Humanizing the Protease-Activated Receptor (PAR) Expression Profile in Mouse Platelets by Knocking PAR1 into the Par3 Locus Reveals PAR1 Expression Is Not Tolerated in Mouse Platelets

Anti-platelet drugs are the mainstay of pharmacotherapy for heart attack and stroke prevention, yet improvements are continually sought. Thrombin is the most potent activator of platelets and targeting platelet thrombin receptors (protease-activated receptors; PARs) is an emerging anti-thrombotic approach. Humans express two PARs on their platelets–PAR1 and PAR4. The first PAR1 antagonist was recently approved for clinical use and PAR4 antagonists are in early clinical development. However, pre-clinical studies examining platelet PAR function are challenging because the platelets of non-primates do not accurately reflect the PAR expression profile of human platelets. Mice, for example, express Par3 and Par4. To address this limitation, we aimed to develop a genetically modified mouse that would express the same repertoire of platelet PARs as humans. Here, human PAR1 preceded by a lox-stop-lox was knocked into the mouse Par3 locus, and then expressed in a platelet-specific manner (hPAR1-KI mice). Despite correct targeting and the predicted loss of Par3 expression and function in platelets from hPAR1-KI mice, no PAR1 expression or function was detected. Specifically, PAR1 was not detected on the platelet surface nor internally by flow cytometry nor in whole cell lysates by Western blot, while a PAR1-activating peptide failed to induce platelet activation assessed by either aggregation or surface P-selectin expression. Platelets from hPAR1-KI mice did display significantly diminished responsiveness to thrombin stimulation in both assays, consistent with a Par3-/- phenotype. In contrast to the observations in hPAR1-KI mouse platelets, the PAR1 construct used here was successfully expressed in HEK293T cells. Together, these data suggest ectopic PAR1 expression is not tolerated in mouse platelets and indicate a different approach is required to develop a small animal model for the purpose of any future preclinical testing of PAR antagonists as anti-platelet drugs.

Generation of hPAR1-KI (Pf4-Cre; Par3 LSL-PAR1 ) mice All mouse studies were approved by the Alfred Medical Research and Education Precinct Animal Ethics Committee (approval number E/1465/2014/M). Pf4-Cre; Par3 LSL-PAR1 mice (hereafter referred to as hPAR1-KI mice) were generated by Ozgene (Perth, Australia). The targeting vector involved exon 2 of Par3 (F2rl2) flanked with a lox-stop-lox and followed by exon 2 of PAR1 (F2r) (Fig 1A). Since exon 1 encodes the signal peptide and exon 2 encodes the entire mature receptor for both PAR1 and Par3 [19], Cre-mediated deletion is predicted to result in the replacement of Par3 with PAR1 and leaves the endogenous Par3 promoter intact ( Fig 1A). The BAC vector comprised unaltered exon 1 of Par3, then exon 2 of Par3 followed by a stop codon and flanked by loxP sites, and followed by PAR1. The vector was injected into C57BL/6J blastocysts, chimeras were bred for germline transmission, and the targeted allele bred to homozygosity (Par3 LSL-PAR1 ). These mice were then bred with Pf4-Cre mice to induce deletion of Par3 and permit the expression of PAR1 specifically in platelets (Pf4-Cre; Par3 LSL-PAR1 , hereafter referred to as hPAR1-KI mice). PCR genotyping of the targeted allele was performed using the following primers: common reverse (AGCTGAAAAATGGAGCGCTTG) with WT forward (TGGGTTCCTCATCCCGTTTG, predicted size 588bp) or mutant forward (TCCTTCACTTGTCTGGCCATC, predicted size 914bp) ( Fig  1A). Deletion of Par3 was confirmed by PCR using the following primers: forward (CAGTGTGTG GTTTTGTTTCACCT) and reverse (GCCAATCACTGCCGGAAAAG,predicted size 963bp) (Fig 1A).

Isolation of mouse and human platelets
All human studies were approved by the Monash University Human Research Ethics Committee (CF07/0141-2007/0025). Blood was collected after written informed consent was obtained from healthy adults (21-50 years old, of both sexes) who had not taken anti-platelet medications in the past 10 d. Mouse and human blood was drawn into syringes containing acid citrate dextrose (ACD; 1:7 v/v). Platelets were isolated by centrifugation as previously described [20].

Megakaryocyte DNA extraction
Bone marrow cells were aseptically flushed from the femora and tibiae of mice using a syringe and a 23-gauge needle containing Dulbecco's Modified Eagle Medium (DMEM, Life Technologies) supplemented with 10% foetal bovine serum (FBS), penicillin, streptomycin and glutamine. Post-harvest, thrombopoietin (0.1 μg/mL) was added to the cells, which were allowed to culture for 4 d at 37°C, 5% CO 2 . Cells were then collected and centrifuged at 50 g for 5 min and the pellet was re-suspended in 4 mL of PBS. The cells were layered over a two-step BSA density gradient and allowed to sediment for 45 min at room temperature. The megakaryocyte-rich population was collected and centrifuged at 50 g for 5 min, and the pellet was treated with DNA extraction buffer (Bioline) and digested at 75°C for 10 min.

Platelet aggregation
Platelet aggregation was measured by light transmission aggregometry in a 96-well plate format. Human and mouse washed platelets (2x10 8 platelets/mL) were added to wells of a clear, flat  bottom, 96-well tissue culture plate (Falcon). Platelets were stimulated with either thrombin (0.01-1 U/mL), PAR1-AP (10-100 μM), PAR4-AP (10-100 μM), or ADP (1-10 μM). The plate was analysed at 37°C in a FLUOstar OPTIMA plate reader (BMG Labtech) using a 595 nm excitation filter, for a period of 50 min (10 read cycles with 5 min double orbital shake period between each read). Ability to aggregate was calculated as ODðNo AgonistÞ À ODðAgonistÞ ODðNo AgonistÞ À ODðblankÞ Â 100 at the time point where aggregation was at a maximum. Optical density was normalised against the blank (maximum) and unstimulated platelets (minimum) and expressed as % maximum.

P-selectin expression
Flow cytometry was used to detect the expression of P-selectin by activated platelets. Human and mouse platelets, prepared as outlined above, were incubated with an anti-P-selectin antibody and stimulated with thrombin (0.01-1 U/mL), PAR1-AP (10-100 μM), or PAR4-AP (30-300 μM) for 15 min at 37°C. Samples were then analysed using a FACSCalibur (Becton Dickinson) flow cytometer and FlowJo software.

Analysis of platelet PAR1 expression
Platelets isolated from either humans or mice (5x10 7 platelets/mL in Tyrode's buffer) were incubated with a PE-conjugated antibody against one of PAR1 (2.5 μg/10 6 cells; ATAP2), CD45 (platelet negative control antibody of matching isotype, 2.5 μg/10 6 cells), or CD41a (platelet positive control antibody of matching isotype, 2.5 μg/10 6 cells). In some experiments, isolated platelets were permeabilized with saponin (0.1% v/v) prior to incubation with antibodies. In these experiments, a PE-conjugated antibody against the cytoplasmic tail of GPIb was included to confirm effective platelet permeabilization [21]. In all cases, samples were analysed using a FACSCalibur flow cytometer and FlowJo Software. Positive expression was measured as a rightward shift in relative fluorescent intensity compared to the isotype control.

Results
Generation of hPAR1-KI (Pf4-Cre; Par3 LSL-PAR1 ) mice We generated mice in which PAR1 (F2R) was knocked into the Par3 (F2lr2) locus and conditionally expressed in platelets upon Pf4-Cre-mediated excision. To this end, exon 2 of PAR1 (encoding the entire mature receptor) was inserted downstream of exon 2 of Par3 flanked by a lox-stop-lox ( Fig 1A). This approach leaves the promoter and exon 1 of Par3 (encoding the signal peptide) intact and results in replacement of Par3 with PAR1 after Cre-mediated excision ( Fig 1A). This is predicted to yield the Par3 signal peptide linked to the complete mature receptor of PAR1 ( Fig 1A). The stability and surface expression of the predicted fusion protein was confirmed in HEK293T cells transiently transfected with vectors expressing the same sequence as that remaining after Cre-mediated excision of the targeted allele, with the predicted Par3/PAR1 fusion protein expressed at similar levels to native PAR1 (S1 Fig). We therefore went on to generate Par3 LSL-PAR1/+ mice and confirmed targeting by Southern blot (S2 Fig) and PCR genotyping (Fig 1B). Heterozygous mice were bred to homozygosity (Par3 LSL-PAR1/LSL-PAR1 ) and crossed with Pf4-Cre mice [22], with experimental mice resulting from crosses of Pf4-Cre;Par3 LSL-PAR1/LSL-PAR1 x Par3 LSL-PAR1/LSL-PAR1 (Fig 1B). Cre positive offspring (Pf4-Cre;Par3 LSL-PAR1/LSL-PAR1 ) are hereafter referred to as hPAR1-KI mice and their Cre negative littermates (Par3 LSL-PAR1/LSL-PAR1 ) as wild-type. Appropriate Pf4-Cremediated deletion in the platelet/megakaryocyte lineage was confirmed by PCR of DNA isolated from megakaryocytes versus tail biopsy (Fig 1C). We used a forward primer placed upstream of the loxP site and a reverse primer in each of Par3 and PAR1 as indicated ( Fig  1A). As predicted, a band of expected size was only generated in megakaryocyte DNA from hPAR1-KI mice. Sequencing of this PCR product confirmed the predicted targeting and Cremediated deletion (S3 Fig).
Platelets from hPAR1-KI mice do not respond to PAR1-selective activation Mice from these crosses were born at the expected rates, were normal in weight and appearance, and had normal whole blood counts (S1 Table). We assessed PAR1 function in platelets isolated from hPAR1-KI mice using two distinct markers of platelet activation: platelet aggregation and P-selectin expression. PAR1-selective activation by a PAR1-AP failed to aggregate platelets from hPAR1 mice at concentrations up to 100 μM (100 times the concentration required to cause platelet activation in human platelets [23]) (Fig 2A). This lack of response was identical in platelets from wild type mice ( Fig 2B). As expected, PAR1-AP induced a robust aggregation response in human platelets (Fig 2C). A similar pattern of responses was observed when P-selectin expression was used as the marker of platelet activation, with PAR1-AP unable to elicit any response in platelets isolated from either hPAR1-KI (Fig 2D) or wild type (Fig 2E) mice, but a robust response in human isolated platelets (Fig 2F). Together, these data indicate lack of functional PAR1 in platelets from hPAR1-KI mice.
Lack of PAR1 expression on platelets from hPAR1-KI mice We next examined PAR1 expression in platelets from hPAR1-KI mice. We were unable to detect PAR1 on the surface of platelets from hPAR1-KI mice by flow cytometry (Fig 3A). As with the functional experiments, platelets from wild type mice and humans served as effective negative and positive controls for PAR1 surface expression by this method, respectively (Fig 3B  and 3C). In all experiments, the platelet specific marker, CD41, and normal forward/side scatter properties were used to confirm the integrity of the platelets being assessed (Fig 3). We also tested for PAR1 expression in permeabilized platelets. In these experiments, an antibody against the cytoplasmic tail of GPIbα [21] was used as a positive control for successful permeabilization. As predicted, this antibody only recognised an epitope on the internal surface of the platelet membrane, with no staining of intact platelets and strong staining in permeabilized platelets from both mice and humans (S4 Fig). However, we were unable to detect any PAR1 in permeabilized platelets from hPAR1-KI mice (Fig 4A). Again, platelet integrity was confirmed by CD41 expression in all cases, while permeabilization was confirmed by detection with the anti-GPIbα-tail antibody (Fig 4). Platelets from wild type mice and humans again served as controls (Fig 4B and 4C), indicating the anti-PAR1 antibody still binds its epitope in permeabilized cells.
Finally, we also examined PAR1 expression by Western blot. Here, a band at the predicted size of approximately 75 kD was detected in lysates of human platelets, but not of wild-type or hPAR1-KI mouse platelets (Fig 5), mirroring the flow cytometry data. Together, these findings indicate PAR1 is not expressed at detectable levels in platelets from hPAR1-KI mice.

PAR3 expression and function is disrupted in platelets from hPAR1-KI mice
Our genetic strategy is predicted to disrupt Par3 expression in platelets. This was confirmed genetically by PCR genotyping and sequencing (Fig 1). Regardless, given the lack of PAR1 expression or function in platelets from hPAR1-KI mice, we therefore also probed for Par3 protein in these platelets. We never detected Par3 expression by Western blot in human platelets, but routinely observed a band of the previously reported size (approximately 50kDa) in wild type mouse platelets (Fig 5). This band was absent in platelet lysates from all hPAR1-KI mice examined (Fig 5), further confirming disruption of Par3 in hPAR1-KI mice. To further examine this, we tested for the predicted functional effects of Par3-deficiency.
Platelets from Par3-/-mice exhibit diminished responsiveness to thrombin [24]. Here, we observed a similar effect in platelets from hPAR1-KI mice. We examined platelet aggregation and P-selectin expression in response to thrombin. Platelets from hPAR1-KI mice exhibited an approximately 3-fold decrease in sensitivity to thrombin in both assays when compared to platelets from wild type mice (Fig 6). Specifically, aggregation of hPAR1-KI platelets was only achieved at thrombin concentrations of 0.3 U/mL and above ( Fig 6A) compared with 0.1 U/ mL and above for wild type platelets (Fig 6B). A similar pattern emerged when examining Pselection expression in response to thrombin: aggregation of hPAR1-KI platelets was only achieved at thrombin concentrations of 1 U/mL ( Fig 6D) compared with 0.3 U/mL and above for wild type platelets (Fig 6E).
Importantly, hPAR1-KI platelets responded normally to other agonists, PAR4-AP and ADP (Fig 7). When comparing platelets from hPAR1-KI and wild type mice, near-identical concentration-responses were observed for ADP-induced aggregation (Fig 7A and 7B), PAR4-APinduced aggregation (Fig 7D and 7E), and PAR4-AP-induced P-selectin expression (Fig 7G  and 7H). These functional data are consistent with the phenotype observed in platelets from Par3-/-mice and, when combined with the failure to detect Par3 expression in hPAR1-KI platelets, suggest correct targeting and Par3 deletion in hPAR1-KI mice.

Discussion
PARs are leading targets for new anti-platelet drugs, exemplified by the recent approval of the PAR1 antagonist, vorapaxar, and the early clinical development of the PAR4 antagonist, BMS-986120. Given the increased focus of examining PAR antagonists as anti-platelet agents, there Whole cell lysates of platelets isolated from hPAR1-KI mice, wild type mice, or humans were probed for either PAR1 or Par3 protein by Western blot. PAR1 was readily detected in human platelets but not in platelets from wild type or hPAR1-KI mice. In the same samples, Par3 was detected in platelets from wild type mice but not in platelets from either hPAR1-KI mice or humans. Actin was used as a protein loading control for all lysates. Blots shown are representative of n = 3 experiments. is much interest in developing an appropriate small animal model for in vivo testing of such drugs. This has proved to be a major limitation in the field, as the PAR1/PAR4 expression profile of human platelets is only shared amongst primates. Here, we attempted to create a genetically-modified mouse that mimics the PAR1/PAR4 expression profile of human platelets. To do this, we knocked the human PAR1 gene in to the mouse Par3 locus in a manner that provided for platelet-specific replacement of Par3 with PAR1. Correct targeting was confirmed genetically and was apparent functionally. Yet despite this, we were unable to detect expression or function of PAR1 in platelets from these mice, indicating the approach used here is insufficient to 'humanize' platelet PAR expression in mice by inducing PAR1 expression in place of Par3.
Forced PAR1 expression has been previously achieved using transgene-based approaches in other cell types. For example, mouse PAR1 (Par1) was specifically expressed in endothelial cells via the Tie2 promoter/enhancer [25] and in cardiomyocytes via the αMHC promoter [26]. In addition, PAR1 has been expressed in mouse mammary gland epithelium [27]. In contrast to these successes, a previous attempt failed to express either Par1 or PAR1 transgenes in mouse platelets via multiple platelet-specific promoters [18]. The reasons for the inability to express PAR1 in mouse platelets remain unknown. However, given the limitations and lack of control associated with transgene-based approaches, we used a knock-in approach in which PAR1 was targeted to the Par3 locus, essentially resulting in replacement of Par3 with PAR1, and with expression governed by the endogenous mouse Par3 promoter.
Correct genetic targeting was confirmed by multiple methods, including Southern blot analysis of founders, PCR genotyping across the insert, and PCR analysis of correct Pf4-Cre-mediated excision in the DNA in isolated megakaryocytes. Correct targeting was also apparent by the absence of Par3 expression in platelets from hPAR1-KI mice. The disruption of Par3 expression was further supported by functional analyses of these platelets that revealed a phenotype remarkably similar to that observed in platelets from Par3 -/mice [24], with an approximately 3-fold decrease in sensitivity to thrombin-induced platelet activation. Together, this detailed analysis indicates correct targeting occurred in these mice and resulted in disruption of Par3 expression. However, no replacement PAR1 expression or function was detected.
The reason behind this lack of expression of PAR1 in mouse platelets is unknown. Our initial functional studies showed that hPAR1-KI platelets did not respond to a PAR1-specific agonist but responded normally to other agonists, suggesting that the receptor was either absent or unable to signal. We probed extensively for protein expression but were unable to detect PAR1 in platelets from hPAR1-KI mice via flow cytometry or Western blot. The lack of detection of PAR1 even in permeabilizsed platelets (by flow cytometry) or in whole cell platelet lysates (by Western blot) strongly suggest that the PAR1 protein is either not translated or is very rapidly degraded. The inability of mouse platelets to ectopically express PAR1 so far appears unique, but remains unexplained and could arise from a number of critical stages in the development of mature PAR1 protein. While possible that the Par3 signal peptide is insufficient to guide PAR1 to the cell surface [28], the highly conserved nature of signal peptides coupled with our observation of robust surface expression of the fusion protein in HEK293T cells suggest otherwise. Other potential explanations include as yet unidentified mechanisms that appear unique to the mouse platelet, such as poor or defective N-linked glycosylation of the receptor leading to receptor instability and signaling dysfunction [29,30]. Since the primary aim of this study was to generate a research tool for in vivo thrombosis studies, we did not explore the precise mechanism behind the lack of tolerance for PAR1 expression in mouse platelets. Whether or not replacement of the entire human PAR1 gene in the mouse Par3 locus will be sufficient to drive expression in mouse platelets remains unknown. Regardless, it appears that other approaches will need to be examined if a non-primate model is to be used for pharmacological screening of drugs targeting platelet PARs. and 4R in red, the loxP site in yellow, and PAR1 exon 2 in blue. Red text indicates the sequence output of the PCR product from primers 3F and 4R, which was identical to the predicted genomic DNA sequence. (B) Annotated sequence of the predicted coding sequence and protein chimera, comprising the Par3 signal peptide (grey text) and PAR1 mature receptor (blue text). Red text indicates sequenced PCR product amplified from genomic DNA of hPAR1-KI mice, aligning to the predicted region of PAR1 coding sequence.