutagenesis primers were as follows (only the sense primers are listed): D81Y, CAGCTCCATTACTTCAAAATGCATTATTATGATGGCAATAATTTGCTTG; D89A, GATGGCAATAATTTGCTTGATGGCTTAGAACTCTCCACA; and D122V, AAGATGAACTGATTAACATAATAGTTGGTGTTTTGAGAGATGATGAC. The presence of the desired mutation and the absence of a second mutation were verified by DNA sequencing.

COS-1 cells were transfected with the wild-type and different mutant MCFD2 expression vectors and labeled with ³⁵S methionine-cysteine (TRAN³⁵S LABEL; MP Biomedicals, Solon, OH) at 28 hours after transfection as previously described.³  Immunoprecipitation with antimyc, anti-MCFD2, and anti-LMAN1 antibodies was performed as previously described.¹⁰  Proteins were separated in 4% to 12% Criterion Bis-tris gels run with the Mes running buffer (Bio-Rad, Hercules, CA) and visualized by exposing to a Kodak Bio-Max film (Rochester, NY) using a Kodak Biomax Transcreen LE.

Immunofluorescence staining of HeLa cells transfected with different MCFD2 expression vectors was previously described.³  Images were viewed on a Leica DMRXE confocal microscope (Wetzlar, Germany) using a 40× oil-immersion objective with a 1.25 numeric aperture. Confocal images were acquired using the Leica Confocal Software, version 2.61.

The ABO glycosyltransferase gene around the common O1 allele (c.261delG) was amplified using primers MO-46 and MO-57 as described.¹³  The region around the rare O2 allele (c.G802A) was amplified using the following primers: O2-S, AGATCCTGACTCCGCTGTTC; O2-AS, CACAAGTACTCGGGGGAGAG. The presence of a homozygous O allele, as detected by DNA sequencing, was scored as blood type O. Other genotypes were not distinguished further and scored as non-O. In our patient samples, no O2 genotype was observed.

Washed human platelets were obtained from 7 individuals from 5 Italian families (Table 4), with mutations in LMAN1 or MCFD2, and from 15 healthy subjects, as previously described.¹⁴  Briefly, blood was collected in ACD (85 mM trisodium citrate, 71 mM citric acid, and 111 mM dextrose, pH 4.5) and centrifuged at 400g for 15 minutes at 20°C to obtain platelet-rich-plasma (PRP). After removal of platelet-poor plasma, platelet pellets were washed 3 times with Tyrode-17.5% albumin solution with 1 μM PGE1 as an inhibitor of platelet activation and then resuspended at 10⁹/mL in the same buffer supplemented with 1 μM PGE1 and 1 μL/mL apyrase (kindly provided by Dr R. L. Kinlough-Rathbone, McMaster University, Hamilton, ON) and 1 mM PMSF (phenylmethanesulphonyl fluoride). FV antigen levels were determined in both plasma and platelet lysates by an enzyme-linked immunosorbent assay (ELISA) using a sheep anti-FV antibody (Affinity Biologicals, Ancaster, ON). Plasma FV and platelet FV were normalized to the hematocrit and platelet counts, respectively, to compare the relative reduction of FV in each pool among F5F8D patients and controls. Of note, the size of the platelet (plt) FV pool in our study (∼ 3.7% of total circulating FV, or ∼ 637 ng/10⁹ plts in healthy controls) is significantly smaller than in previous reports (2500-7730 ng/10⁹ plts by a radioimmunoassay¹⁵  and 900-1300 ng/10⁹ plts by a polyclonal antisera-based ELISA¹⁶ ). This discrepancy could be due to intrinsic differences in antigenicity between platelet and plasma FV. Consistent with this notion, FV has been shown to be substantially modified in platelets.^17,18 

We compiled a list of the current and all previously reported patients with known mutations in either LMAN1 or MCFD2 and their corresponding FV and FVIII levels (Tables 2,3). In cases where FV and FVIII levels were measured multiple times, only the means are listed. Two-tailed Student t test for independent samples, assuming equal variance, was used to determine differences between the mean activity levels. The relationship between FV and FVIII was measured using the Pearson correlation test.

We analyzed 15 patients with F5F8D from 11 previously unreported families. Table 1 summarizes the factor V and factor VIII levels of the affected individuals. The proband in family B19 is of Afro-Caribbean origin. Family B20 is a consanguineous Turkish family with 2 affected siblings. Family B21 is an Iraqi Chaldean family. Family B22 is from Saudi Arabia. Families B23 to B25 are 3 unrelated Iranian Jewish families. Families B26 to B28 are from Italy. A novel homozygous MCFD2 mutation was identified in family B19 (c.374-375insGA). This mutation occurs in exon 3 and leads to frameshift and a premature stop of translation. A novel homozygous MCFD2 missense mutation (c.241G>T) was identified in family B22. This mutation causes a substitution of tyrosine for a highly conserved aspartic acid residue at amino acid position 81 (D81Y). Novel homozygous LMAN1 mutations (c.795delC and c.1356delC) were identified in both affected siblings in families B20 and B21, respectively. A previously reported mutation (c.149+5G>A) was identified in both affected siblings of family B27. The LMAN1 mutations in families B20 and B21 are single-nucleotide deletions in exon 8 and exon 11 that both result in a frameshift and a premature stop of translation. Analysis of families B23 to B25 identified homozygosity in all affected individuals for 2 previously reported mutations (c.1149+2T>G and c.89–90insG).²  The proband of family B26 carries a mutation (c.2T>C) commonly found in patients of Italian origin.¹⁹  Although no mutations were identified by DNA sequencing of the MCFD2 and LMAN1 genes in family B28, Western blot analysis demonstrated the absence of detectable LMAN1 protein (data not shown), suggesting an LMAN1 regulatory or splice mutation missed by the exonic sequence analysis.

The D81Y mutation is one of 5 missense mutations identified in MCFD2 to date, all of which result in an amino acid substitution at a highly conserved amino acid residue in one of the 2 EF hand domains. Two of these missense mutations (D129E and I136T), both located in the second EF hand domain, have been shown to disrupt LMAN1 binding.³  To test the effect of the remaining 3 missense mutations (D81Y and 2 previously reported mutations, D89A⁶  and D122V¹¹ ), each was expressed as a myc-tagged fusion protein and transfected into COS-1 cells. Antimyc antibody was used to specifically identify the missense mutant proteins. All 3 missense mutants failed to coimmunoprecipitate with LMAN1 (Figure 1A). In contrast, the wild-type MCFD2 (both the endogenous and the myc-tagged proteins) are readily immunoprecipitated in complex with LMAN1, as detected by both anti-LMAN1 and anti-MCFD2 antibodies (Figure 1A). By immunofluorescence staining, all 3 missense mutant MCFD2 proteins are detected by anti-myc antibody in transfected cells, but fail to colocalize with LMAN1 (Figure 1B), in contrast to wild-type MCFD2. The staining pattern for all 3 missense MCFD2 mutations resembles that of protein disulfide isomerase (PDI), an ER marker (Figure 1B), similar to the pattern previously observed for the D129E and I136T mutations.³ 

To date, at least 15 MCFD2 mutations and 32 LMAN1 mutations have been reported, including the 4 new mutations identified here. In addition, unidentified mutations may exist in regulatory regions of LMAN1 that result in no mRNA accumulation.⁶  Data on FV and FVIII levels are available for a total of 46 patients with MCFD2 mutations and a total of 96 patients with LMAN1 mutations (Table 2–3).^2-6,11,19-27 Figure 2 shows a comparison of FV and FVIII levels between patients with MCFD2 mutations and patients with LMAN1 mutations. We observed a small but statistically significant difference in the distribution of FV and FVIII levels between these 2 classes of patients, with lower levels for both factors in the patients with MCFD2 mutations compared with those with LMAN1 mutations (mean values of 9.6 vs 13.7 for FV [P < .001] and 10.0 vs 16.0 for FVIII [P < .001]). FVIII levels also exhibit a wider distribution, which is consistent with previous observations of healthy subjects and LMAN1 carriers.²⁸  In contrast, no significant differences were observed between male and female patients for either FV or FVIII levels (Figure S1, available on the Blood website; see the Supplemental Materials link at the top of the online article). This includes analyses of the whole patient population, as well as subgroups divided by gene mutations (LMAN1 versus MCFD2). Although averages are slightly higher in all female groups, this trend is not statistically significant.

FVIII, but not FV, is known to be affected by ABO blood type, with lower levels in blood group O individuals thought to result from reduction in plasma von Willebrand factor.^29,30  We thus tested the possibility that differences in ABO blood group might partially explain the distribution of FVIII levels observed in Figure 2. For all patients listed in Tables 2–3 for whom DNA samples were available, genotyping was performed for the 2 most common O genotypes (O1 and O2, together accounting for > 99% of group O³¹ ) to distinguish blood group O from non-O types. No significant difference in FV and FVIII levels was observed in patients with blood group O compared with non-O blood groups (Figure S2).

If the observed variation in factor levels is the result of biologic differences in the effect of loss of function for MCFD2 versus LMAN1, then plasma FV levels might be expected to correlate with FVIII levels for each individual patient. To test this hypothesis, we performed a Pearson correlation analysis on the FV/FVIII levels. Moderate correlation was observed in both patients with MCFD2 mutations (r = 0.53, P = .001) and those with LMAN1 mutations (r = 0.322, P = .001; Figure 3), suggesting that deficiencies in LMAN1 or MCFD2 exert a similar impact on FV and FVIII. A nonparametric Spearman correlation analysis yielded similar results (data not shown).

Circulating FV exists in both plasma and platelets. The platelet FV pool has been shown to originate from endocytosis of plasma FV in humans^18,32 ; although in mice, platelet FV is derived exclusively from biosynthesis within the megakaryocytes.^33,34  To address the question of whether LMAN1 and MCFD2 mutations exert differential effects between these 2 pools, we measured platelet FV antigen levels in 4 patients with LMAN1 mutations and 3 patients with MCFD2 mutations, as well as 15 healthy controls. Platelet FV levels are reduced to the same extent as FV levels in plasma, for both the LMAN1 group (15% of control for plasma and 13% for platelets) and the MCFD2 group (8% of control for both plasma and platelets; Table 4).

Here we report 4 MCFD2 and 5 LMAN1 mutations. The mutation in family B19 appears to be the result of a duplication event, with insertion of a GA dinucleotide in a region that normally contains a string of 3 GAs. This mutation results in a frameshift that deletes both EF hand domains of MCFD2. The mutation in family B22 (D81Y) is the fifth missense mutation identified in MCFD2, and the second identified in the first EF hand domain. In the current report, we demonstrate that this mutation, as well as 2 other missense mutations identified previously, disrupts the binding to LMAN1 (Figure 1). Taken together with results of previously reported D129E and I136T mutations, all missense mutations identified to date are located in the EF hand domains and result in loss of LMAN1 binding. The 2 mutations identified in families B23 to B25 are consistent with the previously reported founder effect.²  To date, all F5F8D patients in Jewish families have been found to carry either the c.1149+2T>G (Tunisian Jews) or the c.89-90insG mutation (Middle Eastern Jews), with the combined prevalence of one of these mutations in these populations estimated at 1:100 000.^12,28  The c.2T>C mutation in LMAN1 identified in family B26 has been reported only in patients of Italian origin; therefore, it likely represents a founder allele in the Italian population. The mutation in family B27 (c.149+5G>A) is thus far the most frequently identified mutation in MCFD2. Previous studies have shown that this is a recurring mutation that arose independently in different geographic regions.^6,11 

The difference observed between LMAN1 and MCFD2 patients in the distribution of plasma levels for FV and FVIII is unlikely due to a systematic error in measurement as these data were collected at multiple centers around the world. However, a consistent trend toward high or low values from one of the centers contributing a relatively large group of patients could potentially bias the results, particularly given the smaller sample size for the MCFD2 group. Of note, most of the patients in a recent report from India (included in our analysis) have MCFD2 mutations,¹¹  with measurements clustered at the high end of the range for both FV and FVIII. However, if data from these patients are removed, the difference in levels between the MCFD2 and LMAN1 groups increases even further in significance (mean values of 8.4 vs 13.7 for FV [P < .001] and 9.3 vs 16.0 for FVIII [P < .001]). A recent report of an additional 9 F5F8D patients from 5 families is notable for levels in the lower end of the range, particularly for FVIII (at 1% or lower for 6 of 9 patients).³⁵  Mutations have not been identified in these patients. In addition to laboratory-specific bias in factor activity measurements and biologic differences between MCFD2 and LMAN1, other factors may contribute to variations in plasma FV and FVIII levels among certain groups of patients, including common genetic modifiers or environmental factors unique to the corresponding population. However, sex and ABO blood type do not appear to be important modifying factors (Figure 3). Interestingly, FVIII levels in F5F8D patients with type O blood are not further reduced compared with non-O blood types.

Platelet FV levels are reduced to the same extent as plasma FV levels, for both the LMAN1 group and the MCFD2 group (Table 4). A prior study on 2 F5F8D patients is consistent with our results, demonstrating reduced factor Xa binding in platelet releasates to approximately 40% of the healthy control.³⁶  The platelet FV pool in humans is derived via endocytosis of plasma FV,^18,32  whereas in the mouse, platelet FV is synthesized de novo in the megakaryocytes.^33,34  Recent studies suggest that endocytosis of FV is a specific, clathrin-dependent, and probably receptor-mediated mechanism.³⁷  Our observations of similarly reduced FV levels in platelets and plasma of F5F8D patients suggest that the receptor for uptake of plasma FV into megakaryocytes is not saturated at physiologic concentrations of plasma FV. Alternatively, a mechanism may exist that regulates the rate of FV uptake into megakaryocytes to match the plasma FV level. Consistent with either of these models, the amount of intracellular FV in cultured megakaryocytes has been shown to be dependent on the exogenous factor V concentrations in the culture media.³⁸  Our results would predict that other conditions that alter plasma FV level should also alter the platelet level, including severe liver diseases or autoimmune clearance of plasma FV. Although a number of patients with FV inhibitor antibodies have been described,^39-42  bleeding symptoms are highly variable, and direct measurement of platelet FV antigen in this setting has not been reported.

The finding that MCFD2 mutations are generally associated with lower levels of FV and FVIII suggests that MCFD2 may play a more direct role in transporting FV and FVIII, perhaps initiating the interaction with these 2 cargo proteins in the ER and/or recruiting them to ER exit sites. Consistent with this hypothesis, MCFD2 with a missense mutation (D129E) that disrupts the LMAN1-MCFD2 interaction can still be cross-linked to FVIII.¹⁰  The lectin activity of LMAN1 may function to further stabilize the receptor-cargo complex by binding to the sugar residues of FV and FVIII. Alternatively, LMAN1 may merely function to carry MCFD2 and FV/FVIII cargo to the ER exit sites for packaging into COPII vesicles. In this model, LMAN1 deficiency would act only through the resulting secondary MCFD2 deficiency and higher levels of FV/FVIII in LMAN1^−/− patients could reflect FV/FVIII transport by the small, residual pool of intracellular MCFD2. A recent study suggests that MCFD2 is dispensable for the interaction of LMAN1 with the lysosomal enzymes cathepsin C and cathepsin Z.⁴³  Thus, MCFD2 could function as a selective cargo adaptor for FV and FVIII, with LMAN1 deficiency resulting in subclinical deficiencies for additional cargo proteins that directly interact with LMAN1 or are linked to adaptors other than MCFD2. This hypothesis awaits experimental confirmation.

The authors thank Dr MariaTeresa Bajetta and Dr Rossana Lombardi for their assistance during the experiments performed in Milan, and the Genomic Medicine Biorepository at Cleveland Clinic Foundation for processing some patient samples.

This work was partially supported by grants from the National Institutes of Health (PO1 HL057346 [D.G., R.J.K., B.Z.], R37 HL039693 [D.G.], and HL052173 [R.J.K.]), a Career Development Award from the National Hemophilia Foundation (B.Z.), Fondazione Italo Monzino, Milan, Italy (POC(2005-2007)BS-1/2005), and Telethon Foundation, Italy (grant nos. GGP030261 and GGP06155). R.J.K. and D.G. are investigators of the Howard Hughes Medical Institute.

National Institutes of Health

Contribution: B.Z., M.S., C.Z., and A.Y. performed experiments; B.Z., F.P., and D.G. designed the research and analyzed results; B.Z. made the figures; P.P. performed statistical analyses; B.Z., M.S., F.P., and D.G. wrote the paper; other authors provided vital new reagents and clinical data, and critical review of the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Bin Zhang, Genomic Medicine Institute, Lerner Research Institute, Cleveland Clinic Foundation, Cleveland, OH 44195; e-mail: zhangb@ccf.org.