The exact origin of the first serotype of AAV (AAV1) is still unknown, as it was not initially isolated from tissues, but as a contaminant of Ad stocks ^[29][30][33,34], and its antibodies were found both in humans and non-human primates (NHPs) ^[31][35]. This serotype uses sialic acid as its primary cellular surface receptor ^[32][36], and AAV receptor (AAVR) as a coreceptor ^[33][37]. According to Rabinowitz J.E. et al., AAV1 does not bind heparin as it lacks R585 and R588, the amino acid residues required for such binding, and thereby it cannot be purified using heparin ^[29][34][33,38]. Zolotukhin S. et al., developed a protocol for AAV1 chromatographic purification by iodixanol gradient centrifugation and anion-exchange chromatography ^[35][39]. It can also be purified using mucin columns, as it can bind the sialic acid residues in mucin ^[36][40]. Moreover, recombinant AAV1 (rAAV1) was not found to contain any detectable post-translational modifications (PTMs), according to a systematic analysis conducted on ten AAV serotypes by Mary B. et al. ^[37][41], and it was the first viral vector to be approved for use in gene therapy ^[38][42].

2.2. AAV2

4.2. AAV2

AAV2 is considered the most studied serotype among all AAVs ^[39][45]. It was first discovered in 1965 as a contaminant of simian Ad preparations ^[40][52]. Later, in 1998, its primary cellular receptor, heparan sulfate proteoglycan (HSPG), was identified by Summerford C. and Samulski R.J. ^[41][53], and the amino acid residues providing its affinity to HSPG were suggested, afterwards, as R585 and R588 ^[42][54]. Accordingly, rAAV2 can be purified using heparin column affinity chromatography ^[43][55]. Nevertheless, the binding of AAV2 to its primary receptor was found to be insufficient for cell entry, so several coreceptors were later identified for it ^[40][52], including the human fibroblast growth factor receptor 1 (FGFR1) ^[44][56], αVβ5 ^[45][57] and α5β1 integrins ^[46][58], hepatocyte growth factor receptor (HGFR) ^[47][59], laminin receptor (LR) ^[48][60], and CD9 ^[49][61]. The capsid of rAAV2 is reported to acquire multiple PTMs, including ubiquitination, phosphorylation, SUMOylation, and multiple-site-glycosylation ^[37][41].

2.3. AAV3

4.3. AAV3

The third serotype of AAV (AAV3) was originally isolated from humans ^[30][39][34,45]. Similar to AAV2, this serotype uses HSPG, FGFR1, and LR receptors ^[48][50][60,74], along with the human HGFR (hHGFR) receptor ^[48][60]. Iodixanol gradient ultracentrifugation along with ion exchange chromatography have been used for AAV3 purification ^[51][75]. PTMs of rAAV3 capsid include acetylation, phosphorylation, and glycosylation ^[37][41]. Due to its inadequate transduction efficiency in vitro and in murine cell lines, AAV3 was mostly overlooked as a choice for gene therapy ^[52][76]. However, as it has been later found to use hHGFR as a coreceptor, it showed extremely efficient transduction of human liver cancer cells as well as human and NHP hepatocytes ^[52][53][76,77].

2.4. AAV4

4.4. AAV4

The fourth serotype of AAV (AAV4) is considered one of the most antigenically distinct serotypes ^[54][83]. It has reportedly originated in NHPs ^[55][84], mainly in green African monkeys ^[56][85], as antibodies to its viral particles have been detected in their sera ^[57][58][86,87]. A study of the AAV4 structure showed that its capsid surface topology shares a significant similarity with that of human parvovirus B19 and Aleutian mink disease virus ^[54][83]. AAV4 uses a-2,3-O-linked sialic acid for cell binding and infection ^[59][88]; accordingly, mucin columns can be used for AAV4 purification, based on its ability to bind sialic acid residues in mucin ^[29][36][33,40]. In addition, as this serotype lacks heparin-binding activity, it cannot be purified using heparin column affinity chromatography like AAV2, however, ion-exchange chromatography procedures have been developed and proven a high purification efficiency ^[60][89].

2.5. AAV5

4.5. AAV5

As it was first isolated in 1983 from male genital lesions, AAV5 became the only AAV serotype to be isolated directly from a human tissue ^[61][93]. This serotype is considered the most genetically divergent of all AAVs ^{[2][62][63][64]}[6,94,95,96], with a variety of unique characteristics, such as the distinct size and function of its ITR regions ^[18][22], utilizing herpes simplex virus (HSV) as its helper virus for human infections ^[62][94], and using an atypical endocytic route as a pathway for viral entry ^[29][33]. Another distinctive feature of AAV5 is its ability to transduce cells that cannot be transduced with AAV2, an exclusive advantage for gene therapy uses ^[65][66][67][97,98,99]. AAV5 was also found to use sialic acid as its primary receptor ^[59][68][69][88,100,101], along with platelet-derived growth factor receptors (PDGFR) α and β as coreceptors ^[70][71][102,103].

2.6. AAV6

4.6. AAV6

The classification of the sixth serotype of AAV (AAV6) is still a matter of controversy, as it presents high genomic similarity with both AAV1 and AAV2 serotypes, however, it has still been assigned its own serotype numbering ^[29][31][33,35]. AAV6 has a serological profile almost identical to that of AAV1, and shares its sequence of coding region with a homology percentage of 99%, along with multiple regions identical to those of AAV2 ^[72][113]. Accordingly, it was suggested to be a naturally occurring hybrid resulting from homologous RECOMation between AAV1 and AAV2 ^{[31][39][72][73]}[35,45,113,114]. AAV6 was first isolated from a human Ad preparation ^[39][72][74][45,113,115], and similar to AAV1, was found to bind sialylated proteoglycans, mainly α2,3-/α2,6-linked sialic acid, as its primary receptor, as well as binding heparan sulfate ^[32][75][76][36,116,117]. As for its coreceptor, it binds epidermal growth factor receptor (EGFR) ^[77][118]. The only reported PTM of rAAV6 is acetylation of its capsid proteins ^[37][41].

2.7. AAV7

4.7. AAV7

The seventh serotype of AAV (AAV7) was first isolated in 2002 from NHP tissues, specifically, from Rhesus macaque monkeys ^[31][78][35,129]. Its mechanisms of cell binding and cell entry are still unknown ^[40][79][52,130], but it is established that this serotype does not bind heparin, or any other glycan in general ^[80][131]. Capsid proteins of rAAV7 undergo multiple PTMs, including glycosylation, primarily, along with phosphorylation, SUMOylation, and acetylation ^[37][41]. A study by Calcedo R et al., investigated the epidemiology of AAV-neutralizing antibodies in the worldwide population, and found that the seroprevalence of AAV7 antibodies is relatively low in humans, an advantage of this serotype to be used in clinical applications ^[81][132]. Viral vectors based on AAV7 proved a high efficiency of transduction for skeletal muscle cells in murine models, similar to that achieved by AAV1, and higher than AAV2 ^[82][133]. This serotype also proved a strong tropism to hepatocytes in murine ^[83][92] and human ^[83][92] tissues. In CNS of NHPs, AAV7 viral vectors were found to achieve a robust transduction mainly in cortical and spinal tissues ^[84][134].

2.8. AAV8

4.8. AAV8

Similar to AAV7, the eighth serotype of AAV (AAV8) was first isolated in 2002 from Rhesus macaque monkeys ^[31][78][35,129]. As a primary receptor, AAV8 binds LR, the same receptor used by AAV2 and AAV3 ^[39][48][45,60]. Various procedures for rapid and scalable purification of AAV8 have been developed since its discovery, including, for example, dual-ion-exchange chromatography ^[85][86][137,138], or iodixanol gradient centrifugation ^[87][88][139,140]. Phosphorylation, glycosylation, and acetylation are the three PTMs reported for rAAV8 capsid proteins ^[37][41]. AAV8 is best known for its strong tropism to liver cells, and accordingly, its transduction efficiency of hepatocytes, which is far stronger and faster than those of all other AAV serotypes in different models, including murine, canine, and NHP ^{[83][89][90][91][92][93][94][95][96]}[92,112,141,142,143,144,145,146,147].

2.9. AAV9

4.9. AAV9

The ninth serotype of AAV (AAV9) was first identified in a human isolate in 2004, and was named a new serotype as it had a serological profile distinct from the previously known AAVs, however, it was suggested to be closely related to clades containing AAV7 and AAV8 ^[78][97][129,155]. As a primary receptor, AAV9 uses terminal N-linked galactose ^[98][99][156,157], and it is also suggested to bind a putative integrin, along with LR as coreceptors ^[48][98][60,156]. Scalable simple purification protocols have been developed for AAV9 purification, including ion-exchange chromatography ^[20][24] and sucrose gradient centrifugation ^[100][158]. Capsid of rAAV9 has one of the highest totals of PTMs, including multiple ubiquitination, phosphorylation, SUMOylation, and glycosylation modifications, along with acetylation ^[37][41]. In most tissues, AAV9 seems to achieve cell transduction with efficiency superior to other AAVs ^[29][78][33,129]. For example, in a study aimed to investigate AAV1–9 distribution following systemic delivery in a murine model, AAV9 has shown rapid-onset, the best genome distribution, and the highest protein levels, in comparison with all other AAVs ^[83][92].

2.10. AAV10 and AAV11

4.10. AAV10 and AAV11

The ninth and tenth serotypes of AAV (AAV10 and AAV11) were first found and described in 2004 in NHP isolates, namely from cynomolgus monkeys, with capsid proteins of great resemblance to AAV8 and AAV4, respectively ^[30][101][34,177], resulting in serological cross-reactivity with those two serotypes ^[102][178]. However, antisera against AAV10 and AAV11 were not found to have any cross-reactivity against those of AAV2, which recommended them as good viral vector candidates for gene therapy in individuals having antibodies against the latter ^[29][33]. It remains unknown what cellular receptors and coreceptors AAV10 and AAV11 use for cell binding and entry ^[33][74][37,115]; therefore, procedures describing their purification protocols are generally based on iodixanol gradient centrifugation ^[103][179]. Similar to AAV9, capsid of rAAV10 has one of the highest totals of PTMs, including mainly multiple glycosylation and phosphorylation modifications, along with ubiquitination, SUMOylation, and acetylation ^[37][41].

2.11. AAV12

4.11. AAV12

The twelfth serotype of AAV (AAV12) was first isolated from a simian Ad stock, and then characterized as a novel serotype, as it exhibited distinctive biological and serological properties ^[104][185]. Although it was proven that AAV12 does not use heparan sulfate proteoglycans or sialic acids for attachment and cell entry, it remains unknown how exactly it binds target cells ^[101][104][177,185]. However, according to a study investigating components of a potential receptor complex for AAV12, mannose and mannosamine were suggested as components of such complex, as they inhibited AAV12 cell transduction ^[105][186].

The thirteenth serotype of AAV (AAV13) is another simian Ad that appears to bind HSPG, although its primary cell receptor remains unknown ^{[106][107][108]}[189,190,191]. It was also found to share structural similarity with AAV2 and AAV3, making it the closest related AAV to those two serotypes, with a capsid conserving all AAV capsids’ structural features ^[109][188], but there are only limited data on this serotype tropism and transduction efficiency ^[108][191].

In addition to the natural AAV serotypes described above, novel AAV vectors have been developed during the last two decades, and are still being developed ^[110][192]. Using different engineering strategies, novel hybrid vectors have been generated in order to enhance their transduction, modulate their immunogenicity, or limit their tropism to specific cells ^{[111][112][113][114][115]}[193,194,195,196,197]. There has been different types of such engineered novel vectors, including mosaic, chimeric, and combinatorial vector libraries ^[29][33]. Mosaic vectors have multiple subunits of various serotypes in their capsid, chosen according to their properties of receptor binding and intracellular trafficking ^[29][116][33,198].

3. AAV as Viral Vectors for Gene Therapy Applications

3.1. AAV Viral Vectors for Gene Therapy of the CNS

5.1. AAV Viral Vectors for Gene Therapy of the CNS

AAV-mediated CNS gene therapy was first believed to predominantly target neurons, with lesser chance to affect other cells and tissues in the CNS ^[117][224], a fact that was later proven to be inaccurate in many studies ^{[84][118][119][120]}[134,160,214,225]. For a variety of neurodevelopmental and neurodegenerative diseases, AAV-mediated gene therapy has been tested using different administration routes, including intraparenchymal, intrathecal, intracerebroventricular, and intracisternal injection, many of which have shown promising results ^[121][226]. Parkinson’s disease (PD) has been one of the most studied neurological disorders as a target for AAV-mediated gene therapy ^{[117][122][123]}[224,227,228]. In a phase I clinical trial, Kaplitt et al., investigated an AAV-mediated gene therapy approach for advanced PD patients, where serotype 2 was used as a vector for unilateral subthalamic delivery of the glutamic acid decarboxylase (GAD) gene ^[124][229]. Providing a significant improvement in motor function scores up to 12 months after surgery, the approach was found to be safe and well-tolerated in patients. A similar double-blind, controlled, randomized clinical trial conducted later by LeWitt et al., investigated the same AAV2-GAD vector for bilateral subthalamic delivery in advanced PD patients, and showed similar results of safety and improvement of motor function ^[125][230]. Bartus et al., also tested the bilateral stereotactic delivery of AAV2-neurturin in PD patients of an open-label clinical trial, the initial obtained data of which supported the feasibility, safety, and good tolerance of the approach as a potential treatment for PD ^[126][231]. However, bilateral intra-striatal infusion of an AAV2 vector containing the aromatic L-amino acid decarboxylase (AADC) gene in moderately advanced PD patients led to an improvement in PD rating scales that was associated with a risk of intracranial hemorrhage in patients, along with headaches ^[127][232]. As for sustainable transgene expression, a clinical trial conducted by Mittermeyer G. et al., investigated the potentials of rAAV2 carrying the aromatic L-amino acid decarboxylase gene (rAAV2-AADC) ^[128][233]. In the trial, 10 patients with moderately advanced PD received bilateral infusions of recombinant vector into the putamin, showing good tolerance and a stable expression of transgene that lasted for the following 4 years, although higher vector doses were suggested for further studies. In children with AADC deficiency, the same delivery route for the same vector (rAAV2-AADC) was also assessed in an open-label phase I/II clinical trial ^[129][234]. The therapy was well-tolerated in general, providing evidence for potential improvement of motor function.

3.2. AAV Viral Vectors for Gene Therapy of Respiratory Diseases

5.2. AAV Viral Vectors for Gene Therapy of Respiratory Diseases

For over 20 years, AAV vectors have been vastly investigated for gene therapeutics of respiratory diseases, both in preclinical experiments and human clinical trials ^[130][131][283,284]. However, a key limitation was that many AAV serotypes cannot efficiently transduce airway epithelial cells through the apical surface, suggesting molecular modifications of such serotypes to enhance their transduction efficiency ^[114][196]. Cystic fibrosis (CF) represents one of the most studied diseases as a target for AAV-mediated gene therapy ^[132][285]. In rabbits, AAV has proven to promote an efficient and stable gene transfer of the cystic fibrosis transmembrane conductance regulator (CFTR) gene into the airway epithelium, indicating, as a result, the vector potential to be used for gene therapy ^[133][286]. Subsequently, since 1998, clinical trials have started using recombinant AAV viral vectors to target different sites of the airway epithelium of CF patients, using variant administration routes ^{[132][134][135]}[285,287,288]. The very first clinical trial to do so used the maxillary sinuses for delivery of recombinant AAV2 containing the CFTR gene (AAV2-CFTR) in ten CF patients, demonstrating a safe, successful transduction of targeted cells and a detected function restoration of the sinuses ^[135][288]. Soon after, a following phase II double-blind clinical trial tested unilateral administration of the same vector into the sinus for 23 CF patients, with an in-patient control, achieved by administering a placebo drug into the other sinus ^[136][289]. The approach again showed safety and good tolerance, although it did not confirm clinical effectiveness of the treatment. Aitken et al., also tested AAV2-CFTR in a phase I clinical trial for twelve mild CF patients using aerolization by nebulation for delivery, confirming the approaches safety, but failing to yield an effective clinical treatment ^[137][290]. Further trials on AAV2-CFTR with single or repeated dosing have yielded similar results of safety and good tolerance, providing little evidence of clinical treatment after intranasal and endobronchial delivery ^{[138][139][140]}[291,292,293]. In order to optimize such therapeutic approaches and produce a functional CFTR in CF patients, further modifications of recombinant AAV vectors have since been developed ^[141][142][294,295]. Alpha-1 antitrypsin (α1AT or AAT) deficiency is another lung and liver disease that has been extensively studied as a target for AAV-mediated gene therapy both in vivo and in humans ^[131][143][284,296]. Reportedly, intravenous delivery of recombinant AAV vector carrying the human alpha-1 antitrypsin gene (AAV-hAAT) in the murine model resulted in potentially therapeutic serum levels of AAT ^[144][145][297,298]. Further research found that intrapleural delivery of rAAV2-hAAT or rAAV5-hAAT could achieve higher AAT levels both in lungs and serum compared to intramuscular delivery in C57BL/6 mice, with the rAAV5 showing 10-fold higher effectiveness than rAAV2 ^[146][147][299,300]. However, rAAV6/2-hAAT has been shown to transduce murine lung cells even more efficiently than rAAV5 both in vivo (in murine lung cells) and in vitro (in human airway epithelial cell culture) ^[148][301].

3.3. AAV Viral Vectors for Gene Therapy of Muscle Diseases

5.3. AAV Viral Vectors for Gene Therapy of Muscle Diseases

Having a large, body-distributed mass, and myofibers with a long half-life time makes muscles an attractive target for gene therapy, also considering the minimal invasiveness of the intramuscular delivery route ^[82][149][133,309]. Muscular dystrophies (MD) constitute a large group of muscle diseases that have been studied for a long time and proved to be a good target for AAV-mediated gene therapy ^[150][310]. For example, AAV1 carrying the follistatin gene, an antagonist of muscle growth negative regulator, has been shown to promote sustained improvement in muscle size and strength in NHPs following intramuscular administration ^[151][311]. There has also been many reports of both histopathological and functional sustained muscle restoration as a result of using AAV-mediated therapeutic approaches in animal models of Duchenne muscular dystrophy (DMD) ^[152][153][312,313], and Limb-Girdle muscular dystrophies (LGMD) ^{[154][155][156][157][158]}[314,315,316,317,318].

3.4. AAV Viral Vectors for Gene Therapy of Cardiovascular and Blood Diseases

5.4. AAV Viral Vectors for Gene Therapy of Cardiovascular and Blood Diseases

Cardiovascular diseases make another attractive target for gene therapy, being a leading cause of death globally with their high incidence and mortality rates ^[159][160][128,321]. As mentioned before, AAV serotypes 6, 8, and 9 have been shown to efficiently transduce cardiac cells in different animal models, and accordingly, they have been used for therapeutic purposes in different cardiovascular diseases ^{[159][161][162][163][164]}[126,128,167,168,322]. In the search for an effective gene therapy platform for heat failure (HF), White J et al., published a study in 2011, in which they suggested a novel technique for AAV-mediated myocardial gene therapy using molecular cardiac surgery ^[159][128]. Using AAV6 in an ovine model, the suggested approach resulted in a global transgene expression, that was cardiac-tropic and substantially more robust and targeted, compared to that of intramuscular or intracoronary injection. Furthermore, a randomized phase I/II AAV1-based clinical trial for heart failure treatment was conducted in 2013, where the researchers used a sarcoplasmic reticulum calcium ATPase gene (SERCA2a), the product of which was suggested to play a key role in HF pathology ^[165][323]. As a result, adverse effects, including death, were found to be highest in the placebo group, and lowest in the high-dose group, with evidence of long-term transgene expression. However, in the low-dose and mid-dose groups, adverse effects were also found to be high but delayed.

3.5. AAV Viral Vectors for Gene Therapy of Liver Diseases

5.5. AAV Viral Vectors for Gene Therapy of Liver Diseases

AAV-based gene therapy plays a significant role in liver diseases, as it is, in some cases, an alternative to the only effective therapy, which is liver transplantation ^[166][324]. In addition to the previously mentioned diseases affecting the liver, that have been targeted by AAV-base gene therapy, such as CF, AAT deficiency, and hemophilia, there are other liver diseases that have been targeted and continue to be, using AAV8, mainly. Wilson’s disease (WD) is a rare autosomal recessive disease caused by mutations in the copper transporter gene, ATP7B, resulting in copper accumulation mainly in the liver, along with some other organs ^[167][168][325,326]. A recent in vivo study, conducted by Murillo et al., proved the efficacy of an AAV-based gene therapy for WD in a murine model of the disease ^[169][327].

3.6. AAV Viral Vectors for Gene Therapy of Endocrine Disorders

5.6. AAV Viral Vectors for Gene Therapy of Endocrine Disorders

Being a major health issue for humans all around the world, with relatively high prevalence estimates, endocrine disorders have always been in the focus of research for therapeutic approaches and techniques ^[170][171][331,332]. Accordingly, rapid developments of diagnostic techniques, along with the better understanding of endocrine disorders’ pathophysiology and molecular bases, helped in developing different therapies, such as hormonal replacement therapies, for example ^[171][172][332,333]. However, as these approaches have mostly focused on improving the patient’s quality of life, reducing or reversing symptoms, rather than targeting the underlying defect, the resulting effect has not always been sufficient, which highlighted the need for gene therapy approaches, including AAV-based ones ^[172][333]. Type 1 diabetes mellitus (T1DM) is one of the endocrine autoimmune disorders that has been extensively studied as a target of gene therapy ^[173][334]. T1DM is characterized by self-destruction of insulin-secreting islet β cells, resulting from a wide variety of causative factors ^[174][335]. In non-obese diabetic (NOD) mice, administration of recombinant adeno-associated virus serotype 8 carrying DNA of mouse insulin promoter (dsAAV8-mIP) has been shown to prevent hyperglycemia in a dose-dependent manner ^[175][336]. High levels of mouse interleukin 10 (mIL-10) achieved following rAAV2-IL-10 intramuscular administration also proved to have a positive effect in NOD mice by decreasing autoimmunity, and thereby hyperglycemia ^[176][337]. Similarly, negative regulation of the immune response by programmed death ligand 1 (PDL1) was achieved in NOD mice following intraperitoneal delivery of AAV8-PDL1, which protected β cells ^[177][338].

3.7. AAV Viral Vectors for Gene Therapy of Cancer

5.7. AAV Viral Vectors for Gene Therapy of Cancer

Despite the variety of therapeutic approaches developed so far for cancer, including chemotherapy, radiotherapy, surgery, and different medications, cancer is still a huge health issue and a leading cause of death worldwide ^[178][179][342,343]. Several AAV-mediated cancer gene therapy approaches have been reported so far, including suicide gene, RNA-interference, and anti-angiogenesis gene therapies ^[178][342]. In AAV-mediated suicide gene therapy, an AAV-based system with herpes simplex virus type-1 thymidine kinase and ganciclovir (AAVtk/GCV) has been used. This system was found to exhibit a significant in vitro and in vivo tumor-suppressor efficacy in human head and neck cancer xenografts in a murine model ^{[180][181][182]}[344,345,346], and murine models of bile duct cancer ^[183][347] and bladder carcinoma ^[184][348].