A viable mouse model of factor X deficiency provides evidence for maternal transfer of factor X

Background:Activated factor X (FXa) is a vitamin K-dependent serine protease that plays a pivotal role in blood coagulation by converting prothrombin to thrombin. There are no reports of humans with complete deficiency of FX, and knockout of murine F10 is embryonic or perinatal lethal. Objective:We sought to generate a viable mouse model of FX deficiency. Methods:We used a socket-targeting construct to generate F10-knockout mice by eliminating F10 exon 8 (knockout allele termed F10tm1Ccmt, abbreviated as ‘−’; wild-type ‘+’), and a plug-targeting construct to generate mice expressing a FX variant with normal antigen levels but low levels of FX activity [4–9% normal in humans carrying the defect, Pro343→Ser, termed FX Friuli (mutant allele termed F10tm2Ccmt, abbreviated as F)]. Results:F10 knockout mice exhibited embryonic or perinatal lethality. In contrast, homozygous Friuli mice [F10 (F/F)] had FX activity levels of ∼5.5% (sufficient to rescue both embryonic and perinatal lethality), but developed age-dependent iron deposition and cardiac fibrosis. Interestingly, F10 (−/F) mice with FX activity levels of 1–3% also showed complete rescue of lethality. Further study of this model provides evidence supporting a role of maternal FX transfer in the embryonic survival. Conclusions:We demonstrate that, while complete absence of FX is incompatible with murine survival, minimal FX activity as low as 1–3% is sufficient to rescue the lethal phenotype. This viable low-FX mouse model will facilitate the development of FX-directed therapies as well as investigation of the FX role in embryonic development.


Introduction
The hemostatic process is a complex, highly regulated and dynamic sequence of events that ultimately results in the localized formation of a stable clot in response to vascular damage [1]. Activated coagulation factor X (FXa; a vitamin Kdependent, two-chain plasma serine protease) catalyzes the conversion of prothrombin to thrombin and occupies a pivotal role in the coagulation cascade. Under physiological conditions, cleavage of FX to form FXa is initiated by the FVIIa/ tissue factor (TF) complex and maintained by the FIXa/FVIIIa complex.
Several lines of evidence suggest that FX/FXa may have functions other than hemostasis, such as promoting cytokine release from endothelial cells and initiating signaling via protease activated receptors (PARs) [2,3]. Interestingly, no humans with deletion of both F10 genes have been reported. Instead, all of the described mutations are either missense mutations or compound heterozygotes with a deletion or stop codon on one allele and a missense mutation on the other, suggesting that complete deficiency of FX is incompatible with life (http://www.hgmd.cf.ac.uk/ac/index.php) [4]. Similar findings have been described in mice, in which targeted disruption of the F10 gene is associated with loss of 50% of homozygous affected animals in utero, and loss of the remaining affected mice as a result of bleeding complications within a few days of birth [5].
In this study, we sought to develop a viable mouse model of FX deficiency and to determine whether very low levels of FX activity could rescue the embryonic and perinatal lethality. We used a Ôplug and socketÕ targeting strategy; the ÔsocketÕ targeting construct eliminated exon 8 and its 3¢ flanking sequence following homologous recombination, to give rise to F10-knockout mice (knockout allele F10 tm1Ccmt abbreviated as Ô-Õ; wild-type F10 allele abbreviated as Ô+Õ) exhibiting embryonic and perinatal lethality [6]. We chose exon 8 as almost 50% of the reported F10 mutations are located within this exon. We used a second (plug) targeting construct to reconstitute the F10 gene with a variant exon 8 containing a Pro 343 fi Ser substitution (FX Friuli, allele F10 tm2Ccmt abbreviated as F), which, in humans homozygous, results in FX activity levels of 4-9% and normal antigen levels [7]. Homozygous FX Friuli mice [F10 (F/F)], with expression directed by the endogenous F10 promoter, yielded FX activity levels of 5.5%, and complete rescue of embryonic and perinatal lethality of FX deficiency. We used this genetic approach to further define the role of FX in embryonic survival and generated mice with the F10 ()/F) genotype. We demonstrate that FX activity levels of 1-3% were sufficient for complete rescue of embryonic and perinatal lethality, and provided evidence for a contribution of maternal transfer of FX activity to embryonic survival. These mice can be used in studies of novel therapies for FX deficiency as well as probing the role of FX in processes such as sepsis and metastasis.

Isolation of murine F10 genomic DNA
Designed from the exon 8 sequence of the murine F10 gene, a 350-bp fragment from 129/SvJ mouse genomic DNA (Stratagene, La Jolla, CA, USA) was amplified using polymerase chain reaction (PCR) and sequenced. This was used as a probe to screen a bacterial artificial chromosome (BAC) library derived from HindIII, partially digested embryonic stem (ES) 129/SvJ DNA (Genome Systems, St Louis, MO, USA). A 7-kb NheI-digested fragment was isolated from a positive BAC clone containing exons 7 and 8, the intervening intron, and the flanking region downstream of exon 8. This fragment also was used to design and construct the F10 socket-targeting vector.

Construction of the F10 socket-and plug-targeting vectors
Engineered in the pPD20 plasmid vector (provided by Dr Randy Thresher, UNC-Chapel Hill, USA), the F10 sockettargeting construct consists of 5¢ (2.8-kb including intron 6, exon 7, and part of intron 7 of the murine F10 gene) and 3¢ [1.6-kb containing the sequence downstream of the putative polyadenylation signal (ATTAAA) and the overlapping termination codon (TAA) of the F10 gene] homologous sequences flanking a functional neomycin (Neo) gene and a partially deleted, non-functional HPRT minigene (DHPRT) (Fig. 1A) [6]. The F10 plug-targeting construct cloned in pBY9 plasmid has the identical 5¢arm as the F10 socket-targeting vector; its 3¢arm contains a 1.4-kb overlapping HPRT sequence with the pPD20 plasmid (Fig. 1E). In the plug-targeting construct, exon 8 has been modified by site-directed mutagenesis at the codon for amino acid 343 (in mouse FX protein) to generate the Pro 343 fi Ser (FX Friuli) mutation (CCC fi TCC) [7].

Electroporation, selection, analysis of ES cell clones and DNA extraction
HPRT-deficient murine ES cells (E14TG2a) were grown as described [8] and electroporated with 100 lg of SacII-linearized F10 socket-targeting vector and G418-selected. For the F10 plug-targeting construct, HPRT-deficient murine ES cells heterozygous for deletion of exon 8 (from the original targeting event) were electroporated as above and selected based on G418 sensitivity and growth in HAT medium [9]. ES cell clone DNA as well as liver DNA from neonates was isolated using the Easy DNA Kit (Invitrogen, Carlsbad, CA, USA). All ES cell clones were identified by PCR and/or Southern blot. For adult mice, genomic DNA was extracted from blood samples using the QIAamp DNA Blood Mini Kit (Qiagen, Valencia, CA, USA).  Generation of F10 ()/)), F10 (F/F), and F10 ()/F) mice The production of all chimeric mice was performed by the Transgenic & Chimeric Mouse Facility of the University of Pennsylvania (USA). Correctly targeted F10 (+/)) and F10 (+/F) murine ES cell clones were expanded, micro-injected into C57BL/6 blastocysts, and implanted into the uterine horns of pseudo-pregnant mice. The chimeric males were mated with wild-type C57BL/6 females to generate heterozygous F10 (+/)) or F10 (+/F) offspring, which were intercrossed to obtain homozygous F10 ()/)) or F10 (F/F) mice, respectively. F10 ()/F) mice were generated by crossing homozygous F10 (F/F) females with heterozygous F10 (+/)) males.
Reverse transcription-polymerase chain reaction and Northern blotting of total RNA isolated from liver tissues Total RNA was isolated from liver of F10 (+/+) and F10 ()/)) newborn mice, using TRIzol reagent (Invitrogen), reverse transcribed into first-strand cDNA using oligo(dT) primers (Invitrogen) and PCR-amplified for exons 1-6 and exon 8. For Northern blot analysis, total RNA (20 lg) was isolated from liver of neonates, and hybridized with a F10 cDNA fragment corresponding to exons 1-6. A b-actin cDNA probe was used as a control. Band intensity was determined by densitometry.

Activated partial thromboplastin time and prothrombin time FX assays
Citrated plasma collected from neonates and adult mice (4-6 weeks old) was used to measure FX coagulation activities by activated partial thromboplastin time (aPTT) and prothrombin time (PT) assays (diluted 1:40 for aPTT, 1:80 for PT), as previously described [11]. Clotting times were converted to percent normal FX activity using a pooled normal mouse plasma standard curve.

Histologic analysis of FX-deficient mice
Adult mice were sacrificed by CO 2 inhalation; organs were harvested and fixed in 10% formalin overnight at 4°C. Tissues were embedded in paraffin, sectioned (5 lm per section), and stained with hematoxylin and eosin, Prussian blue, or MassonÕs Trichrome stain.

Construction of mice with targeted disruption of exon 8
With the expectation that a deletion of murine F10 would result in lethality, which would then require rescue to produce a viable FX-deficient mouse model, we chose to use the Ôplugand-socketÕ targeting strategy originally described by Detloff et al. [6]. As well as being different from the previously described FX deficiency mouse model [5], this approach has the advantage that it allowed us to inactivate and then replace the targeted region with another sequence of interest, which would then express a low-activity FX variant under the control of the endogenous promoter and its regulatory sequences. Using a targeting construct (socket) designed to delete exon 8 (Fig. 1A), we electroporated HPRT-deficient murine ES cells and identified correctly targeted ES clones (8/192 screened, 4%) that were expanded and micro-injected into C57BL/6 blastocysts. The resulting chimeric males were mated with wild-type C57BL/6 females. Subsequently, heterozygous F10 (+/)) mice were intercrossed to obtain homozygous F10 ()/)) offspring.

Knock-in of F10 Friuli and characterization of mice homozygous for variant FX
Clearly, the embryonic and perinatal mortality of the F10 ()/)) mice limit their utility for studies of FX therapeutics. We therefore sought to rescue the mortality by restoring a low level of FX expression under the control of the endogenous F10 promoter. We selected an exon 8 mutation from the human mutation database that is associated with reduced but not absent FX activity, FX Friuli. This missense mutation (Pro 343 fi Ser) in the catalytic domain exhibits normal antigen but reduced activity levels of 4-9% [7]. Given the extensive sequence and functional conservation in the catalytic domain of FX across species, we were able to generate a plug-targeting construct containing the identical mouse F10 gene mutation, expected to have a similar functional effect. We carried out a retargeting experiment on the original, targeted murine ES cells to generate mice expressing the FX Friuli mutation. The use of two selectable markers results in a high rate of correctly retargeted ES cells; 9/48 clones examined (19%), in contrast to 8/192 (4%) from the original targeting event, were correctly targeted as assessed by PCR and Southern blot analyses. The results of the F1 intercross of F10 (+/F) mice are summarized in Fig. 3A [compared to results of the F1 intercross of F10 (+/ )) mice in Fig. 2A]. These show that F10 (F/F) mice are born at the expected frequency based on Mendelian inheritance, confirming that the Friuli variant rescues embryonic lethality. Secondly, perinatal lethality is also rescued; indeed, survival is comparable to that of littermates with the F10 (+/F) genotype (80% at 18 months). Determination of FX antigen and activity levels in the F10 (F/F) mice reveals normal antigen levels as judged by Western blot (Fig. 3B), and intrinsic and extrinsic activity of 5.5 ± 1.8% and 5.5 ± 1.9%, respectively (Fig. 3C). As expected, F10 (+/F) mice exhibited approximately 50% intrinsic/extrinsic activity (Fig. 3C). Thus, the murine model of FX Friuli is similar to the human mutation in terms of phenotype, activity, and antigen levels. Furthermore, the reproductive fitness of F10 (F/F) mice is indistinguishable from that of the wild-type mice; female F10 (F/F) mice have litter sizes similar to those of wild-type females (data not shown).

Survival of mice with very low FX levels [F10 ()/F)]
The homozygous Friuli mice [F10 (F/F)], although useful as a model for FX deficiency, nevertheless failed to define a lower limit of FX activity required for survival. To generate and characterize mice with even lower levels of FX, we crossed the F10 (F/F) females with F10 (+/)) males to generate F10 ()/F) offspring [F10 (F/F) females rather than F10 (+/)) females were used in this mating scheme to confirm the reproductive competence of these mice]. Such  matings showed the expected genotypic distribution (Table 1, P > 0.2). Intrinsic and extrinsic FX activity in the F10 ()/F) mice was 1.1 ± 0.1% and 2.9 ± 0.2%, respectively ( Table 1). Postnatal survival of the F10 ()/F) is normal, clearly demonstrating that FX activity levels as low as 1-3% are adequate to rescue embryonic and perinatal mortality in F10 ()/)) mice. One hundred percent of F10 ()/F) mice survived for the first postnatal month and 80% have survived for at least 15 months.

Histologic analysis of FX-deficient mice
The F10 (F/F) and F10 (-/F) mice demonstrated relatively normal survival through the postnatal period. However, a previous report documents that both low TF (1%) and low FVII (1%) mice exhibit cardiac fibrosis shown to have resulted from recurrent episodes of hemorrhage into the myocardium from cardiac vessels [12]. We therefore examined cardiac tissue from 10-month-old F10 (F/F) mice, and observed similar findings -iron deposition in the myocardium (Prussian blue staining) and co-localized fibrosis (MassonÕs Trichrome staining; Fig. 4E-L) -in contrast to age-matched littermates ( Fig. 4A-D). These findings are consistent with the observations made by Pawlinski et al. [12] in low TF and low FVII mice. Examination of F10 (F/F) mice at an earlier time point (three months of age) failed to disclose evidence of iron deposition or cardiac fibrosis (data not shown). These results suggest that both age and genotype of the mice are important contributors to development of cardiac fibrosis in these FXdeficient mice.

Discussion
Severe FX deficiency has a low incidence, of the order of one case per 500 000, and complete absence of FX has never been reported. A recently compiled database (http://www. hgmd.cf.ac.uk/ac/index.php) [4] confirms this finding. These data in humans are consistent with the data presented here; mice with homozygous deletion of the F10 gene show a reduction in numbers occurring as early as E11.5 and are thus born in reduced numbers (approximately 50% of expected numbers), exhibiting 100% perinatal mortality.
Because of an interest in developing new approaches for the treatment of bleeding disorders [13][14][15][16], and to analyze the potential role of FX in pathophysiological processes such as sepsis and metastasis [17,18], we wished to develop a viable murine model of severe FX deficiency. The only extant model [15] requires in utero transplantation of wild-type fetal liver, a cumbersome manipulation and not transmissible from one generation to the next. The plug-and-socket methodology

Prussian blue
F10 utilized here facilitates repeated modification of a specific locus as well as expression of the gene of interest under the control of the endogenous promoter, thus preserving the temporal and spatial patterns of expression during embryonic development. Exon 8 was chosen because we wanted to target a region of the F10 gene that encodes variants with a range of activities, as nearly half of the reported patients with FX deficiency have mutations in exon 8. We restored FX expression using a FX variant with extremely low activity and normal antigen levels, FX Friuli [7]. Based on extensive conservation of the catalytic domain between species, we hypothesized that (i) the mutant murine protein would have activity levels similar to that of the human protein, and (ii) these levels would be adequate to rescue both embryonic and postnatal lethality. Both of our hypotheses proved to be correct; the FX Friuli [F10 (F/F)] mice have normal antigen levels, and activity levels quite similar to those of homozygous affected humans (5.5% in mice, 4-9% in humans), and these levels are adequate to rescue both embryonic and postnatal lethality. From F10 ()/F) mice, we also established that FX activity levels of only 1-3%, and/or antigen levels of 50%, are adequate for such rescue. At the molecular level, the murine Friuli mutation in exon 8 most likely causes a similar structural perturbation in the catalytic domain as has been proposed for its human counterpart [19], thus affecting its catalytic potential. Clearly, further studies will help to define the role of this mutation in the mechanistic and molecular interactions of FX within the coagulation cascade.
The Friuli mice, viable but with low circulating levels of FX, can be used to address the role of maternal clotting factor levels on embryonic survival, a point of uncertainty in the literature. Although our data are indirect, a comparison of two intercrosses [F10 ()/F) intercross and F10 (+/)) intercross; Table 2] strongly suggests that embryonic survival of FX-deficient mice is improved by the presence of biologically active FX in the mother. For example, F10 ()/ F) and F10 (+/)) mothers had distinct levels of FX activity (1-3% and 50%, respectively) but similar antigen levels (50%), suggesting that presence of maternal FX antigen is not sufficient to improve embryonic survival. Thus, the partial penetrance of the embryonic lethality in F10 null embryos is, at least partly, a function of maternal transfer of FX. A probable biological significance of maternal transfer of FX (and other coagulation factors in general) may be the prolongation of survival of embryos with FX or other coagulation deficiencies that would otherwise succumb.
Moreover, although relative embryonic survival of F10 knockout mice [F10 ()/))] is directly dependent on the specific genotype of the mother (with respect to FX activity levels), our data suggest that embryonic and postnatal survival of F10 ()/F) mice is independent of the motherÕs genotype. This is clearly shown in both the F10 (F/F) · F10 (+/)) cross and the F10 ()/F) intercross (Tables 1 and 2, respectively), where F10 ()F) mice were born in numbers that establish lack of embryonic lethality.
In summary, we have shown that both embryonic and postnatal lethality of FX deficiency can be rescued by Ôknock-inÕ of a variant FX molecule with reduced activity and normal antigen levels (FX Friuli), and that this mouse model accurately models the mild human bleeding diathesis. Moreover, F10 ()/F) mice had FX antigen levels of 50% and activity levels of 1-3%, and also survived the embryonic and postnatal periods. Our strategy has made it possible to separate the roles of antigen and activity in embryonic development. It represents a genetic approach to assessing the role of maternal transfer of FX biological activity in embryonic survival, and demonstrates that mere presence of FX antigen is not sufficient. Additionally, these mice provide a viable model for those interested in studying novel therapies for FX deficiency, for investigation of the causes of embryonic lethality in FX-deficient mice and for the role of FX in biological processes such as sepsis and metastasis. )/F n (%) )/) à n (%) )/F cross )/F 89 29 (33%) 58 (65%) 2 (2%) <0.001 *All parental genotypes exhibit 50% antigen levels. Values are compared to normal Mendelian distribution (25%, 50%, 25%) using chi-squared test. à P value = 0.0045 of F10 (+/)) intercross compared to F10 ()/F) intercross in Fig. 2A (FisherÕs exact test).