Targeted insertion of an anti-CD2 monoclonal antibody transgene into the GGTA1 locus in pigs using FokI-dCas9

Abstract

Xenotransplantation from pigs has been advocated as a solution to the perennial shortage of donated human organs and tissues. CRISPR/Cas9 has facilitated the silencing of genes in donor pigs that contribute to xenograft rejection. However, the generation of modified pigs using second-generation nucleases with much lower off-target mutation rates than Cas9, such as FokI-dCas9, has not been reported. Furthermore, there have been no reports on the use of CRISPR to knock protective transgenes into detrimental porcine genes. In this study, we used FokI-dCas9 with two guide RNAs to integrate a 7.1 kilobase pair transgene into exon 9 of the GGTA1 gene in porcine fetal fibroblasts. The modified cells lacked expression of the αGal xenoantigen, and secreted an anti-CD2 monoclonal antibody encoded by the transgene. PCR and sequencing revealed precise integration of the transgene into one allele of GGTA1, and a small deletion in the second allele. The cells were used for somatic cell nuclear transfer to generate healthy male knock-in piglets, which did not express αGal and which contained anti-CD2 in their serum. We have therefore developed a versatile high-fidelity system for knocking transgenes into the pig genome for xenotransplantation purposes.

Figure 1

Targeting of GGTA1 in WT pig fetal fibroblasts using FokI-dCas9. (A) Target region within GGTA1 exon 9. Green arrows indicate target sites for guide RNAs GT-3 and GT-4. Red arrowheads indicate binding sites for primers GTFS-F1 and GTFS-R1, used to amplify the target region after genome editing. (B) Flow cytometric analysis of αGal expression (IB4-FITC) on fibroblasts, three days after co-transfection with expression vectors for FokI-dCas9, GT-3 and GT-4. A substantial proportion of the co-transfected cells (red line) showed a marked reduction in αGal expression compared to control vector-transfected cells (blue line); black line, unstained WT fibroblasts. (C) Sequence analysis of the target region, amplified from genomic DNA isolated from the pool of co-transfected cells. One clone contained a 25 bp deletion around the predicted cleavage site (arrowhead). The predicted amino acid sequences of WT and mutated α1,3-galactosyltransferase are shown below the sequence, indicating an altered stretch of 33 residues (italics) and truncation of 85 residues in the mutant protein.

Figure 2

Validation of the anti-CD2 mAb GGTA1 targeting construct. (A) Top, GGTA1 genomic structure. Middle, knock-in backbone showing 5′ and 3′ homology arms (5′-HA and 3′-HA), neomycin resistance cassette (NeoR), multiple cloning site (MCS), and polyadenylation signal (pA). Into this was cloned (bottom) the CMV immediate early enhancer (CMV IE), mouse H-2K^b promoter, intron, and coding regions for the heavy and light chains of anti-CD2 mAb diliximab (αCD2-HC and αCD2-LC) linked by a furin cleavage site-F2A ribosome skip signal (f2A). The isotype of diliximab is human IgG3. (B) Detection of anti-CD2 mAb diliximab secreted by stably transfected COS-7 cells. Human leukocytes were incubated with culture supernatant, and mAb binding to CD3⁺ T cells was detected with anti-human IgG3. Red line, supernatant (S/N) from COS-7 cells transfected with the anti-CD2 knock-in construct; blue line, positive control (62.5 ng/ml purified diliximab); black line, supernatant from vector-transfected cells.

Figure 3

Analysis of pig fetal fibroblast anti-CD2 knock-in clone #3. (A) Absence of αGal expression by clone #3 (red line); blue line, WT fibroblasts; black line, unstained WT fibroblasts. (B) Presence of anti-CD2 mAb diliximab in the supernatant of clone #3 (red line) detected as described in the legend to Fig. 2; blue line, positive control (20 ng/ml purified diliximab); black line, supernatant from WT fibroblasts. (C) PCR analysis to confirm correct targeting in clone #3 and in two piglets generated from clone #3 by somatic cell nuclear transfer. The schematic diagram (top) shows the expected genomic configuration for integration of the knock-in construct in GGTA1; the upstream and downstream homology arms (HA) are shown in yellow. Two primer pairs (red arrowheads; 5′ = UKI-F3/UKI-R2; 3′ = 117-F/1123-R), each with one primer outside and one primer within the targeting construct, were used with genomic DNA isolated from clone #3 fibroblasts (lanes 1), WT fibroblasts (lanes 2), two cloned piglets (lanes 3 and 4), and one WT piglet (lanes 5). Clone #3 and both cloned piglets generated upstream and downstream products of the expected size (1506 bp, lanes 1, 3 and 4, left hand gel; and 947 bp, lanes 1, 3 and 4, right hand gel, respectively), which were confirmed by sequencing. MW, molecular weight markers (λ/HindIII + θX174/HaeIII).

Figure 4

Anti-CD2 knock-in pigs. (A) Healthy appearance of the pigs. (B) Absence of αGal expression on peripheral blood leukocytes (PBL) (red line); blue line, WT leukocytes; black line, unstained. (C) Presence of anti-CD2 mAb diliximab in the serum (diluted 1:50) of a knock-in pig (red line), detected as described in the legend to Fig. 2; blue line, positive control (WT pig serum spiked with 200 ng/ml purified diliximab); black line, WT pig serum.