Molecular Genetics and Metabolism 114 (2015) 281–293

Contents lists available at ScienceDirect

Molecular Genetics and Metabolism journal homepage: www.elsevier.com/locate/ymgme

Nonclinical evaluation of CNS-administered TPP1 enzyme replacement in canine CLN2 neuronal ceroid lipofuscinosis Brian R. Vuillemenot a,⁎, Derek Kennedy a, Jonathan D. Cooper b, Andrew M.S. Wong b, Sarmi Sri b, Thom Doeleman b, Martin L. Katz c, Joan R. Coates c, Gayle C. Johnson c, Randall P. Reed d, Eric L. Adams d, Mark T. Butt e, Donald G. Musson a, Joshua Henshaw a, Steve Keve a, Rhea Cahayag a, Laurie S. Tsuruda a, Charles A. O'Neill a a

BioMarin Pharmaceutical Inc., Novato, CA, USA King's College London, London, UK c University of Missouri, Columbia, MO, USA d Northern Biomedical Research, Muskegon, MI, USA e Tox Path Specialists, Frederick, MD, USA b

a r t i c l e

i n f o

Article history: Received 11 July 2014 Received in revised form 2 September 2014 Accepted 3 September 2014 Available online 16 September 2014 Keywords: CLN2 Neuronal ceroid lipofuscinosis Batten disease Tripeptidyl peptidase-1 Intracerebroventricular Intrathecal

a b s t r a c t The CLN2 form of neuronal ceroid lipofuscinosis, a type of Batten disease, is a lysosomal storage disorder caused by a deficiency of the enzyme tripeptidyl peptidase-1 (TPP1). Patients exhibit progressive neurodegeneration and loss of motor, cognitive, and visual functions, leading to death by the early teenage years. TPP1-null Dachshunds recapitulate human CLN2 disease. To characterize the safety and pharmacology of recombinant human (rh) TPP1 administration to the cerebrospinal fluid (CSF) as a potential enzyme replacement therapy (ERT) for CLN2 disease, TPP1-null and wild-type (WT) Dachshunds were given repeated intracerebroventricular (ICV) infusions and the pharmacokinetic (PK) profile, central nervous system (CNS) distribution, and safety were evaluated. TPP1-null animals and WT controls received 4 or 16 mg of rhTPP1 or artificial cerebrospinal fluid (aCSF) vehicle every other week. Elevated CSF TPP1 concentrations were observed for 2–3 days after the first ICV infusion and were approximately 1000-fold higher than plasma levels at the same time points. Anti-rhTPP1 antibodies were detected in CSF and plasma after repeat rhTPP1 administration, with titers generally higher in TPP1-null than in WT animals. Widespread brain distribution of rhTPP1 was observed after chronic administration. Expected histological changes were present due to the CNS delivery catheters and were similar in rhTPP1 and vehicletreated animals, regardless of genotype. Neuropathological evaluation demonstrated the clearance of lysosomal storage, preservation of neuronal morphology, and reduction in brain inflammation with treatment. This study demonstrates the favorable safety and pharmacology profile of rhTPP1 ERT administered directly to the CNS and supports clinical evaluation in patients with CLN2 disease. © 2014 Elsevier Inc. All rights reserved.

1. Introduction

Abbreviations: aCSF, artificial cerebrospinal fluid; ANOVA, analysis of variance; CHO, Chinese hamster ovary; CLN2, ceroid lipofuscinosis neuronal type 2; CNS, central nervous system; CSF, cerebrospinal fluid; D, deep tissue; ELISA, enzyme linked immunoabsorbent assay; ERT, enzyme replacement therapy; EU, endotoxin unit; H&E, hematoxylin and eosin; GFAP, glial fibrillary acidic protein; IAR, infusion associated reaction; Iba1, ionized calcium-binding adapter molecule 1; ICV, intracerebroventricular; IT, intrathecal; IT-C, intrathecal-cisternal; IT-L, intrathecal-lumbar; MEG, medial ectosylvian gyrus; MRI, magnetic resonance imaging; NBF, neutral buffered formalin; PBS, phosphate buffered saline; PK, pharmacokinetic; PV, parvalbumin; rhTPP, recombinant human tripeptidyl peptidase1; S, superficial tissue; TPP1, tripeptidyl peptidase-1; WT, wild-type. ⁎ Corresponding author at: BioMarin Pharmaceutical Inc., 105 Digital Drive, Novato, CA 94949, USA. E-mail address: [email protected] (B.R. Vuillemenot).

http://dx.doi.org/10.1016/j.ymgme.2014.09.004 1096-7192/© 2014 Elsevier Inc. All rights reserved.

The CLN2 form of neuronal ceroid lipofuscinosis is characterized by neurodegeneration and progressive functional decline, resulting in death by adolescence [6]. CLN2 disease is caused by CLN2 gene mutations leading to the loss of tripeptidyl peptidase-1 (TPP1) activity [18, 28]. In the absence of TPP1, lysosomal storage material accumulates in the CNS and other organs. The onset of symptoms begins at approximately 3 years of age, with a subsequent rapid progressive decline in neurological function over a period of several years [6]. At disease end stage children are no longer able to communicate, are immobile, suffer from chronic seizures, and must be tube fed. There is currently no disease-modifying therapy. A gene therapy trial involving an adenoassociated viral vector to deliver human CLN2 to the CNS of CLN2 patients showed evidence of slowing disease progression, but resulted in

282

B.R. Vuillemenot et al. / Molecular Genetics and Metabolism 114 (2015) 281–293

serious adverse effects [35]. Neural stem cell therapy has also been attempted in CLN2 patients, and although well tolerated, limited evidence of efficacy has been observed with this approach [27]. Enzyme replacement therapy (ERT) has been efficacious for treating lysosomal storage disorders that affect non-CNS tissues [24,36], although no effective ERTs are available for lysosomal storage disorders primarily affecting the CNS. TPP1 ERT is a potential therapy for CLN2 disease. Recombinant human (rh)TPP1 is produced as a mannose-6phosphorylated pro-enzyme lacking activity [20]. After binding to cell surface mannose-6-phosphate receptors, the pro-enzyme is taken up into the lysosome where it is cleaved to the active enzyme [11,34]. For successful rhTPP1 ERT, sufficient CNS concentrations will be necessary. An obstacle for effective ERT is enzyme delivery to the CNS, since TPP1 does not cross the blood–brain barrier. Infusion of rhTPP1 into the cerebrospinal fluid (CSF) to circumvent the blood–brain barrier in monkeys was well tolerated and resulted in enzyme distribution throughout the CNS [31]. Previous studies have demonstrated rhTPP1 pharmacological effects when delivered to the CSF in murine [5] and canine [32] disease models. CNS administration of enzymes in animal models has resulted in minimal safety findings [8,10,30]. In these studies, the implantation and presence of the CNS delivery catheters was responsible for most of the observed histological changes [4]. TPP1-null Dachshunds contain a frameshift mutation in the TPP1 gene (canine homolog of CLN2) and are a relevant animal model of CLN2 disease [1]. Affected animals display progressive neurological decline requiring euthanasia at 10–11 months old [15,26]. We have previously demonstrated that repeat intrathecal (IT) injection results in brain distribution of active rhTPP1 and reduced accumulation of lysosomal storage [32]. However, the relatively rapid administration of rhTPP1 associated with IT injections resulted in infusion associated reactions (IARs) consisting of a hypersensitivity response with clinical signs of facial swelling, hyperemia, urticarial, vomiting, hypotension, and tachycardia. Due to the IARs, animals could only be dosed with this approach until 7 months of age. This study was conducted to characterize the pharmacology and safety of rhTPP1 administered by slow infusion to the CSF to enable human clinical trials. Since infusion was expected to result in lower plasma concentrations than bolus injection, we hypothesized that IARs would be prevented, allowing for chronic administration. Pharmacokinetics (PK), CNS distribution, immunogenicity, safety, and neuropathology of ICV-infused rhTPP1 were evaluated in TPP1-null and wildtype (WT) Dachshunds and are described in this report. Additional clinical evaluations on this cohort of animals have demonstrated the potential of this treatment for efficacy in human CLN2 disease patients through the attenuation of neurodegeneration, functional improvements, and lifespan extension [14,33].

2. Materials and methods 2.1. Study design Animal procedures were conducted at the University of Missouri and followed NIH and USDA guidelines for the care and use of animals in research. All methods were conducted according to protocols reviewed and approved by the University of Missouri Institutional Animal Care and Use Committee. Dogs were housed in an AAALAC accredited facility and were socialized daily. Animals received ICV rhTPP1 every other week from approximately 2.5 months of age until reaching end-stage disease, with age-matched WT controls euthanized alongside TPP1null animals. Since brain size does not increase proportionately with body mass, fixed dose levels of 4 or 16 mg rhTPP1 were used and not normalized to body weight. Nine TPP1-null and 10 homozygous WT animals weighing 1.2–2.8 kg at study start and 3.3–6.7 kg at euthanasia were assigned to six treatment groups of 3 or 4 animals per group as shown in Table 1.

2.2. rhTPP1 and vehicle preparation and synthesis Recombinant human pro-form TPP1 (rhTPP1) produced in CHO cells [20] was formulated in artificial CSF (216.5 mM NaCl, 0.80 mM MgSO4, 3.01 mM KCl, 1.40 mM CaCl2, 0.80 mM Na2HPO4, 0.20 mM NaH2PO4, pH 7.3) to a concentration of 3.35 or 13.47 mg/mL. aCSF was administered to vehicle control animals. All rhTPP1 preparations had a specific activity N 4.5 units/mg total protein, Kuptake in TPP1-null human fibroblasts of b10 nM, and endotoxin of b0.2 EU/mL as determined by Limulus amebocyte lysate assay. 2.3. Catheter implantation surgery Animals had catheters implanted into a cerebral lateral ventricle (ICV, either left or right ventricle was used) and the lumbar subarachnoid space (IT-L) to enable rhTPP1 or vehicle infusion and serial CSF collection, respectively. Catheters were placed using MRI-guided stereotactic localization. Each catheter was connected to a subcutaneous port enabling repeat dose administration and/or CSF collection. Computed tomography of the brain was performed after ICV infusion of Iohexol contrast agent (240 mg/mL; Omnipague, GE Healthcare, Oslo, Norway) to confirm catheter location. IT-L catheter patency was confirmed by CSF collection. Animals recovered from surgery for at least 7 days prior to their first infusion. 2.4. rhTPP1 administration rhTPP1 was initially infused via the ICV catheter every other week. In many animals, reactions to the ICV delivery device eventually caused the catheter to lose patency, and growth of the skull, to which the catheter is anchored, resulted in the retraction of the catheter tip into the parenchyma. The ICV catheters were assessed before each infusion by computed tomography. If the ICV catheter had dislocated or lost patency, rhTPP1 infusion continued via the IT-L catheter. After being used for repeated infusions, many IT-L catheters eventually lost patency due to fibrotic reactions. If both catheters were no longer patent, the dose was administered as a bolus injection over a 2 minute period into the subarachnoid space at the cerebellomedullary cistern (IT-C). The total number of doses administered to each animal, as well as the route of administration of each dose, is summarized in Table 2. Animals were fasted overnight prior to dose administration and pretreated with diphenhydramine (2 mg/kg IM; WestWard Eatontown, NJ) 30 min before infusion. Additional doses of diphenhydramine were administered in the event of IARs. rhTPP1 or vehicle was infused every other week at a continuous rate of 1.2 mL over 2 h (0.6 mL/h) followed by a flush of PBS equaling the catheter/port volume at the same rate. The infusion duration was extended for the 16 mg-treated groups after the third or fourth dose to 4 h at 0.3 mL/h to mitigate IARs. WT controls received approximately 20 doses over 10 months. Dose administration continued in the TPP1-null animals until clinical signs progressed to end-stage disease. In some animals, CNS reactions associated with the implanted ICV delivery devices, including meningitis and/or obstructive hydrocephalus, mandated euthanasia before disease end-stage. Animals received a total of 15–29 doses of rhTPP1 administered every other week for 29 to 57 weeks. 2.5. Clinical evaluations Findings from morbidity and mortality observations and food consumption were recorded daily. Body weights were measured prior to surgery and prior to each infusion. Comprehensive neurological and physical examinations were conducted before surgery and weekly thereafter throughout the study. Physical examinations included body weight, heart rate, respiration, body temperature, auscultation, abdominal palpation, gait, behavior, and general appearance of the eyes, ears, oral cavity, and skin.

B.R. Vuillemenot et al. / Molecular Genetics and Metabolism 114 (2015) 281–293

283

Table 1 Study design. Group

Genotype

Treatment

Dose (mg)

Infusion duration

No. of animals

1 2 3

WT WT WT

aCSF vehicle rhTPP1 rhTPP1

– 4 16

4 3 3

4 5 6

TPP1-null TPP1-null TPP1-null

aCSF vehicle rhTPP1 rhTPP1

– 4 16

2 2 2 4 2 2 2 4

h h h (first 3 or 4 infusions) h (remaining infusions) h h h (first 3 or 4 infusions) h (remaining infusions)

3 3 3

aCSF, artificial cerebrospinal fluid; rhTPP1, recombinant human tripeptidyl peptidase 1; WT, wild-type.

Neurologic examinations included the assessment of general sensory and motor functions, cranial nerve function, and spinal reflexes, and the presence of hyperesthesia [22]. Indirect ophthalmoscopy was conducted before the first dose and monthly thereafter. Euthanasia was performed when end-stage disease, defined as the loss of cognition, severe mentation abnormalities, loss of visual tracking, drug-refractory myoclonic jerks and inability to eat without significant assistance, was reached. 2.6. Blood and CSF sampling Blood was collected from a peripheral vein before doses 1 (study Week 1, approximately 2.5 months of age), 10 (study Week 19, approximately 6.5 months of age) and 20 (study Week 39, approximately 11 months of age) for CBC and plasma chemistry. Plasma was collected immediately before infusion, at 1 h into the infusion, at the end of infusion, and at 0.5, 1, 4, 12, 24, 48, 72, 120 and 168 h post-infusion for PK assessment. These serial samples were collected after the first infusion (study Week 1 at approximately 2.5 months of age), tenth infusion (study Week 19 at approximately 6.5 months of age) and twentieth infusion (study Week 39 at approximately 11 months of age). CSF was collected via IT-L catheter before dose 1, at 1 h into and at the end of infusion, and at 0.5, 1, 4, 12, 24, 48, 72, 120 and 168 h postinfusion for PK assessment. Serial CSF samples were also collected before dose 10 and at the same time points post-dose for two animals administered the 16 mg dose. CSF analysis for total cell count and protein concentration was performed before the first dose, at necropsy, and when clinical signs suggested meningitis. 2.7. rhTPP1 analysis rhTPP1 pro-enzyme was measured in plasma, CSF, and CNS tissue by enzyme linked immuno-absorbent assay (ELISA). This assay is specific for the recombinant human form of TPP1 and does not detect the endogenous canine TPP1 present in WT animals. One ELISA was used to measure the pro-enzyme form of rhTPP1 in plasma and CSF, and a second ELISA to assess the catalytically active form in CNS tissue. Tissue ELISA results were normalized to grams of total protein in each homogenate,

as determined with a BCA protein assay kit (Pierce, Rockford, IL). PK parameters were calculated from plasma and CSF exposure data using WinNonlin software (Version 6.1, Pharsight Corporation, Mountain View, CA).

2.8. Total anti-rhTPP1 antibody analysis An electrochemiluminescent immunoassay (Meso Scale Discovery, Rockville, MD) was used to evaluate total anti-rhTPP1 antibodies in plasma and CSF. Samples were screened for reactivity and considered reactive if their average relative light unit value exceeded the plate screening cut point. Presumptive positive samples underwent confirmation by the determination of cut point. Cut points for antibody screening, confirmation, and titer were determined prior to screening study samples. True positive samples were semi-quantified by titration by lowering the detection signal below a defined limit after adding excess rhTPP1.

2.9. Necropsy Forty-eight hours post-final dose, animals were euthanized and brains cut into serial 4 mm coronal sections. Selected brain structures were sampled using a 4 mm diameter biopsy punch and frozen for rhTPP1 analysis. Samples consisted of superficial (b4 mm deep) and deep (N4 mm deep) pericruciate cortex, occipital cortex, cerebellum, striatum, medulla, midbrain, pons, hypothalamus, and thalamus. Superficial samples were immediately proximal to CSF, while deep samples were just interior to the superficial samples. The remainders of the brain sections were fixed in 10% neutral buffered formalin (NBF) and processed for histology. The heart, lung, kidney, and liver were collected and weighed. Samples of each tissue were fixed in 10% NBF for histopathology. Spinal cord segments at C3, T7, and L1 were collected. Each segment was divided and one piece frozen and the other fixed. For 6 animals (one animal each from Groups 1, 3, and 4 and all 3 animals from Group 6), additional tissues were collected and fixed for more extensive histopathology (aorta, cervix, colon, esophagus, gallbladder, lymph nodes, ovary,

Table 2 Dose numbers and routes of administration used for each animal. Group

Animal

Number of doses/route

Group

Animal

No. of doses/route

1

A B C D E F G H I J

3 ICV, 18 IT-L (21 total) 5 ICV, 9 IT-L, 1 IT-C (15 total) 15 ICV, 7 IT-L, 3 IT-C (25 total) 15 ICV, 5 IT-L (20 total) 5 ICV, 7 IT-L, 6 IT-C (18 total) 10 ICV, 4 IT-L, 6 IT-C (20 total) 19 ICV (19 total) 5 ICV, 4 IT-L, 12 IT-C (21 total) 9 ICV, 11 IT-L (20 total) 5 ICV, 16 IT-L, 4 IT-C (25 total)

4

K L M

11 ICV, 8 IT-L (19 total) 12 ICV, 3 IT-L (15 total) 5 ICV, 8 IT-L, 2 IT-C (15 total)

5

N O P Q R S

6 ICV, 7 IT-L, IT-C (22 total) 3 ICV, 20 IT-C (23 total) 11 ICV, 14 IT-L (25 total) 5 ICV, 6 IT-L, 18 IT-C (29 total) 19 ICV, 1 IT-L, 3 IT-C (23 total) 15 ICV, 8 IT-L, 4 IT-C (27 total)

2

3

ICV, intracerebroventricular; IT-C, intrathecal-cisternal; IT-L, intrathecal-lumbar.

6

284

B.R. Vuillemenot et al. / Molecular Genetics and Metabolism 114 (2015) 281–293

uterus, pancreas, pituitary gland, skin/subcutis, stomach, thymus, thyroid with parathyroid, trachea, urinary bladder and testes). Frozen tissues were analyzed for rhTPP1 by ELISA. Fixed samples were stained for histopathology and evaluated by a board-certified veterinary pathologist specializing in CNS pathology blind to treatment and genotype. 2.10. Neuropathological evaluation Coronal sections, with a 7 μm thickness, were cut from the thalamus and medial ectosylvian gyrus (MEG) of the cerebral cortex at the ICV catheter level. Parallel series of sections (every eighth section) through each region were Nissl stained to reveal cytoarchitectural organization, or immunostained for glial fibrillary acidic protein (GFAP) as a marker of astrocytosis, ionized calcium-binding adapter molecule 1 (Iba1) as a marker of microglial activation, or parvalbumin (PV) as a marker of GABAergic interneurons. Additional sections were examined for autofluorescent storage. All evaluations were performed blind to genotype and treatment. Cortical thickness measurements were obtained from Nissl stained sections using StereoInvestigator software (Microbrightfield Inc., Williston, VT, USA). Optical fractionator counts of the number of Nissl stained lamina V neurons, and of the number of PV positive interneurons in laminae II–VI were made in the medial ectosylvian gyrus (MEG), again using StereoInvestigator software. Beginning with a randomly chosen starting section, these counts were made in three regularly spaced sections using a grid size of 375 × 375 μm, dissector frame 168 × 94 μm and × 40 objective (Nissl stained neurons), or a grid size of 400 × 400 μm, dissector frame 330 × 185 μm and ×20 objective (PV stained interneurons). Only neurons with a clearly identified morphology that fell within the dissector frame were counted. For all optical fractionator estimates, the mean coefficient of error (CE) of individual estimates was calculated according to the method of Gundersen and Jensen [12] and was less than 0.08 in all these analyses. The mean size of lamina V pyramidal neurons in the MEG was obtained by measuring cross sectional area of at least 100 pyramidal neurons per animal using StereoInvestigator software [7]. The expression of GFAP and Iba1 was measured by semi-quantitative thresholding image analysis [16], using Image-Pro Plus 5.0 software (MediaCybernetics, Chicago, IL, USA) (N = 3 animals per treatment group). Statistical significance was determined by one-way ANOVA with post-hoc Bonferroni, with P ≤ 0.05 considered significant. 3. Results 3.1. rhTPP1 was generally well-tolerated There were no changes in body weight, weight gain, hematology, plasma clinical chemistry, or CSF cell count and chemistry due to rhTPP1 administration in any animals. Mild to moderate IARs were observed after infusion of 16 mg rhTPP1 over 2 h in WT and TPP1-null animals in association with infusions 3 through 5 (Weeks 5 to 9). Clinical signs included facial edema, erythema, urticaria, hypotension, diarrhea, and vomiting. These reactions subsided with additional doses of diphenhydramine (2 mg/kg, IM) and resolved completely within 24 h. Extension of the infusion from 2 to 4 h eliminated IARs. After repeated rhTPP1 infusions, dosing over a 2 minute period via cisternal injection was well tolerated with no IARs. Clinical signs of inflammation associated with ICV and/or IT-L catheters were observed, including cervical hyperesthesia, lumbar hyperesthesia, CSF pleocytosis, and edema around the catheters. These reactions occurred in all animals with no apparent dose or temporal relationship, resolved within 24 h of dosing, and were likely related to the CNS catheters [4]. There were no clinical signs indicative of off target effects, exaggerated pharmacology, or attributable to rapid catabolism of accumulated lysosomal storage material by rhTPP1. Neurological

examinations performed on these animals indicate that rhTPP1 delayed or prevented the appearance of neurodegenerative signs caused by CLN2 disease, as described fully in an additional manuscript [14]. 3.2. ICV rhTPP1 infusion resulted in favorable PK and CNS exposure rhTPP1 pro-enzyme was detected at high levels in CSF following the first ICV infusion (Fig. 1A). CSF rhTPP1 exposure peaked within 3 h after the infusion, with the mean CSF Cmax approximately 3-fold higher in the 16 mg-treated animals than at the 4 mg dose. There was a high level of inter-animal variability in exposure at each dose level. Post-dose reductions in CSF rhTPP1 followed bi- or triphasic kinetics, with the majority cleared from the CSF within 168 h after the end of infusion. The CSF concentration remained above the in vitro Kuptake, the concentration at which uptake is half maximal (~2 nM; [20]), for 2 to 3 days post-dose. CSF rhTPP1 profiles in TPP1-null and WT animals were similar. CSF was sampled for PK from two 16 mg-treated animals after dose 10, administered as a 4 hour infusion. Extension of the infusion from 2 to 4 h increased the time to maximal CSF rhTPP1. Peak CSF concentrations and elimination profiles following dose 10 were similar dose 1. rhTPP1 pro-enzyme was also present in plasma following the first ICV infusion (Fig. 1A). Maximal plasma levels were approximately 3 orders of magnitude below CSF levels. The mean plasma Cmax was approximately 3.4-fold greater following the 16 mg dose than the 4 mg dose. There was a high level of inter-animal variability in exposure at a given dose level. Elimination of rhTPP1 from the plasma followed bior triphasic kinetics, with the majority cleared within 80 h after the end of infusion for both dose levels. Plasma rhTPP1 levels following doses 10 and 20 were significantly lower than those after dose 1. Mean CSF and plasma PK parameters are presented in Tables 3 and 4. rhTPP1 was detected along the length of the spinal cord (Fig. 1B) and in all of the brain structures sampled (Fig. 1C). In most structures, tissue collected from the superficial layers and from deeper within the structure had similar rhTPP1 levels, indicating substantial enzyme penetration. Overall, mean tissue levels were similar between the two dose levels. The higher mean rhTPP1 concentrations observed at the 4 mg than the 16 mg dose in some tissues was likely due to the high interanimal variability in tissue rhTPP1 concentration. 3.3. Anti-rhTPP1 antibodies were detected in CSF and plasma Antibodies against rhTPP1 were detected in CSF and plasma (Fig. 2). Ant-rhTPP1 antibodies were first detected in plasma from a 4 mgtreated WT animal at 14 days after the first dose and were present in CSF and plasma of all animals that received rhTPP1 within 1 to 4 months following the first dose. Titers generally increased over the first 2 to 4 months of dosing and then remained level for the remainder of the study. The increase in antibody titers correlated with the decrease in rhTPP1 detected in the plasma, but not CSF. Anti-rhTPP1 titers were approximately 10- to 100-fold higher in the TPP1-null than in the WT animals. Titers were similar between CSF and plasma and both dose levels. No anti-rhTPP1 antibodies were detected in baseline samples collected before the first rhTPP1 dose or in vehicle-treated animals at any time point. 3.4. Histological changes were present due to the CNS delivery system There were no rhTPP1-related changes in organ weights or gross lesions in the visceral organs. Microscopic findings were limited to the CNS and were largely associated with the CNS catheters (Fig. 3). Inflammation, perivascular infiltrates, and focal fibrosis were present adjacent to the ventricles and spinal cord and along the catheter tracks in all animals after rhTPP1 or vehicle administration. Some microscopic findings were more pronounced in rhTPP1-treated animals than in those which received vehicle, including the loss of ependymal cells, ventricular inflammation, and edema along the ICV catheter track. Microscopic

B.R. Vuillemenot et al. / Molecular Genetics and Metabolism 114 (2015) 281–293

285

Fig. 1. rhTPP1 exposure following ICV infusion. Exposure data from TPP1-null and WT animals at each dose level were similar and were grouped together for this analysis. A) rhTPP1 was present at high levels in CSF and lower levels in plasma following the first ICV infusion of 4 or 16 mg. B) rhTPP1 was distributed over the spinal cord after repeat administration. C) rhTPP1 was widely distributed in the brain after repeat administration. At necropsy, brains were sectioned into serial 4 mm coronal sections. From the appropriate sections, two 4 mm tissue punch samples were collected, with one of the samples from each structure superficial (b4 mm deep) and the other deep (N4 mm deep) in location. rhTPP1 concentrations in the brain and spinal cord were normalized to the total protein present in each sample. Results from animals administered 4 mg rhTPP1 are shown in light gray, while those from animals dosed at the 16 mg level are shown in dark gray. Error bars represent SD. D, deep; S, superficial.

changes associated with rhTPP1 were similar to those observed after the administration of other proteins directly to the CNS [4] and were highly variable between and within groups. The changes associated with rhTPP1 infusion were not associated with adverse clinical effects. Histological changes related to CLN2 disease progression were observed in CNS tissues from all TPP1-null animals, including the accumulation of

lysosomal storage, neuronal loss or degeneration, and gliosis. Gross brain lesions observed in two animals consisted of neutrophilic inflammation along the catheter track in a 4 mg rhTPP1-treated TPP1-null animal and an inflammatory mass at the tip of the IT-L catheter in a vehicle-treated WT animal. There were no rhTPP1-related microscopic findings in non-CNS tissues.

Table 3 Summary of rhTPP1 CSF PK parameters.a Dose (mg)

Dose number

Nb

AUC0–t (ng-h/mL)

AUC0–∞ (ng-h/mL)

Cmax (ng/mL)

Tmax (h)

t1/2 (h)

Vz (mL)

CL (mL/h)

1.88 × 106 (5.98 × 105) 6.45 × 106 (3.45 × 106) 1.52 × 107 (1.49 × 107)

1.88 × 106 (5.98 × 105) 6.45 × 106 (3.45 × 106) 1.52 × 107 (1.49 × 107)

4.73 × 105 (2.46 × 105) 1.46 × 106 (9.13 × 105) 2.29 × 106 (2.71 × 106)

2.75 [2.5, 3] 3 [2.5, 5] 6.5 [5, 8]

10.2 (5.34) 5.86 (3.39) 2.71 (0.0184)

35.9 (27.9) 54.5 (96.9) 6.71 (9.49)

2.25 (0.717) 5.66 (9.79) 1.71 (2.42)

4

1

2

16

1

6

16

10

2

AUC0–t, area under the CSF concentration–time curve from time 0 to the last measurable concentration; AUC0–∞, area under the CSF concentration–time curve from time 0 to infinity; Cmax, maximum concentration; Tmax, time of maximum concentration; t1/2, half-life; Vz, volume of distribution based on terminal phase; CL, clearance. a Values reported as mean (SD), except for Tmax which is reported as median [min, max]. b CSF samples for PK analysis were not collected from all animals due to the loss of patency of the IT-L catheter.

286

B.R. Vuillemenot et al. / Molecular Genetics and Metabolism 114 (2015) 281–293

Table 4 Summary of rhTPP1 plasma PK parameters.a,b Dose (mg)

Dose number

N

AUC0–t (ng-h/mL)

AUC0–∞ (ng-h/mL)

Cmax (ng/mL)

Tmax (h)

t1/2 (h)

Vz (mL)

CL mL/h

1.85 × 103 (1.90 × 103) 3.80 × 101 (5.04 × 101) 5.15 × 103 (1.91 × 103) 1.42 × 103 (2.69 × 103) 7.25 × 101 (2.30 × 101)

1.89 × 103 (1.89 × 103) 7.92 × 101 (2.83 × 101) 5.18 × 103 (1.91 × 103)

5.27 × 102 (6.95 × 102) 1.71 × 101 (2.34 × 101) 1.78 × 103 (9.07 × 102) 1.68 × 102 (3.09 × 102) 1.81 × 101 (5.52 × 100)

2 [1, 2] 2 [1, 6] 2.25 [2, 6] 8 [4, 8] 4 [4, 4.5]

10.4 (13.9) 8.12 (9.90) 8.97 (1.89) – (–) 9.75 (–)

7.33 × 104 (1.29 × 105) 7.69 × 105 (9.96 × 105) 4.69 × 104 (4.67 × 104) – (–) 1.29 × 106 (–)

3.51 × 103 (1.95 × 103) 5.39 × 104 (1.93 × 104) 3.29 × 103 (2.54 × 103) – (–) 9.19 × 104 (–)

4

1

6

4

10

3c

16

1

6

16

10

4d

16

20

3e

1.74 × 102 (–)

AUC0–t, area under the plasma concentration–time curve from time 0 to the last measurable concentration; AUC0–∞, area under the plasma concentration–time curve from time 0 to infinity; Cmax, maximum concentration; Tmax, time of maximum concentration; t1/2, half-life. a Values reported as mean (SD), except for Tmax which is represented as median [min, max]. b Due to the loss of patency of the IT-L catheter for CSF sample collection, additional plasma samples, compared to CSF samples, were collected for analysis. c N = 2 for AUC0–∞, t1/2, Vz and CL. d N = 0 for AUC0–∞, t1/2, Vz and CL. e N = 1 for AUC0–∞ and t1/2.

3.5. Neuropathological evaluation indicates that rhTPP1 attenuated CLN2 disease progression The impact of rhTPP1 administration upon CLN2 disease neuropathology was assessed CNS tissues from all animals. Immunostaining for the astrocyte marker GFAP revealed pronounced CNS astrocytosis in vehicle-treated TPP1-null animals. This contrasted with the much paler GFAP immunoreactivity present in astrocytes in vehicle-treated WT animals. Within the thalamus, rhTPP1 had very little effect upon astrocytosis in TPP1-null animals, and moderately increased GFAP staining in WT animals. A dose-dependent increase in cortical GFAP immunoreactivity was seen in rhTPP1 treated WT animals that remained very low compared to that in TPP1-null animals (Fig. 4A). In TPP1-null animals, there was a trend to a decrease in cortical astrocytosis in

animals that received 16 mg rhTPP1, but not the 4 mg dose (Fig. 4A). Persisting astrocytes in 16 mg rhTPP1 treated TPP1-null animals exhibited a smaller more palely stained soma with thinner processes, suggesting lower activation. Thresholding image analysis indicated reduced cortical astrocytosis in 16 mg rhTPP1 treated TPP1-null animals (Figs. 4B, C). In contrast to the effects upon astrocytosis, Iba1 immunostaining revealed little effect of rhTPP1 upon cortical microglial activation (Fig. 5A). Intensely stained Iba1 positive microglia with a swollen cell body and thickened processes were evident throughout the cortex and thalamus of TPP1-null animals with very little difference in Iba1 staining or microglial morphology between groups (Fig. 5A). There was no increase in Iba1 immunoreactivity in response to rhTPP1 in either TPP1-null or WT animals. However, thresholding image analysis revealed a non-

Fig. 2. Anti-rhTPP1 antibodies were detected in CSF and plasma after repeat every other week administration of rhTPP1. TPP1-null animals are indicated by open symbols, while WT animals are indicated by closed symbols. No anti-rhTPP1 antibodies were present in samples collected prior to the first infusion. (A) CSF anti-rhTPP1 at the 4 mg dose level. (B) CSF antirhTPP1 at the 16 mg dose level. (C) Plasma anti-rhTPP1 at the 4 mg dose level. (D) Plasma anti-rhTPP1 at the 16 mg dose level.

B.R. Vuillemenot et al. / Molecular Genetics and Metabolism 114 (2015) 281–293

287

Fig. 3. CNS catheter-related findings were observed in CNS histology. Representative H&E stained brain sections after repeat administration of aCSF vehicle are shown in A and F, while those following the 4 mg rhTPP1 administration are shown in B–E. Inflammatory reactions largely due to the presence of and infusion through the CNS catheters were observed. (A) ICV catheter track with minimal inflammation in a vehicle-treated WT animal. The circular space is the lumen of the catheter. (B) Severe inflammation adjacent to the ICV catheter track in a 4 mg rhTPP1-treated TPP1-null animal. (C) Multiple foci of inflammation and neovascularization (purple areas) in tissue adjacent to the lateral ventricle in a WT animal that received 4 mg doses of rhTPP1. The lateral ventricle is labeled. (D) Chronic inflammation and fibrosis in the choroid plexus adjacent to the site of rhTPP1 administration in a WT animal that received 4 mg doses of rhTPP1. (E) Minimal inflammation around periventricular blood vessels in a 4 mg rhTPP1-treated TPP1-null animal. (F) Inflammatory mass surrounding the ITL catheter in a vehicle-treated WT animal. The arrows indicate the border between the mass (IM) and the adjacent spinal cord (SC) which appears to be compressed. The circular space is the lumen of the catheter. The scale bar indicates 200 μm (B–D) or 500 μm (A, E, F).

statistically significant dose-dependent trend to decreased Iba1 staining in enzyme treated TPP1-null animals (Figs. 5B, C). Nissl staining revealed pronounced cortical atrophy in TPP1-null animals at disease end-stage in both rhTPP1 and vehicle-treated dogs (Fig. 6). However, neuronal morphology was preserved in the rhTPP1treated animals. In vehicle-treated TPP1-null animals the lamina V pyramidal neurons appeared distended and full of storage material with thickened twisted dendrites (Fig. 6A). In contrast, in rhTPP1-treated animals the morphology of these neurons more closely resembled WT animals, with a smaller cell soma and thinner straighter dendrites (Fig. 6A). Counts

of parvalbumin-positive interneurons, affected in CLN2 disease, revealed a pronounced dose-dependent neuroprotective trend of rhTPP1 treatment within the cortex (Fig. 6C). Neuron counts in TPP1-null animals revealed moderate dose-dependent effects of rhTPP1 upon neuron survival (Fig. 6D). The area of lamina V cortical neurons was significantly reduced in 4 mg- and 16 mg-treated TPP1-null animals to a similar size as vehicletreated WT controls (Fig. 6E). Confocal microscopy confirmed the effects of rhTPP1 upon lysosomal storage, with less storage material generally present in cortical and thalamic neurons at both dose levels, although still above WT levels

288

B.R. Vuillemenot et al. / Molecular Genetics and Metabolism 114 (2015) 281–293

Fig. 4. Reduction in cortical astrocytosis after rhTPP1 administration. (A) Immunostaining for glial fibrillary acidic protein (GFAP) reveals the extent of astrocytosis in the cortex of WT and TPP1-null animals that received 4 or 16 mg recombinant rhTPP1 or vehicle (N = 3). The low level of astrocytosis evident in vehicle-treated WT animals was only moderately increased in rhTPP1-treated WT animals, with astrocytes displaying a predominantly protoplasmic morphology (insets). The pronounced astrocytosis evident in vehicle-treated TPP1-null animals was attenuated in 16 mg rhTPP1-treated animals, which exhibited paler GFAP immunoreactivity. This effect was also apparent at higher magnification with fewer and smaller astrocytes in 16 mg rhTPP1-treated TPP1-null animals compared to the hypertrophied morphology of fibrous astrocytes in the cortex of either vehicle-treated or 4 mg rhTPP1-treated TPP1-null animals. Scale bars = 200 μm (lower magnification) or 20 μm (higher magnification). (B, C) Thresholding image analysis indicated a non-statistically significant trend to reduced GFAP immunoreactivity in the 16 mg rhTPP1-treated TPP1-null animals in both superficial (B) and deep layers (C).

(Fig. 7). Taken together, these data reveal that rhTPP1 treatment had positive effects upon several key CLN2 disease related neuropathologies. 4. Discussion We have evaluated the safety, PK, and CNS distribution of chronic repeat-dose rhTPP1 in an animal model of CLN2 disease, the TPP1-null Dachshund. This assessment indicates a low likelihood of drug related adverse events when administered via ICV infusion, with observed safety findings largely related to the implantation and long term presence of the CNS delivery catheters as well as the canine immune response to a heterologous human protein. An advantage of assessing safety in TPP1-null animals is that the evaluation may be more relevant to CLN2 disease patients undergoing similar neurological changes than assessments in normal animals. The underlying disease process may affect rhTPP1 PK and reactions to the recombinant protein. Inclusion of WT animals allowed for the characterization of effects independent of changes due to disease progression. There were minimal safety findings

after repeat CNS administration. No changes in clinical examinations, clinical pathology, organ weights, or histology were associated with the administration of the recombinant protein. In addition, there were no effects due to the rapid catabolism of accumulated lysosomal storage materials that likely resulted from rhTPP1 administration. IARs observed in the 16 mg treated animals early on in the course of their treatment during and after rhTPP1 infusion coincided with the appearance of CSF and plasma antibodies. These reactions were less severe than those observed following repeat bolus intrathecal injection in these animals [32]. The lower severity is likely due to the reduced systemic exposure resulting from infusion as compared with bolus injection. These reactions were manageable by extending infusion time and premedication with diphenhydramine and were likely due to the canine immune response to the heterologous recombinant human protein. The human and canine forms of TPP1 are approximately 95% homologous [34]; the 5% difference was sufficient to trigger an immune response in WT animals with normal canine TPP1 expression. The lack of the endogenous canine TPP1 likely triggered a more robust immune response

B.R. Vuillemenot et al. / Molecular Genetics and Metabolism 114 (2015) 281–293

289

Fig. 5. rhTPP1 administration did not influence cortical microglial activation. (A) Immunostaining for ionized calcium-binding adapter molecule 1 (Iba1) reveals the extent of microglial activation in the cortex of WT and TPP1-null animals that received rhTPP1 or vehicle (N = 3). Pronounced microglial activation present in vehicle-treated TPP1-null animals was not affected by rhTPP1 administration. TPP1-null animals from all treatment groups displayed intensely stained microglia with swollen cell bodies and thickened processes, although these were less abundant in 16 mg rhTPP1-treated animals. In WT animals there was also no overt difference between the distribution and morphology of Iba1 positive microglia, which displayed a smaller cell soma with many branched processes. Scale bars = 200 μm (lower magnification) or 20 μm (higher magnification). (B, C) Thresholding image analysis revealed a dosedependent trend to reduced area of Iba1 immunoreactivity in rhTPP1-treated TPP1-null animals, but this did not reach statistical significance in either superficial (B) or deep layers (C).

in TPP1-null animals. This is consistent with the higher antibody titers in TPP1-null animals compared to WT controls. In a previous study, severe infusion associated reactions (IARs) were observed after repeat intracisternal injections of rhTPP1 in TPP1-null Dachshunds [32]. None of the animals in the present study that received bolus injections exhibited such reactions. Since the animals that received bolus injections in this study had previously received multiple ICV/IT-L infusions, this indicates that repeated slow infusions of rhTPP1 can result in desensitization. In contrast to our earlier Dachshund study, animals in this study only received the bolus injections later in the course of their treatment, after they were apparently desensitized by the repeat ICV/IT-L infusions. Based on these findings, it may be possible to achieve higher dose levels of ICV rhTPP1 without IARs by initially treating with low doses and then gradually increasing the dose. The IARs observed in Dachshunds are not necessarily predictive of similar responses in CLN2 disease patients. IARs in human patients have been reported for systemically administered ERTs, and protocols have been established to treat them which may be applicable

if systemic hypersensitivity reactions occur in patients receiving rhTPP1 [9,21]. Any safety signals associated with the treatments, except for manageable IARs, were not rhTPP1-related. Histological changes due to the CNS delivery devices, including inflammation, edema, and perivascular infiltration, were present in all animals, including vehicle controls. These changes were most prominent adjacent to the catheters and delivery site and are typical effects of direct CNS administration of proteins [4]. We observed similar catheter-related effects in our work with ICV administered rhTPP1 in monkeys [31]. It is likely impossible to implant and administer therapeutics through a CNS catheter without causing histological changes to the surrounding tissue. However, CNS catheters have a long history of safety for repeat administration of drugs to human patients [17,29]. Maintaining patent ICV catheters for dose administration in juvenile Dachshunds for chronic periods was challenging; to our knowledge this represents for the first time the implanted ICV catheters that have been used in such animals. In many cases, ICV catheter patency was lost due

290

B.R. Vuillemenot et al. / Molecular Genetics and Metabolism 114 (2015) 281–293

Fig. 6. Neuronal morphology was preserved following rhTPP1 administration. (A) Nissl staining was performed on the cortex from WT and TPP1-null animals that received infusions of rhTPP1 or vehicle (N = 3). This revealed pronounced cortical thinning in TPP1-null animals that was not affected by rhTPP1. However, lamina V neuronal morphology was preserved in rhTPP1 animals by rhTPP1 treatment, more closely resembling that in WT animals than the swollen and distended morphology in vehicle-treated TPP1-null animals (insets). Scale bars = 200 μm (lower magnification) or 50 μm (higher magnification). (B) Cortical thickness measurements indicated no statistically significant impact of rhTPP1 upon cortical atrophy, as is also evident in the lower power micrographs in panel A (the cortex in sections from TPP1-null animals of all three treatment groups, bottom row, display marked thinning compared to sections from WT animals, top row). (C, D) Counts of the number of parvalbumin-positive interneurons (C) and Nissl stained lamina V pyramidal neurons (D), revealed a non-significant trend to increased neuron number in rhTPP1-treated TPP1-null animals. (E) Measurements of the area of Nissl stained lamina V neurons revealed that these neurons were significantly smaller in rhTPP1-treated TPP1-null animals compared to vehicle-treated controls.

to the buildup of debris (likely inflammatory cells) in the lumen of the catheter, or the catheter tip dislocated from the ventricle due to the growth of the skull, to which the catheter is anchored. In such cases, dosing continued via the IT-L and/or IT-C routes. The challenges in

maintaining the ICV catheter may be related to the smaller area of the ventricular system in dogs compared to humans and lack of room for error as well as the reactivity of the dog immune system to the heterologous human enzyme. ICV catheters have a long history of successful

B.R. Vuillemenot et al. / Molecular Genetics and Metabolism 114 (2015) 281–293

291

Fig. 7. Lysosomal storage burden was reduced in rhTPP1-treated animals. Confocal micrographs from unstained sections through the cortex (A) and thalamus (B) reveal the impact of 4 mg or 16 mg recombinant human TPP1 (rhTPP1) treatment compared to animals that received artificial cerebrospinal fluid vehicle (vehicle-treated). In WT animals from all treatment groups very little punctate storage material was present in any cell type, although several blood vessels contained autofluorescent material (indicated by red arrowheads). This phenotype was more pronounced in the cortex than in the thalamus, and was presumably due to the incomplete perfusion of this tissue. In comparison, vehicle-treated TPP1-null animals showed a pronounced intracellular accumulation of autofluorescent storage material, which was revealed to be punctate in nature at higher magnification (insets). Administration of 4 or 16 mg rhTPP1 partially reduced the amount of storage material, with less punctate material present within cells (insets), although this was not reduced to the level in WT animals. Compared to the cortex (A), the clearance of storage material appeared to be more pronounced in the thalamus (B).

and safe use in pediatric patients [19], so the management of the ICV catheters that will be used to administer rhTPP1 to CLN2 patients should be more straightforward. Although the differences in the route of administration may have affected the CNS distribution of rhTPP1 in the dogs, CSF and plasma exposure should be similar between routes, as we observed previously in monkeys [31]. The relatively small number of animals per group and the early age at which the treatments started are additional potential limitations when extrapolating the results of this study to human CLN2 patients. Due to their rarity and low reproductive capacity, the number of TPP1-null Dachshunds available for enrollment to the study was limited. Given the low variability in disease progression observed in this animal model, an N = 3 per group, including both TPP1-null and WT control

animals, was sufficient to characterize the pharmacological effects and safety of rhTPP1. Enzyme treatments commenced at the earliest possible age in the dogs, before the onset of clinical neurodegeneration. It is expected that CLN2 patients who will receive rhTPP1 will display CLN2 disease that has progressed further, past the onset of neurological symptoms. All things being equal, it is likely that the earlier the age at which treatment starts in these patients, the more likely it will be efficacious. Additional studies in the dog model that begin rhTPP1 at a later age may add relevance to the likely clinical scenario. There were no findings related to exaggerated pharmacology or offtarget effects of rhTPP1, similar to our safety assessment of rhTPP1 in monkeys [31]. The safety of rhTPP1 is likely related to its biochemistry and PK profile. rhTPP1 was administered as an inactive pro-enzyme

292

B.R. Vuillemenot et al. / Molecular Genetics and Metabolism 114 (2015) 281–293

that required lysosomal uptake for activation. In the CSF and plasma, rhTPP1 is inactive [11]. After ICV infusion of rhTPP1, approximately 99% remained in the CNS, further reducing the chances of systemic safety findings. The activity of TPP1 is specific to the cleavage of tripeptides from the free amino termini of oligopeptide substrates, resulting in the catabolism of peptides that accumulate in the lysosome [23]. Once activated, off-target effects are unlikely due to the substrate specificity and the fact that the mature enzyme only has a significant activity at acidic lysosomal pH. Taken together, the evaluation of ICV rhTPP1 in Dachshunds indicates a low risk of drug-related safety findings. All potential safety findings are clinically manageable and not expected to result in significant concerns for human CLN2 disease patients. Our neuropathological evaluation indicated that rhTPP1 affected many CLN2 disease related pathologies. Long lasting reduction of astrocytosis occurred. Neuron survival analysis revealed positive trends that did not reach statistical significance, likely due to the relatively small number of animals in each group. The rhTPP1 treatments increased the lifespan of these animals in a dose-responsive manner [14]. The average lifespans of TPP1-null dogs treated with vehicle, 4 mg rhTPP1, or 16 mg rhTPP1 were 42, 53, and 62 weeks, respectively. The differences in the ages of the animals when the histological samples were collected may have affected the neuropathological evaluation; the results may have been more significant if all animals were of the same age at euthanasia. Since the animals were maintained until disease end-stage, the impact upon neuropathology is promising, especially the improvement in neuronal morphology and reduction of storage burden. These data indicate that rhTPP1 ERT has strong potential to impact brain pathology in human CLN2 disease patients. Following ICV infusion, rhTPP1 was detected at high levels in the CSF and remained elevated above the Kuptake (the concentration at which uptake is half maximal, ~ 60–120 ng/mL) for two to three days after the infusion, an extended period that enabled distribution into the CNS. Enzyme was also detected in plasma, although maximal concentrations were approximately three orders of magnitude less than those in CSF. rhTPP1 was likely present in plasma due to the natural outflow of CSF through the venous sinuses across the arachnoid granulations or lymphatics [3,25]. Movement of the enzyme out of the CNS into the systemic circulation, and the consequent immune response, likely caused the IARs observed. Plasma rhTPP1 levels were lower after doses 10 and 20 compared to post-dose 1, correlating with the presence of anti-rhTPP1 antibodies. However, CSF concentrations were similar after dose 10 compared to dose 1. It is possible that the reduction of plasma exposure as the study progressed was due to the neutralization of rhTPP1 by the antibodies. If so, the positive CNS pharmacological effects sustained in the presence of these antibodies may indicate that the higher CSF enzyme concentrations were able to overcome any neutralizing effects of the antibodies. Since immunogenicity to human proteins in animals typically does not translate well to human patients [2], it will be important to characterize the relationships between antirhTPP1 antibodies, CSF/plasma exposure, and any IARs observed with efficacy in the clinic. The plasma and CSF pharmacokinetic profile observed in the Dachshunds was very similar to that observed after ICV infusion to monkeys [31]. rhTPP1 was detected in all areas of the CNS analyzed after repeat administration. The widespread CNS distribution pattern of rhTPP1 is likely due to the circulation of CSF around the brain and spinal cord. ICV infusion of rhTPP1 places the enzyme close to the primary CSF source, the choroid plexus [37]. The precise distribution pattern necessary to stabilize or reverse CLN2 disease progression is unknown. However, it is likely that the more extensive the CNS distribution achieved in CLN2 disease patients, the more likely that therapeutic benefits will result. CLN2 disease related neurodegeneration is primarily characterized by declines in cognition, motor skills, coordination, and visual ability [13]. Based on these declines, necessary sites in the brain in which to restore TPP1 activity to treat the disease likely include the cerebral cortex, thalamus, striatum, and cerebellum. In these sites, rhTPP1 was detected

in both the superficial tissue (often in direct contact with CSF) and deep tissue (N 4 mm removed from CSF flow) layers. The presence of rhTPP1 in deep tissue samples, as well as the extensive distribution in different structures, offers promise that ICV administration may lead to the restoration of TPP1 activity over much of the CNS in CLN2 disease patients. 5. Conclusions We have demonstrated the favorable safety and pharmacology resulting from repeat ICV administration of rhTPP1 in TPP1-null and WT Dachshunds. Safety findings were largely due to the implanted CNS catheters. The high CSF enzyme concentrations resulted in distribution throughout the CNS, leading to reduction in lysosomal storage accumulation and improvements in CNS cellular phenotypes. In conjunction with our studies indicating the attenuation of disease progression and functional improvements in this animal model [14,33], this work offers strong evidence for the potential safety and efficacy and supports the currently ongoing clinical evaluation of ICV administered rhTPP1 in patients with CLN2 disease. Conflicts of interest Brian R. Vuillemenot, Derek Kennedy, Donald G. Musson, Joshua Henshaw, Steve Keve, Rhea Cahayag, Laurie S. Tsuruda, and Charles A. O'Neill are employees of BioMarin Pharmaceutical Inc. and stockholders in BioMarin Pharmaceutical Inc. Jonathan D. Cooper, Andrew M.S. Wong, Sarmi Sri, Thom Doeleman, Martin L. Katz, Joan R. Coates, Gayle C. Johnson, Randall P. Reed, Eric L. Adams, and Mark T. Butt are paid consultants of BioMarin Pharmaceutical Inc. Acknowledgments The authors thank Becky Schweighardt, Anu Cherukuri, Brian Long, Julia Winston, Jeff Peng, Pascale Tiger, Michael Tomlinson, Lani Castaner, Camille Flournoy, Christine Sibigtroth, Melissa Carpentier, Molly Williams, Beth Taylor, Jill Zeller, Rhonda Rus, and Bob Boyd. rhTPP1 and vehicle were provided by BioMarin Pharmaceutical Inc. This work was funded by BioMarin Pharmaceutical Inc. References [1] T. Awano, M.L. Katz, D.P. O'Brien, et al., A frame shift mutation in canine TPP1 (the ortholog of human CLN2) in a juvenile Dachshund with neuronal ceroid lipofuscinosis, Mol. Genet. Metab. 89 (2006) 254–260. [2] V. Brinks, D. Weinbuch, M. Baker, Y. Dean, P. Stas, S. Kostense, B. Rup, B. Jiskoot, Preclinical models used for immunogenicity prediction of therapeutic proteins, Pharm. Res. 30 (2013) 1719–1728. [3] M. Bulat, M. Klarica, Recent insights into a new hydrodynamics of the cerebrospinal fluid, Brain Res. Rev. 65 (2011) 99–112. [4] M.T. Butt, Morphologic changes associated with intrathecal catheters for direct delivery to the central nervous system in preclinical studies, Toxicol. Pathol. 39 (2011) 213–219. [5] M. Chang, J.D. Cooper, D.E. Sleat, et al., Intraventricular enzyme replacement improves disease phenotypes in a mouse model of late infantile neuronal ceroid lipofuscinosis, Mol. Ther. 16 (2008) 649–656. [6] M. Chang, J.D. Cooper, B.L. Davidson, et al., CLN2, in: S.E. Mole, R.E. Williams, H.H. Goebel (Eds.), The Neuronal Ceroid Lipofuscinoses (Batten Disease), Oxford University Press, New York, 2011, pp. 80–109. [7] J.D. Cooper, A. Messer, A.K. Feng, et al., Apparent loss and hypertrophy of interneurons in a mouse model of neuronal ceroid lipofuscinosis: evidence for partial response to insulin-like growth factor-1 treatment, J. Neurosci. 19 (7) (1999) 2556–2567. [8] A.D. Dierenfeld, M.F. McEntee, C.A. Vogler, et al., Replacing the enzyme alpha-Liduronidase at birth ameliorates symptoms in the brain and periphery of dogs with mucopolysaccharidosis type I, Sci. Transl. Med. 1 (2) (2010) 60ra89. [9] A.H. El-Gharbawy, J. Mackey, S. DeArmey, et al., An individually, modified approach to desensitize infants and young children with Pompe disease, and significant reactions to alglucosidase alfa infusions, Mol. Genet. Metab. 104 (2011) 118–122. [10] B.R. Felice, T.L. Wright, R.B. Boyd, et al., Safety evaluation of chronic intrathecal administration of idursulfase-IT in cynomolgus monkeys, Toxicol. Pathol. 39 (2011) 879–892. [11] A.A. Golabek, P. Wujek, M. Walus, et al., Maturation of human tripeptidyl-peptidase I in vitro, J. Biol. Chem. 279 (2004) 31058–31067.

B.R. Vuillemenot et al. / Molecular Genetics and Metabolism 114 (2015) 281–293 [12] H.J. Gundersen, E.B. Jensen, The efficiency of systematic sampling in stereology and its prediction, J. Microsc. 147 (Pt 3) (1987) 229–263. [13] A. Jalanko, T. Braulke, Neuronal ceroid lipofuscinoses, Biochim. Biophys. Acta 1793 (2009) 697–709. [14] M.L. Katz, J.R. Coates, C.M. Sibigtroth, et al., Enzyme replacement therapy attenuates disease progression in a canine model of late-infantile neuronal ceroid lipofuscinosis (CLN2 disease), J. Neurosci. Res. (2014), http://dx.doi.org/10.1002/jnr.23423 (Epub ahead of print). [15] M.L. Katz, J.R. Coates, J.J. Cooper, D.P. O'Brien, M. Jeong, K. Narfstrom, Retinal pathology in a canine model of late infantile neuronal ceroid lipofuscinosis, Invest. Ophthalmol. Vis. Sci. 49 (2008) 2. [16] C. Kielar, L. Maddox, E. Bible, et al., Successive neuron loss in the thalamus and cortex in a mouse model of infantile neuronal ceroid lipofuscinosis, Neurobiol. Dis. 25 (1) (2007) 150–162. [17] M.B. Kosier, P. Minkler, Nursing management of patients with an implanted Ommaya reservoir, Clin. J. Oncol. Nurs. 3 (1999) 63–67. [18] M. Kousi, A.E. Lehesjoki, S.E. Mole, Update of the mutation spectrum and clinical correlations of over 360 mutations in eight genes that underlie the neuronal ceroid lipofuscinoses, Hum. Mutat. 33 (2012) 42–63. [19] K. Kramer, M. Smith, M.M. Souweidane, Safety profile of long-term intraventricular access devices in pediatric patients receiving radioimmunotherapy for central nervous system malignancies, Pediatr. Blood Cancer 61 (2014) 1590–1592. [20] L. Lin, P. Lobel, Production and characterization of recombinant human CLN2 protein for enzyme-replacement therapy in late infantile neuronal ceroid lipofuscinosis, Biochem. J. 357 (2001) 49–55. [21] S.E. Lipinski, M.J. Lipinski, A. Burnette, T.A. Platts-Mills, W.G. Wilson, Desensitization of an adult patient with Pompe disease and a history of anaphylaxis to alglucosidase alfa, Mol. Genet. Metab. 98 (3) (2009) 319–321. [22] M.D. Lorenz, J.R. Coates, M.R. Kent, Handbook of Veterinary Neurology, 5th edition Elsevier, St. Louis, 2011. 2–36. [23] H. Oyama, T. Fujisawa, T. Suzuki, B.M. Dunn, A. Wlodawer, K. Oda, Catalytic residues and substrate specificity of recombinant human tripeptidyl peptidase I (CLN2), J. Biochem. 138 (2005) 127–134. [24] G. Parenti, C. Pignata, P. Vajro, M. Salerno, New strategies for the treatment of lysosomal storage diseases, Int. J. Mol. Med. 5 (2013) 691–706. [25] M. Pollay, The function and structure of the cerebrospinal fluid outflow system, Cerebrospinal Fluid Res. 21 (7) (2010) 9. [26] D.N. Sanders, S. Kanazono, F.A. Wininger, et al., A reversal learning task detects cognitive deficits in a Dachshund model of late-infantile neuronal ceroid lipofuscinosis, Genes Brain Behav. 10 (2011) 798–804.

293

[27] N.R. Selden, A. Al-Uzri, S.L. Huhn, T.K. Kock, D.M. Sikora, M.D. Nguyen-Driver, D.J. Guillaume, J.L. Koh, S.H. Gultekin, J.C. Anderson, H. Vogel, T.L. Sutcliffe, Y. Jacobs, R. D. Steiner, Central nervous system stem cell transplantation for children with neuronal ceroid lipofuscinosis, J. Neurosurg. Pediatr. 11 (2013) 643–652. [28] D.E. Sleat, R.J. Donnelly, H. Lackland, et al., Association of mutations in a lysosomal protein with classical late-infantile neuronal ceroid lipofuscinosis, Science 277 (1997) 1802–1805. [29] N. Sundaresan, N.D. Suite, Optimal use of the Ommaya reservoir in clinical oncology, Oncology 3 (1989) 15–22. [30] C.H. Vite, P. Wang, R.T. Patel, R.M. Walton, et al., Biodistribution and pharmacody(rhIDU) in namics of recombinant human alpha-L-iduronidase mucopolysaccharidosis type I-affected cats following multiple intrathecal administrations, Mol. Genet. Metab. 103 (2011) 268–274. [31] B.R. Vuillemenot, D. Kennedy, R.P. Reed, et al., Recombinant human tripeptidyl peptidase-1 infusion to the monkey CNS: safety, pharmacokinetics, and distribution, Toxicol. Appl. Pharmacol. 277 (2014) 49–57. [32] B.R. Vuillemenot, M.L. Katz, J.R. Coates, et al., Intrathecal tripeptidyl-peptidase 1 reduces lysosomal storage in a canine model of late infantile neuronal ceroid lipofuscinosis, Mol. Genet. Metab. 104 (2011) 325–337. [33] R.E. Whiting, K. Narfström, G. Yao, et al., Enzyme replacement therapy delays pupillary light reflex deficits in a canine model of late infantile neuronal ceroid lipofuscinosis, Exp. Eye Res. 125C (2014) 164–172. [34] A. Wlodawer, S.R. Durell, M. Li, H. Oyama, K. Oda, B.M. Dunn, A model of tripeptidylpeptidase I (CLN2), a ubiquitous and highly conserved member of the sedolisin family of serine-carboxyl peptidases, BMC Struct. Biol. 3 (2003) 8. [35] S. Worgall, D. Sondhi, N.R. Hackett, B. Kosofsky, M.V. Kekatpure, N. Neyzi, J.P. Dyke, D. Ballon, L. Heier, B.M. Greenwald, P. Christos, M. Mazumdar, M.M. Souweidane, M. G. Kaplitt, R.G. Crystal, Treatment of late infantile neuronal carotid lipofuscinosis by CNS administration of a serotype 2 adeno-associated virus expressing CLN2 cDNA, Hum. Gene Ther. 19 (2008) 463–474. [36] K. Wyatt, W. Henley, L. Anderson, et al., The effectiveness and cost-effectiveness of enzyme and substrate replacement therapies: a longitudinal cohort study of people with lysosomal storage disorders, Health Technol. Assess. 16 (2012) 1–543. [37] M.W. Zappaterra, M.K. Lehtinen, The cerebrospinal fluid: regulator of neurogenesis, behavior, and beyond, Cell. Mol. Life Sci. 69 (2012) 2863–2878.