Genetic Dissection of Quantitative Trait Loci for Hemostasis and Thrombosis on Mouse Chromosomes 11 and 5 Using Congenic and Subcongenic Strains

Susceptibility to thrombosis varies in human populations as well as many inbred mouse strains. Only a small portion of this variation has been identified, suggesting that there are unknown modifier genes. The objective of this study was to narrow the quantitative trait locus (QTL) intervals previously identified for hemostasis and thrombosis on mouse distal chromosome 11 (Hmtb6) and on chromosome 5 (Hmtb4 and Hmtb5). In a tail bleeding/rebleeding assay, a reporter assay for hemostasis and thrombosis, subcongenic strain (6A-2) had longer clot stability time than did C57BL/6J (B6) mice but a similar time to the B6-Chr11A/J consomic mice, confirming the Hmtb6 phenotype. Six congenic and subcongenic strains were constructed for chromosome 5, and the congenic strain, 2A-1, containing the shortest A/J interval (16.6 cM, 26.6 Mbp) in the Hmtb4 region, had prolonged clot stability time compared to B6 mice. In the 3A-2 and CSS-5 mice bleeding time was shorter than for B6, mice confirming the Hmtb5 QTL. An increase in bleeding time was identified in another congenic strain (3A-1) with A/J interval (24.8 cM, 32.9 Mbp) in the proximal region of chromosome 5, confirming a QTL for bleeding previously mapped to that region and designated as Hmtb10. The subcongenic strain 4A-2 with the A/J fragment in the proximal region had a long occlusion time of the carotid artery after ferric chloride injury and reduced dilation after injury to the abdominal aorta compared to B6 mice, suggesting an additional locus in the proximal region, which was designated Hmtb11 (5 cM, 21.4 Mbp). CSS-17 mice crossed with congenic strains, 3A-1 and 3A-2, modified tail bleeding. Using congenic and subcongenic analysis, candidate genes previously identified and novel genes were identified as modifiers of hemostasis and thrombosis in each of the loci Hmtb6, Hmtb4, Hmtb10, and Hmtb11.


Introduction
Cardiac and peripheral vascular diseases are the leading cause of death, with thrombosis as the precipitating event. Numerous studies [1][2][3][4][5][6] show that inherited risk factors contribute significantly to the development of acute myocardial infarction, ischemic stroke, atherosclerosis, thromboembolism and peripheral arterial occlusion. Progress has been made in diagnosis and treating these diseases, but only few of the genetic factors has been identified that explain thrombosis variation.
Approaches to identify genetic determinants of phenotypic traits have advanced rapidly in the last 15 years. Previously, identification relied on single gene spontaneous mutations that caused an overt phenotype [7][8][9]. The development of strategies to delete or insert human genes into mice led to the association of specific gene products with specific phenotypes [10][11][12]. Results of some of these deletions or insertions were surprising; in some cases, there was no obvious phenotype where one was anticipated, and in others, unanticipated phenotypes arose. In addition, murine strain backgrounds were shown to modify the phenotypes of expressed genes [13,14]. With the realization that most common chronic diseases in humans, such as atherosclerosis, obesity or diabetes, are multigenic with more than one gene and even unanticipated gene interactions contributing to disease, new approaches for gene identification and characterization were needed. The development of statistical analyses packages to map complex traits in mice has provided an approach to such complex diseases [15]. These approaches allow for the mapping of loci of genes that contribute quantitatively to a particular phenotype or pathology, a quantitative trait locus (QTL) [16]. The loci of a number of genes for obesity [17][18][19][20][21] and atherosclerosis [22][23][24][25][26][27][28] have been successfully mapped by such QTL analyses. The first step in QTL analysis is the identification of two phenotypically disparate strains, performing the strain intercross, and phenotyping and genotyping F2 progeny. Both inbred strains and recombinant congenic strains (strains with essentially the same background except for selective differential chromosomal segments) [29], have been used in such QTL mapping. Only a few studies [14,30] have investigated risk factors for thrombotic disease using mouse genetic approaches.
An approach that we have utilized is to analyze QTL with the chromosome substitution strains (CSS) [31]. In consomic strains the differential variation is a whole chromosome rather than just a chromosomal segment. This approach has been applied to the B6 and A/J mice, which have been characterized to have differences in many traits [32,33]. A panel of 21 CSS was constructed by a "marker-assisted" breeding program. Each of the 21 chromosomes were substituted individually with an A/J chromosome [34] into the B6 background. Often many more QTL are identified with the CSS than with genome-wide scans. In a previous study [35], we found that the two inbred strains, C57BL/6J (B6) and A/J, had marked differences in hemostatic and thrombotic responses. In a tail bleeding/ rebleeding assay, rebleeding time (clot stability) was longer in A/J mice compared to B6 mice. In addition, A/J mice had a rapid occlusion time after FeCI 3 injury of the carotid and reduced patency. The tail bleeding/rebleeding assay, a reporter of hemostasis and thrombosis, was used as screening assay for a panel of consomic strains (B6-Chr1-19, X, Y A/J ), which had one chromosome from A/J mice in a B6 background [35]. The B6-Chr11 A/J (CSS-11) and B6-Chr5 A/J (CSS-5) had a phenotype similar to A/J mice, long clot stability. We identified the QTLs, Hmtb6, Hmtb4, and Hmtb5, for hemostasis and thrombosis on mouse chromosomes 11 and 5 in an F 2 intercross from CSS-11 X B6 and CSS-5 X B6 [36]. The goal of this study was to fine map the QTLs, Hmtb6, Hmtb4, and Hmtb5, using congenic and subcongenic mouse strains. Here we report confirmation of the QTLs for clot stability in Hmtb6 and Hmtb4 and the QTL for bleeding in Hmtb5. In addition, another QTL for bleeding, Hmtb10, and a QTL for carotid occlusion time after injury, Hmtb11, were identified in the proximal region.

Mice
The parental inbred mouse strains C57BL/6J (B6, 000664) and A/J (000646), and the C57BL/6J-Chr 5A/J/NaJ (CSS-5, 004383) and C57BL/6J-Chr 11A/J/NaJ (CSS-11, 004389) mouse strains were purchased from the Jackson Laboratory (Bar Harbor, ME) and maintained in the Biological Resources Unit facility at the Cleveland Clinic. Congenic and subcongenic strains were generated from CSS-5 and CSS-11 consomic mice using marker-assisted backcrossing [31]. Mice were housed in sterilized isolator cages with a 14hr/10hr light/dark cycle and ad libitum access to water and food. The bleeding/ rebleeding assay was performed on both male and female mice at 6-9 wk. of age. The Cleveland Clinic Institutional Animal Care and Use Committee approved this study and procedures were followed in accordance with institutional guidelines.

Genotyping
Genomic DNA was prepared from ear punches of the mice and genotyping performed using polymerase chain reaction (PCR) for microsatellite markers (Mouse Mappairs, Invitrogen, Carlsbad, CA). PCR was performed using HotstarTaq Master Mix Kit (QIAGEN, Valencia, CA). The PCR products were detected by electrophoresis on 10% polyacrylamide gels (National Diagnostics, Atlanta, GA) and visualized by ethidium bromide staining. Microsatellite markers and primers for the RLF markers were used to identify mice by marker assisted selection [31,34] (Tables S1 and S2). Marker and genomic coordinates of genes were determined from the Mouse Genome Database (MGD), 2012 [37]. The following terms were used for the annotated analysis: hemostasis, thrombosis, clotting, platelets, fibrinolysis, bleeding, blood and coagulations. Genes were subjected to functional annotation programs, David Bioinformatics Resources 6.7 (http:// david.abcc.ncifcrf.gov/home.jsp) [38] and the MGD, 2012 [37].

Tail Bleeding/Clot Stability Assay
Phenotyping was performed using the bleeding/rebleeding assay as previously described [35]. The mice were anesthetized with ketamine/xylazine (90 mg/kg, 10 mg/kg), the tail prewarmed for 5 minutes in 10 mL of saline at 37°C in a water bath. The tail was lifted from the saline and a 5 mm tail segment amputated and immediately returned to the saline. Bleeding time was measured as the time between the start of the bleeding and cessation of the bleeding. Clot stability time was measured as the time between the cessation of the bleeding and the start of the second bleeding.

FeCl 3 -induced Carotid Injury model
To induce thrombus formation in the carotid artery, a ferric chloride (FeCl 3 ) model of vessel injury was employed as previously described [35]. Mice were anesthetized with ketamine/xylazine (90 mg/kg, 5 mg/kg), a midline cervical incision was made and the left common carotid artery isolated by blunt dissection. The flow probe (Transonic Systems, model 0.5PSB) was placed under the artery and when a stable baseline was reached, a 0.5 × 2 mm strip of filter paper saturated with 10% FeCl 3 solution was applied to the surface of the carotid artery for 3 minutes. Occlusion time was determined from the addition of the FeCl 3 solution to the occlusion of the artery (minimum blood flow). There was no difference in baseline blood flow data in carotid arteries among the mouse strains. The flow probe (Model 0.5PSB,Transonic Systems, Ithaca, NY) was in place from baseline measurements to several minutes after the stable occlusion had been reached, or stopped at 30 min if it had not occluded. Blood flow was recorded every 10 sec (Model TS420, Transonic Systems).

CaCl 2 Abdominal Aorta Induced Aneurysm (AAA) Formation
AAA was induced by periaortic application of CaCl 2 [39]. Mice at age 6~8 weeks were anesthetized (90mg /kg ketamine/ 5mg/kg xylazine) and the abdominal aorta between the renal arteries and bifurcation of the iliac arteries was isolated from surrounding tissue. An image of the aorta was acquired by Spot Image 4.0.3 through a cooled CCD camera attached to an Olympus SZ-PT dissection microscope. After image acquisition, cotton gauze (1 x 0.5 cm) that had been soaked in 0.5 M CaCl 2 solution was applied to the external surface of the aorta. After 10 minutes the aorta was rinsed with 0.9% sterile saline and the incision was closed. Three weeks later, after the mice were anesthetized and the aorta exposed after dissecting away the surrounding fatty tissue and scarring adhesions, images were collected as above. The aorta diameter was measured from the before and after CaCl 2 images by ImagePro software.

Statistics
Data are presented as the mean ± SEM. One-way ANOVA and Newman-Kuels post-test were used to determine statistical differences between groups. A P value of <0.05 was considered significant.

Chromosome 11 Congenic and Subcongenic Strains
Our previous study, using F2 progeny from a cross of CSS-11 and B6, identified a QTL for clot stability time in the bleeding/rebleeding assay on mouse distal chromosome 11 that peaked at the genetic marker D11Mit336 [36]. To verify this QTL (Hmtb6), congenic strains were constructed by crossing the CSS-11 mice and B6 mice to produce the congenic strain 6A-4 and the subcongenic strain 6A-2. The position of the A/J regions is indicated in Figure 1A, and boundary markers are summarized in Table 1. The minimum congenic interval is the interval between markers with A/J genotypes on the B6 background. The maximum interval is the interval between two markers with B6 genotypes just outside the minimum interval. The interval between minimum and maximum intervals are of unknown genotype [29]. The minimum A/J interval for 6A-4 is 20.8 Mbp, and for 6A-2, 2.9 Mbp.

Bleeding Time in Chromosome 11 Congenic and Subcongenic Mice
The homozygous congenic and subcongenic mice were phenotyped using the tail bleeding/rebleeding assay as described in the Methods. The higher bleeding time value for CSS-11 mice was not statistically significant (P > 0.05) compared to B6 mice, and the bleeding times in 6A-4 mice and 6A-2 mice were similar to the bleeding time in the B6 mice ( Figure 1B).

Clot Stability Time in Chromosome 11 Congenic and Subcongenic Mice
Similar to the CSS-11 parental strain (417±71 sec, n=10), the congenic mice (6A-4) exhibited significantly longer (P<0.05) clot stability times (311±40 sec, n=27) compared to the B6 strain (203±25 sec, n=27) ( Figure 1C). The clot stability time was 53% longer in the congenic strain than for B6. To further narrow the QTL interval, a subcongenic strain, 6A-2, was developed ( Figure 1A, Table 1). The clot stability time (454 ± 57 sec, n=12) in this subcongenic strain mice was 2.2-fold longer (P=0.01) compared to B6 mice ( Figure 1C), and compared to B6 mice, the 2A strain accounts for 100% of the increase in clot stability time for CSS-11, whereas the 4A strain accounts for only 50% of the increase suggesting the trait is isolated to the D11Mit258 to D11Mtt336 region. Although the region from D11Mit258 and D11Mit336 was genotyped as A/J, we cannot completely rule out the uncertain regions as harboring the gene(s) responsible for the phenotype of long clot stability. Hmtb6-2A F1, clot stability time (382 ± 65, n=11), was 2-fold longer (P=0.05) than for B6 mice, suggesting dominant inheritance of the QTL. These results demonstrate that this distal region on chromosome 11 contains a genetic factor (or factors) that dominantly controls the clot stability time. This Hmtb6 QTL was narrowly mapped to a minimum region of 2.9 Mbp. In this region 107.6-110.5 Mbp region on Chromosome 11 there are 25 protein-coding genes (Table S3, blue area) and when the two adjoining regions of uncertain genotype were included, there are 429 genes (Table S3). Annotation of these genes Identified 14 genes additional genes related to hemostasis, thrombosis, clotting, platelets, coagulation, bleeding, blood, or fibrinolysis ( Table 2). The genes in the A/J fragment are Abca5, Apoh, Gna13, and the genes in the regions of uncertain genotype are Ace, Adam11, Cntnap1, Gjc1, Itga2b, Itgb3, Pecam1, Plcd3, Cd300lf, Hn1, and Jmjd6 (Table 2A).

Chromosome 5 Congenic and Subcongenic Strains
Congenic and subcongenic strains were constructed to verify the overlapping QTL on chromosome 5, Hmtb4 and Hmtb5, for clot stability and bleeding. Congenic strain 3A-2 and The minimum congenic interval is the interval between markers with A/J genotypes on the B6 background. The maximum interval is the interval between two markers with B6 genotypes just outside the minimum interval. The interval between minimum and maximum intervals are of unknown genotype [29]. Marker and genomic coordinates were determined from the Mouse Genome Database (MGD), 2012 [37].

Bleeding Time in Chromosome 5 Congenic and Subcongenic Strains
Bleeding time in the CSS-5 mice (63±7, n=7) was significantly shorter than for B6 mice (121±13, n=28) as previously reported [35] and the congenic strain, 3A-2 (60±9, n=17) also had the same phenotype. The A/J fragment of 3A-2 strain was slightly longer than for the 2A-1 strain with an additional 35.6 Mbp region between D5Mit197 and rs6297441 ( Table 1). The bleeding time of the 2A-1 was the same as the time in B6 mice suggesting the additional region of 3A-2 harbored the short bleeding time. This narrow region of 3A-2 may be the site for the CSS-5 short bleeding time phenotype ( Figure 3A), and the Hmtb5 locus identified in the QTL mapping [36]. Thus, for the Hmtb5 locus, the minimum A/J interval on 3A-2 strain, 100.2-108. 6 Mbp that did not overlap with 2A-1, there are 78 protein-coding genes (Table S4), and functional annotation analysis revealed 11 relevant genes ( Table 2).
In the proximal region of chromosome 5, bleeding times for the subcongenic strains 4A-2 and 4A-1 mice were similar to B6 mice ( Figure 3A). The 3A-1 mice, carrying the A/J interval from the proximal end to the marker D5Mit394, had a longer bleeding time (251±51, n=15, P< 0.01) that was 2-fold longer than A/J or B6 mice ( Figure 3A). This suggests an additional locus for bleeding time on chromosome 5 and is consistent with the QTL analysis that identified a suggestive locus at the proximal end of the chromosome. This locus was named Hmtb10 with the minimum interval of 32.9 Mbp. There are 266 protein-coding genes (Table S5), and the annotation analysis identified 15 genes (Table 2). Thus, there are two loci on chromosome 5 for bleeding, one in the distal region (Hmtb5) and one in the proximal region (Hmtb10).

Carotid and Abdominal Aorta Vascular Injury
Hemostasis and thrombosis are complex polygenic processes. While QTL link genomic regions to phenotypic data, multiple traits within the QTL region may be responsible for a particular phenotype [40][41][42]. To further characterize the phenotype of the Hmtb4, Hmtb5, and Hmtb10 loci, carotid occlusion time after FeCl 3 injury ( Figure 4A) and abdominal aortic aneurysm formation after CaCl 2 injury ( Figure 4B) were assessed in the congenic strains. In the carotid vascular injury model, a thrombus forms and occludes blood flow. The time required for the occlusion time can detect imbalances in coagulation and platelet functions. The strains 6A-1, 3A-1 and 3A-2, had occlusion times that were not statistically different from B6 or CSS-5 mice ( Figure 4A), but the values for 3A-1 and 3A-2 mice were closer to the CSS-5 mice. The 4A-2 strain had a prolonged occlusion time that was 4-fold longer than B6 mice (P < 0.01) and similar to the CSS-5 strain. In addition, the 4A-2 strain had the lowest response to the CaCl 2 injury and the increase in diameter of the congenic was similar to the CSS-5 mice. The results of these two models, carotid injury and abdominal aorta injury, revealed that another locus, related to thrombosis modification, occurs within this region on chromosome 5, Hmtb11. Compared to 3A-1 mapped at rs1680965 and D5Mit394, the 4A-2 congenic was genotyped at rs1680965 and D5Mit197, only a difference of 10 Mbp. Within this additional region there were only 10 additional protein coding genes and they did not associate with the annotation criteria (Table S7). One possibility is that the regions from 0-21.4 Mbp (rs1680965) in the two congenic strains, 3A-1and 4A-2, were genetically different. In this region (0-21.4 Mbp) there are 89 protein-coding genes (Table S8). Annotation analysis identified 17 genes in this region (Table 2).

CCS-17 Backcross with Chromosome 5 Congenics
Previously, we reported an interaction of the genes on chromosome 5 and chromosome 17 that modified bleeding and clot stability in the tail bleeding/rebleeding assay [35]. To test if we could determine if this interaction was in the proximal or distal region on chromosome 5, the 3A-1 (proximal) mice and 3A-2 (distal) mice were crossed with CSS-17 mice ( Figure 5A). When the 3A-1 mice with the long bleeding times were crossed with CSS-17 mice, the 17 x 3A-1 mice now had short bleeding times (60±9, n=9) suggesting an inhibition by chromosome 17 on the long bleeding times of 3A-1 mice. When the 3A-2 mice with the short bleeding times were crossed with CSS-17 mice, the 17 x 3A-2 mice had longer bleeding times (115±31, n=24) compared to the 3A-2 mice suggesting an interaction of chromosome 17 with this region. Thus, the genes of A/J chromosome 17 interacted with Hmtb5 and Hmtb10 of chromosome 5 to modify bleeding time.
For clot stability time ( Figure 5B), there was no difference between, the 3A-2 and 17 x 3A-2 or between 3A-1 and 17 x 3A-1. The clot stability time in the F1 progeny of the CSS-5 and CSS-17 mice was similar to the value for B6 mice [35], and in the cross of CSS-17 and CSS-5 (17 x 5) the long clot stability time was recovered. Taken together this suggests that long clot stability time requires two alleles for expression, two in the CSS-5 or two from CCS-17 or one allele from each. The one allele from chromosome 17 and either the locus Hmtb5 (3A-2) or Hmtb10 (3A-1) in the cross were sufficient for the expression of the phenotype.

Discussion
In this study, we confirmed the previously identified QTL, Hmtb6, Hmtb4, and Hmtb5 for hemostasis and thrombosis on mouse chromosomes 11 and 5 in congenic and subcongenic strains. Two additional QTL have also been identified, Hmtb10 and Hmtb11. To our knowledge, this is the first study to confirm QTLs associated with hemostasis and thrombosis using congenic and subcongenic mouse strains. Figure 6 summarizes the QTL and potential candidate genes. The phenotype of the congenic and subcongenic strains demonstrates that the gene (or genes) underlying Hmtb6 QTL is located in the distal region on mouse chromosome 11, with a minimum interval of 2.9 Mbp. This region contains 25 known and predicted genes and corresponds closely to a syntenic region on human chromosome 17, from q12 to q25.3. No QTL has been identified in this region in human populations for hemostasis and thrombosis. The only gene known to be related to hemostasis and thrombosis on mouse chromosome 11 is Serpinf2 (α 2 -antiplasmin), an inhibitor of plasmin, the primary  enzyme for clot lysis. Serpinf2 is located at 75.2 Mbp and was excluded from the candidates for this trait from our previous study [36]. This conclusion was further confirmed in our current study. Interestingly, Mohlke et al. [43] identified a QTL, Mvwf as a modifier of von Willebrand factor (VWF), that regulates plasma VWF level on mouse chromosome 11, in a region between markers Ngfr and Hoxb9 (95.5-96.2 Mbp). They further found that alterations in Galgt2, a candidate at the QTL locus at 95.8 Mbp on chromosome 11, influenced plasma VWF levels [44]. This locus was noted in the Hmtb6 QTL in the previous study [36], and our current study clearly demonstrates that Mvwf is located proximal to the Hmtb6 locus and outside the maximal interval (100-121.8 Mbp), indicating that a gene (or genes) other than Mvwf on chromosome 11 regulates hemostasis and thrombosis. Annotation analysis revealed 3 genes in the Hmtb6 QTL, Abca5, Apoh, and Gna13. Abca5 deficient mice develop a dilated cardiomyopathy with large thrombi [45,46]. Gna13 plays a role in platelet -granule release [47][48][49]. Apoh, more commonly known as beta-2-glycoprotein 1, binds to cardiolipin, inhibits platelets activation, and is involved in anticoagulation [50][51][52]. These genes have not been commonly associated with thrombosis or hemostasis in humans, and may be novel modifiers. In addition to Abca5 there are 4 other ATP-binding cassette proteins, and besides Gna13, Rgs9 is also associated with G-protein signaling. Three genes are involved in calcium transport, Cacng1, Cacng4 and Cacng5, and three genes code for protein kinases, Map2k6, Prkar1a, and Prkca. Wipi1 and Arsg are associated with the Golgi and endoplasmic reticulum. In the regions of uncertain genotypes of Hmtb6, there were three genes, Itga2b, Itgb3, and Pecam-1 that are associated with bleeding and thrombosis. Mutations of Itga2b and Itgb3 are assoicated with Glanzmann thrombasthenia, the most common inherited disease of platelet function in humans, causing a bleeding disorder [53]. Pecam-1 deficient mice have reduced occlusion time in the formation of a thrombus [54].
Previously [36] we detected a significant locus, Hmtb4, on mouse chromosome 5 for clot stability with a 1-LOD confidence interval 55-66 cM (112.9-127.6 Mbp). From the congenic mapping we confirmed the phenotype of this locus with increased clot stability similar to the consomic strain CSS-5, compared to the parental strain, B6 mice. Functional annotation analysis identified 18 relevant genes. Genes associated with human syndromes or diseases include: polymorphisms in Hps4 and Vps33a genes, which are, associated with Hermansky-Pudlak syndrome type 4 and the absence of platelet dense granules [9,55]. Orai, the major calcium channel in the platelet plasma membrane, has recently been reported to play a role in platelet activation and thrombus formation [56]. Ptpn11 is a member of the protein tyrosine phosphatase family, and mutations are identified with Noonan syndrome, one of the most common genetic disorders of congenital heart defects [37,57]. Mice homozygous null for P2rx4, a purinergic receptor, ligand gated ion channel 4, have hypertension, abnormal artery morphology and vascular remodeling [58]. The Sh2b3, adaptor protein 3 or LNK deficient mice have normal tail bleeding time, but in 60% of the deficient mice rebleeding occurred within 60 seconds, suggesting the clot was unstable. In the 3A-2 congenic strain tail bleeding was reduced and rebleeding inhibited, but similar to the Lnk-/-mice, FeCl 2 -induced occlusion time was higher in the CSS-5 mice than in B6 mice and similar to the 3A-2 mice. Selplg (Pselectin) is a cell adhesion molecule stored in -granules in platelets and Weibel-Palade bodies in endothelial cells and when these cells are activated P-selectin is released to the surface [59,60]. P-selectin is important in the recruitment of leukocytes to sites of injury [61]. Adam1a and Adam1b have disintegrin metallopeptidase domains that have been implicated as inhibitors of blood coagulation, but these two proteins have not previously been associated with coagulation [62]. From the previous study mapping the QTLs, the region of the 126.8 -138. 6 Mbp was outside the 95% confidence region for the Hmtb4 [36]. The annotated gene in this region was PAI-1, and we did not find any difference in mRNA expression of PAI-1 in the parent strains, and PAI-1 antigen was not different in the congenic strains. It's not likely this is the candidate gene in this QTL. Although considered outside the region of confidence from the previous QTL mapping, there are two known genes that could be candidate genes, Pf4 and Pdgfr, in the proximal region of uncertain genotype in 3A-2.
Hmtb5 locus for bleeding with 1-LOD confidence interval, 45-75 cM (113-128 Mbp) at the distal end of chromosome 5, was previously identified [36]. The congenic strain, 3A-2 with the A/J interval of 100.2-108.6 Mbp, had a short bleeding time similar to the CSS-5 mice suggesting this region was the site of the suppressed bleeding time phenotype in CSS-5. In the subcongenic strain, 2A-1, bleeding time was similar to B6 mice. In the region (100.2-108.6Mbp) of 3A-2 that did not overlap with 2A-1 there were 11 annotated genes for Hmtb5. A potential candidate gene in this region is Hpse, heparanase, alpha-endoglucuronidase, a the major enzyme for heparin sulfate degradation. Diabetes and vascular injury increase the expression of heparanase and transgenic mice with overexpression of heparanase increase thrombosis after mild vascular injury [63,64].
This study has identified a new QTL, Hmtb5 for bleeding in the proximal region of chromosome 5. From the congenic mapping we confirmed the locus identified at the RFLP marker, rs1680965, in the QTL mapping for long bleeding time. Annotation analysis identified 15 genes in this region. Two of these genes are associated with blood pressure regulation, Drd5 [62,65] and Nos3 [66][67][68]. The cappuccino deficiency, a spontaneous mutation of Bloc1s4, is associated with reduced platelet dense granules and prolonged bleeding [69,70]. Another gene of interest is EMILIN1. We identified EMILIN2, with a high homology to EMILIN1, in the QTL on chromosome 17 [71,72], in the Hmtb8 locus and found a thrombosis phenotype. The results of the tail bleeding/rebleeding assay and FeCl 3 injury model in EMILIN1 deficient mice were not different than for B6 mice (Table S9) suggesting EMILIN1 is not a causative gene in this QTL.
An additional new QTL was uncovered, Hmtb11 in the congenic strain, 4A-2, that had an increased occlusion time in the carotid injury model and was protected from abdominal aneurysm after CaCl 2 injury, which suggests another locus in the proximal region. Genes of particular note from the annotation search are CD36 that is associated with platelet glycoprotein IV deficiency [73]. Fgl2, fibrinogen-like protein2, contributes to immunologically mediated thrombosis in experimental and human viral hepatitis [74].

Limitations of study and future directions
Using congenic and subcongenic mice from two consomic strains, CSS-11 and CSS-5, three QTLs and two additional QTLs were fine-mapped for thrombosis and hemostasis modifiers. Sixty-four potential candidate genes were identified. The limitation of the study is that the causative gene has yet to be identified. Further fine mapping is needed with additional subcongenic strains, as well as expression QTL analysis and quantitative RT-PCR. Future directions will include identification of underlying polymorphisms located in the coding, noncoding or regulatory regions and translating these findings to variation in human genes to assess the relationship of genetic variation to hemostasis and thrombosis risk.