Gene Editing in Hematopoietic Progenitor Stem Cells

The genome editing using engineered nuclease has strategically transformed the idea of gene therapy for monogenic diseases including in hematopoietic stem and progenitor cells (HSPCs) (Osborn et al., 2016; Yu et al., 2016). The genome editing technology enables to create a site specific double-strand break (DSB) by the engineered nucleases that programmable triggering the cell’s endogenous repair machinery to edit the genome in a site-specific manner via the non-homology end joining repair (NHEJ) and the homology directed repair (HDR) mechanisms(Branzei and Foiani, 2008). The approach allows the precise alteration of the disease-causing alleles at the specific locus making it a permanent event that maintains the phenotypical gene expression under the control of endogenous regulatory elements.

Over the past decade, three major classes of engineered nucleases have been used for genome editing, including zinc-finger nucleases (ZFNs) (Kim et al., 1996; Urnov et al., 2010), transcription activator-like effector nucleases (TALENs) (Li et al., 2011; Miller et al., 2011) and CRISPR–Cas9 (clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated (Cas) protein 9) (Hsu et al., 2014; Sander and Joung, 2014; Tsai and Joung, 2016; Wiedenheft et al., 2012). ZFNs and TALENs are fusions between arrays of ZF or TALE DNA-binding domains and the dimerization-dependent FokI nuclease domain. The both of ZFN and TELENs nucleases exclusively rely on protein–DNA interactions to mediate site-specific recognition of genomic DNA sequences which requires complex protein engineering for each new targets (Kim and Kim, 2014). By contrast, CRISPR–Cas9 nuclease is a RNA-guided endonuclease. Through the guidance of a 23 nucleotides linked to CRISPR-domain RNA (gRNA), CRISPR-Cas9 finds the complementary protospacer DNA target in a genome where it cuts the double stranded DNA precisely 3 base pairs upstream of a PAM (Protospacer Adjacent Motif). The broken DNA ends generated by those nucleases are repaired either by NHEJ resulting in small insertion/deletions (indels) to disrupt target allele, or by HDR to precisely replace desired nucleotides required with delivery homologous DNA template. Compared to ZFNs and TALEN, the CRISPR/Cas9 system has rapidly become the most promising genome editing tool with demonstrated advantages including simplicity, easy programming, low cost and potential multiplexed editing (Bannikov and Lavrov, 2017; Brunetti et al., 2018; Salsman and Dellaire, 2017; Tsai and Joung, 2016) (Minkenberg et al., 2017). Despite of the genome editing holds tremendous promise for the developing novel gene therapy, the technique has been shown to be more refractory in HSPCs than any other cell types due to their quiescent status associated with low activity of the HDR machinery, and prone to DSB induced toxicity. However since first publication of using the ZFNs editing on human CD34+ cell (Genovese et al., 2014), the substantial developments have been made in last few years to circumvent the problems.

Optimization of gene editing efficiency in HSPCs

In vitro expansion of HSPCs

Since all nucleases targeted gene editing occurs through cell cycle progress, the increased stimulation HSPCs ex vivo can make them more permissive to editing components. However, increasing stimulation can also promote cell differentiation. To circumvent this, the compounds that agonist HSPC self-renew while maintain their primitive phenotypes have been discovered and applied in the culture (Boitano et al., 2010; Fares et al., 2014; Goessling et al., 2011). Using the compounds in HSPCs culture, researchers have achieved significantly increased percentage of edited HSPCs in vitro and also increased human cell engraftment in vivo (Charlesworth et al., 2018; Genovese et al., 2014). In a recent published study, Psatha et al. have described 5 days HSPC culture condition, in which StemRegenin 1(SR1) was used with a small molecule Ly2228820 (SL), the p38?MAPK14 inhibitor (Psatha et al., 2017). Using this culture condition, they have successfully expanded highly engraftable CD34+/CD38?/CD90+ primitive HSPC cells. They then tested if using SR1 and SL condition can also expand edited HSPCs effectively. For this, they cultured edited HSPCs for additional 5 days after the editing, and found that the edited CD34+/CD38?/CD90+ primitive HSPCs can be effectively expanded in vitro without any loss of editing efficiency. Moreover, the expansion of edited cells gave rise to a more than 2-fold higher engraftment compared to their unexpanded counterparts, showing the same editing rate (Psatha et al., 2018). The study highlights a possible way to obtain sufficient engraftable HSPCs by expanding edited cells effectively ex vivo in presence of SR1 and SL. However, this study was conducted using the NHEJ directed gene editing strategy for disrupting the genomic locus. It would be important to know if the presence of SR1 and SL in culture can also increase the HDR directed gene editing. And also convincing evidence on long-term in vivo engraftment from significantly expanded HSPCs is needed to ensure no oncogenic burden associated with ex vitro expansion.

Delivering the editing components

In clinical application setting, the approach for delivering nucleases or other components into HSPCs should be transient to avoid the cytotoxicity engendered by prolonged endonucleases activity and immune responses. Therefore, a “hit-and-run” approach is used, whereby the nuclease complex is transient expression. The mostly used method for delivering DNA or RNA encoding engineered nucleases is via nuclear transfection. The transfection of plasmid DNA encoding the nucleases to HSPCs has been used with success on editing targeted loci (Holt et al., 2010; Mandal et al., 2014). However the main concern from this approach is its potential random integration into the genome which could lead to cytotoxicity in HSPCs and their progenies. And DNA related cytotoxicity, such as cyclic GMP-AMP synthase induced pathway (Sun et al., 2013), could lead to high toxic to primitive HSPCs. Therefore, the transfection of mRNA encoding nucleases synthesized in vitro has become an optimal alternative approach (Liang et al., 2015; Wang et al., 2015). It has emerged from recent studies that the mRNA transfection approach indeed has provided an increased efficiency in genome editing in HSPCs (De Ravin et al., 2016; Kuo et al., 2018; Schiroli et al., 2017). In addition, Cas9 can be delivered as the protein or as the precomplexed ribonucleoproteins (RNPs) by mixing gRNAs with Cas9 protein (Dever et al., 2016; Liang et al., 2015). The approach serves in protecting gRNAs from degradation, and reducing cytotoxicity caused by naked RNA stimulated innate immunity. The improved editing efficiency based on such approach has been achieved in targeting HSPCs shown in recent studies (Bak et al., 2017; Kuo et al., 2018; Schiroli et al., 2017; Vakulskas et al., 2018).

Apart from above components, a safe and efficient delivering DNA donor template into edited cells is crucial for achieving HDR process. Several donor template platforms have been used. Single-stranded DNA oligonucleotide (ssODN) donor has been shown as a simple and effective approach in genome editing for correction of single-nucleotide mutation in HSPCs, such as for Sickle cell disease (SCD) (DeWitt et al., 2016).  Integration defective lentiviral vector (IDLV), that allow incorporating large DNA template, has been used in the ZFN genome editing to target the IL2RG mutations and the adenosine deaminase (ADA) gene (Genovese et al., 2014; Joglekar et al., 2013). However, those early studies showed a very limited gene targeting efficiency in HSPCs, suggesting that IDLV could be more cytotoxicity to HSPCs. The efficiency of IDLV in targeted integration in HSPCs can be significantly improved by using cyclosporine H, which is shown in a very recent study (Petrillo et al., 2018). Certain adenoviral serotypes (Ad5) can transduce human HPSCs and deliver large transgene cassettes (Li et al., 2013) (Saydaminova et al., 2015). However the concern that residual of adenovector particles could be highly immunogenic which may prevent its potential use in clinical application therapy. Recombinant adeno-associated viral vectors (rAAVs) have been shown to naturally mediate HR in mammalian cells without stimulating DSB (Barzel et al., 2015; Mingozzi and High, 2011; Moser and Hirsch, 2016). Hence, rAAV vectors are emerged as ideal delivery approach due to their wide range of tropism, high transduction rate and very low immune response. In particular, the rAAV6 vector has been shown to provide more efficient and robust genome-editing in HSPCs than other delivery vectors shown in recent therapeutic potential application studies (De Ravin et al., 2017; De Ravin et al., 2016; Kuo et al., 2018; Moser and Hirsch, 2016; Schiroli et al., 2017). However, relative small packaging capacity in rAAV6 has limited its use for delivering cassata larger more than 4.5 kb including the both homology arms. To improve the packaging capacity, Bak and Proteus (Bak and Porteus, 2017) have developed a dual-AAV6 donor vector system that enables site-specific integration of large transgene cassette up to 6.5 kb into primary T cells and HSPCs with long-term repopulation capacity. Overall, the conditions for delivery the components used in gene editing should always be optimised for each targeted gene to achieve most efficient targeting and minimum cytotoxicity. A comprehensive detailed protocol using CRISPR/Cas9 with rAAV6 as templet vehicle for HDR-targeted editing in HSPCs has been published by Bak and Daniel recently (Bak et al., 2018), which could be also served as a guide for implement gene editing technique for other nucleases

Improve the HDR

Unlike NHEJ pathway which occurs throughout the cell cycle, the HDR event is restricted in S/G2 phases of cell cycle which makes the HDR process much less efficient than NHEJ (Gutschner et al., 2016; Heyer et al., 2010). Therefore, inhibiting nuclease activity at G1/M phases and resting cells at S/C2 phases may improve HDR efficiency. The concept has been experimentally tested by Gutschner and colleagues (Gutschner et al., 2016). In which, the hGemCas9 system is generated by incorporating the human geminin domain which allows the nuclease to be ubiquitinated and degraded by APC/Cdh1 complex in G1 and late M phase, therefore leading to increased hGemCas9 activity in S/G2 phases. Using this cell-cycle-tailored hGemCas9 system, Gutschner et al have achieved an increased rate of HDR up to 1.87 fold compared to wild-type Cas9 in cell lines. A further development based on this approach was published recently by Lomova et al. (Lomova et al., 2018). In their study the hGemCas9 was used in combination with a cell synchronization compound RO-3306 which functioning in transiently arresting cells at S/G2 phase via inhibiting CDK1(Vassilev et al., 2006). It was shown from Lomova’s study that the ratio of HDR/NHEJ was increased to four-fold on human CD34+ cells compared to the controls in vitro, and a significant improvement of edited HSPCs in immune-deficient mice. The improved HDR has also been achieved by directly inhibiting the NHEJ pathway through targeting DNA ligase IV, a key enzyme in the NHEJ pathway, using the inhibitor Scr7 (Hu et al., 2018; Maruyama et al., 2015). Although high increased efficiency in HDR has been achieved in human cell lines and cancer cells, so far, there has been no published data of using Scr7 on human HSPCs.

The assessment off-target sites

Although ideal engineered nucleases would have singular genome-wide specificity, unintended off-targets can occur, particularly at loci with homologous to the intended targeting site. Several the off-target detection methods have been used in HSPCs gene editing studies. An early developed assay is based on using the silico prediction off-targets sites that have degree of similarity to the on-targets, and then followed by targeted sequencing (Fine et al., 2014) (Hsu et al., 2013). This initially developed method is still mostly used in the HSPCs editing studies as it is more practicable assay. However the fundamental limitation with this approach is it is not designed to identify off-target sites in an unbiased manner as the sites that not fit the computational criteria will not be discovered,

To achieve unbiased off-target detection, the cell based genome-wide assays have been developed. On of such assay used in HSPCs editing studies is Integrase-defective lentiviral vector (IDLV) capture assay, which was designed to capture IDLV into sites of nuclease-induced DSBs. Then clustered sites of integrations are recovered by linear amplification-mediated PCR (LAM-PCR) and mapped using high-throughput sequencing (Gabriel et al., 2011). Although the IDLV capture can directly identify DSBs in living cells, it is relatively insensitive, owing to its low absolute integration efficiencies that require positive selection to overcome (Gabriel et al., 2011). And the assay may have high background due to event of random integration IDLVs into cellular genomes even in the absence of nuclease-induced DSBs (Gabriel et al., 2011).

Whole genome sequencing (WGS) has been proposed as an unbiased method for defining engineered nuclease specificity. Although this method is useful for the analysis of single-cell clones (Veres et al., 2014), it lacks sensitivity, particularly for those low frequencies off-target in a population cells (Tsai and Joung, 2014). With existing deep sequencing technology, it is impractical to perform WGS on millions of cellular genomes, and it is inadequate to confirm the off-target sites at < 0.1% in a population cells (Tsai and Joung, 2016). In considering the limitations in those off-target assays, it would be necessary to use combined approaches shown as in Kuo’ study (Kuo et al., 2018) to ensure confidence in safety of therapeutic strategies by gene editing in HSPCs.

Therapeutic potential of HSPC gene editing

Non-homologous end joining-based strategy

NHEJ DNA repair pathway triggered by engineered nucleases is the active random repair process, leading to the alteration of nucleotide sequencing at the specific site via in-frame deletions, insertions. Sine it is not involved in harnessing the HDR machinery, it has become a viable genome editing option for correcting gene mutations. Two HSPC targeted loci, chemokine coreceptor 5 (CCR5) and BCL11A, have received the most early attention as their potential therapeutic benefits via NHEJ process. The concept of editing CCR5 was intrigued by the report that the transplantation of a donor HSCs with a naturally occurring CCR5 mutation confers a loss of detectable HIV-1 RNA and proviral DNA in a HIV patient (Hutter et al., 2009). Holt et al. first published the report of the successful disrupting CCR5 using the ZFNs (Holt et al., 2010). In their study, NSG mice transplanted with ZFN-modified HSPCs underwent rapid selection for CCR5(-/-) cells when challenged with CCR5-tropic HIV-1, showed significantly lower HIV-1 level compared to the controls. Several studies publishes later have also demonstrated the feasibility of CCR5 disruption in HSPCs that lead to resistance to HIV infection in vivo model (DiGiusto et al., 2016; Li et al., 2013; Xu et al., 2017). Among them, DiGiusto et al. conducted a preclinical study to assess efficacy and safety of the ZFN-based CCR5 disruption in HSPCs on the clinical-scale delivering CCR5-specific ZFN-mRNA to normal adult HSPCs. In which, they demonstrated effective biallelic CCR5 disruption of 40-60% in liquid culture cells, and in up to 72.9% of modified colony forming units from edited HSPCs. The edited HSPCs preserved long-term multiple lineage potential in vivo with no demonstrated potential for tumorigenesis or leukemagenesis (DiGiusto et al., 2016). Based on this, further safety and feasibility studies are ongoing in subjects infected with HIV-1 (NCT02500849@clinicaltrials.gov).

Targeting HSPCs genomic locus BCL11A via NHEJ gene editing has been developed for potential treatment of the ?-hemoglobinopathies, which are inherited monogenic blood disorders due to the mutations in ?-globin gene causing either Thalassemis (abnormal haemoglobin production) or sickle cell disease (SCD) (abnormal haemoglobin tetramer) (Steinberg, 1999). The observed fact of that the severity of both conditions can be ameliorated by the induction of Fetal haemoglobin (HbF) (Collins et al., 1995) led to discover the BCL11A transcription factor as a repressor for HbF (Bauer and Orkin, 2015), and BCL11A erythroid-specific enhancer, GATAA in association with fetal-to-adult haemoglobin switching (Canver et al., 2015), which could be targeted for inducing HbF in HSPCs for potential treatment of those conditions. To this end, Bjurstom et al. conducted the genome editing strategy to disrupt the BCL11A exon2 in HSPCs using the engineered nucleases ZFNs, TALENs or CRISPR-Cas9 (Bjurstrom et al., 2016). It was shown in their study that the ZFNs gave rise to more allelic disruption in the targeted locus which is associated with increased levels of HbF in erythroid cells derived from nuclease-treated CD34+ cells in vitro. However, a low level of disruption in the BCL11A gene in bone marrow (4%) was observed after engraftment into NSG mice. Using the ZFNs approach Chang et al. performed study to compare targeted disruption of the BCL11A, either within exon 2 or at the GATAA motif (Chang et al., 2017). It was shown from their study that the allelic disruption of GATAA not only give rise to robust long-term engraftment leading to elevated level of HbF expression in erythroid cells, but also prevent the adverse effect of erythroid enucleation seen in the BCL11A exon2 ablation. Using same strategy, a comprehensive preclinical study has been carried out in HSPCs from adult donors and two patients with ?-Thalassemia Major (Psatha et al., 2018). The modification of GATAA motif in mobilized CD34+ cells from ?-thalassemia patients resulted in a readily detectable increased ?-globin with a preferential increase in G-gamma, leading to an improved phenotype that likely to give a survival advantage for maturing erythroid cells. A phase 1/2 clinical trial for correcting the ?-thalassemia phenotype by genome editing is currently being evaluated by the same group.

Homologous recombination based strategy

In larger majority genetic blood diseases, the homologous directed repair strategy is required for correcting genotype, with delivering exogenous DNA template. The process is much more challenging than NHEJ-based gene editing due to its low efficiency, particular in targeting primitive HSPCs. However, the promising progress in targeted integration in HSPCs for some PIDs has been made in recent years.

Interleukin-2 receptor common gamma-chain (IL2RG)

The first attempt using the ZFNs for gene knockin in HSPCs was demonstrated by Genovese et al. (Genovese et al., 2014). In this study, two genomic loci, AAVS1 “safe harbour” or a mutational hotspot of IL2RG were targeted with a GFP cassette delivered with IDLV vector. Although there was 24-26% indels found in the ZFN target sites, only 5% GFP+ colonies were found in colony-forming cells (CFU) assay. At 8 weeks after transplantation edited CD34+ cells into NSG mice, the frequency of 2% GFP+ cells were found among primitive and committed progenitors in the BM. To improve gene targeting efficiency, Genovese et al. tailored the culture condition by extending cell activation time making them more permissive for the editing molecules, and by adding the compounds into the culture to preserve the stemness in primitive HSPCs (Genovese et al., 2014). The modified protocol indeed gave rise to significantly increased GFP+ cells (?2-fold) in primitive cell population in vitro and also in vivo in long-term engrafted HSPCs. Using improved the ZFNs protocol Genovese et al. performed the IL2RG gene correction in CD34+ cells derived from SCID-X1 patient with delivering IDLV vector consisting of the exons 5-8 IL2RG cDNA and a PGK-GFP cassette flanked by homologous sequences. In which, they found 3% GFP+ cells in primitive HSCs and up to 11% GFP+ in committed progenitors in liquid culture. The CFU assay yield 3 GFP+ colonies out of 100 scored, with a myeloid progeny colony showed reconstituted normal IL2RG protein expression. The data from this study highlighted the problem with targeting primitive HSCs for homologous recombination. A recent development in targeting integration of IL2RG has been demonstrated by same group (Schiroli et al., 2017). In order to establish therapeutic potential of a gene correction strategy for the treatment of SCID-X1, a humanise SCID-X1 mouse model was used to evaluate efficacy and safety of the edited HSPCs in a preclinical setting. To improve editing efficiency, they made the modification on the ZFN mRNA by incorporating the base analogs to prevent recognition by cellular sensors that associated with the activation of the interferon-responsive genes by exogenous RNAs. This modification results in a significant reduced cytotoxicity caused by in vitro electroporation of the ZFN mRNA, leading to high HDR (25%) in CD34+ cells derived from a SCID-X1 patient. By changing to use AAV6 as donor DNA vehicle following the ZFN mRNA electroporation, they achieved up to fivefold higher HDR-mediated gene editing in the most primitive CD34+ CD133+ CD90+ cells over the IDLV vehicle approach. It was also demonstrated in this study that optimised clinical relevant protocol is transferable to the clinic scale, showing reproducible editing efficiency even in a large scale 2.5x107 HSPCs. More importantly, the edited cells preserved long-term engraftments in NSG mice, showing an average 12% HDR in HSPCs at 23 weeks end point, which exceeded the threshold (10%) of functional HSPCs required for fully reconstitute immune function at a standard transplant dose established in the their study (Schiroli et al., 2017). The off-target assay did not detect significant amounts of modification above the threshold of sensitivity in any of the off-target sites identified previously by genome-wide screening for the ZFN set (Gabriel et al., 2011). Based on these data, it would be interesting to see if the optimised protocol could lead to adequate editing efficiency in HSPCs derived from the SCID-X1 patient, which could paves the way to translation HSPCs gene editing into the therapy.

X-linked chronic granulomatous disease (X-CGD)

Two recent studies published by De Ravin et al. presented the promising results on the targeted integration of CYBB gene encoding gp91phox for the treatment of X-CGD (De Ravin et al., 2017; De Ravin et al., 2016). Their initial study (De Ravin et al., 2016) was based on the ZFNs targeted integration of transgene into a genomic “safe habour”AASV1 with the aim to overcome insertional mutagenesis by the viral vector gene therapy, where 3 X-CGD patients underwent the gene therapy developed myelodysplasia due to the integration at MDS-EV11 locus (Stein et al., 2010). De Ravin et al. carried out extensive experiments firstly to explore the optimised conditions in clinical relevant approaches for delivery of the ZFNs, and AAV6 delivery of donor construct containing promoter-less Venus marker cDNA into the intron 1of the PPP1R12C gene at AAVS1 locus. The results from their study showed up to 58% Venus-positive HSCs in vitro and 6–16% human cell marking were observed following engraftment into NSG mice. Using their optimised approach, they then targeted HSPCs derived from X-CGD/gp91phox patients with donor constructs containing either a promoter less gp91phox (2A-2A-gp91), or gp91phox driven by a synthetic MND promoter (MND-91). Although the both approaches showed a similar targeted integration efficiencies (~15% gp91phox expression), a robust functional correction through MND promoter, rather than the endogenous PPP1R12C promoter was obtained with significant high MFI of gp91expression and DHR oxidase activity in edited HSPCs in vitro. At 8 weeks following transplantation of edited HSPCs into NSG mice, the MND-91 and 2A-2A-gp91 corrected HSPCs grafted average 3.7±4.2% and 10.7±4.2% of human CD45+ cells respectively from bone marrow gp91expressing cells. Since the gene therapy corrected cells in X-CGD patients do not entail a selective advantage, the question is if the level of reconstituted gp91expressing cells achieved in this study would be sufficient for the disease phenotype correction. Nerveless, the data presented in the study has provided the first promising alternative approach in correction of X-CGD. However, long-term efficiency in vivo still remain to be established, and the safety issue of disrupting PPP1R12C gene encoding for a phosphatase in stem cells also need to be determined.

In a later study led by the same group (De Ravin et al., 2017), De Ravin et al. have achieved the targeted correction of the point mutation C676T X-CGD using CRISPR/Cas9 and delivering single strand oligo nucleotide (ssODN). The C676T mutation, accounted for 6% of X-CGD patients, occurs at the exon 7 of CYBB gene resulting in a premature stop codon and an inactive gp91phox protein. Following experiments to optimise the targeting CYBB 676 locus in normal CD34+ cells, they achieved level of HDR editing efficiency even within primitive (CD34+CD133+CD90+) HSPCs at ~30%, which is high than any previously reported. In CD34+ cells derived from CYBB 676 patient, they achieved targeted repair of >20% of HSPCs and restored gp91expression to 31% in myeloid cells differentiated from edited HSPCs, which resulted in restoration of the function of NADPH oxidase activity and superoxide radical production. Analysing of transplantation of gene-repaired X-CGD HSPCs into NSG mice at 8 and 20 weeks, they demonstrated not only improved stable human engraftment and corrected CYBB alleles, but also the production of functional mature human myeloid and lymphoid cells for up to 20 weeks. The off-target sequencing analysis on computationally predicted off- target sites in edited CD34+ cells from the patient revealed one single indel (>3 bp) at the RP11-454H19.2 gene at a high read depth 1,200,000x, but not at 10,000 read depth. However, one single indel observed in the uncorrected healthy control CD34+ HSPCs, indicating that this could be due to amplification/sequencing errors at high level of coverage. Whole-exome sequencing at 800× coverage of corrected patient CD34+ HSPCs also failed to detect any off-targets. Using same approach, De Ravin et al. have tried to correct a second X-CGD patient with CYBB 676 mutation (De Ravin et al., 2017). Although the gene repair efficiency was achieved in a similar level to the patient 1 in vitro, a less than 50% of the gene repair rate was observed after transplantation into NSG mice. This has highlighted the necessity of careful validation of editing condition at every level to achieve a consistent outcome. Nerveless, this study presented a viable approach in correction of a missense mutation in HSPCs by targeted integration that restore gene function under the control of the genes endogenous promoter.

X-linked hyper-IgM syndrome (XHIM)

XHIM is a primary immunodeficiency due to mutations in CD40 ligand gene (CD40L) expressed on the activated T cells. The mutated CD40L fail to bind CD40 on B cells which affect immunoglobulin class switch recombination that represented by the absence of IgG, IgA, IgE with normal to elevated IgM. XHIM patients are susceptible to bacterial infection, with development of autoimmunity and malignancies in some X-HIM individuals (Hayward et al., 1997; Levy et al., 1997). XHIM can be treated by allogenic HSCT, but has been associated with some sever site effects. Although the experimental gene therapy using viral vector in XHIM mouse model showed the correction of immune defect, the mice developed abnormal lymphoproliferation due to unregulated expression of the gene from ectopic genomic loci (Brown et al., 1998; Sacco et al., 2000). Therefore, using gene editing tools in targeted integration of XHIM gene under control of its endogenous promoter has become an optimal alternative approach for treating the disease. Using the TALEN as targeted gene editing approach Hubbard et al. have first demonstrated the feasibility in restoration of normal expression of CD40L and rescued IgG class switching in XHIM patient T cells (Hubbard et al., 2016). A later study by Kuo et al. developed the both TALEN and CRISPR/Cas9 platforms to achieve site-specific editing of a human CD40L cDNA, at the 5’UTR of the gene allowing bypassing all known disease-causing mutations in XHIM (Kuo et al., 2018). The both approaches were tested in T cells derived from XHIM patient. Although the TALEN approach resulted in CD40L expression at the baseline in unstimulated cells, an up-regulated CD40L expression to >20% was detected upon anti-hCD3/anti-hCD28 immune stimulation which is comparable to stimulated T cells from healthy donors. The corrected XHIM T cells demonstrated a normal receptor-binding activity to recombinant chimeric CD40-muIg. The data highlighted that a proportionally small number of gene-corrected T cells in XHIM may be sufficient to allow enough class-switching to ameliorate the disease. In CRISPR/Cas9 treated XHIM T cells, high rates of targeted gene integration was attained with restore physiologically-regulated CD40L expression and function. In targeting CD34+ cells from healthy donor, Kuo et al. have shown that both platforms gave rise to a similar level of allelic disruption rate in samples from 8 biological replicate, 4 PBSC donors (29.1 ± 7.8% with TALEN, average 33% with CRISP/Cas9). A relative high targeted gene integration rate was observed in CRISP/Cas9 treated cells, particular when gRNA and Cas9 protein delivered as RNP (to 20.8 ± 6.6%). By adding the adenovirus helper protein that co-introduced as mRNA during electroporation with TALENs or CRISPRs, a 2-fold enhanced gene modification was achieved. However, this augment effect was not observed in engrafted NSG mice in vivo. Following transplantation of edited cells into NSG mice at 12-20 weeks, the targeted gene integration was detected in the bone marrow from 80% of mice, with integration rates ranged from 0.3% to 22%, a mean of 4.4% across all treatment groups. The analysing of thymus from engrafted mice showed 60% mice had thymic reconstitution,

With frequency of engraftment trending higher in those analysed at 5 months compared to 3 months post-transplant. The off-target activity was not detected based on silico predicted off-target sites for both TALENs and CRISPR in K562 edited cells. However, using IDLV capture approach in TALENs edited K562 cell, three off-target loci (OT1, OT2 and OT3) were observed. High-throughput sequencing of off-target sites in HSPCs and K562 cell treated with TALENs mRNA demonstrated statistically significant gene disruption at OT1 in HSPCs, and OT2 in both cells. However, there was no off-target site identified in CRISPR treaded cells using another cell based assay GUIDE-seq, which was designed recently with a high sensitivity for detecting the off-target sites mutagenized by Cas9-gRNA (Tsai et al., 2015). Taking together, CRISPR approach showed some advantages over TALENs in targeting integration of XHIM gene. Overall, this study paves an important step toward to developing a curative therapy for XHIM through site-specific gene correction.

The major hurdles in HSPCs gene editing

Despite the genome editing holds tremendous promise for the developing novel gene therapy, HSPCs targeted editing is still in its infancy, and many issues regarding this new technology are remained to be addressed before translating it into safe clinical application.

One major hurdle facing in HSPCs targeted editing is low efficiency, particularly in vivo following transplantation of edited cells, where the engrafted cells and frequency of edited cells decline significantly within 8 to 12 weeks and continuously decline in prolonged period. This suggests the “real” long-term HSPCs either have failed to undergo genome editing due to their quiescence and more resistance to homologous recombination, or they have been damage by DSBs due to exposure to nuclease and lost their self-renew property underwent apoptosis.

Cite this page