Original Article Seven-year follow-up of durability and safety of AAV CNS gene therapy for a lysosomal storage disorder in a large animal Sara Marcó,1,2,5,6 Virginia Haurigot,1,2,5,6 Maria Luisa Jaén,1,2,5 Albert Ribera,1,2,5 Víctor Sánchez,1,2 Maria Molas,1,2,5 Miguel Garcia,1,2,5 Xavier León,1,2,5 Carles Roca,1,2,5 Xavier Sánchez,1,2 Joan Bertolin,1,2 Jennifer Pérez,1,2 Gemma Elias,1,2 Marc Navarro,3 Ana Carretero,3 Martí Pumarola,3 Anna Andaluz,3,4 Yvonne Espada,3,4 Sonia Añor,3,4 and Fatima Bosch1,2,5 1Center of Animal Biotechnology and Gene Therapy, Edifici H, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain; 2Department of Biochemistry and Molecular Biology, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain; 3Department of Animal Medicine and Surgery, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain; 4Hospital Clínic Veterinari, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain; 5CIBER de Diabetes y Enfermedades Metabólicas Asociadas (CIBERDEM), 28029 Madrid, Spain Delivery of adeno-associated viral vectors (AAVs) to cerebro- Sanfilippo syndrome type A. The Sanfilippo syndrome comprises spinal fluid (CSF) has emerged as a promising approach to a group of monogenic, autosomal recessive neurodegenerative dis- achieve widespread transduction of the central nervous system orders caused by mutations in enzymes involved in the stepwise (CNS) and peripheral nervous system (PNS), with direct appli- degradation of the glycosaminoglycan (GAG) heparan sulfate cability to the treatment of a wide range of neurological dis- (HS).7 As a consequence, undegraded HS accumulates in the lyso- eases, particularly lysosomal storage diseases. Although studies somes of cells, causing cell dysfunction and, eventually, cell in small animal models have provided proof of concept and ex- death.7,8 Although the enzyme whose deficiency results in each periments in large animals demonstrated feasibility in bigger subtype of Sanfilippo disease (types A–D) is different, all 4 sub- brains, there is not much information on long-term safety or types share common pathophysiological underlying mechanisms durability of the effect. Here, we report a 7-year study in as well as similar clinical signs and prognosis.7 MPSIII is consid- healthy beagle dogs after intra-CSF delivery of a single, clini- ered mainly a neurodegenerative disease, with progressive cogni- cally relevant dose (2  1013 vg/dog) of AAV9 vectors carrying tive and motor impairment that mainly correlate with cortical the canine sulfamidase, the enzyme deficient in mucopolysac- and cerebellar atrophy accompanied by a compensatory enlarge- charidosis type IIIA. Periodic monitoring of CSF and blood, ment of the lateral ventricles in brain magnetic resonance imaging clinical and neurological evaluations, and magnetic resonance (MRI) images.9–13 There is, however, mild somatic involvement, and ultrasound imaging of target organs demonstrated no with hepato- and splenomegaly, recurrent ear and upper respira- toxicity related to treatment. AAV9-mediated gene transfer re- tory tract infections, frequent diarrhea, and facial dysmorphisms. sulted in detection of sulfamidase activity in CSF throughout Disease progresses with age, leading to the death of affected indi- the study. Analysis at tissue level showed widespread sulfami- viduals, generally in their second decade of life.7,14 dase expression and activity in the absence of histological find- ings in any region of encephalon, spinal cord, or dorsal root In the case of lysosomal storage disorders (LSDs), most therapeutic ganglia. Altogether, these results provide proof of durability strategies rely on the principle of cross-correction, where soluble lyso- of expression and long-term safety for intra-CSF delivery of somal enzymes with mannose-6-phosphate (M6P) residues present in AAV-based gene transfer vectors encoding therapeutic pro- the extracellular compartment can be taken up by M6P receptor- teins to the CNS. mediated endocytosis into neighboring cells. Hence, transduction of large proportions of cells with gene therapies is not necessary to achieve full therapeutic efficacy. INTRODUCTION The development of central nervous system (CNS)-targeted gene therapies for the treatment of neurological disorders is a very Received 16 March 2021; accepted 29 September 2021; active field of research in both academia and industry.1–6 CNS-tar- https://doi.org/10.1016/j.omtm.2021.09.017. geted gene therapy is a modality of particular interest for those 6 These authors contributed equally indications for which access of the target to the CNS from the cir- Correspondence: Prof. Fatima Bosch, Center of Animal Biotechnology and Gene culation is limited. Our initial interest in the field of CNS gene Therapy, Edifici H, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain. therapy was for mucopolysaccharidosis type IIIA (MPSIIIA) or E-mail: fatima.bosch@uab.es 370 Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 ª 2021 The Author(s). This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). www.moleculartherapy.org Different ways of delivering genes to the CNS in vivo have been ations of efficacy and safety. Therefore, it is ultimately data from described so far. The simplest approach is based on the intrave- larger animal species that best inform product behavior and distri- nous (i.v.) administration of adeno-associated viral vectors bution, in particular durability and long-term safety. In a previous (AAVs) because of the ability of certain serotypes, such as serotype short-term feasibility study in healthy dogs, we demonstrated that 9 (AAV9), to cross the blood-brain barrier (BBB) and transduce our approach was scalable to larger brains with a surgical proced- the CNS.15,16 This method is used to deliver Zolgensma, an ure—intracerebroventricular delivery—that is standard practice in FDA- and EMA-approved gene therapy that targets motor neu- pediatric neurosurgery.34,46 In the present study, we report 7 years rons of the spinal cord for the treatment of spinal muscular atro- of follow-up of 3 healthy dogs injected in the CSF via cisterna ma- phy in pediatric patients.17 However, i.v. administration may not gna with canine Sgsh-encoding AAV9 vectors at a clinically rele- be the route of administration (ROA) of choice to reach the vant dose. Our study demonstrates that the procedure is safe CNS in other neurodegenerative diseases that require widespread and results in high levels of SGSH activity in CSF detectable 7 transduction of the CNS, or for the subset of patients that are sero- years after a single vector delivery, in the absence of any adverse positive for anti-AAV antibodies at high titers.18 Moreover, the events. Transgene expression and activity were detectable within high doses required for efficacy when using this ROA8,19–21 may the CNS and the peripheral nervous system (PNS) at the end of trigger immunological responses against vector capsid.22 Clinical the study. Vector persistence and transgene expression were also trials with systemically delivered AAVs are ongoing for MPSIIIA detectable in liver. To the best of our knowledge this is the longest and MPSIIIB (NCT02716246, NCT04088734, NCT04360265, and safety and durability follow-up in a large animal species after CNS- NCT03315182). directed gene transfer. Vectors encoding for lysosomal enzymes can also be delivered RESULTS directly to the CNS by intraparenchymal injection or by administra- General tolerability and safety of intra-CSF AAV9-sulfamidase tion into the cerebrospinal fluid (CSF). The first method, currently administration under clinical testing for MPSIIIA (NCT01474343, NCT02053064, Three male healthy beagle dogs (dogs 1–3) were administered 2  and NCT03612869) and MPSIIIB (NCT03300453),23,24 requires a 1013 vg (2.78  1011 vg/ mL of dog brain) of AAV9 vectors carrying complex surgical procedure that involves the deposit of small volumes an optimized version of canine sulfamidase coding sequence under of AAV at two depths/site through 6–8 burr holes to cover a signifi- the control of the CAG promoter (AAV9-Sgsh) through cisterna ma- cant proportion of brain volume.23–25 However, the limited diffusion gna injection.34 We previously reported the 3-month follow-up for of the vector from the site of injection when delivered to most CNS these dogs.34 Here, we extend this original observation up to a total regions other than the thalamus or striatum26–30 and the restricted of 82 months (7 years). number of injections that can be performed safely have resulted in limited proof of efficacy in clinical trials.23,24,31–33 Seven years after intra-CSF AAV9-Sgsh administration, treated dogs remained clinically well, with no signs of adverse events. CSF samples Alternatively, we and others have demonstrated in several animal were tested regularly for elevations in white blood cell (WBC) counts models the advantages of delivering vectors, in particular AAV9, or total protein (TP) levels as indicators of CNS inflammation. Red directly into the CSF to achieve efficient and widespread CNS trans- blood cells (RBCs) were counted in parallel to monitor for potential duction.34–42 Whereas intraparenchymal administration results in blood contamination during sample withdrawal. CSF WBC and TP uneven distribution of AAV vectors, with very high vector genomes levels remained within normal range throughout the study (Table at the point of administration quickly dropping with distance,43 1). Only 1 in >50 CSF samples obtained from all 3 dogs showed intra-CSF delivery of AAV9 vectors ensures widespread, even distri- WBC counts above the upper limit of normal (Table 1). This obser- bution of transduced cells throughout the brain and spinal cord.34,42 vation was made in a sample obtained from dog 1 51 months post- Additional enhancement of therapeutic protein distribution is vector delivery, had only a 2 magnitude, and was accompanied by achieved for secretable transgenes, since transduced cells secrete pro- normal TP and no symptomatology (Table 1). Likewise, 3 samples teins to the CSF efficiently.34,35,39,44 Furthermore, after intra-CSF presented TP levels slightly above the 25 mg/dL limit but showed administration, part of the vector passes from the CSF to the circula- no increase in WBCs, and dogs had no symptomatology (Table 1). tion and transduces the liver, which can secrete the therapeutic pro- Together with our previous monitoring throughout the first 3 months tein into the bloodstream.34–36,39,44,45 post-vector administration,34 the present observation extending up to 82 months post-test article delivery argue against the development of Using an AAV9 vector encoding sulfamidase (Sgsh), the enzyme subclinical acute or chronic inflammatory processes in the CNS of the deficient in MPSIIIA, we were the first to report whole-body dogs injected with vectors encoding species-specific sulfamidase. This correction of a LSD after delivery of AAVs to the CSF of mice.34 is in clear contrast with what has been observed previously in dogs Although rodents provide an invaluable tool to perform proof- and non-human primates (NHPs) administered vectors encoding hu- of-concept studies, anatomical/physiological species differences man transgenes34,47 and underscores the importance of using species- impose several limitations on non-clinical development. For specific transgenes for safety studies, particularly when evaluating example, the short lifespan of rodents precludes long-term evalu- long-term tolerance to a treatment. Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 371 Molecular Therapy: Methods & Clinical Development Table 1. CSF cell counts and TP levels after intra-CSF delivery of 2  1013 vg of AAV9-Sgsh to healthy beagle dogs Dog ID Measure (reference values) Months post-injection 4 6 8 10 12 16 20 24 28 32 38 46 51 61 65 72 82 RBC (0 cells) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Dog 1 WBC (<5 cells/mL) 0 0 1 1 0 1 0 0 2 NA 1 0 11 0 0 4 0 TP (<25 mg/dL) 19.3 24.8 22.7 20.2 19 18.2 20.2 20.6 22.9 21.4 21 23 21.5 26.6 26 20.4 22.2 RBC (0 cells) 59* 0 0 0 0 0 0 0 0 NA 0 0 0 0 0 0 0 Dog 2 WBC (<5 cells/mL) 1* 1 4 0 5 1 1 1 1 NA 0 0 1 5 0 0 0 TP (<25 mg/dL) 19.4* 22.8 21.6 19.1 21.1 22.6 23.2 20.9 21.9 NA 20.5 22.2 21.3 21.2 24 20.7 25.9 RBC (0 cells) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Dog 3 WBC (<5 cells/mL) 0 1 2 0 1 1 1 0 1 NA 1 0 0 1 0 0 0 TP (<25 mg/dL) 19.6 18.9 19.4 21.7 18.9 22.6 21.7 24.6 20.5 17.4 20.1 21.3 20.9 21.8 24.2 21 20.9 *Sample contaminated with blood. White blood cell (WBC) and total protein (TP) values were corrected taking into account blood contamination. RBC, red blood cell; NA, sample not available. The values outside the reference interval are italicized. To further assess potential toxicities, blood tests were regularly per- transverse T2-weighted scans were obtained for all dogs 60 and formed. Although several isolated deviations in clinical chemistry 82 months after vector delivery; representative images are shown in or hematological parameters were observed in different samples Figure S3 and Figure 1B. Similar to the encephalon, no abnormalities from all 3 dogs throughout the 7-year follow-up, likely reflecting sub- could be identified in any of the dogs after systematic analysis of all optimal sample quality, abnormalities were considered either to be of the images obtained. clinically irrelevant magnitude or not to follow a pattern compatible with test article- or procedure-related toxicities, which agreed with In addition, comprehensive abdominal ultrasounds were performed the general health status of the animals (Tables S1 and S2). 34 months after intra-CSF AAV9-Sgsh vector injection. Special focus was placed on the analysis of the liver, the main peripheral organ CNS and liver imaging and functional neurological evaluation transduced by AAV9 vectors after delivery to the CSF of small and As part of the long-term safety evaluation, MRI scans of the enceph- large animal models.34–37,39,40,45 Liver parenchyma was homogeneous alon and the spinal cord, as well as abdominal ultrasounds, were and presented normal echogenicity in the 3 dogs. Livers were of obtained at different time points after vector administration in all in- normal size and echostructure, and there was no evidence of alter- jected dogs. ations in hepatic vascularization or the blood flow through the portal vein. Brain MRI scans were performed 34 (Figure S1) and 82 (Figure 1A) months after vector delivery, in which dorsal T2-weighted images All animals remained clinically well throughout the whole follow-up and transverse and sagittal T1-weighted images after intravenous period. To further evaluate the possibility of toxicities specifically in administration of the contrast agent gadolinium were acquired for the nervous system, full neurological evaluations were performed in all dogs. Systematic analysis of the scans revealed no abnormalities all dogs by a diplomate veterinary neurologist 52 and 82 months after in the encephalon of any of the AAV-injected dogs in any of the AAV9-Sgsh delivery to the CSF (Table S3). The neurological exami- MRI sequences studied. Scans showed no evidence of mass effect, nations included observation of the mental status, posture, and gait asymmetries, or significant alterations in signal intensity at any of as well as evaluation of postural reactions, spinal reflexes, and cranial the time points studied (Figure 1A; Figure S1) and were indistinguish- nerves and sensory evaluation (pain sensation and detection of areas able from those of healthy uninjected dogs (Figure S2). After admin- of hyperesthesia) (Table S3). In consonance with the MRI results, no istration of the contrast, no signs of enhancement could be observed signs of diffuse or focal lesions were observed in any of the AAV9-in- in any region of the encephalon or the meninges in any of the dogs jected dogs. (Figure 1A; Figures S1 and S2). Time course of sulfamidase activity in the CSF and serum In a previous study, we observed that after delivery of AAV9 vectors As previously reported,16 a clear increase in sulfamidase activity to the CSF of dogs through either cisterna magna or lateral ventricle over basal levels was detectable in the CSF of all 3 dogs as soon injection, the lumbosacral intumescence and the cauda equina were as 1 week after vector administration. The time to peak activity after the sections of the spinal cord with the highest levels of transgene vector delivery varied depending on the dog. Dogs 1 and 3 showed a expression.34 Hence, we focused on these regions of the spinal cord rapid raise in enzymatic activity, which peaked 2 and 4 weeks post- for the evaluation through MRI. Sagittal and transverse T1-weighted injection, respectively (Figure 2A).34 Dog 2 showed a slower (pre- and post-contrast administration) and dorsal, sagittal, and kinetics of transgene expression and had peak values of SGSH 372 Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 www.moleculartherapy.org Figure 1. Preserved structure of the encephalon and the spinal cord after intra-CSF AAV9-Sgsh delivery Adult beagle dogs (dogs 1–3) received a dose of 2  1013 vg of AAV9 vectors encoding canine sulfamidase through intracisternal administration. (A) Representative images of the magnetic resonance imaging (MRI) of the encephalon of dogs 1–3 performed 82 months after vector delivery. Scans were performed with a 0.2-T permanent open magnet system. Left and middle: transverse and sagittal T1-weighted images, respectively, obtained after intra- venous administration of a gadolinium-containing contrast agent. Right: dorsal T2-weighted images ob- tained prior to contrast injection. No abnormalities were observed in the encephalon of any of the AAV9-Sgsh- injected dogs in any of the MRI sequences. (B) Repre- sentative images of the MRI analysis of the spinal cord performed 82 months after vector delivery, focused on the lumbosacral intumescence and the cauda equina, as previous work had determined that these were the por- tions of the spinal cord most efficiently transduced after intra-CSF administration of AAV9 vectors to dogs. A 0.4-T scanner was used to obtain dorsal (left), sagittal (middle), and L2–L3 transverse (right) T2-weighted images. Similar to the encephalon, no abnormalities could be identified in any of the dogs after systematic analysis of all the images obtained. Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 373 Molecular Therapy: Methods & Clinical Development Figure 2. Follow-up of sulfamidase activity in the CSF of AAV9-Sgsh-injected dogs (A) CSF samples were withdrawn regularly from dogs 1–3 for a period of 7 years and assayed for sulfamidase ac- tivity. All AAV9-injected dogs showed a substantial in- crease in SGSH activity in the CSF over baseline levels, and activity remained high for the whole of the follow-up period. As a reference, the basal CSF sulfamidase activity quantified in CSF samples from 15 untreated healthy beagle dogs was 0.47 ± 0.05 nmol/17 h/mL of CSF. p.i., post injection. The vertical dashed line indicates the first 3 months of treatment. (B) Fold-change increase over basal levels in CSF sulfamidase activity at the end of the study (82 months post-injection) represented for each dog. Seven years after treatment, CSF sulfamidase ac- tivity remained at least 2.4-fold higher than at baseline in all AAV9-injected dogs shown. (C) Sulfamidase activity in the CSF at the basal time point from 15 untreated healthy beagle dogs and at the end of the study (M82, 82 months post-injection) from the three AAV9-injected dogs. Re- sults are shown as mean ± SEM. A two-tailed t test was used for statistical comparison of CSF sulfamidase ac- tivity at the endpoint (82 months post-injection) with the basal CSF sulfamidase activity. ***p < 0.001. We and others have previously reported that upon administration of AAV9 vectors to the CSF of mice, cats, dogs, or NHPs, a portion of the vector passes to the circulation and trans- duces the liver, which can become a source of circulating enzyme.34–42,45 However, no obvious increase in sulfamidase activity was detected in the serum of any of the animals in- activity at 1–3 months post injection that were lower than those jected with AAV9-Sgsh, likely because of the high basal level of sulfa- observed in dogs 1 and 3 (Figure 2A).34 At the end of this first phase midase activity in the serum of healthy dogs. of the study (90 days) dogs had an average sulfamidase activity in the CSF 8.2-fold higher than the average basal levels. By 1 year Post-mortem analysis of vector biodistribution and transgene post injection all dogs had reached similar levels of SGSH activity expression in the CNS and DRG in the CSF (3.40, 2.87, and 3.83 nmol/17 h/mL CSF for dogs 1, 2, At the end of the in-life portion of the study, samples were collected and 3, respectively), which was on average 7.1-fold higher than from different areas of the encephalon (n = 30/dog), spinal cord (n = that documented in uninjected dogs (0.47 ± 0.05 nmol/17 h/mL, 8/dog, covering cervical, dorsal, and lumbar portions), and dorsal n = 15 dogs) (Figure 2A). After the first 24 months, CSF SGSH ac- root ganglia (DRGs) (cervical, dorsal, lumbar) for quantification of tivity remained fairly stable in dogs 2 and 3 for the whole of the 7- vector genome copy number (VGCN) and sulfamidase mRNA levels. year follow-up period. By year 7, dogs 2 and 3 had levels of SGSH One non-injected dog was used as negative control. Similar to what activity in the CSF that were 6.2-fold and 4.1-fold higher than base- was previously observed in shorter (3 month) studies in dogs,34 quan- line (Figures 2A and 2B). In contrast, in dog 1 CSF SGSH activity tification of VGCN demonstrated widespread distribution of the vec- progressively decreased with time, stabilized around 4 years post in- tor within the CNS and PNS ganglia (Figure 3). In the encephalon, jection, and was 2.4-fold higher than baseline at the end of the 83% and 93% of samples from dogs 2 and 3 tested positive for follow-up period (Figures 2A and 2B). The average value of CSF sul- AAV9-Sgsh vector genomes (Figure 3A). The percentage was slightly famidase activity for the 3 dogs 7 years post vector delivery is sta- lower for dog 1, in which 70% of samples had copy numbers above 0.1 tistically highly significant (p value = 0.0005) compared to the vector genomes/diploid genome (vg/dg) (Figure 3A). In the spinal average value measured in the CSF of 15 untreated dogs (3-fold cord, values were more similar among all 3 dogs; we failed to detect increase) (Figure 2C). To the best of our knowledge, these expres- vector genomes in only 1 of 10 spinal cord samples from dog 1 (Fig- sion data constitute the longest follow-up of transgene expression ure 3B). DRGs showed relatively low VGCN, and for dog 1, 2 out of 3 after delivery of AAVs to the CNS of a large animal. samples tested negative, although this observation could be limited by 374 Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 www.moleculartherapy.org Figure 3. Widespread detection of vectors within the CNS and PNS ganglia Vector gene copy number was analyzed in tissue punches from multiple regions of the encephalon (A), spinal cord (B), and dorsal root ganglia (C) obtained during necropsy of dogs 1–3. md, medial; dr, dorsal; lt, lateral; vn, ventral; cn, central; R, right; L, left; ca, cervical anterior; cc, cervicocaudal; ict, cervicothoracic intumescence; ta, thoracic anterior; tm, thoracic medial; tc, thoracic caudal; ils, lumbosacral intumescence; ceq, cauda equina; drg, dorsal root ganglia; c, cervical; t, thoracic; l, lumbar; Et, epithalamus; Th, thalamus; Hyp, hypothalamus. vg/dg, vector genomes/diploid genome. sampling, as the few DRGs collected had to be distributed among the through CSF-mediated diffusion,34 although some of these deep brain different analyses (Figure 3C). structures, such as the striatum, the epithalamus, and the thalamus, had very low to no detectable mRNA levels (Figure 4A). Expression Codon optimization of the transgene allowed for differentiation of of AAV9-Sgsh vectors also persisted throughout the years in the spi- vector-derived sulfamidase mRNA from endogenous transcripts. nal cord of all 3 dogs, from the cervical region down to the cauda Despite the slight differences in VGCN, all 3 dogs showed similar equina, on average at levels similar to those documented in the brain levels and distribution of sulfamidase expression in the encephalon, of the respective animals except for dog 3, in which they seemed to be with at least 90% of positivity among all samples analyzed (Figure 4A). slightly higher (Figure 4B). DRGs were also positive for transgene The regions positive for Sgsh expression covered the whole volume of expression (Figure 4C). Although some of the DRG samples from the encephalon, from areas close to the cisterna magna (site of injec- dog 1 had tested negative for vector genomes, expression of codon- tion into the CSF), such as cerebellum, pons, and medulla oblongata, optimized canine sulfamidase was detectable at low levels in all sam- to areas distant from the injection site, such as frontal cortex and ol- ples. Overall, these results showed persistent, widespread AAV9 factory tract (Figure 4A). Moreover, sulfamidase expression was de- vector distribution and transgene expression within the CNS and tected in deeper areas of the brain, consistent with vector distribution PNS ganglia of a large animal up to 7 years after delivery to the CSF. Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 375 Molecular Therapy: Methods & Clinical Development Figure 4. Persistent vector-derived sulfamidase expression in the CNS and PNS ganglia Optimized canine sulfamidase mRNA expression was measured in tissue punches obtained from multiple regions of the encephalon (A), spinal cord (B), and dorsal root ganglia (C) obtained from dogs 1–3 7 years after vector injection. In (A), histograms depict the relative sulfamidase expression (normalized to housekeeping gene) in all areas of the encephalon analyzed, ordered as per their location in the left hemisphere, as indicated in the schematic representation included below the histogram. md-FC, medial frontal cortex; dr-FC, dorsal frontal cortex; lt-FC, lateral frontal cortex; OT, olfactory tract; St, striatum; md-PC, medial parietal cortex; dr-PC, dorsal parietal cortex; lt-PC, (legend continued on next page) 376 Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 www.moleculartherapy.org Distribution of active sulfamidase within CNS several sections were obtained at 3 different levels of each ganglion. To fully map distribution of active sulfamidase in the CNS of dogs 7 At each level, the first section was used for H&E staining and the years after gene transfer, we measured SGSH activity in tissue extracts consecutive sections were used for immunohistochemistry against obtained from the encephalon (n = 30/dog) and the spinal cord (n = GFAP and Iba1, making a total number of 9 sections analyzed per 8/dog; covering cervical, dorsal, and lumbar portions). The enzymatic DRG, and were compared to the findings in a control, non-injected assay used cannot distinguish endogenous from vector-derived sulfa- dog by a diplomate veterinary pathologist. In all H&E sections midase. Contrary to CSF, where background SGSH activity is very from all 3 AAV-injected dogs, neuronal cell bodies of normal appear- low, allowing for sensitive monitoring of transgene expression, tissue ance and distribution and in normal numbers were observed, with no SGSH activity in the CNS is relatively high, which lowers the sensi- evidence of neuronal fragmentation or neuronophagia (Figure 6; Fig- tivity of the assay to detect increments over baseline. The brain and ure S8). GFAP reactivity was observed in satellite cells and Schwann spinal cord of 15 control animals not injected with the SGSH vectors cells, and Iba1 reactivity was very weak and limited to mononuclear were used to obtain reference values for endogenous activity for each cells scattered among the neuronal bodies and occasionally in cells region analyzed, with a sample size of at least 10 samples per region, grouped in nodules of mononuclear cells, which was a rare finding totaling >400 measurements. In agreement with biodistribution and (Figure 6; Figure S8). Figure 6 shows representative images of a lum- transgene expression data, increases in sulfamidase activity over the bar DRG obtained from dog 2; DRG toxicities have been described to levels detected in control dogs were documented in >70% of the affect mostly the lumbar region and depend on transduction levels,48 encephalon and spinal cord samples obtained from each of the in- and dog 2 was the dog with the highest levels of SGSH activity in CSF. jected dogs (Figure 5A). The magnitude of the increase over control levels varied significantly depending on the region analyzed and To complete the analysis, we studied spinal cord samples correspond- was frequently—but not always—consistent across animals. For ing to the same regions (cervical, thoracic, and lumbar) with the same example, in all 3 dogs, the increments observed in the meridial histological tools. Tissue architecture was preserved in all injected portion of the frontal cortex were significant, whereas in the dien- dogs; the distribution of GFAP reactivity was homogeneous, with cephalon they were very modest (Figure 5A). In contrast, the vari- an abundance of astrocytes but with no sign of gliosis. Iba1 staining ability was higher in pons or medulla oblongata samples, in which was weak and was also homogeneously distributed. Altogether, the values could vary by a factor of 10 among different animals. On samples were considered to have no significant structural changes, the other hand, spinal cord samples showed consistently high levels and the analysis discarded the presence of a chronic inflammatory of SGSH activity in all animals (Figure 5B). For some of the dogs, process. the levels of SGSH activity documented in the spinal cord extracts represented the highest of all the levels measured in that dog. Transgene expression and activity in the liver After vector administration into the CSF, AAV9 vectors have been Histological analysis of the CNS and DRG systematically detected in the liver of mice, dogs, and At sacrifice, we also collected encephalon, spinal cord, and DRG sam- NHPs,34,35,39,45,47,51 indicating that intra-CSF administration of ples for detailed histological analysis. Sections obtained from different AAV9 vectors could potentially transduce the liver, an organ known regions of the CNS in which we had documented high transgene to efficiently produce and secrete proteins to the blood- expression, including frontal, parietal, and occipital cortices, hippo- stream.34,35,39,45 Percutaneous ultrasound-guided liver biopsies were campus, cerebellum, and spinal cord, were analyzed in detail after he- obtained from the 3 dogs injected with AAV9-Sgsh 48 and 54 months matoxylin and eosin (H&E) and Nissl staining and immunostaining after vector administration. Two samples, from the left median and against GFAP and Iba1. Tissues showed normal structure, and no his- left lateral lobes, were obtained from each animal at each biopsy topathological observations were made that would be suggestive of time point (total of 4 biopsies per animal). One non-injected dog long-term product-related toxicities (Figures S4–S7). and one dog injected in the CSF with AAV9-GFP vectors at the same dose as in the present study were used as negative controls. We paid special attention to DRGs, given the recent reports that have uncovered histopathological findings in the DRGs of primates The results of the analysis of transgene expression by quantitative after intravenous or intra-CSF delivery of high doses of AAV vec- real-time PCR in these biopsies were variable. Dog 2 had the highest tors.47–50 DRG samples were taken from each of the AAV-injected levels of expression in both lobes, dog 3 had detectable expression at dogs at three levels: cervical, thoracic, and lumbar. To make the intermediate levels in the left lateral lobe, and dog 1 showed a tran- assessment through the whole dimension of the extracted DRG, script signal in the same range as negative controls (Table S4). The lateral parietal cortex; dr-TC, dorsal temporal cortex; lt-TC, lateral temporal cortex; PL, piriform lobe; Di (Et), diencephalon (epithalamus); Di (Th), diencephalon (thalamus); Di (Hyp), diencephalon (hypothalamus); dr-Hc, dorsal hippocampus; vn-Hc, ventral hippocampus; dr-OC, dorsal occipital cortex; lt-OC, lateral occipital cortex; md-OC, medial occipital cortex; dr-Mes, dorsal mesencephalon; cn-Mes, central mesencephalon; vn-Mes, ventral mesencephalon; dr-Cer (Ver), dorsal cerebellum (vermis); vn-Cer (Ver), ventral cerebellum (vermis); dr-Po, dorsal pons; vn-Po, ventral pons; dr-Cer (Hem), dorsal cerebellum cortex; vn-Cer (Hem), ventral cerebellum cortex; L-MO, left medulla oblongata; R-MO, right medulla oblongata; ca, cervical anterior; cc, cervicocaudal; ict, cervicothoracic intumescence; ta, thoracic anterior; tm, thoracic medial; tc, thoracic caudal; ils, lumbosacral intumescence; ceq, cauda equina; c, cervical; t, thoracic; l, lumbar. Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 377 Molecular Therapy: Methods & Clinical Development Figure 5. Tissue sulfamidase activity levels in the CNS Quantification of sulfamidase activity in the encephalon (A) and spinal cord (B) of dogs administered AAV9-Sgsh. Sulfamidase activity is represented as % of wild-type (WT) activity for each region. To establish this WT reference value, sulfamidase activity was quantified in each CNS area in samples obtained from 15 untreated, healthy beagle dogs (n = 10–15/region). WT reference values are shown as mean ± SEM. *The value of injected dogs is out of the 95% confidence interval (CI) (versus WT); **the value of injected dogs is out of the 99% CI (versus WT). md-FC, medial frontal cortex; dr-FC, dorsal frontal cortex; lt-FC, lateral frontal cortex; OT, olfactory tract; St, striatum; md-PC, medial parietal cortex; dr-PC, dorsal parietal cortex; lt-PC, lateral parietal cortex; dr-TC, dorsal temporal cortex; lt-TC, lateral temporal cortex; PL, piriform lobe; Di (Et), (legend continued on next page) 378 Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 www.moleculartherapy.org corresponding quantification of VGCN in another portion of the to the vector (Table S5). Interestingly, the rise in circulating NAbs same biopsies paralleled the findings of transcript quantification. was most significant in the animal that was seropositive at baseline. Whereas no vector genomes could be detected in the median or lateral The change in levels of CSF NAbs followed a kinetics similar to that lobes of the liver of dog 1, dog 2 showed 1–2 vector genome copies/cell in serum, albeit remaining always at very low titers (1:5–1:80). in each of the 4 different biopsies analyzed, and dog 3 presented detectable vector genomes in both lobes, albeit at lower copy numbers DISCUSSION than dog 2 (Table S4). Altogether, these results suggest that the differ- Genetic diseases that affect the brain constitute important unmet ences in peripheral sulfamidase expression are due to different de- medical needs. LSDs with neurological involvement, in particular, grees of hepatocyte transduction and not to the potential silencing represent a complex therapeutic challenge, as the enzyme replace- of the transgene. ment therapies (ERTs) that might alleviate peripheral disease are inef- fective to treat the CNS because of the presence of the BBB. The op- At the end of the study, vector biodistribution was evaluated in addi- tion of providing the enzyme directly to the CNS has been tested tional liver samples obtained during necropsy. To get more represen- clinically, although the few trials performed so far have revealed the tative data than those provided by biopsies, 20 evenly distributed shortcomings associated with the implantation of the intrathecal de- samples per lobe were collected from each dog (approximately n = livery devices that are required for periodic infusion of proteins to the 60 samples/liver). One non-injected dog was used as negative control. CSF.52,53 CNS-targeted gene therapy represents a promising thera- In agreement with the observations made in biopsy samples, vector ge- peutic option for many of these diseases. Successful transfer to the nomes were only detected in the livers of dogs 2 and 3; for dog 1 all clinic of proof-of-concept studies performed in rodents relies heavily samples except one were negative (Figure 7A). Remarkably, and on the demonstration of feasibility and safety in large animal models with the caveat that samples were analyzed with a different quantita- that are anatomically and physiologically closer to the intended target tive real-time PCR assay, in all 3 dogs VGCN values were very similar population. The volume of the adult mouse brain is on average in biopsy and necropsy samples obtained 3 years apart. AAV9-derived 0.45 cm3,54–57 2,800-fold smaller than that of a human (1,328 transgene expression analyzed by qRT-PCR fully correlated with and 1,205 cm3 for male and females, respectively, at age 4 years VGCN data; the highest levels of transgene expression were detected 9 months to 18 years58). The (beagle) dog, with a brain volume in in dog 2, which had the highest VGCN, followed by dog 3, with almost an adult animal of 70 cm3 and a lifespan of 13 years (S. Yu, no expression detectable in dog 1 (Figure 7B). Finally, sulfamidase ac- 2010, Fed. Am. Soc. Exp. Biol., abstract), offers a better model to tivity was evaluated in an equal number of liver samples (20 per lobe) address clinically relevant questions such as distribution of the ther- and averaged per lobe. Increases over baseline levels, established for apeutic vector or durability of transgene expression. each lobe averaging samples from 15 control dogs, were documented in dogs 2 and 3, at levels that did not correlate with VGCN and expres- Previous studies have unequivocally established the potential of AAV sion data, which likely reflected the limitations of the SGSH activity vectors to correct CNS disease with a variety of gene transfer method- assay in samples with high endogenous activity (Figure 7C). ologies.1 Some of these have also demonstrated feasibility and tolera- bility in large animal models. Few of them, however, have addressed At sacrifice, we also collected liver samples for detailed histological the long-term safety and durability of effect of the different ap- analysis. Liver showed normal structure, and no histopathological al- proaches tested. Here, we report what to the best of our knowledge terations were observed that would be suggestive of long-term prod- is the longest follow-up (almost 7 years) after gene transfer to the uct-related toxicities (Figure S9). CNS of a large animal. We demonstrate stable, multi-year production of active sulfamidase, the enzyme defective in MPSIIIA, in the Baseline anti-AAV9 immunity and response to vector exposure absence of any signs of acute or chronic toxicity. In an attempt to better understand the differences in transduction observed across the dogs, we did a longitudinal evaluation of neutral- Our approach to MPSs is based on the delivery of vectors encoding izing antibodies (NAbs) against AAV9 in paired CSF and serum sam- the missing enzyme into the CSF. The rationale behind this choice ples. At baseline, none of the dogs had anti-AAV9 NAbs in CSF (Table is that the use of this fluid as the route of delivery allows for better dis- S5). Two of the dogs (dogs 2 and 3) were seronegative for circulating tribution of the vectors within the CNS, the main target compartment NAbs, but dog 1 did have anti-AAV9 NAbs at low titers (1:5–1:10) (Ta- for this disease. Although other serotypes have been delivered to the ble S5). After vector administration, NAbs rose significantly in serum, CSF,59,60 we and others have demonstrated that AAV9 mediates effi- with the highest values observed in day 28 samples, and decreased cient and widespread transduction of the encephalon and the spinal thereafter, stabilizing at relatively modest titers years after exposure cord.34,36,38,40 If the therapeutic transgene is secretable, the protein diencephalon (epithalamus); Di (Th), diencephalon (thalamus); Di (Hyp), diencephalon (hypothalamus); dr-Hc, dorsal hippocampus; vn-Hc, ventral hippocampus; dr-OC, dorsal occipital cortex; lt-OC, lateral occipital cortex; md-OC, medial occipital cortex; dr-Mes, dorsal mesencephalon; cn-Mes, central mesencephalon; vn-Mes, ventral mesencephalon; dr-Cer (Ver), dorsal cerebellum (vermis); vn-Cer (Ver), ventral cerebellum (vermis); dr-Po, dorsal pons; vn-Po, ventral pons; dr-Cer (Hem), dorsal cerebellum cortex; vn-Cer (Hem), ventral cerebellum cortex; L-MO, left medulla oblongata; R-MO, right medulla oblongata; ca, cervical anterior; cc, cervicocaudal; ict, cervicothoracic intumescence; ta, thoracic anterior; tm, thoracic medial; tc, thoracic caudal; ils, lumbosacral intumescence; ceq, cauda equina. Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 379 Molecular Therapy: Methods & Clinical Development Figure 6. Histopathological analysis of PNS dorsal root ganglia Representative images corresponding to lumbar DRGs obtained from dogs 1–3 7 years post-vector administration and from a control, uninjected dog after H&E staining (A) and immunohistochemistry for astrocyte marker GFAP (B). Scale bars, 500 mm; insets, 100 mm. 380 Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 www.moleculartherapy.org Figure 7. Dissemination and persistence of AAV9- Sgsh vectors in the liver after intra-CSF delivery to dogs (A) In each dog, vector gene copy number was analyzed in 20 samples per lobe to fully map AAV9 biodistribution in the liver. vg/dg, vector genome/diploid genome. (B) Optimized canine sulfamidase mRNA was quantified in 10 samples per liver lobe in each dog and referenced to the expression of the housekeeping gene Rplp0. (C) Sulfamidase activity was determined in 20 samples per liver lobe per dog and reported as % of wild-type (WT) activity for that lobe. To obtain the WT reference value, endogenous sulfamidase activity was quantified in 1 sample per liver lobe obtained from 15 untreated, healthy beagle dogs (left lateral lobe = 6.61 ± 0.26 nmol/17 h/mg protein; left medial lobe = 6.11 ± 0.19 nmol/17 h/mg protein; and right medial lobe = 6.49 ± 0.26 nmol/17 h/mg protein). Results are shown as mean ± SEM.*The mean value in injected dogs is out of the 95% confidence in- terval (CI) (versus WT); **the mean value of injected dogs is out of the 99% CI (versus WT). In a previous report, we demonstrated that upon delivery of AAV9 vectors encoding hu- man sulfamidase to the CSF of dogs at the vec- tor dose deemed therapeutic in MPSIIIA mice, activity of the enzyme in the CSF peaked within 2–3 weeks.34 This increment was, however, accompanied by signs of CNS inflammation, such as a rise in TP levels and WBC counts in the CSF of injected animals, and subsequently the transgene activity in the CSF dropped. Despite the inflammatory reaction, sulfamidase activity in the CSF always remained above basal levels, and at termination sulfamidase expres- sion was detectable within different areas of the CNS, as well as in the PNS and somatic tis- sues.34 These observations prompted us to use the same vector backbone but replace the hu- man sulfamidase coding sequence by the spe- cies-specific open reading frame (ORF). This approach allowed us to analyze tolerability and durability of expression in the absence of confounding variables, such as immune re- sponses or use of intense immunosuppressive regimes. Additionally, MPSIIIA patients have shown that >98% of subjects are carriers of at least one allele corresponding to a missense mu- tation.7 Thus, most patients are expected to have some degree of tolerance to sulfamidase, as opposed to dogs to the human protein, can generally be detected in the CSF, and distribution by the CSF can many epitopes of which could constitute neoantigens. Therefore, potentially enhance the therapeutic effect. In this regard, intrathecal we consider healthy beagle dogs injected with canine sulfamidase to ERT has offered valuable evidence of the distribution of lysosomal en- be a better predictor of possible outcomes than beagle dogs treated zymes in the CNS via the CSF.61,62 with human sulfamidase. Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 381 Molecular Therapy: Methods & Clinical Development AAV vectors encoding the canine sulfamidase were injected into the The levels in CSF of a transgene expressed and secreted by transduced CSF at the same dose as that of the human sulfamidase, and animals CNS cells is likely to result from a complex interaction of factors, were followed for almost 7 years. The initial kinetics of canine sulfa- including rate of production and clearance, both of which could be midase expression was similar to that of the human transgene, with an influenced by physiological and pathological processes. Along the initial peak within 2–4 weeks post-administration. In contrast to the years, a certain percentage of transduced cells could be lost not neces- study that used human sulfamidase as a transgene, the activity of sul- sarily due to toxicity or inflammation but to physiological cell turn- famidase in the CSF of dogs injected with the canine sulfamidase- over. In this regard, after intra-CSF administration of AAV9 vectors bearing vectors did not drop to pre-treatment levels after a few encoding the reporter gene GFP we observed transduction of neurons weeks.34 On the contrary, CSF sulfamidase activity remained elevated and astrocytes but also of ependymal and leptomeningeal cells.34 In for several years, at steady-state levels that were at least 2.4-fold contrast to neurons, astrocytes as well as ependymal and leptomenin- higher than those detected at baseline. A similar profile of transgene geal cells have a slow turnover.77–80 In particular, the impact of the expression, in which the levels of transgene protein peak a few weeks loss of ependymocytes and leptomeningeal cells that are in direct con- after vector administration before stabilizing at steady-state levels a tact with the CSF as a monolayer lining the ventricles and meninges, few weeks or months later in complete absence of any sign of toxicity, respectively, would most likely be undetectable through the analysis has been observed for other secretable proteins when expressed of the expression of the transgene in tissue punches. This could from ubiquitous or tissue-specific promoters in a variety of or- explain why, despite the observed differences in SGSH levels in gans.34,39,40,47,63–66 The mechanisms underlying this profile of expres- CSF, there are no significant differences in tissue Sgsh expression sion are not fully understood, but failure to stabilize all of the vector among the dogs, with >30 samples/dog of encephalon analyzed. Dif- genomes present in the initial load into stable episomes has been ferences in expression and SGSH activity are more pronounced in the postulated to play a role.67,68 It is worth noting that we used healthy spinal cord, where dog 1 shows the lowest levels for both readouts; it dogs for this study, with normal production of sulfamidase. Although could be the case that cells from the spinal cord are major contribu- for dogs 2 and 3 SGSH activity seemed to have reached steady-state tors to the output of SGSH to the CSF. Alternatively, it may be the case levels by year 2, in dog 1 the decay from peak seemed to continue that transduction was overall less efficient in dog 1. Because in general beyond year 2 and stabilized 4 years post gene transfer. The reasons dog 1 shows lower VGCN, we consider it unlikely that vector for this additional decline are currently unknown. A significant loss of silencing underlies the more pronounced drop in CSF SGSH in this transduced cells due to inflammatory/immune reactions was ruled dog. Other factors, such as changes in the volume of CSF of a magni- out, as we did not observe any clinical, biochemical, or diagnostic tude that could impact the concentration of the enzyme in this fluid sign of inflammation of the CNS during the in-life portion of the over the course of the life of dog 1, seem unlikely.81,82 study. Sacrifice of the animals 7 years after gene transfer gave us the opportunity to comprehensively analyze vector distribution and The administration of the therapeutic product to the most chal- transgene expression and activity at tissue level. In general, dog 1 lenging-to-treat anatomical compartment—the CNS—guarantees showed slightly lower levels of VGCN in the CNS compared to therapeutic levels of the transgene in the organ most severely affected. dogs 2 and 3. However, this difference was not reflected in significant Additionally, the escape of vector to the circulation and subsequent differences in Sgsh mRNA expression levels, which were fairly similar transduction of the liver could provide a peripheral source of the ther- across the encephalon of all 3 dogs. Although sulfamidase expression apeutic protein. Indeed, we were the first to report whole-body was detectable in most deep brain areas, the striatum, epithalamus, correction of a LSD upon delivery of AAV9 vectors to the CSF of and thalamus showed very low to no detectable mRNA levels, which MPSIIIA mice,34 and we have since extended our observations to 3 could be a limitation of intra-CSF-based gene therapy in transducing other indications.35,39,45 In principle, peripheral efficacy would deep brain structures.34,38 It should be borne in mind that MPSIIIA is depend on the maintenance of a certain level of hepatic transduction, caused by mutations in a secretable lysosomal enzyme that can cor- sufficient to support the production of therapeutic levels of the trans- rect the enzymatic deficit of non-transduced cells through cross- gene product. In the 3 dogs treated in this study, we found variable correction. Preclinical and clinical data with intra-CSF ERT for levels of transduction in the liver leading to different levels of trans- several different secretable lysosomal enzymes52,69–76 suggest that gene expression. Although vector genomes were undetectable in liver therapeutic efficacy can be expected not only from direct vector trans- biopsies of dog 1, dog 2 and dog 3 presented vg/dg in all samples. duction but also from the presence of the enzyme in the CSF. Differ- Noticeably, the VGCN remained stable from the time of the biopsies ences in transgene expression were somewhat more pronounced in (48 and 54 months) to the time of sacrifice (82 months) in all dogs. the spinal cord and DRGs. For example, dog 3 seemed to have slightly The correlation with sulfamidase activity measured in post mortem higher levels of Sgsh mRNA expression in spinal cord, but it is difficult tissue extracts was less clear. Whereas in the CSF the baseline amount to interpret whether those differences are of significance, as CSF of SGSH activity is very low, which allows for reliable detection of SGSH activity levels were very similar in dogs 2 and 3. In any case, increases that can be attributed to the administration of sulfami- the comparative analysis of vector distribution and expression data dase-encoding vectors, endogenous levels of SGSH activity in liver across the animals argued against transgene silencing as an explana- parenchyma, and CNS parenchyma, are high, which makes interpre- tion for the different behavior along the years of CSF SGSH in the tation of results more challenging, as small variations can be masked different dogs. by the high endogenous activity. 382 Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 www.moleculartherapy.org Given the absence of any signs that would suggest loss of vector ge- DRG toxicity, characterized by mononuclear cell infiltrate, gliosis, nomes from the liver of dog 1, our observations might be reflecting neuronal loss, and axonopathy, has recently been identified as a com- true differences in liver transduction efficacy. Evidence supporting mon, albeit low in magnitude, histological finding following i.v. and this hypothesis was provided by the longitudinal quantification of intra-CSF but not intraparenchymal administration of high doses anti-AAV9 NAbs in paired CSF and serum samples for all dogs. of AAV vectors to NHPs.47–50 Toxicity is dose dependent48 and can Although the 3 dogs were negative for NAbs in the CSF at baseline, be attenuated by use of microRNA sequences that de-target expres- which explained the similar level of transduction efficacy in the sion from DRGs,85 and frequency is higher after intra-CSF delivery, CNS, dog 1 showed the presence of NAbs in circulation at 1:5–1:10 likely because of the use of tissue-specific promoters when targeting titer. Although low, this titer has been described to be sufficient to peripheral organs.48 Although not completely elucidated, all evidence block liver transduction in large animals and humans.83,84 The seems to point toward high levels of transgene expression due to high- absence of NAbs in the CSF of an animal seropositive for AAV9 is ly efficient transduction of ganglia sensory neurons as the root cause not surprising, given the asymmetrical distribution of NAbs across of toxicity (ganglia have fenestrated capillary, and sensory neurons the BBB, even in patients in whom mild disruption of the BBB is ex- are exposed to CSF).34 Indeed, the VGCN in the DRG was lower pected.34 It is actually possible that dog 1 did have some NAbs in the compared to other short-term studies conducted by us and others af- CSF prior to vector administration but the levels fell below the detec- ter intra-CSF delivery of AAV9 to NHPs.34,50,86 The observation of tion limit of our NAb assay. It is also possible that, even if undetect- higher VGCN shortly after vector delivery is a constant observation able, those low levels of NAbs in CSF partially affected transduction in in the field of in vivo gene transfer with AAVs, irrespective of the dog 1, which seemed to show more regions of the CNS negative for transgene and target organ. Whether the difference in VGCN re- AAV vector genomes than the other dogs. Evidently, if present, the ported in short- versus long-term studies reflects non-functional vec- greatest proportion of those NAbs were neutralized by the high tor genomes that were not taken up by cells and are cleared with time, dose of vectors delivered directly to the CSF. The efficient transduc- stabilization of only an initial fraction of the vector genomes in the tion of the CNS in the presence of peripheral NAbs observed in nucleus, or turnover of some of the targeted cells remains to be deter- dog 1 is in agreement with our earlier work; we showed that in mined for each organ.5,68,87 Importantly, no clinical findings have dogs pre-immunized by systemic exposure to AAV it is possible to been reported to date from clinical trials, and the findings in NHP achieve successful CNS transduction with minimal loss of efficacy, studies remain a histological finding, and no clinical signs have albeit peripheral transduction is completely blocked.34,39 been correlated after analysis of 256 animals.48 When we analyzed the DRGs of our dogs, we could not find any histological signs of If efficacy in mild peripheral MPSIIIA disease was not achieved DRG toxicity. The reasons for this may be several: (1) Findings because of the lack of transduction of the liver, there would always have been reported mostly in NHPs,47–50 with only one report in a be the possibility of providing patients with intravenous ERT, at different species (neonatal piglets49). To the best of our knowledge, expectedly lower doses than those required to treat severe CNS pa- this is the first report investigating AAV-mediated DRG toxicity in thology, and most likely under a treatment regime more favorable dogs after such a long follow-up of 7 years. A recent study reported to patients. Finally, it should be noted that, as the CSF ultimately similar findings at 2.5 years of follow-up after intra-CSF AAV9 treat- drains into the circulatory and lymphatic systems, a certain ment in Krabbe disease dogs88; hence, it is possible that dogs’ DRG amount of the therapeutic protein could also pass to the periphery. neurons are not as highly transduced as primates’. (2) A dose-effect In this regard, studies conducted in a dog model of late infantile relationship has been clearly established, and the doses used in the neuronal ceroid lipofuscinosis (Batten disease) found significant current study are lower than those used in the smaller cynos and rhe- amounts of the soluble lysosomal enzyme tripeptidyl peptidase 1 sus monkeys.48 (3) The timing of our analysis is posterior to the devel- (TPP1) in heart and spleen after delivery of TPP-encoding opment and resolution of toxicity. In any case, the close clinical AAV2 vectors to the CSF, an AAV serotype with minimal periph- follow-up of the dogs in this study, which were subjected to extensive eral dissemination.59 neurological evaluations, suggests that, if present, DRG toxicity had no clinical consequences. Besides the documentation of steady levels of therapeutic transgene through multiple years, another important outcome of this study is In the present study, we chose to use healthy beagle dogs as a large the demonstration of the good tolerability and long-term safety of animal model. Although there is a spontaneous dog model of the approach. Animals remained in an excellent health status MPSIIIA for which an experimental colony has been established throughout the study. We did not document any clinically relevant (MPSIIIA New Zealand Huntaway dog), its use does not provide increase in protein levels or WBC counts at any of the time points any additional benefits for the long-term evaluation of sulfamidase at which CSF was sampled. Likewise, no clinically relevant alterations expression and safety. On one hand, the clinical manifestations and in biochemical or hematological parameters could be attributed to the evolution of the disease seem to differ substantially between humans procedure or test article administration. As further proof of long-term and dogs, which mostly present signs of cerebellar disease, with min- safety, in-life MRI and ultrasound imaging, as well as histopatholog- imal or no behavioral alterations or somatic disease.89–91 On the ical studies of post mortem specimens, revealed no evident abnormal- other hand, the New Zealand Huntaway dog bears a frameshift mu- ities in the architecture of the encephalon, spinal cord, DRGs, or liver. tation with premature codon stop92 resulting in a complete lack of Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 383 Molecular Therapy: Methods & Clinical Development sulfamidase expression, which renders this animal immunologically AAV vector production intolerant to the sulfamidase protein. Acute reactions to the human AAV expression cassette was generated by cloning the cDNA of opti- protein have been well documented in ERT studies performed in mized canine sulfamidase (GeneArt, Life Technologies) into an AAV this animal model.61,93 The intolerance to the canine sulfamidase pro- backbone plasmid containing the ubiquitous CAG promoter (hybrid tein, even if expressed from an AAV vector, can be predicted from ob- of cytomegalovirus [CMV] enhancer and chicken b-actin promoter) servations made in earlier studies in hemophilia B dogs bearing null and the rabbit beta-globin polyA signal (rbGpA). AAV9 vectors were mutations and receiving AAVs encoding for canine factor IX (FIX).94 produced by triple transfection of HEK293 cells and purified with an To enable long-term evaluation of the treatment in New Zealand optimized purification protocol with double cesium chloride gradient Huntaway dogs, the co-administration of immunosuppression would ultracentrifugation that results in vector preps of high purity with be necessary, which could not only compromise the life of dogs but negligible amounts of empty capsids.95 Vectors were titered by quan- also introduce a confounding variable into the evaluation of possible titative real-time PCR and stored at 80 C until use. adverse effects. Moreover, >90% of the mutations in the human SGSH gene described so far are missense mutations, and the report on the CSF collection and intracisternal administration of vectors largest cohort of MPSIIIA patients studied up to now showed that CSF collection and delivery of vectors was performed as previously >98% of subjects were carriers of at least one allele corresponding described.34 Briefly, anesthetized dogs were positioned in lateral re- to a missense mutation.7 Thus, most patients are expected to have cumbency, and a 22-gauge needle was introduced into the cisterna some degree of tolerance to sulfamidase, and therefore healthy beagle magna between the occipital bone and the C1 vertebra. Right place- dogs are a better predictor of possible outcomes than the New Zea- ment of the needle was confirmed by spontaneous flow of crystal- land Huntaway dog. clear CSF, which was collected in sterile tubes and stored at 80 C until further analysis. For vector delivery, a syringe connected to a In conclusion, our study provides strong evidence of the feasibility fluid extension catheter tube with a 3-way stopcock was attached to and long-term durability and safety of using AAV9 vectors to deliver the needle and 1 mL total volume of vector solution was injected genes to the CNS. After the administration of clinically relevant doses over a period of 15 s. The dose used (2  1013 vg/dog) was obtained of vector, the detection of sustained, significant levels of sulfamidase by scaling up the dose previously found to be effective in mice,34 based in the CSF and of widespread transgene expression in the CNS, PNS, on body weight of mice compared to dogs. and liver of treated dogs in the absence of any signs of toxicity after many years of follow-up provides strong evidence supporting the Metabolite analysis and cell counts use of intra-CSF AAV9-mediated gene therapy to treat MPSIIIA General laboratory parameters and CSF TP content were measured by and other neurodegenerative disorders. spectrophotometry with a Cobas Mira Analyzer (Roche) at the Clin- ical Biochemistry Lab of the Veterinary Clinical Hospital, UAB. He- MATERIALS AND METHODS mograms were analyzed with a SCIL VetABC hematology analyzer Animals and CSF cell counts, WBC and RBC, were manually counted with a Male healthy beagle dogs (n = 3; dogs 1 and 2: 9 months, dog 3: Neubauer chamber by the Clinical Hematology Lab. 8 months of age) were purchased from Isoquimen (Spain) and housed at the Veterinary School of the Universitat Autònoma de Barcelona Ultrasonography and magnetic resonance imaging (UAB). Animals’ health was monitored through clinical, hematolog- For ultrasound imaging, anesthetized dogs were positioned in dorsal ical, and biochemical examination. Animals were fed once daily at recumbency and an abdominal ultrasound was performed with a 9:00 AM with 30 g/kg body weight standard dry food (Elite Nutrition, high-frequency linear transducer (LA523, linear array, 18 MHz, My- Nestle). For general anesthesia, dogs were premedicated with an Lab70, Esaote). For brain MRI, anesthetized dogs were positioned in intramuscular administration of 0.05 mg/kg of acepromazine (Equi- ventral recumbency and scans were performed with a 0.2-T perma- promacina, Fatro Ibérica) and 0.2 mg/kg of butorphanol (Torbugesic, nent open magnet system (Vet MR, Esaote). Technical parameters Zoetis). Thirty minutes later, induction was performed by i.v. admin- were (1) T1-weighted images in transverse and sagittal planes: TR = istration of 4 mg/kg of propofol (PropoVet, B. Braun Medical) and 800 ms, TE = 26 ms, FOV = 180  180 mm, thickness = 4.0 mm, 0.5 mg/kg of diazepam (Valium). After endotracheal intubation, spacing = 0.4 mm, matrix = 224  224, and NEX = 3; (2) T2-weighted anesthesia was maintained by inhalation of 2% isoflurane (IsoVet, images in transverse, sagittal, and dorsal planes: TR = 3,000 ms, TE = B. Braun Medical) in 100% oxygen. Throughout all surgical proced- 80 ms, FOV = 180  180 mm, thickness = 4.0 mm, spacing = 0.4 mm, ures, temperature, cardiac and respiratory frequency, capnography, matrix = 224  224, and NEX = 1; and (3) fluid-attenuated inversion arterial pressure, pulse, and electrocardiography were monitored recovery (FLAIR) images in transverse planes: TR = 6,960 ms, TE = with a multifunctional patient Vet Care monitor (B. Braun Medical). 80 ms, TI = 1,800 ms, FOV = 200  200 mm, thickness = 4.0 mm, The neurological evaluation of injected dogs was performed by a spacing = 0.4 mm, matrix = 192  192, and NEX = 1. In addition, diplomate veterinary neurologist from the Veterinary Clinical Hospi- contrast-enhanced T1-weighted images of transverse and sagittal tal of the UAB. All experimental procedures were approved by the planes were acquired after intravenous administration of paramag- Ethics Committee for Animal and Human Experimentation of the netic contrast medium (0.1 mmol/kg gadopentetate dimeglumine UAB. [Magnevist, Bayer]). For spinal cord MRI, anesthetized dogs were 384 Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 www.moleculartherapy.org positioned in dorsal recumbency and scans were performed with a ND-1000 spectrophotometer (NanoDrop). VGCN in 40 ng of total 0.4-T scanner (Aperto Lucent, Hitachi Medical). Technical parame- DNA was determined by quantitative real-time PCR using the Light- ters were (1) T1-weighted images in transverse and sagittal planes: Cycler 480 Probes Master (Roche). For analysis of liver biopsies, TR = 653 ms, TE = 23 ms, FOV = 250  250 mm, thickness = primers and probe specific for sequence within the optimized canine 4.0 mm, spacing = 0.5 mm, matrix = 288  220, and NEX = 2 and sulfamidase cDNA were used: forward primer: 50 -GCA TCA TCG (2) T2-weighted images in transverse, sagittal, and dorsal planes: GCA AGA AAC AC-30 ; reverse primer: 50 -CAG CTT GAT CCT TR = 3,100 ms, TE = 80 ms, FOV = 150  150 mm, thickness = GGT GAT GT-30 ; probe: 50 -ACA CCG AAG AGA ACA GCA 4.0 mm, spacing = 0.5 mm, matrix = 256  224, and NEX = 6. GCG T-30 . The analysis of terminal samples was done with primers Contrast-enhanced T1-weighted transverse and sagittal images were and probe specific for the rbGpA sequence: forward primer: 50 - also obtained. CTT GAG CAT CTG ACT TCT GGC TAA T-30 ; reverse primer: 50 - GAT TTG CCC TCC CAT ATG TCC-30 ; and probe: 50 -CCG Sample collection AGT GAG AGA CAC AAA AAA TTC CAA CAC-30 . A reference Forty-eight and fifty-four months after vector administration, percu- standard curve was built from serial dilutions of a linearized plasmid taneous ultrasound-guided liver biopsies were obtained under general bearing the CAG promoter, the optimized canine sulfamidase cDNA, anesthesia. Briefly, abdominal hair was clipped, and a microconvex and the rbGpA sequences spiked into 20 ng/mL of non-transduced sector transducer operating at 6 or 7 MHz was used to guide liver dog genomic DNA. sampling with a 14 gauge  9 cm Tru-Cut biopsy needle (J0528, SuperCore Biopsy Needle, Argon Medical Devices). Sulfamidase expression Total RNA was purified from tissues homogenized in TriPure Isola- Eighty-two months after vector administration, dogs were pre-medi- tion Reagent (Roche) with the RNeasy Mini Kit (QIAGEN). RNA was cated by intramuscular injection of 0.2 mg/kg butorphanol (Torbuge- quantified in a NanoDrop ND-1000 spectrophotometer (NanoDrop). sic, Zoetis) and 0.05 mg/kg acepromazine (Equipromacina, Fatro cDNA was synthesized with a Transcriptor First Strand cDNA Syn- Ibérica). Thirty minutes later, euthanasia was performed with an thesis Kit (Roche). quantitative real-time PCR was performed with overdose of pentobarbital (150 mg/kg, Dolethal) administered i.v. the LightCycler 480 Probes Master (Roche) with the same optimized The encephalon, the spinal cord, the DRGs, and multiple somatic tis- canine sulfamidase-specific primers and probe used for VGCN quan- sues were collected and either formalin-fixed or snap-frozen and tification in liver biopsies. The values obtained were normalized to the stored at 80 C until use. expression of the dog RPLP0 gene: forward primer: 50 -ACC TCT TTC TTC CAG GCT TTA G-30 ; reverse primer: 50 - CCA CTT TGT CTC Sulfamidase activity CCG TCT TAA T-30 , probe: 50 - ACC ATT GAA ATC TTG AGT CNS and liver samples were sonicated in 100–500 mL of Milli-Q wa- GAT GTG CAG-30 . ter. Enzymatic activity was assayed in supernatants of sonicated sam- ples or CSF with a 2-step protocol using 4-methylumbelliferone- Histology derived fluorogenic substrate (4-MU), as previously described.96 Tissues were fixed for 12–24 h in 10% formalin, embedded in paraffin, Briefly, 30 mg of TP or 10 mL of CSF was first incubated with and sectioned. In brain and spinal cord, H&E and Nissl staining and 10 mmol/L 4-methylumbelliferyl-2-sulfamino-2-deoxy-alpha-D-glu- immunohistochemical staining against Iba1 and GFAP were per- copyranoside sodium salt (4-MU-aGlcNS, Enantia, Spain) and formed in 5-mm sections. For DRG analysis, 3 serial 3-mm sections 10 mL of complete solution (Roche) for 17 h at 47 C. The second in- were obtained, at 3 different levels of the ganglia, with a separation cubation was carried out after the addition of 6 mL of PiCi buffer between series of 50 mm. The first slide of each series was used for (0.4 M Na2HPO4$12H2O-0.2 M tris-sodium citrate 2-hydrate, pH H&E staining, and the consecutive sections were used for immuno- 6.7 + 0.025% [w/v] Triton X-100) in the presence of 10 U/mL of histochemical detection of Iba1 and GFAP, always following the a-glucosidase (Sigma) in 0.2% BSA (pH and heat inactivated, Sigma) same order of tissue sections. They were incubated overnight at 4 C for 24 h at 37 C. After the second enzymatic reaction was stopped by with rabbit anti-GFAP (Z0334, Dako Cytomation) or goat anti-Iba1 increasing the pH, the released fluorescence was measured with a Syn- (Ab5076, Abcam). The secondary antibodies were biotinylated goat ergy HTX fluorimeter (BioTek Instruments). Sulfamidase activity was anti-rabbit immunoglobulin G (IgG) (31820, Invitrogen) or bio- normalized against CSF volume (expressed as nmol/17 h/mL) or TP tinylated rabbit anti-goat IgG (E0466, Dako Cytomation). The ABC content (expressed as nmol/17 h/mg protein) quantified by Bradford peroxidase kit (Pierce) was used for immunodetection, and sections assay (Bio-Rad). were counterstained in Mayer’s hematoxylin. Images were obtained with an Eclipse E800 optical microscope (Nikon). Vector genome copy number CNS, PNS ganglia, or peripheral tissue samples were digested over- Neutralizing antibodies night (ON) at 56 C in 300 mL of Tissue Lysis Solution supplemented Anti-AAV9 NAb titers were determined in dog CSF and serum sam- with Proteinase K (0.2 mg/mL). Total DNA was isolated from super- ples as previously described,97 with an in vitro neutralization assay, at natants with the MasterPure DNA Purification Kit (Lucigen). DNA ISGLOBAL, Barcelona Centre for International Health Research was resuspended in distilled water and quantified with a NanoDrop (CRESIB), Hospital Clínic, Barcelona. Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 385 Molecular Therapy: Methods & Clinical Development Statistical analysis Mucopolysaccharidosis type IIIA: clinical spectrum and genotype-phenotype cor- Results are expressed as mean ± SEM. Statistical comparisons for CSF relations. Ann. Neurol. 68, 876–887. sulfamidase activity were made with unpaired two-tailed t test. Statis- 8. Meyer, A., Kossow, K., Gal, A., Mühlhausen, C., Ullrich, K., Braulke, T., and Muschol, N. (2007). Scoring evaluation of the natural course of mucopolysaccharidosis type tical significance was considered if p < 0.05. Data on sulfamidase ac- IIIA (Sanfilippo syndrome type A). Pediatrics 120, e1255–e1261. tivity in tissues were represented with a confidence interval (CI) of 9. Truxal, K.V., Fu, H., McCarty, D.M., McNally, K.A., Kunkler, K.L., Zumberge, N.A., 95% and 99% between control (non-injected dogs) and injected dogs. Martin, L., Aylward, S.C., Alfano, L.N., Berry, K.M., et al. (2016). A prospective one- year natural history study of mucopolysaccharidosis types IIIA and IIIB: Implications SUPPLEMENTAL INFORMATION for clinical trial design. Mol. Genet. Metab. 119, 239–248. Supplemental information can be found online at https://doi.org/10. 10. Shapiro, E.G., Nestrasil, I., Delaney, K.A., Rudser, K., Kovac, V., Nair, N., Richard, C.W., 3rd, Haslett, P., and Whitley, C.B. (2016). A Prospective Natural History 1016/j.omtm.2021.09.017. Study of Mucopolysaccharidosis Type IIIA. J. Pediatr. 170, 278, 87.e1, 4. 11. Lin, H.Y., Chuang, C.K., Lee, C.L., Tu, R.Y., Lo, Y.T., Chiu, P.C., Niu, D.M., Fang, ACKNOWLEDGMENTS Y.Y., Chen, T.L., Tsai, F.J., et al. (2018). Mucopolysaccharidosis III in Taiwan: We thank the “Servei de Granges i Camps Experimentals” of the Vet- Natural history, clinical and molecular characteristics of 28 patients diagnosed during erinary School at the UAB for technical support. We also thank Aida a 21-year period. Am. J. Med. Genet. A. 176, 1799–1809. Peiró for the NAb determination at Hospital Clínic, Barcelona. This 12. Abreu, N.J., Selvaraj, B., Truxal, K.V., Moore-Clingenpeel, M., Zumberge, N.A., McNally, K.A., McBride, K.L., Ho, M.L., and Flanigan, K.M. (2021). Longitudinal work was supported by funding from Plan Nacional I+D+I from MRI brain volume changes over one year in children with mucopolysaccharidosis the Ministerio de Economía y Competitividad (INNPACTO IPT- types IIIA and IIIB. Mol. Genet. Metab. 133, 193–200. 2012-0772-300000 and SAF2017-86166-R), Generalitat de Catalunya 13. Heon-Roberts, R., Nguyen, A.L.A., and Pshezhetsky, A.V. (2020). Molecular Bases of (2017SGR-01508 and ICREA Academia Award to F.B.), and MPS Es- Neurodegeneration and Cognitive Decline, the Major Burden of Sanfilippo Disease. paña Foundation, Spain, and from the European Union through J. Clin. Med. 9, 344. Regional Development Funds (ERDF). This work is also part of a 14. Enns, G.M., and Huhn, S.L. (2008). Central nervous system therapy for lysosomal storage disorders. Neurosurg. Focus 24, E12. Public-Private Partnership on Gene Therapy between UAB and ES- 15. Foust, K.D., Nurre, E., Montgomery, C.L., Hernandez, A., Chan, C.M., and Kaspar, TEVE Pharmaceuticals. S.M., V.S., and G.E. received predoctoral fel- B.K. (2009). Intravascular AAV9 preferentially targets neonatal neurons and adult as- lowships from Generalitat de Catalunya, Spain. trocytes. Nat. Biotechnol. 27, 59–65. 16. Duque, S., Joussemet, B., Riviere, C., Marais, T., Dubreil, L., Douar, A.M., Fyfe, J., AUTHOR CONTRIBUTIONS Moullier, P., Colle, M.A., and Barkats, M. (2009). Intravenous administration of S.M., V.H., and F.B. designed and supervised experiments and wrote self-complementary AAV9 enables transgene delivery to adult motor neurons. Mol. Ther. 17, 1187–1196. and edited the manuscript. S.M., M.L.J., A.R., V.S., M.M., M.G., C.R., 17. Mendell, J.R., Al-Zaidy, S., Shell, R., Arnold, W.D., Rodino-Klapac, L.R., Prior, T.W., X.S., J.B., J.P., and G.E. generated reagents and performed experi- Lowes, L., Alfano, L., Berry, K., Church, K., et al. (2017). Single-Dose Gene- ments. X.L. generated reagents. M.N., A.C., M.P., A.A., Y.E., and Replacement Therapy for Spinal Muscular Atrophy. N. Engl. J. Med. 377, 1713–1722. S.A. performed experiments. M.P., Y.E., and S.A. analyzed data and 18. Lotfi, P., Tse, D.Y., Di Ronza, A., Seymour, M.L., Martano, G., Cooper, J.D., Pereira, contributed to discussion. F.A., Passafaro, M., Wu, S.M., and Sardiello, M. (2018). Trehalose reduces retinal degeneration, neuroinflammation and storage burden caused by a lysosomal hydro- lase deficiency. Autophagy 14, 1419–1434. DECLARATION OF INTERESTS 19. Ruzo, A., Marcó, S., García, M., Villacampa, P., Ribera, A., Ayuso, E., Maggioni, L., F.B. is an inventor on a patent application for the use of AAV vectors Mingozzi, F., Haurigot, V., and Bosch, F. (2012). Correction of pathological accumu- for the treatment of MPS IIIA. All other authors declare no competing lation of glycosaminoglycans in central nervous system and peripheral tissues of interests. MPSIIIA mice through systemic AAV9 gene transfer. Hum. Gene Ther. 23, 1237– 1246. REFERENCES 20. Fu, H., Dirosario, J., Killedar, S., Zaraspe, K., and McCarty, D.M. (2011). Correction 1. Hocquemiller, M., Giersch, L., Audrain, M., Parker, S., and Cartier, N. (2016). Adeno- of neurological disease of mucopolysaccharidosis IIIB in adult mice by rAAV9 trans- Associated Virus-Based Gene Therapy for CNS Diseases. Hum. Gene Ther. 27, blood-brain barrier gene delivery. Mol. Ther. 19, 1025–1033. 478–496. 21. Meyer, K., Ferraiuolo, L., Schmelzer, L., Braun, L., McGovern, V., Likhite, S., Michels, 2. Gessler, D.J., and Gao, G. (2016). Gene therapy for the treatment of neurological dis- O., Govoni, A., Fitzgerald, J., Morales, P., et al. (2015). Improving single injection CSF orders: metabolic disorders. In Methods in Molecular Biology, F.P. Manfredsson, ed. delivery of AAV9-mediated gene therapy for SMA: a dose-response study in mice and (Humana Press), pp. 429–465. nonhuman primates. Mol. Ther. 23, 477–487. 3. Poletti, V., and Biffi, A. (2019). Gene-based approaches to inherited neurometabolic 22. Mingozzi, F., and High, K.A. (2017). Overcoming the Host Immune Response to diseases. Hum. Gene Ther. 30, 1222–1235. Adeno-Associated Virus Gene Delivery Vectors: The Race Between Clearance, Tolerance, Neutralization, and Escape. Annu. Rev. Virol. 4, 511–534. 4. Giugliani, R., Vairo, F., Kubaski, F., Poswar, F., Riegel, M., Baldo, G., and Saute, J.A. (2018). Neurological manifestations of lysosomal disorders and emerging therapies 23. Tardieu, M., Zérah, M., Husson, B., de Bournonville, S., Deiva, K., Adamsbaum, C., targeting the CNS. Lancet Child Adolesc. Health 2, 56–68. Vincent, F., Hocquemiller, M., Broissand, C., Furlan, V., et al. (2014). Intracerebral administration of adeno-associated viral vector serotype rh.10 carrying human 5. Naso, M.F., Tomkowicz, B., Perry, W.L., 3rd, and Strohl, W.R. (2017). Adeno- SGSH and SUMF1 cDNAs in children with mucopolysaccharidosis type IIIA disease: Associated Virus (AAV) as a Vector for Gene Therapy. BioDrugs 31, 317–334. results of a phase I/II trial. Hum. Gene Ther. 25, 506–516. 6. Verdera, H.C., Kuranda, K., and Mingozzi, F. (2020). AAV Vector Immunogenicity 24. Tardieu, M., Zérah, M., Gougeon, M.-L., Ausseil, J., de Bournonville, S., Husson, B., in Humans: A Long Journey to Successful Gene Transfer. Mol. Ther. 28, 723–746. Zafeiriou, D., Parenti, G., Bourget, P., Poirier, B., et al. (2017). Intracerebral gene ther- 7. Valstar, M.J., Neijs, S., Bruggenwirth, H.T., Olmer, R., Ruijter, G.J., Wevers, R.A., apy in children with mucopolysaccharidosis type IIIB syndrome: an uncontrolled van Diggelen, O.P., Poorthuis, B.J., Halley, D.J., and Wijburg, F.A. (2010). phase 1/2 clinical trial. Lancet Neurol. 16, 712–720. 386 Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 www.moleculartherapy.org 25. Souweidane, M.M., Fraser, J.F., Arkin, L.M., Sondhi, D., Hackett, N.R., Kaminsky, 41. Donsante, A., McEachin, Z., Riley, J., Leung, C.H., Kanz, L., O’Connor, D.M., and S.M., Heier, L., Kosofsky, B.E., Worgall, S., Crystal, R.G., and Kaplitt, M.G. (2010). Boulis, N.M. (2016). Intracerebroventricular delivery of self-complementary adeno- Gene therapy for late infantile neuronal ceroid lipofuscinosis: neurosurgical consid- associated virus serotype 9 to the adult rat brain. Gene Ther. 23, 401–407. erations. J. Neurosurg. Pediatr. 6, 115–122. 42. Sorrentino, N.C., Maffia, V., Strollo, S., Cacace, V., Romagnoli, N., Manfredi, A., 26. Salegio, E.A., Samaranch, L., Jenkins, R.W., Clarke, C.J., Lamarre, C., Beyer, J., Kells, Ventrella, D., Dondi, F., Barone, F., Giunti, M., et al. (2016). A Comprehensive A.P., Bringas, J., Sebastian, W.S., Richardson, R.M., et al. (2012). Safety study of ad- Map of CNS Transduction by Eight Recombinant Adeno-associated Virus eno-associated virus serotype 2-mediated human acid sphingomyelinase expression Serotypes Upon Cerebrospinal Fluid Administration in Pigs. Mol. Ther. 24, 276–286. in the nonhuman primate brain. Hum. Gene Ther. 23, 891–902. 43. Ellinwood, N.M., Ausseil, J., Desmaris, N., Bigou, S., Liu, S., Jens, J.K., Snella, E.M., 27. Bradbury, A.M., Cochran, J.N., McCurdy, V.J., Johnson, A.K., Brunson, B.L., Gray- Mohammed, E.E., Thomson, C.B., Raoul, S., et al. (2011). Safe, efficient, and repro- Edwards, H., Leroy, S.G., Hwang, M., Randle, A.N., Jackson, L.S., et al. (2013). ducible gene therapy of the brain in the dog models of Sanfilippo and Hurler syn- Therapeutic response in feline sandhoff disease despite immunity to intracranial dromes. Mol. Ther. 19, 251–259. gene therapy. Mol. Ther. 21, 1306–1315. 44. Hinderer, C., Bell, P., Vite, C.H., Louboutin, J.-P., Grant, R., Bote, E., Yu, H., Pukenas, B., Hurst, R., and Wilson, J.M. (2014). Widespread gene transfer in the central ner- 28. Gilkes, J.A., Bloom, M.D., and Heldermon, C.D. (2015). Preferred transduction with vous system of cynomolgus macaques following delivery of AAV9 into the cisterna AAV8 and AAV9 via thalamic administration in the MPS IIIB model: A comparison magna. Mol. Ther. Methods Clin. Dev. 1, 14051. of four rAAV serotypes. Mol. Genet. Metab. Rep. 6, 48–54. 45. Roca, C., Motas, S., Marcó, S., Ribera, A., Sánchez, V., Sánchez, X., Bertolin, J., León, 29. Gray, A.L., O’Leary, C., Liao, A., Agundez, L., Youshani, A.S., Gleitz, H.F., Parker, H., X., Pérez, J., Garcia, M., et al. (2017). Disease correction by AAV-mediated gene ther- Taylor, J.T., Danos, O., Hocquemiller, M., et al. (2019). An Improved Adeno- apy in a new mouse model of mucopolysaccharidosis type IIID. Hum. Mol. Genet. 26, Associated Virus Vector for Neurological Correction of the Mouse Model of 1535–1551. Mucopolysaccharidosis IIIA. Hum. Gene Ther. 30, 1052–1066. 46. Piatt, J. (1991). Technique of ventriculostomy. In Neurosurgical Operative Atlas, S.S. 30. Winner, L.K., Beard, H., Hassiotis, S., Lau, A.A., Luck, A.J., Hopwood, J.J., and Rengachary and R.H. Wilkins, eds. (The American Association of Neurological Hemsley, K.M. (2016). A Preclinical Study Evaluating AAVrh10-Based Gene Surgeons), pp. 171–175. Therapy for Sanfilippo Syndrome. Hum. Gene Ther. 27, 363–375. 47. Hordeaux, J., Hinderer, C., Buza, E.L., Louboutin, J.-P., Jahan, T., Bell, P., Chichester, 31. Leone, P., Shera, D., McPhee, S.W., Francis, J.S., Kolodny, E.H., Bilaniuk, L.T., Wang, J.A., Tarantal, A.F., and Wilson, J.M. (2019). Safe and Sustained Expression of D.J., Assadi, M., Goldfarb, O., Goldman, H.W., et al. (2012). Long-term follow-up af- Human Iduronidase After Intrathecal Administration of Adeno-Associated Virus ter gene therapy for canavan disease. Sci. Transl. Med. 4, 165ra163. Serotype 9 in Infant Rhesus Monkeys. Hum. Gene Ther. 30, 957–966. 32. Colle, M.A., Piguet, F., Bertrand, L., Raoul, S., Bieche, I., Dubreil, L., Sloothaak, D., 48. Hordeaux, J., Buza, E.L., Dyer, C., Goode, T., Mitchell, T.W., Richman, L., Denton, N., Bouquet, C., Moullier, P., Aubourg, P., et al. (2010). Efficient intracerebral delivery Hinderer, C., Katz, N., Schmid, R., et al. (2020). Adeno-Associated Virus-Induced of AAV5 vector encoding human ARSA in non-human primate. Hum. Mol. Dorsal Root Ganglion Pathology. Hum. Gene Ther. 31, 808–818. Genet. 19, 147–158. 49. Hinderer, C., Katz, N., Buza, E.L., Dyer, C., Goode, T., Bell, P., Richman, L.K., and Wilson, J.M. (2018). Severe Toxicity in Nonhuman Primates and Piglets Following 33. Ciron, C., Cressant, A., Roux, F., Raoul, S., Cherel, Y., Hantraye, P., Deglon, N., High-Dose Intravenous Administration of an Adeno-Associated Virus Vector Schwartz, B., Barkats, M., Heard, J.M., et al. (2009). Human alpha-iduronidase Expressing Human SMN. Hum. Gene Ther. 29, 285–298. gene transfer mediated by adeno-associated virus types 1, 2, and 5 in the brain of nonhuman primates: vector diffusion and biodistribution. Hum. Gene Ther. 20, 50. Hordeaux, J., Hinderer, C., Goode, T., Buza, E.L., Bell, P., Calcedo, R., Richman, L.K., 350–360. and Wilson, J.M. (2018). Toxicology Study of Intra-Cisterna Magna Adeno- Associated Virus 9 Expressing Iduronate-2-Sulfatase in Rhesus Macaques. Mol. 34. Haurigot, V., Marcó, S., Ribera, A., Garcia, M., Ruzo, A., Villacampa, P., Ayuso, E., Ther. Methods Clin. Dev. 10, 68–78. Añor, S., Andaluz, A., Pineda, M., et al. (2013). Whole body correction of mucopo- lysaccharidosis IIIA by intracerebrospinal fluid gene therapy. J. Clin. Invest. 123, 51. Ballon, D.J., Rosenberg, J.B., Fung, E.K., Nikolopoulou, A., Kothari, P., De, B.P., He, 3254–3271. B., Chen, A., Heier, L.A., Sondhi, D., et al. (2020). Quantitative Whole-Body Imaging of I-124-Labeled Adeno-Associated Viral Vector Biodistribution in Nonhuman 35. Motas, S., Haurigot, V., Garcia, M., Marcó, S., Ribera, A., Roca, C., Sánchez, X., Primates. Hum. Gene Ther. 31, 1237–1259. Sánchez, V., Molas, M., Bertolin, J., et al. (2016). CNS-directed gene therapy for 52. Muenzer, J., Hendriksz, C.J., Fan, Z., Vijayaraghavan, S., Perry, V., Santra, S., Solanki, the treatment of neurologic and somatic mucopolysaccharidosis type II (Hunter syn- G.A., Mascelli, M.A., Pan, L., Wang, N., et al. (2016). A phase I/II study of intrathecal drome). JCI Insight 1, e86696. idursulfase-IT in children with severe mucopolysaccharidosis II. Genet. Med. 18, 36. Gray, S.J., Nagabhushan Kalburgi, S., McCown, T.J., and Jude Samulski, R. (2013). 73–81. Global CNS gene delivery and evasion of anti-AAV-neutralizing antibodies by intra- 53. Marcó, S., Haurigot, V., and Bosch, F. (2019). In Vivo Gene Therapy for thecal AAV administration in non-human primates. Gene Ther. 20, 450–459. Mucopolysaccharidosis Type III (Sanfilippo Syndrome): A New Treatment 37. Samaranch, L., Salegio, E.A., San Sebastian, W., Kells, A.P., Foust, K.D., Bringas, J.R., Horizon. Hum. Gene Ther. 30, 1211–1221. Lamarre, C., Forsayeth, J., Kaspar, B.K., and Bankiewicz, K.S. (2012). Adeno-associ- 54. Chen, X.J., Kovacevic, N., Lobaugh, N.J., Sled, J.G., Henkelman, R.M., and ated virus serotype 9 transduction in the central nervous system of nonhuman pri- Henderson, J.T. (2006). Neuroanatomical differences between mouse strains as mates. Hum. Gene Ther. 23, 382–389. shown by high-resolution 3D MRI. Neuroimage 29, 99–105. 38. Gurda, B.L., De Guilhem De Lataillade, A., Bell, P., Zhu, Y., Yu, H., Wang, P., Bagel, J., 55. Vincent, T.J., Thiessen, J.D., Kurjewicz, L.M., Germscheid, S.L., Turner, A.J., Zhilkin, Vite, C.H., Sikora, T., Hinderer, C., et al. (2016). Evaluation of AAV-mediated Gene P., Alexander, M.E., and Martin, M. (2010). Longitudinal Brain Size Measurements in Therapy for Central Nervous System Disease in Canine Mucopolysaccharidosis VII. App/Ps1 Transgenic Mice. Magn. Reson. Insights 4, 19–26. Mol. Ther. 24, 206–216. 56. Roderick, T.H., Wimer, R.E., Wimer, C.C., and Schwartzkroin, P.A. (1973). Genetic 39. Ribera, A., Haurigot, V., Garcia, M., Marcó, S., Motas, S., Villacampa, P., Maggioni, L., and phenotypic variation in weight of brain and spinal cord between inbred strains of León, X., Molas, M., Sánchez, V., et al. (2015). Biochemical, histological and func- mice. Brain Res. 64, 345–353. tional correction of mucopolysaccharidosis type IIIB by intra-cerebrospinal fluid 57. Ma, Y., Hof, P.R., Grant, S.C., Blackband, S.J., Bennett, R., Slatest, L., McGuigan, gene therapy. Hum. Mol. Genet. 24, 2078–2095. M.D., and Benveniste, H. (2005). A three-dimensional digital atlas database of the 40. Hinderer, C., Bell, P., Gurda, B.L., Wang, Q., Louboutin, J.-P., Zhu, Y., Bagel, J., adult C57BL/6J mouse brain by magnetic resonance microscopy. Neuroscience O’Donnell, P., Sikora, T., Ruane, T., et al. (2014). Intrathecal gene therapy corrects 135, 1203–1215. CNS pathology in a feline model of mucopolysaccharidosis I. Mol. Ther. 22, 2018– 58. Lange, N., Froimowitz, M.P., Bigler, E.D., and Lainhart, J.E.; Brain Development 2027. Cooperative Group (2010). Associations between IQ, total and regional brain Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 387 Molecular Therapy: Methods & Clinical Development volumes, and demography in a large normative sample of healthy children and ado- 75. Nestrasil, I., Shapiro, E., Svatkova, A., Dickson, P., Chen, A., Wakumoto, A., Ahmed, lescents. Dev. Neuropsychol. 35, 296–317. A., Stehel, E., McNeil, S., Gravance, C., and Maher, E. (2017). Intrathecal enzyme replacement therapy reverses cognitive decline in mucopolysaccharidosis type I. 59. Katz, M.L., Tecedor, L., Chen, Y., Williamson, B.G., Lysenko, E., Wininger, F.A., Am. J. Med. Genet. A. 173, 780–783. Young, W.M., Johnson, G.C., Whiting, R.E.H., Coates, J.R., and Davidson, B.L. (2015). AAV gene transfer delays disease onset in a TPP1-deficient canine model 76. Edelmann, M.J., and Maegawa, G.H.B. (2020). CNS-Targeting Therapies for of the late infantile form of Batten disease. Sci. Transl. Med. 7, 313ra180. Lysosomal Storage Diseases: Current Advances and Challenges. Front. Mol. Biosci. 7, 559804. 60. Fu, H., DiRosario, J., Kang, L., Muenzer, J., and McCarty, D.M. (2010). Restoration of central nervous system alpha-N-acetylglucosaminidase activity and therapeutic ben- 77. Nakafuku, M., Nagao, M., Grande, A., and Cancelliere, A. (2008). Revisiting neural efits in mucopolysaccharidosis IIIB mice by a single intracisternal recombinant ad- stem cell identity. Proc. Natl. Acad. Sci. USA 105, 829–830. eno-associated viral type 2 vector delivery. J. Gene Med. 12, 624–633. 78. Johansson, C.B., Momma, S., Clarke, D.L., Risling, M., Lendahl, U., and Frisén, J. 61. Hemsley, K.M., Norman, E.J., Crawley, A.C., Auclair, D., King, B., Fuller, M., Lang, (1999). Identification of a neural stem cell in the adult mammalian central nervous D.L., Dean, C.J., Jolly, R.D., and Hopwood, J.J. (2009). Effect of cisternal sulfamidase system. Cell 96, 25–34. delivery in MPS IIIA Huntaway dogs–a proof of principle study. Mol. Genet. Metab. 79. Coskun, V., Wu, H., Blanchi, B., Tsao, S., Kim, K., Zhao, J., Biancotti, J.C., Hutnick, L., 98, 383–392. Krueger, R.C., Jr., Fan, G., et al. (2008). CD133+ neural stem cells in the ependyma of 62. Beard, H., Luck, A.J., Hassiotis, S., King, B., Trim, P.J., Snel, M.F., Hopwood, J.J., and mammalian postnatal forebrain. Proc. Natl. Acad. Sci. USA 105, 1026–1031. Hemsley, K.M. (2015). Determination of the role of injection site on the efficacy of 80. Kataoka, Y., Cui, Y., Tamura, Y., and Yamada, H. (2004). Proliferation of intra-CSF enzyme replacement therapy in MPS IIIA mice. Mol. Genet. Metab. 115, Leptomeningeal Cells in Delayed Neuronal Death in Gerbils. Acta Histochem. 33–40. Cytochem. 37, 81–86. 63. Arruda, V.R., Stedman, H.H., Haurigot, V., Buchlis, G., Baila, S., Favaro, P., Chen, Y., 81. Di Terlizzi, R., and Platt, S. (2006). The function, composition and analysis of cere- Franck, H.G., Zhou, S., Wright, J.F., et al. (2010). Peripheral transvenular delivery of brospinal fluid in companion animals: part I - function and composition. Vet. J. 172, adeno-associated viral vectors to skeletal muscle as a novel therapy for hemophilia B. 422–431. Blood 115, 4678–4688. 82. Sakka, L., Coll, G., and Chazal, J. (2011). Anatomy and physiology of cerebrospinal 64. Nathwani, A.C., Rosales, C., McIntosh, J., Rastegarlari, G., Nathwani, D., Raj, D., fluid. Eur. Ann. Otorhinolaryngol. Head Neck Dis. 128, 309–316. Nawathe, S., Waddington, S.N., Bronson, R., Jackson, S., et al. (2011). Long-term safety and efficacy following systemic administration of a self-complementary 83. Scallan, C.D., Jiang, H., Liu, T., Patarroyo-White, S., Sommer, J.M., Zhou, S., AAV vector encoding human FIX pseudotyped with serotype 5 and 8 capsid proteins. Couto, L.B., and Pierce, G.F. (2006). Human immunoglobulin inhibits liver trans- Mol. Ther. 19, 876–885. duction by AAV vectors at low AAV2 neutralizing titers in SCID mice. Blood 107, 1810–1817. 65. Hinderer, C., Bell, P., Louboutin, J.-P., Zhu, Y., Yu, H., Lin, G., Choa, R., Gurda, B.L., Bagel, J., O’Donnell, P., et al. (2015). Neonatal Systemic AAV Induces Tolerance to 84. Jiang, H., Couto, L.B., Patarroyo-White, S., Liu, T., Nagy, D., Vargas, J.A., Zhou, S., CNS Gene Therapy in MPS I Dogs and Nonhuman Primates. Mol. Ther. 23, 1298– Scallan, C.D., Sommer, J., Vijay, S., et al. (2006). Effects of transient immunosuppres- 1307. sion on adenoassociated, virus-mediated, liver-directed gene transfer in rhesus ma- caques and implications for human gene therapy. Blood 108, 3321–3328. 66. Hinderer, C., Bell, P., Louboutin, J.-P., Katz, N., Zhu, Y., Lin, G., Choa, R., Bagel, J., O’Donnell, P., Fitzgerald, C.A., et al. (2016). Neonatal tolerance induction enables ac- 85. Hordeaux, J., Buza, E.L., Jeffrey, B., Song, C., Jahan, T., Yuan, Y., Zhu, Y., Bell, P., Li, curate evaluation of gene therapy for MPS I in a canine model. Mol. Genet. Metab. M., Chichester, J.A., et al. (2020). MicroRNA-mediated inhibition of transgene 119, 124–130. expression reduces dorsal root ganglion toxicity by AAV vectors in primates. Sci. Transl. Med. 12, eaba9188. 67. Wang, D., Tai, P.W.L., and Gao, G. (2019). Adeno-associated virus vector as a plat- form for gene therapy delivery. Nat. Rev. Drug Discov. 18, 358–378. 86. Hordeaux, J., Hinderer, C., Goode, T., Katz, N., Buza, E.L., Bell, P., Calcedo, R., Richman, L.K., and Wilson, J.M. (2018). Toxicology Study of Intra-Cisterna Magna 68. Dhungel, B.P., Bailey, C.G., and Rasko, J.E.J. (2021). Journey to the Center of the Cell: Adeno-Associated Virus 9 Expressing Human Alpha-L-Iduronidase in Rhesus Tracing the Path of AAV Transduction. Trends Mol. Med. 27, 172–184. Macaques. Mol. Ther. Methods Clin. Dev. 10, 79–88. 69. Higuchi, T., Shimizu, H., Fukuda, T., Kawagoe, S., Matsumoto, J., Shimada, Y., 87. Arruda, V.R., and Doshi, B.S. (2020). Gene Therapy for Hemophilia: Facts and Kobayashi, H., Ida, H., Ohashi, T., Morimoto, H., et al. (2012). Enzyme replacement Quandaries in the 21st Century. Mediterr. J. Hematol. Infect. Dis. 12, e2020069. therapy (ERT) procedure for mucopolysaccharidosis type II (MPS II) by intraventric- ular administration (IVA) in murine MPS II. Mol. Genet. Metab. 107, 122–128. 88. Bradbury, A.M., Bagel, J.H., Nguyen, D., Lykken, E.A., Pesayco Salvador, J., Jiang, X., Swain, G.P., Assenmacher, C.A., Hendricks, I.J., Miyadera, K., et al. (2020). Krabbe 70. Jones, S.A., Breen, C., Heap, F., Rust, S., de Ruijter, J., Tump, E., Marchal, J.P., Pan, L., disease successfully treated via monotherapy of intrathecal gene therapy. J. Clin. Qiu, Y., Chung, J.-K., et al. (2016). A phase 1/2 study of intrathecal heparan-N-sul- Invest. 130, 4906–4920. fatase in patients with mucopolysaccharidosis IIIA. Mol. Genet. Metab. 118, 198–205. 89. Jolly, R.D., Allan, F.J., Collett, M.G., Rozaklis, T., Muller, V.J., and Hopwood, J.J. 71. Wijburg, F.A., Whitley, C.B., Muenzer, J., Gasperini, S., Del Toro, M., Muschol, N., (2000). Mucopolysaccharidosis IIIA (Sanfilippo syndrome) in a New Zealand Cleary, M., Sevin, C., Shapiro, E., Bhargava, P., et al. (2019). Intrathecal heparan- Huntaway dog with ataxia. N. Z. Vet. J. 48, 144–148. N-sulfatase in patients with Sanfilippo syndrome type A: A phase IIb randomized trial. Mol. Genet. Metab. 126, 121–130. 90. Fischer, A., Carmichael, K.P., Munnell, J.F., Jhabvala, P., Thompson, J.N., Matalon, R., Jezyk, P.F., Wang, P., and Giger, U. (1998). Sulfamidase deficiency in a family 72. Cleary, M., Muschol, N., Couce, M.L., Harmatz, P., Lee, J., Lin, S.-P., Okur, I., Ezgu, F., of Dachshunds: a canine model of mucopolysaccharidosis IIIA (Sanfilippo A). Peters, H., Solano Villarreal, M., et al. (2019). ICV-administered tralesinidase alfa Pediatr. Res. 44, 74–82. (BMN 250 NAGLU-IGF2) is well-tolerated and reduces heparan sulfate accumula- tion in the CNS of subjects with Sanfilippo syndrome type B (MPS IIIB). Mol. 91. Jolly, R.D., Ehrlich, P.C., Franklin, R.J., Macdougall, D.F., and Palmer, A.C. (2001). Genet. Metab. 126, S40. Histological diagnosis of mucopolysaccharidosis IIIA in a wire-haired dachshund. Vet. Rec. 148, 564–567. 73. Calias, P., Banks, W.A., Begley, D., Scarpa, M., and Dickson, P. (2014). Intrathecal delivery of protein therapeutics to the brain: a critical reassessment. Pharmacol. 92. Yogalingam, G., Pollard, T., Gliddon, B., Jolly, R.D., and Hopwood, J.J. (2002). Ther. 144, 114–122. Identification of a mutation causing mucopolysaccharidosis type IIIA in New Zealand Huntaway dogs. Genomics 79, 150–153. 74. Okuyama, T., Eto, Y., Sakai, N., Minami, K., Yamamoto, T., Sonoda, H., Yamaoka, M., Tachibana, K., Hirato, T., and Sato, Y. (2019). Iduronate-2-Sulfatase with Anti- 93. Crawley, A.C., Marshall, N., Beard, H., Hassiotis, S., Walsh, V., King, B., Hucker, N., human Transferrin Receptor Antibody for Neuropathic Mucopolysaccharidosis II: Fuller, M., Jolly, R.D., Hopwood, J.J., and Hemsley, K.M. (2011). Enzyme replacement A Phase 1/2 Trial. Mol. Ther. 27, 456–464. reduces neuropathology in MPS IIIA dogs. Neurobiol. Dis. 43, 422–434. 388 Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 www.moleculartherapy.org 94. Herzog, R.W., Mount, J.D., Arruda, V.R., High, K.A., and Lothrop, C.D., Jr. (2001). 96. Marcó, S., Pujol, A., Roca, C., Motas, S., Ribera, A., Garcia, M., Molas, M., Villacampa, Muscle-directed gene transfer and transient immune suppression result in sustained P., Melia, C.S., Sánchez, V., et al. (2016). Progressive neurologic and somatic disease partial correction of canine hemophilia B caused by a null mutation. Mol. Ther. 4, in a novel mouse model of human mucopolysaccharidosis type IIIC. Dis. Model. 192–200. Mech. 9, 999–1013. 95. Ayuso, E., Mingozzi, F., Montane, J., Leon, X., Anguela, X.M., Haurigot, V., Edmonson, 97. Manno, C.S., Pierce, G.F., Arruda, V.R., Glader, B., Ragni, M., Rasko, J.J., Ozelo, M.C., S.A., Africa, L., Zhou, S., High, K.A., et al. (2010). High AAV vector purity results in sero- Hoots, K., Blatt, P., Konkle, B., et al. (2006). Successful transduction of liver in hemo- type- and tissue-independent enhancement of transduction efficiency. Gene Ther. 17, philia by AAV-Factor IX and limitations imposed by the host immune response. Nat. 503–510. Med. 12, 342–347. Molecular Therapy: Methods & Clinical Development Vol. 23 December 2021 389