Modernising autism spectrum disorder model engineering and treatment via CRISPR-Cas9: A gene reprogramming approach

Abstract

A neurological abnormality called autism spectrum disorder (ASD) affects how a person perceives and interacts with others, leading to social interaction and communication issues. Limited and recurring behavioural patterns are another feature of the illness. Multiple mutations throughout development are the source of the neurodevelopmental disorder autism. However, a well-established model and perfect treatment for this spectrum disease has not been discovered. The rising era of the clustered regularly interspaced palindromic repeats (CRISPR)-associated protein 9 (Cas9) system can streamline the complexity underlying the pathogenesis of ASD. The CRISPR-Cas9 system is a powerful genetic engineering tool used to edit the genome at the targeted site in a precise manner. The major hurdle in studying ASD is the lack of appropriate animal models presenting the complex symptoms of ASD. Therefore, CRISPR-Cas9 is being used worldwide to mimic the ASD-like pathology in various systems like in vitro cell lines, in vitro 3D organoid models and in vivo animal models. Apart from being used in establishing ASD models, CRISPR-Cas9 can also be used to treat the complexities of ASD. The aim of this review was to summarize and critically analyse the CRISPR-Cas9-mediated discoveries in the field of ASD.

Key Words: Autism spectrum disorder, CRISPR-Cas9, Cellular models, Organoids, Animal models, Therapeutic strategies

Core Tip: There are several reviews in the literature explaining the underlying mechanisms contributing to the pathophysiology of autism spectrum disorder by performing several preclinical experiments. Given the significant role of genetics (de novo or inheritable) in the development of autism spectrum disorder, disease specific models should be established for investigating the mechanism involved. Therefore, this review specifically focused on the use of an emerging genomic editing tool, clustered regularly inters-paced palindromic repeats/Cas9, for generating different types of preclinical models as well as new therapeutic options, providing a novel insight into the disease.

INTRODUCTION

Identifying the double helix DNA structure and finding technologies to manipulate it ultimately led to an extensive investigation of genomic structure[1]. Manipulation of genomic structure requires various genomic editing techniques including homing-endonucleases or mega nucleases, zinc finger nucleases, transcription activator like effector nucleases and clustered regularly interspaced short palindromic repeats (CRISPR)-associated protein 9 (Cas9)[2]. Advancement in this field has permitted researchers to alter the DNA of model organisms to obtain the model of interest. In this context, the discovery of the CRISPR-Cas9 system has greatly and enormously expanded the field of study related to the genetic underpinnings of complex and heterogeneous disorders like autism spectrum disorder (ASD)[3]. From bacterial defence systems to genomic engineering tools, CRISPR-Cas9 has been proven beneficial in providing a novel insight into a possible genetic mutation in ASD[4].

The spectrum of disorders under ASD is pervasive. Due to the complexity of this medical condition, it is challenging to determine the diagnostic threshold, making diagnosis difficult. Despite the apparent difficulties connected with identification of ASD aetiologies, intensive genetic investigations have shown that ASD has a substantial genetic basis. Genetic analysis has revealed many susceptibility genes[5]. In addition to this, ASD has been found to be associated with several other disorders such as anxiety, depression, attention-deficit hyperactivity disorder and obsessive-compulsive disorder[6] and with genetic syndromes like Rett syndrome, Angelman syndrome, Timothy epilepsy and Fragile-X syndrome[7,8]. Depending upon the origin of the condition, ASD is diagnosed as syndromic if it is due to specific genetic syndromes with well-defined aetiology, such as Rett syndrome, and it identified as non-syndromic if ASD is diagnosed as the first diagnosis in patients having less-characterized aetiology[9,10].

There are multiple factors involved that contribute to the development of ASD in cases of non-syndromic ASD[10]. Therefore, due to the several aetiologies involved, ASD is considered a heterogeneous group of highly heritable disorders[11,12], and their risk factors could be genetic as well as environmental. A significant role of genetics in the development of the disorder has been known for a long time as confirmed by a meta-analysis of twin studies, which stated that ASD is inherited nearly 64%-91% in monozygotic twins and around 30% in dizygotic twins[13-15].

Modelling of disease at in vitro, in vivo and organoid levels are major avenue of research for investigation of abnormal early brain development because several ASD-associated genes have been found to be highly expressed during prenatal brain development of patients[16]. CRISPR-Cas9 has been successfully used to generate genetically engineered models that could mimic the disorder. At the same time, gene therapies are one of the emerging fields in recent years aimed at curing a wide range of diseases including ASD. Moreover, based on the available genetic information, novel gene therapies have also been created, which may help identify the potential ASD therapy candidates. The advent of CRISPR-Cas9 in gene therapy has been helpful in either silencing the gene using non-homologous end joining or correcting the genetic mutation using non-homologous recombination[17]. These developments have given patients new hope regarding rational treatment against the disease. This article provided an overview of the potential use of CRISPR-Cas9 technology for the establishment of appropriate ASD models along with its application in therapeutic strategies at the genomic level.

GENETIC ARCHITECTURE OF ASD

The genetic background contributing to autism aetiology involves copy number variations, somatic mutations, de novo mutations, single nucleotide variations, insertions, deletions and chromosomal abnormalities[18,19]. These factors interfere with the protein-coding genes involved in neuronal development and several other ASD candidate genes related to critical processes like DNA binding, transcription, postsynaptic density and neuroprotection[20]. Any alteration in well-known ASD-associated genes can ultimately result in impaired working of brain areas responsible for cognitive functions[21,22]. Forkhead box protein 1 (FOXP1) and fragile X messenger ribonucleoprotein 1 (FMR1) are transcription factors and regulating genes. Others, like methyl CpG binding protein 2 (MECP2), tuberous sclerosis 1 (TSC1), SH3- and multiple ankyrin repeats protein 1 (SHANK1), ubiquitin protein ligase E3A and contactin-associated protein-like 2 (CNTNAP2), are involved in a wide range of functions like chromatin remodelling, cell proliferation, maintaining synaptic activity, protein ubiquitination and cell adhesion, respectively. Moreover, mutations in MECP2 and FMR1 are related to genetic syndromes such as Rett syndrome and fragile X syndrome, respectively[3].

The latest advancements in the development of next-generation sequencing have offered opportunities for genetic analysis to elucidate the underlying genetic mechanisms of ASD[23]. Whole exome sequencing has revealed that some biallelic mutations in proximal assembly proteins, phenylalanine hydrolyses and spectrin repeat containing nuclear envelope protein 1 are associated with familial ASD[24]. These genes also include those that are known to control or be controlled by synaptic activity (e.g., MECP2, spectrin repeat containing nuclear envelope protein 1). Genetic analysis using whole genome sequencing has shown that copy number variations and single nucleotide variations result in missense mutations with an overall increase in missense variants, including some ASD risk genes[25]. In addition to this, genome-wide association studies have been able to identify a few potential variants being implicated in the pathogenesis of ASD[26]. Altogether, mutations in specific genes, known to regulate the important biological pathways, neuronal networks, synaptic activity and plasticity, etc, contribute to development of ASD and associated clinical symptoms (Figure 1).

Figure 1 Schematic diagram describing the structure and functioning of clustered regularly interspaced short palindromic repeats-associated protein 9 technique in autism spectrum disorder. In this schematic, we highlighted the mechanism of clustered regularly interspaced short palindromic repeats (CRISPR)-associated protein 9 (Cas9) in recognizing a target using protospacer adjacent motifs sequencing and causing cut at specific point. Following cleavage and forming double stand breaks, repair systems like non-homologous end joining and homology directed repair come into play for avoiding any unspecific mutations. Diverse application of CRISPR-Cas9 has been explained in this diagram for investigating the mechanism involved in autism spectrum disorder pathophysiology. Various potential therapeutic targets for autism spectrum disorder could be investigated using CRISPR-Cas9 technology. sgRNA: Single guide RNA; ASD: Autism spectrum disorder; PAM: Protospacer adjacent motifs; DSB: Double strand breaks; NHEJ: Non-homologous end joining; HDR: Homology directed repair.

STRUCTURE AND FUNCTION OF CRISPR-Cas9

CRISPR-Cas9 is used to cut DNA at predetermined target locations. Although the method has already been revolutionised as a gene editing tool, researchers are constantly exploring new applications. Since being discovered as a bacterial immune system against invading viruses, CRISPR-Cas9 has been adapted as a powerful tool in genomic research. Repeat elements in CRISPR were initially noticed in Escherichia coli by Ishino et al[27]. Contrary to conventional tandem repeats in the genome, the CRISPR repeat clusters were interestingly separated by non-repeating DNA sequences known as spacers. Complete genome sequencing of bacteria and archaea led researchers to determine that these CRISPR elements are adjacent to well-conserved CRISPR-associated genes (Cas)[28]. This whole structure including palindromic repeats, spacer DNA and Cas gene is known as the CRISPR array. After a decade of research, scientists have finally discovered that the spacer DNA sequences belong to viruses[29,30].

The study by Gasiunas et al[31] provided the most significant experimental data about the potential utility of CRISPR systems for bacteria. The concept that the Cas9 enzymes in bacteria can be reprogrammed to target a specific DNA sequence has been the key discovery, which signalled the beginning of CRISPR as a biotechnological gene-editing tool[31,32]. CRISPR RNA and transactivating CRISPR RNA are both vital parts of guide RNA (gRNA) and are required for the functioning of the CRISPR system. Notably, Jinek et al[32] demonstrated that CRISPR-Cas9 could also be guided by single gRNA, a chimeric RNA created by joining transactivating CRISPR RNA and CRISPR RNA[32]. These studies were the reason for adopting CRISPR-Cas9 as a gene editing tool.

The ability of the CRISPR-Cas9 system to produce an autism model and its therapeutic potential are the main topics of this review. In 2012, Doudna and Charpentier[32] found that by using the appropriate template, CRISPR-Cas9 could be used to edit any desired DNA. Depending on how Cas proteins act, the CRISPR-Cas9 system has been divided into type I, type II and type III systems. Type II is the most well-studied and simplest for application in genetic engineering[33].

The Cas9 protein performs the function of genetic scissors in the type II system by producing a double-stranded break (DSB) in the DNA[34]. The Cas9 protein contains two structural lobes, one that aids in recognition (REC) and the other that aids in nuclease activity. The REC lobe consists of REC1 and REC2, which are involved in the recognition of gRNA. The nuclease also has a protospacer adjacent motif (PAM) interacting domain responsible for the binding of Cas9 to targeted DNA. The gRNA is used to target viral DNA in prokaryotes, but when utilised as a gene-editing tool, it can be synthetically constructed to target virtually any gene that needs to be changed.

The three phases of the CRISPR-Cas9 genome editing system are recognition, cleavage and repair[35]. Single gRNA binds to a complementary area on the targeted DNA to begin the recognition process. PAM is a 2–5 base pair sequence that has an “NGG” pattern, where “N” stands for any nucleotide followed by two guanine nucleotides. Once the PAM site is identified, double stranded DNA starts melting at the target site followed by an RNA-DNA hybrid formation. Now, the Cas9 protein is ready to make a DSB at the targeted DNA 3 base pairs upstream to PAM[36]. In the last step, the double stranded blunt ended breaks are repaired by non-homologous end joining and homology directed repair by cellular machinery[34,37,38]. By inserting a donor DNA template with sequence homology at the anticipated DSB site, homology directed repair carries out the precise gene insertion or replacement[39]. The property of CRISPR-Cas9 to either activate genes or to repress genes has been utilised to regulate the transcriptional level of gene expression.

CRISPR-Cas9 MEDIATED GENETIC ENGINEERING OF ASD

Most cases of ASD are idiopathic, with illusive aetiology[40]. The heterogeneous molecular nature of ASD makes it really difficult to understand the associated risk factors and the underlying mechanisms. Modelling ASD is notably challenging due to its multigenic aetiology. Only pertinent and validated disease-specific models could be helpful in discovering novel biomarkers and related therapeutic targets[41].

CRISPR-Cas9 engineered cellular models of ASD

Numerous neurodevelopmental diseases, including ASD are studied using cellular models because of the short experimental period and no ethical concerns and are less expensive. Researchers can create early human brain development, alterations in ASD or any other neurological disorders using in vitro models. Induced pluripotent stem cells (iPSCs), which can grow indefinitely in vitro, can be created by reprogramming somatic cells. Patient-derived cellular models have been validated and are realistic while preserving the genetic makeup of the donor and are an effective tool for deciphering the pathophysiology of ASD. The emergence of the genomic editing tool, CRISPR-Cas9, is helpful in facilitating more efficient in vitro models of ASD considering its genetic background. Using this technique, the researcher can edit primary cultured neural cells or isogenic cell lines by either introducing mutations derived from ASD patients or correcting them. Moreover, this technology reduces genetic background variation and directly correlates the observed symptoms and the associated mutation[42], which further provides information about the role of the particular ASD risk gene in neurodevelopment.

It is known that aberrant neurogenesis and synaptogenesis lead to functional impairments in brain networks in ASD[41]. Therefore, early molecular events during ASD development can be replicated in a model system in neurogenin 2-directed induced iPSCs (for excitatory neurons) that are further differentiated into forebrain glutamatergic neurons. Using that information and based on the whole exome sequencing results of some selected ASD-associated risk genes, the CRISPR-Cas9 approach was used to generate knockout (KO) iPSCs for the functional studies of the following genes: Anosmin 1; FMR2; calcium voltage-gated channel subunit alpha1 C; astrotactin 2; alpha-thalassemia/mental retardation, X-linked; chromodomain helicase DNA binding protein 8 (CHD8); disks large-associated protein 2; teneurin transmembrane protein 1; potassium voltage-gated channel subfamily q member 2; and sodium voltage-gated channel alpha subunit 2 (SCN2A). They revealed that ASD genes could result in similar electrophysiological phenotypes and transcriptional rewiring in the human iPSC-derived excitatory neurons model system[43].

Apart from the role of the SHANK3 gene in synaptogenesis, one of the other consequences of its haploinsufficiency is hyperpolarization-activated cation channelopathy, which contributes to ASD pathogenesis. This impairment was analysed by generating SHANK3 deletion by CRISPR in human embryonic stem cells[44]. Findings also highlighted that iPSC-derived glutamatergic neurons deficient in at least one allele of CNTN5/euchromatic histone lysine methyltransferase 2 resulting in ASD-associated phenotypes presented the increased synaptic activity of excitatory neurons in vitro[45]. In addition, CRISPR mediated inactivation of euchromatic histone lysine methyltransferase 1 in human neurons, which is directly associated with n-methyl-D-aspartate receptor hyperfunction and is implicated in ASD pathophysiology[46].

The major obstacle in the treatment of ASD is testing different drug candidates because of its aetiological heterogeneity. Therefore, an in vitro study has been done using the CRISPR tool for introducing mutations in activity-dependent neuroprotective protein, dead-box helicase 3 X-linked and FOXP1 genes to create a relevant ASD model[47]. Similarly, hemizygous CHD8 (CHD8+/−) iPSC lines were designed to investigate the role of CHD8 in embryo development at the molecular and cellular levels. According to transcriptomic profiling, CHD8 regulates several other genes connected to the development of ASD[48]. In addition to ASD-associated genes, the role of long noncoding RNAs, such as patched domain containing 1-antisense RNA[49]and molybdenum cofactor sulfurase[50], in ASD development was studied using CRISPR technology in human induced pluripotent stem cells (hiPSCs). Cellular models are briefly summarized in Table 1. These aforementioned findings indicate that these ASD associated genes may be a therapeutic target for the treatment of ASD.

Table 1 Summary of clustered regularly interspaced palindromic repeats-associated protein 9 engineered models of autism spectrum disorder.

Ref.	Model	Gene mutation/syndrome	Observed alterations
Cellular
[32]	ES cells	SHANK3/Phelan-McDermid syndrome	Altered neuronal morphology and synaptic connectivity; impaired Ih channels
[33]	iPSCs	CNTN5 or EHMT2/idiopathic ASD	Increased synaptic excitatory neuron activity
[34]	iPSCs	EHMT1/Kleefstra syndrome	Upregulation of NMDAR1; neuronal network impairments
[35]	iPSCs	ADNP, DDX3X and FOXP1/idiopathic ASD	Increased neurogenesis
[36]	iPSCs	CHD8/idiopathic ASD	Dysregulated expression of genes associated with human brain volume or head size
[37]	iPSCs	PTCHD1-AS/idiopathic ASD	Impaired excitatory synaptic function (NMDAR hypofunction); synaptic impairment
[38]	iPSCs	COSMOC/idiopathic ASD	Destabilization of lipid and energy metabolism; affected neuronal maturation
Organoids
[41]	Cerebral	PTEN/idiopathic ASD, CHD8/idiopathic ASD, SUV420H1/ idiopathic ASD	Increased upper layer colossal neurons, cycling progenitor neuron; high outer radial glial cells; increased cortical interneurons; increased newly born deep layer projection neurons
[42]	Cerebral	RAB39b/ idiopathic ASD	Increased NPC proliferation
[44]	Cortical and neurons	MECP2/Rett syndrome	Dysregulation in genes of neuronal and glial cells
[46]	Cortical	UBE3A/Angelman syndrome	Dysfunction in big potassium channel dysfunction causing increased neuronal excitability
[47]	Cortical	TSC1 or 2/Tuberous sclerosis complex	Affected cortical neurons and glial cell development
[49]	Cortico-striatal organoids	SHANK3/ Phelan-McDermid syndrome	Enhanced neuronal excitability; dysregulated expression of protocadherins and zinc-finger genes
[50]	Cortical	FMR1/Fragile X syndrome	Increased number of glial cells and bigger organoid size
Animal Models
[52]	Cynomolgus macaques	SHANK3/ Phelan-McDermid syndrome	Sleep disturbances; increased repetitive behaviour, motor deficit; social and learning impairment; aberrant neural circuit connectivity
[53]	Mice	ARID1B/idiopathic ASD	Social behaviour impairment; altered vocalization; anxiety-like behaviour; neuroanatomical abnormalities; growth impairment
[54]	Mice	CHD8/idiopathic ASD	Cognitive impairment; disrupted pathways involved in neurogenesis, neuroimmune signalling, synaptic processes
[55]	Mice	ASH1L/idiopathic ASD	Dysregulated epigenetic modification; upregulation of neurexin-1α
[56]	Rat	CYFIP1/idiopathic ASD	Extensive changes in white matter; myelin sheath thinning in corpus callosum; abnormal oligodendrocytes; behavioural inflexibility
[57]	Rat	TCF4/idiopathic ASD	Attenuated action potential output; alteration in electrophysiological properties in neurons
[58]	Rat	UBE3A/idiopathic ASD	Deficits in motor coordination as well as learning and memory
[59]	Zebra fish	CHD8/idiopathic ASD	Increased head size; reduction in post mitotic enteric neurons
[60]	Zebra fish	FMR1/Fragile X syndrome	Abnormal behaviour; learning memory deficits; impaired craniofacial cartilage development
[61]	Zebra fish	NR3C2/idiopathic ASD	Disruption in sleep and social functions
[62]	Zebra fish	SHANK3/ Phelan-McDermid syndrome	Reduced social nitration and locomotory activity; repetitive swimming behaviour; reduced levels of post synaptic homer1 and presynaptic synaptophysin

ADNP: Activity-dependent neuroprotective protein; ARID1B: AT-rich interaction domain 1B; ASD: Autism spectrum disorder; ASH1L: ASH1-like histone lysine methyltransferase; CHD8: Chromodomain helicase DNA binding protein 8; CNTN5: Contactin-associated protein-like 5; COSMOC: Molybdenum cofactor sulfurase; CYFIP1: Cytoplasmic FMR1 interacting protein; DDX3X: Dead-box helicase 3 X-linked; EHMT1/2: Euchromatic histone lysine methyltransferase 1/2; FMR1: Fragile X messenger ribonucleoprotein 1; FOXP1: Forkhead box protein 1; ES: Embryonic stem; Ih: Hyperpolarization-activated cation; iPSC: Induced pluripotent stem cell; MECP2: Methyl CpG binding protein 2; NMDAR1: N-methyl-D-aspartate receptor 1; NPC: Neural progenitor cell; NR3C2: Nuclear receptor subfamily 3 group c member 2; PTCHD1-AS: Patched domain containing 1-antisense RNA; PTEN: Phosphatase and TENsin homolog; RAB39b: RAS-related protein Rab-39B; SHANK3: SH3- and multiple ankyrin repeats protein 3; TSC1/2: Tuberous sclerosis 1/2; TCF4: Transcription factor 4.

CRISPR-Cas9 engineered organoids of ASD

The lack of suitable ASD models has always been a hindrance in ASD research because neither 2D cell culture nor animal models can accurately mimic the aetiology of ASD. Therefore, 3D in vitro models like organoids have recently emerged in the field of research. They have been shown to reproduce the gene expression profile, transcriptome, epigenome and disease dynamics of both idiopathic and syndromic ASD[51]. Like other cellular models, iPSC-derived organoids are being used because of no ethical concerns and are preferred over 2D culture and animal models as they can generate more disease-specific models.

This methodology has become even more reliable due to the integration of CRISPR-Cas9 to produce isogenic controls, significantly reducing genetic background differences. Idiopathic ASD has been connected to abnormalities in several genes, and genetic research has found multiple mutations that are linked to this condition[52]. Enhanced neurogenesis in idiopathic ASD has been studied through CRISPR engineered organoid models to create mutations in histone methyltransferase SUV420H1, the tumour suppressor phosphatase and TENsin homolog[53], CHD8 and the GTPase-encoding RAS-related protein Rab-39B[54]. These genes are linked to macrocephalic ASD, and CRISPR-mediated deletion resulted in larger haploinsufficient cerebral organoids in comparison to isogenic control due to overactivation of the P13K-AKT-mTOR pathway[54].

Modelling of syndromic ASD is also being achieved using cerebral organoids to investigate the underlying genetic mechanism. One of the important ASD-associated genes MECP2 is considered critical for early brain development, but its loss-of-function mutations are a common underlying aetiology of Rett syndrome[55], causing severe impairment in human interneurons and ultimately neurogenesis. Human MECP2-KO neurons and cortical organoids were used using CRISPR to investigate its neuropathological function[56,57]. Mutation (deletion) in UBE3A is also related to the pathology of syndromic ASD, and an organoid model derived from human iPSCs demonstrated hyperexcitability in brains contributing to network dysfunction[58].

Similarly, cerebral organoids are used for studying other syndromic ASD, such as a mutation in TSC1/TSC2 genes in CRISPR-engineered human cortical spheroid model[59]. It caused synaptic imbalances, with an increase in γ-aminobutyric acid synapses[60]. Human corticostriatal organoids were studied using CRISPR-generated SHANK3 gene deletion for modelling autism[61]. hiPSC-derived brain organoids with CRISPR-Cas9 induced FMR1-KO, which caused an abnormal increase in astrocyte number, was utilized to model FXS, a syndromic ASD[62]. Various organoid models of syndromic as well as idiopathic ASD is summarized in Table 1.

CRISPR-Cas9 engineered animal models of ASD

Despite the capabilities of in vitro models to recapitulate the basic aetiology of ASD, animal models are preferred as a more fundamental tool to fully understand the complexity involved in ASD. Animal models allow a researcher to investigate behavioural and developmental features in addition to molecular parameters. However, generating an ASD animal model is a time-consuming procedure and involves ethical concerns, but it is helpful in studying neurodevelopmental disorders. Moreover, in the case of ASD, it is helpful in validating the implication of critical genes in the development of ASD.

The emerging CRISPR-Cas9 approach has been a great help in creating various genetic animal models (KO, Knock-in, overexpression and point mutation) to study various ASD-associated genes identified in an individual with ASD. ASD models can be studied in multiple species like rodents including mice, rats, monkeys, fruit flies and zebrafish, depending upon the requirement and purpose of the experiment[63]. CRISPR-mediated generation of mutations in the SHANK3 gene by creating insertions and deletions (indels) in exon 21 led to the development of an ASD model in monkeys and their F1 offspring, showing atypical autistic phenotypes like increased repetitive behaviour along with social and learning deficits[64].

Studies have reported that a CRISPR-mediated mutation in ASD-associated genes such as AT-rich interaction domain 1B[65], CHD8[66] and ASH1-like histone lysine methyltransferase[67] showed ASD-like symptoms in mice. To investigate genes implicated in ASD such as cytoplasmic FMR1 interacting protein[68], transcription factor 4[69] and UBE3A[70] in a rat model created with CRISPR engineered technology was studied. The rats showed autistic phenotypes like alteration in behavioural flexibility, learning ability and memory difficulties.

Similarly, a zebrafish model of ASD using the CRISPR strategy has been used to study the functional role of genes in the development of ASD such as CHD8[71], FMR1[72], nuclear receptor subfamily 3 group c member 2[73]and SHANK3[74]. Major ASD-linked phenotypes observed in these zebrafish models are macrocephaly, hyperactivity, anxiety, impaired social behaviour, sleep disturbances and altered neuronal development (summarized in Table 1).

CRISPR-Cas9-BASED THERAPEUTIC STRATEGIES AND POTENTIAL TARGETS

Over the years, the CRISPR-Cas9 genome editing tool has evolved as a specific delivery tool for delivering genes to the target cells including neural and brain cells. One such benchmark was set by Staahl et al[75], where the engineered variants of the Cas9 ribonucleoprotein complex were delivered to the mice hippocampus, striatum and cortex region and demonstrated the in vivo neuronal gene editing[75]. The advances in the genome editing tool have opened the door for eradicating the genetic mutations underlying severe neurological diseases like ASD.

Several genes that are linked to ASD can be targeted for correction using the CRISPR-Cas9 approach to reduce the disease burden (summarized in Figure 1). The genes that undergo mutations in ASD and ASD-associated monogenic syndromes include calcium voltage-gated channel subunit alpha1 C, FOXP1/2, wingless-related integration site-2, CHD8, homeobox B1, reelin, inner mitochondrial membrane peptidase subunit 2, oxytocin receptor gene, methylenetetrahydrofolate reductase, SHANK2/3, γ-aminobutyric acid type A receptor subunit, homeobox A1, UBE3A, NCK associated protein 1, human serotonin transporter gene, POU class 3 homeobox 2, reduced arabinose yariv1/suppression of tumorigenicity 8, FMR1[76-77], MECP2, TSC1, PTK7, SCN3A and CNTNAP2[78-82]. Some of these genes for monogenic syndromes associated with ASD are targeted using the CRISPR-Cas9 tool in vitro and in vivo; however, many others remain to be explored.

The lack of target specificity or the polygenic form of ASD limits the use of the CRISPR-Cas9 tool as a therapeutic strategy in ASD. The CRISPR-Cas9-based therapeutic strategies that had been explored are summarized in Table 2; They primarily consist of the monogenic form of ASD. One of the studies by Lee et al[83] demonstrated that gold nanoparticle delivery of CRISPR-Cas9 ribonucleoprotein rescued the exaggerated repetitive behaviours in mice caused by fragile X syndrome[83]. The study demonstrated minimal off-target effects, and the editing target used was the metabotropic glutamate receptor subtype 5 gene, one of the overexpressed targets in ASD-associated syndromes[83-85].

Table 2 Summary of clustered regularly interspaced palindromic repeats-associated protein 9 edited therapeutic targets of autism spectrum disorder.

Ref.	In vitro/invivo	Gene mutation/editing method	Observed alterations
[80-82]	BTBR T + tf/J (BTBR), Fmr1 knockout, C57BL/6 mice	mGluR5	Rescued the exaggerated repetitive behaviours in mice caused by fragile X syndrome
[83]	HEK293 cell and Human iPSC (BCRT cell line)	MECP2	Reversal of ASD-associated Rett syndrome-like symptoms
[84]	RX41X iPSC and NOD/SCID female mice	SHANK2	Positive impact on nerve cells was reported like an increase in synapse number, dendritic complexity and length
[85]	C57BL/6 mice, Ube3am-/p+ mice and Ube3am-/pYFP mice on the C57Bl/6	Antisense transcript of UBE3A	Rescued the anatomical and behavioural phenotypes in a mouse model of Angelman syndrome
[86]	HEK293FT cells	FMR1	Fragile X syndrome improved by knocking out the CGG
[89]	Mef2c L35P knock-in mouse	MEF2C	Reversal of autistic-like behaviour

ASD: Autism spectrum disorder; CGG: Cytosine-guanine-guanine; FMR1/Fmr1: Fragile X messenger ribonucleoprotein 1; iPSC: Induced pluripotent stem cell; MECP2: Methyl CpG binding protein 2; MEF2C: Myocyte-specific enhancer factor 2C; mGluR5: Metabotropic glutamate receptor subtype 5; SHANK2: SH3- and multiple ankyrin repeats protein 2.

In another study, the CRISPR-Cas9 tool was used to correct the MECP2 mutations responsible for ASD-associated Rett syndrome via homology directed repair in hiPSCs[86]. Loss-of-function mutations in the SHANK2 gene has been associated with monogenic ASD. CRISPR-Cas9-mediated correction of a nonsense mutation on SHANK2 was demonstrated in iPSCs, and the positive impacts on nerve cells were reported, including an increase in synapse number and dendritic complexity and length[87].

In Angelman syndrome (monogenic form of ASD) caused by deletion of the maternally inherited UBE3A allele, the CRISPR-Cas9 approach was used to knock out the antisense transcript of UBE3A in cultured human neurons and a mouse model. The antisense transcript of UBE3A is a long non-coding RNA that silences the paternal copy of the UBE3A allele and leads to the neurodevelopmental syndrome. The CRISPR-Cas9 approach was used to terminate the long non-coding RNA termed as antisense transcript of UBE3A, which led to the copy of the UBE3A allele available for transcription (activation of UBE3A) and hence rescued the anatomical and behavioural phenotypes in the mouse model of Angelman syndrome[88].

In another study, the CRISPR-Cas9 approach was used to improve fragile X syndrome by knocking out the cytosine-guanine-guanine (CGG) repeats expansion, recovering FMR1 expression in vitro. FMR1 encodes fragile X mental retardation protein, which undergoes epigenetic silencing because of the addition of CGG repeats and excessive DNA methylation, thus the CRISPR-Cas9 approach was used to excise the CGG expansion in the iPSCs[89]. A recent study used the CRISPR-Cas9 tool to activate the extracellular matrix receptor b3 integrin. The study also validated the involvement of b3 integrin haploinsufficiency in the pathophysiology of ASD and ASD-associated fragile X syndrome[90].

All CRISPR-Cas9-based therapeutic strategies established so far mainly comprise the proof of principle studies and have used the conventional homology-directed repair pathway to correct the mutations in the monogenic form of ASD. However, with the advancements in CRISPR-Cas9 genome editing tools, the most recently introduced concept of the base editing technique for more specific genome editing has been explored in fewer studies. One such study used CRISPR-Cas9-based cytidine base editors and the fourth generation base editor system to selectively modify the disco-interacting protein 2a and 2c genes in cell culture. Both of these genes are highly expressed in the central neuron system and known to be associated with ASD[91].

In another study, the CRISPR-mediated cytidine base editor system was used to restore the impairments in social interactions and repetitive behaviours in a knock-in mice model of autism. The de novo mutation in the gene myocyte-specific enhancer factor 2C was introduced in the mice brain, which displayed autistic-like behaviour. With the help of the base editing system the myocyte-specific enhancer factor 2C mutation was eradicated, and the reversal of symptoms was reported in mice[92]. A study demonstrated the use of CRISPR-Cas9 for correcting the mutation in CNTNAP2 in an organoid model derived from patients with syndromic ASD by rescuing the phenotype of organoid overgrowth. This CNTNAP2-organoid model provided an opportunity for further mechanistic inquiry and development of new therapeutic strategies for ASD[93]. Another finding has shown the use of a CRISPR activation-based approach for rescuing abnormalities in SCN2A haploinsufficiency-associated ASD[94]. The CRISPR-Cas9 mediated base editing system is just the beginning of an era of targeted gene modification, which can bring a breakthrough in the treatment of ASD.

A plethora of studies is being conducted worldwide using several targets in cultured cells or in animal models. However, the extrapolation to patients has not been achieved yet. The advances in the techniques leading to improved specificity, targeted delivery and personalized therapeutics will definitely help in the bench-to-bedside conversion of these CRISPR-Cas9 based therapies and help in reducing the disease burden.

CONCLUSION

Understanding brain function and its complexities have only been made possible by emerging genomic engineering tools like transcription activator like effector nucleases, zinc finger nucleases and CRISPR-Cas9. Opportunities for manipulating the genome have created the possibility to generate models for understanding a complex neurological disorder like ASD. Among these genomic editing tools, CRISPR-Cas9 is being considered the most extensive and effective, with the advantages of low mutation rate, high target efficiency and cost-efficient. CRISPR has enabled the creation of models that reproduce exactly the same causal mutations identified in patients, which has made it possible to determine an appropriate and disease-specific drug therapy.

Owing to the heterogeneous nature of ASD, it is difficult to identify the exact cause of ASD in patients as it could be genetic or environmental. No standard medication has been developed for treating ASD, except for aripiprazole and risperidone for irritability and aggressiveness. Thus, creating a reliable model, establishing a causal factor and representing all the characteristics of the disease is difficult. In vitro modelling of ASD has been a great benefit for understanding the underlying mechanism involved in the pathogenesis of ASD. However, it does come with limitations like high heterogeneity among hiPSCs lines. Therefore, reprogramming strategies need to be optimized. CRISPR-Cas9 potentially overcome such limitations by generating isogenic cell lines and increasing the reproducibility of experiments.

To further investigate the pathogenesis of ASD, the genome of animals can be successfully edited to construct a validated KO and knock-in models using CRISPR. These animal models have been reported to present phenotypes, including neuroanatomical, behavioural and morphological characteristics, caused by ASD-associated genes. In that regard, such models are helpful in determining the aetiology of the condition as well as screening appropriate drugs to restore the altered phenotype. Advancement in genomic editing systems is an encouraging indication that could restore the wild-type sequence and potentially be effective in human treatment trials. Utilization of the CRISPR-Cas9 tool is not only limited to the modelling of ASD but also has been helpful in targeting the mutated genes and correcting them.

Based on the available genetic information, ASD-associated genes have been widely explored, but their therapeutic potential is limited to monogenic forms of ASD and remains unexplored in polygenic form of ASD. Also, due to lack of target specificity, genetic therapy using CRISPR-Cas9 is unable to target every ASD- associated gene. Other approaches, such as CRISPR-mediated activation of a gene in which nuclease-deficient Cas9 was fused with a transcriptional activator or the CRISPR-mediated base editor system in gene therapy, have been helpful in restoring and normalizing gene dosage in ASD. However, this method has not been explored well, and optimization of this procedure is necessary before utilization.

Despite advancements in CRISPR-Cas9 tools, there are certain numbers of limitations like off-targeting, delivery method and immunogenicity and associated risks that make it challenging to use in clinical trials. A high frequency of off-targets is a prime concern while using CRISPR for gene therapy because it can lead to further mutations in undesired genomic locations. However, emergence of bioinformatic tools have been helpful in reducing the off-target effects while predicting the off-target modifications. Another major concern is immunogenicity caused by the introduction of Cas9 and delivery methods using viral vectors. Cas9 is derived from Streptococcus pyogenes, which is responsible for various human infections. Therefore, many patients would already harbour pre-existing anti-Cas9 antibodies. Therefore, when it is introduced for therapy purposes in humans, it will be recognised as a foreign antigen. An immune response may develop and cause degradation of Cas9, which would prevent it from gene editing. Another safety concern is the DSBs induced by CRISPR, which often trigger apoptosis. In addition to this, induced DSBs have also resulted in unnecessary massive deletions and rearrangements of sequences, suggesting a significant safety concern for the clinical use of DSB-inducing CRISPR therapy.

Given the challenges involved in using these gene editing techniques, gene therapy is still a distant therapeutic approach. Considering all limitations and the need for improvising CRISPR technology, studies using genomic editing tools is limited to cultured cells or animal models. Extrapolation of such experiments in patients has not been yet achieved. Therefore, the application of results from preclinical studies to the clinical treatment of ASD will require extreme care.