Citation
Transdermal transport and intradermal drug targeting using novel chemical delivery systems

Material Information

Title:
Transdermal transport and intradermal drug targeting using novel chemical delivery systems
Creator:
Chikhale, Prashant J., 1962-
Publication Date:
Language:
English
Physical Description:
xvii, 141 leaves : illustrations; 29 cm.

Subjects

Subjects / Keywords:
Research ( mesh )
Administration, Cutaneous ( mesh )
Administration, Topical ( mesh )
Drug Delivery Systems ( mesh )
Antiviral Agents -- administration & dosage ( mesh )
Antineoplastic Agents -- administration & dosage ( mesh )
Membranes ( mesh )
Chromatography ( mesh )
Models, Biological ( mesh )
Oxidation-Reduction ( mesh )
Department of Medicinal Chemistry thesis Ph. D ( mesh )
Medicinal Chemistry thesis, Ph. D
Dissertations, Academic -- College of Pharmacy -- Department of Medicinal Chemistry -- UF ( mesh )
Dissertations, Academic -- Medicinal Chemistry -- UF
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )
Academic theses ( lcgft )
Academic theses ( fast )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1991.
Bibliography:
Includes bibliographical references (leaves 129-140).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Prashant J. Chikhale.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
029721988 ( ALEPH )
51175216 ( OCLC )

Downloads

This item has the following downloads:


Full Text
TRANSDERMAL TRANSPORT AND INTRADERMAL DRUG TARGETING
USING NOVEL CHEMICAL DELIVERY SYSTEMS
By
PRASHANT J. CHIKHALE
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE
UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1991


Copyright 1991
by
Prashant J. Chikhale


To my mother and father


ACKNOWLEDGEMENTS
I am grateful to my thesis advisor, Professor Nicholas S. Bodor, Ph.D.,
D.Sc., for his advise, guidance, and support during my graduate education. By
working in close association with him and his research group I was exposed to
different aspects of pharmaceutical research which I think are critical for the
development and understanding of scientific thoughts and procedures. During my
third and fourth year of graduate study, I worked with Francisco M. Alvarez,
Ph.D., at Schering-Plough Research, Pembroke Pines, Florida, in the area of in
vitro model development. I appreciate the fine support and guidance provided to
me by Francisco during this part of my thesis work at Schering-Plough Research.
My other committee members, James W. Simpkins, Ph.D., Richard H.
Hammer, Ph.D., Hans Schreier, Ph.D., and Kenneth H. Rand, M.D., were very
supportive towards my graduate education and I wish to thank them for their
advice and encouragement.
I would like to thank Drs. Balasingam Radhakrishnan, Vasu
Venkatraghavan, Ede Marvanyos, Jonnalagadda Sastry, Lazio Prokai, and Emy
Wu for their helpful discussions and assistance during my laboratory work.


Due to the ultimate moral and financial support of my mother, Vasundhara
J. Chikhale, my father, Jayant M. Chikhale, and my grandfather, Madhusudhan S.
Gandhi, I was able to undertake this task of graduate education and therefore they
deserve equal credit for this dissertation.
Last, but not the least, I wish to acknowledge Elsbeth Brunt, without whose
total support and mention this dissertation is not complete.
v


TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS iv
LIST OF TABLES viii
LIST OF FIGURES x
KEY TO ABBREVIATIONS xiv
ABSTRACT xvi
CHAPTERS
I INTRODUCTION 1
Structure of the Skin 1
Structure and Function of the Permeability Barrier of the Skin 4
Protein and Lipid Domains in the Human Stratum Comeum 6
Localization of the Skin-Barrier to Drug-Transport 8
Relationship between Lipid Structure and Barrier Function 9
Drug-Delivery to the Skin via Pro-drugs 14
Drug Metabolism in the Skin 17
Methods for Estimating Transport of Drugs Into and Through
the Skin 22
II SIGNIFICANCE OF THE STUDY 29
Delivery of Acyclovir To the Skin 29
Delivery of 5-Fluorouracil To the Skin 45
Development of the IAM.PC Column 47
Specific Aims 54
ffl EXPERIMENTAL 55
Materials 55
Methods 56
Synthesis of l-methylpyridine-3-carboxylic acid
anhydride diiodide [1] 56


Synthesis of 1-methyl-3-[(9'-guanylmethoxyethoxy)
carbonyl]pyridinium iodide (AQ+) [2] 56
Synthesis of 1-methyl-3-[(9'-guanylmethoxyethoxy)
carbonyl]-1,4-dihydropyridine [3] 57
Synthesis of lipoic-acia ester of acyclovir [4] 58
Synthesis of ethyl ester of 4-isocyanobutyric acid [5] 60
Synthesis of l-(3-ethoxycarbonylpropylcarbamoyl)
-5-fluorouracil [6] 61
Synthesis of l-(3-carboxypropylcarbamoyl)
-5-fluorouracil [7] 62
Synthesis of D,L-cc-lipol [8] 62
Synthesis of lipolyl ester of l-(3-carboxypropyl-
carbamoyl)-5-fluorouracil [9] 63
HPLC Analysis 64
Preparation of Skin Membranes 66
Preparation of Donor Solutions 67
Diffusion Experiments 67
Skin-Content of the Drug 68
Membrane Partition Coefficient Determinations 69
In Vitro Stability in Aqueous Buffer 70
In Vitro Stability in Biological Media 70
IAM.PC Column Chromatography 71
C-18 Column Chromatography 72
IV RESULTS AND DISCUSSION 73
Syntheses 73
Skin Experiments 75
Selection of Delivery Vehicle 75
Skin Penetration and Retention of Acyclovir 77
Chemical Delivery System for Acyclovir Based on Oxidation 80
Chemical Delivery System for Acyclovir Based on Reduction 86
Chemical Delivery System for 5-Fluorouracil 91
Lipophilicity and Partition Coefficients 94
In Vitro Stability 97
Stability of the IAM.PC Column Stationary Phase 100
Selection of Model Solutes and Appropriate Parameters
for Comparison 101
Prediction of n-Alcohol Transport Across Human Skin 102
Permeability of Human Epidermis to Steroids 107
Skin-Permeability of Water-Soluble Drugs 113
Transport of Nucleosides Across Human Epidermal Membrane 120
V SUMMARY AND CONCLUSIONS 125
REFERENCES 129
BIOGRAPHICAL SKETCH 141


LIST OF TABLES
Table Page
1-1 Composition of stratum comeum lipids (44,50) 10
1-2 Composition of mammalian epidermal lipids (57) 13
1-3 Biotransformation reactions by human skin (77) 20
2-1 Lipophilicity of the (alkylaminomethyl)benzoate ester
prodrugs of acyclovir and their susceptibility to
enzymatic hydrolysis in human plasma (141) 37
4-1 Cumulative amounts of acyclovir in the receiver after
application of saturated solution of acyclovir
(33.5 (imol/ml) in propylene glycol to the
hairless-mouse skin in vitro at 32C 77
4-2 Amounts of acyclovir in the skin after application of its
saturated solution (33.5 pmol/ml) in propylene
glycol to the hairless-mouse skin in vitro at 35C 79
4-3 Amounts of acyclovir and AQ+ in the skin and receiver
(cummulative) after application of saturated solution of
acyclovir (33.5 pnol/ml) or A-CDS (45.3 |j.mol/ml) in
propylene glycol to the hairless-mouse skin in vitro
at 320C 81
4-4 Amounts of acyclovir and AQ+ in the skin (at 48 hours) and
receiver (cummulative) after application of saturated
solution of A-CDS (45.3 pmol/ml) in propylene glycol
to the hairless-mouse skin in vitro at 320C for 6 hours.
Donor was removed at 6 hours and skin was extracted
at 48 hours 85
4-5 Amounts of acyclovir and AQ+ in the skin (at 48 hours) and
receiver (cummulative) after application of saturated
solution of A-CDS (45.3 (imol/ml) in propylene glycol
to the hairless-mouse skin in vitro at 32C 85
V i i i


4-6 Amounts of acyclovir and A-LipS2 in the skin and receiver
(cummulative) after application of saturated solution of
acyclovir (35.6 umol/ml) or A-LipS2 (6.0 |imol/ml) in
propylene glycol to the hairless-mouse skin in vitro
at 320C 87
4-7 Amounts of 5-FU in the skin and receiver (cummulative)
after application of saturated solution of 5-FU
(350 |imol/ml) or 5-FU-LipS2 (84 p.mol/ml) in
propylene glycol to the shaved guinea-pig skin
in vip-Q at 320C
92
4-8
Lipophilicity and skin-membrane partition coefficients of
acyclovir and A-CDS from their dilute solutions
in propvlene glvcol in vitro at 320C
95
4-9
Partitioning of acyclovir and A-LipS2 into human epidermal
membrane from their dilute water solutions in vitro
at 320C
96
4-10
Stability of A-CDS, AQ+, A-LipS2, CPCFU, and 5-FU-LipS2,
in vitro at 370C
99
4-11
Permeabilitv of human epidermis to steroids in vitro at 320C
and their relative retention in the IAM.PC and C-18
HPLC columns
108
4-12
Transport of nucleosides across heat-separated human
epidermal membrane in vitro at §20C and their
relative retention in the IAM.PC.MG column,
122


LIST OF FIGURES
Figure Page
1-1 Schematic representation of the human skin (2) 2
1-2 Dual functions of the epidermal lamellar bod)'. Summary of
lipid biochemical and enzyme biochemical studies on
isolated epidermal lamellar bodies (33) 7
1-3 The ceramides of the human stratum comeum (52) 12
1-4 Changes in lipid composition in the stratum comeum (57) 13
1-5 Schematic representation of the use of prodrugs in topical
therapy (63) 14
1-6 Schematic representation of the skin as a metabolic barrier
(E = enzyme)(77) 18
1-7 Two-compartment, side-by-side, in vitro diffusion cells 23
1-8 Equations used for calculating skin-permeability (16) 24
2-1 Chemical structure of acyclovir 30
2-2 Cellular activation of acyclovir 31
2-3 Prodrugs of acyclovir with improved oral absorption (134,135) 34
2-4 Water-soluble ester prodrugs of acyclovir (139) 35
2-5 (Alkylaminomethyl) benzoate ester prodrugs of acyclovir (141) 36
2-6 Delivery of steroid hormones to the skin using 3-spiro-
thiazolidine derivatives (145) 40
2-7 Delivery of acyclovir to the skin using A-CDS 42
2-8 Delivery of acyclovir to the skin using A-LipS2 44
x


2-9 Working hypothesis mechanism for lipoamide
dehydrogenase (153) 45
2-10 Chemical structure of 5-Fluorouracil 46
2-11 Delivery of 5-fluorouracil to the skin using 5-FU-LipS2 48
2-12 Schematic representation of a cell-membrane 49
2-13 The IAM.PC column stationary phase (166) 50
2-14 Synthesis of nucleosil-lecithin (the IAM.PC column
stationary phase) (165) 51
3-1 Reaction scheme for the synthesis of A-CDS 59
3-2 Reaction scheme for the synthesis of A-LipS2 60
3-3 Reaction scheme for the synthesis of 5-FU-LipS2 65
4-1 Diffusion of acyclovir across freshly excised hairless-
mouse skin after application of its saturated
solution (33.5 umol/ml) in propylene glycol
in vitro at 320C 78
4-2 Dermal delivery of acyclovir per unit dose using
acyclovir or A-CDS as a function of time 82
4-3 Dermal delivery of acyclovir per unit dose using
acyclovir or A-LipS2 as a function of time 88
4-4 Dermal delivery of 5-FU per unit dose using
5-FXJ or 5-FU-LipS2 as a function of time 93
4-5 Improvement in delivery of acyclovir (to the hairless-
mouse skin) or 5-FU (to the guinea-pig
skin) using A-CDS, A-LipS2, or 5-FU-LipS2 over
underivatized acyclovir or 5-FU
respectively, as a function of time 94
4-6 Relationship between the stratum comeum membrane
partition coefficients (Km) and its permeability
(Kp) to n-alcohols on a log-log scale 103
4-7 Relationship between the relative retention of
n-alcohols in the C-18 column {K'(C-18)} and
their stratum comeum membrane permeability
coefficients (Kp) on a log-log scale 104
x ¡


4-8 Relationship between the relative retention of
n-alcohols in the IAM.PC column {K'(IAM.PC)}
and their stratum comeum membrane partition
coefficients (Km) on a log-log scale 105
4-9 Relationship between the relative retention of
n-alcohols in the IAM.PC column {K'(IAM.PC)}
and their stratum comeum membrane
permeability coefficients (Kp) on a log-log scale 106
4-10 Chemical structures of the steroids 109
4-11 Relationship between the human epidermal membrane
partition coefficients (Km) and its permeability
(Kp) to the steroids on a log-log scale Ill
4-12 Relationship between relative retention of the steroids
in the C-18 column {K'(C-18)} and their human
epidermal membrane permeability coefficients
(Kp) on a log-log scale 112
4-13 Relationship between relative retention of the steroids
in the IAM.PC column (K'(IAM.PC)} and their
human epidermal membrane partition
coefficients (Km) on a log-log scale 113
4-14 Relationship between relative retention of the steroids
in the IAM.PC column {K'(IAM.PC)} and their
human epidermal membrane permeability
coefficients (Kp) on a log-log scale 114
4-15 Chemical structures of the water-soluble drugs 116
4-16 Relationship between relative retention of the water-
soluble drugs in the IAM.PC column (K'(IAM.PC)}
using ACNfeOH/DPBS lOx (10/10/80) as the
mobile phase and their human epidermal
membrane permeabilities (Kp) on a log-log scale
117
4-17 Relationship between relative retention of the water-
soluble drugs in the IAM.PC column {K'(IAM.PC)}
using MeOH/DPBS lOx (50/50) as the mobile
phase and their human epidermal membrane
permeabilities (Kp) on a log-log scale
.118
4-18
Relationship between relative retention of the water-
soluble drugs in the C-18 column {K'(C-18)} using
MeOH/DPBS lOx (50/50) as the mobile phase and
their human epidermal membrane permeabilities
(Kp) on a log-log scale
x i i
119


4-19 Chemical structures of the nucleosides 121
4-20 Relationship between relative retention of the nucleosides
in the IAM.PC column {K'(IAM.PC)} using
DPBS lOx as the mobile phase and their human
epidermal membrane permeability coefficients
(Kp) on a log-log scale 123
x i i i


KEY TO ABBREVIATIONS
A:
Acyclovir
A-CDS:
1,4-dihydrotrigonelline chemical delivery system for
acyclovir
AQ+:
N-methylnicotinic acid ester of acyclovir
A-LipS2:
Lipoic acid ester of acyclovir
5-FU :
5-Fluorouracil
CPCFU:
Carboxypropylcarbamoyl-5-fluorouracil
5-FU-LipS2:
Lipolyl ester of carboxypropylcarbamoyl-5-fluorouracil
UV:
Ultraviolet
mp:
Melting point
bp:
Boiling point
1HNMR:
Proton Nuclear Magnetic Resonance Spectrum
M.S.:
Mass Spectrum
mol:
Moles
nmol:
Nanomoles
(imol:
Micromoles
mmol:
Millimoles
nm :
Nanometer
mm:
Millimeter
cm:
Centimeter
log :
log 10
XIV


Kobs :
Pseudo first-order rate constant
ti/2:
Half-life
rpm :
Rotations per minute
min:
Minute
h:
Hour
sec:
Second
M:
Moles per liter
pi:
Microliter
X:
Wavelength
r :
Correlation coefficient
IAM.PC :
Immobilized Artificial Membrane
stationary phase
K :
Relative retention or Capacity factor
i.d. :
Internal diameter
Km:
Membrane partition coefficient
Kp:
Permeability coefficient
J:
Steady-state flux
Q:
Donor-cell concentration
5:
Thickness of the skin membrane
Phosphatidyl Choline
xv


Abstract of Dissertation Presented to the Graduate School of the University of
Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of
Philosophy
TRANSDERMAL TRANSPORT AND INTRADERMAL DRUG
TARGETING USING NOVEL CHEMICAL DELIVERY SYSTEMS
By
Prashant J. Chikhale
May 1991
Chairman : Nicholas S. Bodor
Major Department: Medicinal Chemistry
This dissertation examines the feasibility of targeting and localizing
important antiviral agents (like acyclovir) and anticancer agents (like 5-
fluorouracil) specifically to the skin using novel, redox-based chemical targeting
systems. Such approaches should lead to improvement in the effectiveness of
topically administered acyclovir in treating recurrent mucocutaneous herpes
simplex virus infection of type I. Similarly, the basal cell skin carcinomas or
psoriasis can be effectively treated if 5-fluorouracil could be targeted to the intra-
dermal region.
Chemical Delivery Systems (CDS) for acyclovir based on oxidation (the 1,4-
dihydrotrigonelline moiety containing ester; A-CDS) or reduction (the lipoic acid
ester; A-LipS2) in the skin were utilized to enhance the skin-partitioning ability of
acyclovir and use the enzymatic activity of the skin to create metabolic chemical
precursors as reservoirs for the release of acyclovir in the skin.
XVI


Thus, the dermal delivery of acyclovir was improved by 9-fold (p < 0.025)
using A-CDS, and by 37-fold (p < 0.001) using A-LipS2, at 6 hours relative to
underivatized acyclovir, when administered to the hairless-mouse skin, in vitro.
The lipolyl ester of 5-fluorouracil (5-FU-LipS2) also managed to deliver greater 5-
fluorouracil to the hairless-guinea pig skin as compared to underivatized 5-
fluorouracil.
Thus, such drug-targeting CDS could be effectively used to deliver not only
heterocyclic antiviral and anticancer agents, but also other potent therapeutic drugs
and pharmaceuticals 12 the skin.
The dissertation also describes the development of an important, novel, in
vitro, chromatographic model system (the IAM.PC column) containing a cell-
membrane component for studying drug-membrane interactions and predicting
drug-transport across complex biological membrane barriers, particularly the
human skin. It utilizes a covalently bonded lecithin analog on a solid (silica)
stationary support for the purpose of mimicking specific and non-specific
interactions of solutes with cell membranes.
The IAM.PC column was able to demonstrate significant linear relationships
between the relative retention of various solutes on the phospholipid stationary
phase and their permeabilities across the human epidermal membrane, on a log-log
scale. Thus, it could be used as a potential model to study drug-biomembrane
interactions and drug-transport across biological membrane barriers.
XVII


CHAPTER I
INTRODUCTION
The skin on the body of an average adult covers a surface area of
approximately 2 square meters and receives about one-third of the blood
circulating through the body (1). It is one of the most readily accessible organs of
the human body. With a thickness of only a few millimeters (2.97 0.28 mm), the
skin separates the vital organs from the outside environment, serves as a protective
barrier against physical, chemical, or microbial attacks, acts as a thermostat in
maintaining body temperature, plays a role in the regulation of the blood pressure,
and protects the human body against the penetration of ultraviolet rays (2).
Structure of the Skin
Morphologically, the skin is a multilayered organ composed of four major
multilaminate layers: the stratum corneum, the epidermis, the dermis, and the
hypodermis (or the subcutaneous layer) (Figure 1-1) (2).
The hair and an associated vapour layer constitute the outer boundary with
the environment. It is situated on the visible skin surface, which consists of a
lightly shiny lipid film. The quantity of fat ranges between 5-150 pg/cm2 (3)
which corresponds to 0.05-1.5 pm thickness.
1


2
Stratum-
Granulosum
Stratum-
Spinosum
Stratum-
Germinativum
Epidermis
Capillary Network
Sebaceous Gland
Hair Shaft-
Apocrine Sweat
Gland
Hair Follicle
Blood Vessel
Hvpodermis
Figure 1-1 : Schematic representation of the human skin (2).
ti"


3
The next layer is the actual upper surface of the stratum comeum, consisting
of homy lamellae, the comeocytes. They are the non-living cells situated on the
epidermis in the form of many layers of compacted, flattened, dehydrated, and
keratinized cells. They are continuously shed with constant replacement from the
underlying viable epidermal tissue (4). The phenomenon of multiple layering
within this homy layer is regarded as the fundamental factor in inhibiting the
penetration of most substances.
The epidermis and the horny layer are "indented" by hair follicles and
perforated by sweat gland ducts. The epidermal cellular tissue extends along the
follicular canal deep into the dermis, and the sweat gland ducts are lined with their
own endothelial cells, so that practically no dermal surfaces are exposed into which
direct penetration of substances might occur. The intercellular canals may be a
route for percolation of materials through the epidermis (5). These spaces are
represented by acidic glycosaminoglycans which coat the cell-surface and this
intercellular gel (glycocalix) has a high water content which could constitute a
system of canals for permeation. The intercellular spaces represent 15-18 % of the
total epidermal volume (6).
The dermis consists of an aqueous phase as well as structured elements such
as collagen fibers and elastin, both embedded in mucopolysaccharide networks (6).
Fibroblasts, fibrocytes, and histiocytes are embedded in this extensive network of
connective tissue. The connective tissue is slowly regenerated by these cells,
although not completely renewed as the epidermis. The dermis is transversed in


4
various stages by the transport system of the blood and lymph vessels. The system
of capillary vessels extends into the upper dermis, just under the basal membrane,
so that the capillary loops are situated nearer to the surface than the deepest lying
layers of the basalis due to the epidermal undulation. The blood vessel system of
the skin consists of the entering arteries, arterioles, arterial precapillaries, arterial
and venous capillary loops and finally the postcapillaries and venules which
terminate in the skin veins. The nerves also terminate under the epidermis (5).
The dermis is located on the subcutis which is made up of a network of fat
cells. While the collagen fibers of the network form a system of vertical structure
in the female skin, they form additional diagonal cross-striations in the male skin
(7). These fibers arch upwards into the dermis. The subcutis represents a massive
separating layer between the skin and the muscle tissue below.
Structure and Function of the Permeability Barrier of the Skin
"Indeed, the raison d'etre of the epidermis is to make the stratum comeum;
this is its specific biologic mission." (8; p. 387). This process of formation of the
stratum comeum is more correctly called comification rather than keratinization.
In 1877, Fleischer declared that the skin was totally impermeable, a
complete shield against the external world (9). By the turn of the century,
Schwenkenbecker (10) perceived that the skin would admit some substances much
better than others. Later, appreciation for the integrity of the stratum comeum led
to the concept that the full-thickness of this layer is functionally competent (11),


5
which was consistent with the view that it behaved as a homogeneous film (12) (the
"plastic wrap" hypothesis).
Although the importance of lipids for barrier function (13) and the water-
retentive properties (14) of the stratum comeum was well appreciated, it was
Middleton who first noted that it was the organization of lipid into 'shells that
accounted for water retention (15), thus refuting the homogeneous film concept.
Soon thereafter, the existence of separate hydrophilic and hydrophobic pathways
was suggested from physical-chemical observations (16), followed by morphologic
evidence for lipid-protein segregation. Freeze-fracture and thin-section studies
demonstrated lipid bilayers exclusively in the stratum comeum interstices, with the
absence of lipid structures within the comeocyte cytosol (17-19). In addition,
histochemical and cytochemical studies clearly displayed the process of lipid
sequestration to membrane domains (19,20). This can be further deduced from the
ease with which certain nonpolar solvents disperse this tissue into cell suspension,
and the further ability of such solvent extracts to recombine with dispersed cells to
produce a functionally competent tissue (21).
The first direct evidence for lipid localization to intercellular domains was
provided by biochemical analysis of membrane couplets which were prepared
without loss of intercellular lipid. These preparations contained not only multiple
bilayers identical to intact stratum comeum, but they were also lipid-enriched,
accounting for over 80 % of all the lipid in the stratum comeum (22). Moreover,
they displayed both the same lipid distribution of whole stratum comeum (22), and


6
duplicated the x-ray diffraction pattern ascribed to the 'shell' of lipids, previously
thought to fill interfilamentous domains within the comeocyte cytosol (23). One
can assume that the localization of a variety of lipid catabolic enzymes, steroid
sulfatase (24), lipase (25), sphingomyelinase (25), and phospholipases (25), to the
stratum coraeum intercellular domains represents co-localization of hydrolytic
enzymes with their respective lipid substrates in the stratum comeum membrane
domains.
Protein and Lipid Domains in the Human Stratum Comeum
It is now well appreciated that the two-compartment system is formed by the
deposition of epidermal lamellar body contents in intercellular domains at the
granular-comified cell interface (17,18). Subsequent characterization of the lipid
and enzymatic content of lamellar bodies has provided a clear picture of the
molecular events associated with the formation of the stratum corneum
intercellular compartment. Lamellar bodies are enriched in phospholipids (26,
27), free sterols (26), and glycosphingolipids (26), including certain distinctive
acylglycosphingolipid species (28) that may be responsible for the disc-like lipid
bilayers that appear in lamellar bodies (29). Biochemical studies revealed a limited
array of hydrolases (27,30-32), but a striking absence of certain typical lysosomal
enzymes, including P-glucuronidase, galactosidase, and arylsulfatases A and B
(27,32).


7
The lipid catabolic enzymes that have been demonstrated in lamellar bodies
seem ideally suited to bring about the transformation of the polar lipid contents of
lamellar bodies to the nonpolar species that eventually reside in the stratum
comeum (Figure 1-2) (33). However, the regulation, timing, and location of these
degradative events is still not certain. According to some studies (34,35), if all
phospholipid and glycolipid species are completely absent from the stratum
comeum, then degradation must occur very soon after secretion. Yet others have
found some glycolipids and phospholipids in the lower stratum comeum
(20,30,36), suggesting that lipid transformations in the intercellular spaces may be
a more gradual process.
"Pro-Barrier1 Lipids:
Glycolipids. Free Sterols,
Phospholipids
Catabolic Enzymes:
Acid Phosphatase,
Proteases, Lipases,
Glycosidases
Conversion of "Pro-Barrier" Lipids to
Non-Polar Products (Lipases, Glucosidases)-
Glycolipids
Ceramides
FFA
1) Release ol Oesmosomes into Intercellular
Space (Lipases)
2) Degradation of Non-Lipid Intercellular
Species (Acid Phosphatase, Proteases)
Barrier Function
Desquamation
Figure 1-2 : Dual functions of the epidermal lamellar body. Summary of lipid
biochemical and enzyme biochemical studies on isolated epidermal lamellar bodies
(33).


8
In addition to lipid catabolic enzymes, the lamellar body is also enriched in
acid phosphatase (27,31) and proteases (27). Whether these enzymes participate in
the barrier formation is unclear at present, but it is more likely that they participate
in the desquamation process, presumably by degrading extracellular glycoproteins,
desmosomes, etc. (Figure 1-2) (33).
Localization of the Skin-Barrier to Drue-Transport
It was in 1944 that Winsor and Burch first demonstrated the water barrier,
which prevents desiccation of all terrestrial vertebrates, to be localized in the
stratum comeum (37). They sandpapered through human epigastric skin and
observed a sharp rise in the flux of water just as they breached the stratum
comeum. The observation that the stratum comeum represents the barrier to
water loss through the skin was later verified by Blank (38). He used cellophane
tape to strip-off layers of the stratum comeum and observed that water loss did not
increase much until the lowest layers of the stratum comeum were reached.
Therefore, he concluded that the barrier is located at the bottom of the homy layer
(38). These results were subsequently reinterpreted by Scheuplein, who showed
mathematically that Blanks data (38) could be better explained by a model that
regards all layers of the stratum comeum as rate-limiting with equally good
barrier properties (39).
Early solvent extraction experiments indicated that lipids, particularly the
polar lipids in the epidermis, play a vital role in the barrier properties (40,41).


9
Breathnach and co-workers applied the freeze-fracture technique to show that
these lipids form multiple broad bilayers filling the intercellular spaces (42).
These bilayers were recently shown to be present throughout the homy layer
(43,44), which provide the barrier for water diffusion. This was first
demonstrated by Squier, who used horse-radish peroxidase as a probe to monitor
the movement of water in keratinized and non-keratinized epithelia (45). He
observed that the probe penetrated only the most peripheral layers of the stratum
comeum, when applied topically. But, when it was injected subepithelially it
moved upwards freely through the viable epidermis and stopped just at the bottom
of the stratum comeum, where the first intercellular lamellae are located. These
intercellular lipid sheets seem to provide the barrier to water loss through the skin
and limit the penetration of polar solutes from the environment.
Relationship Between Lipid Structure and Barrier Function
The membranes within the stratum comeum do not contain phospholipids or
the usual assortment of fatty acyl chain structures (46-48), unlike most mammalian
cell-membranes (49), which are composed of a variety of phosphoglycerides,
spingomyelin, and cholesterol, that do not constitute an appreciable barrier to
water or small water-soluble molecules (49). The lipid bilayers in the intercellular
lamellae consist of mainly cholesterol, fatty acids, and ceramides (Table 1-1)
(44,50). The free fatty acids include only 7% of unsaturated species and no methyl-
branched components. The most abundant fatty acids are the 22- and 24-carbon


saturated species (44,51). With the exception of ceramide 1, the hydrophobic
chains in the ceramides (Figure 1-3) (52) are almost entirely straight and saturated.
The double bonds in the sphingosine moieties are located at the polar ends of the
ceramide molecules, so that they do not produce kink or perturb the aliphatic
chains. These lipids seem to be ideally suited to form highly ordered, impermeable
membranes and to resist oxidative damage on exposure to air at the skin surface.
Table 1-1 : Composition of stratum comeum lipids (44,50).
Lipid
Pig
Human
Cholesteryl esters
1.7
10.0
Triglycerides
2.8
0.0
Fatty acids
13.1
9.1
Cholesterol
26.0
26.9
Ceramide 1
4.1
3.2
Ceramide 2
16.7
8.9
Ceramide 3
6.9
4.9
Ceramide 4
4.4
6.1
Ceramide 5
4.5
5.7
Ceramide 6
7.6
12.3
Glucosylceramides
1.0
0.0
Cholesteryl sulfate
3.9
1.9
Others
5.7
11.1
There is considerable chain-length diversity among these fatty acids and
ceramides of the stratum comeum, which results in the interdigitation of chains in


the middle of the hydrophobic region. This interaction may be a significant
stabilizing factor in imparting the barrier properties to the stratum comeum.
Ceramide 1 (Figure 1-3) (52) is of special importance to the stratum
comeum barrier function. This acylceramide is one of the principal carriers of
linoleic acid in the epidermis. It is thought to function in the fusion of the disks
extruded from the lamellar granules and in the stabilization of the resulting
multilamellar sheets. It has been postulated that the long co-hydroxyacyl chain of
the acylsphingolipid completely spans one lipid bilayer, while the ester-linked
linoleate is inserted into the adjacent bilayer, thus serving as a molecular rivet in
linking the two membranes together (53,54). The acylceramide and
acylglucosylceramide have been shown to be capable of promoting stacking,
flattening, and fusion of liposomes (55,56).
However, there is a gradient within the stratum comeum itself. Whereas the
stratum compactum still contains significant levels of phospholipids,
glycosphingolipids, and cholesterol sulfate, only the latter persists in the stratum
disjunctum (Table 1-2, Figure 1-4) (57,58). This change in composition
presumably reflects ongoing metabolic activity.


(CH2)i2CH3
NH
l (CH2)29 02C-(CH2)7CH=CH-CH2-CH=CH(CH2)4-CH3
Ceramide 1
0H
(CH2)ir-CH3
NH
(CH2)22 ch3
Ceramide 2
OH
hA^ch2)16_
CH
OH A
^ NH
\( CH2)22ICH3
Ceramide 3
y^-(CH2)12-CH3
0HJs
NH
J-
CHOH(CH2)23CH3
Ceramide 4
OH
(CH2)ir-CH3
NH
\CHOH(CH2)13CH3
Ceramide 5
(CH2)lf-CH3
NH
JCHOH(CH2)21CH3
Ceramide 6
Figure 1-3 : The ceramides of the human stratum comeum (52).


Table 1-2 : Composition of mammalian epidermal lipids (57).
Lipid
Living layers (%)
Stratum comeum (%)
Phospholipids
40
Trace
Sphingolipids
10
35
Cholesterol
15
20
Triglycerides
25
Trace
Fatty acids
5
25
Other
5
10
SG Inner SC Outer SC
LAYER
Figure 1-4 : Changes in lipid composition in the stratum comeum (57).


14
Drug-Delivery to the Skin via Prodrugs
Most pharmaceuticals or drugs are applied to the skin for topical or systemic
absorption (59,60). However, they are absorbed very poorly due to the nature of
the lipid-like barrier of the stratum comeum. A limited number of highly
lipophilic drugs readily partition into the stratum comeum and the rate-limiting
step then transfers to the lower hydrophilic viable layers of the epidermis. The
drug molecule should possess optimal physical-chemical properties, if it has to
demonstrate appreciable penetration across the stratum comeum barrier (61).
The prodrug approach is one method of enhancing the effectiveness of the
drug by improving its penetration into the skin (Figure 1-5) (62,63).
Figure 1-5 : Schematic representation of the use of prodrugs in topical therapy
(63).


15
The fundamental principle governing the improved penetration of
biological membrane barriers by the use of prodrugs is its greater solubility in the
tissue relative to the parent, underivatized drug (62). The use of prodrugs in
dermal delivery has been described (64) to involve the bypass of the membrane
barrier by alteration of the physical-chemical properties of the drug. The concept
(Figure 1-5) (63) emphasizes the diffusion of the prodrug with simultaneous
metabolism. The importance is given to the drug-skin and drug-vehicle
interactions that can be modified by the application of prodrugs (65).
Some classic studies on which to base a rational design for a new prodrug
have examined homologous series of alcohols (66), phenylboronic acids (67),
steroids (68), and salicylate esters (69). These studies form a basis to indicate the
effect of chemical structure on percutaneous absorption, but it is important to
realize that structural modifications may alter both the skin permeability and the
thermodynamic activity of the drug in the vehicle, and these effects cannot be easily
separated.
The most important considerations in the design of a successful prodrug
approach for skin absorption assume that if various promoieties that impart the
desired physical-chemical properties to the drug can be attached to the appropriate
polar, hydrogen-bonding functional groups in the parent drug molecule, then the
attachment should be reversible (70) in vivo, either by chemical or enzymatic
reactions (71) at the appropriate site in the tissue. In addition, if systemic delivery
is not the ultimate goal (for dermal or intradermal targeting), there should be an


1 6
improvement in the delivery of the drug to the target epidermal cells (72), which
may or may not be associated with enhanced transport of the drug across the skin.
However, for systemic delivery, the prodrug approach requires improving
delivery of the drug through the skin.
Various relevant examples that illustrate the use of prodrug systems for
dermal delivery have been based upon esters of acetylsalicylic acid (73),
indomethacin (74) and other non-steroidal antiinflammatory agents. The polar,
nitrogen-containing heterocyclic molecules like 5-fluorouracil, 6-mercaptopurine,
5-fluorocytosine, theophylline have been extensively derivatized into a-
acyloxyalkyl, N-Mannich base, acyl, and a-(bisalkylhetero)alkyl prodrugs for
dermal delivery and were the subject of a recent review (75). It discussed
strategies for the introduction of promoieties onto primary and secondary amines,
amides, imides, hydroxyls, thiols, carbonyls, and carboxylic acid functional
groups. This report (75) also analyzed correlations between solubilities of the
prodrugs and their abilities to deliver the parent molecules across the skin.
Prodrug concepts using local drug-delivery routes specifically to the skin
have been reviewed (65). The concepts for improved dermal delivery to local
target tissues with the use of prodrugs are similar to those for systemic drug
delivery as well as to achieve transdermal (systemic) drug absorption. However,
true site-specificity in drug-delivery to the skin is only possible when the active
drug can be retained by the target site in the microcompartment of the skin.


Thus, the acyloxymethyltheophylline derivatives were observed to deliver
theophylline to the target epidermal cells based upon the results from the inhibition
of epidermal DNA synthesis (75). In addition, the rate of delivery of theophylline
through the skin (given by the in vitro diffusion-cell experiments) correlated with
the delivery of theophylline to the target epidermal cells (75).
However, as outlined above, if the goal is drug-delivery through the skin for
the purpose of systemic absorption, then specific targetting of active sites in the
skin tissue is not desired. Therefore, the physical-chemical properties of the
prodrug that influence its ability to completely penetrate the skin barrier and allow
the release of the parent drug in the systemic microcapillaries becomes an
important prerequisite.
It has been demonstrated conclusively, that if systemic absorption is the goal
of drug-delivery through the skin via prodrugs then increased water and lipid
solubility of the prodrug is necessary (75).
Drug Metabolism in the Skin
The viable epidermal layer, which is below the stratum comeum, is the most
metabolically active layer in the skin (76). Any substance that penetrates the
stratum comeum is subjected to the drug-metabolizing properties of the viable
epidermis (Figure 1-6) (77). This is particularly important to transdermal
delivery because the first-pass effect is now transferred to the skin. The specific
enzyme activity within the skin has been shown to approach and sometimes exceed


1 8
Figure 1-6 : Schematic representation of the skin as a metabolic barrier (E =
enzyme) (77).
that of the corresponding hepatic enzyme (78-81). This metabolic role for the skin
was supported by bioavailability studies with topical glyceryl trinitrate (82) and
efficacy studies with topical cortisol (83). It has been suggested that the rate of
metabolism of benzo[a]pyrene within the skin is the rate-limiting step when
considering percutaneous absorption (84,85).
The potential biotransformation reactions which are known to occur within
the skin are shown in Table 1-3 (77). They are a series of functionalization (phase
I) reactions (oxidations, reductions, hydrolysis) and conjugation (phase II)
reactions (glucuronide and sulfate formation, methylation and glutathione
conjugation).


The enzyme systems in the skin are highly inducible. Cytochrome P-450,
present in the skin at low concentrations, is inducible by topical application of some
agents that also induce hepatic metabolism (86). After 2,3,7,8-tetrachlorodibenzo-
p-dioxin (TCDD) exposure, the activity of aromatic hydrocarbon hydroxylase
increases upto 30-fold and that of 7-ethoxycoumarin deethylase increases 6-fold
(87). Since the skin possesses many of the enzymes that the liver does, it would be
interesting to compare their relative activities. Generally, the activities of the
enzymes in the skin are low compared to that in the liver, typically about 2-6 % of
the hepatic values (87). However, the cutaneous enzyme activities reported were
mostly based on whole skin homogenates. Assuming that these enzymes are located
in the epidermal layer (76,88-90), their true activities range from 80-240 % of
those in the liver (87).
Thus, if a drug diffuses slowly through the epidermis, the skin may serve as
a site of first-pass metabolism. Such metabolism may decrease both the amount of
drug at the local (intradermal) site of action and the amount systemically available.
On the other hand, if absorption is fast, the cutaneous enzymes may become
saturated, in which case a significant amount of the drug may be absorbed into the
systemic circulation without being metabolized (81). Therefore, cutaneous
metabolism can significantly influence drug-delivery into and through the skin.


20
Table 1-3 : Biotransformation reactions by human skin (77).
A. Phase I Reactions
Reaction
Enzymes involved
Substrate
1. Oxidation
1.1. of aliphatic C-atoms
Mixed function oxidase
7,12-Dimethylbenz(a)anthrac-
ene (DMBA)
1.2. of alicyclic C-atoms
Mixed function oxidase
Dehydroepiandrosterone
(DHA) 7aOH-DHA,
7pOH-DHA
1.3. of aromatic rings
Hydroxylases
3,4-Benzopyrene (BP) -
Phenol, Quinone, Dihydrodiol
1.4. of alcohols
Hydroxysteroid dehydro
genases
Cortisol Cortisone
Testosterone A4-Androst-
ene-3,17-dione
np-Estradiol Estrone
1.5. under deamination
MAO
Norepinephrine
1.6. under dealkylation
Deethylase
Demethylase
7-Ethoxycoumarin
Aminopyrine
2. Reduction
2.1. of carbonyl groups
Ketoreductase
Cortisol -* Reichstein's,
epi-E
Progesterone (Allo)preg-
nanediol
Estrone 17P-Estradiol,
5a-DHT, 5a-Androstane-3a,
17p-diol


21
Table 1-3 : Continued
A. Phase I Reactions
Reaction
Enzymes involved
Substrate
2.1. of carbonyl groups
Ketoreductase
5 a-Andros tane-3,17 -dione
-Androsterone, Epiandro-
sterone
2.2. of -C=C-double
bonds
5 a-Reductase
Testosterone 5a-DHT
Progesterone 5a-DHP
3. Hydrolysis
3.1. of ester bonds
Esterases
Diflucortolone-21 -valerate
Betamethasone-21 -valerate
Betamethasone-17-valerate
Flucortin butyl ester
3.2. of epoxides
Epoxidehydratase (EH)
Styrene oxide
B. Phase II Reactions
1. Glucuronide formation
UDPG-transferase
Benz(a)pyrene,
O-aminophenol
2. Sulfate formation
Sulfo-transferase
DHA, A5-Androstene-3(3,17p-
diol
3. Methylation
COMT
Norepinephrine
4. Glutathione-conjugation
Glutathione-S-transferase
Styrene glycol


22
Methods for Estimating Transport of Drugs Into and Through the Skin
In vitro procedures allow the determination of rate of absorption of the
drug molecule directly below the skin membrane, where it is physiologically
improtant. Errors in extrapolating from in vivo rates of urinary excretion are
avoided. For highly toxic compounds, in vitro methodology may be the only way
of obtaining percutaneous absorption data with human skin. The human skin is
unique and no animal model is entirely suitable. The in vitro experiments can be
done with much less effort and in greater numbers because of the simplicity in
methodology.
The three most important considerations in the design of in vitro procedures
are :
(a) The choice of the skin membrane (animal or human) and its preparation.
(b) Appropriately calibrated in vitro diffusion apparatus.
(c) Specific and sensitive analytical method.
Two-chambered diffusion cells (Figure 1-7) have been used since early
times so that a chemical could be applied on one side of the membrane and its rate
of permeation could be obtained from sampling the identical solvent (usually water
or saline) on the other side (68). The donor as well as the receiver side was stirred
to ensure homogeneity of the solutions.
To simulate in vivo situation, one-chambered cells (receptor beneath the
skin) are now commonly used. The investigator can apply the test compound to the


23
SIDE-BI-SIDEDIFFUSION CELL SHOWN ACTUAL SIZE
Figure 1-7 : Two-compartment, side-by-side, in vitro diffusion cells.
skin in a vehicle of choice and also maintain the skin surface at ambient condition of
hydration. Thus, the static cell design of Franz (91) and similar designs have been
widely used (92). The equations that appropriately describe passive drug-transport
across the skin barrier are based upon Fick's laws of diffusion and are defined as
follows in Figure 1-8 (16).


24
1. J = (KmxD/5)Cd = KpxCd
2. D = 52/6xt!ag
3. Km = Cm / Cv
J = Flux of the drug across the membrane (mol/cm2/h).
Km = Partition coefficient of the drug in the membrane.
D = Diffusion coefficient of the drug in the membrane (cm2/h).
Kp = Permeability coefficient of the membrane for the drug (cm/h).
Q= Concentration of the drug in the donor (mol/ml).
8 = Thickness of the membrane (cm),
tiag = Lag-time for diffusion (h).
Cm = Concentration of the drug in the membrane (mol/ml).
Cv = Concentration of the drug in the vehicle (mol/ml).
Figure 1-8 : Equations used for calculating skin-permeability (16).
Since enough human skin for large number of permeability experiments is
usually not available, it is important to study and compare the permeabilities of
human and animal skin. The skin of the rabbit and the mouse is usually most
permeable. The hairless-mouse skin permeability has been reported to be similar
to human skin for some compounds (93,94). However, many apparent
discrepancies between the permeability coefficients of various compounds in
human and animal skin can be explained by the structural differences in the
membranes. Bronaugh et al. (95) have shown that the stratum comeum of the rat
skin is as thick as the human skin. For compounds that penetrate rapidly and do not


25
rely on appendageal diffusion, it may be a good model for human skin, but its many
hair follicles make it a poor model for polar, water-soluble compounds.
Conversely, hairless-mouse skin has a similar hair follicle density, but the stratum
comeum is thinner than the human skin (95) and so it is often more permeable.
The skin of the monkey at the abdominal and ventral forearm testing sites is
sparsely haired, and has been found to be a good model for human skin in
numerous in vivo studies (96,97).
The preparation of the skin membrane is a critical step in the in vitro
experiments. In humans, pigs, rats, and guinea pigs, the dermis is 2-3 mm thick
compared to the epidermis, which is approximately 50-100 mm. If full-thickness
skin is used in the diffusion-cell studies, the thick dermal tissue can present an
artificial barrier, particularly for water-insoluble compounds. Compounds that
are absorbed through the skin in vivo are taken up by blood vessels directly beneath
the epidermis, so they are not required to penetrate the full-thickness of the skin.
Therefore, for relatively hairless-skin, the preparation of an epidermal membrane
by heat separation has been a convenient solution to the problem of skin thickness
(92). Hairy animal skin cannot be separated by this technique because the shafts of
hair leave holes in the epidermis when it is peeled away. In studies with animal skin
that is relatively thin (1 mm or less) such as the mouse, hairless-mouse, and the
rabbit, the preparation of a split-thickness skin using a dermatome is difficult, and
is probably not required.


26
The basic data for in vivo human percutaneous absorption, to which animal
models are compared, were obtained from Feldmann and Maibach (98-100). In
these clinical studies, a specific concentration of radioactive compound (4 mg/cm2)
was applied to a specific anatomical site (the ventral forearm). The area was not
occluded, and the subjects were requested not to wash the area for 24 hours. The
radioactive compounds were applied to the skin in an acetone solution and the
acetone quickly evaporated with a gentle stream of air. Urine was collected for 5
days and assayed for radioactivity. A tracer dose was also given parenterally, and
the percentage of radioactivity in the urine following parenteral administration
was used to correct for the compound that might be excreted by some other route
and for the compound that might be retained within the body.
Comparative in vitro studies show that generally, the skin of common
laboratory animals (mouse, rabbit, rat, and guinea pig) is more permeable than the
skin of man. Skin from the pig and monkey more generally approximates the
permeability of human skin (101,102). Hence, in vitro and in vivo studies with the
pig and monkey skin were observed to correlate well with the in vivo human
percutaneous absorption studies (103,104).
A compound with limited water solubility must be examined carefully when
using in vitro diffusion-cell techniques. This type of substance may seem to
penetrate skin only slightly, when, in fact, the rate-limiting step is not the
penetration into the skin but partitioning from the skin into the aqueous receiver
fluid. Under in vivo conditions, hydrophobic compounds that penetrate the skin


27
are taken up and carried away by the blood in the capillary loops immediately
below the epidermis much earlier than they would appear in the receiver fluid in an
in vitro set-up, specially if the receiver is normal saline or a physiological buffer
solution. It is quite likely that the hydrophobic substance may choose to remain in
the skin rather than partition into the aqueous receiver solution.
The physical-chemical properties of a molecule, in particular its solubility
in oil and water, have been compared with the permeability data in the hope of
finding correlations. Usually, the solubility is expressed as an oil/water ratio (or
its logarithm), and the octanol/water partition coefficients for a number of
compounds have been published by Hansch and Leo (105). Reasonably good
correlations of permeability and oil/water partition coefficients have been obtained
with some homologous series of compounds. If they are applied as dilute solutions
in aqueous vehicles, positive correlations have resulted (106-108). This result is
due, at least in part, to an increased driving force for the drug from the aqueous
vehicle caused by the increase in lipid solubility. Of course, the absolute solubility
in the aqueous vehicle is critical to permeation but is ignored in expressing
solubility data as partition coefficients.
When compounds are applied undiluted, a number of studies have found that
the best correlations have been positive with water solubility (109-111). Under
these conditions, the importance of water solubility of the molecule in promoting
good skin permeation is emphasized rather than its ability to partition from water
into oil.


28
Application of compounds in a saturated solution should, in theory,
overcome the influence of the vehicle on percutaneous absorption. When the
permeation of structurally different compounds was determined from saturated
aqueous solutions, no correlation of skin absorption with mineral oil/water
partition coefficients was observed (16). It is likely that relying on partitioning
data to estimate skin absorption will often be misleading, since other determinants
such as the effect of the vehicle and skin binding are not considered.


CHAPTER H
SIGNIFICANCE OF THE STUDY
The present investigation concentrated on two major aspects pertaining to
transdermal drug-transport and intradermal drug-targeting :
(a) The use of novel redox chemical targeting systems for improving the
delivery of acyclovir (a model antiviral agent) and 5-fluorouracil (a model
anticancer agent) to the skin. In this section, the concept of intradermal
drug-targeting was emphasized to localize the parent drug species in the
skin.
(b) The development and use of the Immobilized Artificial Membrane (IAM)
HPLC column chromatographic model system for studying drug-transport
across human skin as a model biological membrane barrier.
Delivery of Acyclovir To the Skin
Acyclovir (acycloguanosine; 9-(2-hydroxyethoxymethyl)guanine) is an
acyclic, synthetic analog (Figure 2-1) of a naturally occurring purine nucleoside
and exhibits potent in vitro antiviral activity against herpes simplex virus (HSV) of
types I and II (112,113). It is also active against varicella zoster virus (VZV) (114)
and human cytomegalovirus (HCMV) (115). The HSV have the ability to code for
29


30
the specific thymidine kinase, which is capable of phosphorylating acyclovir to a
monophosphate (116,117).
O
Figure 2-1 : Chemical structure of acyclovir.
This capability is essentially absent in uninfected cells and therefore,
acyclovir exhibits high potency and selectivity for HSV infected cells and low
toxicity for the normal host cells (112,114). The monophosphate is then
phosphorylated to the diphosphate via cellular guanylate kinase (GMP) and then to
the triphosphate by other cellular phosphorylating enzymes (112,118). Acyclo-
GTP is a more potent inhibitor of the viral DNA polymerase than of the cellular
DNA polymerase (119). This cellular activation of acyclovir is shown in Figure 2-
2. The viral DNA polymerases use acyclo-GTP as a substrate and incorporate it
into the DNA primer-template to a much greater extent than do the cellular DNA
polymerases (120).


31
Thus, the viral enzyme catalyzes the insertion of the false purine base into
the viral DNA, bringing it to a premature end, since acyclovir does not possess a 3'-
hydroxyl group, which is obligatory for chain elongation.
Acyclovir
HSV coded
Thymidine Kinase
O
GMP
Kinase
O
Acyclo GTP
Cellular Enzymes
Acyclo GDP
Figure 2-2 : Cellular activation of acyclovir.


32
The acyclo-GTP inhibits viral DNA polymerase with a tight binding
constant of 1.6 x 10*9 M (121). Acyclovir is active to a similar extent against HSV I
and II with 50 % inhibition of viral plaque formation at concentrations of 0.06 to
1.8 x 10-6 M for type I and 0.65 to 1.8 x 10-6 M for type II (122). Since two
enzymes are involved in the antiviral activity of acyclovir, resistance can and has
occurred due to either (a) absence of thymidine kinase or its altered sensitivity to
substrates and to the drug (123) or (b) in the ultimate target, the DNA polymerase
(122). Mutants of HSV I without the thymidine kinase are apparently not as
virulent as the wild-type, although those that still have the altered enzyme, are just
as virulent (124).
The topical (local) delivery of acyclovir is an important consideration in the
treatment of cutaneous herpes simplex viral infections of type I. The HSV are
capable of infecting almost any cutaneous site, including the scalp, toes, knees,
elbow, and hand (125). If the patient's recurrences are mild, then topical
formulations are favorable for the treatment of recurrent herpes labialis (125).
Unfortunately, the clinical trials conducted with acyclovir 5 % ointment have
failed to show efficacy (126-128). The modified aqueous cream with 5 %
acyclovir has shown some promise in a recent trial (125), but no definite efficacy.
Gibson et al. (129) had previously demonstrated efficacy in a small prophylactic
cream trial, and Fiddian et al. (130) had originally shown efficacy in an earlier
trial in England.


33
The reports on the ineffectiveness of the drug in animals and in humans in
certain therapeutic situations has been attributed to the major problem of delivery
(131,132) of acyclovir across the stratum comeum to reach the infection site in the
epidermal cells of the skin. Acyclovir has several non-optimal physical-chemical
properties which present barriers to its effective delivery. It has low aqueous (1.2
mg/ml at pH 7.0 and 250Q (133) and low lipid (log P = -1.47; octanol buffer
partition coefficient) (141) solubility. Thus, it shows inadequate skin partitioning
ability which could lead to a large concentration drop from the stratum comeum
surface to the basal layer of the epidermis when the drug is topically applied.
Hence, improved penetration of acyclovir through the stratum comeum and the
epidermal layers (Figure 1-1) has been recommended (132).
The 6-deoxy-6-amino derivative (Figure 2-3) (134,135) of acyclovir has
been studied as a prodrug to improve the oral bioavailability of acyclovir (134). It
is deaminated to acyclovir by adenosine deaminase (135). However, it resulted in
only modest increases in acyclovir plasma levels relative to those achieved by
acyclovir upon oral dosing of rats and dogs (134).
Krenitsky et al. (136) made the 6-deoxyacyclovir (Figure 2-3). This
prodrug was found to be 18 times more water-soluble than acyclovir and is rapidly
oxidized by xanthine oxidase to acyclovir in vivo. After oral administration to rats
and human volunteers, it was rapidly absorbed and resulted into 5-6 times greater
bioavailability than acyclovir (136-138).


34
6-deoxyacyclovir
6-deoxy-6-
aminoacyclovir
Acyclovir
Figure 2-3 : Prodrugs of acyclovir with improved oral absorption (134,135).
The compound is also susceptible to undergo oxidation by aldehyde oxidase
to give the inactive 8-hydroxy-6-deoxyacyclovir (Figure 2-3) (134,135), but this
non-activating oxidation plays only a minor role in comparison to the activating
oxidation by xanthine oxidase to acyclovir (136).


35
Highly water-soluble esters as prodrugs of acyclovir (Figure 2-4) (139)
have been described which may permit administration of large quantities of the
drug for either topical use as eye drops or parenteral administration.
Intramolecular complex formation with acyclovir to improve its water solubility
for developing parenteral formulations have been attempted (140).
1, R = C0CH2NH2.HC1
2, R = C0CH(CH3)NH2.HC1
3, R = C0CH2CH2NH2.HC1
4, R = COCH2CH2COONa
5, R = COCH2N3
6, R = COCH(CH3)NHCOOCH2C6H5
7, R = COCH2CH2NHCOOCH2C6H5
8, R = COCH2CH2COOH
Figure 2-4 : Water-soluble ester prodrugs of acyclovir (139).


36
However, it is unlikely that they would improve the delivery of the antiviral
agent across the membrane barriers since there is no significant improvement in
their lipid-solubility over acyclovir.
Examples of highly water-soluble prodrugs of acyclovir (Figure 2-5) (141)
with high susceptibility to undergo enzyme catalyzed hydrolysis were recently
described (141).
N
>
HN
h.n^n^n
l,v^0R
I R = H
ii
CH2N(CH3)2
III R=C-4 CH2N(CH3)2
V R =
VI R =
yj ch2n(c2h5)2
o _yCH2N(C3H7)2
II /==<
O
II /=>
vn R=c^ y-CH2N(C3H7)2
,CH2N(C4H9)2
vm r =
IX
O /CH2
II /=\
o
R c~\^
x R= c-Qkch2<^0
0 / \ / V
X! R = l^J-cuX^y
Figure 2-5 : (Alkylaminomethyl) benzoate ester prodrugs of acyclovir (141).


37
These prodrugs which are (alkylaminomethyl) benzoate esters of acyclovir
showed increased hydrophilicity (excess of 10 % w/v) and lipophilicity compared
to that of acyclovir (log P = 1.47) as determined by their partitioning between n-
octanol and 0.05 M phosphate buffer, pH 7.4 (Table 2-1) (141).
Table 2-1 : Lipophilicity of the (alkylaminomethyl) benzoate ester prodrugs of
acyclovir and their susceptibility to enzymatic hydrolysis in human plasma (141).
Ester
11/2 in human plasma
(min)
logP
Acyclovir
-1.47
II
7.0
-0.94
III
33
-0.92
IV
7.5
-0.37
V
25
-0.35
VI
0.8
0.60
VII
57
0.60
VIII
2.3
1.50
IX
4.6
-0.04
X
3.7
-0.05
XI
8.5
-0.11
According to Table 2-1, there is no correlation between lipophilicity of the
various synthesized (alkylaminomethyl) benzoate ester prodrugs of acyclovir and
their ability to hydrolyze in human plasma (due to the esterases). However, there
appears to be a significant difference in their susceptibilities towards hydrolysis by
the plasma esterases depending on the position of the alkylaminomethyl group on


38
the benzoate moiety. Structures II, IV, VI, and VIII (Figure 2-5) with the
alkylaminomethyl group at the 3 position on the benzoate group, were observed to
hydrolyze faster (Table 2-1) than the corresponding molecules with identical
substitutions on the 4 position (structures III, V, and VII in Figure 2-5). Thus, the
alkylaminomethyl substitution at position 3 may allow easy approach by the plasma
esterase enzymes to readily catalyze hydrolysis of those substrates whereas, the
corresponding alkylaminomethyl substitution at position 4 may be hindering the
enzyme approach. This indicates the sensitivity of changes in the structural
configuration of these esters to their susceptibility towards hydrolysis by the
plasma esterases.
Thus, to overcome the major drawbacks of acyclovir, namely its
ineffectiveness against recurrent disease (142,143) due to poor human skin
penetration (144), we have been involved in the design, synthesis, and evaluation of
novel chemical targeting systems for this important antiviral agent, which would
(a) carry acyclovir into the skin due to their better partitioning ability, and (b) use
the metabolic ability of the skin to trap the drug-delivery systems for a sustained
release of acyclovir m the skin. The objective is to deliver the drug to and not
through the skin so as to achieve high local skin concentration of acyclovir where
the herpes simplex virus infects the epidermal skin cells.
In our laboratory, it was demonstrated that the spirothiazolidine derivatives
of endogenous steroid hormones like hydrocortisone (73,145), testosterone, and
progesterone (146) could offer unique advantages in delivering the parent drugs


39
specifically to the skin. Natural amino acid cysteine and its derivatives were used
for their synthesis (Figure 2-6) (145) and the compounds were found to be
susceptible towards enzymatic hydrolysis of the likely iminium ion intermediate
(Figure 2-6) (145) formed after spontaneous cleavage of the carbon-sulfur bond
(147).
It was observed that the thiazolidine derivatives of potent steroidal
hormones readily partitioned into the skin, followed by ring-opening to form the
iminium ion intermediate which then binds to the skin as shown in Figure 2-6 (145)
followed by sustained release of the parent hormone in the skin. (73,145,146).
Thus, it was perceived that to achieve high local skin concentration of
acyclovir to improve its clinical effectiveness, it is necessary to design carrier
mechanisms to lock the precursors in the form of chemical reservoirs, from which
there would be a sustained release of the antiviral agent in the skin.
A 1,4-dihydrotrigonelline moiety containing ester of acyclovir, A-CDS,
was previously designed and successfully synthesized (148,149) to serve as a redox
chemical delivery system for improving the delivery of the antiviral agent to the
brain. Two important features of this redox system were noted. The relative
lipophilicity studies based on HPLC elution indicated that the A-CDS was about 30
times (log P = 2.12) more lipophilic than acyclovir (log P = 0.64). Hence, it
demonstrated facile penetration across the blood-brain barrier. In addition, it was
rapidly oxidized to the quaternary metabolite, AQ+, effectively locking this polar
species inside the rat brain, from which slow release of acyclovir in the brain was


40
COAHs
C02C2H5
3-Spirothiazolidine
derivative
SKIN
Parent Steroid
Hormone
Figure 2-6 : Delivery of steroid hormones to the skin using 3-spirothiazolidine
derivatives (145).


41
observed. It was this dramatic separation between the physical-chemical properties
of A-CDS and its oxidized metabolite, AQ +, which led to an increased flux of A-
CDS into the brain, and a decreased efflux of AQ+ from the brain.
Thus, the idea of using A-CDS for the delivery of acyclovir to the skin was
based on the above findings (Figure 2-7). It was thought that this novel approach
towards dermal delivery would not only exploit the favourable dual physical-
chemical properties of the carrier system described, but also the metabolic ability
of the skin, as a biomembrane. Therefore, the possibility of enhancing skin
concentration of AQ+ and consequently that of A was investigated (150).
The principle (Figure 2-7) of delivering acyclovir to the skin is based upon
administering the relatively more lipophilic A-CDS to the skin. After facile
penetration across the stratum comeum, the A-CDS would rapidly oxidize to form
the quaternary metabolite, AQ+, in the skin due to the presence of oxidative
enzymes analogous to the NAD+/NADH redox co-enzyme systems (151,152). AQ+
would then be slowly hydrolyzed by the non-specific esterases to release acyclovir
in the skin. The net result would be accumulation of AQ + from which there would
be a sustained release of the antiviral agent in the skin. The released acyclovir
would then undergo activation leading to the formation of the active triphosphate
form of acyclovir, resulting in the inhibition of HSV DNA polymerase and hence
its replication. Finally, acyclovir and the charged carrier Q+ will be systemically
absorbed and rapidly eliminated due to their polar characteristics.


42
I
ch3
Elimination via systemic circulation
Figure 2-7 : Delivery of acyclovir to the skin using A-CDS.


43
Another carrier system, the lipoic-acid ester of acyclovir (A-LipS2) was
designed. Its delivery to the skin, however, is based upon a principle (Figure 2-8)
similar to that of the spirothiazolidine derivatives previously discussed
(73,145,146). After rapid skin-penetration A-LipS2 is expected to undergo a
reductive disulfide ring-opening by entering the lipoamide-dehydrogenase redox
cycle (Figure 2-8). Due to this interaction with the active site of the enzyme
system, the free thiol is then capable of forming intermolecular disulfide bonds
within the skin. This is followed by the release of the antiviral agent in the skin due
to the hydrolytic action of the non-specific esterases or the lipases (Figure 2-8).
The A-LipS2, after rapid partitioning into the skin is assumed to enter the
lipoamide dehydrogenase redox co-enzyme cycle (Figure 2-9) (153) via reduction
or simple charge transfer to open the intramolecular disulfide bond, followed by
an intermolecular disulfide covalent bond formation in the skin (Figure 2-8).
Acyclovir would then be slowly released in the skin due to the hydrolytic action of
either non-specific esterases or lipases (Figure 2-8).
Thus, both the carrier systems, A-CDS based on oxidation and A-LipS2
based on reduction, were expected to generate metabolic reservoirs for sustained
release of the antiviral agent, acyclovir in the skin, for improving its dermal
delivery.


44
FAST
O
IN PG (SATURATED SOLUTION)
O
ii y\/
(CH^CXK V
SLOW
ESTERASES
SKIN
ACYCLOVIR
k \/s
OH
I
INTRACELLULAR ACTIVATION
I
INHIBITION OF VIRAL
DNA POLYMERASE
LIP (SH)
I
INHIBITION OF VIRAL REPLICATION
SYSTEMIC ELIMINATION
Y
ENTERS THE NORMAL LIPOAMIDE
DEHYDROGENASE CYCLE
+ NAD LIP (S)2 + NADH
+ H+
Figure 2-8 : Delivery of acyclovir to the skin using A-LipS2.


45
Lip(SH)2 + NAD+
Up(S)2 + NADH + H+
Figure 2-9 : Working hypothesis mechanism for lipoamide dehydrogenase (153).
Delivery of 5-Fluorouracil To the Skin
5-Fluorouracil (5-FU) (Figure 2-10) is used as an antitumor agent
(154,155). It belongs to the class of fluorinated pyrimidines whose
antiproliferative activity is due to its nucleotides 5-fluorouridine triphosphate and


46
5-fluorodeoxyuridine monophosphate (154). The conversion of 5-FU to these
cytotoxic anabolites occurs intracellularly. However, when 5-FU is used to treat
solar keratoses and multiple superficial basal cell carcinomas (156) of the skin, it
demonstrates minimal partitioning across the stratum comeum, due to poor lipid
solubility.
o
Figure 2-10 : Chemical structure of 5-Fluorouracil.
The antitumor activity, affinity, and toxicity of 5-FU has been modified by
the introduction of a carbamoyl group (157-159). l-Hexylcarbamoyl-5-FU
(HCFU) was synthesized as a masked form of 5-FU (157) and was found to be more
active than 5-FU against experimental solid tumors in mice following oral
administration. Further, acute, sub-acute, and chronic toxicity tests have shown
that the HCFU is less toxic than FU. Hence, various 1-alkylcarbamoyl derivatives
of 5-FU have been synthesized (157,160) mainly for oral administration and some
of them have been evaluated for their transdermal delivery potential (161).


47
O
0=LnH(CH2)3-CCfc(CH2)5
Lipolyl ester of CPCFU
SKIN
O
F
ESTERASES
0J-NH-(CH2^-cc&(CH2)5
Lipolyl ester of CPCFU skin
covalently bound in the skin
5-FU
O
0=LnH-(CH2)3-CQ2H
CPCFU
HYDROLYSIS
COj
B2N- (CH2)3-C02H
4-Aminobutyric acid
Figure 2-11 : Delivery of 5-Fluorouracil to the skin using 5-FU-LipS2.


48
Therefore, a relatively more lipophilic derivative, the lipolyl ester of
carboxypropylcarbamoyl 5-FU (5-FU-LipS2) was synthesized and evaluated for its
ability to deliver 5-FU to the skin. The principle of delivery (Figure 2-11) being
similar to that of A-LipS2 discussed earlier is outlined below.
Development of the 1AM.PC Column
The use of in vitro models to understand and predict complex, in vivo,
biological phenomena is justified by virtue of their simplicity. If appropriately
designed, they can provide time conserving and inexpensive alternatives to
experiments using animals and humans.
Cell-membranes (Figure 2-12) interact with virtually every type of
biomolecule (162). All solutes including drugs, sugars, amino acids, peptides, etc.,
interact with the components of cell-membranes during their uptake andtransport.
The association of molecules with membranes can be due to either specific or non
specific interactions (163).
Specific interactions similar to those in affinity chromatography, involve
polar interactions between the solute and some part or parts of the membrane. Non
specific binding, on the other hand, requires solubilization of the molecule (like in
partition chromatography) in the hydrophobic environment of the bilayer.
Phospholipid headgroups contribute substantial molecular volume to the
membrane interface because lecithin is the major membrane forming lipid in most
cells.


49
Figure 2-12 : Schematic representation of a cell-membrane.
These lipid headgroups are barriers (164) to transport of some solutes, but can be
significant binding site to others. This type of molecular selectivity which occurs
in biomembranes was the basis for developing the solid chromatographic support
composed of a membrane lipid (165).


50
Thus, the IAM.PC column consisting of a monolayer of an analog of lecithin
(monomyristoyl lysolecithin) as the stationary phase, chemically (covalently)
bonded to silica through propylamine (Figure 2-13) (166) was synthesized (Figure
2-14) (165).
o
7A
o
1 5A
H3cNCH2-qH2
H,C O
o=f£-o
O
H2q^~ qH- h2
o o
I I
o=c c=o
I I
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
_
h2c
ch2
h2c
ch2
h3c
1
c=o
nh2
NH
NH
NH.
ch2
ch2
ch2
CK
ch2
ch2
ch2
CH;
ch2
ch2
ch2
CH;
+ H3q^
H3cNCH2-q(H2
H.C7 O
0=/0~'
h2<\-
o
ch-ch2
o
1
o
o=c
o
II
-o-
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h3c
1
c=o
NH,
NH
NH.
ch2
ch2
ch!
ch2
ch2
ch!
ch2
ch2
CH;
+ H3<\
H3C-N-CH2-qH2
h2c o
0=F-0~
o
H2qch-ch2
o
1
o
1
o=c
o
II
-o-
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h2c
ch2
h3c
1
c=o
nh2
nh2
NH
NH
ch2
ch2
ch2
CK
ch2
ch2
ch2
CK
ch2
ch2
ch2
CK
Silica Surface
Figure 2-13 : The IAM.PC column stationary phase (166)


51
1,12-Dodecanedicar boxy lie
acid + DCC
THF, 25C
18 h
Dodecanedicarboxylic acid
anhydride (cyclic anhydride)
CHC13, 25C Monomyristoyl
DMAP, 48 h lysokdthin
Lecithin-imidazolide
CHC13> 25C, 1-2 h
* Lecithin-COOH
Carbonyldiimidazole
CHCI3,25C Nucleosil-300
18-24 h (7NH2)
Nucleosil-Lecithin
(IAM.PC column stationary phase)
Figure 2-14 : Synthesis of nucleosil-lecithin (the IAM.PC column stationary phase)
(165).
Attempts to mimic or reproduce the cell-membrane environment on a solid
support, such as the IAM column, could lead to a potential in vitro model to study
molecular processes occurring in natural membranes. From a practical standpoint,
it can be industrially used for rapid screening of various pharmacological agents.
Therefore, this dissertation describes efforts towards developing the IAM
column for studying drug-membrane interactions and hence, its applicability in


52
predicting drug-transport across human skin as a model biological membrane
barrier.
The elucidation of transdermal-transport of various drug molecules across
the human epidermal membrane has become an important part of pharmaceutical
science in recent years. This is particularly because many pharmaceutical
industries are interested in delivering potent biologically active drug-molecules
through the skin for their direct availability in the systemic circulation (the central
compartment). Thus, the drugs could be protected from the undesired first-pass
effect due to metabolism by the liver upon oral administration. Therefore, it is
desirable to develop an in vitro model system which would easily predict the
transdermal availability of drugs.
In addition, the epidermal cell-membranes (Figure 1-1) (2) have been
demonstrated to contain significant amount of phospholipids (57,58). Therefore,
the correlation of drug-transport across the human epidermal membrane with the
retention of the molecules in the IAM.PC column could be justified.
Various membrane models and their advantages or disadvantages for skin
penetration studies were described in a review (167) including excised animal and
human skin, artificial (eg. polymeric sheets) and natural (eg. egg shell)
membranes. However, it was concluded that most of these models fall short of
actually mimicking the in vivo percutaneous absorption process.
Cell culture offers the ideal combination of uniformity and reproducibility
of an artificial system while retaining the biological significance of a recently


53
excised piece of skin tissue. Hence, human skin keratinocyte cultures have allowed
formation of a fully differentiated epidermal layer including a morphologically
distinct stratum comeum (168). An alternative approach of similar attractiveness
would be a validated in vivo isolated perfused skin model (169).
The use of the rotating diffusion cell and associated lipid membrane as an in
vitro model system for percutaneous absorption has been documented (170). The
artificial membrane was formed with isopropyl myristate (IPM), a lipid chosen to
be representative of those in the stratum comeum. The transport resistances of this
membrane to 8 model penetrants were contrasted with those for excised human
cadaver skin. The results indicated that, although some degree of correlation
between the two was evident, the predictability of the IPM membrane model could
be improved (170). Therefore, in another study (171) three alternative lipid
models were chosen to mimick the stratum comeum (the epidermal penetration
barrier). These were dipalmitoyl phosphatidylcholine, linoleic acid, and
tetradecane. For the permeants with diverse physico-chemical properties, the
tetradecane membrane appeared to offer the best correlation with human skin
(171).
Although phosphatidylcholine-coated silica was reported (172) as a useful
stationary phase for HPLC determination of partition coefficients between octanol
and water, such physically adsorbed lipids on a silica support would result into
decreased stability of the stationary phase. Other examples of using phospholipids
in liposome systems (173,174) to estimate the potential partitioning of various


54
solutes into biological membranes have been reported. However, no literature
exists on the use of covalently bonded phosphatidylcholine molecules onto solid
silica surface as a HPLC model system for predicting drug-transport across human
skin.
Specific Aims
Thus, the main objectives of this dissertation could be stated as follows :
1. To design and synthesize redox-based chemical targeting systems for
acyclovir and 5-fluorouracil with improved skin partitioning ability.
2. To evaluate the ability of these targetting systems to deliver the parent drugs
specifically to the skin. Thus, to improve their dermal delivery.
3. To develop the I AM column as an in vitro model system for predicting drug-
transport across human skin.


CHAPTER III
EXPERIMENTAL
Materials
Chemicals of reagent grade were purchased from Aldrich (Milwaukee, WI)
or Sigma (St. Louis, MO) chemical company and were used without further
purification. Ultraviolet spectra were recorded on a Cary 210 Spectrophoto-meter
and the Proton Nuclear Magnetic Resonance spectra on a Varan EM 390
Spectrometer. Chemical shifts were reported as parts per million (5) relative to an
internal standard, tetramethylsilane. Infrared spectra were recorded on a
Beckman Acculab MX 620 double-beam Spectrophotometer. Elemental Analyses
were performed by Atlantic Microlabs (Atlanta, GA). Melting points were
determined using electrothermal melting point apparatus and were uncorrected.
Acyclovir was obtained from Burroughs-Wellcome Co., North Carolina. The
IAM.PC or IAM.PC.MG capped high pressure liquid chromatographic columns
(12 |xm, 10 cm x 4.6 mm i.d.) was obtained from Regis Chemical Co., Morton
Grove, IL.
55


56
Methods
Synthesis of 1-methvl pvrdine-3-carboxylic acid anhydride diiodide ill
Nicotinic acid was reacted with phosgene in benzene in the presence of
triethylamine to form the pyridine 3-carboxylic acid anhydride. This reaction was
carried out according to the previously published literature method (175). The
anhydride was then reacted with methyl iodide in acetonitrile under anhydrous
conditions (148). A mixture of nicotinic anhydride 3.5 g (0.015 mol.) and methyl
iodide 5.44 g (0.038 mol.) was refluxed in acetonitrile, overnight. The orange
crystals formed were filtered, washed with dry ether, and dried in a desiccator
under vacuo. Yield 85 %; mp. 213-215 0C. 1HNMR (90 MHz, d6-DMSO) 8
pyridine protons 8.03-9.36 (8H, multiplet), methyl protons 4.45 (6H, singlet). IR
(KBr) -CH3 stretch (broad, 2800-3200 cm-i), anhydride stretch (1750, 1810 cm-i).
Analysis calculated for C14H14N2O3I2: % C, 32.81; H, 2.73; N, 5.47; I, 49.61;
Found C, 32.66; H, 2.59; N, 5.60; 1,49.66.
Synthesis of l-methyl-3-i(9'-guanylmethoxyethoxyfcarbonvnpyridinium iodide
(AQ1I21
This reaction was carried out according to the previously published
literature method (166). To a solution of 0.5 g (2.2 mmol.) of acyclovir in 20 ml.
dimethylformamide, was added 1.14 g (2.2 mmol.) of [1] and a catalytic amount
0.07 g (0.6 mmol.) of 4-dimethylaminopyridine. The reaction mixture was stirred
at room temperature for 4 days under nitrogen until all of the acyclovir was
consumed. As the reaction proceeded, which was monitored by TLC, the orange


57
color of the anhydride [1] was replaced by a yellow color. The reaction mixture
was filtered and the filtrate evaporated under vacuo. The residue was washed with
dry acetone and dry ether to remove 4-dimethylaminopyridine, unreacted
acyclovir, anhydride, and other impurities, and dried in a desiccator under vacuo.
Yield 65 %; mp. 201-2020C. 1HNMR (90 MHz; d-DMSO) 5 pyridine protons
8.12-9.48 (4H, multiplet); purine Cs proton 7.93 (1H, singlet); purine C3-NH2and
Ci-OH or N2-NH tautomeric protons 6.51 (3H, broad singlet); -CH20-5.45 (2H,
singlet); N+-CH3 4.51 (3H,
singlet); -CH202C- 3.84-4.06 (2H, broad triplet); -OCH2- 3.42-3.51 (2H, broad
triplet). UV characteristics were determined in methanol at 254 nm.
Synthesis of l-methvl-3-rf9l-guanylmethoxyethoxy)carbonyll-1.4-dihydropyrid-
ine (A-CDS) T31
To a solution of 1.58 g (3.3 mmol.) of [2] in 120 ml of degassed water was
added 1.69 g (20.1 mmol.) of sodium bicarbonate, all at once. The mixture was
stirred at 00C and 2.33 g (13.38 mmol.) of sodium dithionite was added over a 5
minute period. The reaction was maintained under nitrogen all the time. The
product [3] is insoluble in water and formed light yellow-colored crystals over the
aqueous layer. The crystals were collected, washed with ice-cold water, dry ether,
and dried in a desiccator under vacuo. Yield 54 %; mp. 163-1650C. iHNMR (90
MHz; d-DMSO) 5 purine Cs proton 7.81 (1H, singlet); pyridine C2 proton 6.81
(1H, singlet); purine C3-NH2 and Ci-OH or N2-NH tautomeric protons 6.48 (3H,


58
broad singlet); pyridine proton 5.60 (1H, doublet); -CH2O 5.36 (2H, singlet);
pyridine C5 proton 4.48-4.81(lH, multiplet); -CH2O2C 3.84-4.24 (2H, multiplet); -
OCH2 and pyridine C4 protons 3.21-3.81 (4H, multiplet); N-CH3 2.81 (3H,
singlet). UV characteristics were determined in methanol at 254 and 370 nm.
Analysis calculated for C15H22N6O6: % C, 47.11; H, 5.79; N, 21.98; Found C,
46.97; H, 5.59; N, 22.27.
Figure 3-1 shows the general reaction scheme for the synthesis of A-CDS.
Synthesis of lipoic-acid ester of acyclovir ('A-LipS2>) T41
Dimethylformamide (20 ml) was heated to 600C and 0.225 g (1.0 mmol.) of
acyclovir was dissolved in it. The solution was cooled to room temperature and
0.206 g (1.0 mmol.) of dl-thioctic acid, 0.206 g (1.0 mmol.) of
dicyclohexylcarbodiimide and 0.122 g (1.0 mmol.) of 4-dimethylaminopyridine
were added to it. The reaction was allowed to run at room temperature under
nitrogen for 4 days and was monitored by TLC. The reaction mixture was filtered
and the filtrate was evaporated under vacuo to remove DMF. The solid residue
was washed with dry ether and then with dry benzene and the solvents were
evaporated under vacuo to remove all residual DMF. The residue was finally
washed with cold methanol to remove unreacted acyclovir, dl-thioctic acid, and
other impurities. Yield 50 %; M.S. (FAB; NBA) = 414 (M+l).
Figure 3-2 shows the general reaction scheme for the synthesis of A-LipS2.


59
Figure 3-1 : Reaction scheme for the synthesis of A-CDS.


60
O
Acyclovir
+
d,l-thioctic acid
DMF DCC
RT, N2 DMAP
4 days
Figure 3-2 : Reaction scheme for the synthesis of A-LipS2.
Synthesis of ethvl ester of 4-isocvanobutvric acid f51
Ethyl 4-aminobutyrate hydrochloride (40.0 g; 238.6 mmol) was suspended
in 250 ml of toluene. Phosgene gas, dissolved in toluene (Fluka; 20 % solution;


61
124.2 ml; 238.6 mmol) was added to it at room temperature with stirring and the
reaction mixture was stirred at 80 OC for 1 hour under nitrogen. Phosgene solution
(60 ml) was added and the reaction mixture was again stirred at 800C for 30
minutes. Additional 60 ml of phosgene solution was added and the reaction
mixture was further stirred at 800C for 30 minutes. Boiling was continued for 1
more hour and at all times nitrogen was bubbled through the reaction mixture.
The solution was cooled and the solid yellow residue was filtered. The solvent was
evaporated under vacuo to give a yellow residue which was distilled to give [5].
Yield 15 g (40 %); bp = 810C/6 mm Hg. 1HNMR (90 MHz; CDC13) 6 4.1 (2H,
quartet, -C02-CH2), 3.3 (2H, triplet, =N-CH2), 2.35 (2H, triplet, -CH2-C02-), 1.9
(2H, multiplet, -C-CH2-C), 1.25 (3H, triplet, -CH3).
Synthesis of l-(3-EthoxycarbonvlpropvlcarbamovlV5-Fluorouracil T61
A mixture of 5-Fluorouracil (5 g; 38.4 mmol), [51 (6 g; 38.4 mmol) in 20 ml
pyridine was stirred for 3 hours at 90C under nitrogen. The reaction mixture was
then cooled, pyridine was evaporated under vacuo, and the residue was dissolved in
150 ml of dichloromethane. The organic layer was washed with 2N HC1 (100 ml),
then water, and dried over MgSC>4. Evaporation of dichloromethane and washing
with ether gave a white solid compound. Yield 7.7 g (69 %); mp. 135-1360C.
iHNMR (DMSO-d6) 5 9.1 (1H, broad singlet, -CO-NH-C-), 8.3 (1H, doublet, C6-
H), 4.05 (2H, quartet, -C02-CH2), 3.3 (2H, multiplet, -N-CH2), 2.31 (2H, triplet, -


62
CH2-CO2-), 1.85 (-C-CH2-C-), 1.2 (3H, triplet, -CH3). Analysis calculated for
C11H14FN3O5: % C, 45.99; H, 4.91; N, 14.63; Found C, 45.93; H, 4.91; N, 14.58.
Synthesis of l-(3-CarboxvpropvlcarbamovD-5-Fluorouracil 171
Concentrated HC1 (40 ml) and 5.75 g (20 mmol) of [6] were stirred at 80C
for 1 hour. Cooling at 10C gave crystals, which were filtered, washed with cold
water, and dried. Yield 3.65 g (71 %); mp. 157-1590C. 1HNMR (DMSO-d): d
1.76 (2H, quintet, -C-CH2-C-), 2.27 (2H, triplet, -CH2-CO2-), 3.32 (2H, quartet, -
N-CH2-), 8.39 (1H, doublet, 6-H), 9.16 (1H, triplet, -CO-NH-), 11.40 (2H, broad,
3-H and -CO2H).
Synthesis ofDL-q-Lipol 181
DL-a-Lipoic acid (2.06 g; 10 mmol) was placed in a flame-dried 250 ml,
round-bottom flask, fitted with a stirring bar and dropping funnel and closed with
a septum. Chloroform (70 ml) was added and catecholborane in THF (50 ml; 50
mmol) was then added dropwise. The mixture was refluxed for 5 hours until the
reduction was complete, as monitored by TLC. Nitrogen atmosphere was
maintained during the entire reaction with an oil-bubbler. Twenty ml of cold
water was added dropwise and the organic solvents were removed in vacuo.
Dichloromethane (50 ml) was then added and the solution was extracted with one
25 ml portion of water followed by six 25 ml extractions using 1 N NaOH solution
to remove catechol. The dichloromethane solution was washed with water (2 x 25


63
ml), dried with magnesium sulphate, filtered, and evaporated. Yield 1.8 g (94 %).
iHNMR (CDCI3): 3.6 (3H, multiplet, ring -CH=, and -CH2-O-), 3.15 (2H, triplet,
ring -CH2-SS-), 2.4 (2H, multiplet, ring -CH2-C-SS-), 1.5 (8H, broad multiplet,
alkyl -(CH2)4-).
Synthesis of Lipolyl ester of l-(3-Carboxvpropvlcarbamovl)-5-Fluorouracil 91
To 60 ml of dichloromethane (containing 10 ml of dimethylformamide) at
50C with stirring was added £71 (2.6 g; 10 mmol), 18] (1.92 g; 10 mmol), and a
catalytic amount of 4-dimethylaminopyridine. Into this mixture,
dicyclohexylcarbodiimide (6.2 g; 30 mmol) dissolved in 20 ml dichloromethane
was added in a period of 10 minutes and the stirring was continued for 1 day. The
reaction mixture was filtered and the filtrate was poured into 300 ml water. The
dichloromethane layer was separated, washed with 3 x 30 ml of 2 N HC1, 3 x 30 ml
of water, 3 x 30 ml of 5 % sodium bicarbonate solution, and finallly with 3 x 30 ml
of water. The organic layer was separated, dried over magnesium sulphate,
filtered, and evaporated in vacuo. The oily residue was purified using column
chromatography (benzene/ethyl acetate; 3/1 solution and silica gel, E. Merck, No.
7734, Kieselgel 60). The oily product was crystallized from methanol to give a
white powder. Yield 1.1 g (25.4 %); mp. 210-2150C. iHNMR (CDCI3): 5 1.35
(4H, multiplet, -C-(CH2)2-C-), 1.58 (2H, quintet, -C-CH2-C-02C-), 1.63 (2H,
multiplet, -S-C-CH2-), 1.85 (1H, dddd, 4'p-H), 1.88 (2H, quintet, -C-CH2-C-),


64
2.34 (2H, triplet, -C-CH2-C02-), 2.40 (1H, dddd, 4'a-H), 3.00-3.17 (2H, multiple!,
-S-S-CH2-), 3.40 (2H, quartet, -N-CH2-), 3.50 (1H, multiplet, 3'cc-H), 8.40 (1H,
doublet, 6-H), 9.03 (1H, triplet, -CO-NH-), 9.17 (1H, broad, 3-NH). M.S. (FAB;
NBA) = 433 (M+l). Figure 3-3 shows the general reaction scheme for the
synthesis of 5-FU-LipS2.
HPLC Analysis
Cation exchange high pressure liquid chromatography was used to analyze
acyclovir, AQ+, and A-CDS. The column, Partisil SCX 10 (im, 22 cm x 4.6 mm
i.d., and a vydac guard column packed with SCX EE 3855 (Du Pont) was linked to
a Spectra-Physics HPLC system consisting of SP 8810 precision isocratic pump, SP
8450 UV/VIS detector, SP 4290 integrator, and SP 8780 autosampler. The mobile
phase consisted of methanol: 1 mM NH4H2P04 (15 : 85), pH 4.00, at a flow rate of
0.9 ml/min. The retention times for acyclovir, A-CDS, AQ+, and A-LipS2 using
the above conditions were observed to be 6.0, 8.0, 10.5, and 7.8 min. respectively.
They were detected at 254 nm and the injection volume was 20 |il.
Reverse-phase HPLC was used for the detection for 5-FU, CPCFU, and 5-
FU-LipS2. They were detected using ASI C-18, 5 p.m, 22 cm x 4.6 mm i.d. at 265
nm with the mobile phase consisting of acetonitrile : water (30 : 70) at a flow rate
of 0.8 ml/min. Their retention times were 4.0, 5.2, and 8.4 min. respectively.


65
coa
H2N(0^)3 C02 C2H5 OCN(CH2)3a^c2h5
Toluene
3-4 h, 80C, N2 [5]
5 -FU
3 h, 90C, N2
Pyridine
HC1
lh, 80C
O
F
o^_NH_(CH2)3-Cp2H
[7]
HOOC (CH2 )4
Catecholborane
CHCI3
5 h, Reflux, N2
HO-(CH2)5
[8]
[7] + [8]
DCC, DMAP
DMF
24 h
O
Lipolyl ester of CPCFU
[9]
Figure 3-3 : Reaction scheme for the synthesis of 5-FU-LipS2.


66
Preparation of Skin Membranes
Three different types of skin membranes were used in vitro to assess the
transport of drugs and their chemical targetting systems into and through the skin;
the full-thickness, freshly excised, hairless-mouse skin, the shaved, epilated,
full-thickness, guinea-pig skin, and the heat-separated, human epidermal
membrane.
Female hairless-mice (SKH-HR-1 strain, Temple University) were
sacrificed by cervical dislocation and the full-thickness skin was removed from the
abdomen and back. The fat layer below the dermis which was gently removed
using forceps was discarded and the full-thickness skin was used for the in vitro
diffusion or partition experiments, immediately.
Human skin from cadavers (Thigh or abdominal region; North Regional
Transplant Services, Lansing, MI) previously dermatomed was gently swirled in
distilled water (preheated to 6(X>C) for about 2-3 min. The epidermal membrane
(about 100 (i. thick) was then carefully separated from the dermis in a glass-trough
containing distilled water at room temperature. The stratum comeum surface was
then placed flat against a plastic sheet and the dermal side of the epidermal layer
was covered with an absorbent paper saturated with 0.9 % NaCl. This human
epidermal membrane was used immediately or stored at -100C and used within 24
hours for the in vitro diffusion or partition experiments.
Guinea-pigs (Harlan Sprague Dawley; 3-4 weeks) were anesthetized using
pentobarbital (i.p.). The hair was shaved with electric clippers and epilated with


67
the depilatory cream, NairR. They were then sacrificed by heart-puncture using
pentobarbital. The full-thickness skin was removed from the abdomen and back,
which was used for the in vitro diffusion experiments, immediately.
Preparation of Donor Solutions
Excess drug and the chemical targetting systems were sonicated in
propylene glycol for 30 minutes and 0.8-1.0 ml of this saturated solution was
used in the donor compartment of diffusion-cells in experiments involving the
hairless-mouse and guinea-pig skin. Acyclovir, 5-FU, and their delivery systems
were observed to be stable in propylene glycol as the donor vehicle.
For the transport studies using human epidermal membrane, excess drug
was stirred in phosphate buffered saline, pH 7.1, for 24 hours at 320C in an
incubator. The suspension was filtered and 5 ml of the filtrate was used as the
donor solution. All the drugs that were evaluated for their transport
characteristics across the human epidermis were stable in the aqueous donor
vehicle.
Diffusion experiments
Two-compartment, vertical diffusion cells (Kresco Enggineering, Palo
Alto) of 7.1 cm2 surface area were used for hairless-mouse and guinea-pig skin
penetration studies. The donor was about 0.8-1.0 ml of saturated solution of the
drug or the chemical delivery system in propylene glycol and the receiver about 40


68
ml of phosphate buffer, pH 6.5. The entire system was incubated at 320C in the
incubator (Lab-Line Instruments, IL) and the receiver was stirred magnetically at
200 rpm. At appropriate time intervals, an aliquot (0.5 ml) of the receiver solution
was sampled for direct analysis by HPLC and was replaced with equal volume of
fresh buffer. The samples were either analyzed immediately or kept frozen at -
10C until analysis.
For the diffusion studies involving human epidermal membrane, side-by-
side diffusion cells (Science Glass Co., Miami, FL) of 1.77 cm2 surface area with
donor and receiver cell volume of 5 ml each were used (Figure 1-7). The donor
was a saturated solution of the drug or the delivery system in either propylene
glycol, dipropylene glycol, or phosphate buffered saline, pH 7.1 and the receiver
was saline. The cells were unstirred. At appropriate time intervals, the entire
receiver solution was sampled for HPLC analysis and the receiver was filled with
equal volume of fresh saline.
The appropriate equations (based upon Fick's laws of diffusion) are shown
in Figure 1-8 (16). These equations were used to obtain the permeability
parameters from the in vitro diffusion experiments.
Skin-Content of the Drug
At various time intervals, the circular piece of skin (7.1 cm2) from the
diffusion experiments was sampled and the donor solution on the skin was
completely removed with 50 % methanol. It was then cut into fine pieces with a


69
pair of scissors, homogenized with 1.0 ml acetonitrile, centrifuged at 13,000 rpm
for 5 minutes (Beckman Microfuge 11), and the supernatant was either
appropriately diluted with acetonitrile before or directly injected into the HPLC
system.
Membrane Partition Coefficient Determination
Dilute solution of the drug or the chemical delivery system (about 100
nmol/ml) in an aqueous medium or propylene glycol (5.0 ml) was allowed to
remain in contact with the stratum comeum side of the skin membrane in a two-
compartment diffusion cell assembly at 320C. The dermis was blocked with
aluminium foil and the receiver was kept empty to prevent diffusion through the
membrane. At various time intervals, an aliquot (0.1 ml) of the donor solution was
sampled for HPLC analysis. The donor aqueous solution was directly analyzed
while the donor propylene glycol solution was appropriately diluted with
acetonitrile for HPLC analysis. The partitioning of the compound in the
membrane was estimated from the decrease in concentration in the donor solution
at equilibrium. The concentration of the drug in the membrane was divided by that
remaining in the donor vehicle at equilibrium to yield Km, the membrane partition
coefficient.


70
In Vitro Stability in Aqueous Buffer
Forty milliliters of Phosphate buffer, 0.05 M, pH 7.4, containing the drug
(about 100 nmol/ml) was shaken at 370C in a water bath (model YB-531; American
Scientific Products). At various time intervals, an aliquot (0.1 ml) of the sample
was withdrawn and injected directly into the HPLC system for analysis.
In Vitro Stability in Biological Media
Rat whole blood, human plasma, and hairless-mouse skin homogenate in
0.05 M phosphate buffered saline, pH 7.4 were used for assessing the relative
abilities of the various chemical targetting systems to release the parent drugs in the
presence of hydrolytic enzymes in biological media under near physiological
conditions.
Male, white rats (Sprague-Dawley, IN) were humanely sacrificed by
decapitation and their whole blood collected into heparinized tubes. A stock
solution of the drug (0.1 ml of 10 |imol/ml) in acetonitrile was added to 4.9 ml of
the rat whole blood. Five milliliters of the whole blood containing the drug was
carefully shaken at 37 0C. At various time intervals, an aliquot (0.1 ml) sample was
pipetted into centrifuge tubes containing 0.9 ml cold acetonitrile, centrifuged at the
maximum speed of 13 for 5 minutes, and the supernatant was directly analyzed by
HPLC.
Freshly excised, full-thickness, hairless mouse skin was cut into fine pieces,
weighed, and homogenized with 0.05 M phosphate buffered saline, pH 7.4 to give a


71
50% homogenate. A stock solution of the drug (0.1 ml of 10 p. mol/ml) in
acetonitrile was added to 4.9 ml of the 50% skin-homogenate and the final 5 ml of
the skin-homogenate containing the drug was shaken at 32 0C. At various time
intervals, an aliquot (0.1 ml) was sampled and pippetted into centrifuge tubes
containing 0.9 ml cold acetonitrile, centrifuged at the maximum speed of 13 for 5
minutes, and the supernatant was injected into the HPLC system.
Fresh human blood was collected into heparinized centrifuge tubes and
centrifuged at the speed of 50 for 5 minutes. The fresh plasma was separated and
made up with 0.05 M phosphate buffered saline, pH 7.4 to give 85 % plasma. A
stock solution of the drug (0.1 ml of 10 p. mol/ml) in acetonitrile was added to 4.9
ml of 85% plasma. The final 5 ml of human plasma containing the drug was shaken
at 370C. At various time intervals, an aliquot (0.1 ml) was sampled and pippetted
into centrifuge tubes containing 0.9 ml cold acetonitrile, centrifuged at the
maximum speed of 13,000 rpm for 5 minutes, and the supernatant was injected
directly into the HPLC system.
I AM Column Chromatography
The 12 p. IAM.PC or the IAM.PC.MG capped column (containing methyl-
glycolate capped LAM-phophatidylcholine stationary phase), 15 cm x 4.6 mm i.d.,
was connected to the Waters HPLC system consisting of 510 pump, 712 Wisp
automatic injector, Lambda-Max Model 480 LC Spectrophotometer, and a 730
Data Module. A dilute solution of the drug (50-100 nmol/ml) in the mobile phase


72
was injected into the column. The relative retention, also called the capacity factor,
K', of the test compound was calculated using the equation, K' = (tr = to)/to, where tr
is the retention time in minutes of the test compound and to that of the unretained
compound, citric acid. The test soutes were detected by UV at A. =210, 220, or 254
nm except the alcohols which were detected by refractive index detection using
Refractomonitor III operating at 0.5 mA. The mobile phase compositions used
were 0-30 % acetonitrile in Dubelco's Phosphate Buffered Saline 10 x dilution
(DPBS 10 x), pH 7.1, at a flow rate of 1.0 or 2.0 ml/min. The column was washed
with acetonitrile : water (20 : 80) at 1.0 ml/min for 1 hour, before and after
analysis, and was preserved (stored) by washing it with acetonitrile at 1.0 ml/min
for 1 hour.
C-18 Column Chromatography
The Waters jiBondapak 10 p, 15 cm x 4.6 mm i.d. column was used with
similar HPLC conditions described in the section on IAM Column
Chromatography, above. Generally, mobile phase compositions containing higher
percentage of organic modifyer were required to elute various compounds from
the C-18 stationary phase. Thus, the alcohols were eluted using acetonitrile : DPBS
10 x (30 : 70) and the steroids were eluted with acetonitrile : DPBS 10 x (50 : 50) as
the mobile phase.


CHAPTER IV
RESULTS AND DISCUSSION
Syntheses
All the compounds were successfully synthesized based upon the results
obtained from Nuclear Magnetic Resonance Spectra, Mass Spectrometry, and
Elemental Analyses.
Generally, the coupling of acyclovir with the acylating agent was the most
difficult and a slow step compared to all the other synthetic steps. The reason for
this difficult and slow rate of acylation of acyclovir is its low solubility in most
inorganic and organic solvents used in synthetic reactions. In addition, the
relatively low reactivity of the acyclic alcohol functional group of acyclovir makes
it even more less amenable for derivatization.
Therefore, it was observed that for esterification of acyclovir, activation of
the acid, namely to its anhydride was required not only for increasing the rate of
the reaction towards completion, but also increasing the yield of the product.
Thus, the trigonelline anhydride was first made and then coupled to acyclovir to
give AQ+. The acylation of acyclovir could have been easily achieved using
nicotinic anhydride. However, in the subsequent step, the quatemization of the
tertiary nitrogen in the pyridine moiety of the nicotinic acid ester of acyclovir
73


74
could have also resulted into methylation of the N-2 and N-7 reactive centers on the
guanine moiety. Therefore, to circumvent this problem of formation of undesired
products, the trigonelline anhydride was specifically synthesized in our laboratory
and was reported in the literature (148) for the first time.
Although, this reaction could be carried out in pyridine as the reaction
solvent, dimethylformamide was observed to be a better solvent. However, both
solvents have their inherent disadvantages. Dimethylformamide is extremely
difficult to evaporate due to its high boiling point (1530Q, and requires exhaustive
washing steps with ether and benzene to remove it from the final product.
Pyridine, on the other hand, can be evaporated again with excessive washing steps
using ether or benzene. However, the low solubility of acyclovir and other
reactants in pyridine was the primary limiting factor in reducing the efficiency of
the synthetic reactions using this solvent. Thus, it was observed that perhaps
dimethylformamide due to its greater solubilizing ability behaved as a better
solvent for carrying out the reactions.
Acylation of acyclovir with a simple unhindered, long-chain acid, such as
lipoic acid was again observed to be a slow, low-yield process. This reaction was
carried out using dimethylformamide as the reaction solvent and equimolar
amounts of acyclovir, lipoic acid, dicyclohexylcarbodiimide, and 4-
dimethylaminopyridine. The reaction was observed to proceed a little faster upon
warming the mixture to 400C.


75
Skin Experiments
The in vitro experiments were conducted using freshly prepared skin
membranes in the two-compartment diffusion-cell assembly. This particular
laboratory set-up allows for simultaneous monitoring of dermal (local) or
transdermal delivery of various solutes when administered to the skin from the
donor side.
The full-thickness, hairless-mouse skin was chosen as a model membrane to
study and compare the ability of acyclovir and its chemical delivery systems to
localize delivery of acyclovir to the skin. When freshly excised and used
immediately, this tissue provides a rapid and simple in vitro method to evaluate and
demonstrate transport and metabolism (hydrolytic enzymes maintain activity) (177-
179) of various solutes in the layers of the skin. This tissue when used in
combination with the two-compartmental diffusion-cell assembly, serves as an
appropriate, standardized, reliable, and reproducible method by which concurrent
transport and metabolism of different solutes or drugs could be easily monitored.
The method is easy to perform and the availability of the hairless-mouse is plenty
and relatively inexpensive. The shaved, guinea-pig skin and heat-separated human
epidermal membranes were similarly used.
Selection of Delivery Vehicle
Propylene glycol was chosen as the appropriate vehicle for the delivery of
acyclovir and its chemical delivery systems. Saturated solution of the drug in


76
propylene glycol was used in the donor compartment. Therefore, the
thermodynamic activity of the drug was maximized and was nearly equal to the
saturation solubility of the drug in propylene glycol so that maximum driving
force of the drug in the membrane could be achieved. By using saturated solutions
the vehicle effect in affecting the partitioning of the compounds in the membrane is
eliminated and therefore helps in direct comparison of the pure ability of the
molecular species to interact with the skin membrane.
Propylene glycol was chosen as the donor vehicle since it is widely used in
formulating topical drug products. A-CDS is unstable in aqueous solutions and
undergoes a water catalyzed hydroxyl group addition on the C(, atom of the
dihydro trigonelline moiety (180). Therefore, it could not be administered to the
skin in an aqueous (buffered) donor vehicle.
Thus, the compounds demonstrated optimal combination of their membrane
partitioning effect, Km, and their concentration Cd(M-mol/ml), in the donor phase.
This is necessary because according to Fick's laws of diffusion, the flux of the drug
across the membrane is directly proportional to its membrane partition coefficient
and its concentration in the donor phase facing the membrane (Figure 1-8) (16).
This balance between the product of Km and Cd is important since they compromise
the actual magnitudes of each other. Thus, if Cd increases, indicating that the
vehicle has a greater affinity for the drug than the drug has for the skin membrane
then Km tends to decrease and vice versa.


77
Therefore, to monitor the appearance of the drug in the receiver on the
other side of the skin, that is, to observe sufficient, quantifiable amounts of the
drug after diffusion through the skin-membrane, these interactions between the
drug, vehicle, and the skin are important and have to be adjusted appropriately.
Skin Penetration and Retention of Acyclovir
The results obtained upon administering a saturated solution of acyclovir in
propylene glycol to the freshly excised full-thickness hairless-mouse skin, using the
two-compartment diffusion cells in vitro, are shown in Table 4-1 and Figure 4-1.
The drug was analyzed by HPLC at various time intervals in the receiver according
to the description in the methods section.
Table 4-1 : Cumulative amounts of acyclovir in the receiver after application of
saturated solution of acyclovir (33.5 p mol/ml) in propylene glycol to the hairless-
mouse skin in vitro at 320C.
Time (hours)
Acyclovir in receiver
(nmol/cm 2)a
2
0.2+0.1
4
0.4 + 0.2
5
0.4 + 0.2
6
0.5 + 0.1
8
0.8+0.2
10
0.9+0.3
12
1.5 +0.4
a Average of three determinations s.d.


78
Small quantities of acyclovir could be detected in the receiver compartment
over an extended period of time. This indicated limited penetration of acyclovir
through the skin and was expected due to its unfavourable physical-chemical and
membrane permeation properties (133). Such reduced ability to partition into the
stratum comeum has indeed been reported (131,132) and one of the primary
objectives of this dissertation was to improve upon it.
Hence, low values of flux (8.82 x 10-5 jimol/cm2/h) and permeability
coefficient (2.44 x 10-6 cm/h) for the diffusion of acyclovir through the hairless-
mouse skin were observed.
Time (hours)
Figure 4-1 : Diffusion of acyclovir across freshly excised hairless-mouse skin after
application of its saturated solution (33.5 (J. mol/ml) in propylene glycol in vitro at
320C.


79
During the diffusion experiments, when the skin and the receiver phase
were analyzed at various time intervals for the drug content, acyclovir showed
greater tendency to penetrate the skin completely. Therefore, higher levels of
acyclovir could be detected in the receiver as compared to those in the skin tissue
(Table 4-1, 4-3, and 4-6).
Table 4-2 : Amounts of acyclovir in the skin after application of its saturated
solution (33.5 |imol/ml) in propylene glycol to the hairless-mouse skin in vitro at
350C.
Time (hours)
Acyclovir in skin
(nmol/cm2)a
3
0.74+0.16
6
1.41 + 0.37
9
2.84 + 0.37
12
3.88 + 0.41
a Average of three determinations s.d.
This greater propensity of the antiviral agent to go through the skin may be
responsible for its lack of efficacy in treating the cutaneous HSV I infection as
reported by many workers (127,131,132). Analysis of drug-content in the
receiver as well as in the skin tissue at the same time-point is critical because it
helps in better understanding the fate of the drug under investigation (181).


80
Table 4-2 shows the amounts of the antiviral agent in the skin when it was
applied to the skin in the in vitro experiment at 35 0C. As compared to the values at
32C (Table 4-1), the amounts are higher at 35C (Table 4-2), as expected.
Chemical Delivery System for Acyclovir Based on Oxidation
Table 4-3 shows the amounts of acyclovir or AQ+ (in nmoles per unit
diffusional surface area) in the skin and receiver that were found after application
of A-CDS as a saturated solution in propylene glycol to the hairless-mouse skin in
Y-ifr-Q-
The reduced, relatively more lipophilic A-CDS demonstrated rapid and
facile skin-partitioning followed by fast oxidative metabolic step probably due to
the presence of NAD+-NADH redox coenzyme systems to form increasingly high
levels of the oxidized metabolite, AQ+, in the skin. The levels of AQ+ in the skin
increased about 14 fold from 6 to 48 hours. However, compared to the oxidative
metabolic reaction, the ester hydrolysis of AQ+ to release acyclovir in the skin
mediated by the action of the non-specific esterases was observed to be relatively
much slower. As a result of this rate difference in the kinetic processes, there was
locking" of AQ+ in the skin from which a slow release of acyclovir was observed.
Thus, the ratio of AQ+/A in the skin was found to increase from 7.5 to 50 to 73 at 6,
24, and 48 hours respectively, indicating faster oxidation relative to ester
hydrolysis in the hairless-mouse skin.


81
Hence, A-CDS managed to improve the delivery of acyclovir to the skin per
unit dose by 9 fold (p < 0.025) at 6 hours, compared to underivatized acyclovir.
This improvement was 4 and 3 fold (p < 0.025) over acyclovir at 24 and 48 hours,
respectively (Figure 4-2, 4-5). The improvement factor is a pure measure of the
superior ability of A-CDS to deliver acyclovir to the skin compared to
underivatized acyclovir under similar experimental conditions and normalizes the
effects due to the dose of the compound applied to the skin in the same delivery
vehicle.
The A-CDS, according to Table 4-3, also delivered acyclovir into the
receiver compartment, indicating that an increase in systemic delivery may occur
in vivo. This could be due to formation of a large fraction of AQ+ and acyclovir in
Table 4-3 : Amounts of acyclovir and AQ+ in the skin and receiver (cumulative)
after application of saturated solution of acyclovir (33.5 ji mol/ml) or A-CDS (45.3
pmol/ml) in propylene glycol to the hairless-mouse skin in vitro at 320C.
Compound
Dose Time
Acyclovir (nmol/cm2)a
AQ+ (nmol/cm2)a
(nmol/cm 2)
(h)
Skin
Receiver
Skin
Receiver
Acyclovir
3896
6
0.4 + 0.2
0.4 + 0.2
24
1.2+0.5
3.8 + 1.5
48
1.7 0.5
7.3 1.4
A-CDS
4989
6
4.7 + 2.0
5.4 +0.2
35+14
6+1
24
5.8 + 1.8
6.3 +0.5
289 +17
37 +4
48
6.7 + 1.6
7.8 +0.3
488 +35
142 + 23
a Average of three determinations s.d.


82
20
x
.h
>
_o
o
%
<4-1
o
>
I
13
<3
10 -
using acyclovir
using A-CDS
Figure 4-2 : Dermal delivery of acyclovir per unit dose using acyclovir or A-CDS
as a function of time.
the extracellular compartment of the skin because release of AQ+ and acyclovir in
the receiver was observed. If the AQ+ and acyclovir were specifically
concentrated into the skin cells then they would not have been detected in such high
amounts in the receiver. Their more polar characteristics would have prevented
their efflux out of the intracellular compartment.
When acyclovir was applied to the skin, the ratio of its amounts in the skin to
receiver decreased from 1 to 0.3 to 0.2 at 6, 24, and 48 hours respectively, as seen
in Table 4-3. This greater tendency for acyclovir to go through the skin may


83
indicate its presence predominantly in the extracellular compartment when
underivatized acyclovir is administered to the skin. Its lower lipid solubility may
not allow for sufficient penetration of the cell-membranes of the epidermal cells.
The A-CDS, on the other hand, altered this kinetic profile most probably
due to intracellular localization of AQ+ and hence acyclovir. Due to A-CDS, the
ratio (skin/receiver) of acyclovir was found to remain constant and was 0.87, 0.92,
and 0.86 at 6, 24, and 48 hours, respectively. Thus, a steady-state in the
distribution of acyclovir between the skin and receiver had reached after the
application of A-CDS to the skin. This steady-state must have been a direct result
of a locked" fraction of acyclovir in the intracellular compartment from which
facile release of acyclovir in the receiver would not have occurred. The
corresponding ratios (skin/receiver) of AQ+ as a result of A-CDS application to the
skin were 5.8, 7.8, and 3.4 at 6, 24, and 48 hours, respectively. Thus, the AQ+ was
more preferentially locked" in the skin than acyclovir. Therefore, the overall
effect of applying A-CDS to the skin was the establishment of a reservoir" of the
metabolic precursor (AQ+) of acyclovir with a consequent improvement in the
delivery of acyclovir to the skin.
The results seem to indicate that the concentration of AQ+ in the skin is an
important determinant of the availability of acyclovir in the skin. It is worthwhile
to note that as the ratio (skin/receiver) of AQ+ decreases with time (from 5.8 to 7.8
to 3.4) (Table 4-3), so does the improvement in acyclovir delivery to the skin due
to A-CDS over underivatized acyclovir (from 9 to 4 to 3 fold) (Figure 4-5).


Full Text
TRANSDERMAL TRANSPORT AND INTRADERMAL DRUG TARGETING
USING NOVEL CHEMICAL DELIVERY SYSTEMS
By
PRASHANT J. CHIKHALE
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE
UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1991

Copyright 1991
by
Prashant J. Chikhale

To my mother and father

ACKNOWLEDGEMENTS
I am grateful to my thesis advisor, Professor Nicholas S. Bodor, Ph.D.,
D.Sc., for his advise, guidance, and support during my graduate education. By
working in close association with him and his research group I was exposed to
different aspects of pharmaceutical research which I think are critical for the
development and understanding of scientific thoughts and procedures. During my
third and fourth year of graduate study, I worked with Francisco M. Alvarez,
Ph.D., at Schering-Plough Research, Pembroke Pines, Florida, in the area of in
vitro model development. I appreciate the fine support and guidance provided to
me by Francisco during this part of my thesis work at Schering-Plough Research.
My other committee members, James W. Simpkins, Ph.D., Richard H.
Hammer, Ph.D., Hans Schreier, Ph.D., and Kenneth H. Rand, M.D., were very
supportive towards my graduate education and I wish to thank them for their
advice and encouragement.
I would like to thank Drs. Balasingam Radhakrishnan, Vasu
Venkatraghavan, Ede Marvanyos, Jonnalagadda Sastry, Lazio Prokai, and Emy
Wu for their helpful discussions and assistance during my laboratory work.

Due to the ultimate moral and financial support of my mother, Vasundhara
J. Chikhale, my father, Jayant M. Chikhale, and my grandfather, Madhusudhan S.
Gandhi, I was able to undertake this task of graduate education and therefore they
deserve equal credit for this dissertation.
Last, but not the least, I wish to acknowledge Elsbeth Brunt, without whose
total support and mention this dissertation is not complete.
v

TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS iv
LIST OF TABLES vüi
LIST OF FIGURES x
KEY TO ABBREVIATIONS xiv
ABSTRACT xvi
CHAPTERS
I INTRODUCTION 1
Structure of the Skin 1
Structure and Function of the Permeability Barrier of the Skin 4
Protein and Lipid Domains in the Human Stratum Comeum 6
Localization of the Skin-Barrier to Drug-Transport 8
Relationship between Lipid Structure and Barrier Function 9
Drug-Delivery to the Skin via Pro-drugs 14
Drug Metabolism in the Skin 17
Methods for Estimating Transport of Drugs Into and Through
the Skin 22
II SIGNIFICANCE OF THE STUDY 29
Delivery of Acyclovir To the Skin 29
Delivery of 5-Fluorouracil To the Skin 45
Development of the IAM.PC Column 47
Specific Aims 54
m EXPERIMENTAL 55
Materials 55
Methods 56
Synthesis of l-methylpyridine-3-carboxylic acid
anhydride diiodide [1] 56

Synthesis of l-methyl-3-[(9'-guanylmethoxyethoxy)
carbonyl]pyridinium iodide (AQ+) [2] 56
Synthesis of 1-methyl-3-[(9'-guanylmethoxyethoxy)
carbonyl]-1,4-dihydropyridine [3] 57
Synthesis of lipoic-acid ester of acyclovir [4] 58
Synthesis of etnyl ester of 4-isocyanobutyric acid [5] 60
Synthesis of l-(3-ethoxycarbonylpropylcarbamoyl)
-5-fluorouracil [6] 61
Synthesis of l-(3-carboxypropylcarbamoyl)
-5-fluorouracil [7] 62
Synthesis of D,L-a-lipol [8] 62
Synthesis of lipolyl ester of l-(3-carboxypropyl-
carbamoyl)-5-fluorouracil [9] 63
HPLC Analysis 64
Preparation of Skin Membranes 66
Preparation of Donor Solutions 67
Diffusion Experiments 67
Skin-Content of the Drug 68
Membrane Partition Coefficient Determinations 69
In Vitro Stability in Aqueous Buffer 70
In Vitro Stability in Biological Media 70
IAM.PC Column Chromatography 71
C-18 Column Chromatography 72
IV RESULTS AND DISCUSSION 73
Syntheses 73
Skin Experiments 75
Selection of Delivery Vehicle 75
Skin Penetration and Retention of Acyclovir 77
Chemical Delivery System for Acyclovir Based on Oxidation 80
Chemical Delivery System for Acyclovir Based on Reduction 86
Chemical Delivery System for 5-Fluorouracil 91
Lipophilicity and Partition Coefficients 94
In Vitro Stability 97
Stability of the IAM.PC Column Stationary Phase 100
Selection of Model Solutes and Appropriate Parameters
for Comparison 101
Prediction of n-Alcohol Transport Across Human Skin 102
Permeability of Human Epidermis to Steroids 107
Skin-Permeability of Water-Soluble Drugs 113
Transport of Nucleosides Across Human Epidermal Membrane 120
V SUMMARY AND CONCLUSIONS 125
REFERENCES 129
BIOGRAPHICAL SKETCH 141

LIST OF TABLES
Table Page
1-1 Composition of stratum comeum lipids (44,50) 10
1-2 Composition of mammalian epidermal lipids (57) 13
1-3 Biotransformation reactions by human skin (77) 20
2-1 Lipophilicity of the (alkvlaminomethyl)benzoate ester
prodrugs of acyclovir and their susceptibility to
enzymatic hydrolysis in human plasma (141) 37
4-1 Cumulative amounts of acyclovir in the receiver after
application of saturated solution of acyclovir
(33.5 (imol/ml) in propylene glycol to the
hairless-mouse skin in vitro at 32°C 77
4-2 Amounts of acyclovir in the skin after application of its
saturated solution (33.5 fimol/ml) in propylene
glycol to the hairless-mouse skin in vitro at 350C 79
4-3 Amounts of acyclovir and AQ+ in the skin and receiver
(cummulative) after application of saturated solution of
acyclovir (33.5 pmol/ml) or A-CDS (45.3 |imol/ml) in
propylene glycol to the hairless-mouse skin in vitro
at 320C 81
4-4 Amounts of acyclovir and AQ+ in the skin (at 48 hours) and
receiver (cummulative) after application of saturated
solution of A-CDS (45.3 pmol/ml) in propylene glycol
to the hairless-mouse skin in vitro at 320C for 6 hours.
Donor was removed at 6 hours and skin was extracted
at 48 hours 85
4-5 Amounts of acyclovir and AQ+ in the skin (at 48 hours) and
receiver (cummulative) after application of saturated
solution of A-CDS (45.3 (imol/ml) in propylene glycol
to the hairless-mouse skin in vitro at 32°C 85
v i i i

4-6
Amounts of acyclovir and A-LipS2 in the skin and receiver
(cummulative) after application of saturated solution of
acyclovir (35.6 iimol/ml) or A-LipS2 (6.0 jimol/ml) in
propylene glycol to the hairless-mouse skin in vitro
at 320C 87
4-7 Amounts of 5-FU in the skin and receiver (cummulative)
after application of saturated solution of 5-FU
(350 pjnol/ml) or 5-FU-LipS2 (84 pmol/ml) in
propylene glycol to the shaved guinea-pig skin
in vitro at 32°C 92
4-8
Lipophilicity and skin-membrane partition coefficients of
acyclovir and A-CDS from their dilute solutions
in propvlene glvcol in vitro at 320C
95
4-9
Partitioning of acyclovir and A-LipS2 into human epidermal
membrane from their dilute water solutions m vitro
at 320C
96
4-10
Stability of A-CDS, AQ+, A-LipS2, CPCFU, and 5-FU-LipS2,
in vitro at 370C
99
4-11
Permeabilitv of human epidermis to steroids in vitro at 320C
and their relative retention in the IAM.PC and C-18
HPLC columns
108
4-12
Transport of nucleosides across heat-separated human
epidermal membrane in vitro at 320C and their
relative retention in the IAM.PC.MG column
122

LIST OF FIGURES
Figure Page
1-1 Schematic representation of the human skin (2) 2
1-2 Dual functions of the epidermal lamellar bod)'. Summary of
lipid biochemical and enzyme biochemical studies on
isolated epidermal lamellar bodies (33) 7
1-3 The ceramides of the human stratum comeum (52) 12
1-4 Changes in lipid composition in the stratum comeum (57) 13
1-5 Schematic representation of the use of prodrugs in topical
therapy (63) 14
1-6 Schematic representation of the skin as a metabolic barrier
(E = enzyme)(77) 18
1-7 Two-compartment, side-by-side, in vitro diffusion cells 23
1-8 Equations used for calculating skin-permeability (16) 24
2-1 Chemical structure of acyclovir 30
2-2 Cellular activation of acyclovir 31
2-3 Prodrugs of acyclovir with improved oral absorption (134,135) 34
2-4 Water-soluble ester prodrugs of acyclovir (139) 35
2-5 (Alkylaminomethyl) benzoate ester prodrugs of acyclovir (141) 36
2-6 Delivery of steroid hormones to the skin using 3-spiro-
thiazolidine derivatives (145) 40
2-7 Delivery of acyclovir to the skin using A-CDS 42
2-8 Delivery of acyclovir to the skin using A-LipS2 44
x

2-9 Working hypothesis mechanism for lipoamide
dehydrogenase (153) 45
2-10 Chemical structure of 5-Fluorouracil 46
2-11 Delivery of 5-fluorouracil to the skin using 5-FU-LipS2 48
2-12 Schematic representation of a cell-membrane 49
2-13 The IAM.PC column stationary phase (166) 50
2-14 Synthesis of nucleosil-lecithin (the IAM.PC column
stationary phase) (165) 51
3-1 Reaction scheme for the synthesis of A-CDS 59
3-2 Reaction scheme for the synthesis of A-LipS2 60
3-3 Reaction scheme for the synthesis of 5-FU-LipS2 65
4-1 Diffusion of acyclovir across freshly excised hairless-
mouse skin after application of its saturated
solution (33.5 umol/ml) in propylene glycol
in vitro at 320C 78
4-2 Dermal delivery of acyclovir per unit dose using
acyclovir or A-CDS as a function of time 82
4-3 Dermal delivery of acyclovir per unit dose using
acyclovir or A-LipS2 as a function of time 88
4-4 Dermal delivery of 5-FU per unit dose using
5-FXJ or 5-FU-LipS2 as a function of time 93
4-5 Improvement in delivery of acyclovir (to the hairless-
mouse skin) or 5-FU (to the guinea-pig
skin) using A-CDS, A-LipS2, or 5-FU-LipS2 over
underivatized acyclovir or 5-FU
respectively, as a function of time 94
4-6 Relationship between the stratum comeum membrane
partition coefficients (Km) and its permeability
(Kp) to n-alcohols on a log-log scale 103
4-7 Relationship between the relative retention of
n-alcohols in the C-18 column (K'(C-18)} and
their stratum comeum membrane permeability
coefficients (Kp) on a log-log scale 104
x ¡

4-8
Relationship between the relative retention of
n-alcohols in the IAM.PC column {K'(IAM.PC)}
and their stratum comeum membrane partition
coefficients (Km) on a log-log scale 105
Relationship between the relative retention of
n-alcohols in the IAM.PC column {K'(IAM.PC)}
and their stratum comeum membrane
permeability coefficients (Kp) on a log-log scale 106
4-10 Chemical structures of the steroids 109
4-11 Relationship between the human epidermal membrane
partition coefficients (Km) and its permeability
(Kp) to the steroids on a log-log scale Ill
4-12 Relationship between relative retention of the steroids
in the C-18 column {K'(C-18)} and their human
epidermal membrane permeability coefficients
(Kp) on a log-log scale 112
4-13 Relationship between relative retention of the steroids
in the IAM.PC column (K'(IAM.PC)} and their
human epidermal membrane partition
coefficients (Km) on a log-log scale 113
4-14 Relationship between relative retention of the steroids
in the IAM.PC column {K'(IAM.PC)} and their
human epidermal membrane permeability
coefficients (Kp) on a log-log scale 114
4-15 Chemical structures of the water-soluble drugs 116
4-16 Relationship between relative retention of the water-
soluble drugs in the IAM.PC column (K'(IAM.PC)}
using ACN&OH/DPBS lOx (10/10/80) as the
mobile phase and their human epidermal
membrane permeabilities (Kp) on a log-log scale 117
4-17 Relationship between relative retention of the water-
soluble drugs in the IAM.PC column {K'(IAM.PC)}
using MeOH/DPBS lOx (50/50) as the mobile
phase and their human epidermal membrane
permeabilities (Kp) on a log-log scale 118
4-18 Relationship between relative retention of the water-
soluble drugs in the C-18 column (K'(C-18)} using
MeOH/DPBS lOx (50/50) as the mobile phase and
their human epidermal membrane permeabilities
(Kp) on a log-log scale 119
x i i

4-19 Chemical structures of the nucleosides 121
4-20 Relationship between relative retention of the nucleosides
in the IAM.PC column {K’(IAM.PC)} using
DPBS lOx as the mobile phase and their human
epidermal membrane permeability coefficients
(Kp) on a log-log scale 123
x i j i

KEY TO ABBREVIATIONS
A:
Acyclovir
A-CDS:
1,4-dihydrotrigonelline chemical delivery system for
acyclovir
AQ+:
N-methylnicotinic acid ester of acyclovir
A-LipS2:
Lipoic acid ester of acyclovir
5-FU :
5-Fluorouracil
CPCFU:
Carboxypropylcarbamoyl-5-fluorouracil
5-FU-LipS2:
Lipolyl ester of carboxypropylcarbamoyl-5-fluorouracil
UV:
Ultraviolet
mp:
Melting point
bp:
Boiling point
1HNMR:
Proton Nuclear Magnetic Resonance Spectrum
M.S.:
Mass Spectrum
mol:
Moles
nmol:
Nanomoles
(imol:
Micromoles
mmol:
Millimoles
nm :
Nanometer
mm:
Millimeter
cm:
Centimeter
log :
log 10
XIV

Kobs :
Pseudo first-order rate constant
*1/2:
Half-life
rpm :
Rotations per minute
min:
Minute
h:
Hour
sec:
Second
M:
Moles per liter
pi:
Microliter
X:
Wavelength
r :
Correlation coefficient
IAM.PC :
Immobilized Artificial Membrane
stationary phase
K’ :
Relative retention or Capacity factor
i.d. :
Internal diameter
Km:
Membrane partition coefficient
Kp:
Permeability coefficient
J:
Steady-state flux
Q:
Donor-cell concentration
5:
Thickness of the skin membrane
XV

Abstract of Dissertation Presented to the Graduate School of the University of
Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of
Philosophy
TRANSDERMAL TRANSPORT AND INTRADERMAL DRUG¬
TARGETING USING NOVEL CHEMICAL DELIVERY SYSTEMS
By
Prashant J. Chikhale
May 1991
Chairman : Nicholas S. Bodor
Major Department: Medicinal Chemistry
This dissertation examines the feasibility of targeting and localizing
important antiviral agents (like acyclovir) and anticancer agents (like 5-
fluorouracil) specifically to the skin using novel, redox-based chemical targeting
systems. Such approaches should lead to improvement in the effectiveness of
topically administered acyclovir in treating recurrent mucocutaneous herpes
simplex virus infection of type I. Similarly, the basal cell skin carcinomas or
psoriasis can be effectively treated if 5-fluorouracil could be targeted to the intra-
dermal region.
Chemical Delivery Systems (CDS) for acyclovir based on oxidation (the 1,4-
dihydrotrigonelline moiety containing ester; A-CDS) or reduction (the lipoic acid
ester; A-LipS2) in the skin were utilized to enhance the skin-partitioning ability of
acyclovir and use the enzymatic activity of the skin to create metabolic chemical
precursors as reservoirs for the release of acyclovir in the skin.
XVI

Thus, the dermal delivery of acyclovir was improved by 9-fold (p < 0.025)
using A-CDS, and by 37-fold (p < 0.001) using A-LipS2, at 6 hours relative to
underivatized acyclovir, when administered to the hairless-mouse skin, in vitro.
The lipolyl ester of 5-fluorouracil (5-FU-LipS2) also managed to deliver greater 5-
fluorouracil to the hairless-guinea pig skin as compared to underivatized 5-
fluorouracil.
Thus, such drug-targeting CDS could be effectively used to deliver not only
heterocyclic antiviral and anticancer agents, but also other potent therapeutic drugs
and pharmaceuticals Iq the skin.
The dissertation also describes the development of an important, novel, in
vitro, chromatographic model system (the IAM.PC column) containing a cell-
membrane component for studying drug-membrane interactions and predicting
drug-transport across complex biological membrane barriers, particularly the
human skin. It utilizes a covalently bonded lecithin analog on a solid (silica)
stationary support for the purpose of mimicking specific and non-specific
interactions of solutes with cell membranes.
The IAM.PC column was able to demonstrate significant linear relationships
between the relative retention of various solutes on the phospholipid stationary
phase and their permeabilities across the human epidermal membrane, on a log-log
scale. Thus, it could be used as a potential model to study drug-biomembrane
interactions and drug-transport across biological membrane barriers.
XVII

CHAPTER I
INTRODUCTION
The skin on the body of an average adult covers a surface area of
approximately 2 square meters and receives about one-third of the blood
circulating through the body (1). It is one of the most readily accessible organs of
the human body. With a thickness of only a few millimeters (2.97 ± 0.28 mm), the
skin separates the vital organs from the outside environment, serves as a protective
barrier against physical, chemical, or microbial attacks, acts as a thermostat in
maintaining body temperature, plays a role in the regulation of the blood pressure,
and protects the human body against the penetration of ultraviolet rays (2).
Structure of the Skin
Morphologically, the skin is a multilayered organ composed of four major
multilaminate layers: the stratum corneum, the epidermis, the dermis, and the
hypodermis (or the subcutaneous layer) (Figure 1-1) (2).
The hair and an associated vapour layer constitute the outer boundary with
the environment. It is situated on the visible skin surface, which consists of a
lightly shiny lipid film. The quantity of fat ranges between 5-150 |ig/cm2 (3)
which corresponds to 0.05-1.5 pm thickness.
1

2
Stratum-
Granulosum
Stratum-
Spinosum
Stratum-
Germinativum
Epidermis
Capillary Network
Sebaceous Gland
Hair Shaft-
Apocrine Sweat
Gland
Hair Follicle
Blood Vessel
Hv podermis
Figure 1-1 : Schematic representation of the human skin (2).
fti

3
The next layer is the actual upper surface of the stratum comeum, consisting
of homy lamellae, the comeocytes. They are the non-living cells situated on the
epidermis in the form of many layers of compacted, flattened, dehydrated, and
keratinized cells. They are continuously shed with constant replacement from the
underlying viable epidermal tissue (4). The phenomenon of multiple layering
within this homy layer is regarded as the fundamental factor in inhibiting the
penetration of most substances.
The epidermis and the horny layer are "indented" by hair follicles and
perforated by sweat gland ducts. The epidermal cellular tissue extends along the
follicular canal deep into the dermis, and the sweat gland ducts are lined with their
own endothelial cells, so that practically no dermal surfaces are exposed into which
direct penetration of substances might occur. The intercellular canals may be a
route for percolation of materials through the epidermis (5). These spaces are
represented by acidic glycosaminoglycans which coat the cell-surface and this
intercellular gel (glycocalix) has a high water content which could constitute a
system of canals for permeation. The intercellular spaces represent 15-18 % of the
total epidermal volume (6).
The dermis consists of an aqueous phase as well as structured elements such
as collagen fibers and elastin, both embedded in mucopolysaccharide networks (6).
Fibroblasts, fibrocytes, and histiocytes are embedded in this extensive network of
connective tissue. The connective tissue is slowly regenerated by these cells,
although not completely renewed as the epidermis. The dermis is transversed in

4
various stages by the transport system of the blood and lymph vessels. The system
of capillary vessels extends into the upper dermis, just under the basal membrane,
so that the capillary loops are situated nearer to the surface than the deepest lying
layers of the basalis due to the epidermal undulation. The blood vessel system of
the skin consists of the entering arteries, arterioles, arterial precapillaries, arterial
and venous capillary loops and finally the postcapillaries and venules which
terminate in the skin veins. The nerves also terminate under the epidermis (5).
The dermis is located on the subcutis which is made up of a network of fat
cells. While the collagen fibers of the network form a system of vertical structure
in the female skin, they form additional diagonal cross-striations in the male skin
(7). These fibers arch upwards into the dermis. The subcutis represents a massive
separating layer between the skin and the muscle tissue below.
Structure and Function of the Permeability Barrier of the Skin
"Indeed, the raison d'etre of the epidermis is to make the stratum comeum;
this is its specific biologic mission." (8; p. 387). This process of formation of the
stratum comeum is more correctly called comification rather than keratinization.
In 1877, Fleischer declared that the skin was totally impermeable, a
complete shield against the external world (9). By the turn of the century,
Schwenkenbecker (10) perceived that the skin would admit some substances much
better than others. Later, appreciation for the integrity of the stratum comeum led
to the concept that the full-thickness of this layer is functionally competent (11),

5
which was consistent with the view that it behaved as a homogeneous film (12) (the
"plastic wrap" hypothesis).
Although the importance of lipids for barrier function (13) and the water-
retentive properties (14) of the stratum comeum was well appreciated, it was
Middleton who first noted that it was the organization of lipid into 'shells' that
accounted for water retention (15), thus refuting the homogeneous film concept.
Soon thereafter, the existence of separate hydrophilic and hydrophobic pathways
was suggested from physical-chemical observations (16), followed by morphologic
evidence for lipid-protein segregation. Freeze-fracture and thin-section studies
demonstrated lipid bilayers exclusively in the stratum comeum interstices, with the
absence of lipid structures within the comeocyte cytosol (17-19). In addition,
histochemical and cytochemical studies clearly displayed the process of lipid
sequestration to membrane domains (19,20). This can be further deduced from the
ease with which certain nonpolar solvents disperse this tissue into cell suspension,
and the further ability of such solvent extracts to recombine with dispersed cells to
produce a functionally competent tissue (21).
The first direct evidence for lipid localization to intercellular domains was
provided by biochemical analysis of membrane couplets which were prepared
without loss of intercellular lipid. These preparations contained not only multiple
bilayers identical to intact stratum comeum, but they were also lipid-enriched,
accounting for over 80 % of all the lipid in the stratum comeum (22). Moreover,
they displayed both the same lipid distribution of whole stratum comeum (22), and

6
duplicated the x-ray diffraction pattern ascribed to the 'shell' of lipids, previously
thought to fill interfilamentous domains within the comeocyte cytosol (23). One
can assume that the localization of a variety of lipid catabolic enzymes, steroid
sulfatase (24), lipase (25), sphingomyelinase (25), and phospholipases (25), to the
stratum comeum intercellular domains represents co-localization of hydrolytic
enzymes with their respective lipid substrates in the stratum comeum membrane
domains.
Protein and Lipid Domains in the Human Stratum Comeum
It is now well appreciated that the two-compartment system is formed by the
deposition of epidermal lamellar body contents in intercellular domains at the
granular-comified cell interface (17,18). Subsequent characterization of the lipid
and enzymatic content of lamellar bodies has provided a clear picture of the
molecular events associated with the formation of the stratum corneum
intercellular compartment. Lamellar bodies are enriched in phospholipids (26,
27), free sterols (26), and glycosphingolipids (26), including certain distinctive
acylglycosphingolipid species (28) that may be responsible for the disc-like lipid
bilayers that appear in lamellar bodies (29). Biochemical studies revealed a limited
array of hydrolases (27,30-32), but a striking absence of certain typical lysosomal
enzymes, including P-glucuronidase, galactosidase, and aryl sulfatases A and B
(27,32).

7
The lipid catabolic enzymes that have been demonstrated in lamellar bodies
seem ideally suited to bring about the transformation of the polar lipid contents of
lamellar bodies to the nonpolar species that eventually reside in the stratum
comeum (Figure 1-2) (33). However, the regulation, timing, and location of these
degradative events is still not certain. According to some studies (34,35), if all
phospholipid and glycolipid species are completely absent from the stratum
comeum, then degradation must occur very soon after secretion. Yet others have
found some glycolipids and phospholipids in the lower stratum comeum
(20,30,36), suggesting that lipid transformations in the intercellular spaces may be
a more gradual process.
"Pro Barrier" Lipids:
Glycolipids. Free Sterols,
Phospholipids
Catabolic Enzymes:
Acid Phosphatase,
Proteases, Lipases,
Glycosidases
Conversion of "Pro-Barrier" Lipids to
Non-Polar Products (Lipases, Glucosidases)-
Glycolipids —
Phospholipids
-â–º Ceramides-
FFA<
1) Release of Oesmosomes into Intercellular
Space (Lipases)
2) Degradation of Non-Lipid Intercellular
Species (Acid Phosphatase, Proteases)
Barrier Function
Desquamation
Figure 1-2 : Dual functions of the epidermal lamellar body. Summary of lipid
biochemical and enzyme biochemical studies on isolated epidermal lamellar bodies
(33).

8
In addition to lipid catabolic enzymes, the lamellar body is also enriched in
acid phosphatase (27,31) and proteases (27). Whether these enzymes participate in
the barrier formation is unclear at present, but it is more likely that they participate
in the desquamation process, presumably by degrading extracellular glycoproteins,
desmosomes, etc. (Figure 1-2) (33).
Localization of the Skin-Barrier to Drug-Transport
It was in 1944 that Winsor and Burch first demonstrated the water barrier,
which prevents desiccation of all terrestrial vertebrates, to be localized in the
stratum comeum (37). They sandpapered through human epigastric skin and
observed a sharp rise in the flux of water just as they breached the stratum
comeum. The observation that the stratum comeum represents the barrier to
water loss through the skin was later verified by Blank (38). He used cellophane
tape to strip-off layers of the stratum comeum and observed that water loss did not
increase much until the lowest layers of the stratum comeum were reached.
Therefore, he concluded that the barrier is located at the bottom of the homy layer
(38). These results were subsequently reinterpreted by Scheuplein, who showed
mathematically that Blank’s data (38) could be better explained by a model that
regards all layers of the stratum comeum as rate-limiting with equally good
barrier properties (39).
Early solvent extraction experiments indicated that lipids, particularly the
polar lipids in the epidermis, play a vital role in the barrier properties (40,41).

9
Breathnach and co-workers applied the freeze-fracture technique to show that
these lipids form multiple broad bilayers filling the intercellular spaces (42).
These bilayers were recently shown to be present throughout the homy layer
(43,44), which provide the barrier for water diffusion. This was first
demonstrated by Squier, who used horse-radish peroxidase as a probe to monitor
the movement of water in keratinized and non-keratinized epithelia (45). He
observed that the probe penetrated only the most peripheral layers of the stratum
comeum, when applied topically. But, when it was injected subepithelially it
moved upwards freely through the viable epidermis and stopped just at the bottom
of the stratum comeum, where the first intercellular lamellae are located. These
intercellular lipid sheets seem to provide the barrier to water loss through the skin
and limit the penetration of polar solutes from the environment.
Relationship Between Lipid Structure and Barrier Function
The membranes within the stratum comeum do not contain phospholipids or
the usual assortment of fatty acyl chain structures (46-48), unlike most mammalian
cell-membranes (49), which are composed of a variety of phosphoglycerides,
spingomyelin, and cholesterol, that do not constitute an appreciable barrier to
water or small water-soluble molecules (49). The lipid bilayers in the intercellular
lamellae consist of mainly cholesterol, fatty acids, and ceramides (Table 1-1)
(44,50). The free fatty acids include only 7% of unsaturated species and no methyl-
branched components. The most abundant fatty acids are the 22- and 24-carbon

saturated species (44,51). With the exception of ceramide 1, the hydrophobic
chains in the ceramides (Figure 1-3) (52) are almost entirely straight and saturated.
The double bonds in the sphingosine moieties are located at the polar ends of the
ceramide molecules, so that they do not produce kink or perturb the aliphatic
chains. These lipids seem to be ideally suited to form highly ordered, impermeable
membranes and to resist oxidative damage on exposure to air at the skin surface.
Table 1-1 : Composition of stratum comeum lipids (44,50).
Lipid
Pig
Human
Cholesteryl esters
1.7
10.0
Triglycerides
2.8
0.0
Fatty acids
13.1
9.1
Cholesterol
26.0
26.9
Ceramide 1
4.1
3.2
Ceramide 2
16.7
8.9
Ceramide 3
6.9
4.9
Ceramide 4
4.4
6.1
Ceramide 5
4.5
5.7
Ceramide 6
7.6
12.3
Glucosylceramides
1.0
0.0
Cholesteryl sulfate
3.9
1.9
Others
5.7
11.1
There is considerable chain-length diversity among these fatty acids and
ceramides of the stratum comeum, which results in the interdigitation of chains in

the middle of the hydrophobic region. This interaction may be a significant
stabilizing factor in imparting the barrier properties to the stratum comeum.
Ceramide 1 (Figure 1-3) (52) is of special importance to the stratum
comeum barrier function. This acylceramide is one of the principal carriers of
linoleic acid in the epidermis. It is thought to function in the fusion of the disks
extruded from the lamellar granules and in the stabilization of the resulting
multilamellar sheets. It has been postulated that the long co-hydroxyacyl chain of
the acylsphingolipid completely spans one lipid bilayer, while the ester-linked
linoleate is inserted into the adjacent bilayer, thus serving as a molecular rivet in
linking the two membranes together (53,54). The acylceramide and
acylglucosylceramide have been shown to be capable of promoting stacking,
flattening, and fusion of liposomes (55,56).
However, there is a gradient within the stratum comeum itself. Whereas the
stratum compactum still contains significant levels of phospholipids,
glycosphingolipids, and cholesterol sulfate, only the latter persists in the stratum
disjunctum (Table 1-2, Figure 1-4) (57,58). This change in composition
presumably reflects ongoing metabolic activity.

(CH2)lá-CH3
NH
l— (CH2)29— 02C-(CH2)7—ch=ch-ch2-ch=ch—(CH2)4-CH3
Ceramide 1
0H /V
(CH2)ir-CH3
NH
(CH2)22 ch3
Ceramide 2
OH
°H^CH2)l6_
CHi
OH A
^ NH
\—( ch2)22—ch3
Ceramide 3
(CH2)12 ch3
NH
J-
CHOH—(CH2)23—CH3
Ceramide 4
(CH2)ir-CH3
NH
\—CHOH—(CH2)
13—CH3
Ceramide 5
OH
(CH2)ir-CH3
NH
J—CHOH—(CH2)21—CH3
Ceramide 6
Figure 1-3 : The ceramides of the human stratum comeum (52).

Table 1-2 : Composition of mammalian epidermal lipids (57).
Lipid
Living layers (%)
Stratum comeum (%)
Phospholipids
40
Trace
Sphingolipids
10
35
Cholesterol
15
20
Triglycerides
25
Trace
Fatty acids
5
25
Other
5
10
SG Inner SC Outer SC
LAYER
Figure 1-4 : Changes in lipid composition in the stratum comeum (57).

14
Drug-Delivery to the Skin via Prodrugs
Most pharmaceuticals or drugs are applied to the skin for topical or systemic
absorption (59,60). However, they are absorbed very poorly due to the nature of
the lipid-like barrier of the stratum comeum. A limited number of highly
lipophilic drugs readily partition into the stratum comeum and the rate-limiting
step then transfers to the lower hydrophilic viable layers of the epidermis. The
drug molecule should possess optimal physical-chemical properties, if it has to
demonstrate appreciable penetration across the stratum comeum barrier (61).
The prodrug approach is one method of enhancing the effectiveness of the
drug by improving its penetration into the skin (Figure 1-5) (62,63).
Figure 1-5 : Schematic representation of the use of prodrugs in topical therapy
(63).

15
The fundamental principle governing the improved penetration of
biological membrane barriers by the use of prodrugs is its greater solubility in the
tissue relative to the parent, underivatized drug (62). The use of prodrugs in
dermal delivery has been described (64) to involve the bypass of the membrane
barrier by alteration of the physical-chemical properties of the drug. The concept
(Figure 1-5) (63) emphasizes the diffusion of the prodrug with simultaneous
metabolism. The importance is given to the drug-skin and drug-vehicle
interactions that can be modified by the application of prodrugs (65).
Some classic studies on which to base a rational design for a new prodrug
have examined homologous series of alcohols (66), phenylboronic acids (67),
steroids (68), and salicylate esters (69). These studies form a basis to indicate the
effect of chemical structure on percutaneous absorption, but it is important to
realize that structural modifications may alter both the skin permeability and the
thermodynamic activity of the drug in the vehicle, and these effects cannot be easily
separated.
The most important considerations in the design of a successful prodrug
approach for skin absorption assume that if various promoieties that impart the
desired physical-chemical properties to the drug can be attached to the appropriate
polar, hydrogen-bonding functional groups in the parent drug molecule, then the
attachment should be reversible (70) in vivo, either by chemical or enzymatic
reactions (71) at the appropriate site in the tissue. In addition, if systemic delivery
is not the ultimate goal (for dermal or intradermal targeting), there should be an

1 6
improvement in the delivery of the drug to the target epidermal cells (72), which
may or may not be associated with enhanced transport of the drug across the skin.
However, for systemic delivery, the prodrug approach requires improving
delivery of the drug through the skin.
Various relevant examples that illustrate the use of prodrug systems for
dermal delivery have been based upon esters of acetylsalicylic acid (73),
indomethacin (74) and other non-steroidal antiinflammatory agents. The polar,
nitrogen-containing heterocyclic molecules like 5-fluorouracil, 6-mercaptopurine,
5-fluorocytosine, theophylline have been extensively derivatized into a-
acyloxyalkyl, N-Mannich base, acyl, and a-(bisalkylhetero)alkyl prodrugs for
dermal delivery and were the subject of a recent review (75). It discussed
strategies for the introduction of promoieties onto primary and secondary amines,
amides, imides, hydroxyls, thiols, carbonyls, and carboxylic acid functional
groups. This report (75) also analyzed correlations between solubilities of the
prodrugs and their abilities to deliver the parent molecules across the skin.
Prodrug concepts using local drug-delivery routes specifically to the skin
have been reviewed (65). The concepts for improved dermal delivery to local
target tissues with the use of prodrugs are similar to those for systemic drug
delivery as well as to achieve transdermal (systemic) drug absorption. However,
true site-specificity in drug-delivery to the skin is only possible when the active
drug can be retained by the target site in the microcompartment of the skin.

Thus, the acyloxymethyltheophylline derivatives were observed to deliver
theophylline to the target epidermal cells based upon the results from the inhibition
of epidermal DNA synthesis (75). In addition, the rate of delivery of theophylline
through the skin (given by the in vitro diffusion-cell experiments) correlated with
the delivery of theophylline to the target epidermal cells (75).
However, as outlined above, if the goal is drug-delivery through the skin for
the purpose of systemic absorption, then specific targetting of active sites in the
skin tissue is not desired. Therefore, the physical-chemical properties of the
prodrug that influence its ability to completely penetrate the skin barrier and allow
the release of the parent drug in the systemic microcapillaries becomes an
important prerequisite.
It has been demonstrated conclusively, that if systemic absorption is the goal
of drug-delivery through the skin via prodrugs then increased water and lipid
solubility of the prodrug is necessary (75).
Drug Metabolism in the Skin
The viable epidermal layer, which is below the stratum comeum, is the most
metabolically active layer in the skin (76). Any substance that penetrates the
stratum comeum is subjected to the drug-metabolizing properties of the viable
epidermis (Figure 1-6) (77). This is particularly important to transdermal
delivery because the first-pass effect is now transferred to the skin. The specific
enzyme activity within the skin has been shown to approach and sometimes exceed

1 8
Figure 1-6 : Schematic representation of the skin as a metabolic barrier (E =
enzyme) (77).
that of the corresponding hepatic enzyme (78-81). This metabolic role for the skin
was supported by bioavailability studies with topical glyceryl trinitrate (82) and
efficacy studies with topical cortisol (83). It has been suggested that the rate of
metabolism of benzo[a]pyrene within the skin is the rate-limiting step when
considering percutaneous absorption (84,85).
The potential biotransformation reactions which are known to occur within
the skin are shown in Table 1-3 (77). They are a series of functionalization (phase
I) reactions (oxidations, reductions, hydrolysis) and conjugation (phase II)
reactions (glucuronide and sulfate formation, methylation and glutathione
conjugation).

The enzyme systems in the skin are highly inducible. Cytochrome P-450,
present in the skin at low concentrations, is inducible by topical application of some
agents that also induce hepatic metabolism (86). After 2,3,7,8-tetrachlorodibenzo-
p-dioxin (TCDD) exposure, the activity of aromatic hydrocarbon hydroxylase
increases upto 30-fold and that of 7-ethoxycoumarin deethylase increases 6-fold
(87). Since the skin possesses many of the enzymes that the liver does, it would be
interesting to compare their relative activities. Generally, the activities of the
enzymes in the skin are low compared to that in the liver, typically about 2-6 % of
the hepatic values (87). However, the cutaneous enzyme activities reported were
mostly based on whole skin homogenates. Assuming that these enzymes are located
in the epidermal layer (76,88-90), their true activities range from 80-240 % of
those in the liver (87).
Thus, if a drug diffuses slowly through the epidermis, the skin may serve as
a site of first-pass metabolism. Such metabolism may decrease both the amount of
drug at the local (intradermal) site of action and the amount systemically available.
On the other hand, if absorption is fast, the cutaneous enzymes may become
saturated, in which case a significant amount of the drug may be absorbed into the
systemic circulation without being metabolized (81). Therefore, cutaneous
metabolism can significantly influence drug-delivery into and through the skin.

20
Table 1-3 : Biotransformation reactions by human skin (77).
A. Phase I - Reactions
Reaction
Enzymes involved
Substrate
1. Oxidation
1.1. of aliphatic C-atoms
Mixed function oxidase
7,12-Dimethylbenz(a)anthrac-
ene (DMBA)
1.2. of alicyclic C-atoms
Mixed function oxidase
Dehydroepiandrosterone
(DHA) — 7aOH-DHA,
7pOH-DHA
1.3. of aromatic rings
Hydroxylases
3,4-Benzopyrene (BP) -â–º
Phenol, Quinone, Dihydrodiol
1.4. of alcohols
Hydroxysteroid dehydro¬
genases
Cortisol Cortisone
Testosterone -â–º A4-Androst-
ene-3,17-dione
17P-Estradiol Estrone
1.5. under deamination
MAO
Norepinephrine
1.6. under dealkylation
Deethylase
Demethylase
7-Ethoxycoumarin
Aminopyrine
2. Reduction
2.1. of carbonyl groups
Ketoreductase
Cortisol Reichstein's,
epi-E
Progesterone (Allo)preg-
nanediol
Estrone 17P-Estradiol,
5a-DHT, 5a-Androstane-3a,
17p-diol

21
Table 1-3 : Continued
A. Phase I - Reactions
Reaction
Enzymes involved
Substrate
2.1. of carbonyl groups
Ketoreductase
5 a-Andros tane-3,17 -dione
-•-Androsterone, Epiandro-
sterone
2.2. of -C=C-double
bonds
5 a-Reductase
Testosterone -â–º 5a-DHT
Progesterone —► 5a-DHP
3. Hydrolysis
3.1. of ester bonds
Esterases
Diflucortolone-21 -valerate
Betamethasone-21 -valerate
Betamethasone-17-valerate
Rucortin butyl ester
3.2. of epoxides
Epoxidehydratase (EH)
Styrene oxide
B. Phase II - Reactions
1. Glucuronide formation
UDPG-transferase
Benz(a)pyrene,
O-aminophenol
2. Sulfate formation
Sulfo-transferase
DHA, A5-Androstene-3(3,17p-
diol
3. Methylation
COMT
Norepinephrine
4. Glutathione-conjugation
Glutathione-S-transferase
Styrene glycol

22
Methods for Estimating Transport of Drugs Into and Through the Skin
In vitro procedures allow the determination of rate of absorption of the
drug molecule directly below the skin membrane, where it is physiologically
improtant. Errors in extrapolating from in vivo rates of urinary excretion are
avoided. For highly toxic compounds, in vitro methodology may be the only way
of obtaining percutaneous absorption data with human skin. The human skin is
unique and no animal model is entirely suitable. The in vitro experiments can be
done with much less effort and in greater numbers because of the simplicity in
methodology.
The three most important considerations in the design of in vitro procedures
are :
(a) The choice of the skin membrane (animal or human) and its preparation.
(b) Appropriately calibrated in vitro diffusion apparatus.
(c) Specific and sensitive analytical method.
Two-chambered diffusion cells (Figure 1-7) have been used since early
times so that a chemical could be applied on one side of the membrane and its rate
of permeation could be obtained from sampling the identical solvent (usually water
or saline) on the other side (68). The donor as well as the receiver side was stirred
to ensure homogeneity of the solutions.
To simulate in vivo situation, one-chambered cells (receptor beneath the
skin) are now commonly used. The investigator can apply the test compound to the

23
SIDE-BI-SIDE- DIFFUSION CELL SHOWN ACTUAL SIZE
Figure 1-7 : Two-compartment, side-by-side, in vitro diffusion cells.
skin in a vehicle of choice and also maintain the skin surface at ambient condition of
hydration. Thus, the static cell design of Franz (91) and similar designs have been
widely used (92). The equations that appropriately describe passive drug-transport
across the skin barrier are based upon Fick's laws of diffusion and are defined as
follows in Figure 1-8 (16).

1.
J = (Km x D / 5) Cd = Kp x Cd
D = 82 / 6 x tiag
24
2.
3.
Km — Cm / C
V
J = Flux of the drug across the membrane (mol/cm2/h).
Km = Partition coefficient of the drug in the membrane.
D = Diffusion coefficient of the drug in the membrane (cm2/h).
Kp = Permeability coefficient of the membrane for the drug (cm/h).
Q= Concentration of the drug in the donor (mol/ml).
8 = Thickness of the membrane (cm),
tiag = Lag-time for diffusion (h).
Cm = Concentration of the drug in the membrane (mol/ml).
Cv = Concentration of the drug in the vehicle (mol/ml).
Figure 1-8 : Equations used for calculating skin-permeability (16).
Since enough human skin for large number of permeability experiments is
usually not available, it is important to study and compare the permeabilities of
human and animal skin. The skin of the rabbit and the mouse is usually most
permeable. The hairless-mouse skin permeability has been reported to be similar
to human skin for some compounds (93,94). However, many apparent
discrepancies between the permeability coefficients of various compounds in
human and animal skin can be explained by the structural differences in the
membranes. Bronaugh et al. (95) have shown that the stratum comeum of the rat
skin is as thick as the human skin. For compounds that penetrate rapidly and do not

25
rely on appendageal diffusion, it may be a good model for human skin, but its many
hair follicles make it a poor model for polar, water-soluble compounds.
Conversely, hairless-mouse skin has a similar hair follicle density, but the stratum
comeum is thinner than the human skin (95) and so it is often more permeable.
The skin of the monkey at the abdominal and ventral forearm testing sites is
sparsely haired, and has been found to be a good model for human skin in
numerous in vivo studies (96,97).
The preparation of the skin membrane is a critical step in the in vitro
experiments. In humans, pigs, rats, and guinea pigs, the dermis is 2-3 mm thick
compared to the epidermis, which is approximately 50-100 mm. If full-thickness
skin is used in the diffusion-cell studies, the thick dermal tissue can present an
artificial barrier, particularly for water-insoluble compounds. Compounds that
are absorbed through the skin in vivo are taken up by blood vessels directly beneath
the epidermis, so they are not required to penetrate the full-thickness of the skin.
Therefore, for relatively hairless-skin, the preparation of an epidermal membrane
by heat separation has been a convenient solution to the problem of skin thickness
(92). Hairy animal skin cannot be separated by this technique because the shafts of
hair leave holes in the epidermis when it is peeled away. In studies with animal skin
that is relatively thin (1 mm or less) such as the mouse, hairless-mouse, and the
rabbit, the preparation of a split-thickness skin using a dermatome is difficult, and
is probably not required.

26
The basic data for in vivo human percutaneous absorption, to which animal
models are compared, were obtained from Feldmann and Maibach (98-100). In
these clinical studies, a specific concentration of radioactive compound (4 mg/cm2)
was applied to a specific anatomical site (the ventral forearm). The area was not
occluded, and the subjects were requested not to wash the area for 24 hours. The
radioactive compounds were applied to the skin in an acetone solution and the
acetone quickly evaporated with a gentle stream of air. Urine was collected for 5
days and assayed for radioactivity. A tracer dose was also given parenterally, and
the percentage of radioactivity in the urine following parenteral administration
was used to correct for the compound that might be excreted by some other route
and for the compound that might be retained within the body.
Comparative in vitro studies show that generally, the skin of common
laboratory animals (mouse, rabbit, rat, and guinea pig) is more permeable than the
skin of man. Skin from the pig and monkey more generally approximates the
permeability of human skin (101,102). Hence, in vitro and in vivo studies with the
pig and monkey skin were observed to correlate well with the in vivo human
percutaneous absorption studies (103,104).
A compound with limited water solubility must be examined carefully when
using in vitro diffusion-cell techniques. This type of substance may seem to
penetrate skin only slightly, when, in fact, the rate-limiting step is not the
penetration into the skin but partitioning from the skin into the aqueous receiver
fluid. Under in vivo conditions, hydrophobic compounds that penetrate the skin

27
are taken up and carried away by the blood in the capillary loops immediately
below the epidermis much earlier than they would appear in the receiver fluid in an
in vitro set-up, specially if the receiver is normal saline or a physiological buffer
solution. It is quite likely that the hydrophobic substance may choose to remain in
the skin rather than partition into the aqueous receiver solution.
The physical-chemical properties of a molecule, in particular its solubility
in oil and water, have been compared with the permeability data in the hope of
finding correlations. Usually, the solubility is expressed as an oil/water ratio (or
its logarithm), and the octanol/water partition coefficients for a number of
compounds have been published by Hansch and Leo (105). Reasonably good
correlations of permeability and oil/water partition coefficients have been obtained
with some homologous series of compounds. If they are applied as dilute solutions
in aqueous vehicles, positive correlations have resulted (106-108). This result is
due, at least in part, to an increased driving force for the drug from the aqueous
vehicle caused by the increase in lipid solubility. Of course, the absolute solubility
in the aqueous vehicle is critical to permeation but is ignored in expressing
solubility data as partition coefficients.
When compounds are applied undiluted, a number of studies have found that
the best correlations have been positive with water solubility (109-111). Under
these conditions, the importance of water solubility of the molecule in promoting
good skin permeation is emphasized rather than its ability to partition from water
into oil.

28
Application of compounds in a saturated solution should, in theory,
overcome the influence of the vehicle on percutaneous absorption. When the
permeation of structurally different compounds was determined from saturated
aqueous solutions, no correlation of skin absorption with mineral oil/water
partition coefficients was observed (16). It is likely that relying on partitioning
data to estimate skin absorption will often be misleading, since other determinants
such as the effect of the vehicle and skin binding are not considered.

CHAPTER H
SIGNIFICANCE OF THE STUDY
The present investigation concentrated on two major aspects pertaining to
transdermal drug-transport and intradermal drug-targeting :
(a) The use of novel redox chemical targeting systems for improving the
delivery of acyclovir (a model antiviral agent) and 5-fluorouracil (a model
anticancer agent) to the skin. In this section, the concept of intradermal
drug-targeting was emphasized to localize the parent drug species in the
skin.
(b) The development and use of the Immobilized Artificial Membrane (IAM)
HPLC column chromatographic model system for studying drug-transport
across human skin as a model biological membrane barrier.
Delivery of Acyclovir To the Skin
Acyclovir (acycloguanosine; 9-(2-hydroxyethoxymethyl)guanine) is an
acyclic, synthetic analog (Figure 2-1) of a naturally occurring purine nucleoside
and exhibits potent in vitro antiviral activity against herpes simplex virus (HSV) of
types I and II (112,113). It is also active against varicella zoster virus (VZV) (114)
and human cytomegalovirus (HCMV) (115). The HSV have the ability to code for
29

30
the specific thymidine kinase, which is capable of phosphorylating acyclovir to a
monophosphate (116,117).
Figure 2-1 : Chemical structure of acyclovir.
This capability is essentially absent in uninfected cells and therefore,
acyclovir exhibits high potency and selectivity for HSV infected cells and low
toxicity for the normal host cells (112,114). The monophosphate is then
phosphorylated to the diphosphate via cellular guanylate kinase (GMP) and then to
the triphosphate by other cellular phosphorylating enzymes (112,118). Acyclo-
GTP is a more potent inhibitor of the viral DNA polymerase than of the cellular
DNA polymerase (119). This cellular activation of acyclovir is shown in Figure 2-
2. The viral DNA polymerases use acyclo-GTP as a substrate and incorporate it
into the DNA primer-template to a much greater extent than do the cellular DNA
polymerases (120).

31
Thus, the viral enzyme catalyzes the insertion of the false purine base into
the viral DNA, bringing it to a premature end, since acyclovir does not possess a 3'-
hydroxyl group, which is obligatory for chain elongation.
Acyclovir
HSV - coded
Thymidine Kinase
Acyclo - GMP
GMP
Kinase
O
Acyclo - GTP
Cellular Enzymes
Acyclo - GDP
Figure 2-2 : Cellular activation of acyclovir.

32
The acyclo-GTP inhibits viral DNA polymerase with a tight binding
constant of 1.6 x 10*9 M (121). Acyclovir is active to a similar extent against HSV I
and II with 50 % inhibition of viral plaque formation at concentrations of 0.06 to
1.8 x 10-6 M for type I and 0.65 to 1.8 x 10-6 M for type II (122). Since two
enzymes are involved in the antiviral activity of acyclovir, resistance can and has
occurred due to either (a) absence of thymidine kinase or its altered sensitivity to
substrates and to the drug (123) or (b) in the ultimate target, the DNA polymerase
(122). Mutants of HSV I without the thymidine kinase are apparently not as
virulent as the wild-type, although those that still have the altered enzyme, are just
as virulent (124).
The topical (local) delivery of acyclovir is an important consideration in the
treatment of cutaneous herpes simplex viral infections of type I. The HSV are
capable of infecting almost any cutaneous site, including the scalp, toes, knees,
elbow, and hand (125). If the patient's recurrences are mild, then topical
formulations are favorable for the treatment of recurrent herpes labialis (125).
Unfortunately, the clinical trials conducted with acyclovir 5 % ointment have
failed to show efficacy (126-128). The modified aqueous cream with 5 %
acyclovir has shown some promise in a recent trial (125), but no definite efficacy.
Gibson et al. (129) had previously demonstrated efficacy in a small prophylactic
cream trial, and Fiddian et al. (130) had originally shown efficacy in an earlier
trial in England.

33
The reports on the ineffectiveness of the drug in animals and in humans in
certain therapeutic situations has been attributed to the major problem of delivery
(131,132) of acyclovir across the stratum comeum to reach the infection site in the
epidermal cells of the skin. Acyclovir has several non-optimal physical-chemical
properties which present barriers to its effective delivery. It has low aqueous (1.2
mg/ml at pH 7.0 and 250Q (133) and low lipid (log P = -1.47; octanol - buffer
partition coefficient) (141) solubility. Thus, it shows inadequate skin partitioning
ability which could lead to a large concentration drop from the stratum comeum
surface to the basal layer of the epidermis when the drug is topically applied.
Hence, improved penetration of acyclovir through the stratum comeum and the
epidermal layers (Figure 1-1) has been recommended (132).
The 6-deoxy-6-amino derivative (Figure 2-3) (134,135) of acyclovir has
been studied as a prodrug to improve the oral bioavailability of acyclovir (134). It
is deaminated to acyclovir by adenosine deaminase (135). However, it resulted in
only modest increases in acyclovir plasma levels relative to those achieved by
acyclovir upon oral dosing of rats and dogs (134).
Krenitsky et al: (136) made the 6-deoxyacyclovir (Figure 2-3). This
prodrug was found to be 18 times more water-soluble than acyclovir and is rapidly
oxidized by xanthine oxidase to acyclovir in vivo. After oral administration to rats
and human volunteers, it was rapidly absorbed and resulted into 5-6 times greater
bioavailability than acyclovir (136-138).

34
6-deoxyacyclovir
6-deoxy-6-
aminoacyclovir
Acyclovir
Figure 2-3 : Prodrugs of acyclovir with improved oral absorption (134,135).
The compound is also susceptible to undergo oxidation by aldehyde oxidase
to give the inactive 8-hydroxy-6-deoxyacyclovir (Figure 2-3) (134,135), but this
non-activating oxidation plays only a minor role in comparison to the activating
oxidation by xanthine oxidase to acyclovir (136).

35
Highly water-soluble esters as prodrugs of acyclovir (Figure 2-4) (139)
have been described which may permit administration of large quantities of the
drug for either topical use as eye drops or parenteral administration.
Intramolecular complex formation with acyclovir to improve its water solubility
for developing parenteral formulations have been attempted (140).
1, R = C0CH2NH2.HC1
2, R = COCH(CH3)NH2.HCl
3, R = C0CH2CH2NH2.HC1
4, R = COCH2CH2COONa
5, R = COCH2N3
6, R = COCH(CH3)NHCOOCH2C6H5
7, R = COCH2CH2NHCOOCH2C6H5
8, R = COCH2CH2COOH
Figure 2-4 : Water-soluble ester prodrugs of acyclovir (139).

36
However, it is unlikely that they would improve the delivery of the antiviral
agent across the membrane barriers since there is no significant improvement in
their lipid-solubility over acyclovir.
Examples of highly water-soluble prodrugs of acyclovir (Figure 2-5) (141)
with high susceptibility to undergo enzyme catalyzed hydrolysis were recently
described (141).
N
>
HN
h2n^n^n
l,°v^0R
I R = H
II R = £-{3
CH2N(CH3)2
III R=C-4 CH2N(CH3)2
V R =
VI R =
yj— ch2n(c2h5)2
o _/CH2N(C3H7)2
II /=\
o
II /=>
vn R=c—^ ^-ch2n(c3h7)2
vm r =
IX
o /CH2
ii /=\
o
R" c~\^f
X R= C-^^-CH2f<^0
o
o
xi R= ^0”ch
Figure 2-5 : (Alkylaminomethyl) benzoate ester prodrugs of acyclovir (141).

37
These prodrugs which are (alkylaminomethyl) benzoate esters of acyclovir
showed increased hydrophilicity (excess of 10 % w/v) and lipophilicity compared
to that of acyclovir (log P = - 1.47) as determined by their partitioning between n-
octanol and 0.05 M phosphate buffer, pH 7.4 (Table 2-1) (141).
Table 2-1 : Lipophilicity of the (alkylaminomethyl) benzoate ester prodrugs of
acyclovir and their susceptibility to enzymatic hydrolysis in human plasma (141).
Ester
11/2 in human plasma
(min)
logP
Acyclovir
-1.47
II
7.0
-0.94
III
33
-0.92
IV
7.5
-0.37
V
25
-0.35
VI
0.8
0.60
VII
57
0.60
VIII
2.3
1.50
IX
4.6
-0.04
X
3.7
-0.05
XI
8.5
-0.11
According to Table 2-1, there is no correlation between lipophilicity of the
various synthesized (alkylaminomethyl) benzoate ester prodrugs of acyclovir and
their ability to hydrolyze in human plasma (due to the esterases). However, there
appears to be a significant difference in their susceptibilities towards hydrolysis by
the plasma esterases depending on the position of the alkylaminomethyl group on

38
the benzoate moiety. Structures II, IV, VI, and VIII (Figure 2-5) with the
alkylaminomethyl group at the 3 position on the benzoate group, were observed to
hydrolyze faster (Table 2-1) than the corresponding molecules with identical
substitutions on the 4 position (structures III, V, and VII in Figure 2-5). Thus, the
alkylaminomethyl substitution at position 3 may allow easy approach by the plasma
esterase enzymes to readily catalyze hydrolysis of those substrates whereas, the
corresponding alkylaminomethyl substitution at position 4 may be hindering the
enzyme approach. This indicates the sensitivity of changes in the structural
configuration of these esters to their susceptibility towards hydrolysis by the
plasma esterases.
Thus, to overcome the major drawbacks of acyclovir, namely its
ineffectiveness against recurrent disease (142,143) due to poor human skin
penetration (144), we have been involved in the design, synthesis, and evaluation of
novel chemical targeting systems for this important antiviral agent, which would
(a) carry acyclovir into the skin due to their better partitioning ability, and (b) use
the metabolic ability of the skin to trap the drug-delivery systems for a sustained
release of acyclovir in the skin. The objective is to deliver the drug to and not
through the skin so as to achieve high local skin concentration of acyclovir where
the herpes simplex virus infects the epidermal skin cells.
In our laboratory, it was demonstrated that the spirothiazolidine derivatives
of endogenous steroid hormones like hydrocortisone (73,145), testosterone, and
progesterone (146) could offer unique advantages in delivering the parent drugs

39
specifically to the skin. Natural amino acid cysteine and its derivatives were used
for their synthesis (Figure 2-6) (145) and the compounds were found to be
susceptible towards enzymatic hydrolysis of the likely iminium ion intermediate
(Figure 2-6) (145) formed after spontaneous cleavage of the carbon-sulfur bond
(147).
It was observed that the thiazolidine derivatives of potent steroidal
hormones readily partitioned into the skin, followed by ring-opening to form the
iminium ion intermediate which then binds to the skin as shown in Figure 2-6 (145)
followed by sustained release of the parent hormone in the skin. (73,145,146).
Thus, it was perceived that to achieve high local skin concentration of
acyclovir to improve its clinical effectiveness, it is necessary to design carrier
mechanisms to lock the precursors in the form of chemical reservoirs, from which
there would be a sustained release of the antiviral agent in the skin.
A 1,4-dihydrotrigonelline moiety containing ester of acyclovir, A-CDS,
was previously designed and successfully synthesized (148,149) to serve as a redox
chemical delivery system for improving the delivery of the antiviral agent to the
brain. Two important features of this redox system were noted. The relative
lipophilicity studies based on HPLC elution indicated that the A-CDS was about 30
times (log P = 2.12) more lipophilic than acyclovir (log P = 0.64). Hence, it
demonstrated facile penetration across the blood-brain barrier. In addition, it was
rapidly oxidized to the quaternary metabolite, AQ+, effectively locking this polar
species inside the rat brain, from which slow release of acyclovir in the brain was

40
CO^Hj
CO2C2H5
3-Spirothiazolidine
derivative
SKIN
Parent Steroid
Hormone
Figure 2-6 : Delivery of steroid hormones to the skin using 3-spirothiazolidine
derivatives (145).

41
observed. It was this dramatic separation between the physical-chemical properties
of A-CDS and its oxidized metabolite, AQ +, which led to an increased flux of A-
CDS into the brain, and a decreased efflux of AQ+ from the brain.
Thus, the idea of using A-CDS for the delivery of acyclovir to the skin was
based on the above findings (Figure 2-7). It was thought that this novel approach
towards dermal delivery would not only exploit the favourable dual physical-
chemical properties of the carrier system described, but also the metabolic ability
of the skin, as a biomembrane. Therefore, the possibility of enhancing skin
concentration of AQ+ and consequently that of A was investigated (150).
The principle (Figure 2-7) of delivering acyclovir to the skin is based upon
administering the relatively more lipophilic A-CDS to the skin. After facile
penetration across the stratum comeum, the A-CDS would rapidly oxidize to form
the quaternary metabolite, AQ+, in the skin due to the presence of oxidative
enzymes analogous to the NAD+/NADH redox co-enzyme systems (151,152). AQ+
would then be slowly hydrolyzed by the non-specific esterases to release acyclovir
in the skin. The net result would be accumulation of AQ + from which there would
be a sustained release of the antiviral agent in the skin. The released acyclovir
would then undergo activation leading to the formation of the active triphosphate
form of acyclovir, resulting in the inhibition of HSV DNA polymerase and hence
its replication. Finally, acyclovir and the charged carrier Q+ will be systemically
absorbed and rapidly eliminated due to their polar characteristics.

42
I
ch3
Elimination via systemic circulation
Figure 2-7 : Delivery of acyclovir to the skin using A-CDS.

43
Another carrier system, the lipoic-acid ester of acyclovir (A-LipS2) was
designed. Its delivery to the skin, however, is based upon a principle (Figure 2-8)
similar to that of the spirothiazolidine derivatives previously discussed
(73,145,146). After rapid skin-penetration A-LipS2 is expected to undergo a
reductive disulfide ring-opening by entering the lipoamide-dehydrogenase redox
cycle (Figure 2-8). Due to this interaction with the active site of the enzyme
system, the free thiol is then capable of forming intermolecular disulfide bonds
within the skin. This is followed by the release of the antiviral agent in the skin due
to the hydrolytic action of the non-specific esterases or the lipases (Figure 2-8).
The A-LipS2, after rapid partitioning into the skin is assumed to enter the
lipoamide dehydrogenase redox co-enzyme cycle (Figure 2-9) (153) via reduction
or simple charge transfer to open the intramolecular disulfide bond, followed by
an intermolecular disulfide covalent bond formation in the skin (Figure 2-8).
Acyclovir would then be slowly released in the skin due to the hydrolytic action of
either non-specific esterases or lipases (Figure 2-8).
Thus, both the carrier systems, A-CDS based on oxidation and A-LipS2
based on reduction, were expected to generate metabolic reservoirs for sustained
release of the antiviral agent, acyclovir in the skin, for improving its dermal
delivery.

44
FAST
O
IN PG (SATURATED SOLUTION)
O
ii /\y
(CH2)400'
SLOW
ESTERASES
SKIN
ACYCLOVIR
k \/s
OH
1
INTRACELLULAR ACTIVATION
I
INHIBITION OF VIRAL
DNA POLYMERASE
LIP (SH)
I
INHIBITION OF VIRAL REPLICATION
SYSTEMIC ELIMINATION
Y
ENTERS THE NORMAL LIPOAMIDE
DEHYDROGENASE CYCLE
+ NAD ^— - LIP (S)2 + NADH
+ H+
Figure 2-8 : Delivery of acyclovir to the skin using A-LipS2.

45
Lip(SH)2 + NAD+ Lip(S)2 + NADH + H+
Figure 2-9 : Working hypothesis mechanism for lipoamide dehydrogenase (153).
Delivery of 5-Fluorouracil To the Skin
5-Fluorouracil (5-FU) (Figure 2-10) is used as an antitumor agent
(154,155). It belongs to the class of fluorinated pyrimidines whose
antiproliferative activity is due to its nucleotides 5-fluorouridine triphosphate and

46
5-fluorodeoxyuridine monophosphate (154). The conversion of 5-FU to these
cytotoxic anabolites occurs intracellularly. However, when 5-FU is used to treat
solar keratoses and multiple superficial basal cell carcinomas (156) of the skin, it
demonstrates minimal partitioning across the stratum comeum, due to poor lipid
solubility.
o
Figure 2-10 : Chemical structure of 5-Fluorouracil.
The antitumor activity, affinity, and toxicity of 5-FU has been modified by
the introduction of a carbamoyl group (157-159). l-Hexylcarbamoyl-5-FU
(HCFU) was synthesized as a masked form of 5-FU (157) and was found to be more
active than 5-FU against experimental solid tumors in mice following oral
administration. Further, acute, sub-acute, and chronic toxicity tests have shown
that the HCFU is less toxic than FU. Hence, various 1-alkylcarbamoyl derivatives
of 5-FU have been synthesized (157,160) mainly for oral administration and some
of them have been evaluated for their transdermal delivery potential (161).

47
O
0=L.NH—(CH2)3-CCk—(CH2)5
Lipolyl ester of CPCFU
SKIN
O
F
ESTERASES
O^-NFMCH^-cCfc—(CH2)5
Lipolyl ester of CPCFU skin
covalently bound in the skin
5-FU
O
0=LnH-(CH2)3-CQ2H
CPCFU
HYDROLYSIS
CÜ2
H2N- (CH2)3-C02H
4-Aminobutyric acid
Figure 2-11 : Delivery of 5-Fluorouracil to the skin using 5-FU-LipS2.

48
Therefore, a relatively more lipophilic derivative, the lipolyl ester of
carboxypropylcarbamoyl 5-FU (5-FU-LipS2) was synthesized and evaluated for its
ability to deliver 5-FU to the skin. The principle of delivery (Figure 2-11) being
similar to that of A-LipS2 discussed earlier is outlined below.
Development of the 1AM.PC Column
The use of in vitro models to understand and predict complex, in vivo,
biological phenomena is justified by virtue of their simplicity. If appropriately
designed, they can provide time conserving and inexpensive alternatives to
experiments using animals and humans.
Cell-membranes (Figure 2-12) interact with virtually every type of
biomolecule (162). All solutes including drugs, sugars, amino acids, peptides, etc.,
interact with the components of cell-membranes during their uptake andtransport.
The association of molecules with membranes can be due to either specific or non¬
specific interactions (163).
Specific interactions similar to those in affinity chromatography, involve
polar interactions between the solute and some part or parts of the membrane. Non¬
specific binding, on the other hand, requires solubilization of the molecule (like in
partition chromatography) in the hydrophobic environment of the bilayer.
Phospholipid headgroups contribute substantial molecular volume to the
membrane interface because lecithin is the major membrane forming lipid in most
cells.

49
Figure 2-12 : Schematic representation of a cell-membrane.
These lipid headgroups are barriers (164) to transport of some solutes, but can be
significant binding site to others. This type of molecular selectivity which occurs
in biomembranes was the basis for developing the solid chromatographic support
composed of a membrane lipid (165).

50
Thus, the IAM.PC column consisting of a monolayer of an analog of lecithin
(monomyristoyl lysolecithin) as the stationary phase, chemically (covalently)
bonded to silica through propylamine (Figure 2-13) (166) was synthesized (Figure
2-14) (165).
Figure 2-13 : The IAM.PC column stationary phase (166)

51
1,12-Dodecanedicar boxy lie
acid + DCC
THF, 25°C
18 h
Dodecanedicarboxylic acid
anhydride (cyclic anhydride)
CHC13, 25°C Monomyristoyl
DMAP, 48 h lysolecithin
Lecithin-imidazolide
CHC13> 25°C, 1-2 h
1 Lecithin-COOH
Carbonyldiimidazole
CHCI3,25°C Nucleosil-300
18-24 h (7NH2)
Nucleosil-Lecithin
(IAM.PC column stationary phase)
Figure 2-14 : Synthesis of nucleosil-lecithin (the IAM.PC column stationary phase)
(165).
Attempts to mimic or reproduce the cell-membrane environment on a solid
support, such as the IAM column, could lead to a potential in vitro model to study
molecular processes occurring in natural membranes. From a practical standpoint,
it can be industrially used for rapid screening of various pharmacological agents.
Therefore, this dissertation describes efforts towards developing the IAM
column for studying drug-membrane interactions and hence, its applicability in

52
predicting drug-transport across human skin as a model biological membrane
barrier.
The elucidation of transdermal-transport of various drug molecules across
the human epidermal membrane has become an important part of pharmaceutical
science in recent years. This is particularly because many pharmaceutical
industries are interested in delivering potent biologically active drug-molecules
through the skin for their direct availability in the systemic circulation (the central
compartment). Thus, the drugs could be protected from the undesired first-pass
effect due to metabolism by the liver upon oral administration. Therefore, it is
desirable to develop an in vitro model system which would easily predict the
transdermal availability of drugs.
In addition, the epidermal cell-membranes (Figure 1-1) (2) have been
demonstrated to contain significant amount of phospholipids (57,58). Therefore,
the correlation of drug-transport across the human epidermal membrane with the
retention of the molecules in the IAM.PC column could be justified.
Various membrane models and their advantages or disadvantages for skin
penetration studies were described in a review (167) including excised animal and
human skin, artificial (eg. polymeric sheets) and natural (eg. egg shell)
membranes. However, it was concluded that most of these models fall short of
actually mimicking the in vivo percutaneous absorption process.
Cell culture offers the ideal combination of uniformity and reproducibility
of an artificial system while retaining the biological significance of a recently

53
excised piece of skin tissue. Hence, human skin keratinocyte cultures have allowed
formation of a fully differentiated epidermal layer including a morphologically
distinct stratum comeum (168). An alternative approach of similar attractiveness
would be a validated in vivo isolated perfused skin model (169).
The use of the rotating diffusion cell and associated lipid membrane as an in
vitro model system for percutaneous absorption has been documented (170). The
artificial membrane was formed with isopropyl myristate (IPM), a lipid chosen to
be representative of those in the stratum comeum. The transport resistances of this
membrane to 8 model penetrants were contrasted with those for excised human
cadaver skin. The results indicated that, although some degree of correlation
between the two was evident, the predictability of the IPM membrane model could
be improved (170). Therefore, in another study (171) three alternative lipid
models were chosen to mimick the stratum comeum (the epidermal penetration
barrier). These were dipalmitoyl phosphatidylcholine, linoleic acid, and
tetradecane. For the permeants with diverse physico-chemical properties, the
tetradecane membrane appeared to offer the best correlation with human skin
(171).
Although phosphatidylcholine-coated silica was reported (172) as a useful
stationary phase for HPLC determination of partition coefficients between octanol
and water, such physically adsorbed lipids on a silica support would result into
decreased stability of the stationary phase. Other examples of using phospholipids
in liposome systems (173,174) to estimate the potential partitioning of various

54
solutes into biological membranes have been reported. However, no literature
exists on the use of covalently bonded phosphatidylcholine molecules onto solid
silica surface as a HPLC model system for predicting drug-transport across human
skin.
Specific Aims
Thus, the main objectives of this dissertation could be stated as follows :
1. To design and synthesize redox-based chemical targeting systems for
acyclovir and 5-fluorouracil with improved skin partitioning ability.
2. To evaluate the ability of these targetting systems to deliver the parent drugs
specifically to the skin. Thus, to improve their dermal delivery.
3. To develop the I AM column as an in vitro model system for predicting drug-
transport across human skin.

CHAPTER III
EXPERIMENTAL
Materials
Chemicals of reagent grade were purchased from Aldrich (Milwaukee, WI)
or Sigma (St. Louis, MO) chemical company and were used without further
purification. Ultraviolet spectra were recorded on a Cary 210 Spectrophoto-meter
and the Proton Nuclear Magnetic Resonance spectra on a Varían EM 390
Spectrometer. Chemical shifts were reported as parts per million (5) relative to an
internal standard, tetramethylsilane. Infrared spectra were recorded on a
Beckman Acculab MX 620 double-beam Spectrophotometer. Elemental Analyses
were performed by Atlantic Microlabs (Atlanta, GA). Melting points were
determined using electrothermal melting point apparatus and were uncorrected.
Acyclovir was obtained from Burroughs-Wellcome Co., North Carolina. The
IAM.PC or IAM.PC.MG capped high pressure liquid chromatographic columns
(12 p.m, 10 cm x 4.6 mm i.d.) was obtained from Regis Chemical Co., Morton
Grove, IL.
55

56
Methods
Synthesis of 1-methvl pvridine-3-carboxvlic acid anhydride diiodide ill
Nicotinic acid was reacted with phosgene in benzene in the presence of
triethylamine to form the pyridine 3-carboxylic acid anhydride. This reaction was
carried out according to the previously published literature method (175). The
anhydride was then reacted with methyl iodide in acetonitrile under anhydrous
conditions (148). A mixture of nicotinic anhydride 3.5 g (0.015 mol.) and methyl
iodide 5.44 g (0.038 mol.) was refluxed in acetonitrile, overnight. The orange
crystals formed were filtered, washed with dry ether, and dried in a desiccator
under vacuo. Yield 85 %; mp. 213-215 0C. 1HNMR (90 MHz, d6-DMSO) 5
pyridine protons 8.03-9.36 (8H, multiplet), methyl protons 4.45 (6H, singlet). IR
(KBr) -CH3 stretch (broad, 2800-3200 cm-l), anhydride stretch (1750, 1810 cm-i).
Analysis calculated for C14H14N2O3I2: % C, 32.81; H, 2.73; N, 5.47; I, 49.61;
Found C, 32.66; H, 2.59; N, 5.60; 1,49.66.
Synthesis of l-methyl-3-i(9'-guanylmethoxyethoxy)carbonynpyridinium iodide
(AO±) 121
This reaction was carried out according to the previously published
literature method (166). To a solution of 0.5 g (2.2 mmol.) of acyclovir in 20 ml.
dimethylformamide, was added 1.14 g (2.2 mmol.) of [1] and a catalytic amount
0.07 g (0.6 mmol.) of 4-dimethylaminopyridine. The reaction mixture was stirred
at room temperature for 4 days under nitrogen until all of the acyclovir was
consumed. As the reaction proceeded, which was monitored by TLC, the orange

57
color of the anhydride [1] was replaced by a yellow color. The reaction mixture
was filtered and the filtrate evaporated under vacuo. The residue was washed with
dry acetone and dry ether to remove 4-dimethylaminopyridine, unreacted
acyclovir, anhydride, and other impurities, and dried in a desiccator under vacuo.
Yield 65 %; mp. 201-2020C. 1HNMR (90 MHz; d6-DMSO) 5 pyridine protons
8.12-9.48 (4H, multiplet); purine Cs proton 7.93 (1H, singlet); purine C3-NH2 and
Ci-OH or N2-NH tautomeric protons 6.51 (3H, broad singlet); -CH20-5.45 (2H,
singlet); N+-CH3 4.51 (3H,
singlet); -CH202C- 3.84-4.06 (2H, broad triplet); -OCH2- 3.42-3.51 (2H, broad
triplet). UV characteristics were determined in methanol at 254 nm.
Synthesis of l-methvl-3-rf9t-guanylmethoxyethoxy)carbonyll-1.4-dihydropyrid-
ine (A-CDS) T31
To a solution of 1.58 g (3.3 mmol.) of [2] in 120 ml of degassed water was
added 1.69 g (20.1 mmol.) of sodium bicarbonate, all at once. The mixture was
stirred at 00C and 2.33 g (13.38 mmol.) of sodium dithionite was added over a 5
minute period. The reaction was maintained under nitrogen all the time. The
product [3] is insoluble in water and formed light yellow-colored crystals over the
aqueous layer. The crystals were collected, washed with ice-cold water, dry ether,
and dried in a desiccator under vacuo. Yield 54 %; mp. 163-1650C. *HNMR (90
MHz; dó-DMSO) 5 purine Cs proton 7.81 (1H, singlet); pyridine C2 proton 6.81
(1H, singlet); purine C3-NH2 and Ci-OH or N2-NH tautomeric protons 6.48 (3H,

58
broad singlet); pyridine proton 5.60 (1H, doublet); -CH2O 5.36 (2H, singlet);
pyridine C5 proton 4.48-4.81(lH, multiplet); -CH2O2C 3.84-4.24 (2H, multiplet); -
OCH2 and pyridine C4 protons 3.21-3.81 (4H, multiplet); N-CH3 2.81 (3H,
singlet). UV characteristics were determined in methanol at 254 and 370 nm.
Analysis calculated for C15H22N6O6: % C, 47.11; H, 5.79; N, 21.98; Found C,
46.97; H, 5.59; N, 22.27.
Figure 3-1 shows the general reaction scheme for the synthesis of A-CDS.
Synthesis of lipoic-acid ester of acyclovir (A-LipS2l T41
Dimethylformamide (20 ml) was heated to 600C and 0.225 g (1.0 mmol.) of
acyclovir was dissolved in it. The solution was cooled to room temperature and
0.206 g (1.0 mmol.) of dl-thioctic acid, 0.206 g (1.0 mmol.) of
dicyclohexylcarbodiimide and 0.122 g (1.0 mmol.) of 4-dimethylaminopyridine
were added to it. The reaction was allowed to run at room temperature under
nitrogen for 4 days and was monitored by TLC. The reaction mixture was filtered
and the filtrate was evaporated under vacuo to remove DMF. The solid residue
was washed with dry ether and then with dry benzene and the solvents were
evaporated under vacuo to remove all residual DMF. The residue was finally
washed with cold methanol to remove unreacted acyclovir, dl-thioctic acid, and
other impurities. Yield 50 %; M.S. (FAB; NBA) = 414 (M+l).
Figure 3-2 shows the general reaction scheme for the synthesis of A-LipS2.

o
II
C~ OH
59
2
COCl2/Benzene
2 N(C2H5)3, N2
Figure 3-1 : Reaction scheme for the synthesis of A-CDS.

60
Acyclovir
DMF DCC
RT, N2 DMAP
4 days
Figure 3-2 : Reaction scheme for the synthesis of A-LipS2.
Synthesis of ethvl ester of 4-isocvanobutvric acid Í51
Ethyl 4-aminobutyrate hydrochloride (40.0 g; 238.6 mmol) was suspended
in 250 ml of toluene. Phosgene gas, dissolved in toluene (Fluka; 20 % solution;

61
124.2 ml; 238.6 mmol) was added to it at room temperature with stirring and the
reaction mixture was stirred at 80°C for 1 hour under nitrogen. Phosgene solution
(60 ml) was added and the reaction mixture was again stirred at 800C for 30
minutes. Additional 60 ml of phosgene solution was added and the reaction
mixture was further stirred at 800C for 30 minutes. Boiling was continued for 1
more hour and at all times nitrogen was bubbled through the reaction mixture.
The solution was cooled and the solid yellow residue was filtered. The solvent was
evaporated under vacuo to give a yellow residue which was distilled to give [5].
Yield 15 g (40 %); bp = 810C/6 mm Hg. 1HNMR (90 MHz; CDC13) 8 4.1 (2H,
quartet, -C02-CH2), 3.3 (2H, triplet, =N-CH2), 2.35 (2H, triplet, -CH2-C02-), 1.9
(2H, multiples -C-CH2-C), 1.25 (3H, triplet, -CH3).
Synthesis of l-(3-EthoxycarbonvlpropvlcarbamovlV5-Fluorouracil T61
A mixture of 5-Fluorouracil (5 g; 38.4 mmol), [51 (6 g; 38.4 mmol) in 20 ml
pyridine was stirred for 3 hours at 90°C under nitrogen. The reaction mixture was
then cooled, pyridine was evaporated under vacuo, and the residue was dissolved in
150 ml of dichloromethane. The organic layer was washed with 2N HC1 (100 ml),
then water, and dried over MgSC>4. Evaporation of dichloromethane and washing
with ether gave a white solid compound. Yield 7.7 g (69 %); mp. 135-1360C.
iHNMR (DMSO-d6) 5 9.1 (1H, broad singlet, -CO-NH-C-), 8.3 (1H, doublet, C6-
H), 4.05 (2H, quartet, -C02-CH2), 3.3 (2H, multiplet, -N-CH2), 2.31 (2H, triplet, -

62
CH2-CO2-), 1.85 (-C-CH2-C-), 1.2 (3H, triplet, -CH3). Analysis calculated for
C11H14FN3O5: % C, 45.99; H, 4.91; N, 14.63; Found C, 45.93; H, 4.91; N, 14.58.
Synthesis of l-(3-Carboxvpropylcarbamovl)-5-Fluorouracil 171
Concentrated HC1 (40 ml) and 5.75 g (20 mmol) of [6] were stirred at 80°C
for 1 hour. Cooling at 100C gave crystals, which were filtered, washed with cold
water, and dried. Yield 3.65 g (71 %); mp. 157-1590C. 1HNMR (DMSO-dó): 3
1.76 (2H, quintet, -C-CH2-C-), 2.27 (2H, triplet, -CH2-CO2-), 3.32 (2H, quartet, -
N-CH2-), 8.39 (1H, doublet, 6-H), 9.16 (1H, triplet, -CO-NH-), 11.40 (2H, broad,
3-H and -CO2H).
Synthesis of DL-q-Lipol 181
DL-a-Lipoic acid (2.06 g; 10 mmol) was placed in a flame-dried 250 ml,
round-bottom flask, fitted with a stirring bar and dropping funnel and closed with
a septum. Chloroform (70 ml) was added and catecholborane in THF (50 ml; 50
mmol) was then added dropwise. The mixture was refluxed for 5 hours until the
reduction was complete, as monitored by TLC. Nitrogen atmosphere was
maintained during the entire reaction with an oil-bubbler. Twenty ml of cold
water was added dropwise and the organic solvents were removed in vacuo.
Dichloromethane (50 ml) was then added and the solution was extracted with one
25 ml portion of water followed by six 25 ml extractions using 1 N NaOH solution
to remove catechol. The dichloromethane solution was washed with water (2 x 25

63
ml), dried with magnesium sulphate, filtered, and evaporated. Yield 1.8 g (94 %).
iHNMR (CDCI3): 3.6 (3H, multiplet, ring -CH=, and -CH2-O-), 3.15 (2H, triplet,
ring -CH2-SS-), 2.4 (2H, multiplet, ring -CH2-C-SS-), 1.5 (8H, broad multiplet,
alkyl -(CH2)4-).
Synthesis of Lipolyl ester of l-(3-Carboxvpropylcarbamovl)-5-Fluorouracil Í91
To 60 ml of dichloromethane (containing 10 ml of dimethylformamide) at
50C with stirring was added j7] (2.6 g; 10 mmol), 18] (1.92 g; 10 mmol), and a
catalytic amount of 4-dimethylaminopyridine. Into this mixture,
dicyclohexylcarbodiimide (6.2 g; 30 mmol) dissolved in 20 ml dichloromethane
was added in a period of 10 minutes and the stirring was continued for 1 day. The
reaction mixture was filtered and the filtrate was poured into 300 ml water. The
dichloromethane layer was separated, washed with 3 x 30 ml of 2 N HC1, 3 x 30 ml
of water, 3 x 30 ml of 5 % sodium bicarbonate solution, and finallly with 3 x 30 ml
of water. The organic layer was separated, dried over magnesium sulphate,
filtered, and evaporated in vacuo. The oily residue was purified using column
chromatography (benzene/ethyl acetate; 3/1 solution and silica gel, E. Merck, No.
7734, Kieselgel 60). The oily product was crystallized from methanol to give a
white powder. Yield 1.1 g (25.4 %); mp. 210-2150C. iHNMR (CDCI3): 5 1.35
(4H, multiplet, -C-(CH2)2-C-), 1.58 (2H, quintet, -C-CH2-C-02C-), 1.63 (2H,
multiplet, -S-C-CH2-), 1.85 (1H, dddd, 4'p-H), 1.88 (2H, quintet, -C-CH2-C-),

64
2.34 (2H, triplet, -C-CH2-C02-), 2.40 (1H, dddd, 4'a-H), 3.00-3.17 (2H, multiple!,
-S-S-CH2-), 3.40 (2H, quartet, -N-CH2-), 3.50 (1H, multiplet, 3’cc-H), 8.40 (1H,
doublet, 6-H), 9.03 (1H, triplet, -CO-NH-), 9.17 (1H, broad, 3-NH). M.S. (FAB;
NBA) = 433 (M+l). Figure 3-3 shows the general reaction scheme for the
synthesis of 5-FU-LipS2.
HPLC Analysis
Cation exchange high pressure liquid chromatography was used to analyze
acyclovir, AQ+, and A-CDS. The column, Partisil SCX 10 (im, 22 cm x 4.6 mm
i.d., and a vydac guard column packed with SCX EE 3855 (Du Pont) was linked to
a Spectra-Physics HPLC system consisting of SP 8810 precision isocratic pump, SP
8450 UV/VIS detector, SP 4290 integrator, and SP 8780 autosampler. The mobile
phase consisted of methanol: 1 mM NH4H2P04 (15 : 85), pH 4.00, at a flow rate of
0.9 ml/min. The retention times for acyclovir, A-CDS, AQ+, and A-LipS2 using
the above conditions were observed to be 6.0, 8.0, 10.5, and 7.8 min. respectively.
They were detected at 254 nm and the injection volume was 20 |il.
Reverse-phase HPLC was used for the detection for 5-FU, CPCFU, and 5-
FU-LipS2. They were detected using ASI C-18, 5 |i.m, 22 cm x 4.6 mm i.d. at 265
nm with the mobile phase consisting of acetonitrile : water (30 : 70) at a flow rate
of 0.8 ml/min. Their retention times were 4.0, 5.2, and 8.4 min. respectively.

65
coa
H2N—(CH^—C02 —C2H5 — OCN—(CH2)3—CO2—C2H5
Toluene
3-4 h, 80°C, N2 [51
5 -FU
3 h, 90°C, N2
Pyridine
HC1
lh, 80°C
O
F
NH-(CH2)3-Cp2H
[7]
HOOC— (CH2 )4
Catecholborane
CHCI3
5 h, Reflux, N2
HO-(CH2)5
[8]
[7] + [8]
DCC, DMAP
DMF
24 h
O
Lipolyl ester of CPCFU
[9]
Figure 3-3 : Reaction scheme for the synthesis of 5-FU-LipS2.

66
Preparation of Skin Membranes
Three different types of skin membranes were used in vitro to assess the
transport of drugs and their chemical targetting systems into and through the skin;
the full-thickness, freshly excised, hairless-mouse skin, the shaved, epilated,
full-thickness, guinea-pig skin, and the heat-separated, human epidermal
membrane.
Female hairless-mice (SKH-HR-1 strain, Temple University) were
sacrificed by cervical dislocation and the full-thickness skin was removed from the
abdomen and back. The fat layer below the dermis which was gently removed
using forceps was discarded and the full-thickness skin was used for the in vitro
diffusion or partition experiments, immediately.
Human skin from cadavers (Thigh or abdominal region; North Regional
Transplant Services, Lansing, MI) previously dermatomed was gently swirled in
distilled water (preheated to 6(PC) for about 2-3 min. The epidermal membrane
(about 100 (i. thick) was then carefully separated from the dermis in a glass-trough
containing distilled water at room temperature. The stratum comeum surface was
then placed flat against a plastic sheet and the dermal side of the epidermal layer
was covered with an absorbent paper saturated with 0.9 % NaCl. This human
epidermal membrane was used immediately or stored at -100C and used within 24
hours for the in vitro diffusion or partition experiments.
Guinea-pigs (Harlan Sprague Dawley; 3-4 weeks) were anesthetized using
pentobarbital (i.p.). The hair was shaved with electric clippers and epilated with

67
the depilatory cream, NairR. They were then sacrificed by heart-puncture using
pentobarbital. The full-thickness skin was removed from the abdomen and back,
which was used for the in vitro diffusion experiments, immediately.
Preparation of Donor Solutions
Excess drug and the chemical targetting systems were sonicated in
propylene glycol for 30 minutes and 0.8-1.0 ml of this saturated solution was
used in the donor compartment of diffusion-cells in experiments involving the
hairless-mouse and guinea-pig skin. Acyclovir, 5-FU, and their delivery systems
were observed to be stable in propylene glycol as the donor vehicle.
For the transport studies using human epidermal membrane, excess drug
was stirred in phosphate buffered saline, pH 7.1, for 24 hours at 320C in an
incubator. The suspension was filtered and 5 ml of the filtrate was used as the
donor solution. All the drugs that were evaluated for their transport
characteristics across the human epidermis were stable in the aqueous donor
vehicle.
Diffusion experiments
Two-compartment, vertical diffusion cells (Kresco Enggineering, Palo
Alto) of 7.1 cm2 surface area were used for hairless-mouse and guinea-pig skin
penetration studies. The donor was about 0.8-1.0 ml of saturated solution of the
drug or the chemical delivery system in propylene glycol and the receiver about 40

68
ml of phosphate buffer, pH 6.5. The entire system was incubated at 320C in the
incubator (Lab-Line Instruments, LL) and the receiver was stirred magnetically at
200 rpm. At appropriate time intervals, an aliquot (0.5 ml) of the receiver solution
was sampled for direct analysis by HPLC and was replaced with equal volume of
fresh buffer. The samples were either analyzed immediately or kept frozen at -
10°C until analysis.
For the diffusion studies involving human epidermal membrane, side-by-
side diffusion cells (Science Glass Co., Miami, FL) of 1.77 cm2 surface area with
donor and receiver cell volume of 5 ml each were used (Figure 1-7). The donor
was a saturated solution of the drug or the delivery system in either propylene
glycol, dipropylene glycol, or phosphate buffered saline, pH 7.1 and the receiver
was saline. The cells were unstirred. At appropriate time intervals, the entire
receiver solution was sampled for HPLC analysis and the receiver was filled with
equal volume of fresh saline.
The appropriate equations (based upon Fick's laws of diffusion) are shown
in Figure 1-8 (16). These equations were used to obtain the permeability
parameters from the in vitro diffusion experiments.
Skin-Content of the Drug
At various time intervals, the circular piece of skin (7.1 cm2) from the
diffusion experiments was sampled and the donor solution on the skin was
completely removed with 50 % methanol. It was then cut into fine pieces with a

69
pair of scissors, homogenized with 1.0 ml acetonitrile, centrifuged at 13,000 rpm
for 5 minutes (Beckman Microfugeâ„¢ 11), and the supernatant was either
appropriately diluted with acetonitrile before or directly injected into the HPLC
system.
Membrane Partition Coefficient Determination
Dilute solution of the drug or the chemical delivery system (about 100
nmol/ml) in an aqueous medium or propylene glycol (5.0 ml) was allowed to
remain in contact with the stratum comeum side of the skin membrane in a two-
compartment diffusion cell assembly at 320C. The dermis was blocked with
aluminium foil and the receiver was kept empty to prevent diffusion through the
membrane. At various time intervals, an aliquot (0.1 ml) of the donor solution was
sampled for HPLC analysis. The donor aqueous solution was directly analyzed
while the donor propylene glycol solution was appropriately diluted with
acetonitrile for HPLC analysis. The partitioning of the compound in the
membrane was estimated from the decrease in concentration in the donor solution
at equilibrium. The concentration of the drug in the membrane was divided by that
remaining in the donor vehicle at equilibrium to yield Km, the membrane partition
coefficient.

70
In Vitro Stability in Aqueous Buffer
Forty milliliters of Phosphate buffer, 0.05 M, pH 7.4, containing the drug
(about 100 nmol/ml) was shaken at 37°C in a water bath (model YB-531; American
Scientific Products). At various time intervals, an aliquot (0.1 ml) of the sample
was withdrawn and injected directly into the HPLC system for analysis.
In Vitro Stability in Biological Media
Rat whole blood, human plasma, and hairless-mouse skin homogenate in
0.05 M phosphate buffered saline, pH 7.4 were used for assessing the relative
abilities of the various chemical targetting systems to release the parent drugs in the
presence of hydrolytic enzymes in biological media under near physiological
conditions.
Male, white rats (Sprague-Dawley, IN) were humanely sacrificed by
decapitation and their whole blood collected into heparinized tubes. A stock
solution of the drug (0.1 ml of 10 jimol/ml) in acetonitrile was added to 4.9 ml of
the rat whole blood. Five milliliters of the whole blood containing the drug was
carefully shaken at 37 0C. At various time intervals, an aliquot (0.1 ml) sample was
pipetted into centrifuge tubes containing 0.9 ml cold acetonitrile, centrifuged at the
maximum speed of 13 for 5 minutes, and the supernatant was directly analyzed by
HPLC.
Freshly excised, full-thickness, hairless mouse skin was cut into fine pieces,
weighed, and homogenized with 0.05 M phosphate buffered saline, pH 7.4 to give a

71
50% homogenate. A stock solution of the drug (0.1 ml of 10 p. mol/ml) in
acetonitrile was added to 4.9 ml of the 50% skin-homogenate and the final 5 ml of
the skin-homogenate containing the drug was shaken at 32 0C. At various time
intervals, an aliquot (0.1 ml) was sampled and pippetted into centrifuge tubes
containing 0.9 ml cold acetonitrile, centrifuged at the maximum speed of 13 for 5
minutes, and the supernatant was injected into the HPLC system.
Fresh human blood was collected into heparinized centrifuge tubes and
centrifuged at the speed of 50 for 5 minutes. The fresh plasma was separated and
made up with 0.05 M phosphate buffered saline, pH 7.4 to give 85 % plasma. A
stock solution of the drug (0.1 ml of 10 p. mol/ml) in acetonitrile was added to 4.9
ml of 85% plasma. The final 5 ml of human plasma containing the drug was shaken
at 370C. At various time intervals, an aliquot (0.1 ml) was sampled and pippetted
into centrifuge tubes containing 0.9 ml cold acetonitrile, centrifuged at the
maximum speed of 13,000 rpm for 5 minutes, and the supernatant was injected
directly into the HPLC system.
I AM Column Chromatography
The 12 p. IAM.PC or the IAM.PC.MG capped column (containing methyl-
glycolate capped LAM-phophatidylcholine stationary phase), 15 cm x 4.6 mm i.d.,
was connected to the Waters HPLC system consisting of 510 pump, 712 Wisp
automatic injector, Lambda-Max Model 480 LC Spectrophotometer, and a 730
Data Module. A dilute solution of the drug (50-100 nmol/ml) in the mobile phase

72
was injected into the column. The relative retention, also called the capacity factor,
K', of the test compound was calculated using the equation, K' = (tr = to)/to, where tr
is the retention time in minutes of the test compound and to that of the unretained
compound, citric acid. The test soutes were detected by UV at A. =210, 220, or 254
nm except the alcohols which were detected by refractive index detection using
Refractomonitor III operating at 0.5 mA. The mobile phase compositions used
were 0-30 % acetonitrile in Dubelco's Phosphate Buffered Saline 10 x dilution
(DPBS 10 x), pH 7.1, at a flow rate of 1.0 or 2.0 ml/min. The column was washed
with acetonitrile : water (20 : 80) at 1.0 ml/min for 1 hour, before and after
analysis, and was preserved (stored) by washing it with acetonitrile at 1.0 ml/min
for 1 hour.
C-18 Column Chromatography
The Waters pBondapak 10 p, 15 cm x 4.6 mm i.d. column was used with
similar HPLC conditions described in the section on IAM Column
Chromatography, above. Generally, mobile phase compositions containing higher
percentage of organic modifyer were required to elute various compounds from
the C-18 stationary phase. Thus, the alcohols were eluted using acetonitrile : DPBS
10 x (30 : 70) and the steroids were eluted with acetonitrile : DPBS 10 x (50 : 50) as
the mobile phase.

CHAPTER IV
RESULTS AND DISCUSSION
Syntheses
All the compounds were successfully synthesized based upon the results
obtained from Nuclear Magnetic Resonance Spectra, Mass Spectrometry, and
Elemental Analyses.
Generally, the coupling of acyclovir with the acylating agent was the most
difficult and a slow step compared to all the other synthetic steps. The reason for
this difficult and slow rate of acylation of acyclovir is its low solubility in most
inorganic and organic solvents used in synthetic reactions. In addition, the
relatively low reactivity of the acyclic alcohol functional group of acyclovir makes
it even more less amenable for derivatization.
Therefore, it was observed that for esterification of acyclovir, activation of
the acid, namely to its anhydride was required not only for increasing the rate of
the reaction towards completion, but also increasing the yield of the product.
Thus, the trigonelline anhydride was first made and then coupled to acyclovir to
give AQ+. The acylation of acyclovir could have been easily achieved using
nicotinic anhydride. However, in the subsequent step, the quatemization of the
tertiary nitrogen in the pyridine moiety of the nicotinic acid ester of acyclovir
73

74
could have also resulted into methylation of the N-2 and N-7 reactive centers on the
guanine moiety. Therefore, to circumvent this problem of formation of undesired
products, the trigonelline anhydride was specifically synthesized in our laboratory
and was reported in the literature (148) for the first time.
Although, this reaction could be carried out in pyridine as the reaction
solvent, dimethylformamide was observed to be a better solvent. However, both
solvents have their inherent disadvantages. Dimethylformamide is extremely
difficult to evaporate due to its high boiling point (1530Q, and requires exhaustive
washing steps with ether and benzene to remove it from the final product.
Pyridine, on the other hand, can be evaporated again with excessive washing steps
using ether or benzene. However, the low solubility of acyclovir and other
reactants in pyridine was the primary limiting factor in reducing the efficiency of
the synthetic reactions using this solvent. Thus, it was observed that perhaps
dimethylformamide due to its greater solubilizing ability behaved as a better
solvent for carrying out the reactions.
Acylation of acyclovir with a simple unhindered, long-chain acid, such as
lipoic acid was again observed to be a slow, low-yield process. This reaction was
carried out using dimethylformamide as the reaction solvent and equimolar
amounts of acyclovir, lipoic acid, dicyclohexylcarbodiimide, and 4-
dimethylaminopyridine. The reaction was observed to proceed a little faster upon
warming the mixture to 40°C.

75
Skin Experiments
The in vitro experiments were conducted using freshly prepared skin
membranes in the two-compartment diffusion-cell assembly. This particular
laboratory set-up allows for simultaneous monitoring of dermal (local) or
transdermal delivery of various solutes when administered to the skin from the
donor side.
The full-thickness, hairless-mouse skin was chosen as a model membrane to
study and compare the ability of acyclovir and its chemical delivery systems to
localize delivery of acyclovir to the skin. When freshly excised and used
immediately, this tissue provides a rapid and simple in vitro method to evaluate and
demonstrate transport and metabolism (hydrolytic enzymes maintain activity) (177-
179) of various solutes in the layers of the skin. This tissue when used in
combination with the two-compartmental diffusion-cell assembly, serves as an
appropriate, standardized, reliable, and reproducible method by which concurrent
transport and metabolism of different solutes or drugs could be easily monitored.
The method is easy to perform and the availability of the hairless-mouse is plenty
and relatively inexpensive. The shaved, guinea-pig skin and heat-separated human
epidermal membranes were similarly used.
Selection of Delivery Vehicle
Propylene glycol was chosen as the appropriate vehicle for the delivery of
acyclovir and its chemical delivery systems. Saturated solution of the drug in

76
propylene glycol was used in the donor compartment. Therefore, the
thermodynamic activity of the drug was maximized and was nearly equal to the
saturation solubility of the drug in propylene glycol so that maximum driving
force of the drug in the membrane could be achieved. By using saturated solutions
the vehicle effect in affecting the partitioning of the compounds in the membrane is
eliminated and therefore helps in direct comparison of the pure ability of the
molecular species to interact with the skin membrane.
Propylene glycol was chosen as the donor vehicle since it is widely used in
formulating topical drug products. A-CDS is unstable in aqueous solutions and
undergoes a water catalyzed hydroxyl group addition on the C(, atom of the
dihydro trigonelline moiety (180). Therefore, it could not be administered to the
skin in an aqueous (buffered) donor vehicle.
Thus, the compounds demonstrated optimal combination of their membrane
partitioning effect, Km, and their concentration Cd(lirnol/ml), in the donor phase.
This is necessary because according to Fick's laws of diffusion, the flux of the drug
across the membrane is directly proportional to its membrane partition coefficient
and its concentration in the donor phase facing the membrane (Figure 1-8) (16).
This balance between the product of Km and Cd is important since they compromise
the actual magnitudes of each other. Thus, if Cd increases, indicating that the
vehicle has a greater affinity for the drug than the drug has for the skin membrane
then Km tends to decrease and vice versa.

77
Therefore, to monitor the appearance of the drug in the receiver on the
other side of the skin, that is, to observe sufficient, quantifiable amounts of the
drug after diffusion through the skin-membrane, these interactions between the
drug, vehicle, and the skin are important and have to be adjusted appropriately.
Skin Penetration and Retention of Acyclovir
The results obtained upon administering a saturated solution of acyclovir in
propylene glycol to the freshly excised full-thickness hairless-mouse skin, using the
two-compartment diffusion cells in vitro, are shown in Table 4-1 and Figure 4-1.
The drug was analyzed by HPLC at various time intervals in the receiver according
to the description in the methods section.
Table 4-1 : Cumulative amounts of acyclovir in the receiver after application of
saturated solution of acyclovir (33.5 p.mol/ml) in propylene glycol to the hairless-
mouse skin in vitro at 320C.
Time (hours)
Acyclovir in receiver
(nmol/cm 2)a
2
0.2+0.1
4
0.4 + 0.2
5
0.4 + 0.2
6
0.5 + 0.1
8
0.8+0.2
10
0.9+0.3
12
1.5 +0.4
a Average of three determinations ± s.d.

78
Small quantities of acyclovir could be detected in the receiver compartment
over an extended period of time. This indicated limited penetration of acyclovir
through the skin and was expected due to its unfavourable physical-chemical and
membrane permeation properties (133). Such reduced ability to partition into the
stratum comeum has indeed been reported (131,132) and one of the primary
objectives of this dissertation was to improve upon it.
Hence, low values of flux (8.82 x 10-5 |imol/cm2/h) and permeability
coefficient (2.44 x 10-6 cm/h) for the diffusion of acyclovir through the hairless-
mouse skin were observed.
Time (hours)
Figure 4-1 : Diffusion of acyclovir across freshly excised hairless-mouse skin after
application of its saturated solution (33.5 (J. mol/ml) in propylene glycol in vitro at
320C.

79
During the diffusion experiments, when the skin and the receiver phase
were analyzed at various time intervals for the drug content, acyclovir showed
greater tendency to penetrate the skin completely. Therefore, higher levels of
acyclovir could be detected in the receiver as compared to those in the skin tissue
(Table 4-1,4-3, and 4-6).
Table 4-2 : Amounts of acyclovir in the skin after application of its saturated
solution (33.5 |imol/ml) in propylene glycol to the hairless-mouse skin in vitro at
350C.
Time (hours)
Acyclovir in skin
(nmol/cm2)a
3
0.74+0.16
6
1.41 + 0.37
9
2.84 + 0.37
12
3.88 + 0.41
a Average of three determinations ± s.d.
This greater propensity of the antiviral agent to go through the skin may be
responsible for its lack of efficacy in treating the cutaneous HSV I infection as
reported by many workers (127,131,132). Analysis of drug-content in the
receiver as well as in the skin tissue at the same time-point is critical because it
helps in better understanding the fate of the drug under investigation (181).

80
Table 4-2 shows the amounts of the antiviral agent in the skin when it was
applied to the skin in the in vitro experiment at 35 0C. As compared to the values at
320C (Table 4-1), the amounts are higher at 35°C (Table 4-2), as expected.
Chemical Delivery System for Acyclovir Based on Oxidation
Table 4-3 shows the amounts of acyclovir or AQ+ (in nmoles per unit
diffusional surface area) in the skin and receiver that were found after application
of A-CDS as a saturated solution in propylene glycol to the hairless-mouse skin in
vig-Q-
The reduced, relatively more lipophilic A-CDS demonstrated rapid and
facile skin-partitioning followed by fast oxidative metabolic step probably due to
the presence of NAD+-NADH redox coenzyme systems to form increasingly high
levels of the oxidized metabolite, AQ+, in the skin. The levels of AQ+ in the skin
increased about 14 fold from 6 to 48 hours. However, compared to the oxidative
metabolic reaction, the ester hydrolysis of AQ+ to release acyclovir in the skin
mediated by the action of the non-specific esterases was observed to be relatively
much slower. As a result of this rate difference in the kinetic processes, there was
“locking" of AQ+ in the skin from which a slow release of acyclovir was observed.
Thus, the ratio of AQ+/A in the skin was found to increase from 7.5 to 50 to 73 at 6,
24, and 48 hours respectively, indicating faster oxidation relative to ester
hydrolysis in the hairless-mouse skin.

81
Hence, A-CDS managed to improve the delivery of acyclovir to the skin per
unit dose by 9 fold (p < 0.025) at 6 hours, compared to underivatized acyclovir.
This improvement was 4 and 3 fold (p < 0.025) over acyclovir at 24 and 48 hours,
respectively (Figure 4-2, 4-5). The improvement factor is a pure measure of the
superior ability of A-CDS to deliver acyclovir to the skin compared to
underivatized acyclovir under similar experimental conditions and normalizes the
effects due to the dose of the compound applied to the skin in the same delivery
vehicle.
The A-CDS, according to Table 4-3, also delivered acyclovir into the
receiver compartment, indicating that an increase in systemic delivery may occur
in vivo. This could be due to formation of a large fraction of AQ+ and acyclovir in
Table 4-3 : Amounts of acyclovir and AQ+ in the skin and receiver (cumulative)
after application of saturated solution of acyclovir (33.5 |i mol/ml) or A-CDS (45.3
pmol/ml) in propylene glycol to the hairless-mouse skin in vitro at 320C.
Compound
Dose Time
Acyclovir (nmol/cm2)a
AQ+ (nmol/cm 2) a
(nmol/cm 2)
(h)
Skin
Receiver
Skin
Receiver
Acyclovir
3896
6
0.4 + 0.2
0.4 + 0.2
24
1.2+0.5
3.8 + 1.5
48
1.7 ±0.5
7.3 ± 1.4
A-CDS
4989
6
4.7 + 2.0
5.4 +0.2
35+14
6+1
24
5.8 + 1.8
6.3 +0.5
289 +17
37 +4
48
6.7 + 1.6
7.8 +0.3
488 +35
142 + 23
a Average of three determinations ± s.d.

82
20
x
.h
>
_o
13
C3
<4-1
O
u
>
s
using acyclovir
using A-CDS
10 -
13
3 \
T
I
0 -
J
*K*»*®X
â– 3 3
Í:¡:Í>:Í:ÍS:
24
Time (hours)
Figure 4-2 : Dermal delivery of acyclovir per unit dose using acyclovir or A-CDS
as a function of time.
the extracellular compartment of the skin because release of AQ+ and acyclovir in
the receiver was observed. If the AQ+ and acyclovir were specifically
concentrated into the skin cells then they would not have been detected in such high
amounts in the receiver. Their more polar characteristics would have prevented
their efflux out of the intracellular compartment.
When acyclovir was applied to the skin, the ratio of its amounts in the skin to
receiver decreased from 1 to 0.3 to 0.2 at 6, 24, and 48 hours respectively, as seen
in Table 4-3. This greater tendency for acyclovir to go through the skin may

83
indicate its presence predominantly in the extracellular compartment when
underivatized acyclovir is administered to the skin. Its lower lipid solubility may
not allow for sufficient penetration of the cell-membranes of the epidermal cells.
The A-CDS, on the other hand, altered this kinetic profile most probably
due to intracellular localization of AQ+ and hence acyclovir. Due to A-CDS, the
ratio (skin/receiver) of acyclovir was found to remain constant and was 0.87, 0.92,
and 0.86 at 6, 24, and 48 hours, respectively. Thus, a steady-state in the
distribution of acyclovir between the skin and receiver had reached after the
application of A-CDS to the skin. This steady-state must have been a direct result
of a “locked" fraction of acyclovir in the intracellular compartment from which
facile release of acyclovir in the receiver would not have occurred. The
corresponding ratios (skin/receiver) of AQ+ as a result of A-CDS application to the
skin were 5.8, 7.8, and 3.4 at 6, 24, and 48 hours, respectively. Thus, the AQ+ was
more preferentially “locked" in the skin than acyclovir. Therefore, the overall
effect of applying A-CDS to the skin was the establishment of a “reservoir" of the
metabolic precursor (AQ+) of acyclovir with a consequent improvement in the
delivery of acyclovir to the skin.
The results seem to indicate that the concentration of AQ+ in the skin is an
important determinant of the availability of acyclovir in the skin. It is worthwhile
to note that as the ratio (skin/receiver) of AQ+ decreases with time (from 5.8 to 7.8
to 3.4) (Table 4-3), so does the improvement in acyclovir delivery to the skin due
to A-CDS over underivatized acyclovir (from 9 to 4 to 3 fold) (Figure 4-5).

84
Therefore, as more AQ+ comes out of the skin, there is less of acyclovir formed
from AQ+ in the skin. Hence, AQ+ can be considered a depot-form of acyclovir in
the skin.
However, the net decrease in improvement of acyclovir delivery to the skin
with time due to A-CDS application (Figure 4-5) may have been a result of a
combination of the following processes; extracellular formation of AQ+ and hence
acyclovir resulting into their release from the skin and into the receiver, decrease
in the rate of enzymatic oxidation or hydrolysis of AQ+ or both to
release acyclovir in the skin, increase in the rate of intracellular phophorylation of
acyclovir mediated by the phosphokinases.
Table 4-4 shows the amounts of AQ + and acyclovir formed from A-CDS in
the skin at 48 hours when the saturated donor solution of A-CDS in propylene
glycol was washed from the skin surface at 6 hours during the diffusion
experiment. This washing was carried out by first pipetting and discarding the
donor solution. The residual donor was further removed by similarly washing
(twice) with a solution of 50% methanol in water. The circular piece of skin from
each diffusion-cell was sampled, cut into fine pieces, and the tissue was extracted
with 1 ml acetonitrile as described in the experimental section and analyzed by
cation exchange HPLC for the presence of acyclovir and AQ+. From the levels that
were determined in the skin, it is evident that oxidation of A-CDS to AQ+ in the
skin progresses even after complete removal of A-CDS from the surface of the
skin.

85
Table 4-4 : Amounts of acyclovir and AQ+ in the skin (at 48 hours) and receiver
(cumulative) after application of saturated solution of A-CDS (45.3 pmol/ml) in
propylene glycol to the hairless-mouse skin in vitro at 32 OC for 6 hours. Donor
was removed at 6 hours and skin was extracted at 48 hours.
Time
Acyclovir (nmol/cm2)a
AQ+ (nmol/cm 2)a
(AQ +/Acyclo vir)a
(h)
Skin
Receiver
Skin
Receiver
Skin Receiver
6
6.8 +0.5
5.3 + 0.5
0.8 + 0.01
12
5.3 +0.01
9.0+ 1.0
1.7 ±0.2
24
ND
6.6 + 0.5
30
ND
6.8 + 0.0
36
ND
7.9+ 1.0
48
3.3 ±0.3
ND
44.4 ± 4.2
8.0+ 1.0
13.4 ±1.3
a Average of three determinations ± s.d.
ND = Not detectable.
Table 4-5 : Amounts of acyclovir and AQ+ in the skin (at 48 hours) and receiver
(cumulative) after application of saturated solution of A-CDS (45.3 p.mol/ml) in
propylene glycol to the hairless-mouse skin in vitro at 320C for 48 hours.
Time
Acyclovir (nmol/cm2)a
AQ+ (nmol/cm2)a
(AQ+/Acyclovir)a
(h)
Skin
Receiver
Skin
Receiver
Skin
Receiver
6
5.8 + 0.2
6 ± 1
1.0+0.2
12
5.3 +0.1
11 +2
2.1 + 0.1
24
5.2 + 0.2
37 +4
7.2 + 0.4
30
6.3 + 0.4
60 + 12
9.5 +0.3
36
7.5 + 0.3
85 + 17
11.5 +0.2
48
6.7 ± 1.7
7.8 +0.3
488 ± 35
142 + 22
73.2 ± 3.6
18.2+0.3
a Average of three determinations ± s.d.

86
It clearly shows that exposing the skin for 6 hours to A-CDS leads to the
formation of a fraction of AQ+ and hence, acyclovir in the skin that is localized and
which cannot be easily released in the receiver. This “locked" fraction represented
by 3.3 nmol/cm2 of acyclovir and 44.4 nmol/cm2 of AQ+ in the skin determined at
48 hours (Table 4-4) may be the pure intracellular concentration of acyclovir and
AQ+, respectively, when the A-CDS was applied for 6 hours to the skin (compare
with Table 4-5).
Chemical Delivery System for Acyclovir Based on Reduction
The ability of acyclovir or A-LipS2 to deliver the antiviral agent IQ and
through the freshly excised full-thickness hairless-mouse skin using in vitro two-
compartment diffusion cells was studied and the results obtained are shown in
Table 4-6.
The saturation solubility of the lipoic-acid moiety containing ester for
acyclovir (A-LipS2) in propylene glycol was about 6 times less than that for
acyclovir (Table 4-6). The A-LipS2 demonstrated a greater thermodynamic
driving force into the skin compared to acyclovir in propylene glycol. Therefore,
when applied to the hairless-mouse skin, A-LipS2 achieved high concentrations in
the skin showing a steady increase from 1.5 nmol/cm2 at 2 hours to 3.2 nmol/cm2 at
12 hours. However, the amount of A-LipS2 in the receiver did not show a
significant increase from 1.8 nmol/cm2 at 2 hours to 2.1 nmol/cm2 at 12 hours

87
(Table 4-6). This indicated the preferential ability of A-LipS 2 to be localized in the
skin rather than its diffusivity through the skin.
The most important observation was the ability of this chemical delivery
system to specifically localize high concentration of acyclovir in the skin. No
detectable levels were observed in the receiver compartment (Table 4-6),
indicating the presence of acyclovir in a “locked" compartment in the skin, most
Table 4-6 : Amounts of acyclovir and A-LipS2 in the skin and receiver
(cumulative) after application of saturated solution of acyclovir (35.6 p. mol/ml) or
A-LipS2 (6.0 p.mol/ml) in propylene glycol to the hairless-mouse skin in vitro at
320C.
Compound Dose
(mol/cm2)
Time Acyclovir (nmol/cm2)a A-LipS2 (nmol/cm2)3
(h) Skin Receiver Skin Receiver
Acyclovir
4069
2
0.2+ 0.1
0.3 +0.1
4
0.2+ 0.1
0.4 +0.2
6
0.4+ 0.2
0.4 +0.2
8
1.0+ 0.3
0.8 + 0.2
10
1.3+ 0.2
1.7 +0.4
12
1.1 ±0.4
1.9 ±0.3
A-LipS2
680
2
3.2 ±0.4
NDb
1.5 ±0.2
1.8 ±0.5
4
3.5 + 0.1
ND
2.7+ 0.1
1.4+0.5
6
2.5 + 0.6
ND
2.6+ 0.1
2.2 + 0.3
8
3.1+0.3
ND
1.5+0.1
2.2+ 0.1
10
2.0+ 0.1
ND
3.3+0.0
1.9+ 0.2
12
1.9+ 0.1
ND
3.2+ 0.1
2.1 + 0.2
a Average of three determinations ± s.d.
b Below detection limit, hence could not be detected.

88
Figure 4-3 : Dermal delivery of acyclovir per unit dose using acyclovir or A-LipS2
as a function of time.
presumably in the intracellular phase. Thus, using A-LipS2, more extensive
localization of acyclovir (per unit area per unit dose) occurred in the skin (Figure 4-
6), than using the previously described A-CDS (Table 4-3).
The ratio (A-LipS 2/acyclovir) in the skin increased with time when ALipS2
was administered to the hairless-mouse skin. It was determined to be 0.5,0.8, 1.0,
0.5, 1.6, and 1.7 at 2, 4, 6, 8, 10, and 12 hours respectively. This indicated a higher
rate of accumulation for A-LipS2 relative to the formation of acyclovir in the skin.

89
Hence, A-LipS2 demonstrated capability to act as a “reservoir" for the release of
the antiviral agent in the skin. Hydrolysis of A-LipS2 to release acyclovir in the
skin due to the action of the non-specific esterases or lipases in the skin tissue was
observed to occur rapidly atleast at the initial time period.
According to Table 4-6, at 2 hours the improvement in the delivery of
acyclovir specifically to the skin was found to be 96 fold (p < 0.001). At 4, 6, 8,
10, and 12 hours, this improvement was observed to be 105, 37, 18, 9, and 10 fold
(p < 0.001), respectively (Figure 4-5). The levels of the antiviral agent acyclovir,
in the skin due to the application of A-LipS2 were greater during early time points
at 2, 4, 6, and 8 hours, and were observed to decrease at later periods (10 and 12
hours). This characteristic decrease in the improvement of acyclovir delivery to
the skin, as a function of time could be the result of intracellular phosphorylation
of high amounts of accumulated acyclovir hydrolyzed from A-LipS2 in the
epidermal cells of the skin.
It is assumed in this case that phosphorylation of the accumulated acyclovir
occurs in the excised hairless-mouse skin (the phosphorylation of this nucleoside
has been shown to occur in non-infected host cells) (116,119) and hence, the levels
of the antiviral agent were observed to decrease from 3.2 (at 2 hours) to 1.9
nmol/cm2 (at 12 hours), respectively.
The fact that acyclovir could not be detected in the receiver phase may be
explained either by (a) its localization in a compartment in the skin tissue from
which facile elimination does not occur or (b) during the analytical work-up

90
(especially during the homogenization step), the A-LipS2 present in the skin is
hydrolyzed to acyclovir, implying that only A-LipS2 was being accumulated in the
skin due to the application of A-LipS2 to the hairless-mouse skin, and that the
hydrolytic ability of the cutaneous esterase enzymes was negligible.
If indeed, the intracellular phosphorylation of acyclovir mediated by the
cellular kinases did not occur then the decrease in acyclovir levels in the skin as a
function of time cannot be explained. Therefore, it is likely that the A-LipS2 due to
its extensive solubilizing ability in the skin tissue (see Table 4-9) was localized in
the intracellular compartment and was therefore more susceptible towards
enzymatic attack. The non-specific esterase enzymes were able to release acyclovir
in the intracellular environment, thus making it available for the kinases to
phosphorylate the alcohol. Thus, the levels of the antiviral agent were found to
decrease from 3.2 to 1.9 nmol/cm2 (from 2 to 6 hours) in the skin.
The cells of the epidermal region of the skin have been shown to possess
significant enzymatic activity and this layer is considered to be the enzymatic
barrier in the skin to the transport of various drug molecules with metabolically
sensitive functional groups (177). Furthermore, it has been reported that the
cellular hexokinases within the epidermis (182) are responsible for controlling the
levels of free glucose and its phosphorylated metabolite, the glucose-6-phosphate.
The human epidermal hexokinase acts as a gate that allows free flow of glucose into
the various metabolic channels of the cell (182).

91
Hexokinase systems in the skin and appendages of monkeys (183) were
found to be generally as*active as those reported in the brain (184), suggesting the
presence of an active metabolic process in the cells of the skin. The primate skin
was not capable of catalyzing non-phosphorylating processes for glucose, such as
the conversion of glucose to gluconic acid (by the liver) or that of glucose to
sorbitol (by the fetal liver and seminal vesicle) (183). Hence, phosphorylation of
glucose must be the sole pathway for glucose-utilization in the skin to subsequent
steps in glycolysis, glycogenesis, or mucopolysaccharide synthesis. Therefore,
hexokinase is regarded as one of the important enzymes in this respect. Hence, the
phosphorylation of acyclovir in the non-infected epidermal cells of the hairless-
mouse skin is a likely possibility and the decrease in the levels of the antiviral agent
observed could be explained.
Thus, both the chemical delivery systems, the A-CDS, and the A-LipS2,
managed to improve acyclovir delivery to the hairless-mouse skin, in vitro. The A-
LipS2 (based on reduction) was observed to be significantly greater than the A-
CDS (based on oxidation) in achieving specific delivery of the antiviral agent to the
skin.
Chemical Delivery System for 5-Fluorouracil
The ability of 5-Fluorouracil or its delivery system, the 5-FU-LipS2, to
deliver 5-FU IQ and through the shaved and epilated full-thickness guinea-pig skin
in vitro was studied and the results obtained are shown in Table 4-7.

92
The total amount of 5-FU determined in the skin relative to that diffusing
through the skin was observed to be much greater, using either underivatized 5-FU
or the 5-FU-LipS2 (compare the values in Table 4-2, 4-3, 4-4, 4-5, 4-6, and 4-7).
This could be due to greater volume of the guinea-pig skin barrier compared to
that of the hairless-mouse skin in the previous cases.
The ratio (skin/receiver) of the amounts per unit diffusional skin-surface
area of 5-FU was 21, 11, and 17 at 2, 4, and 6 hours respectively, using
underivatized 5-FU. When 5-FU-LipS2 was administered to the guinea-pig skin,
this ratio was found to be 25, 31, and 15 at 2,4, and 6 hours, respectively.
Table 4-7 : Amounts of 5-FU in the skin and receiver (cumulative)
after application of saturated solution of 5-FU (350 (imol/ml) or
5-FU-LipS2 (84 jimol/ml) in propylene glycol to the shaved guinea-
pig skin in vitro at 32°C.
Compound
Dose
Time
5-FU (|imol/cm2)a
(jimoI/cm2)
(h)
Skin
Receiver
5-FU
50
2
2.1 +0.1
0.10+0.01
4
2.5 +0.1
0.23 + 0.02
6
5.5 ±0.2
0.31 ±0.04
5-FU-LipS2b
12
2
2.0 ±0.1
0.08 ± 0.01
4
2.8 +0.2
0.09 + 0.01
6
3.0 ±0.2
0.20 + 0.01
a Average of three determinations ± s.d.
b The acid metabolite (CPCFU) of 5-FU-LipS2 could not be detected in the
skin or receiver.

93
The improvement of 5-FU delivery to the skin because of administering 5-
FU-LipS2 to the skin showed decreasing relationship with respect to time similar to
that observed with A-CDS and A-LipS2 (Figure 4-5). This improvement was 4,
4.7, and 2.3 fold (p < 0.001) at 2,4, and 6 hours, respectively, and was observed to
decrease with time (Figure 4-5).
o
X
g
I
m
o
S'
>
3
13
a
Q
2 -
jg| using 5-FU
|H using 5-FU-LipS2
X
X
tiff
fill
Hi
â– I
: ;;
Time (hours)
Figure 4-4 : Dermal delivery of 5-FU per unit dose using 5-FU or 5-FU-LipS2 as a
function of time.
The skin and receiver samples were also analyzed for the acid metabolite of
5-FU-LipS2, the l-(3-carboxypropylcarbamoyl)-5-FU (also denoted by CPCFU).

94
However, no CPCFU could be detected either in the skin or in the receiver during
the diffusion experiment. This suggests that probably complete hydrolysis of 5-FU-
LipS2 in the skin occurs to release 5-FU in the skin and hence in the receiver. Thus,
the life-time of CPCFU in the skin may be short and so it could possibly release 5-
FU completely and rapidly in the skin.
Figure 4-5 : Improvement in delivery of acyclovir (to the hairless-mouse skin) or 5-
FU (to the guinea-pig skin) using A-CDS, A-LipS2, or 5-FU-LipS2 over
underivatized acyclovir or 5-FU respectively, as a function of time.
Lipophilicitv and Partition Coefficients
The partition coefficient is determined by the solubility of the drug in the
biological membrane relative to that in the vehicle in which the drug is solubilized

95
and presented to the surface of the membrane. This physical-chemical parameter is
one of the important aspects concerning the design of such chemical targeting
systems which demonstrate improved solubilization in the membrane relative to
that in the vehicle compared to the parent drug molecule.
The partition coefficients for acyclovir and A-CDS (Table 4-8) from their
dilute propylene glycol solutions, in freshly excised, full-thickness hairless-mouse
skin and those for acyclovir and A-LipS2 (Table 4-9) from their dilute water
solutions in heat-separated human epidermal membrane were experimentally
determined at equilibrium using the two-compartment diffusion-cells at 320C.
Table 4-8 also shows the comparison between the relative lipophilicity of acyclovir
and the A-CDS in terms of their relative retention on a C-18, reverse-phase, HPLC
column, at 0 % organic modifyer concentration in the mobile phase.
Table 4-8 : Lipophilicity and skin-membrane partition coefficients of acyclovir
and A-CDS from their dilute solutions in propylene glycol in vitro at 320C.
Compound
Log K' (HPLC)a
Log Khms5
Log Khe0
Acyclovir
0.64 ±0.04
0.82 ±0.15
1.30 ±0.10
A-CDS
2.12 ±0.13
2.03 ±0.26
1.92 ±0.37
All values are averages of three determinations ± s.d.
a Relative lipophilicity based on HPLC elution (relative retention),
b Equilibrium partition coefficient in hairless-mouse skin membrane,
c Equilibrium partition coefficient in human epidermal membrane.

96
A-CDS indicated about 30 fold greater lipophilicity relative to that of
acyclovir using this HPLC method (Table 4-8). The enhanced lipid solubility of A-
CDS must have been responsible for its greater availability in the skin as can be
deduced from the diffusion and skin-retention experiments discussed above, and
therefore its relatively greater solubility in the skin. Thus, the partition coefficient
of A-CDS in the hairless-mouse skin from propylene glycol was determined to be
16 fold greater than the corresponding partition coefficient for acyclovir. In the
human epidermal preparation, A-CDS demonstrated a 4 fold improvement in the
partitioning effect over acyclovir.
Table 4-9 shows a dramatic 8 fold enhancement in the partition coefficient
of the lipoic acid ester of acyclovir over underivatized acyclovir in the human
epidermal membrane. This high partitioning ability of A-LipS2 in the epidermis
must have contributed to its tremendous thermodynamic driving force into the skin
Table 4-9 : Partitioning of acyclovir and A-LipS2 into human epidermal
membrane from their dilute water solutions in vitro at 320C.
Compound
Log Kma
Acyclovir
1.42 ±0.12
A-LipS2
2.32 ±0.11
a Average of three determinations ± s.d.

97
membrane and thus resulting into pronounced improvement of acyclovir delivery
due to high skin concentration of A-LipS2.
Thus, the greater partitioning ability of the chemical delivery systems was a
major contributing factor in helping the antiviral or the anticancer agent to
solubilize in the skin. Whereas, the selective ability of these delivery systems to
undergo redox reactions and to create reservoir forms was the predominant
controlling factor in achieving their high local skin concentrations and thus,
improving the delivery of acyclovir or 5-Fluorouracil specifically to the skin to
achieve their greater intradermal availability, respectively.
In Vitro Stability
An important aspect of pharmaceutical drug-development is the ability to
formulate the drug-specie in appropriate vehicle system(s). This is critically
important because for the ultimate success of a drug product, the stability of the
drug in the formulation should be high for storage purposes. If the principle
ingredient is a derivatized drug as is the case with the described delivery systems,
then an additional factor of importance is the ability of the chemical specie to
release the parent drug, in vivo. In addition, it is also worthwhile to identify and
differentiate between the effect of biological media containing enzyme systems and
a physiological pH effect. Thus, stability studies of the various chemical species in
aqueous buffer at pH 7.4 and in different biological media under near physiological
conditions were carried out.

98
Table 4-10 shows the results obtained upon incubating the drugs in various
media at 370C. The decrease in initial concentration of the drug from the starting
(incubating) solution was monitored by HPLC. The slope of the linear plot (r >
0.995 in all cases) of natural logarithm of the drug concentration remaining versus
time afforded the first-order degradation rate constant, Kobs (in min-i). Thus, the
11/2 values (in min) were calculated from the equation, ti/2 = 0.693/Kobs- The first-
order half lives (Table 4-10) of the various chemical species in pH 7.4 buffer
system were observed to be much greater (than their stabilities in biological media)
and in the range of 2 to 16 hours.
The relative instability (t 1/2 = 2 h) of A-CDS in aqueous buffer solutions
may be due to its susceptibility towards water addition. Water addition across the
5,6-double bond in 1,4-dihydropyridines to form the 6-hydroxy
tetrahydropyridine is an irreversible reaction (185,186). This unwanted instability
of the dihydropyridine moiety could be detrimental to the successful formulation
of such reduced 1,4-dihydropyridines in aqueous vehicles for drug-delivery
purposes. Therefore, the oxidation and water addition processes in various
dihydropyridines were studied theoretically using MNDO/2 (180). Minimizing the
total energy of the structures, the geometries were optimized with respect to all
structural variables. The actual heats of formations of the protonated relative to
the unprotonated forms were observed to predict protonation to occur at C 5. Thus,
the 1,4-dihydropyridines undergo water addition by the nucleophilic attack of
water (or OH-) at C(, and that protonation at C5 is the rate-determining step (176).

99
The A-LipS2 exhibited high stability in the buffer and did not show any sign
of degradation for more than 24 hours. This type of a derivative for acyclovir
would be appropriate and relatively easy to formulate in aqueous vehicle(s).
Whereas, in biological media, it was observed to undergo rapid hydrolysis most
likely by the non-specific esterases to release acyclovir. In 85 % fresh human
plasma its half-life was about 5 minutes, while in the 50 % hairless-mouse skin
homogenate it was about 26 minutes.
Table 4-10 : Stability of A-CDS, AQ+, A-LipS2, CPCFU, and 5-FU-LipS2,
in vitro at 370C.
Compound
Medium
K0bs (min-i)*
11/2 (min)*
A-CDS
pH 7.4 buffer
5.78 ± 0.29 e-3
120± 6
Rat blood
5.33 ± 0.82 e-2
13 ±2
AQ+
pH 7.4 buffer
3.98 ±0.25 e-3
174 ±11
Rat blood
4.32 ± 1.08 e-2
16 ±4
A-LipS2
pH 7.4 buffer
Stable
Days
85 % Human Plasma
1.39 ± 0.28 e-i
5 ± 1
50 % Hairless-mouse
skin homogenate
2.70 ±0.30 e-2
26 ± 3
CPCFU
pH 7.4 buffer
9.96 ± 0.10 e^
696 ± 12
5-FU-LipS2 pH 7.4 buffer
7.22 ± 0.14 e^
960 ±18
* Average of three determinations ± S.E.

100
Stability of the IAM.PC Column Stationary Phase
The IAM.PC column was first subjected to different HPLC conditions.
Various solutes were injected onto the column and their relative retention was
calculated using different chromatographic conditions like changes in flow rate and
mobile phase composition.
The elution profile of the compounds was observed to be highly
reproducible from day-to-day. The IAM.PC column did not show any indication
of deterioration or leakage of its stationary phase component. The operating
pressures at various flow rates using the IAM.PC column were about 300-500 psi
lower than the corresponding pressures using the C-18 column. For example, with
a mobile phase composition of acetonitrile : water (20 : 80) at a flow rate of 1.0
ml/min, the IAM.PC column showed pressure of about 700 psi. Using the same
chromatographic conditions, the C-18 column showed pressure of about 1000 psi.
Apart from this slightly reduced back-pressure in the IAM.PC column, no other
major differences were observed during the IAM.PC column chromatography.
Even over a period of 1 year, the previuosly used IAM.PC columns showed
excellent reproducibility of retention profiles of various solutes. Thus, these
columns exhibited good physical and chemical stability.
The purpose of this part of the investigation was to select appropriate solute
molecules for evaluating their transport across human skin, to assess their
interaction with the phophocholine lipid stationary phase within the IAM.PC
column, and finally to draw appropriate conclusions from correlating the

101
behaviour of the chosen solutes in the IAM.PC column with their ability to be
transported across human skin (187). Finally, the overall goal was to investigate
the possible use of the IAM.PC column as a chromatographic model system to
predict drug-transport across biological membranes (187) due to its ability of
mimicking interactions of solutes with biomembranes, in vitro (187,188).
Selection of Model Solutes and Appropriate Parameters for Comparison
The n-alcohols were first chosen as a set of simple model solutes because of
the known linear relation between their partition coefficients, Km, in the human
stratum comeum and the permeability coefficients, Kp (cm/h) (12, 68). Therefore,
their transport through human skin is primarily governed by their solubilizing
ability in the stratum comeum.
The steroids were the next set of molecules that were evaluated. They are
more complex and bulky molecules than the n-alcohols and their permeation across
human skin has been reported (12) to be a function of both, their Km and diffusion
coefficient, D (cm2/h) in the membrane. Therefore, the selective interaction of
solutes with the IAM.PC stationary phase could be compared to the combined term
of Km and D (in the case of the steroid) and Km (in the case of the n-alcohols).
It was thought that the skin-transport characteristics of more polar, water-
soluble molecules should be evaluated. The ability of the IAM.PC column to
predict such permeabilities could be taken as indication of the use of the membrane-
column to reproduce or mimic specific interactions of various drugs with the

102
amphiphilic cell-membrane component(s) during their transport across biological
membrane barriers. Hence, different water-soluble drugs and polar, heterocyclic,
nitrogen-containing compounds were evaluated
Heat separated human epidermal preparations have been widely utilized for
assessing permeability of human skin to various solutes, in vitro. Therefore, this
well-established technique was used for the diffusion experiments using side-by-
side, two-compartment cells.
The human skin permeability coefficient, Kp, for various drug molecules
was obtained from the in vitro diffusion experiments. It is a characteristic constant
(16), and a pure measure of the physical-chemical interaction of the solute,
dissolved in a solvent, with the biological membrane. Thus, Kp is directly
proportional to Km and D, and inversely proportional to the thickness of the
membrane, h (cm) (Figure 1-8). Since, the relative retention (capacity factor), K'
of the molecule in the IAM.PC column can be considered as a measure of its
interaction with the phospholipid stationary phase, the relationship between Kp and
K' for evaluation of the IAM.PC column as a model system for studying drug-
transport across human skin was considered to be reasonable.
Prediction of n-Alcohol Transport Across Human Skin
The relationships between Km and Kp, K'(C-18) and Kp, K’(IAM.PC) and
Km, K'(IAM.PC) and Kp for the n-alcohols on a log-log scale are shown in Figure
4-6,4-7,4-8, and 4-9, respectively.

103
Log Km
Figure 4-6 : Relationship between stratum corneum membrane partition
coefficients (Km) and its permeability (Kp) to n-alcohols on a log-log scale.
As shown (Figure 4-6), the permeability of the human skin to the n-alcohols
is well predicted by their ability to partition into the membrane (12,68).
Membrane solubilization and its permeability to solutes are highly specific
phenomena entirely dependent upon the interaction of particular solutes with the
membrane components and the membrane as a whole. In the C-18 column (Figure
4-7), the relation between the relative retention of the alcohols and their
permeability across the human stratum corneum shows a biphasic curve.

104
S
Heptanol
•
â– 
â– 
Hexanol
â– 
Pentanol
â– 
â– 
Butanol
â– 
Ethanol
Propanol
â– 2
-1
0 1
Log K'(C-18)
Figure 4-7 : Relationship between relative retention of n-alcohols in the C-18
column {K'(C-18)} and their stratum corneum membrane permeability
coefficients (Kp) on a log-log scale.
The lower alcohols have higher permeabilities as compared to the higher
alcohols according to the predicted straight line if extrapolated to the abscissa
(Figure 4-7). This suggests that the polar interactions with cell-membrane
components that are important for transport of hydrophilic solutes (such as the
lower alcohols) across biological membrane barriers cannot be predicted by their
interaction with the hydrophobic C-18 stationary phase. The C-18 column is able
to demonstrate only non-specific type of hydrophobic interactions.

105
Log K' (IAM)
Figure 4-8 : Relationship between relative retention of n-alcohols in the IAM.PC
column {K'(IAM.PC)}and their stratum corneum membrane partition coefficients
(Km) on a log-log scale.
On the other hand, the IAM.PC column, due to the presence of the
phospholipid head group on the stationary phase is better able to demonstrate polar
interactions with solute molecules, which gives it added selectivity in predicting
drug-membrane interactions. The IAM.PC column is able to show linear
relationships between the relative retention of the alcohols in the column and their
membrane partition coefficients (Figure 4-8) and the membrane permeability
coefficients (Figure 4-9).

106
Log K' (IAM)
Figure 4-9 : Relationship between relative retention of n-alcohols in the IAM.PC
column {K’(IAM.PC)} and their stratum corneum membrane permeability
coefficients (Kp) on a log-log scale.
These linear profiles indicate ability of the IAM.PC column stationary phase
containing the phospholipid to reproduce or mimic the specific and non-specific
interactions of alcohols with cell-membranes. The lower alcohols like methanol,
ethanol, and propanol mainly show specific, polar interactions of their hydroxyl
functionality with the phospholipid head group. Their aliphatic carbon chain is not
long and extensive enough to undergo non-specific solubilization in the
hydrocarbon environment of the lipid portion of the phospholipid molecule.

107
Whereas, the higher alcohols due to their extended methylene groups are better
able to demonstrate both specific and non-specific interactions with the cell-
membrane component in the column. Thus, there is a smooth transition in the
interaction of lower and higher alcohols with the LAM.PC column stationary phase
(Figure 4-9).
Permeability of Human Epidermis to Steroids
Table 4-11 lists the steroids that were evaluated for their permeation across
the heat-separated human epidermal membrane. All the relevant permeation
parameters from the diffusion studies are included. From the experimentally
determined amounts of the steroid diffusing through the skin, the steady-state flux
was obtained, which was used to calculate the permeability coefficients from the
concentration of the steroid in the donor chamber. The diffusional lag-time for the
steroids in the skin membrane was estimated by extrapolating the diffusion curve to
the abscissa, which is a widely used method (16), to calculate the diffusion
coefficient of the steroids. Thus, the partition coefficients were estimated. Table 4-
11 also lists the relative retention of the various steroids in the IAM.PC and the C-
18 column for comparison.
The steroids which were widely ranging in their lipophilicity are shown in
Figure 4-10. They were treated as separate groups according to their structural
characteristics to clarify their structure-transport characteristics.

108
In the 4-pregnene steroids (Figure 4-10), the increase in Kp from cortisone
to progesterone (Table 4-11) was observed to correlate with the reduction in the
number of polar, hydrogen-bonding groups in the steroid molecule. Presence of a
polar ketone (-C=0) or a hydroxyl (-OH) functional group at the Cn, {3 position
Table 4-11 : Permeability of human epidermis to steroids in vitro at 32 0C, and
their relative retention in die IAM.PC and C-18 HPLC columns.
Steroid
Log J
(|imol/cm2/h)
Log Kp
(cm/h)
Log Km
Log D Log K’ Log K'
(cm2/h) (IAM.PC) (C-18)
Prednisone
-4.14
-3.62
0.00
-5.62
-0.50
-0.13
Cortisone
-4.31
-4.14
-0.62
-5.52
-0.43
-0.10
Hydrocortisone
-4.19
-3.90
0.06
-5.96
-0.39
-0.14
Prednisolone
-4.77
-4.47
-0.34
-6.13
-0.38
-0.17
Dexameth asoné
-4.41
-3.70
-0.29
-5.41
-0.20
0.01
Corticosterone
-3.85
-3.37
-0.12
-5.24
-0.17
0.11
Cortexolone
-3.81
-2.90
1.40
-6.30
-0.06
0.14
Spironolactone
-4.51
-3.12
-0.13
-4.99
0.07
0.57
Cortexone
-2.56
-1.61
1.59
-5.20
0.12
0.43
Testosterone
-3.11
-1.63
0.43
-4.06
0.13
0.40
Progesterone
-2.79
-1.44
2.17
-5.61
0.44
0.86
CH3-Androstanolone -4.18
-2.26
1.12
-5.38
0.47
0.71
Estradiol
-4.23
-2.28
1.08
-5.36
0.52
0.34
17a-Ethinyl-E2
-3.45
-1.98
0.66
-4.64
0.63
0.44
Mestranol
-4.07
-1.97
1.99
-5.95
0.97
1.00
causes a significant reduction in skin-permeability due to increase in polarity (in
the case of cortisone, hydrocortisone, and corticosterone). A similar effect was
observed in earlier experiments by other workers (12). The keto group reduces

109
A4-3-keto-steroid
R
Spironolactone
l
Cortisone O COCH2OH OH
Hydrocortisone OH COCH2OH OH
Corticosterone OH COCH2OH H
Cortexolone H COCH2OH OH
Cortexone H COCH2OH H
Testosterone H OH H
Progesterone H COCH3 H
OH
Methylandrostanolone
Figure 4-10 : Chemical structures of the steroids.

110
the skin-permeability to a greater extent as compared to the hydroxyl group (in the
case of cortisone and hydrocortisone). Also, replacement of the C17, a-OH group
(in hydrocortisone) by hydrogen (in corticosterone) leads to a significant increase
in skin-permeability. Thus, it is evident that removal of -OH groups from either
the C11, (3 or the C17, a position, or both on the steroid molecule should lead to
enhancement in the skin-transport ability.
However, a polar functional group at the Cn, p position is more critical in
reducing the skin-permeability properties of the steroid. Corticosterone and
cortexolone are structural isomers differing only in the substitution of -OH group
atCn, P position (for corticosterone) or at C17, a position (for cortexolone). It
was observed from the diffusion studies that the permeability properties of
cortexolone are significantly greater than those of corticosterone. Hence, the skin-
permeability to steroids is highly sensitive to polar functional group substitution at
the C11, P position of the steroid.
It is interesting to note that the IAM.PC column is better able to resolve
corticosterone and cortexolone. The difference in the relative retention of
cortexolone and corticosterone in the IAM.PC column was significantly greater
than that in the C-18 column. This could be attributed to the preferential ability of
the phospholipid moiety of the IAM.PC column stationary phase to interact with
the P face of the steroid. Therefore, the presence of the polar, Cn, p-OH group on
corticosterone perturbed and reduced the non-specific, hydrophobic interactions
between the hydrocarbons of the phospholipid and the lipophilic steroid-ring

center, thus causing less retention. This does not happen in the case of cortexolone
which lacks the Cn, (3 hydroxyl group.
Figure 4-11, 4-12, 4-13, and 4-14 show the relationships between Km and
Kp, K’(C-18) and Kp, K'(IAM) and Km, and K'(IAM) and Kp, on log-log scale,
respectively. These relations clearly show that the partitioning of the steroids in
the human epidermis is not a good predictor of its permeability (Figure 4-11).
Log Km
Figure 4-11 : Relationship between human epidermal membrane partition
coefficients (Km) and its permeability (Kp) to steroids on a log-log scale.

112
Figure 4-12 shows that the partitioning ability of the steroids on the C-18
stationary phase do not predict their human epidermal membrane permeabilities.
This was expected since the C-18 column would not demonstrate specific
interactions of steroids with cell-membranes. Thus, the C-18 column due to its
ability to demonstrate only partition effects (Figure 4-12) of the steroids with the
alkane stationary phase was unable to reproduce the specificity of the interaction of
the steroids with the phospholipid stationary phase in the IAM.PC column (Figure
4-14).
Figure 4-12 : Relationship between relative retention of the steroids in the C-18
column (K'(C-18)} and their human epidermal membrane permeability
coefficients (Kp) on a log-log scale.

113
Figure 4-13 : Relationship between relative retention of the steroids in the IAM.PC
column {K'(IAM.PC)} and their human epidermal membrane partition
coefficients (Km) on a log-log scale.
The IAM.PC column was able to demonstrate significantly better relationship
between the ability of the steroids to be retained in the column versus their
transport ability across the human epidermal membrane (Figure 4-14), but not
with their ability to partition into the membrane (Figure 4-13).
Skin-Permeabilitv of Water-Soluble Drugs
Water-soluble drugs generally show decreasing skin-permeability with
increasing molecular weight (167). The skin-permeabilities to these ionizable

114
Log K' (IAM)
Figure 4-14 : Relationship between relative retntion of the steroids in the IAM.PC
{K'(IAM.PC)} column and their human epidermal membrane permeability
coefficients (Kp) on a log-log scale.
solutes fail to show a particular dependence on their lipophilicities (indicated by
octanol-water partition coefficients) or their equilibrium partitioning into the skin-
membrane. This is because the polar, ionizable, and water-soluble solutes such as
the ones evaluated below depend on the polar, aqueous channels available in the
“intracellular" pathway for diffusion, and hence, the predominant influence of
molecular size and weight on drug-diffusion is observed. That was the original

115
reason for putting forward the theory and concept (16) of two pathways
(intercellular and intracellular) for drug-diffusion in the skin. The “intercellular"
pathway is highly dependent on the ability of lipophilic solutes to partition into and
move along the concentration gradient in the lipid domain of the intercellular
compartment.
Six model water-soluble drugs (Figure 4-15) whose permeabilities through
full-thickness, hairless-rat abdominal skin were previously reported (167) were
evaluated for their IAM.PC column retention. Their molecular weights in the
increasing order are Dopamine (189) < Isoproterenol (247) < Diclofenac (318) <
Papaverine (357) < Diltiazem (451) < Cromolyn (512).
Figure 4-16 shows the relation between log K’ (IAM.PC) and log Kp using a
more polar mobile phase acetonitrile : methanol: DPBS 10 x, pH 7.1 (10 : 10 : 80).
It indicates an inverse linear relationship. Dopamine is more polar in nature than
Diltiazem. Therefore, the interaction of the more polar dopamine with the mobile
phase containing less organic modifyer is enhanced and it is unable to show polar
interaction with the ionized phospholipid head group of the IAM.PC column
stationary phase leading to its less retention and faster elution (Figure 4-16).
Diltiazem, on the other hand, shows greater affinity towards the lipid portion of
the IAM.PC column stationary phase, in the presence of the mobile phase
containing less organic modifyer.

116
ho^Q^NH2HC1
HO
Dopamine HC1
OCH3
\
HO
(-) Isoproterenol HQ
Papaverine HQ
Diltiazem HQ
Cromolyn Na
Figure 4-15 : Chemical structures of the water-soluble drugs.

1 1 7
-6
3
73
s
3
G,
60
â–º2
-8
Dopamine (189.6)
â– 
Isoproterenol (247.7)
Diclofenac (318.1)
Papaverine (357.8)
Diltiazem (451.0)
Log K’ (IAM)
Figure 4-16 : Relationship between relative retention of the water-soluble drugs in
the IAM.PC column {(K'(IAM.PC)} using (ACN/MeOH/DPBS lOx; 10/10/80) as
the mobile phase and their human epidermal membrane permeabilities (Kp) on a
log-log scale.
This relationship is reversed in Figure 4-17 due to the effect of increasing
the concentration (methanol : DPBS 10 x ; 50 : 50) of the organic phase in the
mobile phase. Thus, the more lipophilic solutes (Diltiazem and Papaverine) show
less interaction than the polar solute (Dopamine) with IAM.PC stationary phase.
Such a reversal of retention profile of the solutes ranging in polarity from the
more polar dopamine to the less polar diltiazem, indicates change in the retention

118
mechanism on the phospholipid stationary phase in the IAM.PC column, depending
on the physical-chemical interaction between the drug, mobile phase, and the
phospholipid on the solid silica support. It is worthwhile to note that a similar
interaction occurs when a drug in a vehicle is in contact with the biological
membrane barrier.
-6
a -7H
t
o
^—✓
Q.
ba
2 ^
-2
Diclofenac!
Papaverine I
Dopamine
Isoproterenol
Cromolyn
Diltiazem
-1
Log K' (IAM)
Figure 4-17 : Relationship between relative retention of the water-soluble drugs in
the IAM column {K'(IAM.PC)} using (MeOH/DPBS lOx; 50/50) as the mobile
phase and their human epidermal membrane permeabilities (Kp) on a log-log scale.

119
Dopamine â– 
â– 
Isoproterenol
â–  Diclofenac
â–  Papaverine
â– 
Cromolyn
â– 
Diltiazem
-9 H ■ 1 1 »
-10 12
Log K' (C-18)
Figure 4-18 : Relationship between relative retention of the water-soluble drugs in
the C-18 column {K’(C-18)} using (MeOH/DPBS lOx; 50/50) as the mobile phase
and their human epidermal membrane permeabilities (Kp) on a log-log scale.
Furthermore, this phenomenon cannot exist in the C-18 column due to its
inability to demonstrate solute specific interactions with biomembranes because of
the lack of amphiphilic nature of the C-18 stationary phase in the reverse phase
column. Indeed, in Figure 4-18 with the relatively non-polar mobile phase
(methanol: DPBS lOx ; 50 : 50) and using the C-18 column, the relationship is

120
again reversed with respect to the interaction of the same solutes in the IAM.PC
column (Figure 4-17).
This is an interesting feature of the phospholipid stationary phase in the
IAM.PC column. The reversal of elution profile of the polar and relatively more
non-polar solutes with changes in polarity of the mobile phase may be a direct
result of the changes in the relative magnitude of interaction of the drug with either
the polar, ionized, phosphocholine head group, or with the lipid-like environment
of the hydrocarbon chains of the phospholipid. Hence, the IAM.PC column is
capable of exhibiting dual (polar and non-polar) type of physical-chemical
interactions with solutes. Figure 4-16, 4-17, and 4-18 together seem to indicate the
above specific and non-specific interactions that may be occurring due to
combination of physical-chemical behaviour of the drug, mobile phase, and the
phospholipid. This is analogous to the physical-chemical interactions between the
drug, vehicle, and the biological membrane barrier.
In addition, it should be noted that due to the ionized and polar nature of the
choline head group on the phosphatidylcholine stationary phase, the IAM.PC
column could also demonstrate ion-exchange capability.
Transport of Nucleosides Across the Human Epidermal Membrane
The eight nucleosides that were evaluated for their skin-permeability are
shown in Figure 4-19. They are structurally related in the sense that the
substitution pattern on four sites of the molecule were different. The C4 position of

121
Cytidine
5-Fluorodeoxyuridine Cytosine arabinoside
2’-Deoxycytidine
Trifluorothymidine 3'-Deoxythymidine
5-Iodo-2'-Deoxyuridine
Azidothymidine
Figure 4-19 : Chemical structures of the nucleosides.

122
the heterocyclic base contained either a primary amine or a ketone. The C5 was
substituted with either a hydrogen, a fluorine, a trifluoromethyl, a methyl, or a
iodo group. The sugar had either a 2'-OH or H, and the 3' carbon either a -OH, or
H, or azido functional group. As a result of these substitutions the nucleosides had
a wide range in lipophilicity.
Table 4-12 lists their log relative retention in the IAM.PC.MG capped
column along with their skin-permeation parameters. The K'(IAM.PC.MG)
increases from cytidine to azidothymidine. This increase follows a similar
relationship with the skin-permeability (Kp) of the nucleosides (Figure 4-20).
Table 4-12 : Transport of nucleosides across heat-separated human epidermal
membrane in vitro at 32 0C and their relative retention in the IAM.PC.MG capped
column.
Nucleoside
LogK'a
Log Kp
Log J
Log D
Log Km
(IAM.PC.MG)
(cm/h)
(|imol/cm2/h)
(cm2/h)
Cytidine
-0.82
-7.05
-4.65
5-Fluorodeoxyuridine
-0.76
-5.65
-3.35
-2.52
-3.13
Cytosine arabinoside
-0.74
-5.58
-2.96
-4.43
-3.14
2'-Deoxycytidine
-0.67
-6.62
-3.65
-5.96
-2.66
TFT
-0.24
-5.60
-3.85
-5.43
-2.52
3’-Deoxythymidine
-0.18
-4.37
-2.58
-4.87
-1.50
5-Iodo-2'-Deoxyuridine
-0.05
-2.51
-1.59
AZT
0.11
-1.87
0.25
-5.56
1.69
a Log capacity factors for the nucleosides in the IAM.PC.MG capped column
using the mobile phase DPBS 1 Ox at a flow rate of 1.0 ml/min; UV A = 265 nm and
the retention time of the unretained citric acid was t0 = 2.03 minutes.

123
The partitioning ability of the nucleosides may be directly responsible for their
increasing permeability in the skin membrane. As a result, the relation between
K'(IAM.PC.MG) and K m is similar to K'(IAM.PC.MG) and Kp, on a log-log scale.
The presence of the primary amine in the C4 position along with a
corresponding hydrogen at C5 (in the case of cytidine, cytosine arabinoside, and 2'-
deoxycytidine) has more permeability-reducing property as compared to the
S
o
ex
0.2
Log K' (IAM)
Figure 4-20 : Relationship between relative retention of the nucleosides in the
IAM.PC.MG capped column {K'(IAM.PC.MG)} using DPBS 1 Ox as the mobile
phase and their human epidermal membrane permeability coefficients (Kp) on a
log-log scale.

124
ketone at C4 and a methyl group at C 5 position (in the case of 3’-deoxythymidine
and azidothymidine). The iodo group at C5 enhances permeability due to its
greater lipid solubility (in the case of 5-Iodo-2'-deoxyuridine). The azido group at
3' position of the sugar moiety increases the lipophilicity of the compound and is
reponsible for the enhanced transport ability of azidothymidine across the human
epidermal membrane.

CHAPTER V
SUMMARY AND CONCLUSIONS
Redox-based chemical targetting systems for improving the delivery of the
important antiviral agent, acyclovir to the skin were designed, synthesized, and
evaluated. They were designed to be more lipophilic than acyclovir so as to
facilitate their solubilization in the skin, followed by the metabolic ability of the
skin to form reservoirs of chemical precursors for the slow release of acyclovir in
the skin.
The reduced, 1,4-dihydrotrigonelline ester containing moiety for acyclovir
(A-CDS) was based upon its ability to undergo a dual-physical chemical change in
the molecule, after skin penetration. It was rapidly oxidized to form the
quaternary metabolite, AQ+, from which a slow, rate-limiting hydrolysis by the
non-specific esterases resulted in a steady release of acyclovir in the skin. It was
also noted that the formation of AQ+ in the skin was an important prerequisite for
the availability of acyclovir in the skin. In fact, this was one of the specific aims
which was to be explored for the project and was realized. Thus, the A-CDS was
able to improve the delivery of acyclovir to the hairless-mouse skin, per unit area,
per unit dose, by about 9 fold at 6 hours as compared to the underivatized
acyclovir.
125

126
The lipoic acid ester of acyclovir (A-LipS2), on the other hand, was based
upon its ability to undergo a reduction mechanism followed by a covalent disulfide
bond formation in the skin. This particular chemical delivery system was able to
accumulate in the skin with time. Its large concentration formed a reservoir of A-
LipS2 in the skin and was responsible for the release and localization of acyclovir
specifically in the skin. As a result, the A-LipS2 improved the delivery of
acyclovir per unit dose, by about 96 and 105 fold even at early time periods of 2
and 4 hours, respectively, compared to underivatized acyclovir. Thus, A-LipS2
was selective in targeting delivery of acyclovir to the skin.
The lipolyl ester of carboxypropylcarbamoyl-5-fluorouracil (5-FU-LipS2)
was also designed, synthesized, and evaluated for its ability to deliver 5-
fluorouracil to the hairless-guinea pig skin. Although, much less effective than A-
LipS2, it improved the delivery of 5-fluorouracil to the skin by about 4 fold at 6
hours, compared to underivatized 5-fluorouracil.
In all the above three cases, it was observed that the improvement in drug-
delivery due to the chemical delivery system occurs in the early time period of 2-6
hours. Thereafter, it starts to decline. Such chemical approaches could be useful in
targetting not only antiviral and anticancer agents, but also other potent therapeutic
agents and pharmaceuticals specifically to the skin. It could lead to improvement in
the therapeutic effectiveness of drugs which demonstrate poor skin-transport
properties.

127
The targeting systems were more lipophilic than the parent underivatized
drug as determined by their partitioning ability into the skin membranes. This
enhanced lipophilicity was thought to be responsible for achieving high skin
concentration of the delivery systems. On the other hand, the metabolic processes
in the skin-tissue were responsible for the formation and localization of metabolic
reservoir followed by actual release of the parent drug and hence its availability in
the skin. It was conclusively shown that the targetting systems were capable of
releasing the parent drug under near physiological conditions. The carriers
exhibited very short half-lives in biological media like the rat blood, human
plasma, and skin homogenate. On the other hand, they showed good stability in
aqueous, pH 7.4 buffer system, thus indicating their susceptibility towards
enzymatic degradation, in vivo.
The in vitro model that was developed for studying drug-transport across
human skin was the Immobilized Artificial Membrane Phosphatidylcholine
stationary phase (IAM.PC) HPLC column. This column utilized a solid stationary
phase consisting of an analog of phospholipid (monomyristoyl lysolecithin), which
was covalently linked to the silica surface through propylamine.
In all the examples of drugs studied it was evident that the permeability of
biological membranes to various drugs is an important physical-chemical constant
that should be compared with the retention ability of the drug in the IAM column.
The column has shown promise in predicting the permeability of the human
epidermal membrane, which is the principal barrier to transport of most drugs and

128
pharmaceuticals across human skin either for transdermal or dermal drug-
delivery. This important, novel, in vitro. HPLC method can be a rapid, and simple
method of analyzing potential transdermal candidates (of polar or non-polar
nature) for their ability to be transported across the epidermal membrane barrier
of the human skin. It is also hoped that this model can be useful in reproducing or
predicting amphiphilic drug-cell membrane interactions. Hence, from a
fundamental viewpoint, it could be used to study the specific and non-specific
interactions of various solutes with cell-membranes.

REFERENCES
1. S. W. Jacob and C. A. Francone, Structure and Function of Man. 2nd ed.,
Saunders, Philadelphia, 1970, pp. 55-60.
2. Y. W. Chien, Transdermal Controlled Systemic Medications. Drugs and
the Pharmaceutical Sciences, Vol. 31, Marcel Dekker, New York and
Basel, 1987, pp. 2-4.
3. H. Schaefer, J. Soc. Cosmet. Chem. 24. 331-353, 1973.
4. H. Shaefer, A. Zesch, and G. Stuttgen, Skin Permeability. Springer-
Verlag, New York, 1982, pp. 545-550.
5. R. K. Winkelmann. Br. J. Derm. 81 fSuppl. 41. 11-22, 1969.
6. K. Halprin, A. Ohkawara, and K. Adachi, J. Invest. Dermatol. 49. 559-560,
1967.
7. G. Muller and F. Nurnberger, Arch. Dermatol. Forsch. 244. 171-172,
1972.
8. A. M. Kligman. The Epidermis. W. Montagna and W. C. Lobitz, Jr. (eds.),
Academic Press, New York, 1964, pp. 387-433.
9. R. Fleischer, Habilitation Schrift 81. 1877.
10. A. Schwenkenbecker. Arch. Anat. Physiol. 28. 121-127, 1904.
11. A. M. Kligman and E. Christophers, Arch. Dermatol. 88. 702, 1964.
12. R. J. Scheuplein and I. H. Blank, Phvsiol. Rev. 51.702-747, 1971.
13. I. H. Blank, J. Invest. Derm. 21.259-269, 1953.
14. G. S. Berenson and G. E. Burch, Am. J. Trop. Med. 31. 843-849, 1951.
129

130
15. J. D. Middleton, Br. J. Derm. 80. 437-450, 1968.
16. A. S. Michaels, S. K. Chandrasekaran, and J. E. Shaw, Am. Inst. Chem,
Eng. J. 21. 985-996,1975.
17. P. M. Elias and D. S. Friend, J. Cell Biol. 65. 180-191, 1975.
18. P. M. Elias, N. S. McNutt, and D.S. Friend, Anat. Rec. 189. 577-593, 1977.
19. P. M. Elias, J. Goerke, and D.S. Friend, J. Invest. Derm. 69. 535-546, 1977.
20. P. M. Elias. Int. J. Derm. 20. 1-19. 1981.
21. W. P. Smith, M. S. Christensen, S. Macht, and E. H. Gans, J. Invest. Derm.
78, 7-11,1982.
22. S. Grayson and P. M. Elias, J. Invest. Derm. 78. 128-135,1982.
23. P. M. Elias, L. Bonar, S. Grayson, and H. P. Baden, J. Invest. Derm. 80. 213-
214, 1983.
24. P. M. Elias, M. L. Williams, M. E. Maloney, J. A. Bonifas, B. E.
Brown, S. Grayson, and E. H. Epstein, Jr., J. Clin. Invest. 74. 1414-1421,
1984.
25. G. K. Menon, S. Grayson, and P. M. Elias, J. Invest. Derm. 86. 591-597,
1986.
26. S. Grayson, A. G. Johnson-Winegar, and P. M. Elias, Science 221.962-964,
1983.
27. S. Grayson, A. G. Johnson-Winegar, B. U. Wintroub, E. H. Epstein, Jr.,
and P. M. Elias, J. Invest. Derm. 85. 289-295,1985.
28. P. W. Wertz, D. T. Downing, R. K. Freinkel, and T. N. Traczyk, J. Invest.
Derm. 83. 193-195, 1984.
29. P. W. Wertz and D. T. Downing, Science 217.1261-1262,1982.
30. G. K. Menon, S. Grayson, and P. M. Elias, Clin. Res. 35. 250A, 1987.
31. R. K. Freinkel and T. N. Traczyk. J. Invest. Derm. 80.441-446. 1983.
R. K. Freinkel and T. N. Traczyk, J. Invest. Derm. 85.295-298, 1985.
32.

131
33. P. M. Elias, K. R. Feingold, G. K. Menon, S. Grayson, M. L. Williams, and
G. Grubauer, Skin Pharmacokinetics. Pharmacology and the Skin, Vol. 1,
B. Shroot and H. Schaefer (eds.), Karger, Basel, 1987, p. 3.
34. G. M. Gray and H. J. Yardley, J. Lipid Res. 16. 441-447, 1975.
35. P. A. Bowser and R. J. White, Br. J. Derm. 112. 1-14, 1985.
36. M. A. Lampe, M. L. Williams, and P. M. Elias, J. Lipid Res. 24. 131-140,
1983.
37. T. Winsor and G. E. Burch, Arch. Intern. Med. 74. 428-444, 1944.
38. I. H. Blank, J. Invest. Derm. 45. 249-256, 1965.
39. R. J. Scheuplein, J. Invest. Derm. 67. 31-38, 1976.
40. A. G. Matoltsy, A. M. Downes, and T. M. Sweeney, J. Invest. Derm. 50. 19-
26, 1968.
41. T. M. Sweeney and D. T. Downing, J. Invest. Derm. 55. 135-140, 1970.
42. A. S. Breathnach, T. Goodman, C. Stolinski, and M. Gross, J. Anat. 114.65-
81, 1973.
43. K. C. Madison, D. C. Swartzendruber, P. W. Wertz, and D. T.
Downing, J. Invest. Derm. 88. 714-718, 1987.
44. P. W. Wertz, D. C. Swartzendruber, K. C. Madison, and D. T.
Downing, J. Invest. Derm. 89. 419-425, 1987.
45. C. A. Squier, J. Ultrastr. Res. 43.160-167,1973.
46. H. J. Yardley and R. Summerly, Pharmacol. Ther. 13. 357-383,
1981.
47. H. J. Yardley, Biochemistry and Physiology of the Skin. Vol. I. L. A.
Goldsmith (ed.), Oxford University Press, New York, 1983, pp. 363-381.
48. P. W. Wertz, Biology of the Integument. Vol. II. J. Bereiter-Hahn, A. G.
Matoltsy, and K. S. Richards (eds.), Springer-Verlag, Berlin, 1986, pp. 815-
823.
49. R. Fettypace and D. A. Haydon, Physiol. Rev. 60. 510-550, 1980.

132
50. C. L. Hedberg, P. W. Wertz, and D. T. Downing, J. Invest. Penn. 90. 225-
229, 1988.
51. M. Nicollier, T. Marsengo, J. P. Remy-Martin, R. Laurent, and G. L.
Adessi, J. Invest. Derm. 87. 68-71, 1986.
52. P. W. Wertz and D. T. Downing, Transdermal Drug Delivery. Drugs and
the Pharmaceutical Sciences, Vol. 35, J. Hadgraft and R. H. Guy (eds.),
Marcel Dekker, New York, 1989, pp. 10-11.
53. P. W. Wertz and D. T. Downing, J. Lipid Res. 24. 759-765, 1983.
54. P. W. Wertz, M. C. Miethke, S. A. Long, J. S. Strauss, and D. T.
Downing, J. Invest. Derm. 84. 410-412, 1985.
55. L. Landmann, P. W. Wertz, and D. T. Downing, Biochim. Biophys. Acta
ZZ&, 412-418, 1984. "
56. W. Abraham, P. W. Wertz, and D. T. Downing, Biochim. Biophvs. Acta
939. 403-408, 1988.
57. P. M. Elias, Topical Drug Delivery Formulations. Drugs and the
Pharmaceutical Sciences, Vol. 42, Marcel Dekker, New York, 1990, pp. 17-
24.
58. P. M. Elias, G. K. Menon, S. Grayson, and B. E. Brown, J. Invest. Derm.
91, 3-10,1988.
59. T. Yano, A. Nakagawa, M. Tsuji, and K. Noda, Life Sciences 39. 1043-
1050,1986.
60. K. Thoma. S.T.P. Pharma 3. 39. 1987.
61. J. Hadgraft, Prediction of Percutaneous Penetration : Methods.
Measurements. Modelling. R. C. Scott, R. H. Guy, and J. Hadgraft (eds.),
IBC Technical Services, London, 1990, pp. 252-262.
62. T. Higuchi, Design of Biopharmaceutical Properties Through Prodrugs and
Analogs. E. B. Roche (ed.), APhA Academy of Pharmaceutical Siences
Washington, D.C., 1977, pp. 409-421.
63. J. Hadgraft, Design of Prodrugs. H. Bundgaard (ed.), Elsevier,
Amsterdam, 1985, pp. 271-289.

133
64. V. J. Stella and K. J. Himmelstein, J. Med. Chem. 23. 1275-1282,1980.
65. V. J. Stella, T. J. Mikkelson, and J. D. Pipkin, Drug Delivery Systems:
Characteristics and Biomedical Applications. R. L. Juliano (ed.), Oxford
University Press, New York, 1980, pp. 112-176.
66. R. J. Scheuplein, Pharmacology and the Skin. Advances in Biology of the
Skin, Vol. 12, W. Montagna, E. J. Van Scott, and R. B. Stoughton (eds.),
Appleton-Century-Crofts, New York, 1972, pp. 125-152.
67. W. E. Clendenning and R. B. Stoughton, J. Invest. Derm. 39. 47-49, 1962.
68. I. H. Blank and R. J. Scheuplein, Br. J, Derm. 81 ISuppl. 41.4-10, 1969.
69. D. E. Wurster and S. F. Kramer, J. Pharm. Sci. 50. 288-293, 1961.
70. K. B. Sloan, Adv. Drug Del. Rev. 3. 67-101,1989.
71. W. I. Higuchi, N. A. Gordon, J. L. Fox, and N. F. H. Ho, Drug Dev. Ind.
Pharm. 9. 691-706, 1983.
72. T. Higuchi, Bioreversible Carriers in Drug Design: Theory and
Application. E. B. Roche (ed.), Pergamon Press, New York, 1987, pp. 1-
12.
73. N. Bodor, K. B. Sloan, R. J. Little, S. H. Selk, and L. Caldwell, Int. J.
Pharm. 10. 307-321, 1982.
74. T. Loftsson and N. Bodor, J. Pharm. Sci. 70. 756-758,1981.
75. T. Inagi, T. Muramatsu, H. Nagai, and H. Terada, Chem. Pharm. Bull. 29.
1708-1714,1981.
76. O. D. Laerum, J. Invest. Derm. 52. 204-211,1969.
77. U. Tauber, Dermal and Transdermal Absorption. R. Brandau and B. H
Lippold (ed.), Wissenschaftliche Verlagsgesellschaft mbH, Stuttgart, 1982,
pp. 133-151.
78. P. Mauvais-Jarvis, J. P. Bercovici, D. Crepy, and F. Gauthier, J. Clin.
Invest. 49.31-37. 1970.
79.
R. C. Wester and P. K. Noonan, Int. J. Pharm. 7. 99-110,1980.

134
80. M. D. Rawlins, S. Schuster, P. H. Chapman, V. Shaw, and V. A.
O'Neill, Drug Metabolism in Skin. P. Turner (ed.), Macmillan,
London, 1980, pp. 410-414.
81. P. K. Noonan and R. C. Wester, Percutaneous Absorption. R. L.
Bronough and H. I. Maibach (eds.), Marcel Dekker, New York, 1985,
pp. 65-85.
82. R. C. Wester, P. K. Noonan, S. Smeach, and L. Kosobud, J. Pharm. Sci.
72,745-748, 1983.
83. M. S. Greaves, J. Invest. Derm. 57. 100-107, 1971.
84. L..H. Smith and J. M. Holland, Toxicology 21. 47-57, 1981.
85. J. M. Holland, J. Y. Kao, and M. J. Whitaker, Toxicol. Appl, Pharmacol.
22, 272-280,1984.
86. A. Alvares, D. Bickers, and A. Kappas, Fed. Proc. 32. 235 Abs, 1973.
87. R. Pohl, R. Philpot, and J. Fouts, Drug. Metab. Pispos. 4. 442-450, 1976.
88. J. Bamshad, J. Invest. Derm. 52. 351-352, 1969.
89. G. Weinstein, P. Frost, and S. Hsia, J. Invest. Derm. 51.4-10,
1968.
90. P. H. Chapman, M. D. Rawlins, and S. Schuster, Br. J. Clin. Pharmacol.
4, 393 P, 1977.
91. T. J. Franz, J. Invest. Derm, 64. 190-195, 1975.
92. R. L. Bronaugh, E. R. Congdon, and R. J. Scheuplein, J. Invest. Derm. 76.
94-96, 1981.
93. H. Durrheim, G. L. Flynn, W. I. Higuchi, and C. R. Behl, J. Pharm. Sci.
£9, 781-786, 1980.
94. R. B. Stoughton, Animal Models in Dermatology. H. I. Maibach (ed.),
Churchill-Livingstone, New York, 1975, pp. 121-132.
95. R. L. Bronaugh, R. F. Stewart, and E. R. Congdon, Toxicol. Appl.
Pharmacol. 62. 481-488, 1982.

135
96. R. C. Wester and H. I. Maibach, Toxicol. Appl. Pharmacol. 32,394-398,
1975.
97. R. L. Bronaugh, R. C. Wester, D. Bucks, R. F. Stewart, J. Anderson, and H.
I. Maibach. Food. Chem. Toxicol. 23. 111-114. 1985.
98. R. J. Feldmann and H. I. Maibach, J. Invest. Derm. 52. 89-94, 1969.
99. R. J. Feldmann and H. I. Maibach, J. Invest. Derm. 54. 339-404, 1969.
100. R. J. Feldmann and H. I. Maibach, Toxicol. Appl. Pharmacol. 28. 126-132,
1974.
101. F. N. Marzulli, D. W. C. Brown, and H. I. Maibach, Toxicol. Appl.
Pharmacol. Suppl. 3.79-83. 1969.
102. A. H. McGreesh, Toxicol. Appl. Pharmacol. Suppl. 2.20-26, 1965.
103. M. J. Bartek, J. A. La Budde, and H. I. Maibach, J. Invest. Derm. 58. 114-
123, 1972.
104. M. J. Bartek and J. A. La Budde, Animal Models in Dermatology. H. I.
Maibach (ed.), Churchill-Livingstone, New York, 1975, pp. 103-120.
105. C. Hansch and A. Leo, Substituent Constants for Correlation Analysis in
Chemistry and Biology. Wiley, New York, 1979, pp. 171-330.
106. M. S. Roberts, R. A. Anderson, and J. Swarbrick, J. Pharm. Pharmacol.
22, 677-683,1977.
107. R. J. Scheuplein, J. Invest. Derm. 45. 334-346, 1965.
108. J. E. Treheme. J. Phvsiol. 133. 171-180.1956.
109. F. N. Marzulli, J. F. Callahan, and D. W. C. Brown, J. Invest. Derm. 44.
339-344, 1965.
110. R. J. Scheuplein, Handbook of Physiology : Reactions to
Environmental Agents. D. H. K. Lee (ed.), Williams and Wilkins,
Baltimore, 1977, pp. 299-322.
111.
H. Tsuruta, Ind. Health 13. 227-236, 1975.

136
112. G. B. Elion, P. A. Furman, J. A. Fyfe, P. De Miranda, L. Beauchamp,
and H. J. Schaeffer, Proc. Natl. Acad. Sci. USA 74. 5716-5720, 1977.
113. H. J. Schaeffer, L. Beauchamp, P. De Miranda, G. B. Elion, and D. J.
Collins, Nature 212. 583-585, 1978.
114. G. B. Elion. Am. J. Med. 73.7-13. 1982.
115. S. A. Plotkins, S. E. Starr, and C. K. Bryan, Am, J. Med. 73. 257-261,
1982.
116. J. A. Fyfe, P. M. Keller, P. A. Furman, R. C. Miller, and G. B. Elion, J.
Biol. Chem. 253. 8721-8727, 1978.
117. G. B. Elion. J. Antimicrob. Chemother. 12 TSuppl. B1. 9-17. 1983.
118. W. H. Miller and R. L. Miller, Biochem. Pharmacol. 31. 3879-3884, 1982.
119. P. A. Furman, M. H. St. Clair, J. A. Fyfe, J. L. Rideout, P. M. Keller,
G. B. Elion, J. Virol. 32. "2-77, 1979.
120. E. R. Kern, J. T. Richards, L. A. Glasgow, J. C. Overall, and P. De
Miranda, Am. J. Med. 73. 125-131, 1982.
121. E. De Clerq and R. T. Walker, Pharmacol. Ther. 26. 1-44, 1984.
122. D. S. Parris and J. E. Harrington, Antimicrob. Agents Chemother. 22,71-
77, 1982.
123. B. A. Larder and G. Darby, J. Virol. 42. 649-658, 1982.
124. V. Veerisetty and G. A. Gentry, J. Virol. 46. 901-908,1983.
125. G. W. Rabom, Acyclovir Therapy for Herpesvirus Infections. D. A. Baker
(ed.), Marcel Dekker, New York, 1990, pp. 51-63.
126. S. L. Spruance, L. E. Schnipper, J. C. Overall, Jr., E. R. Kern, B.
Wester, J. Modlin, G. Wenerstrom, C. Burton, K. A. Arndt, G. L. Chiu, and
C. S. Crumpacker. J. Infect. Pis. 146. 85-90, 1982.
127. S. L. Spruance, C. S. Crumpacker, L. E. Schnipper, E. R. Kern, S.
Marlowe, K. A. Arndt, and J. C. Overall, Jr., Antimicrob. Agents
Chemother. 25. 553-555, 1984.

137
128. G. W. Raborn, W. T. McGaw, M. Grace, S. Samuels, and L. D.
Tyrrell, J. Am. Dent. Assoc. 115. 38-42, 1987.
129. J. R. Gibson, M. R. Klaber, S. G. Harvey, A. Tosti, D. Jones, J. M. Yeo,
Dermatológica 172. 104-107, 1986.
130. A. P. Fiddian, J. M. Yeo, R. Stubbings, and D. Dean, Br, Med. J. 286.
1699-1701,1983.
131. H. C. Krasny, P. De Miranda, M. R. Blum,and G. B. Elion, J. Pharmacol.
Ther. 216. 281-288, 1981.
132. J. M. Yeo and A. P. Fiddian, J. Antimicrob. Chemother. 12 ISuppl. Bl. 95-
103.1983.
133. J. C. Martin, M. A. Tippie, D. P. C. McGee, and J. P. H. Verheyden, L
Pharm. Sci. 76. 180-184, 1987.
134. S. S. Good, H. C. Krasny, G. B. Elion, and P. De Miranda, J. Pharmacol.
Exp. Ther. 227. 644-651, 1983.
135. T. Spector, T. E. Jones, L. M. Beacham, Biochem. Pharmacol. 32. 2505-
2509.1983.
136. T. A. Krenitsky, W. W. Hall, P. De Miranda, L. M. Beauchamp, H. J.
Schaeffer, and P. D. Whiteman, Proc. Natl. Acad. Sci. USA 81. 3209-
3213, 1984.
137. P. Selby, S. Blake, E. K. Mbidde, E. Hickmott, R. L. Powles, K. Stolle,
T. J. McElwain, P. D. Whiteman, and A. P. Fiddian, Lancet 2. 1428-1430,
1984.
138. P. D. Whiteman, A. Bye, A. S. E. Fowle, S. Jeal, G. Land, and J.
Posner, Eur. J. Clin, Pharmacol. 27. 471-475. 1984.
139. L. Colla, E. De Clerq, R. Busson, and H. Vanderhaeghe, J. Med. Chem.
2£, 602-604,1983.
140. R. A. Kenley, S. E. Jackson, J. S. Winterle, Y. Shunko, and G. C. Visor, 2.
Pharm. Sci. 75. 648-653, 1986.
141. E. Jensen and H. Bundgaard, Pharm. Res. 7 (Suppü. S-201,1990.

138
142. L. Corey, A. J. Nahmais, M. E. Guinan, J. K. Benedetti, C. W.
Critchlow, and K. K. Holmes, N. Eng. J. Med. 306. 1313-1319, 1982.
143. M. Shaw, M. King, J. M. Best, J. E. Banatvala, J. R. Gibson, and M. R.
Klaber, Brit. Med. J. 291. 7-9, 1985.
144. D. J. Freeman, N. V. Sheth, and S. L. Spruance, Antimicrob. Agents
Chemother. 29. 730-732, 1986.
145. N. Bodor, Strategy in Drug Research. J. A. K. Buisman (ed.), Elsevier,
New York, 1982, pp. 137-164.
146. N. Bodor and K. B. Sloan, J. Pharm. Sci.71.514-520, 1982.
147. W. M. Schubert and Y. Motoyama, J. Am. Chem. Soc. 87. 5507-5508,
1965.
148. M. E. Brewster, V. Venkatraghavan, E. Shek, and N. Bodor, Svn.
Comm. 17.451-455. 1987.
149. V. Venkatraghavan, E. Shek, R. Perchalski, and N. Bodor, submitted for
publication.
150. P. Chikhale and N. Bodor, J. Pharm. Sci.. 1991.
151. M. F. Powell and T. C. Bruice, Prog. Clin. Biol. Res. 272. 369-385, 1988.
152. N. J. Oppenheimer, R. M. Davidson, T. O. Matsunaga, and B. L. Kam,
Mechanisms of Oxidizing Enzymes. T. P. Singer and R. N. Ondarza (ed.),
Elsevier, New York, 1978, pp. 81-93.
153. C. H. Williams, Jr., C. Thorpe, and L. D. Arscott, Mechanisms of
Oxidizing Enzymes. T. P. Singer and R. N. Ondarza (ed.), Elsevier, New
York, 1978, pp. 3-16.
154. B. A. Chabner, Pharmacological Principles of Cancer Treatment. B. A.
Chabner (ed.), W. B. Saunders Co., Philadelphia, 1982, pp. 183-212.
155. A. F. Cook, M. J. Holman, M. J. Kramer, and P. W. Trown, J. Med Chem.
22, 1330-1335,1979.
156. A. Williams and E. Klein, Cancer 25.454-456,197 .

139
157. S. Ozaki, Y. Ike, H. Mizuno, K. Ishikawa, and H. Mori, Bull. Chem. Soc,
lapaniQ, 2406-2412, 1977.
158. T. Kobari, K. Tan, M. Kumakura, S. Watanabe, I. Shirakawa, H.
Kobayashi, A. Ujiie, Y. Miyama, H. Namekawa, and H. Yamamoto,
Xenobiotica 8. 547-556, 1978.
159. S. Ozaki, Y. Watanabe, T. Nagase, Y. Ike, and H. Mori, Chem. Pharm.
Bull. 34. 893-896, 1986.
160. A. Buur and H. Bundgaard, Int. J. Pharm. 23. 209-222, 1985.
161. H. Sasaki, T. Takahashi, Y. Mori, J. Nakamura, and J. Shibasaki, Int. J.
Pharm. 60. 1-9, 1990.
162. S. Wolf and A. Murray, Composition and Function of Cell Membranes.
Adv. Exp. Med. Biol. Vol. 140, Plenum Press, New York, 1981, pp. 1-
53.
163. A. G. Lee, Comprehensive Medicinal Chemistry. Membranes and
Receptors, Vol. 3, J. C. Emmett (ed.), Pergamon Press, New York, 1990,
pp. 1-43.
164. R. Guy and R. Fleming, J. Coll. Interface Sci. 83. 130-137, 1981.
165. C. Pidgeon and U. Venkataram, Anal. Biochem. 176. 36-47, 1989.
166. R. Markovich, J. Stevens, and C. Pidgeon, Anal. Biochem. 182.1989.
167. J. Houk and R. H. Guy, Chem. Rev. 88.455-471,1988.
168. F. L. Vaughn, R. H. Gray, and I. A. Bernstein, In Vitro 22. 141-149, 1986.
169. Z. Wojciechowski, L. K. Pershing, S. Huether, L. Leonard, S. A.
Burton, W. I. Higuchi, and G. G. Krueger, J. Invest. Derm. 88. 439-446,
1987.
170. J. Hadgraft and G. Ridout, Int. J. Pharm. 39. 149-156,1987.
171. J. Hadgraft and G. Ridout. Int. J. Pharm. 42. 97-104, 1988.
172. K. Miyake, F. Kitaura, N. Mizuno, and H. Terada, J. Chrom. 389. 47-56,
1987.

140
173. F. A. P. C. Gobas, J. M. Lahittete, G. Garofalo, W. Y. Shiu, and D.
Mackay, J. Pharm. Sci. 77. 265-272, 1988.
174. A. M. S. Ahmed, F. H. Farah, and I. W. Kellaway, Pharm. Res. 1. 119-
124, 1985.
175. H. Rinderknecht and M. Gutenstein, Org. Svn. 5. 822-824, 1973.
176. M. Okumura, K. Sugibayashi, K. Ogawa, and Y. Morimoto, Chem.
Pharm. Bull. 37 (5k 1404-1406, 1989.
177. A. Hoelgaard and B. Mollgaard, Advances in Drue Delivery Systems.
Controlled Release Series 1, J. M. Anderson and S. W.
Elsevier, Amsterdam, 1986, pp. 111-120.
178. J. Kao, F. Patterson, and J. Hall, Toxicol. Appl. Pharmacol. 81.502-516,
1985.
179. J. Kao and J. Hall, J. Pharmacol. Exp. Ther. 241. 482-487, 1987.
180. N. Bodor and J. Kaminski, J. Mol. Struct. (TheochemT 163. 315-330,
1988.
181. R. C. Wester and H. I. Maibach, Topical Drue Delivery
Formulations. Drugs and the Pharmaceutical Sciences, Vol. 42, D. W.
Osborne and A. H. Amann (eds.), Marcel Dekker, New York, 1990, pp. 213-
220.
182. K. M. Halprin and A. Ohkawara, J. Invest. Derm. 46. 278-282,1966.
183. K. Adachi and S. Yamasawa, J. Invest. Derm. 46. 473-476,1966.
184. O. H. Lowry and J. V. Passonneau, J. Biol. Chem, 239. 31-42, 1964.
185. U. Eisner and J. Juthan, Chem. Rev. 72. 1-42,1972.
186. R. E. Lyle and P. S. Anderson, Adv. Heterocvcl. Chem. 6. 45-93,1966.
187. P. J. Chikhale, F. M. Alvarez, C. Pidgeon, L. Walton, and J. Kuester,
Pharm. Res. 6 fSuppl.l. S-147, 1989.
188. F. M. Alvarez, P.J. Chikhale, L. Walton, and C. Pidgeon, Pharm. Res. 7
iSuppl.l.S-3. 1990.

BIOGRAPHICAL SKETCH
Prashant J. Chikhale was bom in the small town of Shivrajpur, in the State
of Gujarat, India, on December 19, 1962. He attended primary and high school in
Bombay from 1968-78. He completed a year of junior college in the Parle College,
Bombay from 1978-79 and another year of senior college in Ahmedabad, Gujarat
from 1979-80. He then went on to join the Maharaja Sayajirao University of
Baroda, Gujarat, to pursue the degree of Bachelor of Pharmacy, from 1980-84. In
fall 1984, he joined the Graduate Program of the Arnold & Marie Schwartz
College of Pharmacy and Health Sciences in the Department of Pharmaceutics,
from where he received his degree of Master of Science in Pharmacy. In fall 1986,
Prashant J. Chikhale joined the Ph.D. program of the College of Pharmacy,
University of Florida, in the Department of Medicinal Chemistry.
Prashant J. Chikhale is a member of the American Pharmaceutical
Association (APhA) and the American Association of Pharmaceutical Scientists
(AAPS). He has accepted a post-doctoral research position in the Department of
Pharmaceutical Chemistry at The University of Kansas in Lawrence, Kansas, from
February 1991.
141

I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of Doctor
of Philosophy.
Nicholas S. Bodor, Chairman
Graduate Research Professor of Medicinal Chemistry
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of Doctor
of Philosophy.
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of Doctor
of Philosophy.
Richard H. Hammer
Professor of Medicinal Chemistry

I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of Doctor
of Philosophy.
'! ( p
Hans Schreier
Assistant Professor of Pharmaceutical Sciences
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of Doctor
of Philosophy.
Associate Professor of Pathology and Laboratory Medicine
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of Doctor
of Philosophy.
H i
/ 1/
\0^jh
11
Francisco ll
VI. AlvWez
Manager of Analytical Research, Schering-Plough

This dissertation was submitted to the Graduate Faculty of the
College of Pharmacy and to the Graduate School and was accepted as
partial fulfillment of the requirements for the degree of Doctor of
Philosophy.
May 1991
Dean, Graduate School