Intracerebral administration of recombinant adeno-associated vector (AAV) has been performed in several clinical trials. However, delivery into the brain requires multiple injections and is not efficient to target the spinal cord, thus limiting its applications. To assess widespread and less invasive strategies, we tested intravenous (IV) or intrathecal (that is, in the cerebrospinal fluid (CSF)) delivery of a rAAVrh10-egfp vector in adult and neonate rats and studied the effect of the age at injection on neurotropism. IV delivery is more efficient in neonates and targets predominantly Purkinje cells of the cerebellum and sensory neurons of the spinal cord and dorsal root ganglia. A single intra-CSF administration of AAVrh10, single strand or oversized self-complementary, is efficient for the targeting of neurons in the cerebral hemispheres, cerebellum, brainstem and spinal cord. Green fluorescent protein (GFP) expression is more widespread in neonates when compared with adults. More than 50% of motor neurons express GFP in the three segments of the spinal cord in neonates and in the cervical and thoracic regions in adults. Neurons are almost exclusively transduced in neonates, whereas neurons, astrocytes and rare oligodendrocytes are targeted in adults. These results expand the possible routes of delivery of AAVrh10, a serotype that has shown efficacy and safety in clinical trials concerning neurodegenerative diseases.
Neurodegenerative diseases represent a therapeutic challenge due to the blood–brain barrier and the frequent widespread distribution of the associated lesions: upper and/or lower motor neurons (MNs; amyotrophic lateral sclerosis, spinal muscular atrophy), sensory neurons (Friedriech’s ataxia), brain cerebellum and spinal cord (lysosomal storage diseases). Among these, lysosomal storage diseases (LSD) are good candidates for gene therapy. They are almost always monogenic autosomal recessive disorders without available curative therapy, and their correction is facilitated by secretion/uptake and diffusion of the therapeutic transgenic protein.1 Currently, direct delivery of recombinant adeno-associated vectors (rAAV) into the brain parenchyma is utilized to treat neurodegenerative diseases. The rAAV serotype 2 was initially shown to successfully deliver transgene products to the central nervous system (CNS) after intracerebral injection in rodents.2 Serotypes 1, 2 and 5 were then used with success in large animals models3, 4, 5, 6, 7 and more recently serotype 9 and rh10 in rodents8, 9, 10 and dogs.11 Serotype 9 and rh10 both show increased efficacy compared with the classical serotypes in terms of neurotropism and diffusion around the injection site.8, 11 These successes have led to the application of this approach to the treatment of neurogenetic diseases, and intracerebral rAAVrh10 therapeutic injection is currently being tested in patients with LSD (Sanfilippo type A NCT01474343, Metachromatic Leukodystrophy NCT01801709, Batten disease NCT01161576 and NCT01414985 on ClinicalTrials.gov). However, direct delivery into the brain is invasive and results in robust but relatively local transduction, even with new serotypes, thus requiring multiple parenchymal injections. Such local strategies are well adapted for the treatment of neurological disorders such as Parkinson’s disease, which are localized in specific neuroanatomical regions. By contrast, the treatment of widespread neurological dysfunctions such as LSD requires the efficient transduction of a large volume of brain parenchyma.
Therefore less invasive and more widespread CNS-targeting strategies have been developed, mostly with rAAV9, which has been shown to transduce the CNS of rodents and large animals after systemic12, 13, 14, 15 or intrathecal (IT)16, 17, 18, 19, 20, 21, 22 administration. On the contrary, despite its current utilization in clinical trials, the potency of rAAVrh10 to target the CNS by non-invasive routes such as intravenous (IV) delivery has only been explored in the context of neonatal injection.23, 24, 25 It is noteworthy that the effect of the age at injection on neurotropism has been demonstrated with serotype 9 only.12 Besides, due to their packaging capacity, single-strand AAV (ssAAV) vectors are currently used for most therapeutic transgenes, even though self-complementary AAV (scAAV) vectors are supposed to allow better transgene expression than ssAAV because they do not require DNA synthesis or base pairing between multiple vector genomes.26 The use of oversized genomes in rAAV leads to the encapsidation of various sized fragmented genomes.27, 28 The in vivo efficacy of such oversized ssAAV vectors has been reported in the literature and seems to involve the DNA repair proteins-mediated complementation between the fragmented genomes.27, 29 The compared efficacy between regular ss and oversized sc (o-sc) genomes has not been reported yet.
In the present study, we investigate the pattern of CNS transduction using ss and o-scAAVrh10 vectors encoding green fluorescent protein (GFP), with administration via the intracisternal (that is, in the cisterna magna) or IV routes in neonate and adult Sprague Dawley rats. Our specific objectives are (i) to target CNS with a single non-invasive injection of AAVrh10; (ii) to investigate the effect of the age at injection (adults vs neonates) and the type of vector genome (ss vs o-sc) on the CNS gene transfer efficiency; and (iii) to assess the safety of our approach by evaluating the off-target gene expression and the humoral immune response.
Strong and global CNS transduction after a single intracisternal administration of rAAVrh10
We aim to compare the pattern of transgene expression and the distribution of viral genomes in rat CNS after injection of ssAAVrh10 or o-scAAVrh10 in the cisterna magna of adult or neonate rats. Each vector is encoded for the reporter GFP protein under the control of CAG promoter and is injected at doses of 4 × 1011 vg in adults (2 × 1012 vg kg−1) and of 1011 vg in neonates (2 × 1013 vg kg−1). Four-to-five weeks after the injection, we observe GFP expression in neurons across the whole brain from the frontal lobe to the medulla oblongata in both the age groups and with both the vectors (Figure 1). The proportion of transduced neurons is closely similar to either ss or o-sc vectors (Figures 1a–l). Animals injected when they are neonates (referred as ‘neonates’ for the sake of simplification) display a high level of transgene expression in the whole brain (Figure 1m), while transduced neurons in adults are more numerous in the caudal cortex, the cerebellum and the brainstem, that is, as far as 12 mm rostral to the injection site. In the cerebral cortex, the pattern of transduction is diffuse in neonates, whereas it is patchy in adults, associated with clusters of positive neurons sometimes surrounding vascular lumen (Figure 1a). Purkinje cells (Figures 1c, d, i and j), bulbar MNs and reticular substance neurons (Figures 1e, f, k and l) strongly express GFP in the brainstem regardless of the age at injection. GFP-positive neurons are also present, albeit less numerous in adults, in the cerebral cortex (olfactory, piriform, motor, somatosensorial, visual and retrosplenial regions), the basal ganglia (caudate nucleus), the hippocampus and the thalamic nuclei, as well as in the midbrain, including the inferior and superior colliculi, medial geniculate nucleus and substancia nigra.
The entire spinal cord is transduced for both ages at injection and with both vector constructions, with detection of GFP expression in the dorsal horn sensory neurons and the lower MNs. Medullary sensory ascending and motor descending tracts (Figures 2a and b) and dorsal root ganglia (DRG) sensory neurons (Figure 2d) show robust GFP signals in the full length of the spinal cord for both adults (Figures 2a and b) and neonates (data not shown). The proportion of MNs with GFP expression in adults is high in the cervical (46%±12.1 with ssAAVrh10 and 100% with o-scAAVrh10) and thoracic spinal cord (53%±13.3 with ssAAVrh10 and 51%±12.6 with o-scAAVrh10), while it falls to <25% in the lumbar segment (that is, remote from the injection site) with both the vectors. On the contrary, the transduction of MNs in neonates is robust and consistent in the three segments: 70% (±9.1) and 71% (±9.7) in the cervical part, 98% (±0.9) and 61% (±10.4) in the thoracic part, and 74% (±9.5) and 56% (±11.5) in the lumbar part, for the ss and o-sc AAVrh10 vectors, respectively.
Whatever the location (brain, spinal cord or DRG), neurons represent the main type of transduced cells. However, immunofluorescent colocalization studies show that numerous astrocytes and some oligodendrocytes are also transduced in adults in the vicinity of the injection site (brainstem and cervical spinal cord—Figure 3). On the contrary, only few astrocytes and no oligodendrocyte are transduced at distance from the injection site, and neurons are almost exclusively targeted in neonates (Supplementary Figure S1). Iba1, a microglial cell marker, did not colocalize with GFP in any animal (data not shown).
Importantly, oversized o-sc and ss vector constructions lead to similar GFP expression and similar vector genome copy number in the CNS (Figure 1 and Table 1). The mean vector genome copy number is up to 0.91 (±0.76) vector genome per diploid genome ( vg dg−1) in neonates and up to 6.3 (±4.89) vg dg−1 in adults (detection limit 0.01 vg dg−1). The highest level of transduction is found in the cervical segment of adults, while it is similar in the three segments of neonates (Table 1), which is coherent with the pattern of transgene expression. The small size of the groups and the variation of vector genome quantification observed among all groups cannot allow concluding about differences obtained in adults versus neonates. However, the same variability and similar mean vg dg−1 are observed with ss and o-sc vectors.
Limited CNS transduction after IV administration of rAAVrh10
In this study, we compare the distribution of transgene expression and the amount of viral genomes in the rat CNS after IV administration of ssAAVrh10 or o-scAAVrh10 in adult or neonate (2 × 1013 vg kg−1 in both the age groups). GFP expression in the retina of rats injected IV was monitored in vivo at weekly intervals by fluorescence retinal imaging. Retinal GFP expression was used as an in vivo tracer of gene transfer and was only done in the IV group. In adults, scarce GFP-positive retinal cells are present from 2 weeks p.i. (postinjection) onward only when using o-scAAVrh10 (Figure 4d). The few fluorescent cells detected in vivo are ganglion cells and their corresponding axons as seen on neuroretinal flat mount sections observed by confocal microscopy (Figure 4e). In neonates, numerous GFP-positive retinal cells are present as soon as 3 weeks p.i. (first time point) with both the ss and o-sc vectors (Figures 4f and g) and remained until the killing of the animals. Ganglion cells and retinal pigment epithelial cells are both transduced in neonates. Ss and o-sc vectors lead to the same amount of GFP-positive retinal cells in neonates, while o-scAAVrh10 only transduces the retina in adult.
Animals were euthanized 4–5 weeks after the IV administration to study the pattern of transduction of rAAVrh10. A low number of neurons are transduced in both the age groups. GFP expression is detected in pyramidal neurons of the cerebral cortex, in Purkinje cells of the cerebellum, in trigeminal spinal nuclei neurons and white matter tracts, as well as in trigeminal sensory ganglionic neurons (Figure 5). In all these structures, transduction efficiency is higher in neonates than in adults, and only neurons are targeted. Basal ganglia, hippocampus, thalamic nuclei, midbrain structures and brainstem motor nuclei neurons are weakly or inconsistently transduced (data not shown).
In the spinal cord, transduction is limited to the sensory neurons in adults; sensory neurons and less than one MN per cross-section express GFP in neonates (Supplementary Figure S2). DRG (Supplementary Figures S2c and f) and the corresponding ascendant sensory tracts show the highest GFP expression in both the age groups, whereas lower MNs are inconsistently targeted in neonates only (Supplementary Figures S2b and e). Blood–brain-barrier-free areas such as the DRG or trigeminal ganglion are predominantly transduced in both the age groups. No difference was observed between ss and o-sc vectors regarding the CNS transduction.
In contrast to the CNS, peripheral organs (especially the heart, liver, diaphragm and adrenal gland) are consistently targeted after IV injection in both the age groups; moreover, all the explored striated muscles are highly transduced in neonates (Supplementary Figure S3). Quantitative PCR (qPCR) shows that vector genome copies per diploid genome are low in the CNS with both constructions (0.01–0.1 vg per diploid genome in the three spinal cord segments) and that o-scAAVrh10 yields a higher level of transduction than ssAAVrh10 in the heart of adults and in the liver of neonates (Figure 6a).
Intracisternal and IV administration of rAAVrh10 both cause off-target transduction and immunization
The aim of this study was to target the CNS with a minimally invasive strategy using rAAVrh10. The working hypothesis tested here is that the intracisternal route would lead to a reduced peripheral transgene expression and to a lesser degree of immunization against the transgene product and/or vector when compared with the IV route. GFP off-target expression is detected in the peripheral organs with both administration routes. The observation of GFP and the qPCR results reveal liver and heart as being the main off-target organs in both the age groups and with both the vectors (Figure 6). The number of GFP-expressing cells (Supplementary Figure S3) and vector genome copy numbers per diploid genomes (Figure 6) are surprisingly similar after IV and IT administration in the heart and liver.
The measurement of anti-GFP antibodies by enzyme-linked immunosorbent assay in the sera of rats 2 and 4 weeks after injection shows that both administration routes lead to a strong anti-GFP humoral response in adults as soon as 2 weeks p.i., whereas no anti-GFP antibody is detected in any rat injected neonatally (Supplementary Figure S4). In both adults and neonates, administration of rAAVrh10 also leads to anti-capsid immunization, with detection of higher levels of anti-AAVrh10 neutralization factors in adults compared with neonates and no difference between ss and o-sc vector genomes or between IT and IV routes. At 1 month p.i., the greatest serum dilution with neutralization factors is 1/102–1/5 × 105 in adults (Supplementary Table S1) and 1/10–1/104 in neonates (Supplementary Table S2). Importantly, Hemalun–eosin-stained sections of the CNS, liver and heart did not show any inflammatory lesion, suggesting that the humoral response against the transgene and capsid did not cause adverse effect in the organs that are the most efficiently transduced.
Our results demonstrate efficient transgene expression in the neurons of the brain, the cerebellum, the brainstem (including MNs) and the spinal cord (sensory and MNs) after a single IT injection of either oversized o-sc or ssAAVrh10 in adults and neonates. The surgical procedure was well tolerated both in adults and neonates. Only one minor adverse event occurred in an adult rat that displayed a slight head tilt for 2 days after the injection. Histopathological examination revealed a mild well-circumscribed traumatic lesion due to accidental puncture of the medulla oblongata during cerebrospinal fluid (CSF) sampling. The procedure was otherwise well tolerated in all animals.
Research concerning the IT rAAV-mediated gene transfer started approximately 10 years ago. Initially, efficient DRG and dorsal horn transduction was achieved in rodents with various serotypes, such as ssAAV2,30 scAAV1 and 5,31 ssAAV6,32 ssAAV5 and 8.33 In all these studies, viral suspensions were injected into the lumbar IT space, and transduction was maximal in the lumbar spinal cord DRG and dorsal horns. Transduction of the MNs was not specified in these publications concerning the sensory neurons only. In the study by Vulchanova et al.,33 the injection of 1011 vg in 10 μl to 20 g mice was associated with IV mannitol pretreatment in order to enhance the penetration of AAV5-GFP and AAV8-GFP vectors to the spinal cord parenchyma. They indeed obtained transduction of few MNs using mannitol pretreatment.33 Our findings of consistent MN transduction without any pretreatment show the good neurotropism and penetration capacity of the serotype rh10. The dose and volume of injection that we used (4 × 1011 vg in 50 μl injected to 200 g rats) is comparable to those used in mice, and it is thus probably the serotype that is determinant in our finding of consistent MN transduction. In the study by Fu et al.,34 the intracisternal delivery of 3 × 1010 vg of scAAV2 in 15 μl to mice pretreated with IV mannitol leads to a GFP expression mainly restricted to periventricular regions with only occasional transduction of cortical neurons and lower MNs, suggesting that the serotype rather than the intracisternal route or pretreatment is determinant for MN transduction. When injected into the lumbar IT space (2.5 × 109 vg in 10 μl) or the lateral ventricle of mice (4 × 1010 vg in 3 μl), the neurotropic serotypes 6 and 9 both efficiently and diffusely targeted lower MNs even with moderate doses.16, 35 The efficient MN transduction obtained in rodents with rAAV9 was successfully translated to large animals, such as pigs (4.1012 vg in 1.5 ml),17 cats (1012 vg kg−1 in 1 ml kg−1),22 dogs (2.1013 vg in 1 ml)20 and non-human primates (2.1012 vg in 1.5 ml and 2.1013 vg in 2 ml).18, 19, 21 The data collected on large animals show that serotype 9 efficiently targets MNs in various species and that the intracisternal route or multiple bolus dose infusion along the spinal cord allow distribution of the vector remote from the injection site, demonstrating the feasibility of the strategy in larger models.
The results presented here demonstrate that rAAVrh10 is also an efficient serotype for lower MN targeting and that it can also transduce upper MNs, subcortical neurons, Purkinje cells and brainstem neurons, thus expanding the field of potential therapeutic applications.
The serotype rh10 has been discovered from rhesus monkey tissues, whereas serotype 9 is a human AAV.36 In one study, preexisting antibodies against AAV9 are found in 47% of human patients.37 Data are lacking in human for AAVrh10 but preexistent immunization is expected to be less due to its non-human primate origin. Few publications directly compare rAAV9 and rAAVrh10 under the same conditions. These studies indicate close behaviour and efficiency for CNS targeting after neonatal IV injection,23, 25 and the equal efficiency of rAAV9 and rh10 after direct intracerebral injection.9 Using lumbar IT administration, Wang et al.38 compared serotypes 1, 2, 8, 9 and rh10 and showed that only rAAV9 and rh10 can transduce the full-length spinal cord in mice. These authors reported that both serotypes were equivalent, and efficiently translated their results to a non-human primate using rAAVrh10 with a dose comparable to ours (2.7 × 1012 vg in 250 μl injected to the lumbar IT space of a 600-g marmoset). Interestingly, using the same promoter as in the present study (CAG), Wang et al.38 obtained transduction of predominantly astrocytes in mice and MN in primates. Gray et al.39 also observed low expression of the transgene product in the MN of mice when using the truncated 800 bp CMV/CBA promoter hybrid, whereas we obtain robust MN transduction in rats using the 1.6-kb CAG promoter.
The behaviour of the serotypes after intra-CSF injection is also likely related to the mechanism of entry of the viral particles into the nervous parenchyma. Indeed, the capsid determines the interactions between the virus and the barriers that protect the CNS. Passage from the CSF to the nervous parenchyma can occur by (i) transcytosis across the pia mater, (ii) entry from zones where the pia mater lining is absent, such as in spinal roots or perivascular Virchow–Robin spaces40 or (iii) entry across the ependymal epithelium. We show consistent transduction of DRG, which suggests viral entry through the spinal roots. Besides, transduction of the frontal cerebral cortex, despite the distance from the injection site, can be explained by viral entry through the Virchow–Robin space surrounding superficial penetrating blood vessels of the cortex. The detection of clusters of transduced neurons surrounding vascular lumens in the cortex of adults after IT administration confirms the important role of Virchow–Robin spaces in IT gene transfer. Moreover, this hypothesis is supported by the recent description of the glymphatic system, a clearance mechanism in which CSF enters the brain along para-arterial channels, exchanges with interstitial fluids of the nervous parenchyma and is cleared from the brain along para-veinous pathways.40 The clustering of transduced cells around blood vessels has also been described by Samaranch et al.,19 who also interpret this as due to a vector flow from CSF into the brain through the Virchow–Robin spaces.
The second objective of our study was to assess the effect of the type of vector genome (ss versus oversized o-sc) and the age at injection (neonate versus adult) on the efficacy of transduction and on neurotropism. The superiority of sc vectors was initially described with serotype 2 in the liver but was then qualified following the use of rAAV serotypes capable of highly efficient liver gene delivery, such as rAAV8.26 With rAAVrh10, we observe a similar efficiency of ss and o-sc vectors in the CNS after intracisternal injection. Because of the oversized genome used in our scAAV, only a part of the packaged genome is in a full-length dimerization state leading to a likely reduced efficiency in vivo compared with a regular sc genome. In the context of intra-CSF injection, numerous viral particles are diluted in a small volume of fluid, so the multiple infection of a single cell is thus likely. Complementation of fragmented genome that has been reported in the context of oversized ss genomes27, 29 could thus explain why ss and o-scAAVrh10 are equally efficient when injected at high dose into a restricted compartment, such as the CSF.26, 41
Concerning the effect of the age at injection, we observe, as expected, global CNS transduction after IV injection in neonates as reported previously with rAAV913, 15, 22 and rAAVrh10.23 This is probably due to tight junction immaturity in the neonatal blood–brain barrier. On the contrary, in adults, CNS transduction following IV administration is weak and mostly limited to blood–brain barrier-free zones, suggesting that serotype 10 cannot efficiently cross the endothelial cells by transcytosis. After intracisternal injection, the transgene expression in the rostral brain and caudal spinal cord is stronger in neonates compared with adults. Higher diffusion of viral particles in the immature nervous tissue could be explained by (i) the higher water content of neonate brain parenchyma, (ii) the unique plasticity of the extracellular matrix in immature brain allowing astroglial and neuronal migration and (iii) the high volume injected to the neonates when compared with the adults. Regarding the second point, the matrix metalloproteinases active during the early development allow CNS plasticity by cleavage of extracellular matrix proteins, such as collagen or laminin42 (for a review). An alternative hypothesis involves the transduction of subventricular neural stem cells in direct contact with the CSF and their subsequent migration during brain maturation, including via the rostral migratory stream. Previous studies have shown that rAAV9 will transduce neurons when injected IV into neonates and astrocytes when injected into adults.12 In our study, neurons are predominantly transduced after IV injection in both the age groups. After intracisternal administration, we observe an almost pure neuronal tropism in neonates, while some astrocytes and oligodendrocytes are also transduced in adults.
In conclusion, our results demonstrate efficient and global neuronal transduction with administration of rAAVrh10 with both ss and oversized sc genomes using the intracisternal route. As this serotype is currently being used for intracerebral gene therapy clinical trials, the characterization of its behaviour after non-invasive delivery routes should allow launching new trials for the treatment of diffuse neurodegenerative diseases. The application for the treatment of LSD is especially promising, because secretion-uptake mechanism of the transgenic enzyme enhances its distribution in the CNS and because partial enzymatic restoration is often enough to restore the function of the lysosomes.
Materials and methods
Pseudotyped rAAVrh10 vectors were generated by packaging AAV2-based recombinant ss or oversized sc genomes into AAVrh10 capsids, as previously described.43 Briefly, the vectors were produced by helper virus-free co-transfection in HEK293 cells, using (i) the adenovirus helper and AAV packaging plasmid encoding the adenoviral genes together with the rep2 and cap10 genes and (ii) the AAV2 plasmid containing the gene encoding GFP (under control of CAG, that is, the cytomegalovirus early enhancer element and chicken beta-actin promoter) in an ss or sc genome. The size of sc genome packaged into capsid corresponds to the size of 3 ITR (375 bp) plus the double size of the expression cassette (the CAG promoter and the GFP gene, 5742 bp); the total is 6117 bp. Recombinant vectors (rAAV) were purified by double-CsCl ultracentrifugation followed by dialysis against phosphate-buffered saline (PBS). To measure the transducing activity of rAAV harbouring the GFP gene, HEK293 cells were seeded at 6 × 106 cells per plate 100 mm diameter, were infected with pure or diluted rAAV in Dulbecco’s modified Eagle’s medium and 5% fetal calf serum in the presence of ΔE1-Ad5 (multiplicity of infection of 8). Twenty-four hours later, GFP-positive cells were counted by using a fluorescent microscope (Nikon, Champigny sur-Marne, France). The in vitro transduction efficiency of the ‘oversized-scAAVrh10-CAG-egfp (o-sc-AAVrh10) was three times higher than the ssAAVrh10-CAG-egfp (ratio of vg per transducing units 70 909 for the o-sc and 228 000 for the ss). The integrity of viral DNA was checked on agarose gel in alkaline conditions and as expected in previous reports using the oversized genomes,27 fragments of various sizes were packaged in our o-sc batch.
Physical particles were quantified by dot blot hybridization. Vector titres are expressed as viral genomes per millilitre (vg ml−1). Titres were 5.7 × 1013 vg ml−1 for ssAAVrh10-egfp and 7.8 × 1012 vg ml−1 for scAAVrh10-egfp. To allow proper comparison between ss and o-sc vectors encoding GFP, the vectors were titre-matched (the more concentrated ss vector was extemporarily diluted into PBS), and the same volume of both vectors was accordingly injected to the animals.
Animal care and use
The experimental design is summarized in Table 2. Pregnant and adult (8-week old, female) Sprague Dawley rats were purchased from Janvier SAS Laboratories (Laval, France). Experiments were reviewed and approved by the regional ethics committee (CEEA Pays de la Loire, authorization number CEEA-2010-26). All animal experiments were carried out according to European guidelines for the care and use of experimental animals.
The experimental design is summarized in Table 2. Eight-week-old adult rats (200 g) were anesthetized with 1.5% isoflurane and received 4 × 1012 vg (2 × 1013 vg kg−1) of either ss- or o-scAAVrh10-egfp in 500 μl or PBS (four per group) in the caudal lateral vein using a 26-G vascular catheter. Neonate rats (6–8 g) were anesthetized by hypothermia on postnatal day 2 and received 1011 vg (2 × 1013 vg kg−1) of either ss- or o-scAAVrh10-egfp in 100 μl or PBS (four per group) in the superficial temporal vein using a Hamilton syringe connected to a 33-G bevelled needle. The syringe was connected to a micropump (Micro 4 controller and UltraMicroPump III; World Precision instruments Inc, Sarasota, FL, USA) that allowed a controlled rate of delivery (25 μl min−1).
We performed a pilot study in rats in order to adjust the dose that allowed a broad expression of GFP. We compared a low dose (2 × 1010 vg in adults and 4 × 109 vg in neonates) with a high dose (4 × 1011 vg in adults and 8 × 1010 vg in neonates). The low dose in adults did not allow GFP expression in MNs and was suboptimal in neonates, whereas brain and MN transduction was achieved with the high dose. We therefore carried the study using the highest doses.
Eight-week-old adult rats (200 g) were anesthetized with 1.5% isoflurane and received 4 × 1011 vg (2 × 1012 vg kg−1) of either ss- or o-scAAVrh10-egfp (n=4 per group) or PBS in a total volume of 50 μl. Rats were laid with their head flexed in order to open the atlanto-occipital joint. The skin over the cisterna magna was cut with a scalpel, and the cervical muscles were gently dissected. A 27-G needle was inserted until spontaneous flow of CSF occurred. Forty-to-fifty microlitres of CSF were removed prior to the injection. The vector was delivered manually using a Hamilton syringe connected to a 26-G bevelled needle. Neonate rats (6–8 g) were anesthetized by hypothermia on postnatal day 2 and then percutaneously injected with 1011 vg (2 × 1013 vg kg−1) of either ss- or scAAVrh10-gfp (n=4 per group) or PBS in the cisterna magna with a total volume of 15 μl at a controlled rate of 5 μl min−1.
In vivo GFP fluorescence imaging
GFP expression in rats injected IV was monitored at weekly intervals from 2 weeks p.i. in adults and 3 weeks p.i. in neonates (due to size limitations in rat pups) by fluorescence retinal imaging using a Canon UVI retinal camera connected to a digital imaging system (Lhedioph Win Software, Lheritier SA, Saint-Ouen-l’Aumône, France).
Euthanasia, perfusion and tissue processing
Animals were euthanized 4–5 weeks after injection. They were anesthetized with ketamine (100 mg kg−1) and xylazine (10 mg kg−1) and perfused transcardially with PBS followed by 4% paraformaldehyde. Brains, trigeminal nerves and ganglia, spinal cords, DRG, hearts, livers, adrenals, diaphragms and muscles (biceps femoralis, soleus, supraspinatus, latissimus dorsi) were dissected, postfixed in 4% paraformaldehyde and cryoprotected by overnight incubation in 6% and 30% sucrose. Brains and spinal cords were cut in coronal blocks. Samples were embedded in optimal cutting temperature compound and frozen on dry ice.
GFP expression, immunostaining and fluorescence microscopy analysis
GFP expression profile was explored by direct observation using both epifluorescent microscopy (Nikon ECLIPSE 80i) and laser confocal microscopy (Nikon, C1) equipped with a blue argon laser at 488 nm. The gain setting was optimized using the positive and negative control for each analysed tissue. The specificity of GFP fluorescence was checked on spinal cord tissue of injected animals by western blotting (Supplementary Figure S5) and by GFP immunofluorescence with a rabbit polyclonal anti-GFP antibody (1:100, AB3080; Millipore, Billerica, MA, USA); Alexa 555-conjugated goat anti-rabbit (1:200; Life Technologies, Saint-Aubin, France) was the secondary antibody. The cellular tropism of AAVrh10 in the CNS was explored by immunofluorescence. The primary antibodies were rabbit polyclonal anti-glial fibrillary acidic protein (1:4000, Z0334; Dako, Les Ulis, France), rabbit polyclonal anti-Olig2 (1/100, AB9610; Millipore), mouse monoclonal anti-NeuN (1:800, MAB377; Millipore), and rabbit polyclonal anti-Iba1 (1:500, 019-19741; Wako, Richmond, VA, USA). Sections were washed in PBS and then incubated for 1 h at room temperature with Alexa 555-conjugated goat anti-rabbit or goat anti-mouse immunoglobulin G (1:200; Life Technologies). Sections were washed in PBS, mounted in mowiol (Calbiochem, San Diego, CA, USA) and scanned serially using the argon ion laser (488 nm) to observe GFP signal (green) and with a helium neon laser (543 nm) to observe the Alexa fluor 555 signal (red). For ChAT immunofluorescence, we used a goat polyclonal anti-ChAT antibody (1:100, AB144P; Millipore); the sections were then incubated with a biotinylated rabbit anti-goat antibody (Dako) followed by incubation with Alexa 555-conjugated streptavidin (Life Technologies). Nuclei were counterstained with TO-PRO-3 iodide (Life Technologies). Control immunofluorescence analyses were performed on slides from PBS-injected animals and also on slides with the secondary antibody alone.
Quantification of GFP expression in spinal cord MNs
The GFP-positive MNs were scored in both the adult and neonatal rats injected IT (n=4 per group) on 12 10-μm thick sections separated by 250 μm (4 sections spanning the cervical enlargement, 4 sections spanning the mid-thoracic spinal cord and 4 sections spanning the lumbar enlargement). All MNs positive for fluorescent Nissl stain (Neurotrace Fluorescent Nissl Stain, N21483; Life Technologies) and located within the lamina of Rexed 9 were counted to estimate the percentage of GFP-expressing MNs relative to the total number of Nissl-stained MNs (using the NIS-Elements AR Image analyser, Nikon Imaging Software). Slides from a PBS-injected animal served as a negative control.
Quantification of vector genomes by qPCR
For qPCR analysis, a phenol–chloroform extraction of DNA was carried out on tissue frozen in liquid nitrogen immediately after sampling. The amount of vector genome per diploid genome for each tissue was determined by qPCR with the Fast SYBR green master mix (Applied Biosystems, Waltham, MA, USA) using the standard conditions. Primers were targeted against the GFP transgene (forward 5′-IndexTermGCGAAAGTGGAAAAGCCAAGT-3′, reverse 5′-IndexTermGCCACATCAACAGGACTCTTGTAG-3′, amplicon size 76 bp) and the rat hypoxanthine phosphoribosyltransferase (HPRT) of the host genome (forward 5′-IndexTermACTACAACAGCCACAACGTCTATATCA-3′, reverse 5′-IndexTermGGCGGATCTTGAAGTTCACC-3′, amplicon size 75 bp). Vector genome copies for each sample were quantified according to specific plasmid standard curves and expressed for 2N genomes according to the HPRT copy number measured in the same sample. Amplifications were performed using the StepOnePlus Real Time PCR System (Applied Biosystems). Dissociation curves confirmed amplification of only a single PCR product for all qPCR reactions. DNA extracted from tissue of a PBS-injected rat was amplified as a negative control in each run. Results obtained in PBS-injected rats defined the detection limit of the assay: 0.01 vg dg−1.
Investigation of humoral immune response against the transgene and capsid
An indirect enzyme-linked immunosorbent assay was used to detect GFP-specific antibodies. Plates were coated with GFP (rGFP 11814524001; Roche Diagnostics, Meylan, France) overnight, rinsed and blocked with 5% milk in PBS. Samples in serial dilutions, controls and blanks were added and incubated for 2 h at 37 °C. Horseradish peroxidase-conjugated donkey anti-rat immunoglobulin G (1:5000, r712-035-150; Jackson Immuno Research, Interchim, Montluçon, France) followed by Streptavidin/horseradish peroxidase (1:1000, P 0397; DakoCytomation, France SAS, Trappes, France) was added, and 3,3’,5,5’- Tetramethylbenzidine (BD Biosciences, France SAS, Le Pont-De-Claix, France) was used as substrate. Reactions were stopped with 2 N H2SO4, and reading was determined at 450 nm. Positivity was determined using a cutoff value of 0.200 optical density.
The anti-capsid seroneutralization assay was performed by co-infection of HeLaT cells with rAAVrh10-cmv-lacZ and helper adenovirus 5 with serial dilutions of animal sera. Transduction of HeLaT cells and LacZ expression were prevented in the presence of specific seroneutralizing anti-AAVrh10 antibodies. The proportion of transduced blue cells was finally evaluated after X-gal staining.
To assess differences between the groups, quantitative data were compared using a two-sided, non-paired Wilcoxon–Mann–Whitney test (R Development Core Team (2011). R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, http://www.R-project.org/). A P-value of ⩽0.05 was considered significant.
Haskins ME . Gene therapy for lysosomal storage diseases (LSDs) in large animal models. ILAR J 2009; 50: 112–121.
McCown TJ, Xiao X, Li J, Breese GR, Samulski RJ . Differential and persistent expression patterns of CNS gene transfer by an adeno-associated virus (AAV) vector. Brain Res 1996; 713: 99–107.
Vite CH, Passini Ma, Haskins ME, Wolfe JH . Adeno-associated virus vector-mediated transduction in the cat brain. Gene Therapy 2003; 10: 1874–1881.
Ciron C, Desmaris N, Colle M-A, Raoul S, Joussemet B, Vérot L et al. Gene therapy of the brain in the dog model of Hurler’s syndrome. Ann Neurol 2006; 60: 204–213.
Ciron C, Cressant A, Raoul S, Cherel Y, Hantraye P, De N . Human alpha-iduronidase gene transfer mediated of nonhuman primates: vector diffusion and biodistribution. Hum Gene Ther 2009; 20: 350–360.
Colle M-A, Piguet F, Bertrand L, Raoul S, Bieche I, Dubreil L et al. Efficient intracerebral delivery of AAV5 vector encoding human ARSA in non-human primate. Hum Mol Genet 2010; 19: 147–158.
Ellinwood NM, Ausseil J, Desmaris N, Bigou S, Liu S, Jens JK et al. Safe, efficient, and reproducible gene therapy of the brain in the dog models of Sanfilippo and Hurler syndromes. Mol Ther 2011; 19: 251–259.
Cearley CN, Wolfe JH . Transduction characteristics of adeno-associated virus vectors expressing cap serotypes 7, 8, 9, and Rh10 in the mouse brain. Mol Ther 2006; 13: 528–537.
Klein RL, Dayton RD, Tatom JB, Henderson KM, Henning PP . AAV8, 9, Rh10, Rh43 vector gene transfer in the rat brain: effects of serotype, promoter and purification method. Mol Ther 2008; 16: 89–96.
Piguet F, Sondhi D, Piraud M, Fouquet F, Hackett NR, Ahouansou O et al. Correction of brain oligodendrocytes by AAVrh.10 intracerebral gene therapy in metachromatic leukodystrophy mice. Hum Gene Ther 2012; 23: 903–914.
Swain GP, Prociuk M, Bagel JH, Donnell PO, Berger K, Drobatz K et al. Adeno-associated virus serotypes 9 and rh10 mediate strong neuronal transduction of the dog brain. Gene Therapy 2014; 21: 28–36.
Foust KD, Nurre E, Montgomery CL, Hernandez A, Chan CM, Kaspar BK . Intravascular AAV9 preferentially targets neonatal neurons and adult astrocytes. Nat Biotechnol 2009; 27: 59–65.
Duque S, Joussemet B, Riviere C, Marais T, Dubreil L, Douar A-M et al. Intravenous administration of self-complementary AAV9 enables transgene delivery to adult motor neurons. Mol Ther 2009; 17: 1187–1196.
Bevan AK, Duque S, Foust KD, Morales PR, Braun L, Schmelzer L et al. Systemic gene delivery in large species for targeting spinal cord, brain, and peripheral tissues for pediatric disorders. Mol Ther 2011; 19: 1971–1980.
Wang DB, Dayton RD, Henning PP, Cain CD, Zhao LR, Schrott LM et al. Expansive gene transfer in the rat CNS rapidly produces amyotrophic lateral sclerosis relevant sequelae when TDP-43 is overexpressed. Mol Ther 2010; 18: 2064–2074.
Snyder BR, Gray SJ, Quach ET, Huang JW, Leung CH, Samulski RJ et al. Comparison of adeno-associated viral vector serotypes for spinal cord and motor neuron gene delivery. Hum Gene Ther 2011; 1135: 1129–1135.
Federici T, Taub JS, Baum GR, Gray SJ, Grieger JC, Matthews Ka et al. Robust spinal motor neuron transduction following intrathecal delivery of AAV9 in pigs. Gene Therapy 2012; 19: 852–859.
Samaranch L, Salegio EA, Sebastian WS, Kells AP, Foust KD, Bringas JR et al. AAV9 transduction inthe CNS of non-human primate. Hum Gene Ther 2012; 23: 382–389.
Samaranch L, Salegio Ea, San Sebastian W, Kells AP, Bringas JR, Forsayeth J et al. Strong cortical and spinal cord transduction after AAV7 and AAV9 delivery into the cerebrospinal fluid of nonhuman primates. Hum Gene Ther 2013; 24: 526–532.
Haurigot V, Marcó S, Ribera A, Garcia M, Ruzo A, Villacampa P et al. Whole body correction of mucopolysaccharidosis IIIA by intracerebrospinal fluid gene therapy. J Clin Invest 2013; 123: 3254–3271.
Gray SJ, Nagabhushan Kalburgi S, McCown TJ, Jude Samulski R . Global CNS gene delivery and evasion of anti-AAV-neutralizing antibodies by intrathecal AAV administration in non-human primates. Gene Therapy 2013; 20: 450–459.
Bucher T, Colle M-A, Wakeling E, Dubreil L, Fyfe J, Briot-Nivard D et al. scAAV9 intracisternal delivery results in efficient gene transfer to the central nervous system of a feline model of motor neuron disease. Hum Gene Ther 2013; 24: 670–682.
Miyake N, Miyake K, Yamamoto M, Hirai Y, Shimada T . Global gene transfer into the CNS across the BBB after neonatal systemic delivery of single-stranded AAV vectors. Brain Res 2011; 1389: 19–26.
Hu C, Busuttil RW, Lipshutz GS . RH10 provides superior transgene expression in mice when compared with natural AAV serotypes for neonatal gene therapy. J Gene Med 2010; 12: 766–778.
Zhang H, Yang B, Mu X, Ahmed SS, Su Q, He R et al. Several rAAV vectors efficiently cross the blood-brain barrier and transduce neurons and astrocytes in the neonatal mouse central nervous system. Mol Ther 2011; 19: 1440–1448.
McCarty DM . Self-complementary AAV vectors; advances and applications. Mol Ther 2008; 16: 1648–1656.
Dong B, Nakai H, Xiao W . Characterization of genome integrity for oversized recombinant AAV vector. Mol Ther 2010; 18: 87–92.
Wang Y, Ling C, Song L, Wang L, Aslanidi GV, Tan M et al. Limitations of encapsidation of recombinant self-complementary adeno-associated viral genomes in different serotype capsids and their quantitation. Hum Gene Ther Methods 2012; 23: 225–233.
Hirsch ML, Li C, Bellon I, Yin C, Chavala S, Pryadkina M et al. Oversized AAV transductifon is mediated via a DNA-PKcs-independent, Rad51C-dependent repair pathway. Mol Ther 2013; 21: 2205–2216.
Xu Y, Gu Y, Wu P, Li G-W, Huang L-YM . Efficiencies of transgene expression in nociceptive neurons through different routes of delivery of adeno-associated viral vectors. Hum Gene Ther 2003; 14: 897–906.
Storek B, Harder NM, Banck MS, Wang C, McCarty DM, Janssen WG et al. Intrathecal long-term gene expression by self-complementary adeno-associated virus type 1 suitable for chronic pain studies in rats. Mol Pain 2006; 2: 4.
Towne C, Pertin M, Beggah AT, Aebischer P, Decosterd I . Recombinant adeno-associated virus serotype 6 (rAAV2/6)-mediated gene transfer to nociceptive neurons through different routes of delivery. Mol Pain 2009; 5: 52.
Vulchanova L, Schuster DJ, Belur LR, Riedl MS, Podetz-Pedersen KM, Kitto KF et al. Differential adeno-associated virus mediated gene transfer to sensory neurons following intrathecal delivery by direct lumbar puncture. Mol Pain 2010; 6: 31.
Fu H, Muenzer J, Samulski RJ, Breese G, Sifford J, Zeng X et al. Self-complementary adeno-associated virus serotype 2 vector: global distribution and broad dispersion of AAV-mediated transgene expression in mouse brain. Mol Ther 2003; 8: 911–917.
Dirren E, Towne CL, Setola V, Redmond DE, Schneider BL, Aebischer P . Intracerebroventricular injection of adeno-associated virus 6 and 9 vectors for cell type-specific transgene expression in the spinal cord. Hum Gene Ther 2014; 25: 109–120.
Gao G, Vandenberghe LH, Alvira MR, Lu Y, Calcedo R, Zhou X et al. Clades of adeno-associated viruses are widely disseminated in human tissues clades of adeno-associated viruses are widely disseminated in human tissues. J Virol 2004; 78: 6381–6388.
Boutin S, Monteilhet V, Veron P, Leborgne C, Benveniste O, Montus MF et al. Prevalence of serum IgG and neutralizing factors against adeno-associated virus (AAV) Types 1,2,5,6,8, and 9 in the healthy population: implications for gene therapy using AAV vectors. Hum Gene Ther 2010; 21: 704–712.
Wang H, Yang B, Qiu L, Yang C, Kramer J, Su Q et al. Widespread spinal cord transduction by intrathecal injection of rAAV delivers efficacious RNAi therapy for amyotrophic lateral sclerosis. Hum Mol Genet 2014; 23: 668–681.
Gray SJ, Foti SB, Schwartz JW, Bachabiona L, Taylor-Blake B, Coleman J et al. Optimizing promoters for rAAV-mediated gene expression in the peripheral and central nervous system using self-complementary vectors. Hum Gene Ther 2011; 22: 1143–1153.
Iliff JJ, Lee H, Yu M, Feng T, Logan J, Nedergaard M et al. Brain-wide pathway for waste clearance captured by contrast-enhanced MRI. J Clin Invest 2013; 123: 1299–1309.
Zhou X, Zeng X, Fan Z, Li C, McCown T, Samulski RJ et al. Adeno-associated virus of a single-polarity DNA genome is capable of transduction in vivo. Mol Ther 2008; 16: 494–499.
Milward Ea, Fitzsimmons C, Szklarczyk A, Conant K . The matrix metalloproteinases and CNS plasticity: an overview. J Neuroimmunol 2007; 187: 9–19.
Rabinowitz JE, Rolling F, Li C, Xiao W, Xiao X, Samulski RJ . Cross-packaging of a single adeno-associated virus (AAV) type 2 vector genome into multiple AAV serotypes enables transduction with broad specificity. J Virol 2002; 76: 791–801.
We thank the vector core of the Atlantic Gene Therapies Institute (AGT) in Nantes for the preparation of the rAAV vectors, Oumeya Adjali and Johanne Le Duff for the seroneutralization assay and the Boisbonne Centre for assistance with animal production and care. This work was supported by grants from the Association Française contre les Myopathies (AFM), the National French Academy of Medicine and an additional grant from ‘Investissement d'Avenir—ANR-11-INBS-0011’—NeurATRIS: A Translational Research Infrastructure for Biotherapies in Neurosciences’.
The authors declare no conflict of interest.
Supplementary Information accompanies this paper on Gene Therapy website
About this article
Cite this article
Hordeaux, J., Dubreil, L., Deniaud, J. et al. Efficient central nervous system AAVrh10-mediated intrathecal gene transfer in adult and neonate rats. Gene Ther 22, 316–324 (2015). https://doi.org/10.1038/gt.2014.121
Brain Research (2020)
Journal of Huntington's Disease (2019)
Intrathecal AAVrh10 corrects biochemical and histological hallmarks of mucopolysaccharidosis VII mice and improves behavior and survival
Human Molecular Genetics (2019)
AAVrh-10 transduces outer retinal cells in rodents and rabbits following intravitreal administration
Gene Therapy (2019)
Molecular Therapy (2019)