Heterozygous mutation SLFN14 K208N in mice mediates species-specific differences in platelet and erythroid lineage commitment

Endoribonucleases are a family of proteins responsible for cleaving RNA to regulate and inhibit protein synthesis.¹  The Schlafen (SLFN) family of proteins/genes is made up of 10 mouse and 6 human SLFN genes, all of which possess a characteristic AAA domain coding for DNA helicases, transcription regulators, protein folding regions, and a distinctive SLFN box of unknown function.²  SLFN proteins are divided into 3 subgroups and are highly homologous, classified based on their increasing length. Subgroups II and III contain the aforementioned regions and a SWADL region, whereas subgroup III also possesses an additional helicase region ∼ 400 aa.^2-5 

Reported roles for the SLFN family of proteins include translational control by RNase activity (SLFN14), transfer RNA cleavage as part of the DNA damage response in tumor cells, and T-cell lineage and commitment (SLFN1).⁶  Recently, SLFN14 mutations have been reported in 5 unrelated families worldwide who present with macrothrombocytopenia and associated excessive bleeding.^7-11  Patients with these mutations have a platelet function defect in response to adenosine diphosphate, collagen, and PAR1-peptide and decreased platelet adenosine triphosphate secretion.⁷  Further investigation discovered that SLFN14 colocalizes with ribosomes and causes endoribonucleolytic degradation of ribosomal RNA in cells.¹²  These data, coupled with expression data from several databases, suggest that SLFN14 is responsible for cleavage and regulation of critical RNAs in megakaryocytic and erythroid differentiation. Regulatory RNAs are thought to be critical in hematopoietic lineage commitment, with unique RNA signatures reported in multipotent and bipotent progenitors.^13-15  Pisareva et al revealed that SLFN14 is associated with cleaving RNA and ribosome-bound messenger RNA (mRNA) in an Mg²⁺-dependent and nucleotide triphosphate (NTP)–independent manner.¹⁶  Recent evidence in primary human cells from platelet and erythroid lineages suggests that SLFN14 may function in a similar way in cleaving RNA from the ribosomal unit prior to splitting, influencing hemoglobin production during blood cell development.¹⁷ 

Despite these insights, the mechanistic role of SLFN14 in megakaryocyte (MK) development and hemostasis is unknown. To address this, we generated a global CRISPR-mediated knock-in (KI) mouse model of the patient mutation K219N missense substitution (mouse K208N homolog), which is known to cause thrombocytopenia, and investigated its overall role in hematopoiesis and platelet function. Homozygous KI mice for this mutation (SLFN14^K208N/K208N) did not survive to weaning and, similarly, SLFN14 mutations identified in humans are all of a heterozygous nature; therefore, heterozygous mice were used in all experiments compared with litter-matched wild-type controls (SLFN14^K208N/+ and SLFN14^+/+, respectively). We investigated the SLFN14 K208N mutation through hematological analysis, platelet activation, function, and its overall role in hematopoiesis. SLFN14^K208N/+ mice showed significant differences from SLFN14^+/+ mice in gross hematological analyses. However, unlike the human variants, these mice demonstrate a major defect in erythropoiesis but not megakaryopoiesis or thrombopoiesis.

Differences in the consequences of this missense mutation suggest that SLFN14 is a species-specific regulator of platelet and erythroid lineage commitment. In humans, the mutation causes a defect in thrombopoiesis, whereas a homologous missense mutation in mice causes a significant defect in erythropoiesis. SLFN14^K208N/+ mice present with pronounced microerythrocytosis, hemolytic anemia, splenomegaly, and abnormal thrombus formation in vivo.

A SLFN14 K208N–KI mouse was generated in-house using CRISPR-Cas9 gene editing. All mice were generated on a C57BL/6J background and were bred in heterozygote/wild-type pairs. Animal care and welfare were in accordance with United Kingdom Home Office regulations and the use of Animals (Scientific Procedures) Act 1986. Animals were housed at the Biomedical Services Unit at the University of Birmingham.

All mice were genotyped in-house using DNA extraction from mouse ear clips (DNeasy Blood & Tissue Kits; Qiagen, Manchester, United Kingdom). Polymerase chain reaction (PCR) and Sanger sequencing were used to identify SLFN14 K208N–KI mice following the procedure and conditions outlined in supplemental Tables 1 and 2.

Flow cytometry was performed on a BD Accuri C6 flow cytometer, and results were analyzed using BD Accuri C6 software. Flow cytometry antibodies are listed in supplemental Tables 4.

Mice were exsanguinated under terminal anesthesia by isoflurane/O₂ (5%) gas. Blood was drawn from the inferior vena cava, using a 25-gauge needle, into 1:10 (volume-to-volume ratio) acid citrate dextrose anticoagulant. Washed platelets were prepared as described in supplemental Material.

Washed platelet counts were normalized to 2 × 10⁸/mL with Tyrode’s-HEPES buffer. Aggregation in 300 µL of platelets was measured using a lumi-aggregometer (Chrono-Log, Havertown, PA) at 37°C under stirring conditions (1200 rpm) for 6 minutes post–agonist addition.

Washed platelets at 2 × 10⁷/mL were incubated for 45 minutes on collagen-coated (10 µg/mL) and fibrinogen-coated (100 µg/mL) coverslips under resting or preactivated conditions (0.1 U/mL thrombin). Adhered cells were fixed with 10% formalin, permeabilized, and stained with Alexa Fluor 488–conjugated phalloidin. Images were captured on a Zeiss Epifluorescence microscope and analyzed using a semiautomated machine learning–based workflow.^18,19 

All experiments were double blinded and conducted on 20 to 29g SLFN14^K208N/+ and litter-matched wild-type mice. Mice were anesthetized using isoflurane/O₂ (5%) gas, and 2 to 3 mm of tail tip was excised using a sterile razor blade and placed in prewarmed saline (37°C). Time until first cessation of bleeding was recorded.

Platelet-rich plasma (PRP) from SLFN14^K208N/+ and SLFN14^+/+ mice was adjusted to a final concentration of 2 × 10⁸/mL using platelet-poor plasma and Tyrode’s-HEPES buffer supplemented with 2mM CaCl₂, as previously described.²⁰  Erythrocytes were added for visualization, and clot formation was stimulated by 1 U/mL thrombin. Clots were monitored every 30 minutes for 2 hours, and clot weight and volume were calculated.

Laser-induced injury of cremaster arterioles and FeCl₃-induced injury of carotid arteries were performed and analyzed as previously described.²¹ 

Spleens and decalcified bones were embedded in paraffin wax and sectioned at 5 µm prior to staining with hematoxylin and eosin (H&E) and Perls Prussian blue. Sections were scanned using a Zeiss Axio ScanZ1 slide scanner (Carl Zeiss Ltd., Cambridge, UK). The number of MKs was counted in 10 fields of view taken from 2 femur sections and 1 spleen section per mouse. Sectioning and image analysis were performed double blinded.

Spleens were homogenized and whole bone marrow was flushed from mouse femurs and tibias in 1% fetal bovine serum and 2 mM EDTA in phosphate-buffered saline. Cells were filtered through a 70-µm cell strainer and stained with antibodies as per supplemental Table 4. A total of 50 000 events was collected using a BD Accuri C6 flow cytometer and gated to eliminate dead and cell doublets. Cells were imaged on poly-l-lysine coated coverslips using a Zeiss LSM880 confocal microscope.

Data are presented as mean ± standard error of the mean (SEM), unless stated otherwise. A Student t-test and 2-way analysis of variance were used for platelet activation, and a Mann-Whitney U test was used for in vivo analysis, with P < .05 deemed significant. All analyses were conducted using GraphPad Prism software (v8.4).

Animal care and welfare were in accordance with United Kingdom Home Office regulations and the use of Animals in Scientific Procedures Act 1986 under Project License number P53D52513 (to N.V.M.).

An in-house CRISPR-KI model was developed using homology directed repair (HDR). Human oligonucleotide donor templates of the K219N mutation were coinjected with single guide RNA as per the CRISPR-Cas9 mechanism (Figure 1A). This resulted in the G>T substitution and subsequent K208N amino acid change. Sanger sequencing was used to genotype mice, with Figure 1B showing successful KI of the mutation. A χ² analysis of heterozygote/heterozygote breeding pairs shows significant deviation from Mendelian inheritance, with ∼25% of offspring lost preweaning (P < .0001; Figure 1C). The average prewean loss was 25% across 15 litters, the same proportion of expected homozygote offspring according to Mendel’s law (supplemental Table 3). To assess embryonic lethality, embryos were gathered for observation and genotyping at 12.5 and 14.5 days post–vaginal plug. At embryonic day 12.5 (E12.5), SLFN14^K208N/+ mice were not different from wild-type littermates. However, SLFN14^K208N/K208N embryos were significantly paler, with less-defined vasculature (Figure 1D). E14.5 SLFN14^K208N/K208N embryos were much more pale, with substantially less vascular definition than that seen in the other genotypes (Figure 1D). No SLFN14^K208N/K208N mice survived to weaning for genotyping; therefore, we can deduce that if SLFN14^K208N/K208N pups do survive beyond the critical fetal liver stage at day 14.5, they die shortly after birth. Therefore, in parallel with K219N heterozygous patients, SLFN14 K208N heterozygotes were used in analyses.

SLFN14^K219N/+ patients exhibit macrothrombocytopenia; however, we did not observe any difference in platelet count in homologous SLFN14^K208N/+ mice, but we did detect an increase in size using a flow cytometry counting assay and single positive events of CD41 and Ter119–stained cells (Figure 2A,Bi-ii; P < .0001). Contrary to the patient mutation, for which no effect on erythrocyte production or function was reported, SLFN14^K208N/+ mice exhibit microcytic erythrocytosis, an increase in erythrocyte count accompanied by a reduction in size (Figure 2Biii-iv). We did not observe any difference in leukocyte counts between genotypes (data not shown).

SLFN14^K219N/+ patients have a high immature platelet fraction (IPF).⁷  We assessed IPF in SLFN14^K208N/+ mice using a flow cytometry method and nucleic acid stain SYTO13, which was recently reported to be a more specific marker than its predecessor, Thiazole orange.²²  A 15% increase in the proportion of CD41⁺SYTO13⁺ cells was observed in SLFN14^K208N/+ mice compared with SLFN14^+/+ controls (Figure 2C; P = .0003). Combined with the observed increase in platelet size, these data indicate an increase in the proportion of immature platelets in SLFN14^K208N/+ mice.

Whole blood smears from SLFN14^K208N/+ mice show irregularly shaped smaller erythrocytes (poikilocytes and microcytes) compared with the characteristic plump-shaped cells observed in wild-type controls (Figure 2D). This microcytosis is accompanied by lower hemoglobin levels in SLFN14^K208N/+ mice (Figure 2E).

These results and previous work in T-cell lineage commitment studies identify SLFNs as key drivers in species-dependent hematopoietic lineage commitment. In this case, we observe SLFN14 mutations causing distinct differences in platelet and erythrocyte production and morphology.^3,6 

No alteration in major glycoprotein receptor levels was observed in SLFN14^K208N/+ platelets (Figure 3A). Platelet activation was assessed by flow cytometry, and both genotypes displayed similar α-granule secretion (P-selectin) and integrin αIIbβ3 activation (JON/A) in response to agonists at varying doses (Figure 3B). Platelet function in SLFN14^K208N/+ mice was assessed using light transmission aggregometry. SLFN14^K208N/+ mice display normal platelet function compared with their wild-type littermates in response to the agonists thrombin, collagen, and collagen-related peptide (Figure 3C-E), which are mediated by G protein–coupled receptors and ITAM/receptor tyrosine kinase, and are the 2 main types of activation receptors in platelets.

We next investigated platelet adhesion and spreading on collagen- or fibrinogen-coated surfaces. Under resting and preactivated conditions, no difference in adhesion or cytoskeletal remodeling was observed in SLFN14^K208N/+ platelets (Figure 3F).^18,19 

SLFN14 patients were recruited to studies based on their bleeding phenotypes. To establish whether SLFN14 transgenic mice had a bleeding phenotype, 2 to 3 mm of tail tip was excised, and time to bleeding cessation in prewarmed (37°C) saline was measured. No difference was observed in either genotype, indicating that platelets retain normal function, and abnormal erythrocytes do not impact hemostasis in this model (Figure 4A).

Clot retraction was also investigated in PRP to assess αIIbβ3-mediated platelet function. No visual difference was observed during the time course, and final mean clot weight of 36.0 mg and 30.0 mg for SLFN14^+/+ and SLFN14^K208N/+ mice, respectively, was not significantly different (Figure 4B). This supports our findings that platelet function is maintained and that the SLFN14 K208N mutation in mice does not lead to platelet function defects.

Thrombus formation was assessed in vivo by laser- and FeCl₃-induced injury models. Following laser injury, SLFN14^K208N/+ mice form thrombi at a similar rate to SLFN14^+/+ controls, although they are smaller, with fewer platelets recruited, and embolize more quickly than do those in littermate controls (Figure 4C; supplemental Videos 1 and 2, respectively). FeCl₃ injury to the carotid artery resulted in occlusive thrombus formation in both genotypes (Figure 4D; supplemental Videos 3 and 4). Monitoring the accumulation of fluorescently labeled platelets shows that SLFN14^K208N/+ mice form thrombi at a slightly slower rate and have a tendency for reduced stability, although these were not statistically different from littermate controls (Figure 4Dii-iii). Ultimately, thrombus formation and stability defects in SLFN14^K208N/+ mice are driven by the involvement of other blood cells; together with in vitro platelet function studies, these findings suggest that SLFN14 K208N has only a minor role in mouse platelets.

The spleen acts as a major site for filtering and clearance of blood from the circulation. In classical findings of clearance, platelets undergo phagocytosis controlled by their immunoglobulin G–coated surfaces, rendering Fc receptor–directed clearance by macrophages in the spleen.²³  The spleen can also act as an additional site of hematopoiesis in the event of myelofibrosis or bone marrow scarring in certain pathologies.²⁴  We investigated the spleen to assess differences in blood cell production and clearance. After controlling for body weight, spleens of SLFN14^K208N/+ mice were significantly larger with regard to weight and size compared with those from SLFN14^+/+ littermates (Figure 5A). MK counts from histology sections in the spleen were normal, and the proportion of MKs was unchanged between genotypes by flow cytometry analysis (Figure 5B-C).

Using a similar gating strategy to Koulnis et al, we used CD71 (early erythroid progenitor marker) and Ter119 (mature erythroid marker) to detect erythroid cells and categorized Ter119^hi cells according to size (Figure 5Di).²⁵  There was a significant increase in the proportion of proerythroblasts (ProEs) in heterozygotes (CD71⁺/Ter119⁺; Figure 5Dii). In these spleens, we did not detect any EryA progenitors in either genotype (Figure 5D). However, despite fairly consistent proportions of Ter119^hi cells between genotypes, we observed a difference in the distribution and greater spread of CD71⁺ cells within the EryB gate and an increase in EryCs (most mature erythroid cells) in heterozygotes that was suggestive of erythroid maturation from the intermediate progenitor within the spleen (Figure 5D). Gating strategy for spleen flow cytometry is detailed in supplemental Figure 1.

Our flow cytometry progenitor panel aimed to detect megakaryocyte-erythroid progenitors (MEPs). Consistent with previous findings by Psaila et al, we identified 2 subpopulations by differential expression of CD71 and a small MEP population by coexpression of CD41.²⁶  CD71⁺/CD41⁺ cells were rare in these samples, consistent with previous findings, although we did detect an almost threefold increase in MEPs in SLFN14^K208N/+ mice than in wild-types²⁶  (Figure 5E). In addition, we observed a greater spread of cells within our CD71⁺ population in contrast to the more clustered appearance in controls.

We discovered substantial hemosiderin staining in lysed erythrocytes. These hemoglobin deposits occur naturally as a result of macrophage-mediated clearance of erythrocytes. Using Perls Prussian blue staining, we see a substantial increase in sites of free heme staining, as indicated by the blue areas in Figure 5F. Interestingly, we see staining in the red and white pulp of SLFN14^K208N/+ mice but only red pulp where macrophage clearance occurs in controls. This is also supported by the color difference in heterozygous spleens, which appear significantly darker than SLFN14^+/+ spleens (Figure 5Ai). Here, we hypothesize that erythropoiesis is accelerated to compensate for reduced hemoglobin levels. We suggest that SLFN14^K208N/+ erythrocytes are more prone to hemolysis and that the spleen acts as a secondary site of hematopoiesis, specifically upregulating erythropoiesis. The spleen also acts in cell clearance; as such, it may be unable to recognize the need for these additional erythrocytes in oxygen transport which, in the case of SLFN14^K208N/+ mice, may also be the cause of accelerated hemolysis.

The main site of hematopoiesis is the bone marrow; therefore, to assess discrepancies in hematopoiesis or altered progenitor levels leading to differences in platelet and erythrocyte counts, we examined bone marrow sections and flow cytometry panels, staining for progenitors. Double-blinded evaluation of bone marrow histology sections revealed that MKs were highly populated within the bone marrow, with morphology, including the characteristic polyploid nuclei, consistent across the 2 genotypes (Figure 6A). We analyzed progenitor levels in flushed whole bone marrow from mouse femurs and tibias by flow cytometry. Live single cells were selected based on size and forward scatter to detect progenitor cells and subsequent double-positive populations (determined by antibody staining) gated for analysis (supplemental Figure 2). MKs were highlighted by CD41⁺/CD42⁺ events (MK quadrant); no difference in the percentage of MKs was observed between genotypes (Figure 6B).

Using CD71 and Ter119 as before, there was no significant difference in the proportion of ProEs in the bone marrow (Figure 6Ci-ii). The percentage of Ter119^hi cells was consistent between SLFN14^K208N/+ and wild-type mice; within this population, EryA, EryB, and EryC populations were also unchanged between genotypes (Figure 6Ci-ii). As in spleens, we observed the similar “spread distribution” of variable CD71 expression within the EryB gate; however, in contrast, mature erythrocytes (EryC) were rare. Confocal imaging of this staining pattern is shown in Figure 6Cii.

The predecessor to MKs and ProEs is the MEP. Given the differences that we observed in erythroid cell distribution (Figure 6C) and evidence suggestive of a platelet defect in our patient data, we detected megakaryocyte-erythroblast (MK-EB)–primed MEPs using CD71 and CD41, as before. These double-positive primed MEPs are often difficult to detect in situ within the bone marrow (Figure 6Di).²⁶  We did not observe any difference in the proportion of CD71⁺/CD41⁺ cells in SLFN14^K208N/+ mice (Figure 6Dii). Confocal imaging supports an increase in cellular events in the double-positive channel of heterozygotes and more CD71⁺ single-stained events (Figure 6Diii). The reason for this shift in the MEP population is yet to be determined; however, we hypothesize that it results from SLFN14^K208N/+ MK-EB MEPs being more primed in the erythroid direction, consistent with higher single CD71 positivity in this staining panel (CD71⁺; Figure 6Di). Increased CD71⁺ staining may be expansion of the MEP with preference towards the erythroid lineage, as well as due to increased erythroid progenitors after this stage.

To examine whether the K208N mutation led to aberrant expression in hematopoietic cells of SLFN14^K208N/+ mice, we measured the abundance of SLFN14 mRNA in whole bone marrow by real-time quantitative PCR. Compared with RNA from wild-type controls, SLFN14 mRNA levels were reduced significantly (by ∼50%) in SLFN14^K208N/+ mice (P < .01; supplemental Figure 3).

Furthermore, we considered whether levels of the master transcription factor GATA1 were altered as a result of the K208N mutation. Real-time quantitative PCR was performed to measure GATA1 mRNA levels using cDNA-specific primers for GATA1 and GAPDH as the endogenous control housekeeping gene. GATA1 levels were reduced substantially in RNA from SLFN14^K208N/+ mice compared with litter-matched controls (P < .05; supplemental Figure 3). All primers are given in supplemental Table 5.

SLFN14 is a poorly studied endoribonuclease with suspected roles in cleaving RNA that may contribute to a reported thrombocytopenia and clinical bleeding in multiple unrelated patients/families. Here, we present a CRISPR KI mutation of K208N in mice and establish the role of SLFN14 as a key player in the lineage-commitment pathway, giving rise to species-specific phenotypes in hematopoiesis. In this study, we analyzed heterozygous mutants, because homozygous mutants did not survive to weaning and showed significant deviation from Mendelian inheritance patterns. This suggests that SLFN14 has a critical role in mouse embryogenesis and, particularly, erythropoiesis. SLFN14 K208N is likely to be particularly relevant in future studies of gene-expression profiling in erythrocytes and anemia. E12.5 and E14.5 homozygous embryos were paler than their littermates, showed less distinct vasculature, and did not survive to weaning. These are most likely due to homozygotes’ more severe anemia and hemolysis that results in death shortly after birth.²⁷ 

The K208N mutation in mice presents with a different phenotype than in its homologous human version, suggesting that it plays a critical role at the MEP junction in lineage fate decisions. Previous studies found that endoribonuclease function is critical in RNA regulation in various bacterial species; however, to our knowledge, this is the first discovery of endoribonuclease-mediated species differences in mammals.^28,29 

SLFN14^K219N patients have a high IPF, suggesting that thrombopoiesis is not sufficient to maintain steady levels of platelet production or that platelet clearance is accelerated. Platelets contain residual RNA from their predecessors, MKs, which form beaded extensions into the lumen of bone marrow sinusoids (proplatelets) and subsequent shear forces of the bloodstream that cause release of preplatelets and platelets into the blood.³⁰  RNA content can be used to estimate the rate of thrombopoiesis and platelet turnover by measuring the proportion of reticulated platelets within the circulation. An elevated platelet RNA content signifies newer platelets in the circulation that was previously shown to increase platelet reactivity in cardiovascular events and mortality.³¹  In the case of SLFN14^K208N/+ mice, we infer that the slight increase in platelet size is due to their immaturity (determined by SYTO13 staining), but this is not accompanied by increased platelet reactivity, as shown in our in vitro and in vivo experiments.

Although no platelet defects were observed in in vitro functional studies, SLFN14^K208N/+ mice exhibited reduced thrombus formation in vivo. These defects in the formation and stability of in vivo thrombi are likely attributable to the abnormal erythrocytes in these mice. Erythrocytes are the primary determinant of blood rheology and promote platelet margination, increasing their concentration near endothelium to enable rapid formation of thrombi in response to vessel damage.^32-34  Indeed, previous studies have shown reduced thrombus formation and extended bleeding times in anemic mice.³⁵  Although we did not observe altered hemostasis in SLFN14^K208N/+ mice, changes in the size and number of erythrocytes may explain thrombosis findings. Erythrocyte contribution to platelet activation and thrombin generation should also be taken into consideration, together with their unusual role supporting platelet adhesion in the FeCl₃ model.^35-37  Although SLFN14 patients display excessive bleeding phenotypes, we do not report these similarities in mice.⁷  Bleeding in SLFN14 patients has been characterized and explained by defects in platelet aggregation, but little to no effect on platelet function was found in SLFN14^K208N/+ mice. We believe that this work precedes what is to become extensive research into platelet-erythrocyte interactions in health and disease and reveals potential novel mechanisms in hemolysis and anemia.

In SLFN14^K208N/+ mice, we understand that the spleen acts as a secondary site for erythropoiesis. Here, intermediate progenitors (EryBs) differentiate into mature erythrocytes (EryCs), which are highly populated within the spleen. Loss of CD71 expression is indicative of erythrocyte maturation, and the “spread” appearance of EryB cells in heterozygotes clearly shows this maturation phase. This enhanced erythropoiesis is likely due to severe anemia and hemolysis in these mice attempting to compensate for lower hemoglobin levels. The following questions then arise: at what stage in hematopoiesis do SLFN14 and its mutations cause a shift in lineage commitment to platelet or erythroid directions and how, as an endoribonuclease, does SLFN14 mediate this transition? Our bone marrow flow cytometry data did not show any difference in bone marrow MK, ProE, or MEP numbers, but there are discrepancies within progenitors of the erythroid lineage suggesting that, in hematopoiesis, SLFN14^K208N/+ leads to altered erythropoiesis and defects in erythroid cells. In the bone marrow, SLFN14^K208N/+ Ter119^hi cells expressing CD71 also showed variable expression within the EryB gate, whereas wild-type cells present a more clustered distribution. Platelets and erythrocytes originate from a common progenitor; therefore, pinpointing the exact location of this shift is notoriously difficult. However, we believe that using RNA-sequencing of SLFN14 progenitors (MKs, ProEs, and MK/EB-MEPs) will reveal discrepancies in RNA expression profiles to support our preliminary findings that a reduction in GATA1 mRNA is specifically involved in erythroid development. Future work will establish the mechanistic effects of these mutations on human and murine RNA signatures that are critical in lineage commitment. This will uncover novel insights into SLFN14’s ability to cleave RNAs which perturb RNA metabolism and protein synthesis in MK and erythroid lineages in a species-dependent manner.

The authors thank all technicians and staff at Biomedical Services Unit at the University of Birmingham for housing and husbandry of animals used in this study. They also acknowledge Pip Nicolson for expertise in whole blood smear analysis.

Work in the authors’ laboratories is supported by grants from the British Heart Foundation (PG/16/103/32650, FS/18/11/33443) (N.V.M.) and National Institutes of Health, National Institute of General Medical Sciences grant GM097014 (A.V.P.) and National Heart, Lung, and Blood Institute grant HL146544 (A.V.P. and N.V.M.).

Contribution: R.J.S., C.W.S., A.O.K., and N.V.M. designed the study, designed and performed experiments, and wrote the manuscript; A.B. generated CRISPR mouse colonies; E.J.H. and S.L. contributed to mouse colony maintenance and experiments; V.P.P., A.V.P., and S.P.W contributed intellectually to the study; and all authors reviewed the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Neil V. Morgan, Institute of Cardiovascular Sciences, College of Medical and Dental Sciences, University of Birmingham, Edgbaston B15 2TT, United Kingdom; e-mail: n.v.morgan@bham.ac.uk.