Targeted Repair of CYBB in X-CGD iPSCs Requires Retention of Intronic Sequences for Expression and Functional Correction

Abstract

X-linked chronic granulomatous disease (X-CGD) is an immune deficiency resulting from defective production of microbicidal reactive oxygen species (ROS) by phagocytes. Causative mutations occur throughout the CYBB gene, resulting in absent or defective gp91^phox protein expression. To correct CYBB exon 5 mutations while retaining normal gene regulation, we utilized TALEN or Cas9 for exon 5 replacement in induced pluripotent stem cells (iPSCs) from patients, which restored gp91^phox expression and ROS production in iPSC-derived granulocytes. Alternate approaches for correcting the majority of X-CGD mutations were assessed, involving TALEN- or Cas9-mediated insertion of CYBB minigenes at exon 1 or 2 of the CYBB locus. Targeted insertion of an exon 1–13 minigene into CYBB exon 1 resulted in no detectable gp91^phox expression or ROS activity in iPSC-derived granulocytes. In contrast, targeted insertion of an exon 2–13 minigene into exon 2 restored both gp91^phox and ROS activity. This demonstrates the efficacy of two correction strategies: seamless repair of specific CYBB mutations by exon replacement or targeted insertion of an exon 2–13 minigene to CYBB exon 2 while retaining exon/intron 1. Furthermore, it highlights a key issue for targeted insertion strategies for expression from an endogenous promoter: retention of intronic elements can be necessary for expression.

Sweeney et al. demonstrate TALEN- and CRISPR/Cas9-mediated targeted correction of CYBB mutations in iPSCs from X-CGD patients, and identify a key issue for the design of effective gene targeting and knockin strategies: that retention or inclusion of intronic elements may be necessary for expression from an endogenous promoter.

Introduction

X-linked chronic granulomatous disease (X-CGD) is an immune deficiency caused by mutations in the CYBB gene on the X chromosome, encoding the gp91^phox subunit of phagocyte NADPH oxidase, which is required for production of microbicidal reactive oxygen species (ROS) by phagocytes. Causative mutations for X-CGD occur throughout the 13 exons or adjoining intronic splice sites of the >30-kb CYBB gene, resulting in absent or defective gp91^phox protein and loss of ROS activity.¹ X-CGD patients have recurring, life-threatening bacterial and fungal infections and hyper-inflammation, resulting in granulomatous complications.² Allogeneic hematopoietic stem cell (HSC) transplantation can cure X-CGD, but many patients lack a suitable donor, and graft-versus-host-disease remains a significant risk of allogeneic transplants. Gene transfer correction of autologous HSCs lacks these barriers, but current approaches using retrovirus or lentivirus vectors carry risk of oncogenic insertional mutagenesis due to random vector insertion. More specifically for X-CGD, autologous HSC gene therapy using retrovirus vectors has demonstrated clinical benefit as salvage therapy for life-threatening infection, but long-term gene marking has been low and life-threatening myelodysplasia caused by insertional mutagenesis has been observed with vector derived from murine spleen focus-forming virus.3, 4

We previously demonstrated a “safe-harbor” gene therapy approach for correction of X-CGD patient induced pluripotent stem cells (iPSCs) using zinc-finger nucleases (ZFNs) to mediate targeted insertion of a codon-optimized CYBB minigene into the AAVS1 locus under the control of a constitutive CAG promoter.5, 6 This resulted in constitutive expression of gp91^phox, which restored ROS production upon in vitro differentiation of iPSCs into granulocytes. However, in clinical gene therapy approaches for X-CGD, constitutive ectopic expression of the gp91^phox protein or a consequent aberrant production of ROS in stem cells prior to myeloid differentiation could potentially impair HSC engraftment or otherwise alter stem cell function in an unexpected manner.7, 8

Alternate approaches to address this issue include targeted correction or gene transfer at the CYBB locus, resulting in normal regulation of gp91^phox expression by the endogenous CYBB promoter. Here, we describe several targeted gene transfer strategies for seamless exon replacement of CYBB exon 5 mutations or for transfer of CYBB minigenes to the start sites of exon 1 or exon 2 of the CYBB locus, using site-specific TAL effector nucleases9, 10 (TALENs) or clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9 nuclease11, 12, 13 targeting these sites. The varying effectiveness of these approaches for restoring gene expression and granulocyte function are shown, highlighting an important issue regarding the necessity of intronic elements in the design of targeted gene transfer strategies and donor constructs.

Results

Seamless Targeted Repair of CYBB Exon 5 Mutations

Mutations in exon 5 of CYBB account for approximately 10% of known causative mutations for X-CGD. For seamless repair of all exon 5 mutations, we assessed the targeted replacement of CYBB exon 5 using a donor construct containing the repaired exon 5 sequences interrupted by a puroΔtk gene cassette flanked by piggyBac transposon elements and plasmid(s) expressing either the CYBB5-L4/R4 TALEN pair or CYBB5-g4 CRISPR/Cas9 targeting CYBB exon 5 (Figure 1). X-CGD iPSCs derived from patients with 458T>G mutation or 461A deletion in CYBB exon 5 were nucleofected with pCYBB5-pB-PGK-putk donor plasmid and either CYBB5-L4 and -R4 TALEN expression plasmids or CYBB5-g4 CRISPR expression plasmid. Puromycin-resistant iPSCs were expanded and screened for targeted or off-target vector insertions by PCR (Figure S1A). Precise targeting of the donors to CYBB exon 5 was confirmed by sequencing in iPSC clones that screened positive for targeted insertions without off-target inserts (Figure S1B). Based on PCR screening, overall correction efficiencies of targeted insertion without random insertions were 12%–75% using the TALEN pair or 24%–38% using CYBB5-g4 CRISPR/Cas9 (Table 1). Following transient expression of piggyBac transposase, ganciclovir-selected iPSC clones were screened for excision of the piggyBac-flanked puroΔtk cassette; excision of the piggyBac cassette resulted in reconstitution of the corrected exon 5 for seamless repair of the patient’s CYBB mutation (Figure S1C).

Figure 1.

CYBB Exon 5 Correction Strategy

(A) CYBB locus, with magnified representation of exon 5, including recognition sites for CYBB5-g4 CRISPR and CYBB5 L4/R4 TALENs, and locations of mutations corrected in this study (458T>G and 461A del). Also shown is a portion of the CYBB5 donor plasmid construct, containing homology arms (LHA and RHA) to the introns surrounding exon 5, with exon 5 split into two portions separated by a PGK promoter-driven puroΔtk drug-selection cassette flanked by piggyBac transposon 5′ and 3′ repeats (pB). Splitting of exon 5 divides the recognition sites for the CRISPR or TALENs, so that they do not cut the donor construct before or after insertion. Gray arrowheads denote locations of PCR primers used for screening for random insertions of the donor in which portions of the plasmid backbone outside of the homology arms are retained. (B) CYBB exon 5 and surrounding regions after targeted insertion of the CYBB5 donor construct. White arrowheads denote locations of PCR primers used for screening for targeted insertion of the donor. (C) CYBB exon 5 and surrounding regions after piggyBac transposase-mediated excision of the piggyBac cassette, resulting in constitution of the divided exon 5 for seamless correction of exon 5 mutations. Black arrowheads denote locations of PCR primers used for screening for piggyBac excision.

Table 1.

Gene Correction Efficiencies for X-CGD iPSCs in These Studies

iPSC Line	CYBB Mutationa	Donor/Nuclease	Targeted (%)b	Random (%)c	Corrected (%)d
E5c5	exon 5 (458T>G)	exon 5/TALEN	5/17 (29)	15/17 (88)	2/17 (12)
iGP91-07	exon 5 (458T>G)	exon 5/TALEN	43/44 (98)	11/44 (25)	33/44 (75)
iGP91-07	exon 5 (458T>G)	exon 5/CRISPR	14/17 (82)	13/17 (76)	4/17 (24)
iGP91-09s	exon 5 (461A del)	exon 5/CRISPR	22/24 (92)	15/24 (63)	9/24 (38)
iGP91-06	exon 3 (217C>T)	E1–13 opte/TALEN	1/22 (5)	20/22 (91)	1/22 (5)
iGP91-07	exon 5 (458T>G)	E1–13 opte/TALEN	4/39 (10)	33/39 (85)	4/39 (10)
iGP91-07	exon 5 (458T>G)	E1–13 cDNAf/TALEN	19/22 (86)	7/22 (32)	14/22 (64)
iGP91-07	exon 5 (458T>G)	E2–13g/CRISPR	14/19 (74)	8/19 (42)	10/19 (53)
iGP91-10s	intron 10 (+2 T>A)	E2–13g/CRISPR	9/9 (100)	3/9 (33)	6/9 (67)
iGP91-14s	exon 7 (676C>T)	E2–13g/CRISPR	3/4 (75)	1/4 (25)	2/4 (50)

^aAll X-CGD iPSC lines used in these studies contain CYBB mutations that eliminate gp91^phox expression.

^bFrequency of targeted insertion of the donor vector to the appropriate location in the CYBB locus in puromycin-resistant clones, based on PCR and sequence analysis.

^cFrequency of random insertions of the donor vector in which plasmid sequences outside of the homology arms were retained in puromycin-resistant clones, based on PCR analysis.

^dFrequency of puromycin-resistant clones containing a correctly targeted donor insertion without random insertions.

^eCodon-optimized exon 1–13 cDNA minigene donor.

^fNormal exon 1–13 cDNA minigene donor.

^gCodon-optimized exon 2–13 cDNA minigene donor.

Granulocytes differentiated from seamless gene-corrected iPSCs (Figure 2A) exhibited restoration of gp91^phox protein expression to levels 73%–100% of normal controls based on mean fluorescence intensity (MFI) of the gp91^phox-expressing population by flow cytometry (Figures 2B and S2A). This expression was specific to differentiated granulocytes and monocytes/macrophages based on co-staining with antibody to CD13 myeloid marker, whereas undifferentiated corrected iPSCs lacked gp91^phox expression, demonstrating normal regulation of the corrected protein (Figure 2C). Uncorrected X-CGD myeloid cells lacking gp91^phox expression exhibited lower intensities of CD13 antibody staining than normal or corrected cells expressing gp91^phox, as has been demonstrated previously.¹⁴ iPSC-derived mature granulocytes also demonstrated restoration of ROS production by dihydrorhodamine (DHR) assay, to levels 90%–100% of normal controls based on MFI of the DHR⁺ population (Figures 2D and S2B). In iPSCs with seamless targeted correction of CYBB exon 5, no mutations were detected in the top ten predicted off-target sites for the CYBB5-L4/R4 TALEN pair (Table S1) or CYBB5-g4 CRISPR (Table S2).

Figure 2.

Functional Correction of Granulocytes Differentiated from CYBB Exon 5 Gene-Corrected X-CGD Patient iPSCs

(A) Giemsa-stained cytospin of granulocytes differentiated in vitro from iPSCs, demonstrating characteristic polymorphonuclear morphology of neutrophils or eosinophils (arrows), with other myeloid lineages also present in some differentiations. Shown are cells differentiated from iGP91-07 iPSCs corrected with CYBB5 TALENs (bar, 20 μm). (B) Antibody staining in differentiated granulocytes for gp91^phox protein expression, co-stained with CD13 myeloid surface marker, in uncorrected cells (left) versus gene-correction cells (middle) and normal controls (right). Shown are data from uncorrected and TALEN-corrected E5c5 iPSCs and normal control iPSCs. (C) Undifferentiated gene-corrected iPSCs co-stained for gp91^phox expression and TRA-1-60 as a marker of pluripotency. Data shown are from TALEN-corrected iGP91-07 iPSCs. (D) DHR assay of ROS activity in granulocytes differentiated from uncorrected patient iPSCs (left column) versus gene-corrected cells (middle column) and normal controls (right column). MFI of gated DHR⁺ populations are listed above those populations. Data shown are from TALEN-corrected iGP91-07 (top row) and CRISPR-corrected iGP91-09 iPSCs (bottom row) with normal peripheral blood granulocyte controls.

Targeted Gene Transfer of CYBB Exon 1–13 Minigene to CYBB Exon 1

As an alternate approach for correction of the majority of CYBB mutations, plasmids expressing CYBB1-L3/R3 TALENs targeting the start site of the endogenous CYBB gene were constructed, along with a donor plasmid containing a CYBB minigene (either codon-optimized or normal exon 1–13 cDNA) with a poly(A) signal and a loxP-flanked puromycin resistance gene cassette with a constitutive promoter (Figure 3). X-CGD patient iPSCs were nucleofected with CYBB1-L3/R3 TALEN expression plasmids and the donor plasmid, and puromycin-resistant iPSCs were expanded and screened for targeted or off-target vector insertions by PCR (Figure S3A). Five percent to 10% of puromycin-resistant iPSC clones exhibited targeted insertion without random inserts for the codon-optimized donor or 64% for the normal cDNA donor (Table 1). Precise targeting of the donors to the start site of CYBB exon 1 in iPSC clones was confirmed by DNA sequencing (Figure S3B), which also identified the homologous recombination site in exon 1 for some clones corrected by the codon-optimized donor, based on the switchover from the endogenous exon 1 sequence to the codon-optimized sequence.

Figure 3.

CYBB Exon 1–13 Correction Strategy

(A) CYBB locus, with magnified representation of exon 1, including recognition sites for CYBB1 L3/R3 TALENs. Also shown is a portion of the CYBB-E1–13 donor plasmid construct, containing homology arms (LHA and RHA) to regions flanking the insertion site, surrounding the CYBB exon 1–13 cDNA (either codon-optimized or normal) followed by poly(A) signal (pA) and a loxP-flanked CMV promoter-driven puromycin resistance drug selection cassette (puro). Gray arrowheads denote locations of PCR primers used for screening for random insertions of the donor in which portions of the plasmid backbone outside of the homology arms are retained. (B) CYBB exon 1 and surrounding regions after targeted insertion of the CYBB-E1–13 donor construct. White arrowheads denote locations of PCR primers used for screening for targeted insertion of the donor.

However, upon differentiation of corrected iPSCs into CD13⁺ granulocytes (Figures 4A and S4), there was no detectable gp91^phox expression (Figure 4B) or DHR activity (Figure S5) above the background levels present in uncorrected cells, for either the codon-optimized or normal exon 1–13 cDNA corrections. Similar results were obtained with or without Cre excision of the puromycin resistance expression cassette from the vector insert (data not shown). Because these donor constructs eliminate the downstream intronic and 3′-UTRs of CYBB from the minigene transcript, this suggests that efficient expression from the CYBB promoter may require either regulatory elements within these downstream regions or the process of transcript splicing. In contrast, this same codon-optimized CYBB cDNA was expressed efficiently from an intron-less elongation factor-1α (EF1α) short promoter¹⁵ by a lentiviral vector in X-CGD patient mobilized peripheral blood CD34⁺ hematopoietic stem/progenitor cells (HSPCs), resulting in restoration of ROS activity upon granulocyte differentiation (Figure S6), suggesting that failed gp91^phox protein expression from the exon 1–13 minigene in the CYBB locus is context specific to its placement in CYBB at exon 1 downstream of and relying upon activity of the CYBB promoter and proper pre-mRNA processing for expression.

Figure 4.

Analysis of gp91^phox Protein Expression in Granulocytes Differentiated from X-CGD iPSCs Corrected with CYBB Exon 1–13 Minigene

(A) Giemsa-stained cytospin of granulocytes differentiated in vitro from iPSCs, demonstrating characteristic polymorphonuclear morphology of neutrophils (bar, 20 μm). Shown are uncorrected (left) and codon-optimized minigene-corrected cells (middle) differentiated from iGP91-06 iPSCs, and normal, healthy donor peripheral blood neutrophils (right). (B) Antibody staining for gp91^phox protein expression in granulocytes differentiated from uncorrected X-CGD iPSCs (left column) and CYBB exon 1–13 minigene corrected cells (middle column) versus normal controls (right column). Shown are data from GP91-06 corrected with codon-optimized cDNA (top row), iGP91-07 corrected with codon-optimized cDNA (middle row), and iGP91-07 corrected with normal, un-optimized cDNA (bottom row), with either normal iPSC-derived granulocytes (top row) or normal peripheral blood granulocyte controls (middle and bottom rows).

Targeted Gene Transfer of CYBB Exon 2–13 Minigene to CYBB Exon 2

Hypothesizing that CYBB intron 1 might provide the elements required for expression from the CYBB promoter, we designed CRISPR/Cas9 targeting CYBB exon 2 and constructed a plasmid donor containing a minigene consisting of a codon-optimized CYBB exon 2–13 cDNA with a poly(A) signal and a constitutively expressed puromycin resistance gene (Figure 5). X-CGD iPSCs containing mutations in exon 5, exon 7, or intron 10 were nucleofected with the CYBB2-g3 CRISPR/Cas9 expression plasmid and the donor plasmid. Puromycin-resistant iPSC clones exhibited an overall efficiency of 50%–67% correction without random inserts, based on PCR screening (Table 1; Figure S7).

Figure 5.

CYBB Exon 2–13 Correction Strategy

(A) CYBB locus, with magnified representation of exon 2, including recognition site for CYBB2-g3 CRISPR. Also shown is a portion of the CYBB-E2–13 donor plasmid construct, containing homology arms (LHA and RHA) to regions flanking the insertion site, surrounding the CYBB exon 2–13 codon-optimized cDNA followed by poly(A) signal (pA) and piggyBac transposon 5′ and 3′ repeats (pB) flanking a PGK promoter-driven puroΔtk drug-selection cassette. CYBB cDNA codon optimization in the donor removes the CRISPR recognition site, so that it will not cut the donor construct before or after insertion. Gray arrowheads denote locations of PCR primers used for screening for random insertions of the donor in which portions of the plasmid backbone outside of the homology arms are retained. (B) CYBB exon 2 and surrounding regions after targeted insertion of the CYBB-E2–13 donor construct. White arrowheads denote locations of PCR primers used for screening for targeted insertion of the donor.

Upon granulocyte differentiation, iPSCs corrected with the CYBB exon 2–13 minigene exhibited gp91^phox expression levels 64%–100% of normal control peripheral blood neutrophils, based on MFI of the positive cell populations (Figures 6A and S8). As expected for expression regulated by the endogenous CYBB promoter, gp91^phox expression in corrected cells was specific to monocyte/macrophage and granulocyte lineages, because gp91^phox was co-expressed with CD13 myeloid marker (Figure 6A), whereas undifferentiated iPSCs lacked gp91^phox expression (Figure 6B). Granulocytes from minigene-corrected iPSCs also exhibited DHR activity levels 68%–100% of normal controls, based on MFI of the positive cell populations (Figure 6C). Similar results were observed with or without piggyBac excision of the puromycin gene cassette in exon 2–13 minigene corrected iPSCs (data not shown). In iPSCs with targeted correction of CYBB exon 2, no mutations were detected in the top ten predicted off-target sites for the CYBB2-g3 CRISPR (Table S3).

Figure 6.

Functional Correction of Granulocytes Differentiated from CYBB Exon 2–13 Gene-Corrected X-CGD Patient iPSCs

(A) Antibody staining in differentiated granulocytes for gp91^phox protein expression, co-stained with CD13 myeloid surface marker in uncorrected cells (left) versus gene-corrected cells (middle) and normal controls (right). Shown are data from iGP91-10s patient iPSCs and normal iPSC control. (B) Undifferentiated gene-corrected iPSCs co-stained for gp91^phox expression and TRA-1-60 as a marker of pluripotency. Shown are data from corrected iGP91-14s iPSCs. (C) DHR assay of ROS activity in granulocytes differentiated from uncorrected patient iPSCs (left column) versus gene-corrected cells (middle column) and normal controls (right column). MFI of gated DHR⁺ populations are listed above those populations. Data shown are from iGP91-07 correction (top row), iGP91-10s correction (middle row), and iGP91-14s correction (bottom row), with either normal peripheral blood granulocyte control (top row) or normal iPSC-derived granulocyte controls (middle and bottom rows).

Discussion

We demonstrate here several TALEN- or CRISPR-mediated approaches for correction of mutations in the CYBB gene to restore granulocyte function in iPSCs of patients with X-CGD. An exon replacement strategy using a plasmid donor construct was shown to be effective for seamless correction of both a 458T>G mutation and 461A deletion in exon 5, restoring gp91^phox protein expression and ROS activity upon granulocyte differentiation, to levels comparable to that of healthy donor granulocytes. The ease of TALEN or CRISPR design allows for similar approaches to be generated against additional exons or specific mutations in CYBB, enabling the development of patient-specific gene repair therapeutics for X-CGD. A similar targeted approach for repair of a specific CYBB mutation was recently demonstrated by Flynn et al.¹⁶ using CRISPRs to target insertion of a plasmid-based donor containing CYBB exon 2 and the surrounding intronic sequences to repair a mutation in the splice acceptor site immediately preceding exon 2 in patient iPSCs, further illustrating the efficacy of targeted repair of individual X-CGD mutations without modifying intron sequences.

An alternate approach for generalized repair of CYBB is to target insertion of the entire CYBB cDNA or a minigene comprised of multiple exons into the endogenous CYBB locus, enabling correction of multiple mutations while retaining physiological regulation of gp91^phox protein expression from the CYBB promoter. A previous study by Laugsch et al.¹⁷ demonstrated restoration of gp91^phox expression and ROS activity in iPSCs from a patient with a mutation in exon 7 of CYBB, by utilizing a large bacterial artificial chromosome containing the endogenous sequence of exons 6 through 13 together with the intervening introns, for homologous recombination-mediated replacement of the CYBB mutation without nuclease enhancement. In the current study, TALEN-mediated targeted insertion of a minigene consisting of the normal or codon-optimized cDNA of all 13 exons of CYBB without introns into the start site of exon 1 of the endogenous CYBB gene resulted in no detectable gp91^phox expression, indicating that intronic or other downstream non-coding elements are necessary for expression from the CYBB promoter. In contrast, insertion of a codon-optimized cDNA encompassing exons 2 through 13 into exon 2 of the endogenous gene resulted in normal regulation of expression of gp91^phox and restoration of ROS activity, indicating that inclusion of intron 1 is sufficient for expression of the CYBB minigene from the CYBB promoter.

Numerous studies in mammalian cells have shown that gene expression can be enhanced by inclusion of an intron;18, 19, 20, 21, 22, 23 mechanisms for intronic enhancement of gene expression can include the presence of transcriptional enhancers within intron DNA or splicing-mediated effects on transcription or translation.24, 25 The majority of studies demonstrating intronic enhancement involving human cells, promoters, or genes have used plasmid or viral vectors due to prior limitations in precise genome engineering; however, the recent advent of customized CRISPRs and TALENs has allowed for unprecedented ease of targeted editing of the human genome, enabling us to test multiple targeting strategies and donor constructs including various intron and exon regions for correction of the CYBB locus in patient iPSCs. Our findings from these targeting studies demonstrated a pronounced intronic requirement for CYBB cDNA expression from the endogenous CYBB promoter that was not evident in CYBB expression constructs utilizing other promoters, as we detected no gp91^phox protein expression from the endogenous promoter unless intron 1 was retained in the transcription unit, whereas the same cDNA was shown to be efficiently expressed in HSPCs from either an intron-less EF1α short promoter in a lentiviral expression vector in the current study or from an intron-less synthetic MND promoter after targeted genomic insertion into the AAVS1 safe-harbor site in a previous study.²⁶ It is not clear whether the CYBB promoter merely requires intronic splicing for effective expression or whether there are necessary transcriptional enhancer elements in CYBB intron 1, although a recent publication by Frazão et al.²⁷ suggests a possible NF-κB enhancer element in intron 1, based on the association of a distant upstream NF-κB enhancer site with regions of the CYBB promoter and CYBB introns 1 and 2, and the finding that CYBB mRNA expression was reduced upon inhibition of NF-κB activity. Regardless of the particular intron-associated mechanism involved in CYBB expression, these observations highlight an important general design issue for effective gene targeting and knockin strategies—that donor constructs or targeting strategies may require inclusion or retention of intronic elements in order to achieve therapeutic or even detectable levels of transgene expression from an endogenous promoter.

The current study demonstrates the utility of iPSCs in providing a renewable source of patient-derived cells for modeling of gene correction strategies and vectors. However, at present, the clinical usage of iPSCs for the treatment of hematopoietic disorders remains a challenge due to difficulties with the efficient in vitro production of engraftable HSPCs from human iPSCs,²⁸ despite recent advancements.²⁹ Direct gene correction of engraftable somatic HSPCs could provide more immediate clinical benefit, although the use of plasmid DNA donors for electroporation of HSPCs generally results in substantial cytotoxicity with low correction efficiency. We recently demonstrated that ZFN-mediated targeted insertion of a CYBB minigene into the AAVS1 safe-harbor locus of X-CGD patient mobilized CD34⁺ HSPCs could be achieved efficiently with low cytotoxicity and without drug selection by utilizing an adeno-associated virus (AAV) donor instead of plasmid DNA, resulting in constitutive expression of gp91^phox from the synthetic MND promoter and production of functional neutrophils after engraftment in NSG mice.²⁶ A similar approach for CRISPR/Cas9 targeting CYBB exon 2 with an AAV-based donor for efficient targeted delivery of the CYBB exon 2–13 minigene could provide clinical benefit for the majority of X-CGD patients, while maintaining normal regulation of gp91^phox expression.

Materials and Methods

Approvals for Human Blood Use

Blood from healthy volunteers and X-CGD patients was obtained after written informed consent following the Declaration of Helsinki under the auspices of National Institute of Allergy and Infectious Diseases (NIAID) Institutional Review Board-approved protocols 05-I-0213 and 94-I-0073.

Human iPSC Generation, Characterization, and Maintenance

Transgene-free iPSCs were generated from peripheral blood CD34⁺ HSPCs of X-CGD patients as previously described30, 31 by transduction with Cre-excisable hSTEMCCA-loxP lentivirus³² (iPSC lines iGP91-06 and iGP91-07) or non-integrating CytoTune Sendai viruses (Invitrogen; Thermo Fisher Scientific; iPSC lines iGP91-09s, iGP91-10s, iGP91-14s), or by nucleofection according to manufacturer’s protocol (Amaxa human CD34 cell nucleofector kit; Lonza) of 1 million cells with 4 μg each of pEP4 E02S EN2L (Addgene plasmid 20922; Addgene), pEP4 E02S EM2K (Addgene plasmid 20923), and pEP4 E02S ET2K (Addgene plasmid 20927)³³ episomal plasmid vectors (iPSC line E5c5), followed by derivation and characterization as previously described.30, 31 The iNC-01 iPSC line,³⁰ derived from a healthy donor, was used as a disease-free normal control for some studies. For routine maintenance, iPSCs were cultured on plates coated with ESC-qualified Matrigel (Corning) in NutriStem XF/FF (Stemgent), E8 (Gibco; Thermo Fisher Scientific), or mTeSR1 (STEMCELL Technologies) medium. iPSCs were re-characterized for pluripotency and karyotype after targeted gene correction (Figure S9). Characterization included live stain with anti-human TRA-1-60 primary antibody (BD Biosciences) and Alexa-555 secondary antibody (Molecular Probes; Thermo Fisher Scientific) at 1:200 dilution for 1 hr and by alkaline phosphatase live stain (Molecular Probes; Thermo Fisher Scientific) at 1:500 dilution for 20 min. Two washes with media were performed post-staining before imaging for both TRA-1-60 and alkaline phosphatase. Pluripotency after gene correction was tested by in vitro embryoid body 3-germ layer differentiation as previously described.³⁰ Images were acquired using a Nikon Eclipse Ti-U microscope with DS-Qi1Mc monochrome camera and NIS-Elements BR software. Contrast and brightness adjustments were performed on whole images without other processing.

Design and Construction of TALENs and CRISPRs Targeting CYBB

Pairs of plasmids each expressing individual TALENs were constructed using the Golden Gate TALEN kit³⁴ (Addgene kit 1000000016), according to the kit protocol. CYBB5-L4 TALEN targeting the CYBB exon 5 sequence TGCCCGAGTCAATAATTCTGATCCT was constructed with the following repeat variable diresidues (RVDs): NN HD HD HD NN NI NN NG HD NI NI NG NI NI NG NG HD NG NN NI NG HD HD NG; CYBB5-R4 TALEN targeting TGCCTGTCTCCAAGTT was constructed with the following RVDs: NN HD HD NG NN NG HD NG HD HD NI NI NN NG NG. CYBB1-L3 TALEN targeting the CYBB exon 1 sequence TCAACCTCTGCCACCAT was constructed with the following RVDs: HD NI NI HD HD NG HD NG NN HD HD NI HD HD NI NG; CYBB1-R3 TALEN targeting TGGAGAGCCCCTCAT was constructed with the following RVDs: NN NN NI NN NI NN HD HD HD HD NG HD NI NG.

CRISPR/Cas9 expression constructs were generated by synthesizing oligos (Invitrogen; Thermo Fisher Scientific) containing the 20-bp target sequence with BsmBI-compatible sticky ends, annealing the complementary oligos, and cloning into BsmBI-digested lentiCRISPRv2 plasmid (Addgene plasmid 52961).³⁵ For CYBB5-g4 CRISPR targeting the CYBB exon 5 sequence CACTCTCTGAACTTGGAGAC, the following oligos were used: CYBB5-oligo1: 5′-CACCGCACTCTCTGAACTTGGAGAC-3′; CYBB5-oligo2: 5′-AAACGTCTCCAAGTTCAGAGAGTGC-3′. For the CYBB2-g3 CRISPR targeting the CYBB exon 2 sequence CCCGGTAATACCAGACAAAG, the following oligos were used: CYBB2-oligo1: 5′-CACCGCCCGGTAATACCAGACAAAG-3′; CYBB2-oligo2: 5′-AAACCTTTGTCTGGTATTACCGGGC-3′.

Construction of CYBB Donor Plasmids

The pCYBB5-pB-PGK-putk donor plasmid was constructed with 0.9- to 1-kb left and right homology arms (LHA and RHA) containing the genomic DNA sequences flanking the CYBB exon 5 TALEN and CRISPR target sites together with the piggyBac transposon-flanked phosphoglycerate kinase (PGK) promoter-driven puroΔtk drug-selection gene cassette from pMCSAATPB:puroΔtk,³⁶ which allows for puromycin selection for donor insertion and ganciclovir selection for excision of the piggyBac cassette. Silent mutations (changing GCACTCTCT to GCGTTAAGT) were also introduced in the donor vector to retain the normal amino acid sequence while accommodating the TTAA footprint left by piggyBac transposon excision.

The pCYBB-E1–13opt donor plasmid was constructed with a 1.3-kb LHA containing the genomic DNA sequence immediately upstream of the CYBB start codon followed by a minigene containing a codon-optimized CYBB cDNA⁵ and poly(A) from herpes simplex virus thymidine kinase, then a loxP-flanked 1.7-kb puromycin selection cassette containing a cytomegalovirus (CMV) promoter, puromycin resistance gene cDNA, and bovine growth hormone poly(A), followed by a 1-kb RHA containing the genomic DNA sequence for CYBB exon 1 and part of intron 1. The pCYBB-E1–13cDNA donor plasmid was constructed from pCYBB1–13opt by replacing the optimized CYBB cDNA with the 1.7-kb un-optimized complete CYBB cDNA.

The donor plasmid pCYBB-E2–13 was constructed with an LHA containing the genomic sequence for the last 800 bp of CYBB intron 1, followed by a minigene containing a codon-optimized CYBB cDNA for exons 2–13 and rabbit β-globin pA, then piggyBac transposon 5′ and 3′ repeats, and a RHA containing the first 800 bp of CYBB intron 2. The PGK promoter-puroΔtk cassette was cloned from pCYBB5-pB-PGK-putk into the resulting plasmid between the piggyBac 5′ and 3′ repeats.

Generation of Gene-Corrected iPSC Lines

For gene transfer, 3–5 million iPSCs were detached with Accutase (STEMCELL Technologies), centrifuged at 300 × g for 5 min, and resuspended in 100 μL of Amaxa mouse ES nucleofection solution (Lonza) with 5 μg of donor plasmid DNA and 5 μg each of TALEN plasmids or CRISPR/Cas9 plasmid. Cells were then nucleofected in an Amaxa Nucleofector device using program A-23 and plated onto Matrigel-coated plates with puromycin-resistant DR4 mouse embryonic fibroblasts (MEFs) (MTI-GlobalStem) in NutriStem XF/FF or complete iPSC medium for MEF co-culture.30, 37 Two to 3 days later, cells were selected with 0.25 to 0.5 μg/mL puromycin (Sigma-Aldrich) for up to 7 days. Puromycin-resistant iPSC clones were expanded thereafter on Matrigel-coated plates in E8 or mTeSR1 medium. For Cre excision of loxP-flanked puromycin resistance gene cassette, iPSCs were transfected in one well of a six-well plate with 6 μL of lipofectamine-LTX (Invitrogen; Thermo Fisher Scientific), and 3 μg of EF1a-ZeoR-T2A-CreGFP expression plasmid (containing a Cre-GFP fusion gene from Addgene plasmid 13776³⁸ and a zeocin drug resistance gene, with a T2A element for co-expression) with 3 μL of lipofectamine PLUS reagent, according to manufacturer’s protocol. Cre-transfected cells were selected with 2–4 μg of zeocin (Gibco; Thermo Fisher Scientific) for 2–3 days starting 1 day after transfection. For excision of piggyBac-flanked puroΔtk gene cassette, approximately 2 × 10⁶ iPSCs were transfected with 2 μg of pCMV-PBx excision-only piggyBac transposase expression plasmid (Transposagen Biopharmaceuticals) by nucleofection as above, and cells were selected with 2–4 μM ganciclovir (Sigma-Aldrich) for up to 7 days starting 2 days after transfection. iPSC clones surviving zeocin or ganciclovir selection were expanded as above.

Validation of Targeted Insertion and Excision

Genomic DNA from puromycin-resistant iPSC clones was screened for donor insertion by PCR. Primers used for PCR analysis of targeted or random donor insertion and for piggyBac or loxP excision of the puromycin resistance cassette from the insert are listed in Table S4. Q5 high-fidelity DNA polymerase (New England Biolabs) was used to generate PCR products for sequencing reactions to confirm targeted insertion of the donor and excision of the puromycin resistance cassette. DNA sequencing was performed commercially (Macrogen). Clones screening positive for a targeted insertion with no random insertions were considered to be correctly targeted and were used in subsequent expression and differentiation assays. Gel images were obtained using a ChemiDoc XRS+ imaging system with Image Lab 5.1 software (Bio-Rad).

Analysis of TALEN and CRISPR Off-Target Sites

Genomic DNA from corrected clones that exhibited correct targeted insertion without random insertion was PCR amplified and sequenced for evidence of mutations at predicted off-target sites. The top ten off-target sites for TALEN pairs were predicted using the Paired Target Finder of TAL Effector Nucleotide Targeter 2.0 (TALE-NT; https://tale-nt.cac.cornell.edu/),34, 39 which were sorted by average score for both TALENs. The top ten off-target sites for CRISPRs were predicted using the Feng Zhang laboratory’s CRISPR design tool (http://crispr.mit.edu/). Primers used for PCR screening and sequencing of off-target site mutations for CYBB5-L4/R4 TALENs, CYBB5-g4 CRISPR, and CYBB2-g3 CRISPR are listed in Table S5.

Differentiation and Characterization of Granulocytes from iPSCs

Granulocyte differentiation of iPSCs was performed as previously described.⁶ Granulocyte morphology was assessed by Giemsa stain of cell cytospins.³⁷ Color images of cells were acquired using an EVOS XL Core system (Thermo Fisher Scientific) or Zeiss ICM405 inverted microscope (Zeiss) with 40× objective and a Kodak DC290 digital camera (Eastman Kodak) and Kodak MSD-290 Acquire software in Adobe Photoshop 6 (Adobe Systems). Whole-image adjustments of brightness and contrast were performed using Adobe Photoshop without other processing. For flow cytometry analysis of gp91^phox expression, cells were fixed with 2% paraformaldehyde, permeabilized with 0.1% saponin, and stained at a 1:100 dilution with unconjugated clone 7D5 anti-human flavocytochrome b558 antibody⁴⁰ for 20–30 min followed by a 1:100 dilution of FITC-conjugated secondary antibody for 15–20 min. Some samples were co-stained with APC-conjugated antibody to CD13 myeloid surface marker (BD Biosciences) to confirm myeloid specificity of gp91^phox expression. DHR flow cytometry assays of ROS production were performed as previously described.5, 37 As a normal neutrophil control, in some studies, 400 μL of peripheral blood from healthy donors was used after red blood cell lysis in 4 mL of ACK lysis solution (Quality Biological) for 5 min at 37°C and washing with 1× HBSS. Flow cytometry was performed using a FACSCalibur (BD Biosciences), and analysis was performed using FlowJo software (Tree Star).

Author Contributions

C.L.S. and H.L.M. designed experiments and wrote and edited the manuscript. C.L.S., J.Z., U.C., R.K.M., S.S.D.R., and H.L.M. designed donor constructs and targeting strategies. C.L.S., J.Z., U.C., R.K.M, A.L., A.B., S.B., and J.-W.K. performed research and analyzed data.