Polymorphisms of coagulation factor XIII subunit A and risk of non-fatal hemorrhagic stroke in young white women

(1)

Risk of Nonfatal Hemorrhagic Stroke in Young White Women

Alexander P Reiner, MD, Stephen M Schwartz, PhD, Michele B Frank, MD,

W T Longstreth, Jr, MD, MPH, Lucia A Hmdorff, BA, Gayle Teramura, BS, Fnts R Rosendaal, MD,

Lakshmi K Gaur, PhD, Bruce M Psaty, MD, PhD, David S Siscovick, MD, MPH

Batkground and Purpose—Although family studies have suggested a genetic mfluence on hemorrhagic stroke, the underlymg genetic nsk factors remam poorly defmed Coagulation factor XIII, which is mvolved m hemostasis, fibnnolysis, vascular remodelmg, and tissue repair, represents a candidate gene for hemorrhagic stroke We assessed the potential role of 3 factor XIII subunit A codmg-sequence polymorphisms, along with a promoter polymorphism of plasmmogen activator mhibitor-1 (PAI-1, which is also mvolved in fibrin stabilization and vascular remodelmg), m young white women with hemorrhagic stroke

Methods—Genotype analysis for factor XIII subunit A Val34Leu, Tyr204Phe, and Pro564Leu and for PAI-1 -675 4G/5G was performed m a population-based case-control study of 42 white women aged <45 years with nonfatal hemorrhagic stroke and 345 demographically similar control subjects

Results—Compared with the respective homozygous wild-type genotypes, the Tyr204/Phe204 genotypes (age-adjusted odds raüo [OR] 2 9, 95% 95% CI l l to 7 5) and the Leu564/Leu564 genotype (OR 4 3, 95% CI l 4 to 13 7) were each associated with an increased nsk of nonfatal hemorrhagic stroke The nsk estimate associated with the Phe204 vanant was highest m women with subarachnoid hemorrhage and m nonsmokers, whereas the nsk estimate of the Leu564/Leu564 genotype was highest m women with mtracerebral hemorrhage and in smokers Women who carried either the Phe204 allele or the Leu564/Leu564 genotype m combmation with the PAI-1 5G/5G genotype had a nearly 20-fold increased risk of hemorrhagic stroke (OR 18 9, 95% CI 3 8 to 95 1)

Conduslons—Our findmgs suggest that the Phe204 and Leu564 variants of coagulation factor XIII may be markers for genetic susceptibility to hemorrhagic stroke m women aged <45 years (Stroke. 2001;32:2580-2587.)

Key Words: factor XIII · hemorrhagic stroke · plasmmogen activator mhibitor-1

H

emorrhagic stroke occurs when an artery ruptures and causes blood leakage mto the bram parenchyma or the subarachnoid space Overall, it accounts for «20% of stroke victims and is the most common form of stroke among young adults Data from family studies have suggested the impor-tance of genetic mfluence m hemorrhagic stroke, particularly subarachnoid hemorrhage (SAH) '~3 However, the specific genetic factors are pooily defmed4 Because mtracramal aneurysms are associated with rare mhented connective tissue disorders, most previous candidate gene studies have focused on structural vessel wall protems, such collagen, or, m the case of mtracerebral hemorrhage (ICH), cerebral amyloid deposition 4 5

The events required for a hemorrhagic stroke to become clmically manifest mclude not only the formation of a weakened or abnormal vessel wall but also vessel rupture and hemorrhage Inhented deficiency of coagulation factor XIII is

See Editorial Comment, page 2586

a very rare autosomal-recessive bleedmg disorder character-ized by a particularly high rate of mtracramal hemorrhage (up to 30% of cases)6 Coagulation factor XIII is a proenzyme that circulates äs a heterotetramer composed of 2 catalytic subumts (factor XIII subunit A [factor XIIIA]) and 2 carrier subunits (factor XII subunit B [factor XIIIB]) Durmg the final stages of coagulation, thrombm cleavage of factor XIIIA results m formation of activated factor XIII, a transglutami-nase that cross-lmks adjacent fibrm molecules to increase clot stabihty and resistance to fibnnolysis Factor XIII also participates m extracellular matrix remodelmg, cell adhesion and migration, and tissue repair 7 In a rat model of expen-mental cerebral aneurysm formation, exogenous admimstra-tion of factor XIII abrogated the defective mtimal prolifera-tive response to arterial wall mjury 8 Thus, deficiency of

Received February 22, 2001, final revision received June 14, 2001, accepted June 25, 2001

From the Departments of Medicme (A P R, M B F , W T L , G T , B M P D S S ) , Epidemiology ( A P R , S M S , W T L , L A H , B M P , D S S ) , Neurology (W T L ), Health Services ( B M P ) , Microbiology (L K G ), and Cardiovascular Health Research Unit (W T L , S MS, B M P , D S S ) , University of Washington, Seattle, Fred Hutchmson Cancer Research Center ( S M S ) , Seattle, Wash, and the Hemostasis and Thrombosis Research Center and Department of Chmcal Epidemiology (F R R ), Leiden University Medical Center, Leiden, the Netherlands

Stroke is available at http //www strokeaha.org

(2)

TABLE 1. Genotyping Assays for Factor XIII and PAI-1 Polymorphisms

Polymorphism Factor XIII Val34Leu Factor XIII Tyr204Phe Factor XIII Pro564Leu PAI-1 -675 4G/5G Primers CATGCCTTTTCTGTTGTCTTC TACCTTGCAGGTTGACGCCCCGGGGCACTA GGAAACAGTCTGGTTTGGTAA ACCCCGATGTCATTCAGGACG TCACCTTCTACACCGGGGTCG GACAGCGAGTCTCACAAAGAA CACAGAGAGAGTCTGGACACGTGA TGCAGCCAGCCACGTGATTGTCTAG PCR Product 192 bp (161, 31) 113 bp (91, 22) 116 bp (95, 21) 148 bp (110, 38) Restnction Enzyme Datei ffsal ßsfUI ßseRI

Primer sequences are listed from 5' to 3', and the upper sequence represents the forward primer, the Iower sequence, the reverse primer The underlmed bases were changed from the native sequence to mtroduce a restriction enzyme recognition sequence that would result m digestion of the PCR product to the fragments shown m parentheses for 1 (le, Leu34, Tyr204, Leu564, or 4G) of the 2 alleles Genomic DNA (50 ng) was amplified m a Standard 20-μΙ PCR reaction for 35 cycles of denaturation at 94°C for 30 seconds, annealmg at 50°C for 30 seconds, and extension at 72°C for 60 seconds, followed by a final extension at 72°C for 5 mmutes Aliquots of the PCR products were digested with the appropnate restriction enzyme for 1 hour accordmg to the manufacturer's mstructions (New England

ßiolabs)

coagulation factor XIII may contnbute to the occurrence of hemorrhagic stroke through direct effects on vessel wall mtegnty or aneurysm formation m addition to bleedmg tendency

A common factor XIIIA gene nucleotide Substitution (polymorphism) results m a Val34Leu ammo acid Substitu-tion and was recently associated with mcreased nsk of pnmary ICH m a predommantly elderly patient population 9

Two other common factor XIIIA polymorphisms, Tyr204Phe and Pro564Leu, have been associated with decreased plasma factor XIII levels,10—l2 but their roles m hemorrhagic stroke

have not been previously assessed Therefore, we exammed the association of the factor XIII Val34Leu, Tyr204Phe, and Pro564Leu polymorphisms with hemorrhagic stroke by usmg samples from a population-based case-control study of women aged <45 years Plasminogen activator mhibitor-1 (PAI-1) is a major Inhibitor of fibrmolysis and, hke factor XIII, is involved m fibnn stabilization, vascular remodelmg, and tissue repair 13 The PAI-1 —675 4G/5G promoter

poly-morphism has been associated with Variation m plasma PAI-1 levels14 and with the nsk of early-onset stroke'5 or

cerebro-vascular mortality m women 16 Thus, we additionally

as-sessed potential gene-gene mteractions between the PAI-1 4G/5G polymorphism and the factor XIII polymorphisms

Subjects and Methods

Study Subjects and Data Collection

The source population for the present study was women aged 18 to 44 years residmg m 3 contiguous counties m western Washington state between 1991 and 1995 and represented «*2 2 milhon women-years at nsk 17 18 Ehgible stroke patients free of pnor cerebrovascular

disease were imtially identified by chart review of discharge diag-noses from all 34 Hospitals and Emergency Medical Service mcident reports in the study region Stroke was defmed äs evidence of new focal neurological deficit(s) with no other apparent cause (le, bram tumor, infection, subdural hematoma, seizure, or other neurological condition) lasting >24 hours On the basis of reviews of hospital records, bram imagmg studies, and lumbar puncture results, 141 of the 249 ehgible stroke cases were classified äs hemorrhagic by the study neurologist (W T L ) Of the 141 hemorrhagic strokes, 91 were

further charactenzed äs SAH on the basis of either (1) blood within the subarachnoid space m the presence or absence of a demonstrated aneurysm and no evidence of artenovenous malformation or (2) an aneurysm with blood in other locations (eg, the parenchyma or ventncles) and no evidence of artenovenous malformation For the remaimng 50 women, the strokes were classified äs ICH Six hundred eighty-four control women aged 18 to 44 years without history of cerebrovascular disease were identified from the same geographic area and were frequency-matched to the age distnbution of the cases

Of the 141 identified hemorrhagic stroke cases, 102 were ahve and not disabled at the time of recruitment Eighty (78 4%) of the 102 patients with nonfatal hemorrhagic stroke and 526 (76 9%) of the 684 ehgible control subjects agreed to an m-person interview Demographic charactenstics and medical histones were ascertamed durmg the interview, and blood samples for DNA Isolation were collected from 54 hemorrhagic stroke cases (37 SAH and 17 ICH) and from 391 control subjects Case blood samples were obtamed at least 3 months after the stroke (mean 8 months) All study subjects gave informed consent accordmg to a protocol approved by the Umversity of Washington human subjects committee and participat-mg local hospitals Data collection for the present study did not include Information regardmg the use of ephedra compounds

Factor XIII and PAI-1 Genotyping

Genomic DNA samples were prepared äs descnbed previously ls

Genotyping for the factor XIII Val34Leu polymorphism was pei-formed by polymerase cham reaction (PCR) amphfication of genomic DNA followed by restriction enzyme digestion (PCR-RFLP) with Ddel '9 Similar PCR-RFLP genotypmg methods were

developed for the Tyr204Phe and Pro564Leu factor XIII and PAI-1 4G/5G polymorphisms, äs shown m Table l For all PCR amphfi-cations, positive control DNA samples of known genotype äs well äs a negative control without template DNA were mcluded Ten percent of the study samples underwent repeat genotypmg analysis, and the initial results were confirmed in all cases As an additional validation measure, direct nucleotide sequencmg of the PCR products was performed for selected homozygous or heterozygous mdividuals, m every case, the sequencmg results agreed with the PCR-RFLP genotypmg result

Statistical Analysis

(3)

Το assess the relationship between the inhentance of allehc vanants at one polymorphic locus and the allehc vanants at a second polymorphic locus, pairwise Imkage disequihbiium between factor XIII polymorphic loci was determmed among controls by usmg a permutation proceduie of the expectation-maxirmzation algonthm20 and expressed äs D', which ranges fiom 0 (complete Imkage equihbnum) to l (complete Imkage disequihbrmm) D' is reportedas positive when the mmor allele at one polymorphic locus is associated with the mmor allele at the other locus and is leported äs negative when the mmor allele at one locus is associated with the major allele at the other locus

The association of factor XIII genotype with hemorrhagic stroke was exammed by unconditional logistic regiession adjusted for age and was expressed äs odds ratlos (ORs) and 95% CIs Homozygosity for the more common allele (le, Val34/Val34, Tyr204/Tyr204, or Pio564/Pro564) was used äs the refeience group m the regression models In the case of the Val34Leu and Pro564Leu polymorphisms, 2 dummy variables lepresentmg the heterozygous and rare homozy-gous genotypes were modeled to provide separate estimates of association for each genotype We also performed a multivanate logistic regression model that included all 3 factor XIII poly-morphisms simultaneously to adjust for the potential confoundmg effects of the 2 remaming polymorphisms Because the present study was exploratory m nature, in some mstances we exammed the combmed nsk estimate associated with the presence of heterozygos-ity for Phe204 or homozygosheterozygos-ity for Leu564 on the basis of post hoc analysis of our data

The extent to which associations with factor XIII genotypes were modified by other putative stroke nsk factors was assessed through analyses stratified by these other nsk factors To lest for the presence of mteraction, multiphcative terms were mtroduced mto the logistic regiession models, and the P value was computed for the hkelihood ratio test by companng the model containmg the mteraction term with the model lackmg the mteraction term Logistic regiession models were peiformed by usmg Stata (version 6 O),21 and genetic

data analysis was performed by usmg Arlequm (version 2 000)22 All

statistical testing was 2-sided and perfoimed at the a=0 05 level

Analysis of Factor XIII Alleles Among Healthy Black Subjects

Of the 54 hemorrhagic stroke cases and 391 control subjects with analyzable DNA, race was self-reported m 12 cases and 46 controls äs nonwhite (le, black, Asian, American Indian, or Hispamc) Black women constituted only 5 of the hemoirhagic stroke cases and only 9 of the controls Therefoie, factor XIII allele frequencies among the black population of Seattle were deteimmed by usmg a larger number of blood samples obtamed from 124 healthy black blood donors from the Füget Sound Blood Center Compared with the 345 white control subjects from our case-contiol study, the Leu34 (13% versus 26%, P<0 001), Phe204 (0 4% versus 3 0%, P=0 02), and Leu564 (15% versus 20%, P=005) alleles were significantly less common among the 124 healthy blacks Because of these differences in inteiracial factoi XIII allele frequencies, we excluded the 12 nonwhite hemorrhagic stroke cases and 46 nonwhite control subjects from the present association study to mimmize the possibility of confoundmg due to population admixture Therefore, our present analysis of the relationship between factor XIII genotypes and nsk of hemorrhagic stroke was restncted to the white study subjects with DNA samples (42 hemoirhagic stroke cases and 345 controls)

Results

Characteristics of the Study Subjects

The charactenstics of the hemorrhagic stroke cases and controls are summanzed in Table 2 The mean age was 37 8 years The overwhelmmg majority of the study participants were premenopausal and not usmg oral contraceptives Cig-arette smokmg and hypertension were more commonly re-ported among hemorrhagic stroke cases than among the

TABLE 2. Characteristics of Hemorrhagic Stroke Patients and Control Subjects Charactenstic Age, y Mean Mediän Range Premenopausal, %

Current oral contraceptive use % Current smokers, %

Body mass mdex >30 kg/m2, %

Hypertension, % Diabetes %

Hypercholesterolemia, %

Alcohol consumption ä1 time per week, %

Hemorrhagic Stroke Cases (n =42) 360 370 18-44 881 119 476 9 5 238 4 8 167 425 Control Subjects (n =345) 377 390 19-44 959 111 209 163 87 3 2 164 401

control subjects Of the 42 hemorrhagic stroke cases, 30 were classified äs SAH, and 12 were classified äs ICH Intracramal aneurysms were documented by imagmg studies m 23 (77%) of the 30 women with SAH Artenovenous malformations were documented m 4 (33%) of the 12 women with ICH In the remaming 15 hemorrhagic stroke patients, no vascular source could be identified

Factor XIII Genotypes, Alleles, and Linkage Disequüibrium

The observed genotype distnbutions of the factor XIII Val34Leu, Tyr204Phe, and Pro564Leu polymorphisms and the calculated allele frequencies among the white subjects are shown m Table 3 The genotype distnbutions of all 3 factor XIII polymorphisms were m Hardy-Weinberg equilibnum among the control subjects, mdicatmg the expected relation-ship between genotype frequencies and calculated fiequen-cies among our study population

Among the white control subjects, there was a tendency toward negative Imkage disequihbnum between the Val34Leu and Tyr204Phe polymorphisms (D'= -058, P=0 17) and between the Val34Leu and Pro564Leu poly-morphisms (D' = — 0 19, P=0 33) and a tendency toward positive Imkage disequilibnum between the Tyr204Phe and Pro564Leu polymorphisms (D' = +024, P=0 16) Among the 124 black subjects, the negative Imkage disequihbnum between Val34Leu and Pro564Leu was considerably strenger (D' = -099, P=002)

Association of Factor XIII Polymorphisms With Hemorrhagic Stroke

(4)

TABLE 3. Factor XIII Genotypes in Hemorrhagic Stroke Gases and Controls

Val34Leu

Control subjects (n =345) Hemorrhagic stroke patients Tyr204Phe

Control subjects (n =345) Hemorrhagic stroke patients Pro564Leu

Control subjects (n=345) Hemorrhagic stroke patients

(n=42)

(n =42)

*P value for χ2 test of homogeneity

tOR was adjusted for age reference

n (%) Val 512(742) 65 (77 4) Tyr 669 (97 0) 77 (91 7) Pro 551 (79 9) 57 (67 9) group for OR Alleles n (%) P* Leu 178(258) 19(226) 053 Phe 21 (3 0) 7 (8 3) 0 01 Leu 139(201) 27(321) 001 is women homozygous Genotypes n (%) Val/Val 187(542) 25 (59 5) Tyr/Tyr 324 (93 9) 35 (83 3) Pro/Pro 218(632) 20 (47 6)

for the more

n (%) Val/Leu 138(400) 15(357) Tyr/Phe 21(61) 7(167) Pro/Leu 115(333) 17(405) common allele n (%) Leu/Leu 20 (5 8) 2(48) Phe/Phe 0(0) 0(0) Leu/Leu 12(35) 5(119) P* (95% Cl) Val/Leu 0 42 0 8 (0 4-1 5) Tyr/Phe 003 29(114-75) Pro/Leu 0 01 1 7 (0 9-3 4) ORt (95% Cl) Leu/Leu 0 7 (0 2-3 2) . . Leu/Leu 43(1 4-137) (le VaI34/Val34, Tyr204/Tyr204 or Pro564/Pro564) a moderately mcreased nsk associated with carrymg l copy

of the Leu564 allele (age-adjusted OR l 7, 95% Cl 0 9 to 3 4) compared with the Pro564/Pro564 genotype For subjects who camed at least l copy of the Leu564 allele and/or were heterozygous for Tyr204/Phe204 (62% of cases versus 40% of controls), the age-adjusted OR for hemorrhagic stroke was 2 6 (95% Cl l 3 to 5 0) The combmed prevalence of the Tyr204/Phe204 genotype and/or the Leu564/Leu564 geno-type was 26 2% among the hemorrhagic stroke cases com-pared 9 3% of the controls (l case and l control carned both genotypes) Thus, women carrymg either factor XIII geno-type had a 3-fold mcreased nsk of hemorrhagic stroke (age-adjusted OR 3 3, 95% Cl l 5 to 7 1) In contrast, the presence of l or 2 copies of the Val34Leu polymorphism was associated with a decreased nsk of hemorrhagic stroke (age-adjusted OR «=0 7), but the CIs were compatible with the absence of an association (Table 3)

We also assessed whether the associations between each factor XIII polymorphism and hemorrhagic stroke were influenced by the 2 remaimng factor XIII polymorphisms with the use of multivanate adjustment For each factor XIII polymorphism, adjustment for the other 2 did not appreciably affect the umvanate nsk estimates in Table 3 (data not shown)

For all 3 factor XIII polymorphisms, restnctmg the analy-sis to the women who were either premenopausal or were not usmg oral contraceptives did not appreciably alter the nsk estimates for hemorrhagic stroke When they were analyzed according to hemorrhagic stroke subtype, all 7 hemorrhagic stroke cases carrymg the Phe204 allele belonged to the subgroup of 30 women classified with SAH Thus, the factor XIII Phe204 variant was associated with a nearly 5-fold risk of SAH (age-adjusted OR 4 5, 95% Cl l 7 to 11 8) In contrast, 10 (83%) of the 12 women with ICH camed at least TABLE 4. Factor XIII Phe204 and Leu564/Leu564 Variants m Hemorrhagic Stroke Cases and Controls, by Other Risk Factors

Tyr204Phe Polymorphism Cases (n=42) Risk Factor Overall Current smoking No Yes Hypertension No Yes Alcohol consumption <1 time per week >1 time per week PAI-1 genotype 4G/4G or 4G/5G 5G/5G Y/Y 35 17 18 27 8 18 15 26 9 F/Y 7 5 2 5 2 5 2 4 3 Pro564Leu Polymorphism Controls (n =345) Y/Y 324 258 66 297 27 190 128 253 70 F/Y 21 15 6 18 3 13 8 19 2 OR (95% Cl) 29(114-75) 52(17-164) 0 9 (0 2-5 5) 2 8 (0 95-8 3) 25(04-189) 41 (1 3-128) 1 8 (0 3-9 5) 2 0 (0 6-6 6) 128(16-800) Cases (n =42) P* P/P or P/L 37 009 22 15 089 28 9 043 19 16 019 28 9 L/L 5 0 5 4 1 4 1 2 3 Controls (n =345) P/P or P/L 333 265 68 304 29 195 133 262 70 L/L 12 8 4 11 1 8 3 10 2 ORt (95% Cl)t 35(114-106) 0 (0-6 0) 5 6 (1 3-23 9) 36(1 04-121) 4 7 (0 3-87 8) 50(14-182) 2 0 (0 2-23 6) 1 6 (0 3-7 9) 116(17-792) P* 003 095 054 011

Υ mdicates tyrosme, F, phenylalanme P, prohne, and L, leucme *P value for mteraction

(5)

l copy of the Leu564 allele compared with only 127 (37%) of the 345 controls (age-adjusted OR 9 l, 95% CI l 9 to 42 8)

The nsk of hemorrhag:c stroke associated with Phe204 and Leu564/Leu564 was modified by current cigarette smokmg The association with the Phe204 vanant was largest among nonsmokers, whereas the effect of the Leu564/Leu564 geno-type was most pronounced among smokers (Table 4) The nsk of hemorrhagic stroke associated with either the Phe204 vanant or the Leu564/Leu564 genotype was also particularly high among the subgroup of women with the PAI-1 5G/5G genotype Among the subgroup of women who carned the PAI-1 5G/5G genotype, the presence of either the Phe204 allele or the Leu564/Leu564 genotype was associated with a nearly 20-fold mcreased nsk of hemorrhagic stroke (age-adjusted OR 18 9, 95% CI 3 8 to 95 1) By contrast, there was no mcreased nsk of hemorrhagic stroke in women with either factor XIII genotype who camed the PAI-1 4G/4G or 4G/5G genotype (age-adjusted OR l l, 95% CI 0 6 to 4 6) This resulted in a P value for mteraction of 0 016

Discussion

Our findmgs from a population-based association study among young white women mdicate that 2 common genetic vanants of coagulation factor XIIIA, Phe204 and Leu564, are each associated with a 3- to 4-fold mcreased nsk of nonfatal hemorrhagic stroke Although the 2 mtragenic factor XIII polymorphisms were m partial hnkage disequilibnum, the associations with stroke were mutually mdependent The Leu564 allele has been associated with decreased plasma factor XIII levels,1011 and the Phe204 vanant has been associated with reduced specific activity of plasma factor XIII12 Decreased clot stability may lead to greater suscepti-bihty to mtracramal hemorrhage, particularly among younger individuals, who tend to have lower factor XIII levels 23

The nsk estimates associated with the 2 factor XIII polymorphisms appeared to be modified by cigarette smok-mg, although m opposite directions The risk associated with the Phe204 vanant was greatest among nonsmokers, whereas the effect of the Leu564/Leu564 genotype was most pro-nounced among smokers There was also a Suggestion that the nsk associated with the Phe204 vanant was highest among the women with SAH, whereas the risk associated with the Leu564/Leu564 genotype was most pronounced among the women with ICH Cigarette smokmg has been shown to mcrease the nsk of rupture and hemorrhage m subjects with mtracramal aneurysm 24 25 However, cigarette smokmg is also associated with mcreased plasma factor XIII levels23 äs well äs decreased fibrmolysis 26 Thus, our findmgs suggest that the detnmental effect on clot or vessel wall stability associated with the Phe204 vanant may be "neutrahzed" in cigarette smokers

The risk estimate associated with the presence of either the Tyr204/Phe204 or Leu564/Leu564 genotype was particularly large among women carrymg the PAI-1 promoter 5G/5G genotype PAI-1 is a major physiological mhibitor of fibn-nolysis, and the 5G/5G genotype has been associated with decreased plasma levels of PAI-1 relative to the 4G/4G genotype 14 Thus, the reduced clotting activity associated with the Phe204 and Leu564 factor XIII vanants may be

compounded by mcreased fibrmolysis associated with PAI-1 5G/5G, which potentially mcreases the hkehhood of mtracra-mal hemorrhage In addition, factor XIII and PAI-1 both may participate in vascular remodelmg, tissue repair, and angio-genesis 713 27 28 Delayed wound heahng, mtracramal hemor-rhage, and recurrent miscamage are common features of mhented factor XIII deficiency,6 and an mcreased prevalence of the factor XIII Phe204 allele was recently noted among women with recurrent miscamage '' Thus, genetic Variation of factor XIII may affect the occurrence of hemorrhagic stroke through effects on vessel wall stability and aneurysm formation or rupture, äs well äs decreased clot stability

ICH and SAH are both more common among blacks than whites, and ICH is particularly prevalent in Asian popula-tions It is interestmg to note that the Leu564 allele (which our data suggest may be more strongly related to ICH) is more common among Japanese than white individuals29 Because our data mdicate that both the Phe204 and Leu564 alleles are less common among blacks than whites, there must be other genetic or environmental determmants responsible for the mcreased mcidence of hemorrhagic stroke among blacks

The lack of association between the factor XIII Val34Leu polymorphism and risk of hemorrhagic stroke m the present study contrasts with the association reported among elderly Bntish patients with ICH 9 However, it should be noted that our patient population was considerably younger and that most of the strokes were due to SAH rather than ICH Thus, it is possible that the Val34Leu mutation is more strongly related to ICH than SAH However, another recently pub-hshed study involving a larger number of subjects with both hemorrhagic stroke subtypes did not find any association between the Val34Leu genotype and the risk of hemorrhagic stroke 30

Several limitations of the present study should be noted Because hemorrhagic stroke is an uncommon event in women aged <45 years, the number of cases m the present study is small, which may mcrease the hkehhood of a spunous association The analysis of multiple subgroups also mcreases the possibihty of a false-positive association, therefore, the findmgs with respect to mteraction between factor XIII and other risk factors should be mterpreted cautiously The infrequency of hemorrhagic stroke in young women also precluded the use of a prospective design to lest our hypoth-eses Thus, we were able to study only women who survived at least 3 months after their strokes Because hemorrhagic stroke is associated with a high case-fatahty rate (26% in the present study), our results may be biased because of the exclusion of fatal cases if a particular factor XIII vanant is associated with fatal outcome rather than occurrence of hemorrhagic stroke Fmally, although we hmited our analysis to white subjects, we cannot exclude the possibihty of a spunous association arising that is due to population admixture

(6)

Interpretation of factor XIII levels among the stroke cases measured after the acute event Because previous studies have demonstrated a relationship between the factor XIII Tyr204Phe and Pro564Leu polymorphisms and plasma factor XIII levels,10-12 the determmation of factor XIII genotype actually may be preferable to the measurement of factor XIII activity levels m th:s settmg The advantage of studymg genetic vanants is that they are "fixed at birth" and, therefore, clearly precede disease onset and are not confounded by fluctuations in plasma activity levels related to the acute cerebrovascular event or by other behavioral and environ-mental nsk factors that mfluence factor XIII activity levels 23 In summary, genetic Variation of coagulation factor XIII may be important m the occurrence of hemorrhagic stroke m young white women, particularly m combmation with other genetic and environmental nsk factors These fmdings require confirmation in larger studies, which should mclude fatal and nonfatal hemorrhagic stroke cases, other ethnic groups, and family-based studies that assess genetic hnkage äs well äs association

Acknowledgments

This work was supported by a contract from the National Institute for Child Health and Human Development (NO1-HD-3107) and a grant from the Boeing Co The authors thank Fran Chard, Karen Graham, Carol Handley-Dahl, Judy Kaiser, Marlene Bengeult, Carol Oster-gard, Demse Koriander, Barb Twaddell, Sandy Tronsdal, and Jill Ashman for assistance with the study

References

1 Jousilahti P, Rastenyte D, Tuomilehto J, Sarti C, Vartianen E Parental history of cardiovascular disease and nsk of stroke a prospective follow-up of 14,371 middle-aged raen and women in Fmland Stroke 1997,28 1361-1366

2 Kubota M, Yamaura A, Ono J-I, Itani T, Tachi N, Ueda K, Nagata I, Sugimoto S Is family history an independent nsk factor for stroke7

J Neurol Neurosurg Psychmtry 1997,62 66-70

3 Gatst D, Vaeth M, Tsiropoulos I, Chnstensen K, Corder E, Olsen J, Sorensen HT Risk of subarachnoid haemorrhage m first degree relatives of patients with subarachnoid hemorrhage follow-up study based on national registnes m Denmark BMJ 2000,320 141-145

4 Rastenyte D, Tuorailehto J, Sarti C Geneticsof stroke areview J Neurol Sei 1998,153 132-145

5 Schievmk WI, Michels W, Piepgras DG Neurovascular manifestations of hentable connectwe tissue disorders a review Stroke 1994,25 889-903

6 Anwar R, Miloszewki KJA Factor XIII deficiency Br J Haematol 1999,107 468-484

7 Muszbek L, Yee VC, Hevessy Z Blood coagulation factor XIII structure and function Thromb Res 1999,94 271-305

8 Kang Y, Hashimoto N, Kikuchi H, Yamazoe N, Hazama F Effects of blood coagulation factor XIII on the development of expenmental cerebral aneurysms m rats J Neurosurg 1990,73 242-247

9 Catto AJ, Kohler HP, Bannan S, Stickland M, Carter A, Grant PJ Factor XIII Val34Leu a novel association with pnmary mtracerebral hemor-rhage Stroke 1998,29813-816

10 Kohler HP, Puters TS, Grant PJ Prevalence of three common poly-morphisms in the A-subunit gene of factor XIII in patients with coronary

artery disease association with FXHI activity and antigen levels Thromb Haemost 1999,81 511-515

11 Galhvan L, Markham AF, Anwar R The Leu564 factor XIIIA vanant results m significantly lower plasma factor XIII levels man the Pro564 vanant Thromb Haemost 1999,82 1368-1370

12 Anwar R, Galhvan L, Edmonds SD, Markham AF Genotype/phenotype conelations for coagulation factor XIII specific normal polymorphisms are associated with high or low factor XIII specific activity Blood 1999,93 897-901

13 Kohler HP, Grant PJ Plasmmogen activator Inhibitor type l and coronary artery disease N EnglJ Med 2000,3421792-1801

14 EnkssonP, Kallm B, van't Hooft FM, Bavenhom P, Hamsten A Allele-specific mcrease in basal transcnption of the plasmmogen activator m-hibitor l gene is associated with myocardial mfarction Proc Natl Acad Sa USA 1995,92 1851-1855

15 Endler G, Lalouschek W, Exner M, Mitterbauer G, Hanng D, Mannhalter C The 4G/4G genotype at nucleotide Position —675 in the promoter region of the plasmmogen activator Inhibitor l (PAI-1) gene is less frequent m young patients with minor stroke than controls Br J Haematol 2000,110469-471

16 Roest M, van der Schouw YT, Banga JD, Tempelman MJ, de Groot PG, Sixma JJ, Grobbee DE Plasmmogen activator Inhibitor 4G poly-morphism is associated with decreased nsk of cerebrovascular mortahty molderwomen Circulation 2000,10167-70

17 Schwartz SM, Siscovick DS, Longstreth WT Jr, Psaty BM, Beverly K, Raghunathan TE, Lm D, Koepsell TD Use of low-dose contraceptives and stroke in young women Ann Intern Med 1997,127 596-603 18 Longstreth WT Jr, Rosendaal FR, Siscovick DS, Vos HL, Schwartz SM,

Psaty BM, Raghunathan TE, Koepsell TD, Reitsma PH Risk of stroke m young women and two prothrombotic mutations factor V Leiden and prothrombm gene vanant (G20210A) Stroke 1998,29 577-580 19 Kangsadalampai S, Board PG The Val34Leu polymorphism in the A

subunit of coagulation factor XIII contnbutes to the large normal ränge m activity and demonstrates that the activation peptide plays a role m catalytic activity Blood 1998,922766-2770

20 Slatkm M, Excoffier L Testing for hnkage disequihbnum in genotypic data usmg the EM algonthm Heredity 1996,76 377-383

21 Stata Statistical Software Release 60 College Station, Tex Stata Corp, 1999

22 Schneiders, Roessh D, Excoffier L Arlequm Version 2 000 A Software for Population Genetics Data Analysis Geneva, Switzerland Genetics

and Biometry Laboratory, Umversity of Geneva, Switzerland, 2000 23 Anens RAS, Kohler HP, Mansfield MW, Grant PJ Subunit antigen and

activity levels of blood coagulation factor XIII m healthy mdividuals relation to sex, age, Smoking, and hypertension Artenoscler Thromb Vase Biol 1999,19 2012-2016

24 Juvela S, Porras M, Poussa K Natural history of unruptured intracramal aneurysms probabihty of and nsk factors for aneurysm rupture J Neu-rosurg 2000,3 379-387

25 Longstreth WT Jr, Nelson LM, Koepsell TD, van Belle G Cigarette smokmg, alcohol use, and subarachnoid hemorrhage Stroke 1992,23 1242-1249

26 Newby DE, Wnght RA, Labmjoh C, Ludlam CA, Fox KAA, Boon NA, Webb DJ Endothehal dysfunction, impaired endogenous fibnnolysis, and cigarette smokmg Circulation 1999,991411-1415

27 Schneiderman J, Bordin GM, Engelberg I, Adar R, Seiffert D, Thmnes T, Bernstein EF, Dilley RB, Loskutoff DJ Expression of fibnnolytic genes in atherosclerotic abdominal aortic aneurysm wall a possible mechamsm for aneurysm expansion J Clin Invest 1995,96639-645

28 Dallabnda SM, Falls LA, Faneil DH Factor XHIa Supports micro-vascular endothehal cell adhesion and inhibits capillary tube formation in fibrm Blood 2000,95 2586-2592

29 Suzuki K, Henke J, Iwata M, Henke L, Tsuji H, Fukunaga T, Ishimoto G, Szekelyi M, Ito S Novel polymorphisms and haplotypes m the human coagulation factor XIII A-subumt gene Hum Genet 1996,98393-395 30 Corral J, Iniesta JA, Gonzalez-Conejero R, Villalon M, Vicente V

(7)

Editorial Comment

Hematology-Neurology Connection: Association Between Factor XIII and

Hemorrhagic Stroke in Young Women Through Genetic Polymorphism

Coagulation factor XIII (FXIII) belongs to the enzyme family of transglutammases that catalyze the formation of covalent e-(7-glutamyl)lysme (amide) bonds on cross-lmked protems ' Protein cross-lmkmg is mvolved m several physiological processes, mcludmg hemostasis, wound healmg, and tumor growth FXIII is a proenzyme that is activated by thrombm dunng the late stage of blood coagulation It is a heterotet-ramer consistmg of 2 catalytic A (FXIIIA) and 2 noncatalytic B (FXIIIB) subunits On thrombin activation, the enzyme (le, activated FXIII) catalyzes the formation of covalent bonds among fibrm monomers Activated FXIII also promotes clot retraction2 by bmdmg to platelet GPIIb/IIIa receptors (at-tached to platelet microskeleton via its cytoplasmic domain) and cross-links fibrm with a number of other protems, mcludmg fibronectm and a2-antiplasmin The end result is the formation of sturdier thrombi, more resistant to shear- and fibnnolysis-induced lysis These actions of activated FXIII explam m part why older thrombi are more difficult to lyse with thrombolytic agents FXIII has a long plasma half-hfe of 6 to 10 days, and hmited amounts are reqmred for normal hemostasis (0 05 U/mL) Congemtal FXIII deficiency is exceedmgly rare It is inhented äs an autosomal recessive trait and caused by absence of either subunit Deficiency of the FXIIIA subunit is the most common, however Homozygotes have a lifelong bleedmg diathesis, typically starting with bleeding from the umbilical cord at birth and later character-ized by delayed bleedmg after trauma or surgery and delayed wound bleeding Dunng childhood, patients bruise easily and expenence subcutaneous and mtramuscular hematomas Spontaneous abortion has been reported in affected females, presumably owmg to defective formation of a cytotropho-blastic shell 3 Hemorrhage into the central nervous System occurs more frequently than with other inhented coagulation defects Other conditions mvolvmg defective fibrm formation (eg, dissemmated intravascular coagulation) or fibrm break-down (eg, thrombolytic admimstration) are also associated with central nervous System bleeding These observations suggest the theme that faulty fibrm formation or fibrm breakdown is associated with central nervous System bleed-ing Most cases of congemtal FXIII deficiency are due to mutations m the subunit A gene, often mvolvmg pomt mutations causmg ammo acid Substitution In addition, sev-eral smgle nucleotide polymorphisms of the FXIIIA gene resultmg m smgle ammo acid substitutions have been de-scnbed Genetic polymorphisms are typically encountered m at least 1% of the normal population The dissection of human disease through genetic polymorphisms has become an im-portant research area of vascular biology and other disciplmes m recent years

In the precedmg article, Reiner et al exammed the associ-ation between the Val34Leu, Tyr204Phe, and Pro564Leu variants of FXIIIA and the nsk of nonfatal hemorrhagic stroke m young women by conductmg a population-based case-control study They also exammed possible gene-gene mteractions between FXIIIA polymorphisms and the 4G/5G Promoter polymorphism of the PAI-1 gene P AI-1 is the mam mhibitor of tissue plasmmogen activator and is released from activated endothehal cells and the α-granules of activated platelets The mam results show that women carrymg the Tyr204/Phe204 or Leu564/Leu564 genotype have a sigmfi-cantly higher nsk of hemorrhagic stroke compared with control subjects (age-adjusted ORs of 2 9 and 4 3, respective-ly) This nsk is further magnified when either genotype is combmed with the PAI-1 5G/5G alleles (age-adjusted OR 189) The authors conclude that the Phe204 allele and Leu564/Leu564 genotypes may be markers of genetic sus-ceptibihty to nonfatal hemorrhagic stroke m young white women

(8)

studies raise the hypothesis that some FXIIIA polymorphisms may be somewhat protective agamst thrombosis but are perhaps mvolved m the pathogenesis of hemorrhagic disor-ders These prehmmary findmgs will need to be mvestigated further m future studies

In view of their widespread availability, molecular biology techniques have now entered the clmical tnal arena Studies of genetic polymorphisms abound m practically all disci-plmes of clmical research The underlymg prmciple is based on the notion that phenotypic vanations among mdividuals, includmg anthropometncs, disease susceptibihty, and re-sponse to the environment, are mostly due to naturally occumng vanations in DNA sequence, consistmg of smgle nucleotide polymoφhlsms (SNPs) SNPs are evenly distnb-uted throughout the genome and thus likely flank or are nearby codmg genes 7 Association studies examme whether

polymoφhlc alleles occur more frequently m cases than m controls If a sigmficant association occurs, either the poly-morphism is closely associated with the susceptibihty locus or is m hnkage disequihbnum However, these associations do not necessanly imply a causality link with the disease under study and, like any other case-control study, are susceptible to bias The genetic markers identified through studies must be viewed äs nsk factors, albeit at the molecular

level Two advantages of genetic markers over epidermolog-ical nsk factors or biochemepidermolog-ical markers are that their expo-sure always precedes the disease of mterest and the underly-mg disease cannot influence test results These 2 factors are common sources of bias in classic case-control studies The

vahdity of polymorphisms äs genetic markers can be ascer-tamed only through several studies showing consistent re-sults Perhaps one of the best examples to cite was the demonstration of the angiotensin-convertmg enzyme inser-tion/deletion (I/D) polymorphism äs marker of susceptibihty to cardiovascular disease

Jacques R. Leclerc, MD, FRCP(C), Guest Editor Lilly Research Laboratories Indianapolis, Indiana

References

1 Ichmose A Physiopathology and regulation of factor XII Thromb Haemost 2001 86 57-65

2 Cox AD, Devme DV Factor XHIa bmding to activated platelets is mediated through activation of glycoprotem Ilb-IIIa Blood 1994,83 1006-1016

3 Asahma T, Kobayashi T, Okada Y, Goto J, Terao T Maternal blood coagulation factor XIII is associated with the development of cytotropho-blastic Shell Placenta 2000,21 388-393

4 Wartiovaara U, Perola M, Mikkoka H, Totterman K, Savolamen V, Penttila A, Grant PJ, Tikkanen MJ, Vartiamen E, Karhunen PJ, Peltonen L, Palotie A Association of FXIII Val34Leu with decreased nsk of myocardial mfarction m Finnish rnales Atherosclerosis 1999,142 295-300

5 Elbaz A, Poiner O, Canaple S, Chedru F, Cambien F, Amarenco P The association between the Val34Leu polymorphism in the factor XIII gene and bram mfarction Blood 2000,95 586-591

6 Anens RS, Phillippou H, Nagaswami C, Weisel JW, Lane DA, Grant PJ The factor XIII V34L polymorphism accelerates thrombin activation factor XIII and affects cross-lmked fibrm structure Blood 2000,96 988-995