Cure of Congenital Purpura Fulminans via Expression of Engineered Protein C Through Neonatal Genome Editing in Mice

Abstract

BACKGROUND:

PC (protein C) is a plasma anticoagulant encoded by PROC; mutation in both PROC alleles results in neonatal purpura fulminans—a fatal systemic thrombotic disorder. In the present study, we aimed to develop a genome editing treatment to cure congenital PC deficiency.

METHODS:

We generated an engineered APC (activated PC) to insert a furin-cleaving peptide sequence between light and heavy chains. The engineered PC was expressed in the liver of mice using an adeno-associated virus vector or CRISPR/Cas9 (clustered regularly interspaced short palindromic repeats/clustered regularly interspaced short palindromic repeat–associated 9)-mediated genome editing using an adeno-associated virus vector in vivo.

RESULTS:

The engineered PC could be released in its activated form and significantly prolonged the plasma coagulation time independent of the cofactor activity of PS (protein S) in vitro. The adeno-associated virus vector-mediated expression of the engineered PC, but not wild-type PC, prolonged coagulation time owing to the inhibition of activated coagulation FV (factor V) in a dose-dependent manner and abolished pathological thrombus formation in vivo in C57BL/6J mice. The insertion of EGFP (enhanced green fluorescent protein) sequence conjugated with self-cleaving peptide sequence at Alb locus via neonatal in vivo genome editing using adeno-associated virus vector resulted in the expression of EGFP in 7% of liver cells, mainly via homology-directed repair, in mice. Finally, we succeeded in improving the survival of PC-deficient mice by expressing the engineered PC via neonatal genome editing in vivo.

CONCLUSIONS:

These results suggest that the expression of engineered PC via neonatal genome editing is a potential cure for severe congenital PC deficiency.

Highlights.

• Engineered PC (protein C) significantly prolonged the plasma coagulation time through the inactivation of coagulation FV (factor V).

• Adeno-associated virus vector-mediated expression of engineered PC inhibited pathological thrombosis.

• CRISPR/Cas9 (clustered regularly interspaced short palindromic repeats/clustered regularly interspaced short palindromic repeat–associated 9)-mediated neonatal genome editing using an adeno-associated virus vector enabled the insertion of the target protein in 7% of liver cells.

• Neonatal genome editing improved the survival of PC-deficient mice through the insertion of engineered PC.

Thromboembolic disorders account for 1 in 4 deaths and are the leading cause of mortality worldwide.¹ Thrombogenesis or thrombus formation in the blood is determined by the balance between the activities of coagulation and anticoagulation factors.² The decreased activity of anticoagulation factors, including AT (antithrombin), PC (protein C), and PS (protein S), in the blood is an important predisposing factor for thrombogenicity.³ AT directly inhibits activated coagulation factors including thrombin and coagulation FX (factor X),⁴ whereas PC circulates as a zymogen in the blood and is activated at the site of thrombus formation.^5,6 Thrombin generated via the coagulation factor cascade binds to thrombomodulin on the vascular endothelial cells, thereby leading to the activation of PC via proteolytic cleavage.^5,6 The APC (activated PC) along with the cofactor PS degrades activated coagulation FV (factor V) and FVIII (factor VIII), thereby inhibiting thrombus propagation.^5,6

Genetic abnormality is a major cause of decreased levels of anticoagulation factors in the blood. Heterozygous abnormality of PC (PROC), PS (PROS1), and AT (SERPINC1) genes is a well-known risk factor for venous thromboembolism in adulthood.⁷ Abnormalities in both alleles of PROC or PROS1 cause neonatal purpura fulminans, which is characterized by life-threatening systemic thrombosis and hemorrhagic skin necrosis shortly after birth.^8–10 Additionally, PC-deficient mice died shortly after birth.¹¹ Although early administration of PC preparations could potentially save lives,^12–14 frequent intravenous injections are required because the half-life of PC (PC zymogen, 6–8 hours; APC, 20 minutes)¹⁵ may remarkably shorten during the acute phase.¹⁵ Anticoagulation therapy with warfarin or heparin and PC preparations can be administered for long-term management.^16,17 However, these patients require lifelong treatment and remain at risk of bleeding caused by extensive anticoagulation and thrombosis owing to the underlying disease throughout their lives.¹⁸ Developing innovative therapies for PC or PS deficiency is required, particularly for patients who are homozygotes and combined heterozygotes with severe disease.

Recently, gene therapy and genome editing therapies are attracting attention as novel modalities for curing intractable diseases. Because PC is a protein produced in the liver hepatocytes in a manner similar to the production of coagulation factors, gene therapy and genome editing that are currently being developed for hemophilia B (coagulation FIX [factor IX] deficiency) may apply to PC deficiency. The plasma level of hPC (human PC) was reportedly 63 nmol/L,¹⁹ which is comparable to that of FIX (87 nmol/L).²⁰ Gene therapy for hemophilia B using a hyperactive mutant called the FIX Padua (R338L mutation) has shown effectiveness.^21,22 PC is a serine protease that is primarily present as a zymogen in the blood. The concentration of APC in the blood is ≈40 pmol/L, representing only 1/1700 of the total PC.²³ We hypothesized that gene therapy or genome editing therapy can become a realistic alternative treatment approach if the active form of PC is secreted. We aimed to develop an engineered APC and assessed its application as gene therapy and genome editing therapy targeting PC deficiency in mice.

Materials and Methods

Availability of Data

The original data and materials within the article are available from the corresponding author upon request. PC-deficient mice have been deposited at Riken BioResource Research Center (No. RBRC11381; Ibaraki, Japan). Material transfer agreements are required for the donation of plasmids and mice.

Plasmid Constructs and Adeno-Associated Virus Vector Production

Detailed information of plasmid construction is provided in Supplemental Methods. A plasmid comprising a chimeric promoter (HCRhAAT, which is an enhancer element of the hepatic control region of the ApoE/C1 gene and the human antitrypsin promoter), cDNAs, and the SV40 polyadenylation signal was created to specifically express target gene in hepatocytes by adeno-associated virus (AAV) vector. The DNA fragment was introduced between inverted terminal repeats into the pAAV plasmid to produce the AAV vector. sgRNA (single guide RNA) sequences were designed using an online software provided by Benchling (https://benchling.com; Table S1). For the production of AAV vector that would induce double-strand break, a DNA fragment comprising a chimeric HCRhAAT promoter, SaCas9 (Staphylococcus aureus Cas9) cDNA, the SV40 polyadenylation signal, and a sgRNA sequence driven by the U6 promoter was introduced between inverted terminal repeats into the pAAV plasmid. We simultaneously created a pAAV donor plasmid containing a P2A (2A peptide derived from porcine teschorivus-1) self-cleaving peptide sequence (QGNSEDY, H10Q, S11G, S12N, E23S, N32E, N33D, and H44Y)-conjugated cDNA sequence possessing 1.0 kb homologous arms at the target site for inserting the target gene via homology-directed repair (HDR).

The AAV genes were packaged by triple plasmid transfection of AAVpro293T cells (Takara Bio, Shiga, Japan) to generate the AAV vector (helper-free system), as described previously.²⁴ pHelper plasmid (Takara Bio), the plasmid expressing Rep and serotype 8 capsid (AAV8), and the gene transfer plasmid were simultaneously transfected. AAV vectors were purified from the transfected cells after 72 hours using the ultracentrifugation method, as previously described.²⁵ The titration of recombinant AAV vectors was performed using quantitative polymerase chain reaction, as previously described.²⁶

Plasmid Transfection of the Cells With Wild or Engineered PC

Detailed information is provided in Supplemental Methods.

PC, FV, and FVIII Activities, Coagulation Time, and ELISA

hPC activity (hPC:C) was measured using Berichrom PROTEIN C (Sysmex, Kobe, Japan) by an automated coagulation analyzer (CS-1600; Sysmex). When measuring the existence of APC in the sample, a PC activator (snake venom to activate hPC) was not added to samples. FV activity (FV:C) and FVIII activity (FVIII:C) were measured using a 1-stage clotting-time assay on an automated coagulation analyzer (CS-1600). Prothrombin time (PT) and activated partial thromboplastin time (APTT) were measured by an automated coagulation analyzer (CA-500; Sysmex) using Thrombocheck PT and Thrombocheck APTT (Sysmex). PS-deficient plasma (Affinity Biologicals, Ancaster, Ontario, Canada) was used to determine the role of PS in prolonging the coagulation time via APC. The detailed method for ELISA is described in Supplemental Methods.

Animal Experimentation

All animal experimental procedures were approved by the Institutional Animal Care and Concern Committee of Jichi Medical University (permission number: 20080-04), and animal care was conducted in accordance with the committee’s guidelines and Animal Research: Reporting of In Vivo Experiments guidelines. FVIII-deficient mice (B6;129S4-F8^tm1Kaz/J) were kindly provided by Dr H.H. Kazazian, Jr (University of Pennsylvania, Philadelphia, PA). C57BL/6J male mice were purchased from Japan SLC (Shizuoka, Japan). Sex differences are known in the efficacy of gene transduction with AAV vector to the liver in mice.²⁷ Therefore, all experiments of AAV vector injection into adult mice were conducted with male mice. Meanwhile, because the genotype and sex of the neonatal mice were unknown at birth, all mice were injected with AAV vectors. Animals were maintained in isolators in the specific pathogen-free facility of Jichi Medical University at 23±3 °C with a 12/12-hour light/dark cycle. Detailed information of animal experimentation is provided in Supplemental Methods.

Intravital Microscopy

Intravital microscopy was performed as reported previously.^26,28 Reactive oxygen species–induced thrombus formations at the testicular vein were visualized using laser excitation. Sequential images of the testicular vein (40–60 µm) were obtained using a confocal microscope (Leica TCS SP8; Leica Microsystems, Wetzlar, Germany). Detailed information is provided in Supplemental Methods.

T7 Endonuclease Assay and Quantification of AAV in the Genome and mRNA Expression

Detailed information is provided in Supplemental Methods.

Immunoblotting and Blue Native Polyacrylamide Gel Electrophoresis of Plasma Protein

Detailed information is provided in Supplemental Methods.²⁹

Histological Analysis and Immunohistochemistry

Detailed information is provided in Supplemental Methods.

Tail Clip Assay

The tail clip assay was performed to quantify the bleeding tendency. The tail of the mice anesthetized with isoflurane was clipped 5 mm with a scalpel. The injured tail was placed on filter paper every 10 s. Bleeding time was assessed as the time to stop bleeding. To prevent animal welfare issues, bleeding time was measured up to a maximum of 9 minutes only. Bleeding volume was quantified as the percentage of the red bleeding area in the filter paper.

Statistical Analyses

Specific details on how many independent biological samples or mice were included in an experiment are given in the corresponding figure legends. Sample size is given in the figure legends, and data are presented as mean±SEM. Two-group comparisons for means were performed by 2-tailed unpaired Student t test, and multigroup comparisons for means were performed by 1-way ANOVA with Dunnett or Tukey multiple comparison test or 2-way ANOVA with Bonferroni multiple comparison test. P<0.05 was considered statistically significant. Statistical analyses mentioned above were performed using Graphpad Prism, version 10.3.1 (GraphPad Software, San Diego, CA).

Results

Prolongation of Coagulation Time by the Engineered PC

We attempted to design an engineered hPC to release the activated form. We inserted the furin cleavage sequences (KR, RKR, KRRKR, RKRRKR [2RKR], RKRRKRRKR, RKRRKRRKRRKR, RHQR, or RSKR) or P2A-derived sequence into the thrombin recognition site of mature PC protein (Figure 1A). Following this, we generated HEK293 cells stably expressing hPC and measured hPC:C in the supernatants. To compare the anticoagulant ability of engineered hPCs, we used the previously reported hPC with enhanced activity (PC-QGNSEDY: H10Q, S11G, S12N, E23S, N32E, N33D, and H44Y in light chain)³⁰ as a control. hPC:C was detected in the supernatants of cells expressing wild-type (WT) hPC, hPC-QGNSEDY, and several engineered PCs (Figure 1B). However, hPC:C was not detected in the supernatant of HEK293 cells expressing hPC-RKR, hPC-RHQR, and hPC-P2A (Figure 1B). hPC:C is determined via the cleavage of a synthetic substrate by hPC, which is activated by snake venom in the reagent.³¹ To verify whether our engineered hPCs were secreted in their active form, we measured hPC:C production in the absence of snake venom. Although hPC:C in the supernatants of cells expressing hPC-WT or hPC-QGNSEDY was not measured without snake venom, we detected a significant increase in hPC:C in the supernatants of cells expressing hPC-KRRKR, hPC-2RKR, hPC-RKRRKRRKR, and hPC-RKRRKRRKRRKR (Figure 1C). We further assessed the amount of active hPC-2RKR that was released. We compared the supernatant hPC:C without and with snake venom (Figure S1A). The ratio of hPC:C without snake venom (released as an active form) was ≈40% of the total hPC:C (Figure S1B). The ratio did not differ among 2RKR, RKRRKRRKR, and RKRRKRRKRRKR, while hPC-2RKR showed the highest secretion. Hence, we selected hPC-2RKR for subsequent experiments because of its high secretion efficiency (Figure 1B and 1C).

Figure 1.

Generation of engineered activated hPC (human protein C). A, Schematic presentation of the engineered hPC. An indicated a furin-cleaving peptide sequence was inserted into the thrombin cleavage site of hPC. B, hPC:C levels (mean±SEM [n=3–4]) in the supernatant obtained from the HEK293 cells stably expressing an indicated engineered hPC. C, hPC:C levels (mean±SEM [n=3–4]) measured without snake venom. The P values, 1-way ANOVA with Dunnett multiple comparison test compared with wild-type (WT) hPC. D and E, Normal human plasma was incubated with an indicated concentration of WT hPC, the previously reported engineered hPC (QGNSEDY), or the engineered activated hPC (hPC-2RKR) for 15 minutes. Activated partial thromboplastin time (APTT; D) and prothrombin time (PT; E; mean±SEM [n=3]) were measured using an automated coagulation analyzer. F, Protein S–deficient plasma was incubated with an indicated concentration of hPC or hPC-2RKR for 15 minutes. APTT (mean±SEM [n=3]) was measured using an automated coagulation analyzer. The P values, 2-way ANOVA with Bonferroni multiple comparison test compared with WT hPC. hPC:C: hPC activity expressed as the percentage of pooled human plasma. QGNSEDY: mutations at H10Q, S11G, S12N, E23S, N32E, N33D, and H44Y in hPC. 2RKR indicates RKRRKR; 3RKR, RKRRKRRKR; 4RKR, RKRRKRRKRRKR; EGF, epidermal growth factor-like domain; Gla, gamma-carboxyglutamic domain; and P2A, 2A peptide derived from porcine teschorivus-1.

We examined the anticoagulant potential of hPC-2RKR to measure the inhibition of human plasma coagulation. APTT and PT of normal human plasma was assessed by mixing with an indicated concentration of an hPC or hPC-2RKR from the supernatant. hPC-WT and hPC-QGNSEDY did not affect APTT and PT, whereas hPC-2RKR significantly prolonged PT and APTT in a concentration-dependent manner (Figure 1D and 1E). Moreover, the prolongation of APTT by hPC-2RKR was observed in PS-deficient plasma, suggesting that the anticoagulant activity of hPC-2RKR did not require the cofactor activity of PS (Figure 1F).

APC binds to protease-activated receptor 1 on endothelial cells and exerts cytoprotective effects via intracellular signaling.³² We then examined intracellular signaling in endothelial cells. Phosphorylation of extracellular signal-regulated kinases 1 and 2 was transiently induced by the addition of hPC-2RKR in human umbilical vein endothelial cells (Figure S2).

Inhibition of Pathogenic Thrombus by the Engineered PC In Vivo

To assess the anticoagulant activity of PC-2RKR in vivo, we used a mouse model of thrombus formation induced by reactive oxygen species production; thrombus formation in mice was observed using intravital microscopy. We created an AAV8 vector harboring WT mPC (mouse PC) or the engineered mPC corresponding to hPC-2RKR (mPC-2RKR) and injected the vector in adult C57BL/6J mice at different doses (low dose, 4.0×1010 vg; medium dose, 1.2×10¹¹ vg; and high dose, 4.0×10¹¹ vg). In mice treated with the AAV8 vector harboring mPC (WT), although the mPC antigen levels significantly increased in a dose-dependent manner, plasma coagulation time assessed by APTT remained unaffected even at a high dose (Figure 2A and 2B). By contrast, in mice treated with the AAV8 vector harboring mPC-2RKR, coagulation time was prolonged in a dose-dependent manner; however, compared with mice expressing mPC (WT), those expressing mPC-2RKR showed lower increases in mPC antigen levels (Figure 2A and 2B). Furthermore, in mice expressing mPC-2RKR, a significant decrease in FV:C in plasma was observed (Figure 2C). FVIII:C tended to decrease in a dose-dependent manner in mice treated with mPC-2RKR; however, compared with FV:C, mPC-2RKR expression exerted a lower effect on FVIII:C (Figure 2D). Furthermore, no difference between both groups was observed in the AAV genome in the liver following the vector injection (Figure 2E).

Figure 2.

Prevention of pathogenic thrombus formation by the engineered activated PC (protein C) in mice. Adeno-associated virus (AAV)-8 vectors expressing wild-type mPC (mouse PC) or the engineered activated mPC (mPC-2RKR) under the control of HCRhAAT promoter were intravenously administered into 7-week-old C57BL/6J male mice (low, medium, and high doses: 4.0×1010, 1.2×10¹¹, and 4.0×10¹¹ vg/mouse, respectively). A, Plasma mPC antigen (mPC:Ag) levels (mean±SEM [n=4]) at 4 weeks after vector injection were measured using ELISA. Activated partial thromboplastin time (APTT; B), FV:C (C), and FVIII:C (D) were measured using an automated coagulation analyzer. Values represent mean±SEM (n=4, except FVIII:C in mPC Medium [n=2] and mPC-2RKR High [n=3]). E, AAV genome (mean±SEM [n=4]) in the liver tissue at 8 to 12 weeks post-injection was measured using quantitative polymerase chain reaction. The P values, 2-way ANOVA with Bonferroni multiple comparison test comparing between mPC and mPC-2RKR at the same vector dose. F, Thrombus formation in testicular veins induced by laser-induced reactive oxygen species (ROS) production was observed using intravital confocal microscopy. The green signal indicates platelet thrombus formation shown by DyLight488 signals. Scale bars=20 µm. G, Signal intensities of thrombus formation were quantified using Las X Software and are expressed as the area of thrombus formation in vessel area (%). Values represent mean±SEM (n=3–4). The P values, 1-way ANOVA with Tukey multiple comparison test. HCRhAAT: an enhancer element of the hepatic control region of the ApoE/C1 gene and the human antitrypsin promoter. mPC:Ag: expressed as the percentage of pooled mouse plasma. FV:C: coagulation FV (factor V) activity expressed as the percentage of pooled mouse plasma. FVIII:C: coagulation FVIII (factor VIII) activity expressed as the percentage of pooled mouse plasma. 2RKR indicates RKRRKR; and CD42b, cluster of differentiation 42b.

We examined the therapeutic potential of mPC-2RKR expression via the AAV vector for inhibiting pathological thrombosis in vivo. We assessed thrombus formation in mice treated with the AAV8 vector harboring mPC and mPC-2RKR at a low vector dose (4.0×1010 vg). Although the prolongation in APTT was marginal in mice treated with low-dose AAV 8 vector harboring mPC-2RKR (Figure 2B), mPC-2RKR expression significantly inhibited laser-induced reactive oxygen species–elicited thrombus formation in testicular veins (Figure 2F and 2G; Figure S3; Videos S1 and S3). The mPC (WT) expression did not inhibit reactive oxygen species–induced thrombus formation (Figure 2F and 2G; Figure S3; Videos S1 and S2).

Phenotypic Correction of PC-Deficient Mice via Genome Editing

Because conventional AAV-mediated gene therapy did not contribute to an increase in PC at the neonatal stage (Figure S4), we used genome editing technology to achieve a therapeutic effect during this period. We inserted the donor sequence into the Alb gene to target intron 14 located immediately after the terminal codon sequence.³³ We designed to express a target cDNA as a conjugated gene with Alb by the addition of a self-cleavage P2A peptide sequence instead of the terminal codon sequence of Alb (Figure 3A). The target cDNA was expected to be expressed only via HDR-mediated insertion. We confirmed that the AAV vector-mediated expression of SaCas9 in mice efficiently induced double-strand break at the target site in the liver (Figure S5A) and that it did not decrease plasma albumin levels (Figure S5B).

Figure 3.

Insertion of cDNA sequence by genome editing in neonatal mice. A, Schematic presentation of genome editing experiment. Self-cleaving P2A (2A peptide derived from porcine teschorivus-1) peptide sequence followed by EGFP (enhanced green fluorescent protein) cDNA was inserted into double-strand break (DSB) induced by SaCas9 (Staphylococcus aureus Cas9) with homology-directed repair (HDR) in vivo. B through F, Neonatal C57BL/6J mice received an adeno-associated virus (AAV) donor vector harboring EGFP (2.0×10¹¹ vg/mouse) without (EGFP donor) or with an AAV vector expressing SaCas9 and single guide RNA for intron 14 of Alb locus (6.0×1010 vg/mouse; SaCas9 and EGFP donor). B, mRNA expression of EGFP fused with Alb in the liver was detected by reverse transcription polymerase chain reaction. Representative data from 3 mice are shown. Liver tissues obtained from C57BL/6J were used as negative control. C, Immunoblotting with anti-GFP (green fluorescent protein) antibody in liver tissues obtained from 3 mice. Representative data from 3 mice are shown. Liver tissues obtained from C57BL/6J were used as negative control. D, EGFP expression in liver specimens assessed by EGFP immunostaining was photographed using an all-in-one microscope. The representative photographs were obtained from 4 independent experiments. Higher magnifications of the boxed regions are shown in bottom. Scale bars=200 µm. E, Quantitative evaluation of the EGFP-positive cells in the liver was performed with BZ-X 700 imaging software. Values represent mean±SEM (EGFP donor, n=2; SaCas9 and EGFP donor, n=4). The P value, 2-tailed Student t test. F, Polymerase chain reaction (PCR) analysis of genomic DNA obtained from indicated organs to examine HDR and insertion at DSB by nonhomologous end joining (NHEJ) at 6 weeks post-injection. PCR analysis can distinguish insertion via HDR and that via NHEJ by product size. ITR indicates inverted terminal repeat; and WT, wild type. *A nonspecific band at 60 kDa is observed in all samples.

To examine the genome editing efficacy of this system, we administered AAV8 vectors expressing SaCas9 and sgRNA (6.0×1010 vg) together with AAV8 vectors harboring P2A-fused EGFP (enhanced green fluorescent protein) cDNA sequence (2.0×10¹¹ vg) to neonatal C57BL/6J mice (Figure 3A). We confirmed the expression of EGFP mRNA conjugated with Alb and the efficient cleavage of EGFP protein by P2A sequence (Figure 3B and 3C). Immunofluorescence staining revealed EGFP-positive cells in 7.08±0.92% of all liver cells (Figure 3D and 3E). Remarkably, cDNA was mainly inserted by HDR in the liver (Figure 3F).

Further, we used mPC-2RKR for genome editing therapy to examine whether it would improve the phenotype of homozygous PC-deficient mice (Proc^−/− mice). We generated Proc−/− mice via genome editing and found that they died within a week (Figure 4B; Figure S6). We administered the same doses of AAV8 vectors expressing SaCas9 and sgRNA and AAV8 vectors harboring P2A-fused Proc-2RKR cDNA sequence to newborn pups born from crosses between Proc^+/− mice. However, all Proc^−/− mice died within 2 to 3 days after vector administration (data not shown). Because it was considered that Proc^−/− mice would die before achieving the efficacy of the genome editing treatment, it was necessary to ensure their survival until the appearance of therapeutic effects. We observed that crossing Proc^−/− mice with FVIII-deficient mice (F8^−/− mice) ensured their normal survival (Figure 4B). As reported previously,³⁴ the administration of an antibody against FVIII resulted in the complete disappearance of plasma FVIII in mice for 1 week (Figure S7). Therefore, we periodically administered anti-FVIII antibodies to Proc^+/− pregnant mice, expecting that IgG would be transferred to the neonates through the placenta, following which genome editing treatment could be administered to the neonates (Figure 4A). This method confirmed that genome editing to express mPC-2RKR ameliorated the survival of Proc^−/− mice (Figure 4B). After 6 weeks of genome editing in Proc^−/− mice, mPC antigen levels increased by 286.5±87.6% (Figure 4C) and FV:C decreased significantly (Figure 4D). However, the reduction of FVIII:C (Figure 4E) and prolongation of APTT (Figure 4F) in Proc^−/− mice treated with genome editing was marginal (Figure 4E and 4F). Further, some Proc^−/− mice treated with 10× lower dose of AAV vector could also survive (Figure 4B), but plasma FV:C was not significantly decreased (Figure 4D). By contrast, Proc^−/− F8^−/− mice showed a significant prolongation of APTT and reduction of FVIII:C (Figure 4E and 4F).

Figure 4.

Phenotypic correction of PC (protein C)-deficient homozygotic mice treated with genome editing to insert the engineered activated PC. A, Proc^+/− female mice were treated with an FVIII (factor VIII) antibody (No. GMA8015; 20 µg/mouse) and then mated with Proc^+/− male mice. The antibody administration was repeated every 7 days until delivery. All neonatal mice were treated via genome editing with adeno-associated virus (AAV)-8 vector harboring mPC-2RKR (2.0×10¹¹ vg/mouse [high dose] or 2.0×1010 vg/mouse [low dose]) and AAV8 vector expressing SaCas9 (Staphylococcus aureus Cas9) and single guide RNA for intron 14 of Alb locus (6.0×1010 vg/mouse [high dose] or 6.0×10⁹ vg/mouse [low dose]), together with 1 µg of the FVIII antibody. B, Kaplan-Meier survival curves for Proc^−/− mice (n=13), Proc^−/− mice treated with genome editing (high dose, n=9; low dose, n=10), and Proc^−/− F8^−/− mice (n=10). The P values, log-rank test, compared with Proc^−/− mice. C through F, Plasma mouse PC antigen (mPC:Ag; C), FV:C (D), and FVIII:C (E), and coagulation time (activated partial thromboplastin time [APTT]; F). Values represent mean±SEM (n=5–10, except Proc^−/− Treated [low dose, n=2]). The P values, 1-way ANOVA with Dunnett multiple comparison test compared with the wild-type (WT) C57BL/6J mouse. mPC:Ag: expressed as the percentage of pooled mouse plasma. FV:C: coagulation FV (factor V) activity expressed as the percentage of pooled mouse plasma. FVIII:C: coagulation FVIII activity expressed as the percentage of pooled mouse plasma. 2RKR indicates RKRRKR; E0, embryonic day 0; mAb, monoclonal antibody; mPC, mouse protein C; and P0, postnatal day 0.

Upon activation, PC forms a complex in plasma with serin protease inhibitors such as PC inhibitor and α1-antitrypsin, thereby inhibiting its activity.^35,36 We next examined whether plasma mPC-2RKR expressed by knock-in genome editing would form a complex with other proteins. Two extra bands specifically reacted with anti-mPC antibody were identified in the plasma obtained from PC-deficient mice expressing mPC-2RKR (42.4±5.56%, n=3; Figure S8), suggesting that mPC-2RKR forms a complex with some molecules in the plasma.

We further assessed in vivo thrombus formation and bleeding tendency in Proc^−/− mice treated with genome editing to express mPC-2RKR. We did not observe microvascular thrombus formation in the lung and liver of Proc^−/− mice treated with genome editing (Figure 5A). Compared with WT C57BL/6J mice, Proc^−/− mice treated with genome editing with higher vector dose and Proc^−/− F8^−/− mice exhibited prolonged bleeding time following tail clipping (Figure 5B and 5C). On the contrary, bleeding after tail clipping was not significantly enhanced in Proc^−/− mice treated with genome editing with lower vector dose (Figure 5C and 5D).

Figure 5.

Thrombus formation and bleeding phenotype of PC (protein C)-deficient homozygotic mice treated with genome editing to insert the engineered activated PC. Proc^+/− female mice were treated with an FVIII (factor VIII) antibody (No. GMA8015; 20 µg/mouse) and then mated with Proc^+/− male mice. The antibody administration was repeated every 7 days until delivery. All neonatal mice were treated via genome editing with adeno-associated virus (AAV)-8 vector harboring mPC-2RKR (2.0×10¹¹ vg/mouse [high dose] or 2.0×1010 vg/mouse [low dose]) and AAV8 vector expressing SaCas9 (Staphylococcus aureus Cas9) and single guide RNA for intron 14 of Alb locus (6.0×1010 vg/mouse [high dose] or 6.0×10⁹ vg/mouse [low dose]), together with 1 µg of the FVIII antibody. A, Representative photographs of hematoxylin-eosin staining of liver and lung tissues in an indicated mouse. Scale bars=100 µm. B, Representative photograph of filter paper to assess bleeding after tail clipping in an indicated mouse. C and D, Bleeding time and volume after the tail clip assay. Values represent mean±SEM (n=3–7, except Proc^−/− mice treated with genome editing [Proc^−/− Treated, n=2]). The P values, 1-way ANOVA with Dunnett multiple comparison test compared with the wild-type (WT) C57BL/6J mouse. 2RKR indicates RKRRKR; and mPC, mouse protein C.

Discussion

Purpura fulminans is a rare, life-threatening thrombotic disorder in newborns typically caused by homozygous or compound heterozygous abnormality in PROC or PROS1.^9,10 This disease generally occurs on the first day of life and rapidly progresses through multiorgan failure caused by thrombotic occlusion of vessels. Even if the patients survive with proper and timely diagnosis and treatments, neonates and children with PC or PS deficiency remain at a risk of developing severe thrombotic complications. These patients require lifelong therapy with PC concentrates or fresh-frozen plasma and strict anticoagulation treatments, including warfarin or heparin, under frequent monitoring.³⁷ We succeeded at expressing PC in its active form and showed that genome editing therapy using the engineered PC (PC-2RKR) at the neonatal stage could help in the survival of mice with homozygous PC deficiency. This strategy can provide a cure for patients with purpura fulminans due to PC deficiency.

We succeeded in releasing PC in its active form by inserting the furin cleavage site between the light and heavy chains. PC circulates as a zymogen, and its anticoagulant function can be achieved by releasing the activation peptide through its cleavage with a thrombomodulin and thrombin complex.³⁸ The introduction of a furin cleavage site for the activation of coagulation FVII (factor VII) in gene therapy for an animal model of hemophilia has previously been reported.³⁹ Compared with the previously reported engineered PC (QGNSEDY), which enhances the binding of the gamma-carboxyglutamic domain on the phospholipid surface,³⁰ PC-2RKR engineered in the present study further extended the coagulation time of human plasma. The expression of PC-2RKR in mice resulted in the prolongation of coagulation time owing to the inhibition of FV and FVIII; additionally, it led to the improvement of pathological thrombosis in WT mice and the survival of Proc^−/− mice. Moreover, compared with the Proc^−/− F8^−/− mice, the Proc^−/− mice treated with genome editing to express PC-2RKR exhibited a minimal prolongation of APTT and a lower incidence of thrombus formation along with lesser bleeding time at a lower vector dose, thereby predicting that treatment with PC-2RKR led to fewer hemorrhagic complications at an appropriate dosage.

In the present study, we observed that in WT mice showing an inhibition of pathological thrombosis and Proc^−/− mice exhibiting an improvement of survival by PC-2RKR, albeit FV and FVIII decreased by about only 50% and 10%, respectively. In addition, Proc^−/− F8^−/− mice were viable for >1 year without growth retardation (data not shown), although FVIII inhibition by PC-2RKR was marginal in genome editing therapy or gene therapy. The phenotype of Proc^−/− F8^−/− mice is different from that resulting from the crossbreeding of PC knockout mice with coagulation FXI (factor XI)-deficient mice.⁴⁰ The complete suppression of FXI is insufficient to improve the phenotype, thereby leading to growth retardation and death from thrombosis within 3 months.⁴⁰ Currently, anticoagulation therapy for PC deficiency includes warfarin, heparin, and direct oral anticoagulants, which inhibit the production and activity of coagulation factors.¹⁷ Because warfarin and heparin target multiple coagulation factors to exert anticoagulant activity, bleeding complications are major adverse events.⁴¹ For PC deficiency, targeting FV and FVIII, which are direct targets of PC, may provide an effective anticoagulant treatment. Conversely, loss of FVIII was associated with a prolonged bleeding time. Survival of PC-deficient mice partially improved by crossbreeding them with FXI-deficient mice.⁴⁰ Targeting FXI and administering PC preparations may allow for safer and more effective drug therapy for congenital PC deficiency.

The genome editing therapy targeting liver can be applied to coagulation factor deficiencies and liver metabolic diseases. Because the liver plays a central role in the metabolic processes, the abnormality of a key enzyme can progress to life-threatening conditions, such as urea cycle disorders, glycogen storage diseases, and Wilson disease.^42,43 Liver transplantation remains the sole treatment strategy to cure these diseases.⁴⁴ Although liver transplantation has cured hPC deficiency,^45,46 performing this procedure for all patients may not be possible owing to the lack of a donor or its highly invasive nature.⁴⁷ Gene therapy and genome editing therapy targeting the liver can become alternative therapies to cure these diseases.^33,48,49 Conventional gene therapy with AAV vectors is not expected to achieve a sustained therapeutic effect in newborns because of the potential dilution of the AAV genome during hepatocyte proliferation.⁵⁰

On the other hand, the therapeutic effect of genome editing is expected to sustain for the long term, thereby leading to a disease cure. In the present study, we delivered the genome editing tool Cas9 using an AAV vector; however, Cas9 expression within the cells should be transient to reduce the immunogenicity and off-target effects in the real clinical setting.⁵¹ Recently, the delivery of Cas9 mRNA and sgRNA with lipid nanoparticles has reportedly succeeded in disrupting target gene editing in the liver for human transthyretin amyloidosis.⁵² In the future, simultaneously administering lipid nanoparticle for transient Cas9 expression with an AAV vector as a donor sequence for gene insertion in the liver in vivo would be a better strategy. Moreover, we reported the application of engineered AsCas12f (Acidibacillus sulfuroxidans Cas12f) for neonatal genome editing in mice.⁵³ The small size of AsCas12f allows knock-in genome editing in hemophilia B mice using a single AAV vector, and this strategy is a potential alternative approach.

In the current study, we inserted target genes in ≈7% of liver cells with HDR in in vivo genome editing. HDR appeared to be the main mechanism for inserting the target gene because we could detect genome-edited DNA generated mainly by HDR-mediated insertion. Nonhomologous end joining is the predominant repair mechanism following double-strand break, and the frequency of HDR is less efficient. We have previously reported that nonhomologous end joining is the main mechanism for inserting the ectopic gene in intron 1 of the F9 gene in FIX-deficient mice.²⁶ The efficient HDR-based insertion targeting the Alb locus in the present study differs from that described in our previous report, although a similar method was applied in both experiments. The only difference was that the target site and homology arms included several exon sequences in the present study. The presence of exon sequences in the arms may increase the frequency of HDRs. Further verification is warranted to determine whether HDR efficiencies differ by donor sequence in in vivo genome editing.

This study has several limitations. First, understanding the kinetics and dynamics of PC-2RKR in vivo is warranted. We demonstrated that a part of mPC-2RKR formed a complex with other proteins, albeit we failed to reveal which proteins interacted with PC-2RKR in vivo. APC is inhibited by the interaction with PC inhibitor, but PC inhibitor did not reportedly exist in the plasma of mice.⁵⁴ One candidate protein that inhibits APC in mouse blood is α1-antitrypsin (Serpina1).³⁶ In future studies, we should clarify how PC-2RKR is present and the interaction with other proteins in the blood after the genome editing treatment. Second, the safety of performing genome editing to express APC immediately after diagnosing purpura fulminans must be carefully evaluated because of the bleeding tendency owing to consumptive coagulopathy in the acute phase.⁵⁵ Following the replacement therapy in humans in clinical settings,^12–14 it would be preferable to apply genome editing therapy after the bleeding tendency has ameliorated. Further, we must continue the protein replacement therapy for a certain period after diagnosis because manifestation of the therapeutic effects of genome editing requires several days or weeks. In addition, the level of APCs to be expressed for human therapy remains unclear. In the present study, we successfully expressed a high level of APC in the blood following genome editing. However, the minimum dosage required for the therapeutic activity should be determined before conducting a clinical trial to ensure safety. As a part of PC-2RKR is present as a free form in circulation, the excessive protein in blood may impair physiological hemostasis, similar to anticoagulants. Thus, it is important to compare bleeding complications using our approach with those in clinically established approaches. Finally, we only evaluated the efficacy of the genome editing treatment in a mouse model. It is necessary to verify its safety and efficacy in animal species close to humans, including nonhuman primates.

In conclusion, we have identified a sequence that causes the extracellular secretion of PC in its active form. We succeeded in prolonging the survival of Proc^−/− mice by inserting this sequence into the Alb locus via genome editing. The expression of PC-2RKR, but not WT PC, prevented pathological thrombosis. Even at low efficiency, modifying protein function may improve the disease phenotype. It is important to improve the efficiency and modality of genome editing and the function of the target protein to develop an effective gene therapy for intractable diseases. This study can provide a meaningful approach to gene or genome editing therapies for various diseases.

Article Information

Acknowledgments

The authors acknowledge Dr H.H. Kazazian, Jr (University of Pennsylvania), for providing F8-deficient mice. They greatly acknowledge Dr Koji Suzuki (Suzuka University of Medical Science) for his suggestions to revise the manuscript. They also thank Yaeko Suto, Mika Kishimoto, Tamaki Aoki, Sachiyo Kamimura, Mai Hayashi, Yuiko Ogiwara, Nagako Sekiya, Tomoko Noguchi, Ayano Suto, Hiromi Ozaki, and Hiroko Hayakawa of Jichi Medical University for their technical assistance. Optima XE-90 was subsidized by Japan Keirin Autorace foundation through its promotion funds from Keirin Race. The graphic abstract was created in BioRender.com. Conceptualization: T. Togashi and T. Ohmori. Methodology: T. Togashi, N. Baatartsogt, Y. Nagao, Y. Kashiwakura, O. Nureki, and T. Ohmori. Investigation: T. Togashi, N. Baatartsogt, Y. Nagao, Y. Kashiwakura, M. Hayakawa, T. Hiramoto, and T. Ohmori. Visualization: N. Baatartsogt and T. Fujiwara. Funding acquisition: O. Nureki and T. Ohmori. Project administration: O. Nureki and T. Ohmori. Supervision: E. Morishita, O. Nureki, and T. Ohmori. Writing—original draft: T. Togashi and T. Ohmori. Writing—review and editing: T. Togashi, N. Baatartsogt, Y. Nagao, Y. Kashiwakura, M. Hayakawa, T. Hiramoto, T. Fujiwara, E. Morishita, O. Nureki, and T. Ohmori.

Sources of Funding

This study was supported by the Japan Agency for Medical Research and Development grants JP24bm1223004 (T. Ohmori), JP24bm1323001 (T. Ohmori), JP22am0401005 (O. Nureki and T. Ohmori), JP22ae0201007 (T. Ohmori), and JP22fk0410037 (T. Ohmori) and SENSHIN Medical Research Foundation (T. Ohmori).

Disclosures

T. Togashi, N. Baatartsogt, Y. Kashiwakura, M. Hayakawa, T. Hiramoto, and T. Ohmori are inventors of the patent for the engineered PC sequence used in the present study. The other authors report no conflicts.

Supplemental Material

Table S1

Figures S1−S8

Major Resources Tables

Movies S1−S3

Supplementary Material

atv-44-2616-s001.pdf ^{(4.7MB, pdf)}