From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
Mild Hemophilia A Caused by Increased Rate of Factor VIII A2 Subunit
Dissociation: Evidence for Nonproteolytic Inactivation of Factor VIIIa In Vivo
By S.W. Pipe, A.N. Eickhorst, S.H. McKinley, E.L. Saenko, and R.J. Kaufman
Approximately 5% of hemophilia A patients have normal
amounts of a dysfunctional factor VIII (FVIII) protein and are
termed cross-reacting material (CRM)-positive. FVIII is a
heterodimer (domain structure A1-A2-B/A3-C1-C2) that requires thrombin cleavage to elicit procoagulant activity.
Thrombin-activated FVIII is a heterotrimer with the A2
subunit (amino acid residues 373 to 740) in a weak ionic
interaction with the A1 and A3-C1-C2 subunits. Dissociation
of the A2 subunit correlates with inactivation of FVIII.
Recently, a phenotype of CRM-positive hemophilia A patients has been characterized whose plasma displays a
discrepancy between their FVIII activities, where the onestage clotting assay displays greater activity than the twostage clotting assay. One example is a missense mutation
where ARG531 has been substituted by HIS531. An FVIII cDNA
construct was prepared containing the ARG531HIS mutation
and the protein was expressed in COS-1 monkey cells by
transient DNA transfection. Metabolic labeling with [35S]methionine demonstrated that ARG531HIS was synthesized at
an equal rate compared with FVIII wild-type (WT) but had
slightly reduced antigen in the conditioned medium, suggesting a modest secretion defect. A time course of structural
cleavage of ARG531HIS demonstrated identical thrombin cleavage sites and rates of proteolysis as FVIII WT. Similar to the
patient phenotypes, ARG531HIS had discrepant activity as
measured by a one-stage activated partial thromboplastin
time (aPTT) clotting assay (36% 6 9.6% of FVIII WT) and a
variation of the two-stage assay using a chromogenic substrate (COAMATIC; 19% 6 6.9% of FVIII WT). Partially purified
FVIII WT and ARG531HIS proteins were subjected to functional
activation by incubation with thrombin. ARG531HIS demonstrated significantly reduced peak activity and was completely inactivated after 30 seconds, whereas FVIII WT retained activity until 2.5 minutes after activation. Because the
ARG531HIS missense mutation predicts a charge change to the
A2 subunit, we hypothesized that the ARG531HIS A2 subunit
could be subject to more rapid dissociation from the heterotrimer. The rate of A2 dissociation, using an optical biosensor,
was determined to be fourfold faster for ARG531HIS compared
with FVIII WT. Because the two-stage assay involves a
preincubation phase before assay measurement, an increased rate of A2 dissociation would result in an increased
rate of inactivation and reduced specific activity.
r 1999 by The American Society of Hematology.
P
ment (A3-C1-C2) through a monovalent copper ion-dependent
linkage between the A1 and A3 domains.6
Thrombin activates FVIII through proteolytic cleavage after
ARG740, ARG1689, and ARG372, generating an activated FVIII
(FVIIIa) heterotrimer consisting of the A1 subunit (50 kD) in a
copper ion-dependent association with the thrombin-cleaved
light chain (73 kD) and a free A2 subunit associated with the A1
domain through a weak ionic interaction.7-13 An acidic amino
acid-rich region at the carboxy-terminus of the A1 subunit most
likely interacts with positively charged residues within the A2
subunit to retain the heterotrimeric FVIIIa configuration.14-16
FVIII activity is measured by either a one-stage or two-stage
procedure.17,18 The one-stage assay is based on the ability of
FVIII-containing samples to correct the prolonged activated
partial thromboplastin time (aPTT) of FVIII-deficient plasma.19
The two-stage assay uses the same principle as the one-stage
method, yet is split into two distinct phases.20 In the first phase,
dilutions of FVIII are incubated with factor IXa, factor X, Ca21,
and phospholipid. Activation of FVIII during the first phase
allows it to exhibit its procoagulant activity as a cofactor for
factor IXa, leading to the generation of factor Xa. The
incubation mixture is then subsampled into a second tube
containing Ca21 and a source of prothrombin and fibrinogen,
and the time to fibrin formation is recorded. A modification of
the two-stage assay has also been used in which the second
reaction is incubated with a factor Xa-sensitive chromogenic
substrate where the rate or degree of color change is proportional to the amount of FVIII added in the first stage of the
assay.21
Hemophilia A is a heterogeneous disorder, with severe
phenotypes associated with major disruptions of the FVIII gene.
Patients with severe phenotypes usually have FVIII antigen
levels that are undetectable and are termed cross-reacting
material (CRM)-negative. Other genetic mutations are associ-
LASMA COAGULATION factor VIII (FVIII) functions
within the blood coagulation cascade as a cofactor for
factor IXa in the proteolytic activation of factor X to factor Xa.
A quantitative or qualitative deficiency of FVIII leads to the
phenotype of the bleeding disorder hemophilia A. FVIII is
synthesized as a single-chain polypeptide of approximately 280
kD with the domain structure A1-A2-B-A3-C1-C2.1-3 The A
domains share 35% to 40% amino acid identity and are
homologous to the A-domains of ceruloplasmin. The C domains
also display 35% to 40% amino acid identity to each other and
are homologous to phospholipid-binding proteins, suggesting a
role in phospholipid interactions.1,2,4,5 The B domain shares no
significant homology with any known protein. Intracellular
proteolytic processing within the B domain after residues
ARG1313 or ARG1648 forms a heterogeneous FVIII heterodimer
consisting of approximately 90-to 220-kD heavy chain fragments (A1-A2-B) associated with an 80-kD light chain fragFrom the Departments of Pediatrics and Biological Chemistry,
Howard Hughes Medical Institute, University of Michigan Medical
Center, Ann Arbor, MI; and the Holland Red Cross Laboratory,
Rockville, MD.
Submitted May 11, 1998; accepted August 24, 1998.
Supported by National Institutes of Health (NIH) Grant No. HL52173
and by National Institute of Child Health and Human Development
(NICHD) HD28820.
Address reprint requests to R.J. Kaufman, PhD, Howard Hughes
Medical Institute, University of Michigan Medical Center, 4570 MSRB
II, 1150 W Medical Center Dr, Ann Arbor, MI 48109-0650; e-mail:
[email protected].
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be hereby marked ‘‘advertisement’’ in accordance with 18 U.S.C. section 1734 solely to indicate
this fact.
r 1999 by The American Society of Hematology.
0006-4971/99/9301-0008$3.00/0
176
Blood, Vol 93, No 1 (January 1), 1999: pp 176-183
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
HEMOPHILIA DUE TO INCREASED A2 SUBUNIT DISSOCIATION
ated with inefficient secretion, with patient plasmas having
concomitantly reduced FVIII antigen and activity levels, and
are termed CRM-reduced. Approximately 5% of hemophilia A
patients have considerable FVIII antigen levels (at least 30% of
normal), but FVIII activity levels are significantly reduced,
suggesting a protein dysfunction.22 Interestingly, approximately
40% of the CRM-positive and CRM-reduced hemophilia A
patients contain missense mutations within the A2 domain.23
The A2 domain, representing only approximately 10% of the
entire amino acid sequence of FVIII, therefore contains a
significant clustering of missense mutations resulting in hemophilia A highlighting its functional importance for FVIII
procoagulant activity.
The study of missense mutations has contributed significantly
to our understanding of FVIII structure-function relationships
and the pathophysiology contributing to the hemophilia A
disease phenotype. Missense mutations have been identified at
thrombin cleavage sites critical for functional activation and at
residues important in interaction with von Willebrand factor
(vWF).24-28 However, to date, only two FVIII A2 domain
missense mutations have been characterized as to their mechanism leading to protein dysfunction.27,29 In both cases, the
reduced specific activity could be attributed to reduced interaction with factor IXa.
Recent reports have identified a CRM-positive hemophilia A
phenotype in which the patient plasmas exhibit a familial
discrepancy in which the FVIII procoagulant activity is higher
when measured in a one-stage assay compared with a two-stage
assay.30-36 This discrepancy is unusual, because one-stage and
two-stage assays have been used interchangeably for some time
as a standard determination of FVIII activity and the majority of
patients have similar results as measured by either method.18 In
this report, we have studied the mechanistic basis for one of
these patient phenotypes in which an A2 domain missense
mutation results in substitution of a histidine for arginine at
residue 531.31 Using site-directed mutagenesis, the ARG531HIS
mutation was generated within the FVIII cDNA and the protein
was functionally characterized after expression in transiently
transfected COS-1 cells. ARG531HIS demonstrated only a modest
secretion defect, had reduced specific activity, and had discrepant FVIII activity as measured by either the one-stage or
two-stage methods. Upon thrombin activation, the ARG531HIS
A2 subunit exhibited a fourfold increased rate of dissociation
from the A1/A3-C1-C2 heterodimer. The increased instability
of the ARG531HIS heterotrimer would reduce its specific activity
in both one-stage and two-stage assays but would result in an
increased disadvantage in the two-step procedure due to the
incubation phase in the first step. This hemophilia A phenotype
therefore supports previous in vitro studies that have suggested
that nonproteolytic regulation of FVIIIa activity, via spontaneous A2 subunit dissociation, is important in vivo.
MATERIALS AND METHODS
Materials. Anti-heavy chain FVIII monoclonal antibody (MoAb;
F-8) conjugated to CL-4B Sepharose was a gift from Debra Pittman
(Genetics Institute Inc, Cambridge, MA). FVIII-deficient and normal
pooled human plasma were obtained from George King Biomedical, Inc
(Overland Park, KS). Activated partial thromboplastin (automated
aPTT reagent) and CaCl2 were purchased from General Diagnostics
177
Organon Teknika Corp (Durham, NC). Anti-light chain FVIII MoAbs,
ESH-4 and ESH-8, were purchased from American Diagnostica, Inc
(Greenwich, CT). Human thrombin and aprotinin were purchased from
Boehringer, Mannheim GmbH (Mannheim, Germany). [35S]-methionine (.1,000 µCi/mmol) was purchased from Amersham Corp (Arlington Heights, IL). En3Hance was purchased from Dupont (Boston, MA).
Dulbecco’s modified Eagle’s medium (DMEM), methionine-free
DMEM, fetal bovine serum, biotin N-hydroxy succinimide ester, and
streptavidin-horseradish peroxidase conjugate were purchased from
GIBCO BRL (Gaithersburg, MD). COAMATIC was purchased from
DiaPharma (West Chester, OH).
Plasmid mutagenesis. Mutagenesis was performed within the mammalian expression vector pMT237 containing the FVIII cDNA
(pMT2VIII). Oligonucleotide-directed mutagenesis was used to create a
Spe I-Kpn I polymerase chain reaction fragment in which codon 531
was mutated from CGC to CAC, predicting an amino acid substitution
of histidine for arginine, and was ligated into Spe I-Kpn I–digested
pMT2VIII. The resulting mutant plasmid was designated ARG531HIS.
The plasmid containing the wild-type FVIII cDNA sequence was
designated FVIII WT. All plasmids were purified by centrifugation
through cesium chloride and characterized by restriction endonuclease
digestion and DNA sequence analysis.
DNA transfection and analysis. Plasmid DNA was transfected into
COS-1 cells by the diethylaminoethyl (DEAE)-dextran method as
previously described.38 Conditioned medium was harvested at 64 hours
posttransfection in the presence of 10% fetal bovine serum. Protein
synthesis and secretion were analyzed by metabolically labeling cells at
64 hours posttransfection for 30 minutes with [35S]-methionine (300
µCi/mL in methionine-free medium), followed by a chase for 4 hours in
medium containing 100-fold excess unlabeled methionine and 0.02%
aprotinin. Cell extracts and conditioned medium were harvested and
immunoprecipitations were performed and analyzed as described
previously.38
Protein purification. Partially purified ARG531HIS protein was obtained from 200 mL of conditioned medium from transiently transfected
COS-1 cells by immunoaffinity chromatography,39 yielding 750 to
1,500 ng per purification. FVIII WT protein was purified in parallel
from stably transfected Chinese hamster ovary cells. The proteins eluted
into the ethylene glycol-containing buffer were dialyzed and concentrated against a polyethylene glycol (molecular weight, ,15,000 to
20,000) -containing buffer14 and stored at 270°C.
FVIII activity and antigen assay. FVIII activity was measured by
(1) one-stage aPTT clotting assay on an MLA Electra 750 fibrinometer
(Medical Laboratory Automation, Inc, Pleasantville, NY) by reconstitution of human FVIII-deficient plasma or (2) by modified two-stage assay
using the COAMATIC chromogenic assay according to the manufacturer’s
instructions. For thrombin activation, protein samples were diluted into 50
mmol/L Tris-HCl, pH 7.5, 150 mmol/L NaCl, 2.5 mmol/L CaCl2, and 5%
glycerol and incubated at room temperature with 1 U/mL thrombin.
After incubation for increasing periods of time, aliquots were diluted
and assayed for FVIII activity. One unit of FVIII activity is the amount
measured in 1 mL of normal human pooled plasma. FVIII antigen was
quantified using a sandwich enzyme-linked immunosorbent assay
(ELISA) method using anti-light chain antibodies ESH-4 and ESH-8.40
Recombinant FVIII protein purified in parallel was used as a standard.
Kinetic measurements using biosensor technology. The kinetics of
the A2 subunit dissociation from thrombin-activated FVIII WT and
ARG531HIS was studied by surface plasmon resonance using the IASys
biosensor (Fisons, Cambridge, UK), which measures protein binding
and subsequent dissociation in real time.41 Binding of 1 ng of protein
per square millimeter of the biosensor chip produces a resonance signal
of 200 Arc seconds. MoAb ESH8 (50 µg/mL) in 10 mmol/L sodium
acetate, pH 5.0, was covalently coupled to the activated carboxymethyldextran-coated biosensor cuvette via amino groups using succinimide
ester chemistry.41 The carboxymethyldextran chip and reagents for its
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
178
PIPE ET AL
activation, N-ethyl-N8-(dimethylaminopropyl)carbodiimide hydrochloride and N-hydroxysuccinimide, and deactivation, ethanolamine, were
purchased from Fisons. FVIII WT and ARG531HIS binding to ESH8, their
dissociation from the antibody, and activation by thrombin were
measured in 200 µL of 20 mmol/L HEPES, 0.15 mol/L NaCl, 5 mmol/L
CaCl2, 0.01% Tween 20, pH 7.4. The chip was regenerated by the
addition of 0.1 mol/L glycine, pH 3.0, for 3 minutes, resulting in
complete dissociation of FVIII proteins from the capture ESH8.
Identical signals for reference binding of the FVIII proteins to
immobilized ESH8 were obtained before and after regeneration.
The values of the rate constants for dissociation (koff) of FVIII WT or
ARG531HIS from immobilized ESH8 and those for dissociation of the A2
subunits upon thrombin activation of immobilized FVIII proteins were
determined by fitting the dissociation kinetics data to the following
equation describing a single-phase dissociation process: dR/dt 5
2koffR,42 where R is observed surface plasmon resonance signal. The
fitting procedure was performed using Sigmaplot 1.02 software (Jandel
Scientific, San Raphael, CA).
RESULTS
Synthesis and secretion of ARG531HIS. The synthesis and
secretion of FVIII WT and ARG531HIS was compared by
transient DNA transfection into COS-1 monkey cells. At 64
hours posttransfection, the rates of synthesis were analyzed by
immunoprecipitation of cell extracts from [35S]-methionine
pulse-labeled cells. Intracellular FVIII WT and ARG531HIS were
detected in their single chain forms and migrated at approximately 250 kD (Fig 1, lanes 3 and 5). ARG531HIS exhibited
similar band intensity to the FVIII WT, suggesting that the
missense mutation did not interfere with efficient protein
synthesis. After a 4-hour chase period, FVIII WT was lost from
the cell extract (Fig 1, lane 4) and was recovered from chase
conditioned medium as a 280-kD single chain, a 200-kD heavy
chain, and an 80-kD light chain (Fig 1, lane 8). ARG531HIS was
also lost from the cell extract over the 4-hour chase period
(Fig 1, lane 6), similar to FVIII WT, but had a reduced recovery
from the chase-conditioned medium (Fig 1, lane 9), suggesting
a modest secretion defect. ARG531HIS was also detected within
the chase-conditioned medium in single-chain, heavy-chain,
and light-chain forms of identical molecular mobility as FVIII
WT, suggesting similar posttranslational processing and proper
heavy and light chain association. From the relative band
intensities within the chase-conditioned medium, ARG531HISsecreted protein was determined to be 56% of FVIII WT. FVIII
antigen determinations by ELISA were performed on unlabeled
conditioned media collected from 24 to 64 hours posttransfection. ARG531HIS protein was detected at 46% and 76% of FVIII
WT in two independent transfection experiments, consistent
with the modest secretion defect indicated by the pulse-chase
analysis.
Recombinant-derived ARG531HIS protein demonstrates a similar functional phenotype to patient plasmas. Conditioned
medium was collected from COS-1 cells transiently expressing
FVIII WT and ARG531HIS from 24 to 64 hours posttransfection.
FVIII activity (Fig 2) was measured by a one-stage aPTT
clotting assay or by a modified two-stage method using the
COAMATIC chromogenic assay. Similar to the patient phenotypes reported previously, ARG531HIS had discrepant activity as
measured by the one-stage aPTT clotting assay (36% 6 9.6% of
FVIII WT) as compared with the two-stage chromogenic assay
(COAMATIC; 19% 6 6.9% of FVIII WT). The activity
measurements for the patient plasmas represent all of the
Fig 1. Synthesis and secretion of FVIII WT and ARG531HIS expressed
in COS-1 cells. FVIII WT and ARG531HIS plasmids were transfected into
COS-1 monkey cells. At 64 hours posttransfection, cells were pulselabeled with [35S]-methionine for 30 minutes and cell extracts were
harvested. Duplicate labeled cells were chased for 4 hours in medium
containing excess unlabeled methionine and then cell extracts and
conditioned medium were harvested. Equal proportionate volumes
of cell extract (lanes 1 through 6) and conditioned medium (lanes 7
through 9) were immunoprecipitated with anti–FVIII-specific antibody and equal aliquots were analyzed by SDS-PAGE. Mock indicates
cells that did not receive plasmid DNA. Cell extract pulse (P) and
chase (C). The migration of FVIII from the cell extracts is indicated at
the right by an arrow. FVIII from the conditioned medium is indicated
at the far right as single-chain (SC), heavy-chain (HC), and light-chain
(LC) forms. Molecular weight markers are shown on the left.
reported data obtained from the HAMSTeRS hemophilia A
mutation database.43 The slightly higher two-stage activity
results obtained with the recombinant-derived ARG531HIS are
consistent with those obtained from the few reported patient
plasmas in which a chromogenic assay rather than the classical
two-stage assay was used.31 After immunoaffinity purification
of FVIII WT and ARG531HIS from the conditioned medium,
similar results were obtained. Immunoaffinity purified ARG531HISspecific activity was 62% of FVIII WT by one-stage aPTT
clotting assay as compared with 32% of FVIII WT as determined by the two-stage chromogenic assay.
ARG531HIS demonstrates similar thrombin proteolysis compared with FVIII WT. [35S]-methionine–labeled FVIII WT
and ARG531HIS proteins were immunoprecipitated from chase
conditioned medium of transiently expressing COS-1 cells
labeled at 60 hours posttransfection. After Triton X-100 washes
as described, the immunoprecipitated complexes were incubated with thrombin (0.1 U/mL) for increasing periods of time
at 37°C before sodium dodecyl sulfate-polyacrylamide gel
electrophoresis (SDS-PAGE) analysis. Both FVIII WT and
ARG531HIS were initially detected in their 280-kD single-chain
forms, and dimers of a 200-kD heavy chain in association with
an 80-kD light chain (Fig 3, lanes 2 and 8). Both FVIII WT and
ARG531HIS were sequentially cleaved into a heterotrimer of
fragments consistent with a 50-kD A1 subunit, 43-kD A2
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
HEMOPHILIA DUE TO INCREASED A2 SUBUNIT DISSOCIATION
Fig 2. Recombinant-derived ARG531HIS demonstrates a similiar
phenotype to described ARG531HIS patient plasmas. Plasma activity
data were derived from the HAMSTeRS hemophilia A mutation
database for all reported patients identified with ARG531HIS for which
both one-stage (j) and two-stage (h) results were available. The data
from recombinant-derived ARG531HIS were obtained from assaying the
activity in the conditioned medium from four independent transfection experiments. Activity for ARG531HIS patient plasmas and recombinant-derived ARG531HIS is presented as the percentage of wild-type
(normal plasma or recombinant FVIII WT, respectively).
subunit, and 73-kD thrombin-cleaved light chain, A3-C1-C2
(Fig 3, lanes 3 through 7 and 9 through 13). A 90-kD fragment
appeared immediately after incubation with thrombin consistent
with an A1-A2 heavy chain fragment due to cleavage after
arginine 740. With higher concentrations of thrombin, this
fragment is further cleaved into 50-kD A1 and 43-kD A2
subunits (data not shown). The similar pattern of electrophoretic
mobility and rate of appearance of proteolytic fragments over
time suggests that ARG531HIS, compared with FVIII WT, has
Fig 3. Thrombin proteolysis of FVIII WT compared with ARG531HIS. [35S]-methionine labeled FVIII
WT and ARG531HIS proteins immunoprecipitated from
the chase conditioned medium of transiently expressing COS-1 cells were divided into equal aliquots and
incubated with thrombin (0.1 U/mL) for increasing
periods of time at 37°C. Reactions were terminated
with SDS-PAGE sample buffer and protein fragments
were separated by 10% SDS-PAGE. Time is in minutes, with 0 representing the absence of thrombin.
Mock indicates medium from cells that did not
receive plasmid DNA. FVIII protein forms are indicated at the right as follows: SC, single chain; HC,
heavy chain; LC, light chain; A11A2, A1, and A2,
thrombin-cleaved heavy chain fragments; LC1IIa,
thrombin-cleaved light chain; FVIIIa, predicted thrombin-activated FVIII heterotrimer. Molecular weight
markers (m) are indicated on the left.
179
identical sites of thrombin cleavage and sensitivity to proteolysis.
ARG531HIS exhibits reduced peak activity and increased rate of
inactivation after functional activation by thrombin. Having
demonstrated a similar pattern and sensitivity of ARG531HIS to
thrombin cleavage, the functional consequence of the ARG531HIS
missense mutation on activation and inactivation was examined
in an in vitro functional assay. Equal concentrations of immunoaffinity purified FVIII WT and ARG531HIS were incubated with
thrombin and assayed for FVIII activity by a one-stage aPTT
clotting assay (Fig 4). Upon treatment with thrombin, FVIII WT
was maximally activated within 10 seconds and then inactivated
over the next 5 minutes. ARG531HIS also reached peak activity
after 10 seconds of incubation with thrombin, but at approximately fivefold lower activity compared with FVIII WT. In
addition, ARG531HIS was completely inactivated after 30 seconds
incubation with thrombin, suggesting some increased instability
of the thrombin-activated heterotrimer.
ARG531HIS exhibits increased rate of A2 subunit dissociation
from the thrombin-activated heterotrimer. Because the A2
subunit is retained within the thrombin-activated FVIIIa heterotrimer through a weak electrostatic interaction with the
acidic amino acid rich region at the carboxy terminus of the A1
subunit, we hypothesized that the charge change resulting from
the ARG531HIS mutation would make its A2 subunit susceptible
to more rapid dissociation. The relative rates of A2 dissociation
for FVIII WT and ARG531HIS were determined using an optical
biosensor (Fig 5). An anti-light chain antibody, ESH8, was
covalently immobilized on a carboxymethyldextran-coated biosensor chip. Similar amounts (1.12 ng/mm2) of immunoaffinitypurified FVIII WT or ARG531HIS protein (2.5 nmol/L) were
bound to ESH8 antibody. Unbound material was removed by
washing with buffer and free (nonproteolytic) dissociation from
antibody was measured. The values of dissociation rate constants, koff 5 (8.9 6 0.23) 3 1025 s21, were similar for FVIII
WT and ARG531HIS. Subsequently, thrombin was added to a final
concentration of 1 U/mL. Because the FVIII preparations are
bound to an ESH8-coated chip via their light chains, this
interaction will not be disturbed by thrombin activation and the
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
180
PIPE ET AL
registered by the optical biosensor. The koff values for the FVIII
WT and ARG531HIS A2-subunits were (6.7 6 0.6) 3 1023 s21
and (2.9 6 0.22) 3 1022 s21, respectively. The half lives of
A2-dissociation for FVIII WT and ARG531HIS calculated as
ln2/koff are 103 and 24 seconds, respectively. Because A2dissociation correlates with inactivation, the ARG531HIS thrombinactivated heterotrimer would be expected to inactivate fourfold
faster compared with FVIII WT. Similar results were obtained
via the same method, substituting an anti-A2 domain antibody
(MoAb 413; Holland Red Cross Laboratory) for immobilization
of the FVIII proteins (data not shown). The single exponential
dissociation of the A2 subunit observed for FVIII WT in this
study is comparable to the decay of FVIIIa previously reported.14,44
DISCUSSION
Fig 4. Activation and inactivation of FVIII WT and ARG531HIS by
thrombin. Immunoaffinity purified FVIII WT (h) and ARG531HIS (d)
proteins (0.5 nmol/L) were incubated with thrombin (1 U/mL) at
room temperature and assayed over time for FVIII activity by aPTT.
The results are from a single thrombin activation experiment and are
typical of multiple independent experiments.
A1 subunit will remain associated with the light chain through
the copper ion-dependent linkage between the A1 and A3
domains. Thus, the thrombin-induced release of the A2-subunit
from the heterotrimer can be measured as a dissociation curve
Fig 5. Determination of the kinetic parameters for nonproteolytic
and thrombin-mediated dissociation of FVIII WT and ARG531HIS from
MoAb ESH8. MoAb ESH8 was covalently immobilized to a biosensor
chip at 20 ng/mm2. FVIII WT or ARG531HIS (2.5 nmol/L) were bound to
ESH8 at 1.12 ng/mm2. A resonance response of 200 Arc seconds corresponds to 1 ng of protein bound per square millimeter of the biosensor
chip surface. The kinetics of FVIII WT or ARG531HIS nonproteolytic
dissociation from ESH8 was recorded after replacement of the ligand
by dissociation buffer (at arrow). At the second arrow, thrombin (1
U/mL) was added and thrombin-mediated dissociation of the A2 subunit
from immobilized dimers was followed. The koff values for nonproteolytic and thrombin-mediated dissociation were derived from dissociation kinetic curves as described under Materials and Methods.
The one-stage and two-stage methods for assaying FVIII
activity have been used clinically for more than 40 years.
Several studies have demonstrated little differences in precision
between the two methods.18 The one-stage assay, being technically simpler and easier to automate, has typically been the
method of choice in recent years. Discrepancy in FVIII activity
results have only been described under a few unique conditions.
These have included the measurement of FVIII concentrates
against plasma standards and the measurement of concentrates
against concentrate standards and in the assessment of in vivo
recovery of FVIII concentrate infusions.18 The discrepancies
have included higher two-stage versus one-stage activities or
the converse as investigated in this report. An explanation for
these discrepancies has been elusive, although the Al(OH)3
adsorption step, which is used in the preparation of samples in
the two-stage assay, has been suggested as a major cause of the
discrepancy in assays of concentrates versus plasma. However,
this discrepancy was an average of only 26% higher activity in
the two-stage assay compared with the one-stage assay.45 Other
suggested causes have included the choice of predilution buffer,
thrombin activation of the samples, nonspecific contaminants
such as lipids, and the influence of the presence of vWF.18 In the
latter case, a patient with von Willebrand’s disease demonstrated a discrepancy in FVIII activity in which, even though the
plasma FVIII activity measured by a one-stage assay was
consistent with the antigen level, there was a 75% decrease in
the plasma FVIII activity when a two-stage assay was used.35
The patient was later characterized as having a von Willebrand
disease Normandy defect in which there was a weaker interaction of FVIII with the mutant vWF leading to increased
susceptibility of the FVIII to adsorption by Al(OH)3.36 The
apparent discrepancy was corrected by adding hemophilic
plasma or purified hemophilic vWF.
Several recent reports have now observed these discrepancies
as part of a hemophilia A phenotype. The patients have typically
been mild in phenotype and the discrepancy, a more than
twofold higher activity result by the one-stage assay versus the
two-stage, was observed in all affected family members.31 This
phenotypic subgroup was subjected to DNA analysis to determine a responsible FVIII gene mutation.31 Several single
missense mutations were identified either within the FVIII A1,
A2, or A3 domains. The best characterized missense mutation
was the ARG531HIS mutation chosen for this study. Three
independent reports collected by the hemophilia A mutation
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
HEMOPHILIA DUE TO INCREASED A2 SUBUNIT DISSOCIATION
database have shown that the patients with the ARG531HIS
mutation have FVIII antigen levels from approximately 30% to
100% of normal, suggesting at worst a modest protein secretion
defect.43 All the patients analyzed by both one-stage and
two-stage assays demonstrated at least twofold higher results in
the one-stage assay. Where a chromogenic based assay was also
used, there was still an approximate twofold difference in the
activities, although the chromogenic two-stage results were
somewhat higher than those for the classical two-stage assay.31
The ARG531HIS mutation does not lie within major identified
functional FVIII epitopes, such as the terminal portion of the C2
domain containing the binding site for phospholipids,46,47 the
vWF binding sites (residues 1648-1689 of the A3 domain and
epitopes within the terminal C2 domain),48,49 or the proposed
factor IXa interaction site within the A2 domain (residues
558-565).50,51 Therefore, this particular missense mutation may
require an alternative mechanism for dysfunction not yet
characterized. Insights into the differences between the onestage and two-stage assays allow a hypothesis as to a possible
mechanism. The two-stage assay, divided into two separate
steps, requires a prolonged phase in the first step under
conditions in which FVIII becomes activated and generates the
predicted FVIIIa heterotrimeric structure to exert its cofactor
function with factor IXa. Previous studies have highlighted the
instability of the FVIIIa heterotrimer. The FVIIIa heterotrimer
exhibits a pH-dependent dissociation of the A2 subunit from the
A1/A3-C1-C2 heterodimer that correlates with loss of procoagulant activity.14 The data supporting the idea that the amino acid
region 558 to 565 within the A2 subunit represents a factor IXa
interaction site are consistent with this observation. Porcine
FVIIIa exhibits an increased affinity for its A2 subunit compared with human FVIIIa and, accordingly, demonstrates an
increased specific activity.44,52,53 A genetically engineered recombinant FVIII molecule in which the A2 subunit remains
covalently linked to the heterodimer after activation by thrombin exhibited a fivefold increase in specific activity.54 These
observations highlight the role of A2 dissociation in limiting
FVIIIa activity. The observations are consistent with positively
charged residues within the A2 subunit interacting with acidic
amino acid residues at the carboxy-terminus of the A1 subunit
maintaining a weak electrostatic interaction to preserve the
FVIIIa heterotrimer. Loss of residues 337-372 of the A1 subunit
after cleavage by activated protein C or further proteolysis by
thrombin leads to rapid inactivation of FVIIIa via A2 subunit
dissociation.14,15,55,56 A genetically engineered mutant of recombinant FVIII, containing ARG336ILE, was resistant to proteolytic
cleavage by thrombin and activated protein C and had an
increased specific activity, as determined by a chromogenic
assay, attributable to increased stability of the heterotrimer.12
Many of the amino acid substitutions in the porcine compared
with the human A2 subunit lead to a charge alteration and may
be responsible for the increased stability of the porcine FVIIIa
heterotrimer. Because the two-stage assay involves the preincubation phase before subsampling, an FVIII protein with increased heterotrimeric stability would be predicted to have an
increased activity in the two-stage assay. Consistent with this
finding, when porcine FVIII concentrates were assayed against
human concentrates by the classical two-stage method, the
activity results were two to three times higher than by one-stage
assays.18
181
However, the phenotype characterized in relation to the
missense mutation is one in which not only the
one-stage activity is reduced from that expected by the antigen
levels present, but also the two-stage assay activity results are at
least twofold lower than by the one-stage assay. The ARG531HIS
mutation predicts a loss of charge within the A2 subunit
predicting a weakened electrostatic interaction after thrombin
activation. Thus, FVIIIa dissociation of a ARG531HIS A2 subunit,
compared with FVIII WT, would be hypothesized to be more
rapid. Because the recombinant-derived ARG531HIS protein exhibited a similar phenotype to the patient plasmas, we were able to
analyze the purified protein for its rate of A2 dissociation. The
data using the optical biosensor confirmed the fourfold increased rate of A2 subunit loss after thrombin activation of
ARG531HIS. The reduced peak activity observed for the ARG531HIS
protein compared with FVIII WT can also be attributed to the
increased rate of A2 subunit dissociation. Under the conditions
of the aPTT clotting assay, thrombin was added to the FVIII
samples in buffer and the thrombin-activated samples were then
prediluted before incubation with the clotting assay reagents.
The mutant would undergo spontaneous decay more rapidly
than FVIII WT before assay determination. Accordingly, extrapolating from the initial slope of the inactivation phase, if FVIII
WT was incubated four times longer with thrombin before assay
determination, the remaining apparent peak activity would be
similar to the apparent peak activity observed for the thrombinactivated ARG531HIS protein. Therefore, there is no need to
invoke an additional functional defect for this mutation.
Several other genetic mutations have been described with
discrepant one-stage and two-stage activities. All reported
patients have mild to moderate hemophilia A phenotype and all
of the mutations occur within either the A1, A2, or A3
domains.31 Typically, they involve amino acid substitutions that
alter charge (eg, ARG698TRP) or hydrophobicity (eg, ALA284GLU
and SER289LEU). It can be postulated that these missense
mutations, although not confined to either the acidic region of
the A1 domain or to the A2 subunit itself, may still interfere
with the weak electrostatic interaction at this critical interaction
site, thereby leading to similar instability of the FVIIIa heterotrimer.
Because the patients with the ARG531HIS mutation have a mild
hemophilia A phenotype, the observed FVIII dysfunction
observed in vitro in these assays is apparently also important in
vivo. This is an important observation for several reasons.
Firstly, it is not known in vivo whether FVIIIa procoagulant
activity is limited by spontaneous A2 subunit dissociation or
further proteolysis. The observations from this CRM-positive
mutant and porcine FVIII would suggest that the inherent
instability of FVIIIa also limits its activity in vivo and is either
further compromised by mechanisms that lead to increased A2
dissociation or partially abrogated by mechanisms that lead to
reduced A2 dissociation. Secondly, an increased plasma level of
FVIII has now been identified as a risk factor for thrombosis.57,58 If proteolytic inactivation by activated protein C was of
primary importance in regulating FVIIIa, then one could predict
that a mutation leading to resistance to activated protein C
would also lead to an increased risk of thrombosis. Comprehensive analysis of patients with thrombophilia has failed to
identify any mutations at activated protein C cleavage sites
within FVIII despite the prominent association of factor V
ARG531HIS
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
182
PIPE ET AL
Leiden (resistant to activated protein C) with this cohort of
patients.59 Finally, this study predicts that even minor modifications of the A2 subunit can have major functional impacts both
in vitro and in vivo. This provides further insight into research
efforts directed at producing a new generation of recombinant
FVIII molecules with increased specific activity. Based on the
conclusions from this study and the others summarized here, we
propose that the most significant mechanism of FVIIIa inactivation in vivo is dissociation of the A2 subunit.
REFERENCES
1. Vehar GA, Keyt B, Eaton D, Rodriguez H, O’Brien DP, Rotblat F,
Oppermann H, Keck R, Wood WI, Harkins RN: Structure of human
factor VIII. Nature 312:337, 1984
2. Toole JJ, Knopf JL, Wozney JM, Sultzman LA, Buecker JL,
Pittman DD, Kaufman RJ, Brown E, Shoemaker C, Orr EC: Molecular
cloning of a cDNA encoding human antihaemophilic factor. Nature
312:342, 1984
3. Kaufman RJ, Wasley LC, Dorner AJ: Synthesis, processing, and
secretion of recombinant human factor VIII expressed in mammalian
cells. J Biol Chem 263:6352, 1988
4. Jenny RJ, Pittman DD, Toole JJ, Kriz RW, Aldape RA, Hewick
RM, Kaufman RJ, Mann KG: Complete cDNA and derived amino acid
sequence of human factor V. Proc Natl Acad Sci USA 84:4846, 1987
5. Koschinsky ML, Funk WD, van Oost BA, MacGillivray RT:
Complete cDNA sequence of human preceruloplasmin. Proc Natl Acad
Sci USA 83:5086, 1986
6. Tagliavacca L, Moon N, Dunham WR, Kaufman RJ: Identification
and functional requirement of Cu(I) and its ligands within coagulation
factor VIII. J Biol Chem 272:27428, 1997
7. Eaton D, Rodriguez H, Vehar GA: Proteolytic processing of
human factor VIII. Correlation of specific cleavages by thrombin, factor
Xa, and activated protein C with activation and inactivation of factor
VIII coagulant activity. Biochemistry 25:505, 1986
8. Fulcher CA, Roberts JR, Zimmerman TS: Thrombin proteolysis
of purified factor VIII procoagulant protein: Correlation of activation
with generation of a specific polypeptide. Blood 61:807, 1983
9. Fay PJ, Anderson MT, Chavin SI, Marder VJ: The size of human
factor VIII heterodimers and the effects produced by thrombin. Biochim
Biophys Acta 871:268, 1986
10. Fay PJ: Subunit structure of thrombin-activated human factor
VIIIa. Biochim Biophys Acta 952:181, 1988
11. Fay PJ, Haidaris PJ, Smudzin TM: Human factor VIIIa subunit
structure. Reconstruction of factor VIIIa from the isolated A1/A3C1-C2 dimer and A2 subunit. J Biol Chem 266:8957, 1991
12. Pittman DD, Kaufman RJ: Proteolytic requirements for thrombin
activation of anti-hemophilic factor (factor VIII). Proc Natl Acad Sci
USA 85:2429, 1988
13. Lollar P, Parker CG: Subunit structure of thrombin-activated
porcine factor VIII. Biochemistry 28:666, 1989
14. Fay PJ, Beattie TL, Regan LM, O’Brien LM, Kaufman RJ:
Model for the factor VIIIa-dependent decay of the intrinsic factor Xase.
Role of subunit dissociation and factor IXa-catalyzed proteolysis. J Biol
Chem 271:6027, 1996
15. Fay PJ, Haidaris PJ, Huggins CF: Role of the COOH-terminal
acidic region of A1 subunit in A2 subunit retention in human factor
VIIIa. J Biol Chem 268:17861, 1993
16. O’Brien LM, Huggins CF, Fay PJ: Interacting regions in the A1
and A2 subunits of factor VIIIa identified by zero-length cross-linking.
Blood 90:3943, 1997
17. Denson KWE, Biggs R: Laboratory diagnosis, tests of clotting
function and their standardization, in Biggs R (ed): Human Blood
Coagulation, Haemostasis and Thrombosis. Oxford, UK, Blackwell
Scientific, 1998, p 310
18. Barrowcliffe TW: Comparisons of one-stage and two-stage
assays of factor VIII:C. Scand J Haematol Suppl 41:39, 1984
19. Over J: Methodology of the one-stage assay of factor VIII
(VIII:C). Scand J Haematol Suppl 41:13, 1984
20. Barrowcliffe TW: Methodology of the two-stage assay of factor
VIII (VIII:C). Scand J Haematol Suppl 41:25, 1984
21. Rosen S: Assay of factor VIII:C with a chromogenic substrate.
Scand J Haematol Suppl 40:139, 1984
22. Lazarchick J, Hoyer LW: Immunoradiometric measurement of
the factor VIII procoagulant antigen. J Clin Invest 62:1048, 1978
23. McGinniss MJ, Kazazian HH Jr, Hoyer LW, Bi L, Inaba H,
Antonarakis SE: Spectrum of mutations in CRM-positive and CRMreduced hemophilia A. Genomics 15:392, 1993
24. Arai M, Inaba H, Higuchi M, Antonarakis SE, Kazazian HH Jr,
Fujimaki M, Hoyer LW: Direct characterization of factor VIII in
plasma: Detection of a mutation altering a thrombin cleavage site
(arginine-372—histidine). Proc Natl Acad Sci USA 86:4277, 1989
25. O’Brien DP, Pattinson JK, Tuddenham EG: Purification and
characterization of factor VIII 372-Cys: A hypofunctional cofactor from
a patient with moderately severe hemophilia A. Blood 75:1664, 1990
26. Johnson DJ, Pemberton S, Acquila M, Mori PG, Tuddenham EG,
O’Brien DP: Factor VIII S373L: Mutation at P18 site confers thrombin
cleavage resistance, causing mild haemophilia A. Thromb Haemost
71:428, 1994
27. Aly AM, Higuchi M, Kasper CK, Kazazian HH Jr, Antonarakis
SE, Hoyer LW: Hemophilia A due to mutations that create new
N-glycosylation sites. Proc Natl Acad Sci USA 89:4933, 1992
28. Aly AM, Hoyer LW: Factor VIII-East Hartford (arginine 1689 to
cysteine) has procoagulant activity when separated from von Willebrand factor. J Clin Invest 89:1382, 1992
29. Amano K, Sarkar R, Pemberton S, Kemball-Cook G, Kazazian
HH Jr, Kaufman RJ: The molecular basis for cross-reacting materialpositive hemophilia A due to missense mutations within the A2-domain
of factor VIII. Blood 91:538, 1998
30. Duncan EM, Duncan BM, Tunbridge LJ, Lloyd JV: Familial
discrepancy between the one-stage and two-stage factor VIII methods in
a subgroup of patients with haemophilia A. Br J Haematol 87:846, 1994
31. Rudzki Z, Duncan EM, Casey GJ, Neumann M, Favaloro EJ,
Lloyd JV: Mutations in a subgroup of patients with mild haemophilia A
and a familial discrepancy between the one-stage and two-stage factor
VIII:C methods. Br J Haematol 94:400, 1996
32. Hathaway WE, Christian MJ, Jacobson LJ: Variant mild hemophilia. Discrepancy in one stage and two stage factor VIII assays.
Thromb Haemost 50:357a, 1993 (abstr)
33. Parquet-Gernez A, Mazurier C, Goudemand M: Functional and
immunological assays of FVIII in 133 haemophiliacs—Characterization of a subgroup of patients with mild haemophilia A and discrepancy
in 1- and 2-stage assays. Thromb Haemost 59:202, 1988
34. Gaucher C, Jorieux S, Parquet-Gernez A, Mazurier C: Two
mutations of the factor VIII (FVIII) gene, R527W and R531H, are
identified in 4 unrelated haemophilia A patients with discrepancy in 1and 2-stage assays of FVIII coagulant activity. Haemophilia 2:223a,
1996 (abstr, suppl 1)
35. Montgomery RR, Hathaway WE, Johnson J, Jacobson L, Muntean
W: A variant of von Willebrand’s disease with abnormal expression of
factor VIII procoagulant activity. Blood 60:201, 1982
36. Kroner PA, Foster PA, Fahs SA, Montgomery RR: The defective
interaction between von Willebrand factor and factor VIII in a patient
with type 1 von Willebrand disease is caused by substitution of Arg19
and His54 in mature von Willebrand factor. Blood 87:1013, 1996
37. Kaufman RJ: Vectors used for expression in mammalian cells.
Methods Enzymol 185:487, 1990
38. Pittman DD, Kaufman RJ: Site-directed mutagenesis and expression of coagulation factors VIII and V in mammalian cells. Methods
Enzymol 222:236, 1993
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
HEMOPHILIA DUE TO INCREASED A2 SUBUNIT DISSOCIATION
39. Michnick DA, Pittman DD, Wise RJ, Kaufman RJ: Identification
of individual tyrosine sulfation sites within factor VIII required for
optimal activity and efficient thrombin cleavage. J Biol Chem 269:
20095, 1994
40. Zatloukal K, Cotten M, Berger M, Schmidt W, Wagner E,
Birnstiel ML: In vivo production of human factor VII in mice after
intrasplenic implantation of primary fibroblasts transfected by receptormediated, adenovirus-augmented gene delivery. Proc Natl Acad Sci
USA 91:5148, 1994
41. Johnsson B, Lofas S, Lindquist G: Immobilization of proteins to
a carboxymethyldextran-modified gold surface for biospecific interaction analysis in surface plasmon resonance sensors. Anal Biochem
198:268, 1991
42. O’Shannessy DJ, Brigham-Burke M, Soneson KK, Hensley P,
Brooks I: Determination of rate and equilibrium binding constants for
macromolecular interactions using surface plasmon resonance: Use of
nonlinear least squares analysis methods. Anal Biochem 212:457, 1993
43. Kemball-Cook G, Tuddenham EG: The Factor VIII Mutation
Database on the World Wide Web: The haemophilia A mutation, search,
test and resource site. HAMSTeRS update (version 3.0). Nucleic Acids
Res 25:128, 1997
44. Lollar P, Parker ET, Fay PJ: Coagulant properties of hybrid
human/porcine factor VIII molecules. J Biol Chem 267:23652, 1992
45. Barrowcliffe TW, Kirkwood TB: An international collaborative
assay of factor VIII clotting activity. Thromb Haemost 40:260, 1978
46. Arai M, Scandella D, Hoyer LW: Molecular basis of factor VIII
inhibition by human antibodies. Antibodies that bind to the factor VIII
light chain prevent the interaction of factor VIII with phospholipid. J
Clin Invest 83:1978, 1989
47. Foster PA, Fulcher CA, Houghten RA, Zimmerman TS: Synthetic factor VIII peptides with amino acid sequences contained within
the C2 domain of factor VIII inhibit factor VIII binding to phosphatidylserine. Blood 75:1999, 1990
48. Saenko EL, Shima M, Rajalakshmi KJ, Scandella D: A role for
the C2 domain of factor VIII in binding to von Willebrand factor. J Biol
Chem 269:11601, 1994
183
49. Saenko EL, Scandella D: The acidic region of the factor VIII
light chain and the C2 domain together form the high affinity binding
site for von Willebrand factor. J Biol Chem 272:18007, 1997
50. Fay PJ, Beattie T, Huggins CF, Regan LM: Factor VIIIa A2
subunit residues 558-565 represent a factor IXa interactive site. J Biol
Chem 269:20522, 1994
51. O’Brien LM, Medved LV, Fay PJ: Localization of factor IXa and
factor VIIIa interactive sites. J Biol Chem 270:27087, 1995
52. Lollar P, Parker ET: Structural basis for the decreased procoagulant activity of human factor VIII compared to the porcine homolog. J
Biol Chem 266:12481, 1991
53. Lollar P, Parker CG: pH-dependent denaturation of thrombinactivated porcine factor VIII. J Biol Chem 265:1688, 1990
54. Pipe SW, Kaufman RJ: Characterization of a genetically engineered inactivation-resistant coagulation factor VIIIa. Proc Natl Acad
Sci USA 94:11851, 1997
55. Fulcher CA, Gardiner JE, Griffin JH, Zimmerman TS: Proteolytic inactivation of human factor VIII procoagulant protein by activated human protein C and its analogy with factor V. Blood 63:486,
1984
56. Amano K, Michnick DA, Moussalli M, Kaufman RJ: Mutation
at either Arg336 or Arg562 in factor VIII is insufficient for complete
resistance to activated protein C (APC)-mediated inactivation: Implications for the APC resistance test. Thromb Haemost 79:557, 1998
57. Koster T, Blann AD, Briet E, Vandenbroucke JP, Rosendaal FR:
Role of clotting factor VIII in effect of von Willebrand factor on
occurrence of deep-vein thrombosis. Lancet 345:152, 1995
58. O’Donnell J, Tuddenham EG, Manning R, Kemball-Cook G,
Johnson D, Laffan M: High prevalence of elevated factor VIII levels in
patients referred for thrombophilia screening: Role of increased synthesis and relationship to the acute phase reaction. Thromb Haemost
77:825, 1997
59. Roelse JC, Koopman MM, Buller HR, ten Cate JW, Montaruli B,
van Mourik JA, Voorbert J: Absence of mutations at the activated
protein C cleavage sites of factor VIII in 125 patients with venous
thrombosis. Br J Haematol 92:740, 1996
From www.bloodjournal.org by guest on June 18, 2017. For personal use only.
1999 93: 176-183
Mild Hemophilia A Caused by Increased Rate of Factor VIII A2 Subunit
Dissociation: Evidence for Nonproteolytic Inactivation of Factor VIIIa In
Vivo
S.W. Pipe, A.N. Eickhorst, S.H. McKinley, E.L. Saenko and R.J. Kaufman
Updated information and services can be found at:
http://www.bloodjournal.org/content/93/1/176.full.html
Articles on similar topics can be found in the following Blood collections
Hemostasis, Thrombosis, and Vascular Biology (2485 articles)
Information about reproducing this article in parts or in its entirety may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests
Information about ordering reprints may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#reprints
Information about subscriptions and ASH membership may be found online at:
http://www.bloodjournal.org/site/subscriptions/index.xhtml
Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American Society of
Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.
Copyright 2011 by The American Society of Hematology; all rights reserved.