This document discusses the challenges of conducting preclinical safety studies evaluating biologics administered directly to the central nervous system (CNS) via routes such as intrathecal or intracerebroventricular administration. Key points include the need for control groups to differentiate effects caused by the test article versus the delivery device, customized tissue sampling schemes to thoroughly assess local and distant effects in the brain and spinal cord, and characterization of pharmacokinetics, exposure levels, and potential immunogenicity issues which can impact interpretation of safety results. The document provides recommendations for designing preclinical CNS administration studies to help enable translation of results for first-in-human clinical trials.

1. CONTEMPORARY REVIEW
Safety Evaluation of CNS Administered Biologics—
Study Design, Data Interpretation, and Translation to
the Clinic
Brian R. Vuillemenot,*,1
Sven Korte,†
Teresa L. Wright,‡
Eric L. Adams,§
Robert B. Boyd,§
and Mark T. Butt¶
*Genentech, Inc, South San Francisco, California; †
Covance Laboratories GmbH, Mu¨ nster, Germany;
‡
Dimension Therapeutics, Cambridge, Massachusetts; §
Northern Biomedical Research, Muskegon, Michigan;
and ¶
Tox Path Specialists, Frederick, Maryland
1
To whom correspondence should be addressed at Genentech, Inc., 1 DNA Way, South San Francisco, California 94080. Fax: 650-866-2621.
E-mail: vuillemb@gene.com.
ABSTRACT
Many central nervous system (CNS) diseases are inadequately treated by systemically administered therapies due to the
blood brain barrier (BBB), which prevents achieving adequate drug concentrations at sites of action. Due to the increasing
prevalence of neurodegenerative diseases and the inability of most systemically administered therapies to cross the BBB,
direct CNS delivery will likely play an increasing role in treatment. Administration of large molecules, cells, viral vectors,
oligonucleotides, and other novel therapies directly to the CNS via the subarachnoid space, ventricular system, or
parenchyma overcomes this obstacle. Clinical experience with direct CNS administration of small molecule therapies
suggests that this approach may be efficacious for the treatment of neurodegenerative disorders using biological therapies.
Risks of administration into the brain tissue or cerebrospinal fluid include local damage from implantation of the delivery
system and/or administration of the therapeutic and reactions affecting the CNS. Preclinical safety studies on CNS
administered compounds must differentiate between the effects of the test article, the delivery device, and/or the vehicle,
and assess exacerbations of reactions due to combinations of effects. Animal models characterized for safety assessment of
CNS administered therapeutics have enabled human trials, but interpretation can be challenging. This manuscript outlines
the challenges of preclinical intrathecal/intracerebroventricular/intraparenchymal studies, evaluation of results,
considerations for special endpoints, and translation of preclinical findings to enable first-in-human trials.
Recommendations will be made based on the authors’ collective experience with conducting these studies to enable clinical
development of CNS-administered biologics.
Key words: CNS administration; intrathecal; intracerebroventricular; neurodegeneration; enzyme replacement therapy.
Neurodegenerative diseases represent a major health burden
and inadequately met medical need. This burden includes
Alzheimer’s disease, Parkinson’s disease, amyotrophic lateral
sclerosis, Huntingdon’s disease, lysosomal storage diseases
(LSDs), and other diseases. Systemically administered biologics
have not yet been effective in treating these diseases, largely
due to the challenges in achieving adequate concentrations at
key sites of action in the central nervous system (CNS). Direct
CNS administration circumvents the barriers that keep large
molecules out of the CNS and introduces potential therapies
VC The Author 2016. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved.
For Permissions, please e-mail: journals.permissions@oup.com
3
TOXICOLOGICAL SCIENCES, 152(1), 2016, 3–9
doi: 10.1093/toxsci/kfw072
Contemporary Review
atLibraryonJune27,2016http://toxsci.oxfordjournals.org/Downloadedfrom

2. close to sites of action, and is likely to play an increasingly large
role in addressing these unmet medical needs. Preclinical devel-
opment of CNS-administered therapeutics faces a number of
challenges not typically encountered with systemically admin-
istered therapeutics. The objective of this article is to review the
main challenges of preclinical safety assessment of CNS admin-
istered molecules. Recommendations for the translation of
these studies to enable first-in-human trials will also be made.
Current experience with protein-based therapeutics ad-
ministered directly to the CNS is limited. CNS-administered
biologics have been evaluated preclinically and/or clinically
to treat pain, cancer, neurodegenerative diseases, and lyso-
somal storage diseases (LSDs)]. A summary is presented in
Table 1.
The only currently marketed biologic developed specifically
for direct CNS administration is Ziconitide, a cone snail derived
peptide to treat pain (Williams et al., 2008). Baclofen, a marketed
small molecule therapy approved for treatment of muscle
spasms, was also developed for intrathecal (IT) administration
(Richard and Menei, 2007). Several monoclonal antibodies have
been administered to the cerebrospinal fluid (CSF) off label with
evidence of efficacy in different cancers. In contrast, attempts
to develop therapies for neurodegenerative diseases using CNS
administration of growth and neurotrophic factors have so far
been unsuccessful. LSDs may be caused by genetic deficiency of
lysosomal enzymes, with over two-thirds involving CNS disease
(Schultz et al., 2011). Systemically administered enzyme replace-
ment therapies (ERTs) have been successful in treating nonCNS
symptoms of these diseases, but not the neurological compo-
nents (Hollak and Wijburg, 2014). Several of the approved ERTs
have been used off label for direct CNS administration (Mu~noz-
Rojas et al., 2008, 2010). Several clinical trials are in progress for
CNS-administered ERTs to treat neurodegenerative LSDs (Katz
et al., 2014; Muenzer et al., 2016). Gene therapy strategies using
IT administration have been described in Beutler et al. (2005)
and Hirai et al. (2014). Direct CNS delivery of antisense oligonu-
cleotides (ASOs), used to regulate target mRNA, is also in devel-
opment (Miller et al., 2013). Intracerebroventricular (ICV) and IT
administration of ASOs has resulted in neuronal uptake in mon-
eys and dogs.
BREACHING THE BLOOD BRAIN BARRIER
The blood brain barrier (BBB) represents the primary obstacle to
achieving CNS distribution of large molecule therapeutics ad-
ministered systemically. The BBB consists of tight junctions be-
tween capillary endothelial cells that provide a physical barrier
to the entry of large molecules (Bauer et al., 2014; Tajes et al.,
2014). Numerous transporters in the BBB tightly regulate move-
ment of molecules from the bloodstream to the CNS (Fricker
and Miller, 2004). Through the physical barrier provided by the
tight junctions and the transport barrier, virtually all large mol-
ecules are excluded from the CNS when administered into the
systemic circulation.
Physically breaching the BBB by administering a therapy into
the CSF is one means of achieving CNS distribution. CSF plays
multiple roles including providing brain buoyancy, protecting
the brain from sudden impacts, regulation of solute concentra-
tions and pressure, and elimination of wastes (Sakka et al.,
2011). CSF is secreted continuously at a rate of approximately
0.3 ml/min in adults primarily by choroid plexus ependymal
cells in the ventricles (Oreskovic and Klarica, 2010). It flows out
through the ventricular system into the subarachnoid space,
and around the external surfaces of the brain and spinal cord
(Greitz, 1993). CSF circulates through the brain parenchyma
along perivascular spaces surrounding arteries via the glym-
phatic pathway; this brain-wide perivascular network facilitates
solute exchange between the CSF and interstitial fluid (Iliff et al.,
2012; Yang et al., 2013). The continuous movement of CSF can
distribute therapeutics in the CNS after IT or ICV administra-
tion. The total CSF volume in an adult human is approximately
150 ml, with a total of approximately 500 ml of CSF secreted per
day. Drainage into the systemic circulation via the arachnoid
granulations and lymphatics (Bulat and Klarica, 2011) allows for
maintenance of a stable volume.
TABLE 1. Representative Biological Therapies that Have Been Used for Direct CNS Administration
Indication Molecule Route Development
Phase/Results
References
Pain Ziconitide (Prialt) IT-L Approved Williams et al. (2008)
Breast cancer brain metastasis Trastuzumab (Herceptin) IT-L Off label/P1/2 Oliveira et al. (2011)
Leukemia Rituximab (Rituxan) IT-L P1/2 Jaime-Perez et al. (2009)
Non-Hodgkins lymphoma ICV Off label/P1/2 Rubenstein et al. (2007)
Multiple sclerosis IT-L P2 Bonnan et al. (2014)
Alzheimer’s disease Nerve growth factor (NGF) ICV Off label Eriksdotter Jonhagen et al. (1998)
Parkinson’s disease Glial derived neurotrophic factor
(GDNF)
ICV, IP Clinical trials
stopped after P2
Nutt et al. (2003), Patel et al. (2005)
Amyotrophic lateral sclerosis Brain derived neurotrophic factor
(BDNF)
IT-L No efficacy in P3 Beck et al. (2005)
Vascular endothelial growth factor
(VEGF)
ICV P1/2 Storkebaum et al. (2005)
Mucopolysaccharidosis I Laronidase (Aldurazyme) IT-L Off label Munoz-Rojas et al. (2008)
Mucopolysaccharidosis II Idursulfase-IT IT-L P2/3 Felice et al. (2011), Muenzer et al. (2016)
Mucopolysaccharidosis IIIA Heparan-N-sulfatase IT-L P1/2 Pfeifer et al. (2012)
Mucopolysaccharidosis IIIB Alpha-N-acetylglucosaminidase ICV Preclinical Kan et al. (2014)
Mucopolysaccharidosis VI Galsulfase (Naglazyme) IT-L Off label Mu~noz-Rojas et al. (2010)
Metachromatic leukodystrophy Arylsulfatase A IT-L P1/2 Patil and Maegawa (2013)
CLN2 disease (a form of Batten Disease) Tripeptidyl peptidase I ICV P1/2 Katz et al. (2014), Vuillemenot et al. (2015)
ICV, intracerebroventricular; IP, intrapanenchymal; IT-L, intrathecal lumbar.
4 | TOXICOLOGICAL SCIENCES, 2016, Vol. 152, No. 1
atLibraryonJune27,2016http://toxsci.oxfordjournals.org/Downloadedfrom

3. ROUTES OF DIRECT ADMINISTRATION
TO THE CNS
There are several routes of direct CNS administration, either via
the CSF or directly to tissue. IT administration introduces the
therapeutic into the CSF in the subarachnoid space, between
the arachnoid and pia mater. This is most frequently accom-
plished via the implantation of a catheter into the IT lumbar
(IT-L) space. IT-L administration has a potential disadvantage
that the administered biologic needs to travel a longer distance
to reach the brain than via other direct routes to the CNS.
However, IT-L administration studies in multiple nonclinical
species have demonstrated adequate brain distribution to
achieve pharmacological activity of the therapeutic when ad-
ministered in a large dose volume or with a subsequent catheter
flush with buffer (Dickson et al., 2007; Felice et al., 2011; Xu et al.,
2011). With test articles more dense than CSF, body posture dur-
ing IT-L dose administration (supine vs upright) may lead to dif-
ferences in brain distribution, particularly in semi-bipedal
species, such as monkeys. Delivering test articles to the CSF in
closer proximity to the brain can also be achieved by adminis-
tering into the cisterna magna within the cerebromedullary cis-
tern (IT-cisternal, IT-C). IT-C administration is not often used
clinically because of safety risks but is sometimes used in pre-
clinical studies. Catheters and dosing ports have been im-
planted to the lumbar spine to enable repeat IT-L infusion
administration in multiple animal models. IT-C administration
is typically conducted as a bolus injection. In rodents, IT cathe-
ters terminating in the lumbar space may be introduced
through the cerebromedullary cistern.
ICV administration introduces the therapeutic into the lat-
eral ventricle, in close proximity to the primary CSF production.
The outward flow of CSF from the ventricles may result in wider
CNS distribution than that achieved by IT-L administration
(Vuillemenot et al., 2014). ICV administration is usually accom-
plished by infusion via an implanted catheter and dosing port.
Intraparenchymal (IP) administration introduces the therapy di-
rectly to the brain tissues. Convection enhanced delivery can be
used to increase the distribution achieved with IP administra-
tion through increased pressures (Barua et al., 2014). As com-
pared with IP administration, introducing the therapeutic to the
CSF by IT or ICV administration may produce a broader distribu-
tion pattern.
Both bolus and continuous administration may be utilized
with these different routes. Infusions lasting from several hours
to continuous may be preferable to bolus injection to achieve
the required concentration of therapeutic in the brain safely. If
the rate of infusion is less than the normal turnover of CSF,
then no appreciable changes in CSF volume will result, and
safety concerns due to excessive intracranial pressure can be
minimized. When selecting an appropriate clinical route, it is
important to consider the optimal CNS distribution in the in-
tended patient population. Whenever possible, the intended
clinical route of administration should be used for any pivotal
preclinical safety studies, although alternate routes are often
employed for preliminary studies, and occasionally for pivotal
toxicology studies.
CONSIDERATIONS FOR DESIGN OF
NONCLINICAL CNS ADMINISTRATION
STUDIES
Nonclinical studies involving direct CNS administration are fun-
damentally different than studies with more conventional routes
and several important points should be considered. Adverse ef-
fects in CNS delivery studies are generally not due to the biologic
per se, but changes due to the delivery device alone or with an ad-
ditive effect related to the therapeutic may occur. Inclusion of ve-
hicle and/or device-only control groups is critical to sort out the
causes of any findings. Due to the often limited group size of
these studies, a thorough review of historical control data may be
the only accurate means of assessing study findings.
Recently, the Society of Toxicologic Pathologists published
updated recommendations for sampling the CNS for general
toxicity studies (Bolon et al., 2012). Although possibly sufficient
for general toxicity studies where there is no reason to suspect
an effect on the nervous system, these schemes are not ade-
quate for a study involving direct CNS delivery. The trimming/
embedding/staining scheme for a direct CNS delivery study
should be customized to allow for a thorough assessment of the
local effects on the various structures/cell types of the brain and
spinal cord that may be due to the placement/presence of the
drug and/or the delivery device, as well as more distant effects
that may be due to the device, or distribution of the drug.
Assessment of pharmacokinetics (PK), exposure, and immu-
nogenicity are important for evaluating the dose response of
any pharmacology or toxicity of a CNS administered biologic.
A significant fraction of the test article that is not distributed
into the CNS will enter the systemic circulation within a few
hours via the arachnoid granulations and/or lymphatics. This
occurs through natural CSF turnover and via outflow caused by
increased pressure (Bulat and Klarica, 2011). CSF also drains to
cervical lymph nodes via the glymphatic system, a dural lym-
phatic network (Aspelund et al., 2015). Systemic exposure to
CNS-administered biologics can lead to an immune response.
Anti-drug antibody (ADA) formation may occur when adminis-
tering a human protein to animals. This response may result in
decreased exposure and/or activity or hypersensitivity
reactions. Administering a biologic to the CNS via slow infusion
may reduce the maximal systemic concentrations and reduce
the likelihood of immunogenicity. In addition, pretreatment
with antihistamines has been efficacious in reducing the
incidence of hypersensitivity (Kim et al., 2008; Vuillemenot et al.,
2011).
Sufficient sampling of both plasma and CSF should be in-
cluded to characterize standard PK parameters, while ADAs
may be monitored in serum and/or CSF. To enable collection of
serial samples, a dual catheter/access port setup may be useful,
with one catheter used for dose administration and the other
catheter for CSF collection (Figure 1). CSF samples collected
from the ventricles versus the lumbar region have been shown
to differ in composition and cellularity (Provencio et al, 2005;
Rubalcava and Sotelo, 1995; Torres-Corzo et al., 2009), which
should be considered when interpreting CSF data and compar-
ing between studies. As a backup to lumbar sampling, direct cis-
terna magna sampling may be used to obtain CSF for analysis,
although this is a technically challenging procedure in nonclini-
cal species due to the small volume of this space and close prox-
imity to the spinal cord.
The CNS distribution should be also evaluated to guide the
clinical dose regimen. This can be accomplished through dedi-
cated biodistribution studies or by sampling CNS tissues in
pharmacology and/or toxicology studies to analyze for drug
concentrations. Understanding the relationship between CSF/
systemic PK and CNS exposure is important when designing the
clinical dosing regimen. This is best assessed in animal models,
as serial CSF sampling is generally not possible in clinical
studies.
VUILLEMENOT ET AL. | 5
atLibraryonJune27,2016http://toxsci.oxfordjournals.org/Downloadedfrom

4. A very important final consideration is the selection of ap-
propriate in vivo models to assess the safety, PK, and pharmaco-
logical activity. The species selected should express a similar
target as that in human patients. When administering a human
protein to animals, the molecule should display activity against
the homologous target in the animals. Differences in the activ-
ity or receptor density in the species tested should be consid-
ered when interpreting the results. When developing therapies
for neurodegenerative diseases where the CNS is undergoing
changes that may affect the safety, distribution, and/or activity
of the molecule, it may be informative to conduct safety assess-
ments in an animal disease model undergoing similar changes.
Non-affected controls of the same species can be assessed
alongside to get an idea of the toxicity on a non-diseased CNS
background.
NONCLINICAL SPECIES SELECTION/
INTERPRETATION
When interpreting in vivo data, the relevance of the CNS in the
nonclinical species to the human should be carefully consid-
ered. Biologics have been administered directly to the CNS in all
of the common lab animal species, including mice, rats, dogs,
monkeys, sheep, and pigs. CNS differences between these spe-
cies and human must be considered when designing and inter-
preting direct CNS administration studies. A summary of
different parameters affecting the pharmacology and safety of
CNS-administered molecules is presented in Table 2.
Differences in brain mass are an important limitation of ani-
mal models. The nonclinical species have a significantly
smaller brain than humans. Smaller brains have a higher sur-
face area to volume ratio, potentially increasing relative uptake
from CSF. In addition, the smaller the brain, the less distance a
therapeutic is required to travel to reach all sites of activity
within the CNS. Smaller brains may also display a more severe
reaction to CNS delivery devices, and there is less area in which
to implant these devices. Achieving precise targets within the
smaller brains of animals is more challenging because of the
smaller relative size of these targets. For example, the lateral
ventricle has a mean volume of approximately 25 ml in hu-
mans, but only 0.25 ml in monkeys (Akdogan et al., 2010). Total
CSF volume is also less in the nonclinical species than human.
However, the rate of CSF turnover is similar between dog, mon-
key, and human, but several-fold higher in rodents. Differences
in CSF volume and turnover must be taken into account when
interpreting PK data and scaling to human patients.
MORPHOLOGIC ASSESSMENT OF CNS TISSUES
Morphologic assessment of the nervous system in a study uti-
lizing a direct delivery device requires particular scrutiny in the
areas traversed by the device and the site of deposition of the
therapeutic. Synergy may be observed between the test article
and the delivery system. Complications of surgery or the device
are common, and even in vehicle/device only control animals
there may be numerous microscopic changes (Butt, 2011a).
These changes must be differentiated from what is caused by
the test article.
When evaluating nervous system tissues from a direct CNS
deliver study, the timing of the necropsy, tissue processing, and
staining must be carefully considered. Timing of tissue collec-
tion must capture the full spectrum of potential changes occur-
ring. It is important to assess effects acutely, as early toxicities
may be completely resolved at later times (Switzer, 2011).
Single-dose pilot studies should include multiple tissue collec-
tion times. Morphological assessment is complicated because
there must be sufficient time between device implantation and
dosing to allow for the changes from surgery to resolve. It is
common to observe neuronal necrosis at the site of catheter in-
sertion into the brain, but this should not be confused with an
effect of the tested therapeutic.
Design of the CNS morphological evaluation should be based
on the study objectives, delivery methods, and any knowledge
of the effects of the test article. In addition to brain and spinal
cord, nerves and ganglia may warrant evaluation. Changes to
one part of the nervous system may manifest as changes in
other parts, and all may need to be examined. The brain and
spinal cord should be sectioned to allow for an evaluation that
provides sufficient confidence that any effects on the CNS have
FIG. 1. Dual port catheter system in the cynomolgus monkey. In this setup, the
animal was surgically implanted with access ports and catheters terminating in
the lumbar spine and the cisterna magna. The lumbar catheter/access port,
which has a dosing needle inserted in this picture, was used for dose adminis-
tration, while the cisternal device enabled repeat CSF sampling for toxicokinetic
and other analyses.
TABLE 2. CNS Parameters of Nonclinical Species Compared with Human
Species Mouse Rat Dog Monkey Human
Brain mass (fold human) 0.4 g (0.0004) 2 g (0.002) 72 g (0.072) 100 g (0.1) 1000–1500 g
CSF volume (fold human) 0.04 ml (0.0004) 0.15 ml (0.0015) 12 ml (0.12) 15 ml (0.15) 100–150 ml
CSF turnovers/day (fold human) 12.5 Â (2.5) 28.8 Â (5.76) 5.75 Â (1.15) 4 Â (0.8) 5Â
Posture Quadrupedal Quadrupedal Quadrupedal Semi-bipedal Bipedal
From Pardridge (1991).
6 | TOXICOLOGICAL SCIENCES, 2016, Vol. 152, No. 1
atLibraryonJune27,2016http://toxsci.oxfordjournals.org/Downloadedfrom

5. been determined. For the brain, that involves at least 3 trans-
verse or sagittal sections through the site of administration and
the region traversed by the device, with sufficient additional
sections to characterize any distant effects. It is useful to pro-
duce full transverse sections including both hemispheres in
studies where a device traverses the brain. In general, at least
8–10 full transverse sections are required to evaluate the main
brain regions in any species. For IT studies, multiple sections
near the catheter tip will allow for a complete evaluation of
changes. Inspection of the spinal cord for IT granuloma (Allen
et al., 2006; Butt, 2011a) should be performed. Evaluating trans-
verse and oblique sections of spinal cord increase the sensitivity
for detecting changes. Typically, the spinal cord should be sec-
tioned to include cervical, thoracic, catheter tip, and spinal
cord/cauda equina caudal to the catheter tip regions.
The objectives of the histological evaluation should be taken
into account when determining the choices of tissue preserva-
tion and staining reagents. Intravascular perfusion is
recommended to minimize artifactual changes that often com-
plicate microscopic interpretation (Garman, 2011). Peripheral
nerves are best immediately fixed with a fixative containing
glutaraldehyde to preserve myelin. Tissue sections may be em-
bedded in paraffin or resin or frozen. Although paraffin allows
for more detail than frozen, frozen sectioning may allow for im-
proved immunohistochemistry and use of specialized stains
(Switzer, 2000). Resin embedding is useful for optimal cross sec-
tioning to preserve myelin.
For all studies, paraffin embedded or frozen sections should
be stained with hematoxylin and eosin (H&E) for general evalu-
ation. In the brain and spinal cord, immunohistochemical
stains to reveal astrocyte and microglial reactions can demon-
strate glial cell changes not detectable by H&E. In acute studies,
a stain that increases the sensitivity of detection of neuronal
necrosis should be used, such as Fluoro-Jade (Schmued et al.,
2005) or Cupric silver (Switzer, 2000). Other potentially useful
stains include non-selective silver stains for axons, Luxol fast
blue for myelin, neurofilament protein immunohistochemistry,
and stains for specific neuronal populations. The combination
of a longitudinal section in paraffin (H&E stain) and a cross sec-
tion that has been osmicated, resin embedded, and stained
with toluidine blue provides assessment of axonal degenera-
tion, regeneration, and myelin alterations (Butt, 2011b).
It is virtually impossible to prevent local damage when im-
planting a CNS catheter. Inflammation, haemorrhage, and glio-
sis are frequently encountered adjacent to the delivery device.
Accumulation of fluid around the catheter track may be due to
edema, and/or excess test article. Microscopic changes due to
the delivery device are seldom associated with clinical signs
and are not necessarily adverse, as they may be an unavoidable
consequence of the mode of administration in the animal
model and irrelevant to the intended clinical population. It can
be challenging to differentiate the relative contributions of the
test article and each component of the delivery system, even in
properly controlled studies.
USING NONCLINICAL STUDIES TO ENABLE
FIRST IN HUMAN TRIALS
The intended clinical regimen should guide the nonclinical pro-
gram, with pivotal nonclinical studies should use the same
route as that in the first in human trial. Distribution of the bio-
logic to the target tissues/cells at pharmacologically active con-
centrations should be demonstrated if possible. In addition,
disposition of drug into the systemic compartment, and the re-
lationship between CNS and systemic PK, should be character-
ized. Toxicities revealed in the nonclinical studies should be
monitored for in the clinic. For example, the presence of CNS in-
flammation and elevated CSF white blood cells in nonclinical
studies may lead to monitoring CSF cell counts clinically. The
nonclinical studies should provide information about the ef-
fects of the administration procedure, delivery device, vehicle,
and test article, alone and in combination. It is expected that
there will be a local reaction to CNS administered biologics, so
consideration of the risk/benefit profile in the context of the pa-
tient population is important.
A safe starting dose can be determined using an appropriate
safety factor and the pivotal nonclinical no observed adverse ef-
fect level (NOAEL). For CNS administered biologics, brain mass
or CSF volume can be used to normalize doses between species.
If CSF volume is used, differences in the rates of turnover be-
tween the different species should be considered (Table 2). An
example of clinical safety factors determined based on a mon-
key NOAEL scaled for differences in brain mass is illustrated in
Table 3 (Felice et al., 2011; USFDA, 2005).
Additional nonclinical studies may be needed to support the
use of a delivery device in combination with the therapeutic.
Utilizing the same or similar device in the nonclinical and clini-
cal studies is desirable to demonstrate the safety of the drug-
device combination. For developmental and reproductive
toxicology studies, if warranted, the IV route should be used
(Skov et al., 2007).
FUTURE DIRECTIONS
CNS administration of biologics is likely to play an increasing
role in treating neurodegenerative disease in the future. Careful
consideration of nonclinical program design will insure the suc-
cess of these efforts. Nonclinical studies to enable clinical trials
of CNS administered drugs must consider the clinical regimen,
assessment of exposure in CSF, plasma, and/or CNS tissue, CNS
effects, and the relevance of animal models to human patients.
Nonclinical programs should be designed on a case-by-case ba-
sis, carefully considering the clinical plan and the risk/benefit
profile in the intended patient population. Interpretation of di-
rect CNS administration studies is complicated by the histologi-
cal changes attributable to the route of administration and
presence of delivery devices in the CNS. Therefore, inclusion of
applicable control groups is essential. CNS sampling for histo-
pathological evaluation must be extensive, and may involve
TABLE 3. Calculation of Clinical Safety Factors based on NOAEL from Pivotal Monkey Study
Human (Pediatric; Brain 5 1 kg) Monkey (Brain 5 0.1 kg) Safety margin
Clinical dose, mg mg/kg brain weight Nonclinical NOAEL, mg mg/kg brain weight
10 10 100 1000 100-fold
100 100 100 1000 10-fold
VUILLEMENOT ET AL. | 7
atLibraryonJune27,2016http://toxsci.oxfordjournals.org/Downloadedfrom

6. serial sectioning of the entire brain as well as thorough sam-
pling of spinal cord, dorsal nerve roots, and ganglia. In addition,
use of multiple stains to illuminate specific neuronal changes is
recommended. Prior to entry into first in human trials, a non-
clinical program should describe the safety findings of the bio-
logic in conjunction with the delivery device and vehicle,
support the likely efficacy in the patient population, character-
ize the PK and distribution, and provide rationale for inclusion
of clinical endpoints of safety and efficacy. Therefore, a strong
nonclinical data package is required to support these challeng-
ing but increasingly worthwhile clinical trials.
REFERENCES
Akdogan, I., Kiroglu, Y., Onur, S., and Karabuluti, N. (2010). The
volume fraction of brain ventricles to total brain volume: A
computed tomography stereological study. Folia Morphol. 69,
193–200.
Allen, J., Horais, K., Tozier, N., Wegner, K., Corbeil, J., Mattrey, R.,
Rosi, S., and Yaksh, T. (2006). Time course and role of mor-
phine dose and concentration in intrathecal granuloma for-
mation in dogs. Anesthesiology 105, 581–590.
Aspelund, A., Antila, S., Proulx, S. T., Karlsen, T. V., Karaman, S.,
Detmar, M., Wiig, H., and Alitalo, K. (2015). A dural lymphatic
vascular system that drains brain interstitial fluid and mac-
romolecules. J. Exp. Med. 29, 991–999.
Barua, N. U., Gill, S. S., and Love, S. (2014). Convection-enhanced
drug delivery to the brain: Therapeutic potential and neuro-
pathological considerations. Brain Pathol. 24, 117–127.
Bauer, H. C., Krizbai, I. A., Bauer, H., and Traweger, A. (2014).
“You Shall Not Pass”-tight junctions of the blood brain bar-
rier. Front. Neurosci. 8, 392.
Beck, M., Flachenecker, P., Magnus, T., Giess, R., Reiners, K.,
Toyka, K. V., and Naumann, M. (2005). Autonomic dysfunc-
tion in ALS: A preliminary study on the effects of intrathecal
BDNF. Amyotroph. Lateral Scler. Other Motor Neuron. Disord. 6,
100–103.
Beutler, A. S., Banck, M. S., Walsh, C. E., and Milligan, E. D. (2005).
Intrathecal gene transfer by adeno-associated virus for pain.
Curr. Opin. Mol. Ther. 7, 431–439.
Bolon, B., Garman, R., Pardo, I., Jensen, K., Butt, M., et al. (2012).
STP position paper: Recommended practices for sampling
and processing the nervous system (brain, spinal cord, nerve
and eye) during nonclinical general toxicity studies. Toxicol.
Pathol 41, 1028–1048.
Bonnan, M., Ferrari, S., Bertandeau, E., Demasles, S., Krim, E.,
Miquel, M., and Barroso, B. (2014). Intrathecal rituximab ther-
apy in multiple sclerosis: Review of evidence supporting the
need for future trials. Curr. Drug Tar. 15, 1205–1214.
Bulat, M., and Klarica, M. (2011). Recent insights into a new hy-
drodynamics of the cerebrospinal fluid. Brain Res. Rev. 1,
99–112.
Butt, M. T. (2011a). Morphologic changes associated with intra-
thecal catheters for direct delivery to the central nervous
system in preclinical studies. Toxicol. Pathol. 39, 213–219.
Butt, M. T. (2011b). Evaluation of the adult nervous system. In
Fundamental Neuropathology for Pathologists and Toxicologists:
Principles and Techniques (B Bolon and M Butt, Eds.), pp.
321–338. Wiley and Sons, Hoboken, NJ.
Dickson, P., McEntee, M., Vogler, C., Le, S., Levy, B., Peinovich, M.,
Hanson, S., Passage, M., and Kakkis, E. (2007). Intrathecal en-
zyme replacement therapy: Successful treatment of brain
disease via the cerebrospinal fluid. Mol. Genet. Metab. 91,
61–68.
Eriksdotter Jo¨nhagen, M., Nordberg, A., Amberla, K., B€ackman, L.,
Ebendal, T., Meyerson, B., Olson, L., Seiger, Shigeta, M.,
Theodorsson, E., et al. (1998). Intracerebroventricular infu-
sion of nerve growth factor in three patients with
Alzheimer’s disease. Dement. Geriatr. Cogn. Disord. 9, 246–257.
Felice, B. R., Wright, T. L., Boyd, R. B., Butt, M. T., Pfeifer, R. W.,
Pan, J., Ruiz, J. A., Heartlein, M. W., and Calias, P. (2011).
Safety evaluation of chronic intrathecal administration of
idursulfase-IT in cynomolgus monkeys. Toxicol. Pathol. 39,
879–892.
Fricker, G., and Miller, D. S. (2004). Modulation of drug transpor-
ters at the blood-brain barrier. Pharmacology 70, 169–176.
Garman, R. (2011). Common histological artifacts in nervous sys-
tem tissues. In Fundamental Neuropathology for Pathologists and
Toxicologists Principles and Techniques. (B Bolon and M Butt,
Eds.), pp 191–201. J Wiley and Sons, Hoboken, NJ.
Greitz, D. (1993). Cerebrospinal fluid circulation and associated
intracranial dynamics. A radiologic investigation using MR
imaging and radionuclide cisternography. Acta Radiol. Suppl.
386, 1–23.
Hirai, T., Enomoto, M., Kaburagi, H., Sotome, S., Yoshida-Tanaka,
K., Ukegawa, M., Kuwahara, H., Yamamoto, M., Tajiri, M.,
Miyata, H., et al. (2014). Intrathecal AAV serotype 9-mediated
delivery of shRNA against TRPV1 attenuates thermal hyper-
algesia in a mouse model of peripheral nerve injury. Mol.
Ther. 22, 409–419.
Hollak, C. E., and Wijburg, F. A. (2014). Treatment of lysosomal
storage disorders: Successes and challenges. J. Inherit. Metab.
Dis. 37, 587–598.
Iliff, J. J., Wang, M., Liao, Y., Plogg, B. A., Peng, W., Gundersen, G.
A., Benveniste, H., Vates, G. E., Deane, R., Goldman, S. A., et al.
(2012). A paravascular pathway facilitates CSF flow through
the brain parenchyma and the clearance of interstitial sol-
utes, including amyloid b. Sci. Transl. Med. 4, 147ra111.
Jaime-Pe´rez, J. C., Rodrıguez-Romo, L. N., Gonzalez-Llano, O.,
Chapa-Rodrıguez, A., and Gomez-Almaguer, D. (2009).
Effectiveness of intrathecal rituximab in patients with acute
lymphoblastic leukaemia relapsed to the CNS and resistant
to conventional therapy. Br. J. Haematol. 144, 794–795.
Kan, S. H., Aoyagi-Scharber, M., Le, S. Q., Vincelette, J., Ohmi, K.,
Bullens, S., Wendt, D. J., Christianson, T. M., Tiger, P. M.,
Brown, J. R., et al. (2014). Delivery of an enzyme-IGFII fusion
protein to the mouse brain is therapeutic for mucopolysac-
charidosis type IIIB. Proc. Natl. Acad. Sci. U S A 111,
4870–14875.
Katz, M. L., Coates, J. R., Sibigtroth, C. M., Taylor, J. D., Carpentier,
M., Young, W. M., Wininger, F. A., Kennedy, D., Vuillemenot,
B. R., and O’Neill, C. A. (2014). Enzyme replacement therapy
attenuates disease progression in a canine model of late-
infantile neuronal ceroid lipofuscinosis (CLN2 disease).
J. Neurosci. Res. 92, 1591–1598.
Kim, K. H., Decker, C., and Burton, B. K. (2008). Successful man-
agement of difficult infusion-associated reactions in a young
patient with mucopolysaccharidosis type VI receiving re-
combinant human arylsulfatase B (galsulfase [Naglazyme).
Pediatrics 121, e714–e717.
Miller, T. M., Pestronk, A., David, W., Rothstein, J., Simpson, E.,
Appel, S. H., Andres, P. L., Mahoney, K., Allred, P., Alexander,
K., et al. (2013). An antisense oligonucleotide against SOD1
delivered intrathecally for patients with SOD1 familial amyo-
trophic lateral sclerosis: A phase 1, randomised, first-in-man
study. Lancet Neurol. 12, 435–442.
Muenzer, J., Hendriksz, C. J., Fan, Z., Vijayaraghavan, S., Perry, V.,
Santra, S., Solanki, G. A., Mascelli, M. A., Pan, L., Wang, N.,
8 | TOXICOLOGICAL SCIENCES, 2016, Vol. 152, No. 1
atLibraryonJune27,2016http://toxsci.oxfordjournals.org/Downloadedfrom

7. et al. (2016). A phase I/II study of intrathecal idursulfase-IT in
children with severe mucopolysaccharidosis II. Genet. Med.
18, 73–81.
Munoz-Rojas, M. V., Vieira, T., Costa, R., Fagondes, S., John, A.,
Jardim, L. B., Vedolin, L. M., Raymundo, M., Dickson, P. I.,
Kakkis, E., et al. (2008). Intrathecal enzyme replacement ther-
apy in a patient with mucopolysaccharidosis type I and
symptomatic spinal cord compression. Am. J. Med. Genet.
146A, 2538–2544.
Mu~noz-Rojas, M. V., Horovitz, D. D., Jardim, L. B., Raymundo, M.,
Llerena, J. C., Jr, de Magalh~aes Tde, S., Vieira, T. A., Costa, R.,
Kakkis, E., and Giugliani, R. (2010). Intrathecal administration
of recombinant human N-acetylgalactosamine 4-sulfatase to
a MPS VI patient with pachymeningitis cervicalis. Mol. Genet.
Metab. 99, 346–350.
Nutt, J. G., Burchiel, K. J., Comella, C. L., Jankovic, J., Lang, A. E.,
Laws, E. R., Jr, Lozano, A. M., Penn, R. D., Simpson, R. K., Jr,
Stacy, M., et al. (2003). Implanted intracerebroventricular.
Glial cell line-derived neurotrophic factor. Randomized, dou-
ble-blind trial of glial cell line-derived neurotrophic factor
(GDNF) in PD. Neurology 60, 69–73.
Oliveira, M., Braga, S., Passos-Coelho, J. L., Fonseca, R., and
Oliveira, J. (2011). Complete response in HER2þ leptomenin-
geal carcinomatosis from breast cancer with intrathecal tras-
tuzumab. Breast Cancer Res. Treat. 127, 841–844.
Oreskovic, D., and Klarica, M. (2010). The formation of cerebro-
spinal fluid: Nearly a hundred years of interpretations and
misinterpretations. Brain Res. Rev. 64, 241–262.
Pardridge, W. M. (1991). Peptide Drug Delivery to the Brain. Raven
Press, New York.
Patel, N. K., Bunnage, M., Plaha, P., Svendsen, C. N., Heywood, P.,
and Gill, S. S. (2005). Intraputamenal infusion of glial cell
line-derived neurotrophic factor in PD: A two-year outcome
study. Ann. Neurol. 57, 298–302.
Patil, S. A., and Maegawa, G. H. (2013). Developing therapeutic
approaches for metachromatic leukodystrophy. Drug Des.
Devel. Ther 7, 729–745.
Pfeifer, R. W., Felice, B. R., Boyd, R. B., Butt, M. T., Ruiz, J. A.,
Heartlein, M. W., and Calias, P. (2012). Safety evaluation of
chronic intrathecal administration of heparan N-sulfatase in
juvenile cynomolgus monkeys. Drug Deliv. Transl. Res. 2,
187–200.
Provencio, J. J., Kivisakk, P., Tucky, B. H., Luciano, M. G., and
Ranshoff, R. M. (2005). Comparison of ventricular and lumbar
cerebrospinal fluid T cells in non-inflammatory neurological
disorder (NIND) patients. Neuroimmunology 163, 179–184.
Richard, I., and Menei, P. (2007). Intrathecal baclofen in the treat-
ment of spasticity, dystonia and vegetative disorders. Acta
Neurochir. 97, 213–218.
Rubenstein, J. L., Fridlyand, J., Abrey, L., Shen, A., Karch, J., Wang,
E., Issa, S., Damon, L., Prados, M., McDermott, M., et al. (2007).
Phase I study of intraventricular administration of rituximab
in patients with recurrent CNS and intraocular lymphoma.
J. Clin. Oncol. 25, 1350–1356.
Rubalcava, M. A., and Sotelo, J. (1995). Differences between ven-
tricular and lumbar cerebrospinal fluid in hydrocephalus
secondary to cysticercosis. Neurosurgery 37, 668–671.
Sakka, L., Coll, G., and Chazal, J. (2011). Anatomy and physiology
of cerebrospinal fluid. Eur. Ann. Otorhinolaryngol. Head Neck
Dis. 128, 309–316.
Schmued, L., Stowers, C., Scallet, A., and Xu, L. (2005). Fluoro-
Jade results in ultra high resolution and contrast labelling of
degenerating neurons. J. Brain Res. 1035, 24–31.
Schultz, M. L., Tecedor, L., Chang, M., and Davidson, B. L. (2011).
Clarifying lysosomal storage diseases. Trends Neurosci. 34,
401–410.
Skov, M. J., Beck, J. C., de Kater, A. W., and Shopp, G. M. (2007).
Nonclinical safety of ziconotide: An intrathecal analgesic of a
new pharmaceutical class. Int. J. Toxicol. 26, 411–421.
Storkebaum, E., Lambrechts, D., Dewerchin, M., Moreno-
Murciano, M. P., Appelmans, S., Oh, H., Van Damme, P.,
Rutten, B., Man, W. Y., De Mol, M., et al. (2005). Treatment of
motoneuron degeneration by intracerebroventricular deliv-
ery of VEGF in a rat model of ALS. Nat. Neurosci. 8, 85–92.
Switzer, R. (2000). Application of silver degeneration stains for
neurotoxicity testing. Toxicol. Pathol. 28, 70–83.
Switzer, R. (2011). Fundamentals of neurotoxicity detection. In
Fundamental Neuropathology for Pathologists and Toxicologists
Principles and Techniques. (B Bolon and M Butt, Eds.), pp.
139–157. Wiley and Sons, Hoboken, NJ.
Tajes, M., Ramos-Fernandez, E., Weng-Jiang, X., Bosch-Morato,
M., Guivernau, B., Eraso-Pichot, A., Salvador, B., Fernandez-
Busquets, X., Roquer, J., and Mu~noz, F. J. (2014). The blood-
brain barrier: Structure, function and therapeutic approaches
to cross it. Mol. Membr. Biol. 31, 152–167.
Torres-Corzo, J. G., Tapia-Pe´rez, J. H., Sanchez-Aguilar, M., Della
Vecchia, R. R., Chalita Williams, J. C., and Cerda-Gutie´rrez, R.
(2009). Comparison of cerebrospinal fluid obtained by ventric-
ular endoscopy and by lumbar puncture in patients with hy-
drocephalus secondary to neurocysticercosis. Surg. Neurol. 71,
376–379.
USFDA. 2005. Guidance for industry: Estimating the maximum
safe starting dose in initial clinical trials for therapeutics in
adult healthy volunteers. http://www.fda.gov/downloads/
Drugs/.../Guidances/UCM078932.pdf
Vuillemenot, B. R., Katz, M. L., Coates, J. R., Kennedy, D., Tiger, P.,
Kanazono, S., Lobel, P., Sohar, I., Xu, S., Cahayag, R., et al.
(2011). Intrathecal tripeptidyl-peptidase 1 reduces lysosomal
storage in a canine model of late infantile neuronal ceroid
lipofuscinosis. Mol. Genet. Metab. 104, 325–337.
Vuillemenot, B. R., Kennedy, D., Reed, R. P., Boyd, R. B., Butt, M.
T., Musson, D. G., Keve, S., Cahayag, R., Tsuruda, L. S., and
O’Neill, C. A. (2014). Recombinant human tripeptidyl pepti-
dase-1 infusion to the monkey CNS: Safety, pharmacokinet-
ics, and distribution. Toxicol. Appl. Pharmacol. 277, 49–57.
Vuillemenot, B. R., Kennedy, D., Cooper, J. D., Wong, A. M., Sri, S.,
Doeleman, T., Katz, M. L., Coates, J. R., Johnson, G. C., Reed, R.
P., et al. (2015). Nonclinical evaluation of CNS-administered
TPP1 enzyme replacement in canine CLN2 neuronal ceroid
lipofuscinosis. Mol. Genet. Metab. 114, 281–293.
Williams, J. A., Day, M., and Heavner, J. E. (2008). Ziconotide: An
update and review. Exp. Opin. Pharmacother. 9, 575–583.
Xu, S., Wang, L., El-Banna, M., Sohar, I., Sleat, D. E., and Lobel, P.
(2011). Large-volume intrathecal enzyme delivery increases
survival of a mouse model of late infantile neuronal ceroid
lipofuscinosis. Mol. Ther. 19, 1842–1848.
Yang, L., Kress, B. T., Weber, H. J., Thiyagarajan, M., Wang, B.,
Deane, R., Benveniste, H., Iliff, J. J., and Nedergaard, M. (2013).
Evaluating glymphatic pathway function utilizing clinically rele-
vant intrathecal infusion of CSF tracer. J. Transl. Med. 11, 107.
VUILLEMENOT ET AL. | 9
atLibraryonJune27,2016http://toxsci.oxfordjournals.org/Downloadedfrom