Assessment of Safety and Biodistribution of AAVrh.10hCLN2 Following Intracisternal Administration in Nonhuman Primates for the Treatment of CLN2 Batten Disease

Abstract

CLN2 disease is a fatal, childhood autosomal recessive disorder caused by mutations in ceroid lipofuscinosis type 2 (CLN2) gene, encoding tripeptidyl peptidase 1 (TPP-1). Loss of TPP-1 activity leads to accumulation of storage material in lysosomes and resultant neuronal cell death with neurodegeneration. Genotype/phenotype comparisons suggest that the phenotype should be ameliorated with increase of TPP-1 levels to 5–10% of normal with wide central nervous system (CNS) distribution. Our previous clinical study showed that intraparenchymal (IPC) administration of AAVrh.10hCLN2, an adeno-associated vector serotype rh.10 encoding human CLN2, slowed, but did not stop disease progression, suggesting that this may be insufficient to distribute the therapy throughout the CNS (Sondhi 2020). In this study, we assessed whether the less invasive intracisternal delivery route would be safe and provide a wider distribution of TPP-1. A study was conducted in nonhuman primates (NHPs) with intracisternal delivery to cerebrospinal fluid (CSF) of AAVrh.10hCLN2 (5 × 10¹³ genome copies) or phosphate buffered saline (PBS). No abnormal behavior was noted. CNS magnetic resonance imaging and clinical chemistry data were all unremarkable. Histopathology of major organs had no abnormal finding attributable to the intervention or the vector, except that in one out of two animals treated with AAVrh.10hCLN2, dorsal root ganglia showed mild-to-moderate mononuclear cell infiltrates and neuronal degeneration. In contrast to our previous NHP study (Sondhi 2012) with IPC administration where TPP-1 activity was >2 × above controls in 30% of treated brains, in the two intracisternal treated NHPs, the TPP-1 activity was >2 × above controls in 50% and 41% of treated brains, and 52% and 84% of brain had >1,000 vector genomes/μg DNA, compared to 0% in the two PBS NHP. CSF TPP1 levels in treated animals were 43–62% of normal human levels. Collectively, these data indicate that AAVrh.10hCLN2 delivered by intracisternal route is safe and widely distributes TPP-1 in brain and CSF at levels that are potentially therapeutic.

Clinical Trial Registration:

NCT02893826, NCT04669535, NCT04273269, NCT03580083, NCT04408625, NCT04127578, and NCT04792944.

INTRODUCTION

Neuronal ceroid lipofuscinosis type 2 (CLN2) disease is an autosomal recessive disorder due to mutations in the CLN2 gene on chromosome 11p15, which encodes the lysosomal enzyme tripeptidyl peptidase 1 (TPP-1).^1–7 At least 40 mutations in the CLN2 gene have been identified with two mutations, the splicing mutation IVS5-1G and the nonsense mutation R208X, being the most common, accounting for ∼60–78% of all CLN2 mutations.^1,8,9 The loss of TPP-1 activity leads to accumulation of storage material in lysosomes and resultant neuronal cell death with progressive neurodegeneration.^2,8 Genotype/phenotype comparisons suggest that the severe phenotype should be ameliorated with increase of central nervous system (CNS) TPP-1 levels to 5–10% of normal.^10,11

This might be achieved using an adeno-associated vector (AAV) efficient in transferring genes to the CNS, mediating persistent expression of TPP-1, a secreted protein capable of cross-correcting neighboring cells.^12–14 However, since the manifestations of CLN2 disease are throughout the brain, it is essential that the therapy achieves wide distribution in the CNS.^2,15

Based on efficacy studies in CLN2−/− knockout mice,^16,17 and CNS biodistribution and safety studies in nonhuman primates (NHPs),¹⁸ a phase 1 clinical trial with the AAV serotype rh.10 expressing the normal human CLN2 coding sequence (AAVrh.10hCLN2) was carried out to treat children with CLN2 disease.¹⁵ CNS intraparenchymal (IPC) delivery met the primary endpoint of significantly slowing the clinical progression of neurologic decline, with 42.4–47.5% reduction in the quantitative assessment of the rate of decline of motor + language function compared to two independent natural history control cohorts.^4,5,15,19

However, while IPC administration of AAVrh.10hCLN2 significantly slowed the progression of the disease, it did not completely halt progression, suggesting that the IPC route may be insufficient to distribute the vector/vector-mediated TPP-1 product throughout the CNS. In this context, the IPC route is limited by (1) diffusion of the vector in the parenchyma and (2) the high concentrations of vector at the catheter tip, inducing local inflammation.¹⁵ In contrast, since the cerebrospinal fluid (CSF) bathes the entire CNS, delivery through the intracisternal magna (ICM) route should provide broad access to the brain, allowing higher doses with less safety/toxicity concerns than IPC delivery.^20–34 TPP-1 is a secreted protein,^12–14 and the ICM route provides direct access to the CSF and the cells lining the CSF reservoirs,^20,31 providing sites where AAVrh.10hCLN2 can express TPP-1, allowing it to diffuse throughout the CNS.

To address if the ICM route of delivery yields widespread distribution of TPP-1 at a dose that was safe, we carried out a study in NHPs (African Greens) with ICM administration to the CSF with a single dose of AAVrh.10hCLN2. The study demonstrates that intracisternal administration is safe and leads to broad distribution of vector genome throughout the brain, with elevated levels of TPP-1 protein and enzyme activity in CSF and brain tissue at levels that are potentially therapeutic.

METHODS

Study design

The study was designed to evaluate CLN2 expression following ICM delivery of AAVrh.10hCLN2 (Supplementary Fig. S1) in the presence of immunosuppression to mimic potential human clinical trials. Four naive, African Green NHPs, three males and one female, were randomly assigned to one of two treatment cohorts (AAVrh.10hCLN2, 5.0 × 10¹³ genome copies [gc], or control group administered phosphate buffered saline [PBS]), with periodic assessment and necropsy at 8 weeks after vector or vehicle administration (Table 1).

Table 1.

Assessment of safety and biodistribution of intracisternal delivery of AAVrh.10hCLN2 and biodistribution of AAVrh.10hCLN2 to the nonhuman primate central nervous system ^a

Parameter	Time of Assessment (Study Week)a
	Pre	0	4	7	8
Therapy
Immunosuppressionb	■	■	■	■	■
Vector/vehicle administrationc		■
Safety parameters
Physical and clinical observationsd	■	■	■	■	■
Complete blood countse	■	■	■		■
Serum chemistryf	■	■	■		■
CSFg		■	■		■
Anti-AAVrh.10 immunity—serum and CSFh	■	■	■		■
Behaviori	■	■	■		■
MRIj	■			■
Sacrificek					■
Gross and histopathologyl					■
Biodistribution of vector and protein
TPP-1 levels and functional activity in the CSFm		■	■		■
Vector biodistribution in the CNS and systemic organsn					■
Brain biodistribution of TPP-1 activityo					■

^a n = 4 (1F/3M). Study day 0 was day of vector administration; all observations and safety assessments on day 0 were performed before surgery. Times for assessment for remaining time points were ±2 days, except for “Pre,” which includes blood collections 6–10 weeks before study start, and 1 week before for immunosuppression.

^b Immunosuppression (prednisolone, 1 mg/kg/daily, administered orally) was done for the length of the trial, starting at 1 week before AAV administration.

^c Vector dosing: The vector, AAVrh.10hCLN2 (5 × 10¹³ genome copies, cohort A) or PBS control (cohort B) was administered directly into the cisterna magna as a single dose (1–2 mL and infused at 0.5 mL/min) followed by saline (1 mL) flush. Administration was blinded at dosing.

^d Assessment of general safety parameters, including temperature, pulse, respiratory rate, and weight.

^e Complete blood tests included WBC, RBC, reticulocytes (% and absolute), hemoglobin, hematocrit, MCV, MCH, MCHC, segmented neutrophils (% and absolute), lymphocytes (% and absolute), monocytes (% and absolute), eosinophils (% and absolute), basophils (% and absolute), and platelet count.

^f Serum chemistry tests included ALP, ALT, AST, GGT, albumin, total protein, globulin, total bilirubin, BUN, CK, triglycerides, glucose, LDH, calcium, phosphorus, bicarbonate, amylase, lipase, sodium, chloride, potassium, Na/K, A/G, and B/C ratios, anion gap.

^g CSF was collected at 3 time points per NHP, before vector administration (day 0), and at week 4 and 8 (before necropsy); for analysis, anti-AAVrh.10 total and neutralizing antibodies (footnote g) and quantification of TPP-1 protein/functional activity levels (footnote m)

^g Assessments included total anti-AAVrh.10 and neutralizing AAVrh.10 antibody titers in serum; and total anti-AAVrh.10 titers and neutralizing anti-AAVrh.10 antibody titers in the CSF.

ⁱ At the indicated times, behavioral assessments were performed with videotaping at rest and in responses to a series of standardized challenges, with extraction of quantitative traits as previously.^{18,27,35–37}

^j MRI. MRI of the CNS was performed prevector administration (−1 week) and presacrifice (7 weeks) to assess safety in the CNS.

^k All animals were euthanized 8 weeks postvector administration following perfusion with cold PBS. All major organs were collected and weighed. A detailed histopathology assessment of all organs and tissues was carried out by board-certified veterinary pathologists.

^l Before necropsy, blood and CSF were sampled for analyses, as detailed in footnotes e–g. During the necropsy, the animals were observed for pathology, with organs weighed and examined for gross pathology and abnormalities. The brain was excised, divided into hemispheres, and one hemisphere fixed with 10% formalin for histopathology examination. The other hemisphere was prepared and used for biodistribution, as described in footnotes n, o. Samples of other organs (Supplementary Table S2) were collected at necropsy, fixed with formalin for histopathological analysis, or stored frozen for biodistribution. Additional pathology assessments were performed on three levels of spinal cord (cervical, thoracic, and lumbar) with intact DRG by IHC for inflammation markers (CD3, CD20, Iba1, GFAP, and capase-3) and H&E stains.

^m The CSF collected (footnote g) during the study was also assessed for TPP-1 levels and functional enzyme activity.

ⁿ The left hemisphere of the brain was used for biodistribution and divided into cubes (1 cm³) for determination of vector DNA. Peripheral organs were also assessed for vector DNA.

^o The brain cubes from the left hemisphere (1 cm³) were also used for assessment of TPP-1 biodistribution, using a functional activity assay.

A/G, albumin/globulin; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; B/C, BUN/creatinine; BUN, blood urea nitrogen; CK, creatine kinase; CNS, central nervous system; CSF, cerebrospinal fluid; DRG, dorsal root ganglion; GGT, gamma glutamyl transferase; H&E, hematoxylin and eosin; IHC, immunohistochemistry; LDH, lactate dehydrogenase; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular hemoglobin concentration; MCV, mean corpuscular volume; MRI, magnetic resonance imaging; Na/K, sodium/potassium; PBS, phosphate buffered saline; RBC, red blood cell count; TPP-1, tripeptidyl peptidase 1; WBC, white blood cell count.

One week before vector administration, the NHPs were evaluated by magnetic resonance imaging (MRI) and started on an immunosuppression regimen with prednisolone daily. All NHPs were assessed for TPP-1 levels and activity in the CSF, immunity (anti-AAVrh.10 antibody titers) in serum and CSF, behavior, hematological and serum chemistry parameters at 4 time points (prestudy, days 0, 28, and 56), and brain histopathology at a single end point (8 weeks) postadministration (Table 1). The NHPs were reassessed by CNS MRI scans at 7 weeks after vector administration. Detailed methods regarding the design and production of the vector AAVrh.10hCLN2, vector administration in NHP, in life health assessments, necropsy, all assay methods, and statistical analysis are provided in Supplemental Methods.

RESULTS

General safety and clinical observations

No acute or adverse side effect was noted by the blinded observers during daily/weekly evaluations. No unscheduled mortality or clinical abnormality was observed during the 8-week study. NHPs were assessed at 5 time points: prestudy, day 0 (before dosing), 4 weeks (before blood collection), and week 7 and 8 (before necropsy). Overall, the body weights remain similar throughout the duration of the study (Supplementary Fig. S2A). There was no significant difference in vitals (heart rate, respiratory rate, and body temperature) in the vector-treated group compared to the PBS controls (Supplementary Fig. S2B–D).

Hematology and serum chemistry

Animals were evaluated for an extensive set of serum chemistry and hematologic parameters on day 0 (day of administration), and 4 and 8 weeks postadministration. No significant difference was observed between the treated group and control group in the hematology results (Supplementary Fig. S3). Sporadic changes of hematologic parameters were observed in single animals at several time points of the study and were occasionally outside reference ranges (gray-shaded areas)^27,35,36; these changes did not constitute trends, did not appear to be associated with treatment groups, and did not correlate with clinical abnormalities or gross or microscopic pathologic findings, and were considered to represent normal, naturally occurring variations. There was no significant finding or group difference in the serum chemistry results (Supplementary Fig. S4), nor were the data from the treated cohort outside the normal range of pretreatment levels (gray-shaded areas).^27,35,36

Assessment of NHP behavior

Prevector administration (1 week before and day 0) and at week 4 and 8, the NHP underwent behavioral assessment as previously described.^{18,27,35–37} Blinded videotape analysis by multiple observers of NHP behavior preadministration and postadministration showed no discernible neurological difference in either the vector-treated or control PBS-treated animals (Fig. 1). All NHP scores of “healthy” activities (Supplementary Table S1) were within the historic range of normal NHP behavior, as calculated from prestudy data (Fig. 1). No adverse effect (such as sedation or dyskinesia, Supplementary Table S1) was observed and no NHP demonstrated abnormal deficits (head tilts, comatose, or altered levels of consciousness) at any time postvector administration.

Figure 1.

Effect of intracisternal administration of AAVrh.10hCLN2 vector on NHP safety, as scored from videotaped behavioral assessments at multiple time points. Four treated wild-type NHPs (African Green vervets; three males and one female) were assessed for behavior parameters at 4 time points: before study (Pre), day of administration (day 0), week 4, and week 8 (before necropsy). At indicated times, NHPs were videotaped. Results of 20 behavioral parameters (see Supplementary Table S1 for parameters assessed) for each NHP in the study are shown as sums. Linked data plots over time for each NHP (means of three reviewers, yellow, PBS [M]; green, PBS [M]; red, AAVrh.10hCLN2 [F]; blue, AAVrh.10hCLN2 [M]). Note: no NHP exhibited “Unhealthy” behaviors. NHP, nonhuman primate; PBS, phosphate buffered saline.

Central nervous system magnetic resonance imaging

The data demonstrated that ICM administration of AAVrh.10hCLN2 had no adverse effect, as observed by CNS MRI. MRI scans (T1, T2, T2 FLAIR) performed preadministration (−1 week) and at 7 weeks postadministration of AAVrh.10hCLN2 or PBS were assessed in a blinded manner by a clinical radiologist. All scans were normal, barring one animal that received the PBS control, where there was a T2 hyperintensity observed in the pons, which was present preadministration and postadministration. This bilateral symmetric T2 observation was not present in the T2-FLAIR scans and hence was not a bleed or hematoma.

It is possible that this was due to a pre-existing normal variant or delayed myelination. To examine the region around the injection site, the cisterna magna, we compared multiple levels around the base of the CNS (cerebellum, pons, and C1 spinal cord) for T2 hypersensitivity and found no abnormality. Any needle tract was resolved by 7 weeks postadministration when the MRI scans were performed (Supplementary Fig. S5).

Anti-AAVrh.10 antibodies in serum and CSF

Anti-AAVrh.10 antibodies (total and neutralizing) were determined at predose, 4 weeks, and 8 weeks postadministration in both serum and CSF. Pretherapy, no NHP had detectable anti-AAVrh.10 total or neutralizing antibodies in either serum or CSF. Vector-treated animals were positive for anti-AAVrh.10 total (Fig. 2A, B) and neutralizing (Fig. 2C, D) antibodies in serum and CSF, with higher titers than those found in the animals assigned to the PBS group. In the CSF, while both vector-treated animals were positive for anti-AAVrh.10 total antibodies, only 1/2 NHPs had higher anti-AAVrh.10 neutralizing antibodies than predose levels. As expected, the postdose titers were higher in the serum than in the CSF.

Figure 2.

Assessment of total and neutralizing anti-AAVrh.10 antibody titers in serum and CSF evoked by administration of AAVrh.10hCLN2. Total anti-AAVrh.10 antibody and neutralizing anti-AAVrh.10 antibody titers were determined over time after intracisternal administration of AAVrh.10hCLN2 or PBS (preadministration, and then at 4 and 8 weeks postadministration). The total antibody titer is expressed as the reciprocal of the serum or CSF dilution. The neutralizing antibody titer is expressed as the reciprocal of serum or CSF dilution at which 50% inhibition of AAVrh.10Luc activity was observed. Shown are the results for each NHP in the study, color coded by treatment/gender (yellow, PBS [M]; green, PBS [M]; red, AAVrh.10hCLN2 [F]; and blue, AAVrh.10hCLN2 [M]). (A) Serum total anti-AAV antibody titers. (B) CSF total anti-AAV antibody titers. (C) Serum neutralizing anti-AAV antibody titers. (D) CSF neutralizing anti-AAV antibody titers. Females (n = 1); males (n = 3). The assay limit of detection was 100; sample results below were recorded as 1. AAV, adeno-associated vector; CSF, cerebrospinal fluid.

Macroscopic observations (gross necropsy)

There were no AAV vector-related macroscopic finding. All macroscopic observations were considered incidental, of the type, severity, and/or distribution commonly observed in this species, and/or were of similar incidence and severity across both groups. No significant difference in organ weights (absolute weights, organ/body, and organ/brain weight ratios) were observed between treated and control groups (not shown).

Microscopic observations (histopathology)

A comprehensive histopathological examination of ∼50 different tissue/organs for each NHP (Supplementary Table S2) was performed by two board-certified veterinary pathologists. For most tissues, no abnormal observation was attributed to the treatment (Supplementary Table S3).

For some organs, including lung, heart, kidney, thyroid and esophagus, lesions were noted in one out of two treated NHPs, which were not observed in either control NHPs, but it was the pathologists' opinion that these probably reflected naturally occurring changes. Examples included minimal to mild, multifocal lymphohistiocytic infiltrates in the esophagus, thyroid, kidneys, heart, and lungs. Cortical tubular dilation and medullary mineralization were observed in the kidneys of one treated NHP. The only change observed that was judged to be related to the treatment was a series of neuropathological findings in one (female) of the two AAV-treated NHPs in nerves (sciatic, sural, and glossopharyngeal), spinal cord (cervical, thoracic, and lumbar segments), brain stem (medulla), dorsal root ganglia (DRG), and nerve roots.

These abnormalities were scored based on a system published by Hordeaux et al.³⁸ The changes in the CNS and peripheral nerves consisted of axonopathy (i.e., axonal degeneration, reflected by the presence of axonal spheroids and debris, and swollen myelin sheaths containing macrophages³⁹) and were mild-to-moderate with severity scores of 2 or 3 out of 5. Hypercellularity, comprising mononuclear inflammatory infiltrates and activated glial cells (scores were 2 or 3 out of 5), and neuronal lesions, including small, irregular or angular shaped cells with fading or absent nuclei and cytoplasmic hypereosinophilia, and neuronophagia with or without complete neuronal body obliteration (scores were 1 or 2 out of 5), was also observed in the DRG and nerve roots (Supplementary Fig. S6).

Additional examinations of the lumbar DRG were performed for the AAV-treated female NHP with these lesions and compared to the PBS-treated male NHP, using immunohistochemistry staining for CD3 (T cell), CD20 (B cell), Iba1 (activated microglia/macrophages), GFAP (astrocytes), and cleaved caspase-3 (apoptosis) (Supplementary Fig. S7). In the AAV-treated female, the DRGs displayed lymphocytic infiltrates predominantly composed of CD3-positive T cells (Supplementary Fig. S7B) and, to a lesser extent, CD20-positive B cells (Supplementary Fig. S7D), increased numbers of microglia (Iba1, Supplementary Fig. S7F), as well as astrocytosis and astrogliosis (GFAP, Supplementary Fig. S6H) compared to the PBS controls that had little to no inflammation or activated glial cells (Supplementary Fig. S7A, C, E, G).

The affected NHP also showed multifocal neuronal immunolabeling for cleaved caspase-3 (Supplementary Fig. S7J), in contrast to lack of neuronal immunolabeling in the PBS control (Supplementary Fig. S7I). The findings in the DRG were similar to those reported previously following CNS delivery of AAV vectors.⁴⁰ Aside from axonopathy, the neuronal lesions and hypercellularity observed in the DRG were not observed in the surrounding spinal cord or brain stem in the treated animal.

TPP-1 expression in the CSF

Following administration of 5 × 10¹³ gc vector through the ICM route, CSF was collected from NHPs at 4 and 8 weeks for assessment of TPP-1 protein levels (enzyme-linked immunosorbent assay [ELISA]) and enzyme activity (fluorometric assay) compared to prestudy CSF levels (Fig. 3). CSF TPP-1 protein levels were increased at both the 4- and 8-week time point in both treated NHPs, while TPP-1 levels did not change in the CSF of PBS-treated NHPs over the course of the study (Fig. 3A). The CSF TPP-1 levels obtained in the treated NHPs were 43–62% of normal human CSF TPP-1 levels, based on analysis of the CSF of an 8-year-old male subject without CLN2 disease (Fig. 3A).

Figure 3.

Assessment of human TPP-1 levels and activity in the CSF following AAVrh.10hCLN2 intracisternal administration. Expression of the secreted TPP-1 from the hCLN2 transgene was assessed by ELISA for total protein (ng/mL) and fluorometric assay for functional TPP-1 activity (FU/min/mL). Shown are the results for each NHP in the study, color coded by treatment/gender (yellow, PBS [M]; green, PBS [M]; red, AAVrh.10hCLN2 [F]; and blue, AAVrh.10hCLN2 [M]). The endogenous primate TPP-1 is detected by both assays at low levels. (A) CSF TPP-1 total protein levels. In addition to the NHP samples, CSF from an 8-year-old human male subject normal for CLN2 was also assessed to establish human target levels (purple triangle). (B) CSF TPP-1 functional activity levels. (C) Correlation of CSF TPP-1 protein levels versus TPP-1 functional activity (all data shown as red circles independent of treatment). Females (n = 1); males (n = 3). Note, the PBS male NHP-4 had insufficient volume for both assays and was only tested for functional activity at the 0 time point in panel (B). CLN2, ceroid lipofuscinosis type 2; ELISA, enzyme-linked immunosorbent assay; FU, fluorescence unit; TPP-1, tripeptidyl peptidase 1.

In agreement with the data on protein expression, TPP-1 enzyme activity levels showed the same pattern of elevation in vector-treated NHPs, but not PBS-treated NHPs (Fig. 3B). Elevations in TPP-1 activity at the 8-week time point were ∼2.5-fold and 4.3-fold in the two NHPs for an average elevation of 3.4-fold (Fig. 3B). As expected, there was a direct correlation between TPP-1 levels in the CSF and enzymatic activity (Fig. 3C).

Tissue biodistribution

Assessment of vector DNA demonstrates distribution of AAVrh.10hCLN2 throughout the CNS compared to the PBS controls (Fig. 4A–D). AAV-treated NHP brain cubes displayed vector genome copy numbers (VCN) higher than matching PBS control brain cubes (p < 0.0001, Fig. 4E). In both treated NHPs, the highest VCN per cube was in the inferior portions of the frontal and temporal cortexes and hippocampal/midbrain areas, while deep brain regions thalamus/striatum and posterior occipital/parietal cortex cubes had lower VCN, but significantly higher than PBS controls (Fig. 4A, B, F), demonstrating widespread biodistribution of AAVrh.10hCLN2 across the CNS.

Figure 4.

Biodistribution of vector genome in NHP brain. Shown is a 3-dimension schematic of 1 cm³ cubes of the left hemispheres. Coronal slices (1 cm width) are displayed from the anterior (left) to the posterior (right) end of the NHP brains along the x-axis, with dorsal (top) and ventral (bottom) aspects of the brain along the y-axis. The coronal slices were subdivided into 1 cm³ cubes for analysis; yellow arrow indicates the ICM site of administration to the CSF. Vector DNA copies/μg total DNA were measured. The amount of vector DNA copies is indicated by the color of each cube (see the color scheme key). (A, B) Vector treated. (C, D) PBS treated. (E) AAV genome copy levels in all brain cubes combined for n = 2 AAVrh.10hCLN2-treated NHP and n = 2 PBS-treated NHP. Black dashed line: baseline mean PBS control (combined, n = 2) 1.3 copies/μg DNA. (F) Mean vector biodistribution in brain regions. The cortex and substructures in each cube were identified based on photographs of the coronal slices. VCN data were combined for each substructure and compared for the AAV-treated (red) and PBS-treated controls (yellow). ICM, intracisternal magna; VCN, vector genome copy numbers.

Vector genome levels were >1,000 copies/cube in 52% and 84% of the brain cubes assessed in the two AAV-treated NHPs compared to 0% in the PBS NHP (Fig. 4A–D). We also assessed the VCN in CNS substructures to demonstrate the widespread coverage through the IC route, based on the particular structural contents of each cube, such as thalamus and hippocampus (Fig. 4F). All regions of spinal cord tested also showed high VCN (Supplementary Fig. S8). AAV genomes were detected in systemic organs in both AAV-treated NHPs compared to the PBS-treated control NHP (Supplementary Fig. S8). As expected from previous studies, the liver had high VCN, as did the optic nerve, spleen, heart, and adrenal gland (Supplementary Fig. S8).

TPP-1 distribution in brain

TPP-1 levels in brain cubes were also widely distributed across the brain compared to the PBS controls. Mean TPP-1 activity in the AAV-treated NHPs was >2 × that of background average endogenous TPP-1 levels (164 fluorescence unit/mg), with 50% and 41% of the two treated NHP brains showing higher than background mean TPP-1 levels (Fig. 5A–F). In some brain regions (especially cerebellum), the TPP-1 activity was between 3 × and 6 × above the background PBS levels, indicating high, broad distribution of TPP-1 levels across the brain (Fig. 5F).

Figure 5.

Biodistribution of TPP-1 activity in the brain. Shown is a 3-dimensional schematic of 1 cm³ cubes of the left hemispheres. Coronal slices (1 cm width) are displayed from the anterior (left) to the posterior (right) end of the NHP brains, along the x-axis, with dorsal (top) and ventral (bottom) aspects of the brain along the y-axis. The coronal slices were subdivided into 1 cm³ cubes for analysis; yellow arrow indicates the ICM site of administration to the CSF. TPP-1 functional activity (FU)/mg protein was measured. TPP-1 levels are indicated by the color of each cube (see the color scheme key). (A, B) Vector treated. (C, D) PBS treated. (E) TPP-1 activity levels in all brain cubes combined for n = 2 AAVrh.10hCLN2-treated NHP and n = 2 PBS-treated NHP. Black dashed line: TPP-1 average was 164 FU/min-mg from PBS control (combined, n = 2). 2 × standard deviation ranges from PBS males = 92.1–164 FU/min/mg (gray box). (F) Mean TPP-1 activity biodistribution in brain regions. The cortex and substructures in each cube were identified based on photographs of the coronal slices. TPP-1 levels data were combined for each substructure and compared for the AAV-treated (red) and PBS-treated controls (yellow).

While the TPP-1 activity does not mirror the vector distribution in the CNS in the exact same cubes and substructures (Figs. 4 and 5), this is slightly misleading due to the nature of the TPP-1 protein activity in the CNS. TPP-1 is a secreted protein, circulating in the cellular milieu of the CNS; the localization of the TPP-1 protein will vary from the subcellular localization of the internalized episomal AAV genomes. We and other investigators have observed this in large animal models of CNS disorders.^{27,33,35,41–43}

In addition to the above, TPP-1 is not only secreted but also utilizes the reuptake pathway by CNS cell mannose-6-phosphate receptors to be internalized to lysosomes.¹³ This pathway might explain the differences in the DNA versus TPP-1 protein levels we saw in the vector and protein biodistribution (Figs. 4 and 5). In addition, the cerebellum is the closest CNS structure to the ICM injection site and hence may be higher for TPP-1 protein detection or localization.

DISCUSSION

CLN2 disease is a rare, fatal neurodegenerative disease of early childhood with seizures, ataxia, language delay, vision loss, and early death.^1–7 It is caused by mutations in the CLN2 gene, which encodes lysosomal TPP-1, an enzyme that cleaves tripeptides from the N-terminus of polypeptides imported into the lysosome. TPP-1 is a secreted protein capable of cross-correcting neighboring cells through uptake by the mannose-6-phosphate receptor.^12–14

CLN2 disease globally affects the entire CNS, and thus a challenge in developing a successful gene therapy treatment for CLN2 is the need to achieve widespread transgene expression.^15,44 Genotype/phenotype comparisons suggest that the CLN2 disease phenotype should be ameliorated with increase of TPP-1 levels to 5–10% of normal with wide CNS distribution.^10,11 If AAV-mediated CNS gene therapy could provide sufficient and persistent amounts of TPP-1 throughout the CNS, it could provide a one-time therapy to treat the disease.

AAVrh.10hCLN2 has been previously tested in the clinic using a direct IPC route of administration, demonstrating significant slowing of the progression of the disease compared to two natural history cohorts.¹⁵ In that study, AAVrh.10hCLN2, at a total dose of 2.85 × 10¹¹ to 9.0 × 10¹¹ gc, was administered directly into the brain parenchyma through six burr holes to 12 sites in the brain. This resulted in a 1.3- to 2.6-fold increase in TPP-1 levels in the CSF 12 months after vector administration with slowing of progression of motor and language decline.¹⁵ While these initial clinical results were encouraging and safe, the IPC route of administration likely did not result in optimal distribution of AAVrh.10hCLN2 throughout the brain and the dose delivered did not result in sufficient TPP-1 expression to completely halt progression of the disease.

In addition, CNS MRI assessed within 48 h of vector administration demonstrated T2 hyperintensities with diffusion restriction localized to the sites of vector administration. While these findings did not correlate with any clinical sequela, these localized abnormalities persisted 6–12 months after therapy in most subjects. Studies in NHPs have demonstrated that these MRI T2 hyperintensities are the result of high concentrations of the vector in the region surrounding the catheter tip, resulting in local inflammation.^15,36,45

Based on these observations, while efficacious and safe, catheter-based AAV delivery to the CNS is limited by dose, and it is unlikely that the route of a direct catheter into the CNS parenchyma will be able to provide sufficient distribution of an AAV vector to the CNS to completely halt progression of the disease. In contrast, we hypothesized that ICM administration into the CSF would be safe and provide diffuse distribution throughout the brain,^{21–27,29,33,34} allowing higher doses than the IPC route.

In this study, AAVrh.10hCLN2 was administered to NHP through the ICM route as a single bolus at a higher dose of 5 × 10¹³ gc to evaluate greater biodistribution spread across the CNS. Based on the previous CLN2 gene therapy study¹⁸ and the clinical trial results,¹⁵ we evaluated targeting the CSF in the NHPs to increase the biodistribution of AAVrh.10hCLN2 and TPP-1 in the CNS.

The reason for this was that even with the lower IPC dose used in our clinical study, we saw some MRI abnormalities at the site of administration, likely owing to the high local dose, precluding the use of higher doses using that route.¹⁵ Instead, we decided on the higher 5 × 10¹³ gc dose for the safety study delivering the AAV treatment through the intracisternal route based on our previous study.²⁷ The higher titer is necessary due to dilution of the AAV in the larger volume (∼15 mL CSF)³⁴ compared to small deposits in local brain parenchyma. This AAV dose (3–5 × 10¹³ gc) has been tested by us and other investigators in NHP safety studies.^{27,41,43,46,47} We have used as high as 1 × 10¹⁴ gc doses of AAVrh.10 into NHP CSF without safety issues or complications.²⁷ With regard to safety, the in-life parameters of the study were unremarkable.

Importantly, unlike the MRI findings in the clinical study with IPC administration,¹⁵ there was no abnormality observed in the CNS MRIs attributed to AAVrh.10hCLN2 or the ICM procedure. There was no morbidity/mortality or gross abnormality seen in any animal. In one out of two NHPs treated with AAVrh.10hCLN2, lumbar DRG showed mononuclear cell infiltrates and neuronal degeneration. The DRG abnormalities in terms of observed lesions were similar to those previously attributed to administration of AAV vectors to NHPs,^{35,38,42,48–50} but only in one animal. No other finding attributed to administration of AAVrh.10hCLN2 was observed on macroscopic postmortem examination and histopathologic assessment. As expected, all animals treated with AAVrh.10hCLN2 developed antibodies to the AAVrh.10 capsid.

Assessment of both vector DNA and TPP-1 expression demonstrated that at 5 × 10¹³ gc, AAVrh.10hCLN2 transduction was seen throughout the CNS. Vector genome levels were >1,000 copies/brain cube in 52% and 84% of the brain compared to 0% in PBS. Biodistribution of vector genome levels was not assessed in peripheral nerves as this was primarily a safety study. CSF TPP-1 levels were 43–62% of normal human CSF TPP-1 levels. ICM administration of AAVrh.10hCLN2 also led to wide distribution of TPP-1 levels in the brain and the CSF. The TPP-1 expression levels were >2 × standard deviation of PBS baseline mean in 50% and 41%, in the two NHPs treated with AAVrh.10hCLN2. In contrast to intracisternal administration in this study, we previously assessed IPC administration of the same AAVrh.10hCLN2 vector to the NHP brain.¹⁸

In that study, TPP-1 enzyme activity exceeded 2 × standard deviation above background in only 30% of brain cubes when measured 90 days after vector administration. Comparing that data to our current data demonstrates the superiority of distribution following ICM delivery. Collectively, the safety data and the vector and protein distribution data demonstrate that AAVrh.10hCLN2 delivered by ICM administration is safe and leads to wide, likely therapeutic brain and CSF distribution of TPP-1.

The safety aspects of using spinal needles to deliver the treatments to the cisterna magna provide some caution and risk, but is considered to be safer than direct injections to the brain parenchyma. ICM delivery has technical challenges due to the hidden access site at the base of the skull and it being a less common clinical procedure, unlike the spinal taps. While a direct infusion into the cisterna magna using spinal needles is challenging, it can provide consistent and convenient access with practice in large animal models.

We have not observed injuries or postadministration pathology to the NHPs in our previous ICM delivery studies. The ICM route is being used for several clinical trials (seven current and completed) to deliver AAV and small molecules for CNS disorders. To minimize risk of punctures to the brain stem, catheters can be guided by fluoroscopy or other imaging modalities and iohexol to direct the physicians to the precise site at the opening of the cisterna magna. This procedure has been published for multiple AAV trials using NHPs.^27,29,51

AUTHORs' CONTRIBUTIONS

B.P.D.—study design, conducting of study-related assays, and involved with drafting of the article; J.B.R.—study design, animal studies, data analysis, and involved with drafting of the article; N.S.—vector biodistribution assay development and analysis; I.W.—conducting of study-related assays and behavior studies; N.Y.—conducting of study-related assays and behavior studies; A.G.—conducting of study-related assays; S.M.K.—data analysis and involved with drafting of the article; L.A.H.—clinical radiologist involved with CNS MRI and assessment; R.A.—study veterinarian performed vector administrations and CSF draws; I.C.M.—study pathologist; S.M.—study pathologist; A.N.—vector biodistribution sample analysis coordination; R.K.—study design and data analysis; D.S.—study design, study oversight, data analysis, and involved with drafting of the article; R.G.C.—study design, study oversight, and involved with drafting of the article.

AUTHOR DISCLOSURE

This study was supported, in part, by an STTR grant to LEXEO Therapeutics. D.S., S.M.K., and R.G.C. hold equity in and are consultants to LEXEO. A.N., N.S., and R.K. are employees of LEXEO Therapeutics.

FUNDING INFORMATION

These studies were supported, in part, by NIH grant no. R41 NS117265 and LEXEO Therapeutics, Inc., Laboratory of Comparative Pathology is supported, in part, by Cancer Center Support Grant P30-CA008748 to Memorial Sloan Kettering Cancer Center. Wake Forest Vervet Research Colony is supported, in part, by NIH resource grant P40-OD010965.

SUPPLEMENTARY MATERIAL

Supplemental Methods

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Figure S4

Supplementary Figure S5

Supplementary Figure S6

Supplementary Figure S7

Supplementary Figure S8