ARTICLE IN PRESS
`
`European Journal of Pharmaceutics and Biopharmaceutics xx (2003) xxx–xxx
`
`www.elsevier.com/locate/ejpb
`
`Review article
`
`Cyclosporine A delivery to the eye: A pharmaceutical challenge
`
`F. Lallemanda, O. Felt-Baeyensa, K. Besseghirb, F. Behar-Cohenc, R. Gurnya,*
`
`aSchool of Pharmacy, University of Geneva, Geneva, Switzerland
`bDebiopharm S.A., Lausanne, Switzerland
`cINSERM U450, Paris, France
`
`Received 5 February 2003; accepted in revised form 5 August 2003
`
`Abstract
`
`Systemic administration of cyclosporine A (CsA) is commonly used in the treatment of local ophthalmic conditions involving cytokines,
`such as corneal graft rejection, autoimmune uveitis and dry eye syndrome. Local administration is expected to avoid the various side effects
`associated with systemic delivery. However, the currently available systems using oils to deliver CsA topically are poorly tolerated and
`provide a low bioavailability. These difficulties may be overcome through formulations aimed at improving CsA water solubility (e.g.
`cyclodextrins), or those designed to facilitate tissue drug penetration using penetration enhancers. The use of colloidal carriers (micelles,
`emulsions, liposomes and nanoparticles) as well as the approach using hydrosoluble prodrugs of CsA have shown promising results. Solid
`devices such as shields and particles of collagen have been investigated to enhance retention time on the eye surface. Some of these topical
`formulations have shown efficacy in the treatment of extraocular diseases but were inefficient at reaching intraocular targets. Microspheres,
`implants and liposomes have been developed to be directly administered subconjunctivally or intravitreally in order to enhance CsA
`concentration in the vitreous. Although progress has been made, there is still room for improvement in CsA ocular application, as none of
`these formulations is ideal.
`q 2003 Published by Elsevier B.V.
`
`Keywords: Cyclosporine A; Ocular delivery; Intraocular; Topical; Delivery system; Prodrug; Review
`
`1. Introduction
`
`is a cyclic undecapeptide
`Cyclosporine A (CsA)
`produced by Tolypocladium inflattum Gams and other
`fungi imperfecti. This drug is now routinely used as an
`oral immunosuppressor for organ transplantation. It acts by
`selective inhibition of interleukin-2 release during the
`activation of T-cells and causes suppression of the cell-
`mediated immune response [1]. The resultant
`immuno-
`suppression is non-toxic and reversible when treatment is
`stopped. Therefore, most of the diseases that
`involve
`cytokines or immune – related disorders are potential targets
`of CsA.
`Over the past years, CsA has been evaluated for
`numerous potential applications in ophthalmology. It is
`effective in the treatment of severe intraocular inflam-
`mations affecting the posterior segment of the eye, when
`
`* Corresponding author. School of Pharmacy, University of Geneva, 30,
`quai E. Ansermet, CH-1211 Geneva 4, Switzerland. Tel.: þ 41-22-379-61-
`46; fax: þ 41-22-379-65-67.
`
`E-mail address: robert.gurny@pharm.unige.ch (R. Gurny).
`
`0939-6411/$ - see front matter q 2003 Published by Elsevier B.V.
`doi:10.1016/S0939-6411(03)00138-3
`
`administered systemically by i.v. injection [2] or orally [3].
`Systemic CsA has also shown efficacy in peripheral
`ulcerative keratitis associated with Wegener’s granuloma-
`tosis [4], in severe Grave’s ophthalmopathy [5] and it was
`also effective in preventing recurrence of graft rejection
`after keratoplasty, and this for a long period of time [6]. In
`fact, intraocular fluids (aqueous or vitreous humor) and
`extraocular organs or annexes (cornea, conjunctiva and
`lachrymal glands) can be reached through the systemic
`pathway after oral or i.v. administration. CsA concen-
`trations of 25 – 75 mg/ml were measured in human tears after
`an oral daily administration of 5 mg/kg [7] but deleterious
`side effects such as nephrotoxicity and hypertension may
`occur [8 – 10]. Although it has met numerous difficulties,
`topical ocular delivery should offer a good alternative.
`Despite a poor intraocular penetration, topical CsA has been
`successfully used in a variety of immune-mediated ocular
`surface phenomena like vernal conjunctivitis [11], dry eye
`syndrome [12] and the prevention of corneal allograft
`rejection [13]. An ideal topical ocular formulation must
`fulfill several requirements: the formulation must be well
`
`1
`
`ALL 2014
`MYLAN PHARMACEUTICALS V. ALLERGAN
`IPR2016-01128
`
`

`

`ARTICLE IN PRESS
`
`2
`
`F. Lallemand et al. / European Journal of Pharmaceutics and Biopharmaceutics xx (2003) xxx–xxx
`
`tolerated and easy to administer, increase CsA residence
`time in the eye and avoid systemic absorption (the toxic
`concentration in blood is above 300 ng/ml [14]). In addition,
`the formulation should have a long shelf life and be
`manufactured easily. The main difficulty is that CsA cannot
`be prepared in formulations based on the commonly used
`aqueous ophthalmic vehicles because of both its hydro-
`
`phobicity (log P ¼ 3.0 [15]) and its extremely low aqueous
`
`solubility (6.6 mg/ml [16]). Therefore, in most studies, CsA
`was dissolved and administered in vegetable oils [17 – 19].
`However,
`these media are poorly tolerated, result
`in
`relatively low ocular availability, and have short shelf
`lives. The concentration that has been mostly investigated
`for an eyedrop solution is 1% w/v [20,21] but concentrations
`ranging from 2% w/v [22] to 0.05% w/v [23] have also been
`explored. In all cases, the concentration of the formulation
`remains secondary as long as therapeutic levels in the ocular
`tissues are achieved by the dosage form, immune response
`and inflammation being suppressed at a concentration of
`50 – 300 ng/g of
`tissue [24]. When high intraocular
`concentrations are needed, the drug is also injected directly
`into the eye or periocularly (by the subconjunctival route).
`Numerous formulations were developed to avoid repeated
`injections and achieve controlled release of CsA or to
`enhance efficacy of topical administration. This review
`summarizes the main pharmaceutical systems and devices
`that have been described for topical and intraocular delivery
`of CsA to the eye. It will first present and discuss the topical
`systems developed during the last 15 years. In the second
`part,
`intraocular and subconjunctival devices will be
`reviewed.
`
`2. Topical administration
`
`Most ocular medications may be administered topically
`in order to treat surface as well as intraocular disorders. This
`route is often preferred for the management of various
`pathological diseases that affect the anterior chamber of the
`eye,
`for
`two main reasons:
`it
`is more conveniently
`administered and provides a higher ratio of ocular to
`systemic drug levels. To be administered topically and to
`achieve the necessary patient compliance, CsA must present
`a good local tolerance. Topical CsA in olive oil solution
`induces a burning sensation and an irritative effect on the
`conjunctiva. These side effects have been attributed to the
`vehicles used [25]. Patients did not complain of such
`disorders after application of a 2% w/w CsA ointment, and
`ocular examination revealed no significant lesions [26]. A
`recent study [27] showed that formulations of CsA in peanut
`oil were non-toxic to rabbit eyes. However, an unusual
`corneal deposit [28] was reported in a patient after 5 days of
`topical use of a 1% w/v CsA olive oil eyedrop. This deposit
`was probably due to the precipitation of CsA on the corneal
`surface. No long-term study of surface toxicity is yet
`available.
`
`New developments in the topical delivery of CsA can be
`divided in two general areas of research: new delivery
`systems (solutions, ointments, colloidal carriers and drug-
`impregnated contact lenses) and chemical modifications of
`the drug (prodrugs).
`
`2.1. Solutions and ointments
`
`2.1.1. Oil solutions and ointments
`Several vegetable oils such as arachis [29], castor [30],
`olive [7] and peanut oils [31] have been used to solubilize
`CsA. Petrolatum oils (Cremophor [32]) and ointments
`([23,33] have also been investigated as CsA vehicles for
`topical eye administration. Some authors [24,34] have
`reported that such formulations could achieve, after topical
`administration, therapeutic levels in ocular tissues: Kaswan
`[24] reported concentrations of 4 mg/g in the cornea and 60
`ng/g in the iris 2 h after application. On the other hand, a
`majority of authors [7,17,30,35] have reported none or
`negligible intraocular penetration. Furthermore, oils are
`known to be poorly tolerated by the eye and are therefore
`rapidly evacuated from the ocular surface. Due to its
`lipophilicity [15], CsA has a greater affinity for the vehicle
`than for the cornea, providing a low local availability. Also,
`these vegetable oils may present problems of stability such
`as rancidity [36]. Despite these drawbacks, olive oil is still
`tested in the prevention of corneal graft rejection [13] and is
`still the most frequent reference vehicle cited. A marketed
`ointment formulation for veterinary use (Optimmunew,
`0.2% Cyclosporine USP Ophthalmic Ointment, Schering
`Plough, Welwyn, Herts, UK) is available for the treatment
`of keratoconjunctivitis sicca and ocular surface inflamma-
`tory diseases in dogs [37]. This formulation has not reached
`the human field mainly because of its poor acceptability by
`patients. Tolerance in the veterinary area is evaluated by
`tests (Draize, slit lamp) that do not take into account blurred
`vision and patient discomfort; very important criteria in the
`human field.
`
`2.1.2. Aqueous solutions
`Attempts have been made to improve the solubility of
`CsA in water by complexation of CsA with cyclodextrins or
`penetrations enhancers.
`
`2.1.2.1. Cyclodextrins. Cyclodextrins are complex sugars of
`cyclo-malto-hexose type, exhibiting a lipophilic center
`hidden by an external hydrophilic layer [38]. These
`physicochemical characteristics enable cyclodextrins to
`combine with lipophilic molecules and increase their
`water solubility. CsA combined to a-cyclodextrin was
`solubilized up to 750 mg/ml
`in water [39], which is
`approximately 100-fold higher than for CsA alone. Four
`different formulations were tested and the optimal concen-
`tration for maximum corneal permeability and lowest
`toxicity was found to be 0.025% w/v CsA in 40 mg/ml
`a-cyclodextrin solution. After the application of one drop
`
`2
`
`

`

`ARTICLE IN PRESS
`
`F. Lallemand et al. / European Journal of Pharmaceutics and Biopharmaceutics xx (2003) xxx–xxx
`
`3
`
`every 2 h four times a day on the rabbit eye, this solution
`achieved concentrations of 4.1 ^ 0.4 mg/g in the cornea,
`which was five to 10 times higher than those obtained with a
`10% w/w CsA ointment, and above therapeutic levels. This
`study was confirmed by Cheeks [33], who showed on
`excised rabbit corneas that CsA bound to cyclodextrins
`resulted in higher corneal penetration than an application of
`corn oil solutions. This formulation resulted, however, in a
`very small reservoir effect in the cornea, because of the low
`intrinsic quantity of drug in the formulation and the short
`residence time on the eye surface.
`
`2.1.2.2. Penetration enhancers. Penetration enhancers are
`chemicals that can help solubilize CsA and transiently
`modify the corneal epithelium to promote drug penetration
`through the cornea. Azonew (laurocapram) [40] was used as
`a CsA solvent in order to solubilize the drug and improve its
`delivery to the eye. Clinically significant concentrations of
`CsA were measured in the grafted corneas of rabbits but
`little or no drug was found either in the aqueous humor or in
`w
`resulted in
`the blood of the treated animals. CsA in Azone
`suppression of the severity and incidence of graft rejection.
`This penetration enhancer has, since, been shown to induce
`cytotoxicity on corneal epithelium [41].
`The effect of three other penetration enhancers on the
`transcorneal permeation of CsA was evaluated [42]. Flux
`rates of radiolabeled-CsA across human excised corneas
`were measured in the presence and absence of aqueous
`solutions of benzalkonium chloride (0.01%), dimethyl-
`sulfoxide (DMSO) (20%) or Cremophor (10% and 20%)
`(concentrations expressed in w/v). Cremophor and benzal-
`konium significantly increased flux rates of CsA across
`cornea (Fig. 1) while no change was observed with DMSO.
`Benzalkonium presented a very good tolerance at
`the
`concentration used in eye drops as preservative (0.01% w/v)
`[43], but induced ocular irritation at higher concentration
`(1% w/v) [44]. The topical application of Cremophor has
`been associated with changes of corneal surface structure
`[45] while severe anaphylactic hypersensitivity reactions,
`hyperlipidemia, abnormal lipoprotein patterns, aggregation
`of erythrocytes and peripheral neuropathy were observed
`
`Fig. 1. Influence of benzalkonium (0.01%) on the ex vivo flux values of
`CsA (B) across excised fresh (K) and frozen (A) human corneas after
`topical application of CsA aqueous solution [43].
`
`after systemic absorption [46]. The use of penetration
`enhancers represents a potentially interesting approach, but
`with the serious limitation of the low tolerance of these
`molecules that act by modifying the corneal properties
`(mostly a disruption of the epithelial cell layer of the
`cornea).
`
`2.2. Colloidal carriers
`
`Colloidal carriers are small particles of 100 – 400 nm in
`diameter, suspended in an aqueous solution. Calvo [47] has
`shown that colloidal particles were specifically taken up by
`the epithelial cells of the cornea by endocytosis. The cornea
`then acts as a reservoir, releasing the drug to the surrounding
`tissues. These carriers represent a means of delivering
`lipophilic drugs into hydrophilic tissues. They include
`micelles, emulsions, nanoparticles, nanocapsules and
`liposomes.
`
`2.2.1. Micelles
`CsA was solubilized at 0.1% w/v in isotonic and neutral
`aqueous solution by micelles of the non-ionic surfactant,
`polyoxyl 40 stearate, at a concentration of 2% w/v [48].
`After a single administration on the rabbit eye (50 ml), this
`suspension provided a 60-fold higher concentration in the
`cornea than the 0.1% w/v CsA castor oil control solution.
`Although these results are promising, a certain number of
`points remain to be further investigated. Indeed, polyoxy-
`ethylene stearates are widely used in pharmaceutical
`formulations and cosmetics and are generally regarded as
`essentially non-toxic and non-irritant materials [49], but
`ocular tolerance of the surfactant is not known and has not
`been evaluated in this work. In addition, micelles are often
`unstable and their shelf life must be investigated.
`
`2.2.2. Emulsions
`in the
`Oil-in-water emulsions are particularly useful
`delivery of lipophilic drugs. In vivo data from early studies
`confirmed that emulsions could be effective topical
`ophthalmic drug delivery systems [50], with a potential
`for sustained drug release [51]. With the recent improve-
`ments in aseptic processing, and the availability of new
`well-tolerated emulsifiers (polysorbate-80), emulsion tech-
`nology is currently under evaluation for topical CsA
`delivery. Ding and colleagues have developed a castor oil-
`in water microemulsion [52]. This emulsion, stabilized by
`polysorbate 80, solubilizes up to 0.4% w/w CsA and
`remains stable over 9 months at room temperature. It was
`found to cause only mild discomfort and slight hyperemia
`on the rabbit eyes when applied eight times a day during
`7 days. CsA penetrated into rabbit extraocular tissues
`(cornea, lachrymal glands, conjunctiva) at concentrations
`adequate for local immunosuppression while penetration
`into intraocular tissues was much lower and absorption into
`blood was minimal [53]. These encouraging results allowed
`the formulation to undergo clinical trials of phase II and III
`
`3
`
`

`

`ARTICLE IN PRESS
`
`4
`
`F. Lallemand et al. / European Journal of Pharmaceutics and Biopharmaceutics xx (2003) xxx–xxx
`
`in dry eye disease [23,54]. The phase II trial performed on
`162 patients demonstrated good tolerance of the emulsion
`and significant improvement of ocular signs and symptoms
`of moderate-to-severe dry eye disease [23]. CsA formu-
`lations of 0.05% and 0.1% w/w were selected for evaluation
`in phase III trials. In this pivotal study, RESTASISw
`demonstrated statistically significant and clinically relevant
`increases in Schirmer wetting versus vehicle at 6 months. It
`has received approval (December 2002) from the United
`States Food and Drug Administration (FDA) for RESTA-
`SISw (cyclosporine ophthalmic emulsion, 0.05%) as the
`first and only therapy for patients with keratoconjunctivitis
`sicca whose lack of tear production is presumed to be due to
`ocular inflammation.
`Since epithelial corneal cells exhibit negative charges on
`their surface, Klang [55] hypothesized that a positively
`charged emulsion would interact with the corneal cells and
`prolong residence time on the surface of cornea. As Ding’s
`emulsion [52] was negatively charged, Abdulrazik and
`coworkers [56] made a positively charged emulsion loaded
`with CsA. Positive charges were introduced in the emulsion
`through the insertion of stearylamine (0.12% w/w). Conse-
`quently, the spreading coefficient of this emulsion on the
`cornea was four times higher than that of the negatively
`charged emulsion. It was therefore deduced that
`the
`positively charged submicron emulsion has better wett-
`ability properties on the cornea. After a single dose of the
`positively charged emulsion on the rabbit eye, CsA yielded
`higher maximum concentrations in the conjunctiva and the
`cornea, compared to the emulsion of Ding [52] (Fig. 2).
`Tolerance evaluation and stability tests have to be
`performed but so far the results are encouraging. The
`emulsion should soon be submitted to a phase I clinical trial.
`
`2.2.3. Liposomes
`Liposomes are membrane-like vesicles consisting of
`one or more concentric phospholipid bilayers alternating
`
`aqueous or lipophilic compartments, making them potential
`carriers for lipophilic drugs. Milani [57] applied that
`technology to the ocular delivery of CsA. He obtained a
`40% trapping efficiency of CsA into such vesicles. The
`formulation was tested topically on corneal rat allografts: a
`liposome suspension at a CsA concentration of 0.21 mg/ml
`was administered five times daily on grafted corneas. After
`60 days, a 77% rate of graft survival was achieved with the
`CsA loaded liposomes group while only 45% survival rate
`was observed in the olive oil CsA solution control group
`(Fig. 3). As serum levels were undetectable, the authors
`concluded that the graft was reached only by the topical
`route. However, the potential of liposomes as a topical CsA
`delivery system remains limited because of their short half-
`life on the corneal surface and relatively poor stability.
`A charge-inducing agent
`like stearylamine could be
`introduced in order to improve intraocular penetration and
`drug availability, as shown by Law [58]. Furthermore, large-
`scale manufacture of sterile liposomes is expensive and
`technically challenging, which make liposomes secondary
`candidates for CsA delivery. Subconjunctival and intra-
`ocular injections of CsA loaded liposomes have also been
`investigated (see Section 3).
`
`2.2.4. Nanoparticles
`Nanoparticles, primarily developed for i.v. adminis-
`tration, have demonstrated promising results over the last 10
`years in ophthalmology. These systems are able to
`encapsulate and protect
`the drug against chemical and
`enzymatic degradation, improve tolerance, increase corneal
`uptake and intraocular half-lives. Three main studies have
`been undertaken to evaluate aqueous suspensions of CsA
`loaded nanoparticles.
`Calvo and coworkers [59] have made nanocapsules
`composed of an oily phase (Mygliolw) surrounded by a
`poly-e-caprolactone (PCL) coat. CsA was loaded in the oil
`
`Fig. 2. Concentrations (ng/g or ml) of CsA in different ocular tissues, 60
`min after topical application of the positively charged (B) and negatively
`charged (A) emulsions containing CsA on the rabbit eye [57].
`
`Fig. 3. Percentage of non-rejected grafted corneas after topical treatment by
`CsA-bounded liposomes (A), CsA in olive oil (W) and empty liposomes (K)
`on a rat model and comparison to controls non-treated grafts (X) and
`syngeneic grafts (B) over 60 days [58].
`
`4
`
`

`

`ARTICLE IN PRESS
`
`F. Lallemand et al. / European Journal of Pharmaceutics and Biopharmaceutics xx (2003) xxx–xxx
`
`5
`
`at a concentration of up to 50% (w/w CsA/PCL ratio) to give
`a 10 mg/ml CsA concentration in the final formulation.
`After topical administration, these capsules were taken up
`by corneal epithelial cells [60] and achieved corneal levels
`of CsA that were five times higher than a 10 mg/ml CsA oily
`solution. The delivery system was well tolerated but did not
`provide significant CsA levels at the ocular mucosa for an
`extended period of time. Consequently, this formulation
`failed to prolong corneal graft survival in an experimental
`rat model [61]. Moreover, PCL nanocapsules are degraded
`after autoclaving and g-irradiation [62] and thus must be
`sterilized by aseptic filtration. However, since these
`nanocapsules have a mean size in the range of
`210 – 270 nm the only practical alternative is a totally
`sterile manufacturing process. Although PCL nanoparticles
`did not succeed in prolonging corneal graft survival they
`may be useful in the treatment of extraocular diseases, as the
`cornea would represent a CsA reservoir.
`The ex vivo corneal absorption of CsA loaded
`polyisobutylcyanoacrylate (PACA) nanoparticles and
`nanoparticles in poly(acrylic) acid (carbopol) gel was
`evaluated in bovine corneas [63]. The authors found that
`CsA concentrations in corneas were significantly higher
`with nanoparticles in gel than nanoparticles alone and CsA
`olive oil solution. However, one limitation of the ex vivo
`model is that it does not take into account tear wash and
`lachrymal drainage. Further characterization of nanoparti-
`cles (encapsulation efficiency, size and zeta potential,
`release rates studies) would help an understanding of the
`underlying physiological processes involved in transcorneal
`absorption. PACA nanoparticles are known to penetrate into
`the outer layers of the corneal epithelium causing a
`disruption of the cell membranes [64]. In vivo tolerance
`of these CsA loaded nanoparticles should be investigated.
`Chitosan is a biopolymer obtained by deacetylation of
`chitin that is extracted from crab shells. This polymer is
`positively charged and biodegradable. These properties
`make it a good candidate for ocular delivery. Recently, an
`innovative colloidal drug carrier, made by ionic gelation of
`chitosan, has been described [65]. These nanoparticles
`encapsulated CsA at levels up to 9% of their weight. After
`four applications of 10 ml, the formulation achieved high
`concentrations in vivo (rabbit model) in external ocular
`tissues (cornea and conjunctiva) from 2 to 48 h after
`application (Fig. 4), while other tissues (aqueous humor, iris
`ciliary body, and blood) presented negligible CsA levels.
`These relatively large CsA concentrations in the periocular
`tissues are explained by the corneal and conjunctival surface
`retention due to the positive charges of chitosan. It should be
`noted that no Draize or tolerance tests have been performed
`to evaluate this formulation. However, Felt [66] demon-
`strated excellent tolerance of a topically applied chitosan
`gel. Since steam and dry heat sterilization affect physical
`properties of chitosan [67], and since sterile filtration is not
`possible for such nanoparticles (mean size 283 ^ 24 nm),
`g-sterilization should be investigated for these carriers.
`
`Fig. 4. CsA concentrations (ng/g) in the cornea and conjunctiva after topical
`application of CsA loaded chitosan nanoparticles (B) and control
`formulations over 48 h in a rabbit model (* statistically significant
`differences) [66].
`
`Since, chitosan is a polymer of natural origin, batch to batch
`heterogeneity, with respect
`to molecular weight may
`complicate manufacturing.
`The nanoparticle approach is not yet completely
`satisfactory as the precorneal clearance is still too rapid.
`Of the three CsA colloidal carriers described, the most
`promising, so far, is the chitosan carrier mainly because of
`the therapeutic levels achieved in periocular tissues and its
`good tolerance.
`
`2.3. Solid formulations
`
`Since most solutions are eliminated from the ocular
`surface within a few minutes by normal tear turnover and
`lachrymal drainage, solid systems have been developed in
`order to enhance contact
`time of the drug with the
`extraocular tissues. These include collagen-based systems.
`
`2.3.1. Collagen shields
`This approach consists of the application on the cornea of
`a collagen shield loaded with CsA. Reidy and coworkers
`
`5
`
`

`

`ARTICLE IN PRESS
`
`6
`
`F. Lallemand et al. / European Journal of Pharmaceutics and Biopharmaceutics xx (2003) xxx–xxx
`
`[68] tested such a shield loaded with 4 mg CsA on rabbit
`cornea to enhance drug penetration into the anterior
`chamber by increasing contact time on the cornea. These
`shields are directly applied on the rabbit cornea, and the
`eyelids are kept closed by a piece of tape to avoid rejection
`of the system. The device achieved therapeutic concen-
`trations in the cornea and aqueous humor from 2 to 8 h after
`application with a maximum peak at 4 h; no CsA was
`detected in the blood. Both the corneal and aqueous humor
`concentrations of CsA achieved with the shield were 10-fold
`higher than those obtained with topical CsA-olive oil drops
`and over a time period which was twice longer than the
`control. Collagen shields are useful as drug delivery systems
`and also as bandages after corneal surgery. However,
`several drawbacks of the system must be kept in mind. The
`shield may blur the vision. It is also applied dry on the eye, a
`maneuver that may be poorly tolerated especially in some
`pathologies such as dry eye. Also, the collagen matrix is
`disaggregated by tear fluid and after a few hours fragments
`separate from the main part and may be expulsed. There-
`fore, the collagen shield requires a delicate and experienced
`handling in order not to be adversely applied to the ocular
`surface. Consequently, this device may be difficult for self –
`administration by patients.
`
`2.3.2. Liposomes loaded with CsA incorporated in collagen
`shields
`Topical CsA liposomes are able to enhance corneal graft
`survival [57] but have a too short corneal retention time.
`Combining the positive features of collagen shields with the
`advantages of liposomes may provide a synergistic action
`on the eye. Pleyer [69] showed that such a device was able
`to significantly improve CsA aqueous humor concentration
`at 1 and 3 h when compared to CsA loaded liposomes. But,
`this device did not achieve the aim of prolonging corneal
`and aqueous humor levels over 6 h. In addition, this system
`is complex to manufacture and retains the already
`mentioned drawbacks of the collagen shields.
`
`2.3.3. Collagen particles
`To overcome the disadvantages of shields but maintain
`the benefits of collagen devices, small collagen particles
`loaded with CsA were manufactured and suspended in a
`methylcellulose hydrogel. The particle size is relatively
`
`large: 2 mm2 ^ 0.1 mm2 £ 0.1 mm. A volume of the
`
`suspension containing 0.5 mg CsA was administered on the
`cornea of rabbits. The particles were more effective than
`corn oil
`in delivering CsA to the cornea and anterior
`chamber. Higher concentrations of CsA were found in both
`the cornea and aqueous humor of eyes treated with collagen
`particles. Kinetics of drug penetration were different for
`drug in the collagen vehicle compared to the corn oil. The
`particles produced a CsA peak concentration at 4 h and
`maintained a high level until 8 h (Fig. 5) while corn oil gave
`a peak at 1 h that almost disappeared at 8 h [70]. This
`formulation improved significantly corneal allograft
`
`Fig. 5. Concentrations curves of CsA over 24 h in the cornea and aqueous
`humor after topical application on the rabbit eye of different CsA
`formulations: collagen particles (B), collagen shield (O) and corn oil
`solution (B) [72].
`
`survival compared to corn oil [71] but showed no significant
`improvement compared to the collagen shield. The hydrogel
`exerts a viscosifying effect and enhances retention time on
`the cornea. The main advantage of collagen particles is the
`ease with which the formulation can be applied to the ocular
`surface and its possibility for self-administration. Whether
`the sustained release effect
`is due to the collagen
`formulation or to the methylcellulose hydrogel remains to
`be determined. Kinetic studies comparing CsA collagen
`particles alone and CsA suspended in the methylcellulose
`hydrogel would be very welcome. Tolerance evaluation is
`not mentioned and should be evaluated since the particles
`are quite large.
`
`2.4. Chemical modifications of the molecule: prodrug
`approach
`
`Prodrugs are defined as pharmacologically inactive
`derivatives of drug that are chemically or enzymatically
`converted to the active drug. Such molecules are used to
`modify the physico-chemical properties of the drug or its in
`vivo behavior. To improve its water solubility, a water-
`solubilizing moiety was grafted on a free hydroxyl function
`of CsA [72]. The resulting prodrug could be solubilized in
`isotonic and neutral solutions at a concentration equivalent to
`
`6
`
`

`

`ARTICLE IN PRESS
`
`F. Lallemand et al. / European Journal of Pharmaceutics and Biopharmaceutics xx (2003) xxx–xxx
`
`7
`
`1% CsA w/v that is approximately 10,000 times the solubility
`of CsA alone in water. After topical administration of 25 ml
`of this solution in the rabbit eye, the CsA prodrug is cleaved
`in vivo by tear enzymes and releases free CsA. This resulted
`in CsA concentrations in tears above experimental thera-
`peutic levels for up to 20 min. Objective and subjective
`animal (rabbit) tolerance tests showed an improved tolerance
`over that of classical CsA-in-oil solution. This prodrug is a
`promising candidate in the topical treatment of dry eye
`disease and corneal graft rejection.
`Most of the topical delivery systems discussed here
`did not succeed in achieving aqueous humor therapeutic
`levels but corneal and conjunctival levels were sufficient
`to suppress T cell activation. This route will be useful
`for extraocular diseases. As all devices yielded thera-
`peutic levels, choice between systems will be influenced
`by the degree of tolerance and ease of administration and
`manufacture. Altogether,
`three topical
`systems are
`noteworthy: chitosan nanoparticles, positively charged
`emulsions and CsA prodrugs as they are innovative, well
`tolerated and result
`in therapeutic concentrations
`in
`extraocular tissues in animal models.
`Recently, new non-aqueous solvents were proposed
`for the topical delivery of lipophilic drugs. Perfluorocar-
`bons or fluorinated silicone liquids are chemically and
`biologically inert compounds, and present
`low surface
`tension, excellent spreading characteristics and close-to-
`water refractive indices [73]. These solvents could be an
`alternative to complex topical drug delivery systems.
`The insert approach has not yet been proposed to deliver
`CsA. Inserts are composed of a polymeric support contain-
`ing a drug. They are usually placed in the lower fornix and
`may increase contact time between the preparation and
`the conjunctival tissue to ensure a sustained release. As
`lipophilic drugs can be incorporated in inserts [74], CsA
`could be a good candidate for this delivery system.
`
`3. Other administration sites
`
`The other main routes of administration for ocular
`therapeutics include the subconjunctival, intraocular and
`systemic pathways. They are employed when drugs are not
`absorbed by the topical route or when the vitreous humor is
`the target. We can also briefly mention retrobulbar injection
`of CsA [75] that was reported to enhance effect on transplant
`survival, and direct injections of CsA into the anterior
`chamber [76].
`
`3.1. Subconjunctival administration
`
`The subconjunctival route is an alternative to the topical
`and intraocular delivery of CsA. Following subconjunctival
`injections, the administered drug passes through the sclera
`and into the eye by simple diffusion. Before the injection or
`implantation, the eye is anaesthetized by a topical local
`
`anesthetic. Microspheres and implants have been developed
`and tested after subconjunctival administration.
`
`3.1.1. Microspheres
`In order to maintain high levels of CsA in the cornea and
`aqueous humor, Harper [77] employed microspheres made
`of 50:50 D,L-lactide/glycolide copolymer (PLGA) and
`loaded with CsA. Experimental therapeutic concentrations
`were reached in the cornea and aqueous humor 6 h after
`subconjunctival injection, in a rabbit model, of 0.13 ml of
`the microsphere solution containing a total dose of 2 mg of
`CsA (approx. 15 mg/ml in the final formulation). These
`concentrations were maintained for 2 weeks in the cornea
`after injection. Although these results were encouraging,
`efficacy tests we