CURRENT OPINION
`
`Drugs R D 2007; 8 (3): 133-144
`1174-5886/07/0003-0133/$44.95/0
`
`© 2007 Adis Data Information BV. All rights reserved.
`
`Can Nasal Drug Delivery Bypass
`the Blood-Brain Barrier?
`Questioning the Direct Transport Theory
`
`Frans W.H.M. Merkus1 and Mascha P. van den Berg2
`1 Leiden/Amsterdam Centre for Drug Research, Leiden University, Leiden, The Netherlands
`2 Pfizer Global Research and Development, Sandwich, UK
`
`Abstract
`
`The connection between the nasal cavity and the CNS by the olfactory
`neurones has been investigated extensively during the last decades with regard to
`its feasibility to serve as a direct drug transport route to the CSF and brain. This
`drug transport route has gained much interest as it may circumvent the blood-brain
`barrier (BBB), which prevents some drugs from entering the brain. Approximate-
`ly 100 published papers mainly reporting animal experiments were reviewed to
`evaluate whether the experimental design used and the results generated provided
`adequate pharmacokinetic information to assess whether the investigated drug
`was transported directly from the olfactory area to the CNS. In the analysis the
`large anatomical differences between the olfactory areas of animals and humans
`and the experimental conditions used were evaluated. The aim of this paper was to
`establish the actual evidence for the feasibility of this direct transport route in
`humans.
`Twelve papers presented a sound experimental design to study direct nose to
`CNS transport of drugs based on the authors’ criteria. Of these, only two studies in
`rats were able to provide results that can be seen as an indication for direct
`transport from the nose to the CNS. No pharmacokinetic evidence could be found
`to support a claim that nasal administration of drugs in humans will result in an
`enhanced delivery to their target sites in the brain compared with intravenous
`administration of the same drug under similar dosage conditions.
`
`these papers reveals that the methods used do not
`The connection between the nasal cavity and the
`often utilise the necessary design to determine the
`CNS by the olfactory neurones has been investigat-
`actual drug transport route. In most cases the animal
`ed extensively during the last decades with regard to
`experiments cannot be translated to the human situa-
`its feasibility to serve as a direct drug transport route
`tion. It must not be forgotten that the ultimate goal
`to the brain. This drug transport route has gained
`of the exploration of this direct nose to CNS path-
`much interest as it may circumvent the blood-brain
`barrier (BBB), which prevents drugs from entering way is to be able to apply this delivery route to
`the brain. More than 100 research papers have de-
`humans, facilitating CNS drug concentrations that
`scribed studies on the nose to CNS route (mainly in
`can elicit a central pharmacological effect. All pa-
`animals), with ambiguous results. Careful review of
`pers were reviewed to evaluate whether the experi-
`
`AQUESTIVE EXHIBIT 1117 Page 0001
`
`

`

`134
`
`Merkus & van den Berg
`
`mental design used and the results generated provid-
`ed adequate information to assess whether the inves-
`tigated drug was transported directly from the nasal
`cavity to the CNS.
`
`1. The in Vivo Fate of Intranasal Drugs
`
`Before defining criteria that should be included
`in an experimental design when investigating drug
`transport from the nasal cavity into the cerebrospinal
`fluid (CSF) and/or brain tissue, the in vivo fate of
`intranasal drugs will be described (see figure 1).
`This will serve as a basis for the theoretical experi-
`mental design discussed in the next section.
`Firstly, drugs reach the respiratory epithelium,
`from where they are absorbed into the systemic
`circulation or cleared by mucociliary clearance and
`swallowed. Drugs that are absorbed into the system-
`ic circulation may enter the CNS after passing
`through the BBB. When a nasal drug formulation is
`deposited directly on the olfactory epithelium it is
`possible that drug transport via the olfactory
`neurones occurs. Two possible routes exist by which
`molecules can be transported from the olfactory
`epithelium into the brain and/or CSF. The first is the
`epithelial pathway, where compounds pass paracel-
`lularly across the olfactory epithelium into the peri-
`neural spaces, crossing the cribriform plate and en-
`tering the subarachnoid space filled with CSF. From
`here the molecules can diffuse into the brain tissue
`or will be cleared by the CSF flow into the lymphat-
`ic vessels and subsequently into the systemic circu-
`lation. The second possibility is the olfactory nerve
`pathway, where compounds may be internalised into
`
`Nose
`
`Olfactory
`region
`
`Respiratory
`region
`
`Brain tissue
`
`CSF
`
`Blood
`
`Clearance
`
`Elimination
`
`Fig. 1. Schematic overview of the in vivo fate of drugs following
`nasal administration. CSF = cerebrospinal fluid.
`
`the olfactory neurones and pass inside the neuron
`through the cribriform plate into the olfactory bulb.
`It is possible that further transport into the brain can
`occur by bridging the synapses between the neurons.
`After reaching the brain tissue, drugs are cleared
`either via the CSF flow or via efflux pumps such as
`p-glycoprotein at the BBB[1] into the systemic circu-
`lation. Furthermore, the trigeminal nerve[2] and, in
`animals, the vomeronasal organ[3] also connect the
`nasal cavity with the brain tissue.
`
`2. Experimental Models to Study Nose to
`CNS Drug Transport
`
`2.1 Are the Experiments Realistic for the
`Human Situation?
`
`In most animal studies the investigators suggest
`that nose to CSF/brain transport is also feasible in
`humans. However, there are large anatomical differ-
`ences between animals and man. For instance, in
`rats 50% of the nasal cavity is occupied by olfactory
`epithelium. In humans this is only 3%[4] and, more
`importantly, the olfactory area, located in the roof of
`the nasal cavity, is difficult to reach using nasal
`drops or a nasal spray. Secondly, in most mammali-
`an species (including rodents) the nose contains a
`vomeronasal organ with nerves also connecting di-
`rectly to the brain. However, the existence and func-
`tioning of this organ in humans is still under de-
`bate.[3] Furthermore, many formulations used in
`animal studies contain mucosa-damaging perme-
`ation enhancers (e.g. propylene glycol, ethanol or
`other organic solvents in concentrations up to
`40%),[5-11] while some nasal formulations were used
`in an extremely aggressive way (e.g. spraying a
`formulation over 1 minute using an atomiser con-
`nected to a respiratory pump with high pressure)[5-8]
`or in a large volume.[12-32] When olfactory epithelial
`cells are in contact with these solutions for a long
`time, severe cell structure damage occurs. Aggres-
`sive spraying methods are very painful and severely
`damage the olfactory area.
`Some researchers,[9,33-40] for instance, have ap-
`plied surgical modifications to rats[41] in which the
`trachea of the animal is canulated to maintain respi-
`
`© 2007 Adis Data Information BV. All rights reserved.
`
`Drugs R D 2007; 8 (3)
`
`AQUESTIVE EXHIBIT 1117 Page 0002
`
`

`

`Can Nasal Drug Delivery Bypass the Blood-Brain Barrier?
`
`135
`
`b
`
`Intranasal
`Intravenous
`
`0
`
`5
`
`15
`
`60
`30
`Time (min)
`
`120 180 240
`
`d
`
`40 000
`35 000
`30 000
`25 000
`20 000
`15 000
`10 000
`5000
`0
`
`AUC (nmol · min/L)
`
`0
`
`5
`
`10 20 30 40 60 120 180
`Time (min)
`
`a
`
`1000
`
`800
`
`600
`
`400
`
`200
`
`0
`
`AUC (nmol · min/L)
`
`c
`
`6
`
`5
`
`4
`
`3
`
`2
`
`1
`
`AUC (nmol · min/L)
`
`120
`60
`10 20 30 40 60 120 180
`Time (min)
`Time (min)
`Fig. 2. Hydroxocobalamin area under the concentration-time curve (AUC) for plasma (a, b) and cerebrospinal fluid (c,d) following intranasal
`administration and intravenous infusion in humans (a and c) [n = 5] and in rats (b and d) [n = 8].
`
`0
`
`30
`
`180
`
`240
`
`400
`350
`300
`250
`200
`150
`100
`50
`0
`
`AUC (nmol · min/L)
`
`0
`
`0
`
`5
`
`ration, the oesophagus tied off to prevent swallow-
`ing of the nasal drug formulation, and the nasopala-
`tine duct sealed to prevent any nasal formulation
`from being cleared to the mouth. In addition, the
`nasal perfusion method[42] has been used to deliver
`the formulation.[9,34,36,37,39] The use of such a treat-
`ment would be unrealistic, even unthinkable, in
`humans. Therefore one should exercise caution in
`the interpretation of such animal results.
`A realistic comparison of human and rat data was
`made recently in studies where similar methods
`were used in patients and in rats.[43-45] By adminis-
`tering identical intranasal and intravenous drug for-
`mulations of the lipophilic drug melatonin and the
`hydrophilic drug hydroxocobalamin and using simi-
`lar sampling times, analogous results were obtained
`(figure 2). The drugs investigated were transported
`to the CSF by the systemic pathway: after first being
`absorbed in the systemic circulation, the drug
`reached the CSF via the blood-CSF barrier. The fact
`that only a few human studies report drug concentra-
`tions in the CSF[14,45] or brain (using a positron
`
`emission tomography [PET] scan[46]) is because of
`the difficulties and risks associated with CSF collec-
`tion in humans and the costly and specialised tech-
`nique of PET scanning.
`When investigating the direct transport route a
`distinction must be made between drugs transported
`from the nasal cavity via (i) the systemic circulation
`after crossing the respective BBB and blood-CSF
`barrier, or (ii) via the olfactory epithelium directly
`into the CSF and/or brain tissue. In addition, a nasal
`delivery method appropriate for the species must be
`chosen. Features such as delivery volume, dose fre-
`quency and safety of the formulation need to be
`considered. Delivery volumes that do not overload
`the nasal cavity must be used, and should not exceed
`20µL[47] in rats and 100µL[45] in humans. Investigat-
`ing drug transport requires a pharmacokinetic ap-
`proach, determining and comparing drug concentra-
`tions in the relevant biocompartments after in-
`tranasal drug administration and after a slow
`intravenous infusion, giving plasma concentrations
`similar to those expected after intranasal applica-
`
`© 2007 Adis Data Information BV. All rights reserved.
`
`Drugs R D 2007; 8 (3)
`
`AQUESTIVE EXHIBIT 1117 Page 0003
`
`

`

`136
`
`Merkus & van den Berg
`
`tion. Ideally the drug plasma profiles achieved
`should be similar. This ensures that the rate of
`passive diffusion from the systemic circulation into
`the CNS is the same after both intranasal and intra-
`venous administration. The actual route of drug
`transport can be determined by calculating the rela-
`tive distribution of a drug to the CNS (CSF or brain)
`and plasma after both methods of administration
`according to equation 1:
`
`CNS/plasma ratio = AUCCNS
`AUCplasma
`
`(Eq. 1)
`The CNS/plasma ratio of a compound is ex-
`pressed by the area under the CNS concentration-
`time curve (AUCCNS) divided by the area under the
`plasma concentration-time curve (AUCplasma) fol-
`lowing the same administration route.[48] In case of
`direct drug transport from the nasal cavity, it is
`expected that the CNS/plasma ratio following in-
`tranasal delivery will be significantly higher than
`that after intravenous administration. When this ra-
`tio is equal to or lower than the intravenous one, the
`observed drug transport can be considered to be
`systemic and not via the olfactory neurones.
`All the above-mentioned features necessary for a
`realistic experimental design to study nose to CNS
`drug transport are summarised in table I.
`
`Table I. Criteria for a realistic experimental design when investigat-
`ing nose to CNS drug transport
`The method of nasal administration should be appropriate for the
`animal species used and realistic when compared with the
`human situation. In other words the volume/dose should be
`realistic and the excipients and/or permeation enhancer used in
`the formulation and the administration method used should be
`safe for human application
`In order to compare exactly the transport of the drug via the
`blood-brain barrier and via the olfactory route, the drug should be
`administered both intranasally and intravenously, aiming at
`comparable plasma drug concentrations from both routes
`Drug concentrations in plasma and CNS (CSF and/or brain)
`should be measured and preferably AUC values for plasma and
`CNS (CSF and/or brain) should be calculated
`Drug distribution over the CNS compartment and the systemic
`compartment should be compared following intranasal and
`intravenous delivery
`AUC = area under the concentration-time curve; CSF = cerebro-
`spinal fluid.
`
`2.2 Study Designs Used in the Literature
`
`The literature investigated for this review can be
`divided into four categories: nose to brain, nose to
`CSF, nose to brain and CSF, and pharmacodynamic
`research (table II).
`For each category the following aspects are dis-
`cussed: (i) species used; (ii) the delivery route with
`which the intranasal administration is compared,
`also referred to as the reference route; (iii) the type
`of samples taken and sampling techniques used; and
`(iv) general remarks about the disadvantages and
`advantages of the study design. As the main focus of
`this paper is on transport of drugs, the literature on
`dyes, metals, micro-organisms and tracers like
`WGA-HRP were not reviewed.
`In the majority of the studies drugs were formu-
`lated in an aqueous solution such as saline or buf-
`fered solutions. In 15 papers[14,22,25,49-51,69,70,94-99,114]
`the composition of the formulation was not men-
`tioned, and in seven articles the compound was
`dissolved in a mixture of water, ethanol and propyl-
`ene glycol,[5-8] diluted ethanol[9,10] or a 40% isopro-
`pyl alcohol solution containing 10% sefsol (a skin
`permeation enhancer).[11] Administration of drugs
`into the nose is most often achieved using a piece of
`tubing attached to a microsyringe or micropipette in
`animals and a spray in humans. Typical delivery
`volumes are 5–10 µL/nostril in mice, 10–25 µL/
`nostril in rats and 75–100 µL/nostril in humans.
`These volumes were used as criteria in the eventual
`test described below. In a substantial number of
`investigations in animals[12,15,16,19,20,22,25,26,30-32] and
`humans[13,14,17,18,21,23,24,27-29] (21 of 104), a relatively
`large dose (and in total a large volume) was adminis-
`tered. These studies involved the instillation of mul-
`tiple aliquots of the nasal formulation, taking up to
`30 minutes to administer the complete dose.[12-32]
`
`2.2.1 Nose to Brain Research
`Drug transport from the nasal cavity specifically
`into brain tissue has been studied mainly in rats and
`mice. In most cases the intranasal route of adminis-
`tration was compared with an intravenous bolus
`injection (16 of 26 papers). Brain tissue was ob-
`tained as a whole,[50,52-54] in dissected brain re-
`
`© 2007 Adis Data Information BV. All rights reserved.
`
`Drugs R D 2007; 8 (3)
`
`AQUESTIVE EXHIBIT 1117 Page 0004
`
`

`

`Can Nasal Drug Delivery Bypass the Blood-Brain Barrier?
`
`137
`
`Table II. References for the identified publications per research category
`Category
`References
`Nose to brain
`12,16,30,31,33,34,46,49-68
`Nose to CSF
`6-8,14,25,36-38,40,43-45,69-88
`Nose to brain and CSF
`5,35,39,89-93
`Pharmacodynamics
`9-11,13,15,17-19,21-24,26-29,32,94-113
`Total
`CSF = cerebrospinal fluid.
`
`No. of papers
`27
`32
`8
`37
`104
`
`gions[16,30,33,49,51,52,55-59] or in slices.[30,51,52,55-57,59]
`are significantly different. Finally, this transport
`route has also been investigated on the cellular level
`About half of the studies (13 of 26) investigated the
`using microscopy techniques.[62-66]
`drug absorption in blood[30,33,46,49-51,53,54,57-59] and/or
`drug uptake into other tissues as well.[30,51,52,55,57]
`The disadvantages of nose to brain research are:
`In nine papers the drug uptake into the brain
`(i) the need for many experimental animals; (ii) it is
`following nasal delivery was studied qualitatively
`not applicable in humans, except when using scan-
`and quantitatively.[30,46,51,52,54-57,59] By using a radio-
`ning techniques like PET; and (iii) analysing whole
`labelled compound and autoradiographic analysis,
`brain tissue dilutes the actual drug concentration at
`the drug uptake and distribution throughout the
`the target site and is therefore less informative than
`brain can be visualised. The drug concentration in
`results obtained from specific brain regions. The
`brain tissue was often determined in several brain
`advantages of this approach are that: (i) drug trans-
`areas at a single[51,57] or at three to six time
`port can be studied in a quantitative and qualitative
`points.[30,52,54-56,59] The technique of PET allowed way; and (ii) drug concentrations can be measured at
`Wall et al.[46] to take up to 18 ‘brain samples’ from
`the expected target site.
`each volunteer during a 90-minute post-dose period
`to detect drug uptake into the brain following nasal
`2.2.2 Nose to CSF Research
`administration of (11C) zolmitriptan. High uptake
`The nose to CSF pathway has been investigated
`values were found for the extracranial area in con-
`in a quantitative way only and was mainly studied in
`trast to the inside of the cranium, and hence it was
`rats (23 of 34 papers). All studies compared the
`concluded that zolmitriptan entered the brain after
`intranasal route with intravenous administration
`passing through the BBB, so no direct transport was
`(mainly by bolus injection), except for Chou and
`found.
`Donovan,[71-73] who looked at intra-arterial delivery
`Data analysis using the brain/plasma ratios as
`in rats, Anand Kumar et al.,[6] who investigated the
`described in equation 1 was used in five pa-
`intramuscular route in monkeys, and Born et al.,[14]
`pers,[33,49,50,53,58] whereas three other papers only who compared nasal drug delivery with placebo
`looked at the brain/plasma ratio at a certain time
`treatment in human volunteers. With the exception
`point.[12,60,61] Vyas et al.[54] used the so-called drug
`of one article,[72] all papers report serial blood sam-
`targeting efficiency (DTE) quotient to determine the
`pling per subject, resulting in drug concentration-
`drug transport route into the brain following nasal
`time profiles in plasma or serum. The CSF was
`delivery (see equation 2):
`sampled mainly in a serial manner. However, in 11
`papers, single CSF samples were taken per animal;
`in most cases only one sample was collected at the
`end of the experiment.[25,36-38,40,70,74,75] Only a few
`studies generated CSF concentration-time profiles
`by taking a single CSF sample per animal and using
`several animals per profile.[76-78] Serial CSF sam-
`pling is performed by microdialysis[72,73,79-81] or cis-
`
`(Eq. 2)
`However, this calculation method does not con-
`sider whether or not the two CNS/plasma ratios for
`intranasal (IN) and intravenous (IV) administration
`
`DTE quotient = (AUCCNS /AUCplasma)IN
`(AUCCNS /AUCplasma)IV
`
`© 2007 Adis Data Information BV. All rights reserved.
`
`Drugs R D 2007; 8 (3)
`
`AQUESTIVE EXHIBIT 1117 Page 0005
`
`

`

`138
`
`Merkus & van den Berg
`
`Despite the fact that most of the studies presented
`ternal puncture. This latter method can be done by
`drug concentration-time profiles following nasal
`simply inserting a needle into the cisterna magna of
`and intravenous delivery, only the studies by Anand
`a rat or by placing a rat in a stereotaxic frame in the
`usual upright position[71] or in the supine-70° angle Kumar et al.[5] and Shi et al.[90] used serial sampling
`position.[115] The method described by Chou and
`techniques to obtain either blood or CSF. This is
`Donovan[71] has the advantage of replacement of the
`because brain, blood and CSF were sampled from
`CSF volume by infusing artificial CSF into the third
`the same animal.
`ventricle. In contrast, the method by van den Berg et
`The main disadvantage of this approach is that it
`al.,[115] using the supine-70° angle position instead
`takes a lot of experimental animals to obtain all the
`of the upright position, results in enhanced drug
`data, while the advantage is that more information
`absorption in blood and uptake in CSF following
`on the drug distribution is provided.
`nasal delivery. A disadvantage of these two methods
`is that only small CSF samples can be collected
`2.2.4 Pharmacodynamic Research
`(<50µL), but an important advantage is that these
`The majority of the pharmacodynamic studies
`are not contaminated with blood. Two studies were
`were carried out in humans (20 of 37 papers), some
`performed obtaining serial CSF samples from
`in rats (11 papers) and a few in mice (5 papers).
`humans via an interspinal puncture[14] or using a
`When looking at the reference routes used, it is
`CSF drain in neurosurgery patients.[45] For single
`remarkable to see that all except three stud-
`CSF sampling in rats, methods described by Chou
`ies[94,99,100] used a placebo treatment and only seven
`and Levy[116] and Waynforth and Flecknell[117] were
`investigated the intravenous administration route as
`mainly used. The advantage is that large CSF
`a reference.[32,94,98,100-103] This is in contrast to the
`volumes can be obtained (>75µL); however, there is
`pharmacokinetic studies described above where in-
`a big risk of blood contamination, which may con-
`travenous delivery was the main reference route,
`tribute to the final drug concentration in the CSF
`regardless of the sampling site (brain, CSF or both).
`sample.
`Nearly one-half of the reviewed articles did not
`The disadvantages of studying nose to CSF drug
`report any pharmacokinetic data. The other half
`transport are: (i) when using small animals, highly mainly reported drug plasma concentrations, two
`sensitive analysis methods are required because of
`papers presented plasma and CSF data in rats and
`the small volumes of CSF samples collected; (ii) it is
`humans,[10,104] and one measured only CSF concen-
`difficult to apply to humans; and (iii) drug concen-
`trations in rats.[105]
`trations are not determined at the target site. The
`Because of the wide variety of compounds inves-
`main advantage is the possibility of obtaining a drug
`tigated in this category, a similar diversity of phar-
`concentration-time profile in CSF from one subject,
`macodynamic endpoints was used. Effects were
`which saves animals and reduces the variability in
`studied
`in
`relation
`to memory and
`learn-
`the data.
`ing,[11,13,19,95,96,99] food intake[10,21,24,104,106] and car-
`functioning.[18,29,101]
`diovascular
`Besides
`behavioural[26,32] and clinical effects,[100,107,108]
`changes
`in
`neurotransmitter
`and
`hormone
`levels,[9,27,105]
`electroencephalographic
`record-
`ings[17,23,97,98,102,103,109]
`and
`histological
`changes[15,22,110,111] were also used to determine the
`efficacy of nasal drug delivery.
`A disadvantage of most of the literature reviewed
`for this category is the lack of data on drug concen-
`trations in plasma and the CNS. The advantages are:
`
`2.2.3 Nose to Brain and CSF Research
`From the selection of studied papers, only eight
`report drug concentrations in CSF and brain tissue
`following nasal drug delivery. Five of these studies
`were performed in rats[35,39,89-91] and the other three
`were conducted in monkeys,[5] rabbits[92] or mice.[93]
`All of these articles compared nasal drug delivery
`with the intravenous route of administration, mainly
`given as a bolus injection.
`
`© 2007 Adis Data Information BV. All rights reserved.
`
`Drugs R D 2007; 8 (3)
`
`AQUESTIVE EXHIBIT 1117 Page 0006
`
`

`

`Can Nasal Drug Delivery Bypass the Blood-Brain Barrier?
`
`139
`
`(i) the actual drug efficacy is studied; and (ii) it is
`applicable to study in humans.
`
`3. Putting the Literature to the Test
`
`sound evidence on direct nose to CNS drug transport
`is performed by testing the collection of papers
`listed in table II against the criteria for experimental
`design listed in table I (see also section 2.1).
`
`3.1 Selection Criteria
`
`3.2 Results of the Test
`
`Table III presents the papers that passed the test
`The main question is whether there is solid evi-
`on experimental design, the compound tested, the
`dence for direct nose to CNS transport of an amount
`claim made by the researchers and the actual proof
`of drug sufficient to elicit a pharmacological effect
`for direct nose to CNS transport.
`in humans. This question can only be answered by
`From all the papers screened only 12 used an
`quantitative pharmacokinetic research. Therefore,
`experimental design suitable for investigating quan-
`those papers that investigated this transport route
`titative nose to CNS drug transport. In the ‘nose to
`qualitatively, using only visualisation techniques,
`brain’ category the main reason for excluding papers
`were excluded. Nevertheless, these imaging studies
`do demonstrate uptake of compounds via the olfac- was that only a qualitative approach was used and
`excessively large volumes were delivered. For the
`ork[63]tory pathway in rats and mice. Jansson and Bj¨
`
`papers in the ‘nose to CSF’ and ‘nose to brain and
`observed transcellular transport of fluorescein dex-
`tran (MW = 3000Da) across the olfactory epithelium CSF’ categories the lack of AUC data was the main
`reason for exclusion. Another important reason was
`and further into the connective tissue surrounding
`the use of inappropriate nasal delivery protocols. As
`the olfactory nerves into the olfactory bulb in rats
`a result, no articles from the category ‘nose to brain
`using fluorescence microscopy. Several papers
`and CSF’ passed the test. The lack of plasma and
`have demonstrated transport of small molecules
`aflatoxin B1,[64] CNS drug concentrations in most of the papers
`(MW
`120–320Da)
`like
`dopamine,[52] morphine,[59] within the ‘pharmacodynamic’ category was the
`benzo(a)pyrene,[66]
`reason to exclude these investigations from further
`picolinic acid[55] and different carboxylic acids[56]
`selection. In two studies blood and CSF samples
`into the olfactory bulb in rodents, but did not find
`uptake of these compounds in the rest of the brain. were collected; however, nasal drug delivery was
`not compared with intravenous administration.[10,104]
`Two studies using phosphor imaging techniques
`demonstrated enhanced uptake of interferon-β[57]
`The selection of 12 papers that did pass the test
`and insulin-like growth factor-I[30] following in-
`contained six positive and 11 negative claims on
`direct nose to CNS transport.1 Two of these claims
`tranasal delivery in rats compared with intravenous
`administration. In both studies nasal delivery result- were based on incorrect data analysis, which is
`ed in drug uptake beyond the olfactory bulb into the marked with an ‘X’ in the ‘Proof’ column of table
`cortices, inner brain areas and brainstem and spinal
`III, and are therefore not considered justified. One
`cord. The drug distribution into these latter two
`claim is marked with an ‘X’ because intranasal/
`areas was attributed to drug transport via the trigem-
`intravenous ratios were calculated for plasma and
`inal nerve.[30] However,
`the results of
`these CSF and/or brain tissue instead of CNS/plasma ra-
`tios per delivery route.[53] The claim made by Vyas
`phosphor imaging papers might draw a picture that
`is too optimistic as large volumes (50 and 120µL)
`et al.[54] was considered invalid, because it is not
`were delivered to rats over 20–40 minutes, and the
`clear from their calculations whether their results are
`very large olfactory area in rats compared with
`statistically significant. More importantly, they used
`humans must be considered as well. The search for
`a nasal formulation of 5% zolmitriptan (very water
`
`1
`The total number of claims (n = 17) exceeds the number of papers (n = 12) as some papers tested more than one
`compound.
`
`© 2007 Adis Data Information BV. All rights reserved.
`
`Drugs R D 2007; 8 (3)
`
`AQUESTIVE EXHIBIT 1117 Page 0007
`
`

`

`140
`
`Merkus & van den Berg
`
`Table III. Selection of papers that meet the criteria set in table I
`Reference
`Compound
`Nose to brain (3 papers; 3 compounds tested)
`[14C] Dextromethorphan
`53
`54
`Zolmitriptan
`67
`Morphine
`
`Rat
`Rat
`Rat
`
`Noa
`Yes
`Yes
`
`Species
`
`Claim
`
`Proof
`
`Nose to CSF (9 papers; 14 compounds tested)
`81
`Fluorescein
`80
`Lidocaine
`73
`Procaine
`Tetracaine
`Bupivicaine
`Lidocaine
`FD3b
`Hydroxocobalamin
`Melatonin
`Hydrocortisone
`Hydroxocobalamin
`Estradiol
`Progesterone
`Melatonin
`44
`a No proof for direct nose to CNS transport.
`b Fluorescein isothiocyanate-labelled dextran with a molecular weight of 3.0 kDa.
`CSF = cerebrospinal fluid; I = indicative for direct nose to CNS transport; X = calculation used is incorrect.
`
`Rat
`Rat
`Rat
`Rat
`Rat
`Rat
`Rat
`Human
`Human
`Rat
`Rat
`Rat
`Rat
`Rat
`
`No
`No
`Yes
`Yes
`Yes
`Yes
`No
`No
`No
`No
`No
`No
`No
`No
`
`85
`45
`
`83
`43
`84
`
`X
`X
`I
`
`No
`No
`I
`I
`I
`I
`No
`No
`No
`No
`No
`No
`No
`No
`
`soluble) containing 50% surfactant, which may se-
`verely damage the olfactory epithelial tissue.
`The ‘Proof’ column differs for two more papers
`from the ‘Claim’ column in table III. In the anaes-
`thetics study by Chou and Donovan,[73] the nose to
`CSF transport route was determined by comparing
`CSF/plasma ratios following intranasal and intrave-
`nous delivery as proposed in this review. However,
`no indication of variability of the measured CSF/
`plasma ratios is mentioned. Therefore, it cannot be
`concluded that CSF/plasma ratios for intranasal de-
`livery are significantly higher than those following
`intravenous administration of the local anaesthetics
`lidocaine, procaine, tetracaine and bupivicaine. For
`this reason we have marked these four compounds
`as ‘indicative’ for nose to CSF transport.
`The morphine study by Westin et al.[67] is the
`only one that met all the criteria mentioned in table
`I, except for using a relatively large delivery volume
`(50 µL/nostril). An enhanced brain uptake of mor-
`phine was found in the first 60 minutes following
`nasal delivery in rats as compared with intravenous
`
`infusion. The total brain uptake of morphine at 240
`minutes after delivery was comparable to that for
`intravenous delivery. The authors of this paper
`rightly conclude that the demonstrated olfactory
`transfer of morphine in rats “could be beneficial in
`pain treatment … However, conclusions concerning
`the impact of the olfactory transfer of morphine in
`humans cannot be drawn from the results of rat
`studies” because of the anatomical differences be-
`tween the species.
`Four papers in the ‘pharmacodynamics’ category
`reported enhanced efficacy for intranasal delivery
`over intravenous administration, with comparable or
`higher intravenous drug exposure versus intranasal
`delivery in humans.[98,101-103] Although this suggests
`direct drug transport, the contribution of the olfacto-
`ry pathway to this effect cannot be determined be-
`cause of the lack of pharmacokinetic data. There-
`fore, these studies are not included in table III.
`In summary, 12 of all papers screened presented
`a sound experimental design to study nose to CNS
`transport of drugs. Only two of these papers were
`
`© 2007 Adis Data Information BV. All rights reserved.
`
`Drugs R D 2007; 8 (3)
`
`AQUESTIVE EXHIBIT 1117 Page 0008
`
`

`

`Can Nasal Drug Delivery Bypass the Blood-Brain Barrier?
`
`141
`
`able to provide results that can be seen as an indica-
`tion for direct transport from the nose to the CNS.
`
`4. Conclusion
`In the present review more than 100 papers were
`assessed for pharmacokinetic evidence indicating a
`possible direct drug transport route from the nasal
`cavity to the CNS. Firstly, the articles were screened
`according to the criteria listed in table I on the
`experimental design and methods used, and, second-
`ly, on the method of data analysis used. Only two
`studies in rats received the mark ‘indicative for
`direct nose to CNS transport’ and none of the arti-
`cles was considered to have shown convincing proof
`of a direct pathway. Taken together this is not strong
`pharmacokinetic evidence to claim in humans that
`the intranasal route of administration provides en-
`hanced or direct access of a nasally administered
`drug into the CNS, compared with drug transport via
`the systemic pathway.
`
`Acknowledgements
`The authors used no sources of funding in the writing of
`this article. The authors have no conflicts of interest that are
`directly relevant to the content of this article.
`
`References
`1. Pardridge WM. Drug delivery to the brain. J Cereb Blood Metab
`1997; 17 (7): 713-31
`2. Barnett EM, Perlman S. The olfactory nerve and not the trigemi-
`nal nerve is the major site of CNS entry for mouse hepatitis
`