Zinc finger nucleases for targeted mutagenesis and repair of the sickle-cell disease mutation: An in-silico study

Sickle cell diseases (SCD) or sickle cell anemia (SCA) is a hemoglobinopathy that is mostly common among persons of African descent [1]. SCD arises from a single, point-mutation (base-substitution of Adenine with Thymine in the sixth codon: CAG → CTG) of the gene coding for the beta chain of the Hemoglobin molecule [2]. The phenotypic consequence of this substitution is a replacement of the amino acid glutamic acid (E) with valine (V) [1, 2]. Homozygous expression of this mutant globin genotype (SS) causes SCD, while the heterozygous genotype (AS) is termed the “sickle cell trait” [1, 2]. Unlike the case observed in most normal adult humans where the commonest hemoglobin type (hemoglobin A or Hb A) is a tetramer (which contains 4 subunit proteins- α ₂ β ₂ ^A that are non-covalently bound together), patients with SCD have an adult hemoglobin type with two mutant β subunits (called β ^S) called hemoglobin S (or simply, Hb S) [3, 4]. Hb S has a high predilection to crystallize under conditions of low oxygen-pressure such as may occur following physical or pathological-stress. Specifically, formation of intracellular S crystals causes polymerization of red blood cells, reduced oxygen uptake and or carriage, a as well as clogging of small blood vessels [5, 6]. Overall, although the clinical syndrome of SCD is diversely-wide, its hallmark is a devastating group of symptoms and signs that are collectively known as a ‘sickle cell crisis’ [7, 8]. About 200,000 new born babies within Africa recessively inherent the double autosomal-sickle cell genotype each year— a figure that constitutes approximately 66.6% of the children born with haemoglobinopathies worldwide [9]. Previously studies have shown that persons with heterozygosity for β ^S (Hb SA) or the sickle cell trait are protected against infection by malaria causing protozoa [10]. This, together with findings of an equally high-incidence or common-distribution of the sickle cell trait within the malaria-belt, has led to the proposition that this trait emerged as an evolutionary adaptation of the human-host to infection with plasmodia [11]. The medical management of SCD remains an area of particular challenge [12]. Specifically, despite advances made towards curing SCD through bone-marrow transplantation [13, 14], the resource-intensive nature of this approach has made it impossible for the most affected populations of Africa to access. Thus, care for affected individuals (homozygous SS) here still mostly revolves around syndromic management, with or without agents that increase fetal hemoglobin (Hb F) [15]. Improvements are obviously sought here.

By virtue of its monogenic, point-mutant origin, SCD has attracted several attempts for gene therapy. For instance, as early as 1991, Shesely ED, et al.[16] described a technique for the correction of a human β ^S globin gene to the normal β ^A allele by homologous recombination in the mouse-human hybrid cell line BSM using an oligomer-restriction enzyme construct. In 1998, Lans N, et al.[17, 18] reported ribozyme mediated deletion and augmentation of the sickle-cell (β^S) mutation with fetal haemoglobin levels in the red cells. Selective inhibition of beta-globin RNA transcripts by antisense RNA molecules has equally been tried as a strategy to reduce levels of Hb S polymerization in red blood cells and the symptoms associated with SCD [19‐21]. Pace BS, et al.[21] specifically identified antisense RNA targets in the beta-globin gene other than the homologous regions in gamma-globin, proposing that gene therapy strategies which combine gamma-globin induction along with beta-globin inhibition using antisense vectors may yield more favorable anti-sickling effects longterm. Amosova O, et al.[22],on the other hand, reported third-strand directed repair of the sickle cell mutation using RNA-DNA chimeric oligonuceotides (COs) [23, 24] achieved by shortening the psoralen linker to enhance the specificity of photoadduct formation at the desired mutant T residue site. Despite these notable advances, the place for gene-replacement or repair therapy in SCD has remained rather experimental [25], with no clinical trials of any of the above approaches in human populations yet reported.

Basing on the more-recent developments in targeted mutagenesis (genome-editing) and gene-repair wrought by zinc finger nuclease (ZFN) technology [26‐32], it was hypothesized that the single, point mutation responsible for SCD can be abrogated using similar approaches. Specifically, Zinc finger nucleases-ZFNs, which are artificial, hybrid restriction enzymes created by covalently linking a DNA-binding zinc finger (Zif) domain (composed of three to six finger-arrays) to the non-specific DNA cleavage domain (or simply F_N) of the Flavobacterium okeanokoites bacteria restriction endonuclease, FokI [33]; have recently become a powerful tool for primarily editing host genomes. Following induction of a double strand break (DSB) within the target DNA, repair occurs, either by homologous recombination (HR) or non-homologous end-joining (NHEJ) [26‐32]. Perez et E al. [30]-using engineered ZFNs targeting human CCR5, previously demonstrated establishment of HIV-1 resistance in CD4+ T cells through generation of a doublestrand break (DSB) at predetermined sites in the CCR5 coding region upstream of the natural CCR5D32 mutation. Holmes N et al. [31] have demonstrated control of HIV-1 infection within NSG-mice transplanted with human hematopoietic stem/progenitor cells modified by zinc-finger nucleases targeting CCR5. Most recently, Wilen CB, et al.[32] successfully engineered HIV-Resistant Human CD4+ T Cells using CXCR4-Specific Zinc-Finger Nucleases (ZFN). This evidence, along with on-going improvements in the design and engineering of lentiviral [33, 34] and parvovirus [35] vectors (LV and PV, respectively) for ex-vivo or in-vivo gene-delivery and transduction of erythroid precursors, suggests that the sickle cell mutation may be abrogated in erythroid bone marrow precursor with appropriate ZFNs. Indeed, we are aware that Sangamo Biosciences (http://​www.​sangamo.​com/​index.​html)--one of the leading industries in ZFN-technology, has already focused its ZFN-mediated gene-editing technology to providing a unique solution for the treatment of monogenic diseases like hemophilia and SCA. Their ZFAs or ZFNs are, however, not publically available.

Thus, the specific goal of this study was to construct a database of zinc finger arrays (ZFAs) and engineer ZFNs that respectively specifically bind and cleave within or around the sickle hemoglobin beta (−β^S) gene mutation.

I present here a novel set of 57 zinc finger arrays (ZFAs) and 9 zinc finger nucleases (ZFNs) ― that constitute gene-therapeutic precursors for the targeted mutagenesis and repair of the SCD mutation or genotype. Specifically, although SCD results from a monogenic (Hemoglobin, beta) point-mutation (substitution of A with T) that has attracted extensive interest for target gene-therapy [16‐25], the place for gene-replacement or repair therapy in SCD has remained rather experimental [25], and no clinical trials of any of the above approaches in human populations are yet to be reported. Basing on the more-recent developments in targeted mutagenesis (genome-editing) and gene-repair wrought by zinc finger nuclease (ZFN) technology [26‐32], it was hypothesized that the single, point mutation responsible for SCD can be abrogated using similar approaches. To this end, we sought to construct a database of zinc finger arrays (ZFAs) and engineer ZFNs that specifically bind and cleave within or around the sickle hemoglobin, beta (−β ^S) gene. Now, we present a database of 57 ZFAs (see, Additional file 2 and Figure 1) specific to the β ^A gene (see, Additional file 1 for genomic contextual sequences). Using these ZFAs, we also constructed four ZFNs (shown in Additional file 3 and Figure 2) cleaving specifically within the same β ^A gene. Three of these four ZFN are shown in Table 1. Because the point-mutation of SCD is located at genomic contextual position −69-70-71- bp of the β ^A chain-gene sequences used (that is, within the 6^th codon of the gene), we were equally inspired to design another array of five ZFNs (shown in Additional file 5) specific to >24 bp target-sequences within the adjacent 8,954 bp to the 5′ end of the β ^A chain-gene (see Additional file 4 for the 8,954 bp localized to the 5′ end of the β ^A chain-gene). Two of the ZFN cleaving within these- 8,954 bp but closest to the 5′ end of the β ^A chain-gene, are shown in Table 2, while Figure 3 offers a graphic summary of their distribution.

The above ZFAs and ZFNs may be applied towards the target mutagenesis or repair of the sickle cell mutation, in various ways. Firstly, splicing out the entire genomic region located between positions 5′-82/106, and 3′ -1,333/1,359 or −1,334/1,359 of the β ^S globin gene may be achieved using the ZFNs shown in Table 1. This alone― when followed by the process of non-homologous end-joining (NHEJ) [26‐32] of the residual double strand DNA break (DSB)’s edges could lead to deletion of over 80% bp-sequences of the defective β ^S globin gene. Depending on the efficiency of the gene-delivery and transduction achieved by the vectors [34, 35] used within erythroid precursors, therefore, this strategy alone offers the possibility of reducing the intracellular (either mature red blood cell, RBC or erythroid precursor) expression of the β ^S globin gene, and ultimately curtailing the symptomatology resulting from S polymerization. Note that, because this approach does not involve deletion of the SCD point-mutation which is located at genomic contextual position −69-70-71- bp of the β ^A chain-gene sequences used (that is, within the 6^th codon of the gene), it is only a means to functionally irreversibly inactivate the β ^S globin gene. Thus, as a second option, another may purpose to delete the 5′ region before position −82/106 of the β ^S globin gene, which contains the point-mutation of SCD. This can be achieved using the ZFN in Table 1 that cleaves at position 82/106 and any of the two ZFNs in Table 2 that cleave within the 8,954 bp localized to the 5′ end of the β ^A chain-gene. Alternatively, however, secondary therapeutic events, including say (i) either repair of the target β^S globin gene cleavage-site by either recruiting the homologous recombination (HR) pathway through providing a normal, non-mutant β globin gene template for repair of the spliced ‘mutant (disease causing)’ region [26‐32, 36, 37] (ii) or supplementary replacement of this pathogenic-spliceon with a corresponding 5′ exon of γ-globin in a similar way as Lans N, et al.[17] and Weatherall DJ. [18] did with a 3′ exon, may be sought. Pace BS, et al.[21] has proposed that gene therapy strategies which combine gamma-globin induction along with beta-globin inhibition may yield more favorable anti-sickling effects in the longterm. Lastly, perhaps completely novel gene-therapeutic strategies devised in future may equally explore the target-DNA-binding mechanisms inherent in our ZFAs or ZFN to repair the mutant codon of the β ^S globin gene.

Our study presents a number of limitations that need to be taken into account. First, it is vital to note that the work is limited to only sequence analyses and accompanied by no in-vitro proof of concept studies. This should be attributed to the apparent limited resource-capacity in our Lab, although the same may be ratified by the fact that these very methods [36, 37] have previously been used to successfully assemble ZFAs and engineer ZFNs that are experimentally safe and effective [26‐32]. That said, it is imperative that further optimization be incorporated, including say: (i) modular analysis and assembly [39] to add on one, two or even three other ZFA onto our currently 3 finger-arrays so as to enhance specificity and avoid off-target host-genome toxicity; and (ii) in-vivo assembly and testing for efficacy, say by using either a bacteria-one hybrid (B1H) or yeast one-hybrid (Y1B) system to further inform the best ZFNs to use [40]. Modifications to the cleavage domain in order to generate a hybrid capable of functionally interrogating the ZFN dimer interface so as to prevent homodimerization, while-still enhancing the efficiency of cleavage [41], is also possible. Second, additional pre-clinical experiments employing either humanized-mouse cell models [16] or erythroid precursors cells, are still required to assess the efficacy and safety of these ZFNs and their transducing vectors [34, 35, 42]. Perhaps, those experiments that Wilen CB, et al.[32] recently conducted to assess the safety and efficacy of the Ad5/F35 vector carrying CXCR4-Specific Zinc-Finger Nucleases they used to engineer HIV-resistant human CD4+ T cells, may suffice. Specifically, it is important to (i) compare genome profiles and the hemoglobin, beta gene- sequence-profiles of the erythroid precursors targeted, and (ii) evaluate biophysical profiles of the mutagenic or repair resultant, presumably adult hemoblogin (HBβ ^S→A) produced, including analyzing its crystal structure (see Figure 4 for crystal structure of the normal adult hemoglobin A) [43‐45]Third, its reasonable for one to question how the lentiviral or parvovirus vectors carrying and transducing a diploid (or pair) copy of these ZFN will be used in the clinics within resource limited settings where SCD is most prevalent [25, 46]. Our translational projection or proposition is to have this attempted via in-vivo gene-delivery and transduction, rather than the expensive ex-vivo manipulation. Therefore, sub-dermal, intravenous or intra-osseous routes of injection or infusion for in-vivo use of these vectors as a less-Labour-intensive and affordable gene-delivery and transduction alternative to ex-vivo manipulation must seriously be considered and tried despite the evidence of low efficiency of gene-delivery and transduction offered by the in-vivo routes, when compared to ex-vivo gene delivery [34, 35]. Fourth, zinc finger nucleases targeting host-genes have recently been found to cleave off-target loci [47], and these off-targets-though minimized by reducing the binding energy of ZFN, could not be predicted by in-silico methods [48].

In conclusion, this set of 57 zinc finger arrays (ZFAs) and 9 zinc finger nucleases (ZFNs) ― offers us gene-therapeutic precursors for the targeted mutagenesis and repair of the SCD mutation or genotype. Specifically, the same may be used to either functionally or structurally abrogate the β ^S globin gene. Alternatively, supplementary replacement of this pathogenic-spliceon with a corresponding 5′ exon of γ-globin may be possible. Lastly, novel gene-therapeutic strategies devised in future may equally explore the target-DNA binding and cleaving mechanisms inherent in our ZFAs or ZFN to replace or repair the mutant codon of the β ^S globin gene.

The NCBI gene identity of the Hemoglobin, beta gene is |3043|; whereas the NCBI reference for the 8.95 kb region from base 5190000 to 5198951 is |NT_009237.18|