Professional Documents
Culture Documents
Edited by
James Swrrbrt~k
&hool af Ph#rmacy
University of North G"aro/tna
Chap& Hfll, North Carolina
edited by
David W. Osbome
Anton H. Amann
The Upjohn Company
Kalamazoo, Michigan
informa
healthcare
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
This book contains information obtained from authentic and highly regarded sources. While all reasonable
efforts have been made to publish reliable data and information, neither the author[s] nor the publisher can
accept any legal responsibility or liability for any errors or omissions that may be made. The publishers wish to
make clear that any views or opinions expressed in this book by individual editors, authors or contributors are
personal to them and do not necessarily reflect the views/opinions of the publishers. The information or guid-
ance contained in this book is intended for use by medical, scientific or health-care professionals and is provided
strictly as a supplement to the medical or other professional’s own judgement, their knowledge of the patient’s
medical history, relevant manufacturer’s instructions and the appropriate best practice guidelines. Because of
the rapid advances in medical science, any information or advice on dosages, procedures or diagnoses should be
independently verified. The reader is strongly urged to consult the drug companies’ printed instructions, and
their websites, before administering any of the drugs recommended in this book. This book does not indicate
whether a particular treatment is appropriate or suitable for a particular individual. Ultimately it is the sole
responsibility of the medical professional to make his or her own professional judgements, so as to advise and
treat patients appropriately. The authors and publishers have also attempted to trace the copyright holders of all
material reproduced in this publication and apologize to copyright holders if permission to publish in this form
has not been obtained. If any copyright material has not been acknowledged please write and let us know so we
may rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmit-
ted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented,
including photocopying, microfilming, and recording, or in any information storage or retrieval system, without
written permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.copyright.com
(http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Dan-
vers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a
variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of
payment has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only
for identification and explanation without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
David WbV,Osborne
Anton iJ. Amann
Preface
Contributors
Index
Contributors
.
William .J Adbieks* University of Michigan College of Pharmacy,
Ann. Arbor, Michigan
Eugene R . Cooper Sterling Drug, Inc. , Malverne, Pennsylvania
Rane t. Curl University of Michigan College of En@neering, Ann
Arbor, Michipn
Gary R , Dukes The Upjohn Company, KaXamazoo, Michigan
Peter M , EZias Veterans Administration Medical Center, University
o f California School of Medicine, San Francisco, CaHfornia
Gardon L. FIy nn Cygnus Research Corpomtion, Redwood City,
California
Stig E. Friberg Clarkson University, Potsdam, New Vark
Dougfas A . Hatzanbuhler The Upjohn Company, Kalamazoo, Michigan
Martin Katr Advanced Polymer Systems, Inc ., Redwood City,
California
lbrahim H , KayatI Clarkson University, Patsdam, New York
David H . Lynch Immunex Corporatian , Seattle, Washington
Howard t , Maibath University of Galgomia Sckroal of MeCLicine , San
Francisco, California
I INTRODUCTION
skin. Most drugs are erystdline, and the mere rubbing of sm&X
erystds on the surface of skin. will result in very p m r transfer to
the skin, Delivery from volatile solvents, such as ethanol, and ace-
tone, wnl trmsfes small portions of drugs into the outer layers of
the skin, and the flux will quickly drop off. In fact, m y transpart:
study conducted with such, solvents, will more than likely measure
the itissolution rate of the drug, rather than the transport proper-
ties of skin,
B , Thermodynamic Factors
The drt.idng force for transport is the chemical ptentiitl padient,
and within a single phase, this reduces to the concentration gradi-
ent. To create this gradient within the skin, one normdfy dissolves
the drug in a solvent or vehicle and establishes a cer-tdn concentra-
tion of drug in the outer surface of the skin, Different vehicles
can provide different concentrations of d m g at this interface and,
thus, the d r i ~ n gforce and, hence, the flux will be a function of
the vehicle. In generd, one does not know the concentration of
d m g in the outer surface of the skin, but there is a simple tktermo-
dynamh relationship to compare the concentration of drug in the
skin for different vehicles.
The chemical poten tiat or activity is the continuous vaAable
across interfaces f 1) and, thus, one has the fiirlowing equation at
the skin -vehicle boundary :
Thus, E q . 3 becomes
375
150
1
P
125
Y
.u
-5i
100
0% Salicylic Acid in
4 75 Propylene Gfycal
C ( x , O) = O, C ( h , t ) = O f 81
expr-an2
J = -13
n
2 2 f loll
x = h n=l cos a, f B + B + R]
Practical Considerations far Use of Enhancers
where
a is a r m t of a tan a =
n
o i a 3 4 5 6 z 8 9 10
7
6. In Vivo Evaluation
The most quantitative in d v o evduations of toplcdly applied drugs
have been, done with skin grading, in which blanching of norm&
skin by steroids or the blanching of erythema by nonsteroidal anti-
inBammatow agents is measured. En these models, the onset of re-
Praetieat Considerations for Use o f Enhancers
sponse is related ts lag t i m e , and shorter lag t-imrss arc: ernr red as
good, For disease states, it 3s not clear whether a large, pulsed
deliveq is better, or a slow, prolonged delivew is preferred. AG-
curate knowledge of vehicle delivery rates correlated with clinicFz3
results could lead to new vehicles with very different delivery pro-
files. These systems, in turn, could be evduated clinically to de-
termine the best delivery rate, The methods and skill eer-t&nly ex-
ist to mswer same of these key questions, which undoubtedly will
be mswered as the dosage forms become more contmllable for de-
livery rate,
Topied treatment of skin cancer, wafls, batdness, and so on,
may be possible, not only because of new drugs, but &so because
better vehicles can be designed to enhance delfveq. The skin is
an accessible tissue of large surface area and should be a .rY.iable
port of entry for drugs, Vehicles with supesior delivery will un-
doubtedly play a substantid prole in increasing the usefulness of
this port of entry,
10 Cooper and Patel
Vehicle 1 Vehicle 2
Ingredient (wt%) (wt%l
Table 1.2 Desonide Solubility and Flux Across Hairless Mouse Skin
SoIubMty Flux
Vehicle (mglml) (mgfhr x 104)
REFERENCES
I. EVOLUTION OF C U R R E N T CONCEPTS OF
STRATUM GORHEUM STRUCTURE
It , EPIDERMAL DiFFERENTIATIOM
El . Lamellar Bodies
The lamelliar body, a 0 . 2 to 0 . 3 ~ n ndiameter, ovoid, secretory or-
gmelle, which is considered the centrd actor in the formation, of
the irttercellular "'mortar," is synthesized pdmady within the spi-
nous eel1 m d then displaced to the apex and periphery of the gran-
ular cell ( 2 4 ) . Xn response to an unknown signal, it fuses with the
plasma membrme , secreting its contents into the intercellular spaces,
thereby generating an expanded intercellular compartment that con-
stitutes from 1Q% to 40% of the total volume of this tissue (25).
Thus, secretion is one of two cellular events associated wBh lamel-
lar body ~txwytosis;fusion also occurs, and the ""splicingt' of addi-
tional organelle membrane into the plasma membrane may contribute
a large reservoir of surface area that could explain the stratum
comeurn" remarkable water- holding capacity ( 25) ,
Lamellar bodks appear to contain three types of matedds:
first, sugars, in the form of glycosphingolipids and possibly glyco-
proteins ; second, free sterols and phospholipids; m d tl.rird, a se-
lective array of hydrofgtic enzymes possibe charged with degrading
Impartance of Epidermal Lipids 17
Galaboiffi Enzymes:
Acid P8esphotase.
Dasquamatioa
Proteaser, t~pases, 2) Dsgradation of &oe.LLprd tnterccll~lar
Glycosldases Specks (A& Phos$hatase, Prateares)
PhosphoIipids 40 Trace
Sphingolipids 10 35
Cholesterol 15 20
Triglyce~ides 25 Trace
Fatty acids 5 25
Other -5 10
Totals : I00 100
@Glycosphingotipids
A Phospholipids
Il Cholesterol sulfate
c
Inner SC
SG Outer SC
LAYER
Figure 2.2 C h a n e s in lipid composition that occur during stratum
corneum trmsit. Note that smalX amounts of phospholipids (dia-
monds) , glyeolipids (circles) , and cholesterol sulfate (squares) re-
main in the inner stratum eorneum (SC) (stratum compactum),
whereas only cholesterol sulfate persists in the outer SC. SG , stra-
tum granulosum.
in composition presumably reflects ongoing metabolic activity, fur-
ther dispefling the notion of the stratum corneum as an inert tissue
( 20; see Table 1). Despite the paucity of phospholipids, glyco-
sphinogaGpids, and cholesterol sulfate, these constituents apparently
can arrange themselves into membrane bBayers, possibly by exploit-
ing the ampkipathie properties of sphhgolipids (25,29--313, which
apparently are capable of extensive hydrogen bonding. Yet, the
hydrophobic, long-chain bases (31,32) , the long-chain , fully satu-
rated fatty acids (28,29,31,31; Table 41, and the endchment of
sphingolipids in linoleic acid ( 3 1,33- 35) , all make spkingolipfds par-
ticularly good candidates for epidermal water-proofing (Table 5).
Hence, sphingolipids are bipolar, possessing both a relatively hy-
drophilic terminus, capable of intermolecular bonding that can form
ordered membrane stmctures, and highly hydrophobic moieties that
could. be highly water-repellent. Stilil unresolved is the fate of
these lamellations during the first stages of shedding: X s membrane
bilayer break-up a prerequisite for shedding? Do further changes
in composition or the physicochemical properties of these bilayers
lead to desquamation?
14: 0 3.8
Trace
Trace
TIME (hr)
Figure 2.3 The time course of epidermd lipid biosynthesis and bar-
rier function after treatment with acetone. Note that transepidermal
water loss (TEWL) , totd nonsaponifiable lipid biospthesfs (TNS) ,
and cholesterol (C) synthesis exhibit a parallel return toward normal
over 24 h r (from Ref. 39, reprinted with permission).
Site
Lipid weilfht ($) 6.5 2 0.5 4.3 1 0.8 7.2 10.4 2.0 F 0.6
Polar lipids 4.9 1 1.6 5.2 11.1 3.3 t 0.3 3.2 1 0.89
Neutral lipids 77.7 25.6 65.7 1 1.8 66.4 1 1.4 60.4 t0.9
ACKNOWLEDGMENTS
This work was supported by NXH grant AM 19098 and the Medical
Research Serdce, Veterans Administration, We appreciate the typ-
ing assistance of M r . Bil Chapman and Ms. Sally Michael,
REFERENCES
1. INTRODUCTION
The outermast layer of the skin, the stratum corneurn, has an es-
sential role as a harder against the transport of water and of chem-
ic& and biolo@cd agents ( 1,2). This barder is functionally the
sane, for short pesads, in bath lidng and dead s k h ( 3 1 , vvith 1i-
pids behg responsible for the essentid function (4,5). The bio-
logiicd and medied importance of this cannot be overstated, and the
investigations into the i n d i ~ d u a lstructure and over& organization
of the straturn earneum lipids has been extensive (6-12).
These investigations and the scientific examinations remain an
ongoing process (131, but recent investigr;ltions d3ow a limited dis-
cussion of the relationship between structure and transport of water
through the stratum cornearn. This chapter wiU describe the lay-
ered stmcture and redew its importance far tr~nsdermal,transport.
Fatty Acid
' - - n
meet. White (16) has recently made a very extensive x-ray inves-
tigation into the stratum corneum lipid strmcture, finding a layered
structure only at high temperatures. At physiolo@cal temperatures
solid phases were also present and the structure was complex.
Our discussion will be limited to the features of the layered
structure, which are more complex than indicated by Figure 3 . 1 .
Low-angle x-ray diffractometry shows that in a model such as Elias"
(13) not all of the lipids are positioned with their polar gmups lo-
calized in the polar layer (see Fig. 3 . 1 ) . Some of them are actually
located in the region between the methyl groups, In addition, the
low-angle x-ray diffraction patterns provide information with which
to estimate penetration by water into the space between the lipid
molecules. Order parameters of lipid groups by nuclear magnetic
resonance (PJMR) (17) are used to confirm the x-ray results and to
provide information on the mobility of groups within the lipid mole-
cules.
The results of low-angle x-ray diffraction and NMR, as well as
their interpretation, are comparatively new to the dermatological
constituency; therefore, the methods used and the interpretation of
results will be briefly discussed.
Purity Wt%in
Component Source ( %) mixture
Electron poor
Figure 3.2 X-rays are reflected off of the parallel planes of the
layered structure. These reflections may then be converted to the
interlayer spacing, d , by using Braggts equation.
Added Substance
Table 3.2 The Amount of Water Transferred per Unit Time, Ex-
pressed as mg/(cm2/hr) per Unit Membrane Thickness
In each of the above cases, the flux was determined for water pene-
trating through at least two tablets. The values shown are the mean
2 standard d e ~ a t i o n .
0
CH3\ II CH,\
N-CH S=O
CH,/ CH<
(CH2I,, CH,
44 Friberg et al.
111. SUMMARY
REFERENCES
surfactant and lipid systems. The basis of the power of these tech-
niques is that only the nuclei, which are at resonance, will be di-
rectly observed, This allows observation of selected parts snd fac-
ets of the system, u s u d y by making use of the constituent nudei
of the components (i,e., no perturbation of the system 'by bcorpo-
r&ed foreign ""pmbet' molecules). Substitution of pmtons by deu-
te~iumatoms elither in the aqueous fraction or h the lipid component
has relatively mhas and predictable effects on the system" behaaor
and has been used to great advantage (5-18) in the investigation
of baayer phases formed by lipids and surfactants in the presence
of water.
m i k e the more commonly eneountemd hydrogen (&) and car-
bon ( 1 3 6 ) nuclei, the deutedum ( 2 ~ nucleus
) has a spin quantum
number, I = 1, wit k an associated quadrupoxar moment. This means
that there are three nondegenerate nuclear spin enerm levels (Fig.
2 ) - In an isotropic endronment in which the effects of electric field
gradients at the nudeus are averaged to zero, a shgle msonance is
obt?e~pred. The presence of electdcd field gradients h an axiaIly
symmetdc field, however, leads to changes in the spin e n e r w levels
(see Fig. 21, which leaves the mi = 0 --+ +I and m = -1 -+ 0 transi-
tions with different energies, This has the effect of producing two
&sorption p e a s at resonance separated by a frequency- difference
d v , the size o f which depends on the size of the hteraction with
the electricd field gradient as
ZEEMAN QUADRUPOLAR
Figure 4.2 Nuclear spin energy levels for a spin = 1 nucleus (e.g.,
2 ~ showing
) the allowed transitions (a) for an isotropic and (b) an
axially symmetric electrical field gradient,
Here the bar in Eq. 2b represents the averaging over all possi-
ble orientations of the randomly dispersed lamellar domains. Many
studies have now been carried out to derive order profiles in Iipid
and surfactant systems in lyotropic lamellar phases (5-18). The
form of the profile (Fig. 3) is remarkably consistent for a wide va-
riety of systems showing highest order (SCD values) in the seg-
ments of the amphiphilie chains cIasest to the head group. A rapid
decrease in order, as expressed by Scf), is seen in the methylene
segments near the terminal methyl group, which is located at the
center of the bilayer. The shape of this type of profile is a result
of the packing constraints in the associated bilayer structure, re-
ducing the canformational freedom of the individual chains ( 11,18),
Penetration Enhancer Incorporation in Bilayers
n, X = 2d sin 8 133
tributions from the bilttyer (dbn) and the solvent layer (ds) as
Av = f Av + fiSOAviSO
obs b b
B. Surfactant-Water-Enhancer
The pa&iaf phase diagrams of the host surfactant-water system
containing laurocapxram (Azone) , oleyl alcohol, and propylene glycol,
respectively (Fig. 53, show that the amaunt of pmpylene glycol
that can be accommodated is much smaIter tkm the other two oils,
It can d s o be seen that the heorparatioion of either oleyl alcohol or
Azoxle hcreases the ability of the lamellar phase to aceommodate
water, with the maimurn water content rising from about 50Un the
host system to about 85%in the presence of the enhancer. This is
not observed for soltzb.ilized prepylene glycol for which, in fact, the
water cizpacity of the bilayer phase Is reduced on ineavoratiaion,
Such an enhancement in the water retaining capadty of a bntsyes
system is of Importance in the consideration of penetration enhance-
ment, because it has been shown that the water content of the skin
affects delivery mtes in several cases f 26-28).
Some bformation about the state af surfactant hydration in the
presence of these agents may 'be derived. from the variation of the
quadrupdar splittings of the surfactant - I%HIW $3 system. Studies
made of samples with a f k e d surfstctantlwater mole ratio in the
WATER
fa)
01L
OLEY L ALCOHOL \ \
WATER 632E05
(b)
Figure 4.5 (a) Phase diagram at 298 K for the system C 121E05-
Gzone-water, (b) Partid phase diagrams showing the extent of the
lamellar phases o f the system C laEOg-water with oleyl alcohol and
propylene gXycaX at 298 K .
Penetration Enhancer Incorporation in Bilayers
0 0.5 1.0
f
(a) Moles OiIlMoie Surfactant
MOLES PG
MOLE c 1 2 E 0 5
0 20 40 60 80 300
W A T E R CONTENT f w t . %)
interface with the aqueous compartment, This fact coupled with the
increased structure arising from a greater number of trans config-
urations in the hydrophobic region-a necessary result of packing
of alkyl chains in this fashion---&a lead to more order in the region
of the polyoxyethylene chains to which the water molecules are
bound. The consequence of this is to increase Sb and Avb.
Small-angle x-ray diffraction measurements yielding values of
the characteristic d-spacing (Fig, 8) also indicate structural
changes in the system as the oil content is varied at fixed mole ra-
.
tios of Azonelwater The expected linear increase in d-spacing is
found with increasing water content (Fig. $), dthough the slope of
the plot decreases with an increasing Azonefsurfactant ratio. If the
surfactantlwater ratio is kept constant, an initial increase in the d-
Penetration Enhancer Incorporation in Bilayers
A Z O ~ ~ / C , , E ORatio
~
Second, the model aIso shows synergistic aspects in both the water
and solubBzate retention of the system, both features being charac-
teristic of the r e d system* Extension of the model to include Salty
acid-soap components is being made t s heorporate other- aspects of
skin p H and, composition.
ACKNOWLEDGMENT
This work has been part funded by EOLAS, The Upjohn Company,
and thanks are due to P~ofessolsStig E. Friberg for making avaga-
ble time on his smdl-angle x-ray apparatus,
REFERENCES
I. INTRODUCTION
As both the largest and most visible organ of the human body, the
skin is of unequaled importance in portraying an individudts state
of being. Therefore, the biology and chemistry of the skin and its
appendages is important for both the cosmetic and pharmaceutical
industries. The heterogeneous nature of the skin provldes one of
the diffimlties in studying this tissue, Just as the skin itself com-
prises morpholoQEically different layers, the glands, follicles, and
micmvasculatu~teof the skin provide zones with different charaeter-
istic pmperties. One of these properties is that a thin, noncontin-
uous film of lipid is deposited on the skin surface from the seba-
ceous glands. Although often neglected in the design of a topical
formulation, the presence o f skin surface lipids can significanay in-
fluence the delivery of topical drugs. This infiuence may be either
beneficid to, or detriment& in, obtaining the desired drug delivery
characteristics and, thus, should be conddered , especially for top-
icals applied to sebum-rich areas of the body (e .g. , forehead,
chwks, and scalp). It is impo&-tant to realize that the interaction
between a drug and the skin surface lipid is not only dependent up-
on the physicochemical properties of the drug, but also upon the
physicmhemical nature of the vehicle. Topical agents can be formu-
Iated that will target sebum-rich zones of the skin (e,g., the hair
follicle) or, attematively, a topical agent can be formulated to mini-
mize deIlvery through the hair follicle. Backgmund information on
skin surface lipids and a brief description of the approach neces-
sary to formulate sebum-sdective vehicles will be the focus of this
70 Osborne and Watzenbuhler
The skin is traditionally divlded into three major re@ons: the stra-
tum corneum, the able epidermis, and the dermis, The outermost
of these layers, the stratum corneum, proddes a barfier against
the permeation of most substanees. This layer is composed af non-
viable, keratin-faed cells (squames) that are roughly pentagonal
plates 0.5 vm thick and. 30 to 40 urn across ( 2 ) . Faling the inter-
ceBuIar space between these ceUs are bsayer-stmctured lipids ( 2 ) .
The stmeture of these lipids has proved to be important to the
moisture-retaining absifity of the stratum comeum (3,4) . The viable
epidermis Ees bdow the stratum comeum and consists of stratified
keratinizing epitheliaf cells whose fin& function is to produce the
stratum comeum, This layer does not contdn blood vessels, rely-
ing on noulrlshrnent by cell Buid fmm the deeper dermis layer. The
deepest layer of skin is the dermis, which conslsts of dense, irreg-
ularly arrmged connective tissue, and it is nourished directly by
blood vessels (5). Combined, these layers form the skin and are
connected to the subcutaneous tissue by bundles of collagen fibers.
Embedded in the skin are ecesine sweat glands, apocrine glands,
h d r follicles, mef sebaceous glands, Ecerfne sweat glands are sim-
ple tubular glands distlrlbuted over dmost all of the human body,
Each gland has a secretory part located below the dermis in the
subcutaneous tissue and an excretory duct that ultimatc3liy opens di-
rectly on the skin surface. These glands produce perspiration and
are particularly numerous on the p&m of the hand. Apocrjlne
glands produce eharaeteristic body odors and are primarily located
in the axarz. The secretion of this gland contains chdesterd, ster-
oids, and proteinaceous substances ( 6) and contributes substantially
to the surface lipid on the a i E a , thus, b e h g primarily im.po&ant
to deodorant and antiperspirant hrmulations. When present, tho
apocrhe gland empties into the hElris follicle above the sebaceous
gland.
Two major elassidcations of h d r exist on humans: terminal h&rs
and vellus hizirs. Termin& hairs are the courser h d r s of the scalp
and male tmnk hair, and the root of termhal hairs may extend
more than 3 mrn belaw the skin surface linto the subcutaneous fatty
tissue, Vellus hair is the fi-ne, often unnoticed, body hair that
popujates re@ons such as the forehead, and extends less than 1 mm
Skin Surface L i p i d s and Topical Formulations 71
into the dermis. The hair fiallicle comprises the ha$r, combined with
its surrounding root sheath, The sebaceous gland empties directly
into the upper third of the hair fcirliele, and the cambination of se-
baceous @and and follicle is oAen referred to as the piEosebaeeaus
unit. The sebaceous gland is the major source of skin surface lipid
in the adult. A diagram representing both the skin layers and skin
appendages is shown in Pimre 1.
Because the sebaemus @and is the primary srsctrce of the skin
surface lipids covering large portions of the anatomy, understanding
the properties of this skin appendage becomes important, The se-
baceous glanc;2 is well defined in the embqo, remains relatively smdl
Composition
(wt%) Forehead Cheek Chest I3 ack Side Arm Leg
Free fatty acids 24.1 22.4 28.1 24.9 23.4 36.4 37.8
Wax esters 27.0 24.3 21.7 21.9 17.7 15.8 12.9
Squalene 12.3 13.5 10.7 10.7 8.5 6.9 6.2
Cholesterol 1.2 2.0 1.6 2.1 3.9 4.1 9.4
Cholesterol esters 2.9 3.9 2.9 3.0 5.3 4.4 7.5 0
G;
Total lipidb 160 104 59 38 29 30 19 tr
D
(glcm2) 3
(D
aThe data are averages for three subjects, each examined on two occasions. 9a
b ~ r a p h i cextrapolation indicated that the epidermal lipid contributed an average of 4 g/ern2.
Sourn: From Ref. 11. 2
.v
h)
Skin Surface Lipids and Topical Formulations 75
Odd-carbon-number
chains 22.5 26.2 21.8
Even-carbon-number
chains 77.4 73.8 78.1
Saturated chains 70.9 56.2 69.7
Unsaturated chains 29.9 43.8 30.2
Component Wt%
n-alkanes 6.1
skin surface lipid, although for a given individual, the fatty acid
component of the skin surface lipid remdns relaGvely constant with
time, Thus, each indiddud shows an inverse relationship between
the amount of free fatty acids and the triglycerides charactefistic of
their skin surf ace lipids,
A hydrolysis reaction is the source of one of the important
structurd components of the epidermaf lipids. The enzyme spkingo-
myelinase hydrolyzes sphhgomyelin to ceramide and phosg horylcho-
line 1(17). This enzyme degrades dl cellular gfiosphoBpids before
the cells move from the granular layer of the epidermis into the
stratum cornearn. The ceramides then become one of the mdn corn-
ponents of the layered interceaular lipids found between the termi-
n a y diffemntiated cells of stratum corneum. The other hydmlysis
product, phosphorylcholine , does not fiB the stratum corneurn spaces
but, rather, the phosphorus is thought to be reabsorbed and re-
utaized by the epidermis. This is an explanation for why the skin
surf ace lipids are void of phospholipids ( 18).
]in addition to the substances Ested in Table I, smdl quantities
af other compounds have been found in sebum. Detectable amounts
of mdragens are present in skin surface lipids ( 19) ; however, the
amounts of these steroids are considrtm*ed too smd1 to claim that the
s ~ ~ ~ C ~ glands
O U S play a notable role in the totd body excretion of
androgens, Parafsnic hydrocwbons , such as pdstane , &so have
been shown to be of sebaceous o ~ g i n(21. Their presence is spec-
ulated La be linked to diet.
The influence of diet on skin lipids is another factor that adds
to the chemicd eomple~tyof skin surface lipids. Fasting causes a
50% reduction in the synthesis of fatty acids, w a x esters, ~ n tA- d
glycerides , although not influendng squalene production ( 19) .
L&ewise, the absence o f dietary essential fatty a d d s (linolleic acid)
results in a stratum corneurn with a changed appearance and a sig-
nificantly reduced bar&er function ( 2 ) . Nicolajdes (21) showed that
feeding 1 - [ 1 4 ~ ] octadecane to the rat led to the prtjsence of the
labeled substance in the skin. Although a complete understanding
af skin metabolism is not yet avdable, it is important to remember
that some of the components found in the skin surface lipids are
present simply because the skin was the most efficient route of
""wasteffreclnnova.
When examining the components found h skin surface lipids,
the possibaity that the trace substmces are from exogenous sources
must be considered, The obdous sources are cosmetics, topicd
.
phtarmacmtic&s, toaet Aes , and environment& pollutants AltE-rough
most studies that use skin surface lipids collected from human da-
nors try ta elimbate these eontaminants, the reservoir function of
the stratum cornam m&es it part;icularly difacult to remove totdly
some substances from the skin surface. Beeauae chemicals can be
Skin Surface Lipids and Topical Formulations
One appmach to determining the effect that skin surface lipids have
an topical formulations is to model either the prope&ies or composi-
tion of skin surface lipids by using readily available mater;irals, This
avoids the laborious task af pmling lipids from volunteers and pro-
d d e s the investigator with convenient quantities of lipid mate~a2.
Skin surface lipid models are of interest to not only cosmetic ~ n d
pharmaceutical hdustries but, dso, to the detergent industry, far
which they are used to compare the abnities of laundry formulations
to cXem sosed fabrics. Because of this interest from various fac-
tions of the industsid research community, a number of model skin
surface lipids are cited in the literature, However, most of these
modcls were designed for specific studies, thus, limiting their prac-
ticality in evdurr-ting cosmetic and pharmaceutical topical agents,
The cornpasitions of some of these model lipids will be @ven, and
the advantages and disadvmtages o f each model will be discussed,
in this section,
The F~berg-Osbome model skin surface lipid (34) given fn Ta-
ble 4 is based on the chemical composition elucidation studies that
have been carried out by various investigators, In this model, each
of the major chemical fwtions of human S ~ ~ Uare M represented by
a purified commerciaUg available chemicaf, For instance, the tri-
glycerides are represented by triole.in, whereas the wax: esters are
represented by oXeic acid pdmitic ester. This modrtl has two major
advantages: it chemicafly mimics naturd sebum, and it is a single-
phase liquid that is easy to work with at ambient temperatures.
This second advantage is also the modelrs strongest disadvantage,
because the physical properties of sebum (melting point &out 3S06)
are not obtained. This model was initially used to describe the
phase behador of a simplified cosmetic system when mixed with the
model lipid, l t is in this capacity that a single-phase Xiquid model
becomes essential for practical studies.
Another model, which is based on the chemical, components
found in natural s k h surface lipids, is given in Table 5, Agdn,
the major advantage to this model is that it chemically. mimics the
skin lipids. At room temperature, this mixture is EI waxy solid that
Skin Surface Lipids and Topicat Formulations
Component %
Squalene
Lecithin
Component
Palmitic acid
Myristie acid
Oleic acid
Tripalmitin
Triolein
Palmitic acid palmitic ester
OIeic acid palmitic ester
Cholesterol
Cholesteryl palmitate
Cholesteryl oleate
80 Osborne and Hatzenbuhler
Component %
Olive oil
Coconut oil
Pdmitic acid
Stearic acid
OleSc acid
Paraffin wax
Squdene
Spermaceti
Cholesterol
Total
Component %
Component %
Squalene
Tristearin
Triolein
Stearic acid
Oleic acSd
Cholesterol
Octadecanol
produces both the proteins and lipids that form the skin" barrier.
One aspect of this network, the skin surface lipids, can influence
the topied products that are *plied to the skin, This infuence
may be either beneficid or detrimental, depending upon the forma-
fator" awareness of the physicwfiemicd p m p e ~ i e sof skin surface
lipids and the desired drug delivery characteristics. By matching
the po1ar;ity of the formulation to the poladty of the skin surface
lipids, the hair follicle can Be targeted for d m g delivery, even if
the a m g is inherently too polar to be readay soluble in sebum or
skim surface lipid. The reader is further encouraged to consult
the review articles by Strauss et d. (7) and. WhernIlg (44) for more
detajlled consideration of skin surface lipids
REFERENCES
DAVID W . LYNCH
fmmurrex Corporation, Seattle, Washington
LEE K . ROBERTS University o f Utah School of Medleint;, Salt Lake
City, Utah
The skin, by. weight, is the largest organ of the body, Human
skin serves to provide several important functions. These include
maintaining physic& proteetion (barr-ictr function) against extern&
agents and dessication; r e c e i ~ n gsensory stimuli from the endron-
ment; regulating body temperature and water balance; excreting a
.
variety of substances ; participating in metabolic pathways ( e g. ,
initiation of vitamin ZD synmhesis and subcutaneous fat metabolism) ;
and serving as a earnpartmentalized component of the immune system
to provide protection agdnst certain pathogens, toxins, and neo-
plasia, To protect the host from foreign mate~alsand organisms,
an extremely complex relationship has evolved between the skin an&
the immune system, Many o f the cell types in skin synthesize a
variety af bioactive compounds, many of which have profound ef-
fects, not only on lac& in'tlarnmcltory responses and skin-associated
immune responses but, &so, on systemic immune responses,
The development sf ~frate@esto transdermdly deliver phama-
cologically active compounds is an. exciting new area of research
with far-reaching implications. However, transdermal delivery of
phasmacolo@cdly. active compounds also presents a number of chd-
lenging pmblems because the skin is normillly fairly imperdous to
even relative1y small molecules , T hus , from the pharmacolo@c point
of view, one of the mdor requirements for effective transdermd
dalivem is to accentuate the rate of movement of compounds acm8s
the many different layers of cells and connective tissue that collec-
tively form skin, However, transdermd dellvery of a nuniber aE
88 Lynch and Roberts
Like other organs, the skin is well organized into histolo@cally de-
fined tissues to faeBft(ste the performance af its vafious functions.
The stmctrxral and functiond relationships of the skfn for its im-
munolo@caj. role are discussed in the following:
The skin is di*ded into two mdn layers, the surface epitheli-
urn, termed epidermis, m d the underlying connective tissue layer,
termed the dermis, Beneath the dermis is the hygodemis, which
3s ai layer of loose connective tissue consisting mainly of subcutane-
ous adipose tissuee The hypodermis is lmsely connected to under-
l y h g deep fascia, aponeu~osis,or perhsteum. A schematic repre-
sentation of the skin is presented in Figure 1.
The epidermis is a keratinizing , stratified squamaus epithelium
composed of several distinct eel1 populations. The keratinocytes
represent mare than 90% of that cells d t h i n the epidermis, These
are the keratinizing squamous cells of the self-renewing epithelium
ImmunoloSy of the Skin
7-LA&GERWANS CELL.
STRATUM CORNEUM
INFLAMMATORY CELLS
ADfPOSE TISSUE
ries: B cells, which can differentiate into plasma cells that produce
antigen-specific antibodies; and 2' cells, which are further subdi-
vided into a number of subsets based upon their function, These
include cytotoxic I-' lymphocytes (CTL) , delayed-type hypersensitiv-
ity eells ( l i " r ) ~ ~T-helper
), fTh) cells that aid in the generation of
either cell-mediated immune responses or antibody-mediated immune
responses, and suppressor I\ (T8) eells that act as regulator37 cells
for both cell-mediated and antibody-mediated immune responses.
Functiondly associated with the lymphocytes are antigen-presenting
cells (APC), such as macrophages and dendritic cells, which act to
f'processft and "presentt' antigens to the various subpopulations of
B and T cells, The APG are required for the induction, generation,
and revlation of immune responses. Antigens are "presentedf7to
antigen-spedfic T lymphocytes in the context of major histacornpati-
bnitly complex-encoded eeU-surface molecules that are expressed by
APC (1). The induction and elicitation phases of the immune re-
sponse that are in;i"tiated by the APC causes both the activation and
the clonal expansion of the antigen-specific T cells.
Mature T cells are continudly trafficking between the circula-
tion and various different secondav lymphoid organs such as the
spleen, peripheral lymph nodes, mesenterie lymph nodes, and Pey-
erts patches ( 2 ) . The eontinud surveBlance of different organ sys-
tems in the body by- these recirculating lymphocytes thereby pro-
e d e s an enhanced protection against not only pathogens and toxins,
but also aga;inst neoplszstically transformed somatic cells, This con-
tinud immunosurveillanee is of pairticular importance for those organ
systems, such as the skin and the gastrointestinal tract, that me
in constant. contact with the numerous harmful agents encountered
in the external environment.
The immunosurveBlance capabnities of the immune system are
further enhanced by the compartment&ization of subsets of li" lym-
.
phacytes into defined cf rculatory circuits For example, recent
studies have demonstrated the prefemntial homing of a subset o f T
cells to the Peyerts patches, mesenteAc Igmph nodes, and lamina
propria of the htestine, a feat mediated by a distinct lymphocyte-
binding molecule expressed only within the microvasculature of these
tissues ( 3 ) . These data indicate the existence of a population of
lymphoid cells that &re restricted to gut -associated lymphoid tissue.
Although much work rem&ns to be done, data from a variety of
sources also hdicate the existence of an imrnunoloajical circuit, re-
stricted to the skin, that has been termed skin-associated lymphoid
tissue (SALT). Compartmentalized irnrnunofo@cal circuits thereby
increase the pr0babBit-y of an interaction between an antigen-spscif-
ic I-' eel1 and an antigen-bearing APG by directing effector cells to
the anatomical sitesf s) of antigen deposition ( 4,5) .
Lynch and Raberls
We will now focus on those skin cell populations that either have
been shown to play, or are suspected of pladng, a rale in skin-as-
socialed immune responses. The expedmentd model of aller@c con-
tact dermatitis, referred to as contact bypersensitiety ( C W ) , v\ml
be discussed in detail because this type of skin reaction represents
a major obstacle faced. by the pharmaceutical industry in developing
successful trmsdermal drug delivery systems, The roles played by
individustl cell types in the afferent, or induction, phase (which in-
volves T-cell mtigen-priming by APC) and the efferent, or elicita-
tion, phase (which involves the activation and migration of effector
T cells to sites of antigen deposition) of an immune response will be
outlined. The kinetics and mediators ascribed to CH responses w2l
be compared with those involved in antigen nonspeeidc inflammatory
reactions.
.
nodes (- 9) Whlile in transit in the afferent lymphatics, the LC are
histolo@cillly identified as velled cells. Once in the drdning lymph
nodes, the LC initiate the afferent phases of a CH immune response
by presenting the cell surface-bound antigen to responsive clones
of T cdls (41). The antigen-stimulated T cells undierp clonal ex-
pansion and seed the peripherd lymph nodes with memory. cells that
are capable of mediating the lintense CH responses observed in al-
ler@c contact dermatitis after subsequent chaBenge with the immu-
nizing cant act sensitizer. Thus, under norm& conditions, any agent
that is introduced through the skin (assuming it has the pmperties
of a f o r e i p antigen) can induce or elicit an immune response through
the actions of the LC to stimulate an effector T-cell response,
Although immune responses (whether cell-mediated immunity or
antibody production) are generdy attdbrzted to the actidties of
the effector cells, it must be appreciated that the immune system is
highly rewfated. The induction af Ts-eel1 responses by most anti-
gens serves to regulate the intensity and duration of an immune re-
sponse. Xt has been shown that distinct populations of antigen-spe-
cific Ts cells are capable of inhibiting the afferent and efferent
phases of a CH response (17,48). Thus, the balance between the
actidties of the 1C"~~-ceXl and Ts-cell populations dic-t-inteswhether
.
or not an d e r @ c contact dermatitis reaction will ocmr It has
been suggested that a population of epidermtll APC that are distinct
"Erom LC may selectively elicit antigen-specific Ts-cell responses
( 152, Lariwise, a rsoluble antigen absorbed through the skin that
gains access into the blood may d s o selectively- stimulate a T"~-cell-
dominated immune response (50). This may account for the elicita-
tion of Ts-cell responses to the topied administration of high doses
of contact sensitizing agents C 51) . The immunolo@cal conditfons
and mechanisms responsible fbr the selective elicittation of Ts-eea-
medhted responses are outlined in detall in Section V X of this chap-
ter.
Under norm& conditions, the I;@ are the only cells within the epi-
dermis that express class X I cell surface determinants ( 4 1 ) . This
is consistent with their known ABC function, since class XI mole-
cules are important far antigen presentation and the ee'Ilular inter-
actions required for the induction of an immune response. Gu&aus-
ly, it has been shown that keratinocytes in the skin of patients
with certain skin diseases are induced to express NLA-DR antigens
(18,66-61)). Sim2ar results have been obt&ned expeAmenta2ly
where murine keratinocytcss can be induced to express Xa (Fig, 3 ) .
Interestingly, those conditions in which the ker~tinocylesare in-
duced to express class 11 antigens are genertblily as~ociatedwith the
infiltration of lymphoid eeUs into the skin. For example, y -inter-
feron (y-IFN) can induce keralinacytes to express class I1 antigens,
which suggests that the release of this lymphokine by activated I"
eells may be responsible for inducing expression of these eell-sur-
face molecules f '72). Efowever, experriments in mice suggest that
Immunology of the SkSn
control group in which the sensitization sites were not exdsed. Ex-
cision of the sensitization sites 24 or 48 h r after sensitization re-
sulted in the induction of CH responses in 148 and 61% of the ani-
m d s , respectively, Animals that were unresponsive to the contact
sensitizer a&er excision of the sensitization site were also refractory
to subsequent immunization. 0%major importance was the observa-
tion that most of the contact sensitizer left the site of application
through the venous outflow within minutes of sensitization. As little
as 5% was found remaining 24 h r after application. Morwver, intra-
venous administration of the contact sensitizer resulted in the gen-
.
eration of specific C W S unresponsiveness These data suggest that
epicutaneou s application of reactive chemicals to normal skin results
in the simultaneous induction of bath positive and negative immune
responses, and that antigen must persist at the sensitization site
for an apprapeate pedod to activate an effector response,
This scenario has received considerable support from the re-
sults of number of different investigators who have demonstrarled
that positive and negative immune responses can be induced either
simultaneously, or in isolation, depending upan the immunization
prot;oeol used, For example, pretreatment of animds with cyelopkos-
pharnide (to eliminate suppressor cell precursors) before sensitiza-
tion by conventional skin-painting techniques results in enhanced
GH responses ( 94) . Subcutaneous immunization with haptens or in-
travenously injected haptenated, unfraetionated epidermal cells has
&so induced both positive and negative immune responses (95-97).
This should not be interpreted to mean that both types of immune
responses are Q Z W Q ~ S induced simultaneously. For example, skin-
p&nting with threshold levels of a contact sensitizer leads to induc-
tion of C H responses in the apparent absence of Ts-cell responses
(98). Conversely, skin-painting with supraoptimal doses of hapten
preferentiaEy induce Ts-cell generation (5%). Xt remdns to 'be de-
termined which of these situations will be mimicked with each indi-
Actual antigen that is used in a transdermal. delivery system.
ACKNOWLEDGMENT
The authors would like to thank Janet Swapp, Carol Peck, ~ n d
Patecia Tibola far their excellent secr-etariaf assislmce in the prep-
aration of this m a n u s c ~ p,t
This work was supported by EPlZ Contract 68-01-7288 and Na-
tiand Institutes of Health Grant AM 33543.
REFERENCES
161:1368, 1985,
M, Bigby, T. XCwan, and M. S. Sy, J , I n v e s t , Dermatol,, 89:
495, 1987.
F . Kaning, C , StingE, W . M , Kokoyama, HJ. Yomada, W. L.
Mdoy, E l Tschachler, E l M. Shevach, and S , E. Coligan,
Science, 236: 834, 1881,
L , L, Lanier, N. A . FederspfeX, J. J. Ruitenberg, J , H ,
Phillips, J, P. AUison, D. LSttman, and A . Weiss, J . Exp,
Med, , 165: 1018, 1987.
%2, I?, Haynes, CZln, Res., 34:422, 1986.
Immunology of the S k t n 107
where Ap and AE, are the area fractians of the polar and fipophilic
pathways, respectively; P is the partition coeftfbient of the nongo-
lar phase; Gw is the drug" water solubixty (in mglml); and L is
the thickness of the stratum cornearn (in cm). Bemer and Cooper
suggest that Ap = 0.1 and AL = 0.9. The octanol /water partition
coefficient is used for P , and a vdue of O.Q015 cm is suggested for
the thickness of the stratum corneurn, Using these vafues for the
parameters, J will have the units mg/cm2 * hr.
Berner and Goapelr have &so proposed a three-pardlel-pathway
madel ( 4 1 , which combines not only the polar and lipophiaic pathways,
but &so an oil-water multiXaminate pathway (2). Use of the same
vdues far I , DL, Dp, and Ap as for the two-parallel-path model,
and decreasing the value for AL from 0.9 to 0.5, the equations to
caIculate the upper ( J U ~ ~and ) lower ( J L ~ bounds
~ ) of the max-
imum fiux thmugh the stratum csrneum are e v e n in E q s , 4 and 5*
The values for $Urn, and JL,, have been averaged in this evalu-
ation to [ p r s ~ d ea shgle v&ue for the three-parallel-path model,
V, CONTOUR PLOTS
A-taminiphen ( 10) 151 2.95 0.012a 4.6 0.033 0.032 0.064 0.032 0.038
Atropine (5) 289 0.006 2.4 0.02 0,061 0.081 0,028 0.017 0.05
Benzoic Acid (10) 122 89 3.4(11) 720 439 286 280 38 270
Benzyl Alcohol ( 10) 108 10.5 40(11) 1,060 779 589 727 251 590
Caffeine ( 10) 194 0.955 22(fl) 1.5 12 14 40 23 20
D i d t o x i n (5) 765 0.014 0.01 0.00013 1.3 x 1.7 x lom7 0.00027 0.00017 0.00009
Dextromethorphan (10) 271 13,489 0. 017a 10 30 19 2.8 0.21 14
Diazepam ( 10) 285 631 0.05(14) 4,7 3.4 2.2 7.0 0.6 3.0
Fentanyl (5) 337 200 0.2 2.0 1.9 1.2 22 2.3 5.7
Flurbiprofen
Nydralazhs! HCI (10) 161
N y d m r t i s o n e ( 10) 363
Ibuprofen ( 10) 206
Calc .
DruglRl R2=Rg MW PC CW fluxa
Fundamental equations
liquid solutes
Iog Sw = -1.072logPC + 0.672
log Sw
C
= -logPC - O.OlMP + 1.05
hexamethylmelamines e. g I ., hexmethylneEamine
log S'
W
- 0.007MP + 0.07
= -0.904IogPC
alkyphenols
C
log SW = -0.997 log PC - 0.008 MP + 1.43
-.--- --
Diazepam
Oxazepam
Lorazepam
Temazepam
Desmethyldiazepam
Triazolam
Midazolam
Bromazepam
Chlordiazepoxide
Alprazolam
Clobazam
Figure 7.2 Plot for dmgs of molecular weight 300 and 500. Con-
tour lines are values of predicted transdermal flux in units of pg/
em2 * hr. The oranate and abscissa are natural logarithms of water
solubility (mglml) and oetanollwater partition coefficients, respec-
tively.
h.,
hu
h3
Melting Cale
Ester R MW PC point (OC) CWa Csle. flux
REFERENCES
.
A . F, Kydanieus, and B Bemer, eds, , Transdermal DeTivev
of Dmgs, Vols, 3 - 3 . GIRC Press, Boea Raton, Pla., 1987.
B, 6 , Monkhouse, and A . S . Wuq, Drug Dev, I n d . Pharm,,
24: 183, 1988.
D , W, Osbome, Pharm. Manufact., 3(4):4J, 1986,
. .
B Besner, and E , R Cooper, in T r a n s d e m a l Belivery of
Drugs, V d . 2 ( A , F', Kydonielxs, and B , Bemer, eds,).
CRC Press, Boca Raton, Ra,, pp. 41-62, 1987,
A , S e Mi~haeX~, S 1 K, Chandrasekaran, and 3 , E . Slrzaw,
AlCEfE J , , 22: 985, 1975,
W e 2 , Ailbery, m d 3 , Wadgraft, S, Pharm, Pharmacol, , 31: 229,
1919,
W, J, A l b e q , and J , Hadgraft, J , Pharm, P k a m a c a l , , 31:140,
1879,
J, T, Chou, and P , C. Jurs, in Physic& Ghemicacl Properties
af Dmgs ( 2 3 , W e VdkowsXcy, A , A , Sinkula, and S . C . Val-
vani, eds-). Mared Dekker, New Vork, pp. 1.63-199, 1980.
S, 6 , Valvani, and S N, Vdkowsky , in Physical Chemical
I
The Nemk Index, 10th ed, Merck Ilr Co, , Rahwsy , N ,J., 1983.
S. H e Udkowsky, $3, el Valvarai, W-Y Kuu, and. R-N Bannen-
felser, Arr'zona Database of Aqueous Solubility, 2nd e d , , copy-
dght 1987 by S . flalkswsky*
E, Mukai, K, Arase, M, Hashida, m d W . Sezaki, Int, J.
Pharm. , 25: 95, 1985.
A , MacQondd, A. F, Michaelis, and B. 2;. Senhowski, in Ana-
.
Eyticat Profties of Drug Substances, VoX, 1. ( K Flarey , ed ,> ,
Aeadernjlc Press, Hew Uork , pp, 79-99, 1972.
$3. H . Udkawsky, S. 6 . Vafvani, and T . J , Rosemaur, J,
Pharm. Sci, 72: 866, 1983.
3, A , Akhter, and B. W . Barry, J, Pharm, PharmaeaE, 37:27,
1985.
S, 8 4 , Wallace, J. O, Rura&.is, and W D, Stewart, Can, J,
Pfiarm, Set, 23: 66, 1978.
K. E3. Sfoan, S. A , M, Koeh, K , GI Siver, and F. D. Flowers,
J, Invest, Dematol, , 87: 244, 1886.
R, GarNl, K. Engle, G. Rork, and L, 3. Calidwell, Pharm,
R e s . , 3:225, 1986.
.
K F l o r q , hi Analy tical PpnoffEes o f Drug Substances, Vole 1.
( K , Florey, e d , ) , Academic Press, New York, pp. 397-421,
1872.
A . J. Cumber, and W. 6. J, Ross, Ghem, Biol, Interacttons,
17: 349, 1917.
R , Kdfszan, Quantitative Stmcture-Chrornato~aphicRatention
Relationships, John WSey 8s Sons, New Uork, pp. 232--278,
1987.
W, J , Lambe-t, and L , A . Wright, J . Chromatag., 464: 400, 1989,
B. If, Greenblalt, R, IVI. Arexzdt, D, R. Abemethy, H. G ,
Gjiles, E. M, Sellers, and R 1 . S h a a r , B r , J . Anaesth, $5:
I
985, 1983.
D. J, GreenblaPt, M. DivoXI, D. R. Abernetfiy, H. R , Ochs,
and R. I. Shader, Drug Metab, Rev., 1 4 : 2 5 1 , 1983.
I-I, Bundgaard, A, Buur, S,-6. Chang, and V. W . L. Lee,
Xnl, J . P h a r z , (Amst,), 46:77, 1988,
J, L, Zlatz, CTFA Scientific Manograph Series No. 2 , , CTFA,
Washh@on, D,CI,1983, pp, 29--45,
I). W . Osborne, COLL, Abstract 072, l(318.t ACS Nation& Meet-
ing, New Uork C i t y , Mew Usrk , Aprif 13- 18, 1986,
C. Ackernnann, and C . L. Rynn, Int, J , Pharm, (Amst ,) , 36:
61, 1987.
C , Ackermann, 6 , L, Flynn, and W. M, Smith, Int. J.
fharm. (Amst,), 36:67, 1981,
Kinetic Considerations in the
Design of Surfaclant-Based
Topical Formulations
ANTHONY J . I . WARD* University College Dublin,
Dublin, Irelattd
I, INTRODUCTION
127
128 Ward
I I. EXPERIMENTAL BACKGROUND
A. Passive Drop-on-Fiber
One of the problems of quantitatively examining systems, such as
emulsions, to obtain the kinetic and contained mechanistic informa-
tion has been that of simultaneously defining the volume and inter-
facial area of the system as a function of time. Studies of formula-
tion stability to determine the factors involved in their time evolu-
tion are usually somewhat qualitative and system-specific. An at-
tempt to overcome some of the inherent diffimIties in such measure-
ments has been the development of the passive drop-on-fiber ( 3 , 5 )
technique. This defines the system by fixing in space one droplet
of emdsion size using an inert cylindrical fiber. Provided the dis-
tortions from gravity are negIi@ble, the shape of the axis-symmet.tric
dmp is purely determined by capnlary forces. An analysis of this
type of system ( 3 ) shows that the rate of solubBization into solution
is related in a simple fashion to the relative dimensions of the drop-
let and fiber. Both the volume (V) of the drop and the interfacial
area (A) can be obtained and used to calculate the rate from the
general relationship :
1 dV
Rate = - -
A dt
The way in which the experiment has been performed has mainly
been concerned with systems in which the amount of oil in the dmp
is extremely small compared with the amount required to saturate
the micelles. In this respect, the data obtajned are concerned main-
ly with the initial kinetic p m e s s so that contributions from micelles
containing oil can be neglected.
8. Rotating Disk
Cases in which the solubilizate under investigation is in the form of
a solid that can be fashioned into a shape allows the use of methods
in which the materm is suspended in the solubilizing medium. Dis-
solution of the solubilizate is then folIowed analytically a s a function
of time, with the particular technique being dependent upon the
nature of the system. Radioisotopes, for example, have been used
(1,2) with fatty acid solubilizates. An advantage of this technique
is that it potentiany allows the effects of rheology on the kinetics
to be determined by spinning the solubilizate at different rates.
Surfactant-Based Topical Formulations
Surfactant Cancentration
Do,wCo,w
Rate =
6
OIL
A
Rate = - TI [bla) (C
21. 0
- CMG) [ 31
where b! are the solubilization capacities of the pure oils and the
I
represents the oil mole fractions in the binaly mixtures. The
vslue of the factor P can vafy between zero and (1 + bp/bq), the
former value representing ideal mixing of the solubnizate in the mi-
celle interior, Comparison of data determined from kinetic experi-
ments with values determined at equilibrium show apeement in the
vdue of I" required to fit the data in some eases (&), whereas, in
others, different values are needed in each case. The solubilization
of oil from dodeeane-tetradecane mixtures into micelles of the non-
ionic surfiretant n -dodecylhexaoxyethy1ene glycol ether, f: 12EO6 ,
( 12) shows gmd agreement (Fig. 4a) with a I? value of 1,s f i,e, ,
v e q close to ideal mixing). Xn contrast, solubBization from simBar
mktures into a micellar zwitleAonic surfactant, n -dodecyldirnethylam-
moniurn pmpane sulfonate, shows (Fig. 4b) that, although the equi-
librium data require P = Q, the kinetic date i s fitted only if the
mmimum possible vdue of FZ is used. This discrepancy may be a
result of micelle shape changes in the first stages of the solubniza-
tion process that are not considered in the defivation of E q . 4, A
furf;her manifestation of micellar shape changes that oceurs in the
initial stages of solubilization is the nonlinear temperature depend-
encies observed for the rates in the re@on of the phase-inversion
temperature sf some nonionic surfaelants ( 5,8).
One consequence of the proposed mechanism is that the rate i s
essentidly independent of stirring in the system, unlike the passive
Surfactant-Based Topical Fornula tions
Oils that have sufficient water so1ub;llity are more likely to praduce
kinetics that are dependent upon mechanisms in which a diffu~ion
step is rate-determining . Preliminaq studies ( 7,13) of systems
containing benzene, as either a single component or a component of"
a b i n a v mhtur'e, have indicated that this is tme. Benzene, for
example, is soluble to the extent of about 1500 pprn in water, and
any oil that has a water solubility greater than about 1 0 ' ~w t % may
be regarded as soluble in this context. Oil transport across the
liquid-liquid interface is essentidy that of solubility and is gov-
erned 'by the passive diffusion E q , 2 , Rates determined by the
dmp-on-fiber experiment are for systems under conditions of nini-
ma1 stirring and, therefore, c2o not represent rates that are at the
.
diffusion lirnlit. (i e ,, the diffusion layer thickness , 6 , being on the
order of the dimensions of the ail molecule). Some rates typically
found for soluble oils d i s s o j ~ n gin wafer are presented in Table 1,
fnterestingly, the diffusion layer tltzieknesses d e ~ v e din these
experiments are simnar to those found in membrane pmcesses in
v h o , Furthermore, the ratio of the rates is the same as the ratio
of the oU sdubsitieies in water', as required by E q . 2.
Equation 2 shows that the rate is independent of the surfactant
concentration, inasmuch as it does not affect the act-idty of the oil
isl the aqueous phase. A surfactant concentration scan of the rate
of benzene in contact with aqueaus solutions of sodium dodecyl sul-
fate (Fig. 5) s h ~ w sno i-ncrease untit the total sudactant concentra-
tion is above the CMG. This rntty be understoad in terms of the
relative sizes of the oil "sinksw "prodded by the bufk water and. the
micelles; thus, only when the volume of micdles avdable for solu-
baization is comparable with, or greater than, that from the water
salubaty will there be an observable effect from the surfaatant on
the observed rate, The concentration at which this occurs a 1 oh-
dously depend on the size of the water solubility, GMC, and micelle
volume. The larger the micellar volume and lower the CMG in a
system with oil of a @ven water solubnirty, the Xower will be the to-
tal surlFslctat concentration at which micellar solubaization will have
an important contribution. Factors such as the size of the S U F ~ ~ C -
tant hy drophobe , therefore, will be impor^fiant, as will any factor
that influences the CMC of the surfactant, Sim-ifarly, because non-
ionic surfactants tend to have larger sotubBiz&ion capadties for
oils than do ionic ieurfactmts of simnar rnoleeular volume, this ef-
fect should aceur at lower surfaet ant: concentrations in formulations
based on nonionics.
Surfactant-Based Topical Formulations 135
Benzene
Toluene
p-Xylene
timately with the bilayer took much longer to be removed. This im-
plies that the nature of the oil component location within the sur-
factant association structure i s also of importance in the solubiliza-
tion and related processes. Demonstration of such preference in
the solubilization process in simple micellar solutions is found in ex-
traction from binary mixtures of benzene and eydohexane into either
water o r surfactant solutions (6). The kinetic profile derived from
a drop-on-fiber experiment (Fig. 6) is no longer linear, implying
that the composition of the oil drsp is also time-dependent. A buIk-
scale expesiment, in which the composition of the oil in the micellar
phase was monitored by gas chromatography (131, showed the com-
position to increase to a ratio of about 9: 1 from an inital I: 1 bulk-
oil composition within the first hour of contact between the bulk oil
and micellar phases. Here, the process is determined by the diffu-
sion rates into water, which are in a ratio of approximately 30: 1 in
favor of benzene. Similar effects have recently been found in the
138 Ward
REFERENCES
I, .
3 , A , Shaeiwitz, A . F, -C Chan, E. L. Gussler, and D. F,
Evans, J. CatEoicl Interfizee Sci, , 89: 47, 1981.
2, 6 , Nuang, f), X;". Evans, and E l L. Cussler, J . Colloid Inter-
face Sci, , 82: 499, 1981.
.
3. B J , Carmll, J , Colloid Interface Set, , 79: 126, 1981.
4. Y . C. Chiu, V. 6. Wan, a d W. M , Cheng, rln Structure/
Performance Relationships In Surfactants, ( M e J . Rosen , ed .I.
ACS Symposium Series, p. 23, 1984,
5. B , 9, Carmll, B . 6, Q'Rourke, and A , J. I. Ward, J. Pharm,
PhamacoZ, , 34:281, 1982,
6. A . J, I , Ward, M . C , Carr, and J , Cmdden, J. Colloid Inter-
face Sci. , 106: 558, 1985.
Surfactant-Based Topieat Fomulatians 141
The last two steps in this process have been the foeus af much at-
tention for solid-dosage forms ( 1 , 2 ) and, more recently, for dis-
perse systems ( 3 ) . However, these optimization techniques require
that the responses be smooth and continuous throughout the vari-
able ranges, For topicals, smoothness and continuity must be es-
tablished for the composition ranges before optimizatkn. Thorough.
characterization of the formulation by determining the phase behav-
ior of the system is a way of establishing smmthness and continuity
throughout the composition range. Thus, selection of component
ranges by phase b e h a ~ o rdetermination will be the main focus of
this chapter,
Optimization of Topical Farmulatisns
A, Phase Behavior
Disperse systems a m unique among pharmaeeutic& dosage forms be-
cmse the physicochemied intctrilctions between the components dorn-
inate the stabBity of the system, Small changes in the component
composition can completely destab2ize or dramsllieally improve the
sti3lbQity of the dispersion, A well-characterized example would be
the stable emulsion re@an for the wat~r-p-xylene-decaaxyethylene
glycol nonyphenol ether system, In this systern the stabnfty of the
emulsion dramiltjielnlly increased when the surfactant concentration
was changed, from 3 to 4 wt % ( $1, This dramatic change in physical
stablility was related to the formation of lamellar liquid crystdline
baayers around the oil droplets. The formation of a third phase
(e. g. , a liquid eqstdline phase) is now a common method of h-
creasing emulsion stabiEty (5). Other examples of small composi-
tional changes affecting product properties would include the effect
of pH on, clay suspensions or the effect of solids e o n t e n o n the
melthg point of a suppository,
To understand dispersion stabgity, it must be realized that
changes in stabaity are the result of chmges in the phase behador
of the system, For surfact ant systems , sufficient scientific inves -
tigations have been completed to be@n to correlate emulsion (4,5),
foam (61, and gel (7) stabifity with the equilibrium that exists be-
tween the thermodynamically s a b l e micellar and liquid crystalline
phases. Although the techniques for representing phase b e h a ~ o r
described in this chapter are not limited to surfactant systems, they
have been pdmafly applied to this area of study, Nontraditional
formulations may genertllly be less dependent upon emulsifiers than
current topical creams; however, binary. and ternary phase mpre-
sentations are useful in characterizing any of a vadety of systems,
Most topical f"armu1ations are composed of matedals that fa11 into one
of three groupings: polar components, nonpolar components, and
..
other components (e g , amphiphilie, polymeric, therapeutic, bio-
technolow products) . For any formulation whose components can
be d i ~ d a dinto two or three such groupings, b i n a q or ternary
plots of the phase behavior can be extremdy useful in ehzzracteriz-
ing the physicoehemical interactions between the components of the
system. These physicschernical interactions between the indiv_idual
components ultim~telylead to the stabnity or 1acR of stzibility char-
acteristic of pharmaceuticll.l. dispersions.
1, Binary Systems
Bixlary diagrams are typiedly plotted on Cartesian coordinate grjids
in which the composition is plotted along the abscissa and tempera-
ture is plotted along the ordinate. Phase boundarYies are plotted on
this grid to separate single-phase re@ons fram multiple-phase re-
@ens. An idedized example of a two-component liquid system hav-
ing m upper criticd solution paint is shown in. Figure 1, The left
verticical axis represents the phase behador of pure A over the ten-
perature range of interest, whereas the right ve&ical axis is far
pure B. Both A, and Ea are liquids for this temperature range at
the given pressure. A s expeeted, the so1ubiXity of B in A in-
creases as the temperature is increased, This change in solubgity
is shown by the left phase boundary. Analogously, the changes Jn
the solubility of A in B with increasing temperature is represented
by the right phase boundary, The two-phase liquid-liquid re@on
f&ls between these boundazllies. The phase diagram p r o ~ d e seon-
siderable information about any sample that is mixed within the con-
ditions represented by that diagram, First, for any binary mixture
of components A and B , it can be discerned whether the sample
will be one or f wo liquid phases. Second, if the sample is two
phases, it can be discerned what the composition of each phase w21
be and how much each phase will weigh. For example, consider the
point circled in F i ~ r e1. Overall, ibis sample contains 40- and
60% B. We ertn immediately see that this sample will separate ento
Optimization of Topical Formulations 147
CONIPOSlTlON
Figure 9.7 Idealized binary diagram for a two-component Iiquid sys-
tem having an upper critical point.
two phases at 100°C, with one of the liquid layers having the eom-
position 85%A and 15%B , and the other liquid layer having the
composition 20%A and 80%B . The constant-temperature line that
conneds these three colinear composition points is called a tie line.
For binary systems, all of the tie lines are understood to be parallel
( i .e. , constant temperature) and need not be drawn. Thus, for
m y composition at 100% within the two-phase region, the eomposi-
tion of one liquid layer will be 85%A and 15%B , whereas the other
Iiquid layer wiIl have the composition 20% A and 80% B. The only
diffemnce will be the relative amount of each of these phases as
one moves along the tie line.
To calculate the amount of each layer, remember that the weights
of liquid layer 1 (ml) and liquid layer 2 (m2) must add up to the
total weight of the sample, thus
The sample Ellso has a gross weight percentage of component A that
is designated by y in Pimre 1. However, as stated before this
sample breaks into two phases that have weight percentages of corn-
ponent A equal to y~ and y ~ . Thus, the conservation of' camponent
A &lows us to wAte
which rearranges to
As seen, the weights of the two phases are in the proportian of the
len@lrs of the tie line segments extending from each side of the
sample point. Another way to write this relation is
which both upper and lower critical points exist, the phase bound-
ary (also called the coexistence curve) forms a closed loop. Thus,
as the temperature of a miscible mixture is lowered, the mixture
first separates into two phases and, then, reforms as a miscible
mixture. A very readable article describing these reappearing
phases recently appeared in Scientific Amel-ican (8).
If the temperature range of interest for the two-component sys-
tem is such that one or more solid phases now exist, the binary
diagram has a considerably different appearance. A s seen in Figure
2, two components that are miscible when liquids, and in which only
pure crystalline forms of the two components form as solids, pro-
duce distinctive phase behavior when combined. Again, it is un-
derstood that the pressure for the system is held constant. Above
the curved lines AE and BE, the components are mutua11y soluble,
forming a single liquid phase. Below the horizontal straight line,
the temperature is sufficiently low that no liquid phase exits. Be-
tween the curved lines AE and Be ~s the two-phase re@on in which
the pure components are in equnibrium with the single-phase liquid.
Just as b e k r e , any point in either of these two-phase re@ons will
have a horizontal tie line passing through it. The location at which
this tie line intersects the phase boundary will give the composition
of the liquid phase, whereas the relative amounts of the liquid phase
and solid pure component can be calculated by Eqs, 4 through 6.
A s outlined by Barrow (91, it is instructive to consider what
happens when solutions of two particular concentrations are cooled,
Consider Srst the composition labeled b (60% naphtkdene in benzene)
at 70°C1 As the sample is cooled, it wig remain a liquid ant3 cmss-
ing the phase boundaw labeled as curve BE, Upon cmssing this
boundav the first solid naphthalene crystals will become detectable
in the sample. Note that for this diagram the length of the tie line
segment to the &ght of this composition will not change, whereas
the length. of the left tie line segment will increase with decreasing
.
temperllbure Thus, the amount of solid naphthdene wilt continue
to increase in the sample until the hodzontal line at approximately
-4OC is reached. As stated p r e ~ o u s l y ,below this temperature no
liquid exists; thus, the sample that contained mughly equal amounts
of primaw naphthalene crystds and liquid (2Q% naphthdene and 80%
benzene) at -3OC becomes completely solid at - - P C , consisting of
40% by weight prjlmav naphthalene crystals and 60% by weight of a
fine crystat-gr&ned mixture of 20% naphthalene and 80%benzene
Ci.e, , the errteetic mixture). This behador can be compared with
the phase changes encountered by the composition labeled e (20%
naphthdene) at 5QQC. As the sample is cooled it remains a single
liquid phase until reaching the --.4aG line, st which time the solution
becomes in egusibrium with pure solid benzene and pure solid naph-
thdene. The sample remains at this temperature untiil completely
converted into a bne-grained crystal mixture of the two solids, hav-
ing the composition 208 naphthalene and 80%benzene, The point
labeled E is cdled the eulectie point. Mote the eutectic mixture has
a melting point lower than either of the pure, components.
Although the benzene-naphthalene system is useful in under-
standing eutectics m d identifying eutectics on a binary diagram,
most systems of interest Gontain components that interact with one
another in the solid state, Such interactions result in solids or
semisolids that are mixtures of the two components. Interactions of
this type can also result in the formation of solid-state compounds
consisting of simple mole ratios af the two components. The effect
that formation of campounds and mutual. component solubility has on
the appearance of the binary phase diagram is shown in F i w r e 3 .
The mutual. solid-state solubnity of A with B can. be seen along the
vertical axis of the plot, When B be@ns to solidify at temperature
Optimization af Topical Formulations
I I I r I I r I I I
100% 60160 10076
Whlte Anhydrous
Petrolatum Lanolin
Figure 9.5 Freezing point diagram for the semisolid system white
petrolatum-anhydrous lanolin, Component impurity and wide melting
ranges result in an appearance si@f"rsantly different than for ide-
alized diapams. The phase regions were observed to have the fol-
lowing appearance: (a) single-phase , clear, colorIess liquid ; (b) two
clear Iiquid phases, top colorless, bottom yellow ; (c) single-phase
clear yellow liquid ; (d) white petrolatum with dissolved lanolin ;
( e l cloudy semisolid top phase, colorless clear liquid bottom phase;
(f) homogeneous semisolid ; ( g) two semisolid phases ; top, white,
bottom, yellow ; and (h) two semisolid phases,
t I I 1 I I 1 I
100% 60150 10096
Whlte Anhydrous
Petrofatum Lanolin
Figure 9.6 Freezing point diagram for the semisolid system white
petrolatum-anhydrous lanolin. Comparison with Figure 4 emphasizes
the phase behavior changes that can result from lot-to-lot variabif-
ity of a natural product. The phase regions were observed to have
the following appearance: (a) two clear liquid phases ; top, color-
less, bottom, yellow: (b) multiple phases both liquid and semisolid.
%
White Anhydrws
Petrolatum Lanotin
Figure 9.7 Freezing point diagram for the semisolid system white
petrolatum-anhydrous lanolin. For this particular lot of lanolin,
the phase behavior is similar to the idealized phase behavior for two
miscible solids as shown in Figure 4.
Figure 9.11 Example of the utility of tie lines for determining the
amount and composition of each phase in a two-phase system. The
composition marked X will split into two phases upon reaching equi-
librium. The composition of each phase wiIl be the same as the
composition at the phase boundary-tie line intersection (circled
points), whereas the amount of each phase is proportional to the
geometric distance of the phase boundary-tie Iine intersection
points (circles) and the sample point (X) ,
sample containing 65%A, 22% B , and 13% C wiIl split into a lower
phase with the composition 79% A, 10% B, and 11% C and an upper
phase containing 65% A, 22% B , and 13% C . The amount of each
phase is propo&ional to the geometric distance of each phase bound-
ary from the sample point. By measuring the distances denoted on
the tie line as a and b and knowing the total amount (mass) of the
sample (mtot), the equation for the amount (mass) of the B phase
(i, e. , top phase) would be
m a = m tot [ b / ( a + b ) J
As seen, Eqs. 7 and 8 are identical with Eqs. 6 and 7 and can be
derived andogously once it is realized that the sum of the three
perpendicular distances from any p i n t to the three sides of the tri-
angle i s equal to the height of the triangle.
For the samples that are contained within the three-phase tri-
angle, the composition of each of the phases is given by the point
of intersection between the phase boundaries and the three-phase
triangle. Figure 12 shows an expanded three-phase triangle of the
system shown in Figure 11, Any sample that is Iocated within this
triangle of tie lines will separate into three phases with the a-phase
component having the composition 71% A, 12% B , and 17% C, where-
as the 8 and y phases will have the compositions of 63% A, 19% B,
18% C, and 61% A, 7% B , and 32% 6 , respectively. Thus, once
within the trisulgular three-phase region, only the amwnts of each
phase will change, whereas the compositions of the phases will re-
main the same. The amounts are determined in a fashion analogous
to the method used for two-phase samples. Equations for the
amounts of a phase, B phase, and y phase for the sample denoted
by the open circle in Figure 12 am
m = m Ec/(c + dl1
$ tot
m y = mtot [e/(e + f)5
Ab isotropic solution
solution
( 1; f wt, ratio)
tiquid cr ystal
Eigura 9.16 Dia-am showing the complex phase behador for the
multiple-component water-lipid mixture-TEA :QL system. The mul-
tiple-phase re@ona were observed to have the following appearance
after centrifugation: f a ) cloudy white and elear gray isotropic lay-
ers; (b) cloudy white, clear gray, and anisotropic layers ; (e) cloudy
white and clear cdorless isotropic layers ; ( d ) cloudy white-stable
to centdfrx@ng; ( e ) cloudy white, elear colorless isotropic and
anisotmpic layers ; (f) lamellar and hexagonal Xiquid c q s t a l mixture ;
Ig) isotropic solution and lamellar liquid c q s t a l mixture.
Cloud Temperature
Of agram
5% 10% 15%
SURFAETANT
clear
Temperature
out the range of the component composition, Second, once the suf-
ficient number of samples have been mixed and examined to pmvide
the information required to construct the diagram, detemining the
composition ranges over which to optimize is very straightforward,
Values for the ranges can be read directly from the plots based
upon the phase boundaries. Thus, the determination of the phase
behavior of the fluid or semisolid system is required before the op-
timization of a topical formulation.
Dkgram
number A B C
5% Active, 0,1%preservative
1 92% Water/8%glycerin 40%Emulsifier/60%cwmulsifier on
2 92% Water/@%
glycerin 35%Emulsifier /65%caemuIsifier on
3 92%Water/8%glycerin 30% Emulsifier /TO%coemulsifier Oil
4 86%Water/ 14% glycerin 40%Emuldfier / 60%coemulsifier OiT
5 86%Water / 14% glycerin 35%Emulsifier/ 65%cwmulsifier on
6 86%Water/14%glycerin 30%Emulsifier/7O%coctmulsifier OiT
10%A c ~ v,e 0.1% preservative
7 92%Water/8%glycerin 40%Emulsifier / 60%coemulsifier oil
8 92%Water/8%glycerin 35%Emulsifierf65%eoamulslfier oil
9 92% Water/8%glycerin 30%EmulsiEer/?O%eoemulsifier Oil
10 86%Water114% glycerin 40%EmulsiEer/60%eoemulsifier oil
11 86%Water/ 14% glycerin 35%Emulsiaer/ 65%coemulsifier Oil
12 86%Water/ 14% glycerin 30%Emulsifier/?O%coemuIsifier Oil
15%Active, 0.1% preservative
13 92%Water/8%glycerin 40%Emlsifier/60%coemulsifier on
14 92% Water/8%glycerin 35%Emulsifier /65%coemulsifier OiI
15 92%Water/8%glycerin 30%Emulsifier/70%coemulsifler on
16 86%Water/ 14% glycerin 40%Emulrsifier /60%coemulsifier on
17 86%Water/ 14%glycerin 35%Emulsifier/65%coemdsifier Oil
18 86%Water/ 14% glycerin 30%Emulsifier/?O%coemulsifit?r O i l
lates into the preparation of 738 samples, Dependent on the corn-
glexity of the dispensing and mixing steps and whether or not the
samples will be centrifuged, each sample will require approximately
1Q min to complete, which leads to an estimate of 123 h r (5-118
days) of robotic time, Approximately 2 h r of technician tirne is re-
quired for each day (24 h r of continuous, unattended operation) of
robot time. Thus, the initial samples for this evduation will require
approximately 1 technician-day and 1 robot -week to complete, Ad-
ditional samples mag be required to pinpoint cfiticaX re@ons of the
phase boundary , as described in the following discussion.
the powder to avoid spillage. A lip on the mouth of the tube con-
tdning the powder has proved useful, but requires that the dis-
penshg tube is returned to an exact orientation, This can be ac-
complished by notching both tube and rtack sufficiently to warantee
return to a constant orientation, The tilt an@e at which the tube
is held &so determines the rate at which the solid is dispensed*
Accurate addition of solid depends on an appropriate combination of
tilt mgle , *bration intensity, dbration duration, and a knowledge
of the physical pmperties of the solid,
To pour powder, the robot grips the powder tube and tilts it
to an initial angle over the contrainer. A dbration pulse foIlows,
after whkh the tube is weighed, A lag time usually occurs, repre-
senting the time required for the solid to reach the tip of the tube,
If only a s m d l fraction of the target: weight has been. reached, the
tilt angle is increased, The hand ~ b r a t e sand powder is added.
A. compaAson between the target weight and the amount added is
made. Ass the target weight is approached, both the brati ion in-
tensity and duration are decreased and the pouring routine contin-
ued, When the target weight is obtdned, the pouxlling mutine is
stopped, and the net weight of solid cdeulated, By using this
technique, we have obtdned g w d reproducibsity with target weights
as low as O,O5 g.
3, C o n t a i n e r s , C a p p i n g , and Labeling
Ilirtudfy any contdner is suitable for a laboratory robotic system,
proeded the weight of the f2Eed sample cont&ner does not exceed
the lifthg capacity of the robot* Otherwise, contdners should be
selected based upon the ease of completing pmper mixing within
the container and the washabnity or disposabBity of the contdner,
We have found 16 x 125-mrn disposrzble culture tubes ~ t caps
h to
be ideal for our phase behavior investiigations.
Capphg stations function by either having a deAce that will
grip and rotate the cap whife the tube is held stationary by the
robotic hand, or by having a efedce that tiltill g ~ and
p rotate the
tube, whae the cap is held stationary by the robotic hand, The
former system holds the cap unt2 the tube is repositioned to the
capplistg station, and the cap is torqued onto the tube, Thus, only
one tube is typically uncapped at any @ven time, The latter sys-
tem sows more than one tube to be uncapped during any single
command cycle.
Optimization of Topical Fomulations 183
4, Mixing Samples
After component addition and weighhg, the next stage is to effee-
tively mix each sample, For Buid samples, an automated vortex
mker can be instdled with the robotics system to operate in e&her
a single-action or pump action manner, The length that the sample
is vortexed can eassy be adjusted to meet the requirements of the
system studied, To facaitate mking of more dscous samples, sr
samples that contain components that melt at sBght1-y h i e e r than
ambient temperatures, a heating block can be positioned d t h i n the
work envelope of the robot, and samples may be heated for a pre-
set Xeneh of time, The method of centz.ifu@ng ~ s c o u ssamples
through: a constricted glass tube can also be useful Tar mixing sam-
ples by use of laboratoq robotics. A soft glass tube 5 to 6 in,
Xong is sealed at one end, and a constriction, is placed roughly in
the middle of the tube. After addition of the components, the
mouth of the tube is then flame-sealed, and the mat^ is repeated-
ly centrifuged firam end to end of the tube through the matdx,
Slight annealing may be required after mking by this process (17).
The canstruetian and seding of the constdeted tubes w3l need to
be completed manuafly, although use of robotics is ideally suited for
f XI@ repeafed inverting and centrlfu@ng of the eonstdcted tube,
5, Observation O f Samples
After a we& of robotic effort, the reserzrcher is fsced w.i"eh evaluat-
ing 738 samples and plotting the phase behador of each sample on
one of 18 pseudoternary plots, Xt quickly becomes apparent why it
is necessary that all of the critic& processing steps be completed
using the laboratory- mbotics configuration. If each sample re-
quired a manu& processing step (e.g., mking vvlith a spatula) the
amount of human efforpt; would be unacceptable. Sample evduatiran
should consist of a simple observation to establish the presence or
absence of a homogeneous phase, Thus, each point on each dia-
gram can be quickly platted as a drcle if one phase, and as an x:
if two or more phases are present. The best smmth curve passing
between the xs and os is considered the phase boundary, If a
best mrve cannot be drawn because of lack af data, then add-i(ional
Figure 9.24 Comparison of a rough diagram (a) drawn from an in-
sufficient number of data points and, a finetlized d i a p a n (b) that
has a well-defined phase boundav.
5 - 15 Active
0.1 Preservative
Emulsifier
Goemulsifier
Glycerin
Water
agram as seen in Figure 14 and in Ref. 18. Thus, for most com-
mercial applications, the researcher defines the phase according to
the stability requirements for the product.
Before the experimental design, modeling, alid optimization steps
can be established, the researcher must guarantee that a phase
boundary is not crossed as the composition of each component is
varied. Pseudoternary diagrams have now been constructed to
bracket the constrained ranges of each of the component variables.
If each of the 4 1 samples on each of the 18 diagrams was within the
desired phase boundaw, then the compositional ranges to be op-
timized would be identical with the values of the constraints we
placed on each component. However, if the phase behavior was as
shown in Figure 25, then the compositional ranges would be more
limited. Note that for 5% lo%, and 15%active compound, the 30%:
70%emulsifier/coemulsifier provide a stable emufsion only at higher
than 20%to 25% by weight of the emulsifier-coemulsifier mixture.
Thus, the emulsifier/coemulsifier ratio amst be greater than 30: 70.
For the other two emulsifier/coemulsifier ratios, the mount of oil
can range from 5 to 20 wt%, whereas the amount of emulsifier-co-
emulsifier can range from 10%to 25%. Within the above constrslints,
the water-glycerin mixture can range from 45%to 85% (remember,
the formulation contains a minimum of 5%oil. and 10% emulsifier-eo-
emulsifier). The component ranges based upon the phase behavior
results shown in Figure 25 are given in Table 4.
Qsborne
Figure 9.25 Phase behavior results for the ternary systems listed
in Table 1 for the hypothetical cream system example. X s denote
either clear single-phase regions, two liquid phases (i.e., an un-
stable emulsion) or stable emulsion plus another phase.
Optimization of Topical Formulations
Trial Variable
Optimization of Topical Formulations
Trial Variable
REFERENCES
Preformulaaon, stage
Formulation development stage
Proposed product stage
New product stage,
Established pmduet stage
Re.rriseEI pmduct stage
Preformulatbn Pre-IND
Formula*n development Phase 1-11
Proposed product Phase 111
New product Scaleup and approval
Established product Postapproval
Revised product
The design of each study must begin with a clear objective. With
the objective in mind, the attmutes to Be measured are defined,
and appropriate test methods are defined or developed; the storage
canditions, assay schedules, and types of samples are selected ; and
the timing and method for data evduetion are detemined,
The importance of a clearly stated objective cannot be overem-
phasized. For example, a study to determine the effects of metal
ions on the stability of a new active ingredient in aqueous formula-
tions will be quite diffentnt from a study on an established product
to confirm its stabnity performance profile.
I I. PREEORMULATlON STAGE
A. Profile Studies
The first type is to determine the profile of physical and chemical
properties of the new drug substance. To identie probabIe routes
of demadation, the active ingredient should be subjected to short-
term accelerated studies to produce obvious changes (Table 2). The
storage conditions should be geeared to possible degradative condi-
tions, such as heat, light, maisture, and oxidants. Testing sched-
ules are chosen to allow estimates of stability parameters. The sam-
Tabfe 10.2 Profile Studies
B . Toxicalagy Supplies
A second useful sauree of stability information during the preformu-
lation stage comes from the monitoring of toxicology supplies. These
represent the first rudimentary formulations of the new drug sub-
stance. The objectives are to assure that the potency and level of
significant degradation products are documented during the use of
the supplies. An important added benefit is the development of
supporting data for use during the formulaaon development stage.
A. Formulation Comparison
A convenient means for conducting formulation comparison studies is
to first screen the trial formulations for undedrable physical
changes under accelerated conditions. For example, ointments and
creams may be subjected to temperature cycling to determine the
relative tendency to separate, Topical suspensions should be
screened for changes resulting from particle growth by Ostwdd rip-
ening, for changes in polymorphic forms, or for a tendency to cake
(4). Next, the formulations should be screened for chemical stabil-
ity. By using the assay for the drug substance Pram the previous
stage as a starting point, a stability-indicating assay can be devel-
oped for the most promising formulations. Specificity may be deter-
mined by spiking the assay preparation with like compounds or au-
thentic samples of known or suspected degradation products.
Extensive use can be made of short-term studies at high-tem-
perature and high-humidity conditions (5). For example, the ex-
Stability Testing
Months
A
Months
C. Cllnical Supplies
An additional component of this stage involves the monitoring of
cEnicd supplies (Table 4).
Were, the objectives are to determine the values of the critical
quality parameters of the formulations used to establish safety and
efficacy as a function of time and to ensure that only satisfactory
material is used in the clinic.
Stability Testing
The proposed product stage covers the pedod fmm selection of the
find formulation to NDA fsing ; it roughly corresponds to the pedod
during which Phase Ill clinical studies are conducted, At this
point in the development pmcess, a substantial amount of fundarnen-
tal stabgity informiation will be available. Any inherent sensitftivlities
of the active ingredient have been determined, sipificant degrada-
tion products identified, and major stabiXfty-limiting charactefistics
of the various expedmental formulations have been d e t d e d .
The objectives of this stage are (1) to assess the stabi2ity per-
hrmance of the selected formulation for the purpose of defining
storage conditions and stabiIity -limiting factors ; (2) to eonfirm deg-
radation pathways in the selected formulation; and (3) to establish
the b i t i d expiration dating periad for the contdner-closure system
jU1 which the product is to be marketed.
A greater variety of test data can be collected du&ng this phase
to establish the stabfii2-y-limiting factor(s). Both short;-'term accel-
erated studies and long-term studies at proposed label storage con-
ditions are conducted. The tests and assays to be conducted
skould be essentially the same as those given in Section 1If.G.
Many sf the tests and, assays ineluded in the stabnily protocols uti-
lized during this stage will not be included in the protocols tn later
stages because they will be deemed redundant or unnecessary, In
generd, at least three lots of the proposed formulations are placed
on test in the proposed contaker-closure systems, These lots are
usually produced in production-scale equipment f Table 5) ,
There are eases when scale t s s t h g may not be sclentific&ly
necessary. X f a stable, noncornplex flormulation. (formulation defined
as er f h e d ratio of active ingredient(s1/excipients) is being devel-
oped in several strengths or coxlcentratioians, full testhg should be
conducted on the highest and lowest strength or coneentratbn, The
data from this testing regimen wiU prodde information to fuHy char-
acterize the stabnity performance profs@of formulations at all inter-
mediate strengths. For example, the lowest concentratha would test
the case for which the package surface area per maligram of active
ingredient is at a maximum, whereas the highest concentration would
test the case for which the probabglty of nucleation and crystal
growth I mlutions) or agglomeration (suspensions) is maximized,
Stability Testing
Table 10.5 Proposed Product
The last stage i s the established product stage, whkh continues f"or
as long as the firm manufezetures and markets the gmduct. Dudng
this stage, concern again shifts. At this point, a comprehensive
data base is in place, and the objective beeones one of pedodicany
reassessing the stabntg of' the pmduct .
Because much stabgity data has been generated on the pmduct
during the preceding stages, generally fewer data per product are
required in any @ven tirne span dufing thiss continuation of the
stability testing pmpam.
In general, one lot per year should be placed on test, t o t
selection should be designed ta ro"tte among the various marketed
cantminer-closure systems. The protocol should include tests and
assays far significant q u d t y attributes, as estabashed in predous
stages, The study desim is for full-term studies at label storage
conditions. The schedules t&e into account the relative stability
of the product as well aa the relative sensitidties of the vadous
stability parameters determined during the g r e ~ o u sstages (Table
6) *
Stability Testing 209
VIII. CQNCLUSION
REFERENCES
I. INTRODUCTION
Number of repficates
Human skin
source Formula~on1 Formulation 2
A 4 4
Total 12 12
Ifl. FACTORS
A, Membranes
The literature is full of reports showing that skin abmrption is dif-
ferent in most animals than in humans (Table 2; 3). Human skin
is available, so it should be the membrane of choice.
Human skin should be used as soon as possible. Obviously,
freshness is best. Reports suggest that the freezing of skin
changes fts permeability characterhtics, If skin metabolism data
are warranted, then appropriate procedures and verifications are
warranted.
In Vitro Testing of Topical FormuIations 215
Paraquat
permeability rate (vglcm2)
Water Paraquat
Human
Rat
Hairless rat
Nude rat
Mouse
Hairless mouse
Rabbit
Guinea pig
B. Cell Design
The diffusion cells, connecting lines, and collection chambers should
be made from inert, nonreactive mateAa1. The drug can disappear
thraugh volatifity and apparatus adsorption. Material balance in
the study design is a check on this. Table 3 shows this dvferclnce
( 4 ) . If solubility in reservoir fluid is a problem, then the larger
volume from a Row system may result in relevant data (Table 4; 5).
C , Temperature
It is fortunate that cirrmlating bath water of 37OC results in skin
surface temperature of 3Z4C, just as in humans in vivo. Obviously,
temperabres should be veflied .
D, Dose
This is easily determined by taking a known volume of formulation,
~preadingjt over human skin, and measuring the skin area covered
218 Wester and Maibach
Chemical (%)
Percutaneous absorption
(systemic) 0.045 t 0.037 0.03 t 0.00 5.2 t 1.6
Surface boundlstratum
comeurn 0.036 f 0.005 6.8 2 1.0 0.2 + 0.17
Epidermis and dermis 0.065 2 0.057 5.5 t 0.7 0.61 t 0.25
Total (skin/systemic) 0.15 12.3 6,l
Skin wash /residua2 2.51 2 0.94 71.2 k 5.5 92.4 2 0.8
Apparatus wash 0,006 f 0.005 0.25 t 0.2 0.47 2 0.03
Total (accountabBity) 2.67 83.3 99.0
Doae absorbed
Type (% +- SD)
C , Equilibration
Phy siolo@cam (except for. some psoriasis treatment) a person does
not sit submerged in a tub before appXHng a dermatofo@cal or
transdermcsl d m g , Therefore, it seems irrational to use eguaibra-
tion beyond 30 min to determine if the system is working (bubbles).
Because human skin tends to faXl apart after 24 to 48 Izr in an in
d t m rmn, adding more water time generally seems irrelevant.
I. Receptor Fluid
The pmdous published recammendation is as folXaws ( I ) : Far most
studies, an isotonic mlution buffered to pH 1.4 is a suitabXe and
preferred receptor f"luid; a different pH can be used if it can be
justgIed, fn all Instances, the thermodpamic actidty of dmg in
the receptor Buid should not exceed 10boE its thermodynamrie ae-
t i d t y in the donor medium to maintzlin, a favorable drigr.ing k r c e for
permeation and to assure reasonable and efficient colfection of per-
meant. The receptor medhm may need alteration h r n a str;Ictly
aqueous medium to attdn this endpoint for hydrophobic compounds ,
This factor supercedes coneem for rnahtdning a minim& receiver
volume,
Wester and Maibach
Spironolactone content ( %)
Reservoir Surface
Formulatbn fluid Skin wash Total rwovery
REFERENCES
1985.
.
3, R , C Wester, and H , I , Rlaibaeh , in PereuEaneous Absarp tfan
( R I Bronazxgh, and W. Maibach, eds.). Marcel Dekker, Hew
York, p. 251, 1985.
4 , R , C 1 Wester, M, Mobayen, and H. 1 , Mdbzich, J. Toxicol,
Envlron, Health, 2 1 ; 3 8 7 , 1987.
5, R. G , Wester, W . I. Mdbaeh, J. Surinchak, and X f , A , W .
Bucks, in Pewutanclous Absorption. t R . Branaugh, and H .
Maibacfi, eds,). Marcel Dekker, Mew Uork, p. 223, 1985.
Drug Delivery from Topical Formulations:
heo ore tical Prediction-and Experimental Assessment
WILLIAM J , ADDICKS* and NORMAN D, WEIEaER University of'
Michigan College o f Pharmacy, A n n Arbor, Miclcli~n
RANE L. CURL University of Michigan College of Engineering, Ann
Alnbor, Michigan
GORDON L, FLY FSN C ygnus Research Corporation, Redwood City,
Calr'firnia
I. INTRODUCTION
the vehicle to its interface with the skin and diffusion across the
outer skin. Even in the absence of specific dmg-vehicle-skin in-
teractions, there are special circumstances in which dissolution of
drug suspended in the vehicle o r diffusion through the inner living
strata of the skin may became rate-determining.
A theoretical basis for the study of the release kinetics of
drugs from both suspension and solution ointments for the case in
which release from the vehicle is rate-limiting was established by
Wiguchi (5). Wiguchi first depicted the situation in which the oint-
ment vehicle is initidly saturated with solute, with excess solute
uniformly suspended a s tiny particles, The exact assumptions for
the derivation of the time dependency of release are a s follows:
When Q > > C g , the amount of drug released into a sink bears the
following relationship to time :
where H Is defined as
and
Here, w i s the mass transfer cmfEefent, and, erfC 8) t&es the iden-
tical form as in E q . 9. Notably, E q s , 15 and 16 are asymptotic re-
lationships that are only applicable st long time. The term & is
the finite intercept on the tff- axis.
Rosernm and Ripchi (12) also developed equations for the sit-
uation in whick finite mass transfer resistance is present; here, in
the form of 8 hydrodynamic bboundam layer, Xn this model, release
fmm a suspension m a t ~ xdepends upon two distinct processes.
The first of these invoIves an initid zero-order phase of release in
which the d m g partitions f r o m the matm";uinto the botxndaq layer.
During this period, diffusion through the boundary layer deter-
nines the o v e r d release rate, The duration of this p e ~ o ddepends
mostly upon the drug's external pbaselmatrix partition coefficient
as diffusion in the liquid fam at the surface af the mat* is pre-
dictably fast and should be reasonably simgar from liquid to liquid,
Xn a more generd ease in which the film is any type of m a t e ~ d ,
the duration would be comparably dependent on the perxneantfs dif-
fusidty through it. Regardless and eventudly, a trmsithn pedod
is reached at which the matrix g r a d u a y assumes control of the
process, ultimately with tho re-fease rate being proportiond to the
square r m t of time, Xn this model, it is assumed that $ > > Gs,
where & is once agdn the total drug eoncentration and Cs is the
drug's solubQity in the matrix. The expression for the quantity of
drug released within an area A is
where krm is the thickness of the zone of depletion within the matdx.
This quantity is related to time through. the fallowing expression:
This equation disclose& that the release of the dmg is djrectly pro-
portiond to the drug" ssdubsty in. the matfix and the mass trans-
fer coefficient, Zero-order release is predicted, as the excess o f
solid d m g ensures the mdntenance of a saturated solution at the
releasing surface, When differentiated for time, an expression for
the instantanwus rate of release is obtdnerl,
Drug D e Z i v e ~fmrn Topical Formulations
6. In V ~ V OStudies
When all is considered, mast topical products on the market have
been developed using more clinical research strategies than bench-
top ones, Iln viva assay methods have played an important POI@in
the development of many important topical products, Some of these
studies have invalved the use of animal models; however, the pres-
Drug Delivetly mnz Topical Formulatiuns 241
REFERENCES
.
B Barry, in Dematological Formulations, Marcel Decker, New
York, p. 145, 1983.
.
M. Katz, and B Paulsen , in Handbook of Experimental Phar-
.
macology, New Series, Vol 28, Springer-Verlag, Heidelberg,
1971.
B. Xdsan, J. Pharm. Sci., 64:901, 1975,
G. Flynn, and B . Paulsen, in Percutaneous Absorption: Mech-
anism-Methodology-Drug Delivery, Marcel Dekker , Mew York ,
pp. 431-459, 1985.
T. Niwchi, J . Pharm. Sci., 50:874, 1961.
T. Niguchi, 3. Soc. Cosmet, Chem., 11:85, 1960.
J. Crank, in The Mathematics of Diffusion, 2nd ed, Clarendon
Press, Oxford, p. 48, 1975.
.
M Jaoobs , in Ddffusion Processes, Springer-Verlag, New
York, p. 44, 1935.
W. Higuchi, J . Pharm. Sci,, 51:802, 1962.
D. Paul, and S, McSpadden, J . Membr. Sci., 1:33, 1976.
P, Lee, 3. Membr, Sei, , 7:255, 1980.
T . Roseman, and W. Higuchi, J. Pharm. Sci,, 51:353, 1970.
J. Ayres, and T. Lindstrom, 3. Pharm. Sci,, 66:654, 1977,
R . Guy, and J. Hadgmft, Xnt. J , Pharm,, 6:321, 1980,
6. Flynn, E. Topp, and 6. Amidon, in Topics in Phamaceuti-
cal Science XBBS. Elsevier Science, New York, pp. 313-328,
1985.
J , Crank, in The Mathematics of Diffusion, 2nd ed. Clarendon
Press, Oxford, pp. 44-68, 1975.
W. 3. Addicks, 6 . L. Flynn, and N . Weiner, Pharm. Res., 4:
337, 1987.
D. GemmeE, and J. Mordsan, J , Pharm. Pharmacol. 9:641,
1957.
G. FXynn, and S. Yalkowsky, 3. Pharm. Sci., 6f:846, 1%?2.
G. Reddish, and N . Waes, 3. Am. Pharm. Assoe., 18:576,
1929.
6. Clark, and G . Davies, Br. 3. Dermatol., 00:521, 1949.
L, Loekie, and J. Sprowls, 3 , Am. Pharm. Assoc., 38:222,
1948.
N . BiUups, and R, Sager, Am. 3, Pharm., 137:57, 1965,
B. Poulsen, E. Young, V. Coquilla, and M. Katz, J , Pharm,
Sci. , 57: 928, 1968.
J . Helblian, B. Poulsen, and K. Burdick, Curr, Ther. Res.,
22:713, 1977.
J. Ostrenga, C. Steinnetz, and B. PouIsen, 3. Pharm. Sci.,
60: 1175, 1971.
Drttg Delivery from Topical Formulations 243
v
- log Xi = ( A H f / 2 . 3 RT)[(T, - T)/T,l + lag Y
V
Because trmsport through skin is a diffusion process that de-
pends on a concentration gradient, the application of calculated sol-
ubiXities, befived in part from 6 , to predict the concentration gra-
dients and, hence, diffushn is a l,a@cal progres~ian. Pick's first
law Tor diffusion (Eq, 3) contains terms for the concentrations of
sl s2
the drug %nthe membrane (Ci and C i 1 that constitute the concen-
tra-t_iongradient ( F i g . I). They are difficult to determine. Hence,
Fickfs first law is txsudly expanded into a form CEq. 4) that con-
S,V
t a h s contributions from mare readny measured terms, where Ji
is the flux of drug (i) delivered h r n vehicle f v ) through a mem-
brane that, here, is skin ( s) , D? is the diffusion coefficient of the
1 a1
--
drug in the skin, h, is the thickness of the skin membrane, Ci
s2
and Ci are the concentrations of the drug in that layer of the skin
next to the vehicle and receptor phase or plasma (p) , respectively,
V S,V
Gi is the concentration of the d m g in the vehicle, and Ki is the
distribution or partition coefficient of the drug- between the vehicle
sl v
and skin equal, to G i IGi. En proceedhg from E q . 3 to E q , 4 ,
s,v v sl s2
K i Ci has been substituted for G i m d Ci has been assumed to
approach zero, because sink conditions are mahtahed on the plasma
side of the membrane, that is, ! C = O (see Fig. 1) +
1
V S
log K : ' ~ = log y i / y i E 59
Use of Solubility Parameters in Skin Transport 24 7
-
ues) for chemical structures, derjlved from solub%ity parameters ( 6 )
where 6 = P/V and where TJ molar volume, could be used to cor-
relate b i o a c t i ~ t ywith chemical stmcture. Because the Xi" vdues
were readiry avasable far most organic functional groups and were
additive on a constitutive basis, they could be calculated for most
ehernicd stmetures without resa~tingto additbnal experimentation .
This was in contrast with the Wansch approach ( 10) which used n
vafues dedved from the expedmentaf difference in logarithms of the
partition coefficients of a model chemical (RE) and its substituted
derivative ( R X ; E q , 9).
n
x
= log K
(RX)
- log EI
(R-EI)
M 2
log o = log(BD IDD) = V
D E
12.3 RT(6E - 6D) 2 l 111
.
Iff APPLICATION OF REGULAR SOLUTION THEORY
70 PERCUTANEOUS ABSORPTION PROCESSES
constant
Salicylic acide
0A (7.6)
IPM (8.5)
UCT (10.3)
PRa (12.0)
PG (14.8)
fable 13.1 (Continued)
Solubility
V
Flux (al~
,1 v pisv
Solventc (Ci)s [mglemzh x [cmlh x 103 S?V
1 CSV; (callcm3f 1/21 (mg/cmfo 103 (5 SD)] ( 2 SD)] log Ki
OCT (10.3)
MEGB (12.1)
DMF (12.1)
nlwso (13.0)
-.I
FOR (17.9) 13.7 15.Q (1.4) 1.1 (0.10) - 0.75
2
-.
aSteady-state fluxes, number of cdf run = 3. 3'
bealculated from Eq. 6; R =: 1.98 calldegree, T = 305 K , 6s = 10 ( c ~ / c m ~ ) lfor
/ ~ ;thwphylline 6i = Y
14.0 (cat/cm3)1f2, Vi = 110 cm3lmol; for salicylic acid 6 i = 14.4 (c&/cm3)1/2 or 61 = 11.0 (cal/cm3)1/2 2
for values In parentheses, Vi = 93.9 cm31mole; for 6-mercaptopuhe 61 = 14.4 (callcm3) l12, V i = 81.2 2
cm3lmol; far 5-fluomuracil 61 = 15 (cal/cm3)112, V i = 70.4 cm3/mol. %
COA, oleic acid; IPM, isopropyl myristate; DET, diethyltoluamide; OCT, 1-octanol; PRO, 1-propanol; 2
MEG , methoxyethmol; DNF, dimethylformamide; DMSO , dimethyl sulfoxide; PG , propylene grlycol; EG ,
ethylene glycol; FOR, formamide.
Sources: Data from d Ref. 19; e Ref. 20; Ref. 21; f3 unpublished resulte ; h Ref. 22.
Figure 13.2
6v
Figure 13.4
Figure t 3 . 5
s,v
experimental Pi . This average difference corresponds to the
constant in Eq. 13. If the thickness of the stratum corneum is as-
s
surned to be about 10 m , then Di is about 10-10 cm2/s for most of
the drugs studied, except for D: for salicylic acid, which Is one to
two orders of magnitude greater. The calculated D!, then, is of
about the right magnitude for such polar molecules.
There are five conclusions that can be reached on the basis of
the results given in Table 1 and F f p r e s 2 through 5. First, there
s,v
is generay a minimum in the log experiment& Pi versus 6v curve
that corresponds to those vehicles exhibiting 6v similar to that of
6i. The drug also usually exhibits its greatest mole fraction solu-
s,v
bility in those sane vehicles so that, generally, Pi are inversely
related to solubility ( 30). Second, for single-component vehicles,
those vehicles that exhibit 6v similar to that of & B also cause the most
damage to the skin as sssessed by second application studies. Be-
cause those 6v are similar to tis, they may be solubilizing a portion of
the membrane that is responsible for providing the diffusional resist-
ance to drug permeation and, especinlly, to polar noleeules exhibiting
a 6 i quite different from 68, Third, from the knowledge of the soh-
Use of SolubiEiZly Parameters in Skin Transport 259
s,v
bgities of a. d m g in various vehicles, the relathnship between Pi
and 6v, m d the flux of that drug from one specific vehicle it should
be passib2e ta determhe the ffux of that d m g from other vehicles of
bterest , assuming that the vehicles do noManrage the skin. Fourth,
s?v
the parabdic nature of the relationship between 6v and Pi expected
from remlar solution t h e o ~ yhas been confirmed for a number of dmgs.
Fgtfi, for vcharneleonicic"rsolutes (dmgs) such as saficry-lie acid (see
Fig, 3) that apparently can exhibit two differmt S i values, two differ-
s,v
ent log theoretical K i and, henee, two different log experiment&
P Q 9 Vcurves are generated.
In addition to the results obtained far single-component vehicles,
the rates of delivery. of 5-Buoroumcirl, (5-FU) and 6-mercagtopurhe
( 6-MP 1 from a b b a v component vehicle (oleic acid-pwpflene hTfycol)
.
were ailso determined (Table 2) Here, dthough the log experiment&
pQYV values for the delivery of 5-FU and 6-MP from oleic acid and
s,v
propylene glycol fell an the log theoretical Pi versus 1 5 curve, ~
8rv
none of the log expe~mentaf.Pi vdues for the mktures 4icX. In
each case, startling w i t h mhtures that were rich, in oleie a d d (21,221,
a s the st>lubi^lityof the dmg in each mkture hcreased, its experimen-
s,v
tal Pi v d u e decreased, Also, in each case, them was considerably
more damage to the s k h observed after treatment with the b k a q -
component soluthns than aAer treatment with either aS the 1~iTXgle
components (21,231, regardless of the cSv of the mixture. Thus, en-
hmced delivev of a drug by a mkture of vehicles may be primarny
the result of enhanced damage to the htegrity of the skin, rather
thm to some thermodynamic advantage,
In the investigations af the abnity of prodrugs to deliver their
respective parent dmgs h r n various vehicles through skin, there
was a good correlation between the calculated solubgity parameter of
a series of homologous pmdmgs and their abnity to deliver the par-
ent dmgs through hdrXess mouse skin (24-261, In addition, a ve-
hide effect on rates of delivew was observed that was directly at-
tebutable to solubsity phenomena. Generally, the results from
these articles were qualitatively simBas to the results of Ostrenga
( 9 ) that showed that there was a good correlation between mdar at-
traction constants (F values) and bloaetidty,
The most extensive studies of the effect of reI.llative solubBitles
of a homologous seAes of prodmgs on their abaities to deIiver a
parent d m g through s k h are the studies of ~6,s-bis-aeyloxymelhyl-
6-NIP (6,9-bis-6-MP) and ~ G - a e y l ~ x y r n e t h y l -(6-mono-6-MP)
~-~ pro-
drugs , The acety;loxymethylt through oetanoyloxymethyl abng w i t h
the pivdoyloxymethyl derivatives of both the bis- and mono-series
were synthesized, characte~zed,and the rates at which they deliv-
ered 6-MP through hairless mouse skins were measured. The solu-
Tabte 13.2 The Effect of Binary-Component Vehicles on the Delivery of Drugs Through Hairless Mouse
v s ,V 8 ,-v
Skin: Solubilities ( C i ) , Fluxes (Ji ), and Permeability Coefficients (Pi N
01
0
b
SoXventa 6
,
s
(mole ratios) (callcrn3)112
Ji 9
5,v
Derivative, mg/cm2h x lo3 log pi
R = /vehicle Solubilitya ( 2 SD) em /k
Table 13.3 (Continued)
-
a;?vpb
s,v
Derivative, mglcm2h x 103 log Pi ,
R = /vehicle Solubilitya ( 5 SD) cmlh
through hairless mouse skin (26). The results are very similar to
those found for the acyloxymethyl-type prodrugs of 6-NIP, Because
of stability consideration, these N-Mannick base-type pmdmgs were
tested in only one aprotic vehicle-isopmpyl myristate. A direct
s,v
correlation between log experimental Pi and calculated 6i was ob-
served. Again, the more lipophilie the prodrug (lower 61 value and
Parameters ( ) of
the Fluxes IAn .. rAT.,.r of 6-~vlei'CaJptc•puri:rte
( 6-MP) (PG)
Calculated
on the other. The first paper describes the effect of afkmoic acids
on the delivery of theophylline through human skin, To separate
the purely thermodynamjc effect of the dkanoic aeid vehicles ("'push
effectw resulting from excess free e n e r m ) from the abiXitg of the
Efilkanaic acids to solubaize the penetrant in, the skin ("pull e f f s t T t )
s,v( 1) s,v(2)
a ratio of pmtition coefficients ICi /Ki cztfculated fmm
r e p l a r mlution thwry ( E q . 6) was determined for the delivev af
theophyllhe from the two dkanoic acids b e h g compared. Because
s,v s,v
Pi = Ki .
constant (from Eq 13) , if the experimentally determhed
s , v ( l ) s,v(2)
ratio Pi /Pi was equal to the theoretically determined ratio
s , v ( l ) s,vC2)
Ki IICi then only thermodynamlcaXf.ygenerated push effect8
were responsible for the flux of thwphyllhe from the two solvents,
On the other hand, for example, if the vehicle v( 1) affmted 6& by
m&hg it more polar, then the experbentd ratio would be larger than
.
the theoretical ratio Conversely, if vehicle v( 22 is the one that
causes gs to become more polar for that diffusion exper;irnent, then the
experimental ratio would be smdler than the theoretie81 ratio.
sI
The experimentila results showed that the P i Tor the delivery
of theophyllhe from propionic , hexanoic , octanoic , and a mhture of
wtanoiie and hexmoic acids was inversely related to the 6 , for the
8 l v'
acicE, that is, Pi decreases a s cSv hcreased and bmane more like 6i,
8 v
However, the Pi for the delkery of theaphylline from mhtures of
a m o i c acids c o n t a h h g propionic acid was much higher than expect-
ed based on considerations of the mkturefs solubility paritmeter, At-
s , v ( l ) s,v(2)
so, the experirnentdly determined ratio of Pj /Pi in any
eomparimn of the performance of twa dkanoie aeidi was alwlzgrs smaller
s , v ( l ) s,v(22
than the theoreticdly determined ratio of Ki /Kg if propionic
acid was v(2) and always larger if propionic acid was v( 1). There-
fore, it was concluded that the flux of thwphylline from shgle- or
binary-component mktures of alkanaic acid vehicles was caused by the
thermodynamic push effect, if pr~pionicacid was not included but, if'
pmpiionic aeid was isleluded, then an additiond pull effect by. the ve-
hicle was observed,
The second of the tvvo most recent articles by Cohen and co-
workers describes the effect of alkanoic a d d s on the de1iver;y of
a d e n o ~ b ethrough human s k h , For mktures of hexanaic and pro-
s,v
piionic acids a bell-shaped dependence of observed In Pi on vol-
ume fraction of hexanoic acid was observed. This dependence was
attributed to the same two opposing uadables just identified, fn-
creases in the volume fractbn of hexanoic acid in a vehicle caused
an bcrease in the excess Tree e n e r m of adenosine 31that vehicle
~llnc3,hence, an increased push effect, Qn the other hand, in-
creases in the volume fraction of propionic acid in a vehicle caused
an increase in the flux of propionic acid from that vehicle and,
hence, an increase in the pull effect. A similar result was obtained
for mixtures of propionic acid in isopropyl myristate,
The enhancement caused by the pull effect of propionic acid in
the hexanoic-propionic acid mixtures was determined by first cdeu-
S,V(X) s , v ( ~ )
lating the corresponding ratio of theoretical Ki /Ki and
s , v ( l ) s,v(2)
equating it to a ratio of theoretical Pi /PI ( E q . 13) derived
s,v(l)
from regdar solution theory. The ratios of theoretical Pi 1
Pis*v'2) values were normalized by assuming that the flux of adenosine
from pure hexanoic acid was caused only by a push effect, therefore,
s,v(2) s,v(2)
Ki and, hence, Pi in each calculation is for hexanoie acid.
8,V
The product of the experimental Pi for hexanoic acid and the theo-
ssv( 1) s,v(2)
retical PI /Pi was assumed to give a value for the contribu-
S,V
tion to the total or experimental permeability coefficient Pi,total only
s,v S,V s,v
because of the push effect, or Pi, push. The ratio of Pi,total/Pi,push
was taken as the enhancement factor from the pull effect. The largest
enhancement factor observed was about 6.4 for a mixture containing
0.4 volume fraction of hexanoic acid.
IV. CONCLUSION
REFERENCES
t , INTRODUCTION
The cell type that is most respons.ible for the organizd epider-
mis is the keratinocyte. Other cells such as melanocytes, Langer-
han cells, and macrophages are also present, but they are not in-
volved in forming the anatomical barrier that is characteristic of the
epidermis. The basal cells, which are the only keratinocytes capa-
ble of pmliferation, form a single layer above the dermis. This
layer is connected to the underlying dermis by the basement mem-
brane. Basd cefls differentiate to spinous cetts which form multi-
layers immediately above the basal layer. Keratinocytes in the next
stage sf differentiation are called granular cells because they contain
keratohydin vanules , an important structure in the formation of
keratin. The cells at the ternrind stage of differentiation are the
cornifled cells that form the outermost desquamated l ~ y e r sof the
epidermis. ft is beEeved that tonofsarnents and keratohydin gran-
ules combine in same way to farm the keratin that characterizes the
terminally differentiated keratinocyte. Another cfraractedstic unique
to the epidermis is the formation of desmasomes-structures that
form very. tenacious bonds between the cells in various layers, New
idesrnosomes are bonds between the bas& keratinocyte and the bass-
ment membrane,
Therefore, any chemical that contacts the epidermis is presented
with all of the cellular components of the epidermis, We have devel-
oped a difff-lrentiated keratinocyte culture system that lFsrms a strat-
ified squamous epithelium with morphols@cal. and biochemicd markers
like those observed in an epidermis &med in situ, Keratinocytes
are obtdned as p ~ m a r yisolates (to retain as much of the od@n&
metabolic actidty as possible), seeded on an apgropriste substratum,
allowed to attach and proliferate, then, raised to the air-medium
interface to stimulate further growth and differentiation, This meth -
od creates a microenvironment simgar to that in which a nsrmd epi-
dermis develops, A s a msult , the cells differentiate and farm a
homeostatic relationship with each ather that results .in a tissue that
would be expected So react to a xexliobiotic applied to itits surface in
a manner simBar to the epidermis. We have &so developed methods
of applying test substances to the surface sf this keratinocyte cul-
ture system and have obtdned. data concerning the effects that eer-
tain classes of chemicals have on its metabolism of macmmdecules .
Epidermal cultures of this type have been established using both
rat a d human basal keratinocytes isolated from epidermal fragments,
The system currently used for studying the effects of xenobiotics
applied to the surface of the eulture has as its substratum a corn-
mereiafly a v ~ l a b l enylon membrane, rather than biolo&cal materids
such as collagen.
The early, h i t i d molecular events that t&e glace after topical
application of bis( 6 -ch'foroethyl)sulfide (B CES) to the surface of the
ezllture has been identiffed using this culture system,
2. Collagen Substrata
Various kinds of collagen gels were used as substrata and the re-
sulting attachment and proliferation of keratinoeytes after appropd-
ate seeding observed, The geIs were usually formed in the bottom
of plastic petri dishes and the cells seeded on the surface of the
gels in suffident medium to completely cover the cells.
When rat keratinocytes were seeded onto Vitrogen 100 bovine
dermal collagen, which is composed of 95% type I and 5% type 11 col-
lagens, firm attachment to this substratum resulted (24). However,
3 days after seeding, growth resulted in clusters of cells rather
than in a uniform proliferation over the substratum surface as seen
with plastic substrata. At this time, holes were also produced in
the collagen gel, and the surface began to dissolve. If the culture
was incubated longer, the entire! collagen Iayer disappeared, and
the cells were seen growing directly on the plastic surface of the
contdner.
When rat tail collagen was used as a substratum for the growth
of rat keratinocytes, attachment was observed 6 h r after plating,
compared with 18 to 24 h r on commerciaI plastic vessels (24). By
24 h r , 75% of the surface of this substratum was covered with kera-
tinocytes that had spread and proliferated. A confluent monolayer
was formed in 4 days. However, gels formed from this collagen are
very soft and difficult to handle.
Gels formed from mixtures of Vitrogen 100 and rat tail collagen
were &so investigated as appropriate substrata (24). A mixture
contdning equal amounts of the two types (1: 1, vol/vol) supported
attachment and growth of rat keratinocytes as well as rat tail col-
lagen alone, and also resulted in a firm substratum that would allow
relocation of the entire culture to the dr-medium interface (a pro-
cedure that will be described in a later section). This collagen
mixture also supported stratification and differentiation of submerged
keratinocytes to a greater extent than plastic substrata.
Pore size
Membrane (vm) Matehl Source
4000
a:
LJ
b-
Ld
3000
A
,A
.'.".
2
Lcl
ar 2080
4
3
0
U)
I000
.A
A
LLJ
C)
0
Control P4SQ TCM.I.50 RA-TF ACRO
P200 TCM2QO HA-TF H-T MEM
SUBSTRATUM
6000
a L J Untreated
it-I
-i ELXI LMN
3 rni-iFN
,"'.A
4000
f
LJ
Bf
4
I3,
0
fyl 3,000
\
V1
J
.A
tJ
C_)
Q
Control P450 TCM450 RA--IF" ACRO
P200 TCCM280 HA-TF H-T MEM
SUBSTRATUM
8000
cCJx Untreated
r--
u
mLMN
SO00 rnHFN
3
5%
LLJ
sr: 4000
Q
2
(CT
V1
3;
A
2000
W
0
Q
Control P450 TCM45Q R A - I F ACRO
P2QQ PCM200 HA---TI"' t-4-T MEM
Membranes {MI
Untreated eontrolb 1190 t 279 798 2 125 2.69 t 0.6 1.87 t 0.19
70%DMSO 658 L 138 564 t 110 1.27 t 0.03 1.42 t 0.29
40%ETOW 836 1 8 2 542 2 162 2.05 L 0.78 1.12 + 0.24
Heat killed 83 t 21 0.20 L 0.06
Figure 14.11 Incorporation of (A} [%I] uridine and (B) 11465 leu-
cine by epidermd cultures after a 30-min exposure to BCES in 70%
DNISQ, The cultures were pulse-labeled after removal of BCES fol-
lowed by selected periods of postexposure incubation. BCES con-
centrations (umo~ /crn2) were: 0.0 (solvent control) , open circles ;
0.01, open squares; 1.0, open triangIes; 10, closed circles; 50,
closed squares ; 200, closed triangles. Results are expressed as
percentage of untreated control t SD (n = 9).
corporation by both treated cultures and solvent controls 4-hr PO&-
exposure, pwbably caused by stimulation by the solvent. By 8-ktr
postexposure, all concentrations of B CE S included in this b w r e
caused inhibition of incorporation (P < 0.001). A concentration of
0.01 nmollcm2 caused inhibition for at least: 8-hr postexposure, but
mltures exposed to this dose showed a recovery of metabolic activ-
ity by 24 h r , when compared with the solvent control, However,
concentrations of 0.1 and 0.5 nmol lcmg of BCES resulted in signif -
icisnt hhibition of incorporation of radiolabeled thymidine as ob-
served at 4- and 24-hr postexposure. Fiwre IOB hclcldes csncen-
trations of 1 to 500 nmollcm2 of BCES and enxmmar;izes their effects
on thymidine incorgora~onby epidermal cultures, These concen-
trations caused sipificant inhibition at 4 h r and at 24 h r f P .:
O.001 for both observations), compareid with solvent contmls,
F i p r e 1 l A shows the results of [ 3 ~ uridine
1 incorporation by
cultures after exposure to BCES. Xncorporation of this precursor
after exposure to 0.01 nmatlemz of BCES appezrred to be stimulated,
when compared with solvent controls ( P < 0,OQI). This was evi-
dent 4 h r after exposure and remdned for the full observation peri-
.
od ( 24 h ~ ) A concentration of l. Q nmol / m a caused mild, but sf g-
nificmt , inhibition of hcorporation 24-hr postexposure, after an
initid stimulation of incovoration 4-hr postexposure. The higher
concent rations ( 10- 200 nrns~lcrnz) caused subst antis1 inhibition of
incorporation of this precursor, Similar results were o b t ~ n e din
the heorporation of radiolabeled leucine 'by lifted cultures after ex-
posure to BCES (see Fig. 11B). Exposure to 0.01 nmol /cm2 of
BCES stimulated the incorporation of radiolabeled feucine, compared
with solvent controls (P < O.001). Sipificant inhibition of leucine
incorporation occurred only after exposure to $0 and 200 nmol /ern2
of BCES.
The data in these experiments indicated that only at cornpara-
tively high concentrations of BCES was there sipincant interfer-
ence with RNA and protein metabolism. Inhibition in the uptake of
radiolabeled urYidine and leucine at these concentrations was proba-
bly a result of generalized cytotoxicity caused by massive DNA d-
kyXation and irreversible molecular destruction, Thymidine metabo-
lism, on the other hand, was affected at much lower concentrations.
Theee data suppor2: the condusion that DNA is the most important
target of BCES and are in agreement with observations of a number
of investigators. Crathorn and Roberts f 29) found that exposure
to a concentration causing DNA, damage and interference with repli-
eative abaity in BeLa cells did not bterfere with RNA or protein
synthesis. Lawley and Bmokes (30) observed that the cytotoxic
action. of sulfur mustard was not associated with the inhibition of
growth of BseherSchia colt as measured by RNA. or protein synthesis,
but with inhibition, of DNA, synthesis,
Epidermal Culture la Study Cutaneous Toxicity 295
Over the past 40 years, the ab%ity to control the delivery rate of
active agents to a predetermined site in the human body has been
one of the biggest challenges -nzel by continued innovative solu -
tions-that has faced the rnedicd profession and d m g i n d u s t q .
During this time, some areas of pharmweutical research have
been focused on the controlled delivery of systemic dmgs. SeveraZ
predictable and reliable systems were developed for sylstemic drugs
under the heading of transdermal delivery using the skin as a gor-
tail of" entm ( I ) , Transdermal patches, developed in the 1970s, im-
proved the delivery of drugs such a s nitroglyceAn and scopolamine,
resulting in better control of therapeutic doses, simfler dosage reg-
imens, and fewer side effects than the more traditional oral or par-
enteral administration of the same d m g s , In. general, these deliv-
ery systems have improved, the efficacy and safety of many d m g s
that now may be better administered through the skin.
Although transdermal ddivery systems can be efficient in sup-
plying drugs for systemic effects, they are not practical for eon-
trolling the delivery af materials whose final target is the skin it-
self.
Controlled release of drugs onto the epidermis, with assurance
that the d m g remains primafly. localized and does not enter the
sytstem in significant amounts, is an area of research that has only
recently been addressed with success. No efficient vehicles have
been developed for the controlled and ZocaBzed defivew sf dmgs
300 Nacht and Katz
into the stratum corneum and underlying skin layers. Yet, there
are many instances when epidermal localization of a drug is desira-
ble, but absarpt;ion beyond the epidermis is not.
Corticosteroids are a typical example, Although effective for
skin disorders, their topical application can result in sipificant
systemic &sorption; this may lead to unwanted side effects such as
adrenal suppression or interference with immune functions (2).
Similarly, with sunscreens it is necessary to maximize the amount
of time that the active ingredient is present on the skin surface or
within the outer layers of the epidermis whjile minimizing its trans-
epidermal penetration into the body,
Another problem with the application of topical drugs is that
many vehicles, sueh as ointments, often prove aesthetically unap-
geafing, Greasiness, stickiness, or even discolorations in clothing
can m&e dany wear unpleasant. Thfs frequently results in the pa-
tient" slack of compliance with tredment.
Many of these conventbnd vehicles require high concentrations
of active agents for effective therapy because of their law efficidency
ae delivery systems. As a consequence, irriitation sr allerdc re-
sponses can be elicited in a sipifiicant percentage of users. Qther
disadvantages af existing topical d m g formulations can be uncon-
trolled evaporation of the active ingredient, unpleasant odor, and
the potenthl hcampatibitity of one a r more drugs with each. other or
with the vekiide ,
Thus, the need exists for systems to maximize the amount of
time that an active ingredient, such as a sunscreen, is present,
either on the skin surface or within the epidermis, while minimizing
its transdermd penetration into the body.
Such a new system would pssibly increase the efficacy of topi-
cdly active agents while enhancing product safety,
The ~ierosponge@ polymeric microsphere system utliwely fulfjlls
these requirements, These tiny spongelike spheric& particles (Fig.
1) can entrap active ingredients and then release them onto the
skin over time and in response lo a trigger. They are bioXo@cally
inert, norrirritathg , nonmutagenic , nondllergenic , nontoxic, and
nonbiodegradable. They can extend pmduet stabnfty because of
their unique configuration and improve its aesthetic properties,
I. They are either fully miscible with the monomer mixture or ca-
pable of being made fully miscible by the addition of a minor
amount of a non -water- miscible solvent.
2. They are immiscible with water or, at most, only slightly solu-
ble in it.
3. They are inert ta the monomers and stable when in contact
with any polymerization catdyat used and when subjected to
conditions needed to induce polymerization (such as temperature
.
and radiation)
(Fl
Figure 15.1 (Cantinued)
these cases, an inert liquid, fully nnisdble with the monomers but
immiscible w&h water, is used during palymerisation to form the
pore network. The liquid that has served as a porogen and aecu-
pies the pares of "the formed particles can then be removed leadng
preformed dry micmspheres ,
Subsequently, placement of the functional substance inside the
reservoirs may be achieved by impregnation of the preformed dry
porous polymer particles aceording to techniques, such as contact
ab~osption,assisted, when necessaq , by sdvents to enhance the
absorption rate. The find product is thus prepared in two sequen-
tial steps. First, polymerization is performed using the substitute
p ~ r a g e n . This substila* iis then removed and, replaced by the
function& substmnce.
MatedetIs suitable as substitute porngens are liquid substaxlees
meeting the foregoing cr;Eteda and h a ~ n gthe further chmracte.I.istic
o f being readlily extracted from the particlesF pore network once
polymerization 1s complete, This covers a wide range of substances,
in particular inert, nonpolar o r e n i e salvents,
NacZlt and Katz
2 )***
MONOMER A MONOMER B
Step One: Selecting the Monomers Step Two: Making Monomer Chains
Step Five: Gathering Bunches of Polymeric Step Six: Binding the Bunches
Microspheres to Form a Polymeric
Microsponge", the
Basis of Command
Release" Tee hnology
Characteristic Vdue
Sample: CH-274
Total intrusion volume 1.22 cclg
A. Particle Size
Free- flowing powders with fine aesthetic attributes can be created
by tightly cantrolHng the dkmeter of the microspheres at the time
of polymerization. Typically, diameters between 5 and 300 vm are
feasible, Small particle-sized systems provide the feel of a super-
fine talcum powder, even when they contain substances such as
mineral oil at a 50% payload. Larger-sized particles can be pre-
pared according to the ultimate needs for the products. Particle
size &so has some influence on the release rate of the active ingre-
dient from the Microsponge. The smaller the particle size, the
slower the rate of release (Fig. 4).
A Pragvmmable Topical Delivery System
20
Time (min)
Figure 15. Y Effect of particle size on release rate, Release rates
of declczne from two Microsponge systems of different particle size:
1.4181mh (25 um. squares) and I, 8l%lmin (300 vm, circles) .
B, Pore Diameter and Volume
Pore volume (void volume) determines the amount of active ingredi-
ent that can be entrapped withh the microsphere, Conversely ,
pore dhmeter can have a siwificant effect on the rate of release of
the ingredient. Fimre 5 demonstrates the effect of pore diameter
on the volat2ization rate of menthol. Both pare volume and diameter
are dtal in conlmlling the intensity and duration of effectiveness of
the active ingredient. Pore diameter also affects the m3pation of
the active ingredient from the Microsponge into the vehicle in which
the mated& is dispersed, A s a result, the diiameter of the pores
can have direct impact on the stabirity of the final formulation.
C . Resiliency
By aftedng the d e p e e of monomer cross-linking at the time of po-
lymerization, resniency (dscoelastic properties) of the microsphere
can be modified to produce a beadlet that is softer or firmer accord-
ing to the needs of the find formulation, Cross-linking in excess
of 10%is efficient in mast ~ystems. This allows; ~uffieients t r e n o b
fsr the Mic~ospongeto retain its shape after some or all of the ac-
310 Nacht and Katz
tive ingredient has been removed from the pore network. Increased
cross-linking tends to slow down the rate of release (Fig. 6).
D. Monomer Compositian
Selection of the monomer is dictated bath by the characteristics of
the active ingredient ultimately to be entrapped and by the vehicle
into which it will be dispersed (Fig. 7). Polymers with varying
electrical charges or devees of hydropfiiXicity or lipophilicity may
be prepared to provide Bexibnity in the release of such materids as
lipids, humectants, moisturizers, sunscreens, vitamins, insect repel-
Eants , fragrances, and a variety of pharmacoXo~callyactive ingredi-
ents. Onee entrapped, these ingredients can be formulated into
virtually any product form: powders, gels, ointments, lotions,
creams, liquids.
10 20 30
Time fmin)
Figure 15.7 Release of decane from two Micmsponge systems of
different polymeric composition (MMAIEGDMA : Methyl-methacrylatel
Ethyleneglycoldimeth8cry1ate; St /DVB : Styrene/Divinylbenzene).
31 2 Nachl and Katz
B, Release on Cornnand
El y proper manipuf ation of the aforementioned programmable parame-
ters, Microsponges can be d e s i p e d to rdease @ven amounts of ac-
tive ingredients over time in response to one or more externaf trig-
gers,
1, Pressure Release
Like m ordinary sponge, Microsponge systems release fluid when
pressed or squeezed, '' "ereby replenishing the level of entrapped
f f
2, Temperature Change
Some entrapped materials, such as sunscreens and emsll%ents,can
be too .viscous at rmm temperature to now sgontmeously from the
Microsponge onto the skin. When warmed by the skin temperature,
the sun or other heat source, their dscosity rnay decrease, result -
ing -in an increased Row rate f Fig. 9).
3, Solubility
By taking into account the solubility of the entrapped ingredient,
the Iblicmsponge system can be? programmed to respond to water,
perspiration, or other solvents. For example, dry Microsponges
loaded with a water- soluble ingredient, such as &ntiperspirants or
antiseptics, will release that ingredient in the prcl;senee of water
(Fig. 1a), Release ean also be activated by diffusion, taking into
consideration the partition coefficient of the ingredient between the
Mierasponge and the outside system.
Vf l, SAFETY SUBSTANTIATEON
to!.ion
%vitk
Microsponges @
time (hr)
Figure 15.8 Comparison of human s k h emouieney induced by min-
eral oil in a Mierasponge system vs. microcapsules, Measured with
the gas bearing ellectradynamametes ( R e f , 5) ,
2 4 6
Time (hr)
Figure 15. f 0 Solvent-activated release: release of an antiperspi-
rant in the pmsenee of water from two different Micmsponge sys-
tems.
A Programmable TopieaZ Delivery System
Baseline 2 Week
Figure 15.12 In viva antimicrobial efficacy of benzayl peroAde la-
tion. (Lotion eontdnfng 5 percent BPQ entrapped Jin. a Miemsponge
system tested as descdbedi by Williamson and Kfigman (from Ref, 8).
% Emollient
a Not Rubbed
Gentle Rub
,
,Lotion
micraentrappc?d
with
active
I \ m---a Lotion with "free" active
1/4 112 1 2 3 4 5 6 7
%me After Application (Hours)
% Preference
Less oily 72 28
Less greasy during
application
Disappears faster
Less greasy afterfeel 78 22
Less residue on skin 67 33
A. Sunscreens
High levels of sunscreens can be utilized for longer-lasting product
efficacy to increase protection against sunburn and sun-rel&ted In-
judes (aging and skin cancer). Even at elevated sunscreen con-
centrations, oiliness, irdtancy , and sensitization could be reduced.
A, Programmable Topical Delz'vey System
6. Skin Depigmentation
Skin bleaches, such as hydroquinone, can be stabilized and protect-
ed from oxidtnlion as eddenced by the lack of diacalaration of the
product. This results in impmved efficacy and aesthetic appeal.
E. Antipruritics
Extended actidty is achieved for anti-itching compounds and local
anesthetics used to treat prmr;itus caused by dry skin, eczema,
hemorrhoids, or poison. ivy.
F, Antidandruff
The unpleasant adars af zinc pysjithione and selenium sulfide, as
well as other antidandnB ingredients, are reduced, Irfitation can
afss be lowered, whereas safety and eflictlcy. are extended,
Prolonged actidty with less irdtarxq are okztdned with these active
agents, accompanied By a reduction in odor and peasiness of the
product.
In each of the foregoing cases, the final pmduct can be de-
s i p e d to give distinct and perceivable benefits to the user that
were not p r e ~ o u s l yfeasible because of formulation restr;ictions.
Such statements as '"ease?less," "all-day relief ," and ""new pkesing
frapancef' can now be achievable Consumer beneAts for these types
of products,
Nacht and Katz
REFERENCES
Vesicle Captured
diameter Number of volume Percent
Acronym (vm> lamellae (ml fg) entrappeda
-
E 8.3
0.02 0.10 1.00
Vesicle Diameter (lurn)
B. Process Selection
Liposomes have their own unique problems of scaleup and process
control. These are concerned with adequate supplies of phospholi-
pid raw materials, lipid hydration to achieve the desired capture
volume and particle size, use of pharmaceutically acceptable organic
solvents, removal of solvents (organic phase addition -removal proc-
esses; 22) or detergent (detergent dialysis; 20), and clean up of
unentrapped drug.
Selection of a process should be based on the solubility proper-
ties of the drug in aqueous solution and appropriate solvents, the
liposome / buffer partition coefficient, and the required percantap
entrapment or incorporation.
Topical liposome formulations have the advantage of not requir-
ing the stringent particle size specifications of parenteral products.
If liposome size or heteropneity does not affect in viva performance,
the specifications can be kept quite broad and the process simplified
considerably.
C, Characterization
The characterization of bulk properties of liposome formulations, such
as color, pH, rheolosry , are very similar ta those commonIy employed
for other types of dispersions and suspensions.
When formulating emu'Isions and microemulsions, considerable ef-
fort is placed on making appropriate blends of oil and surfactant-
cosurfactants to achieve a stable dispersion. With phospholipid-
based liposomes that spontaneously form a stable structure, more
emphasis is placed on characterizing liposome-specific properties.
Determination of vesicle size, captured volume, percent entrap-
ment and percent incorporation have already been discussed. One
other prope&y unique to liposomes is the active ingredient release
rate or efflux rate, which has important Implications for the per-
formance and stability of a formulation, It is most simply measured
by monitoring the change in percent entrapment or percent incor-
poration of a formuIation over time under a given set of conditions
(temperature, other excipients , ionic strength, pRsence of biologi-
cal fluids). Average liposome size of a formulation is an important
factor in determining the release rate of small nonelectrolytes (23).
However, mathematical model fitting from experimental determinations
of glucose release indicates that the degree of liposome size hetero-
geneity (polydispersity) does not affect the release rate s i p i f i c m t -
XY ( 2 4 ) .
D. Antimicrobial Preservatives
Topical products are usually preserved by the addition of one or
more antimicrobial preservatives. Preservatives are intrinsicalIy no
different from an active ingredient, which is to say they can be en-
trapped in Xiposomes or incorporated into the bilayer or interfacial
region. The end result is that microbial deterrence at a given con-
centration will be different from that of a simiIar aqueous solution
without liposomes ,
For instance, the commonly used parabens are inactivated by
liposomes as a consequence of partitioning into the bilayer (25).
The extent of inactivation will be a function of partition coefficient
and the lipid concentration. Consequently, the total quantity of
lipophilic preservative must be increased to achieve a satisfactory
germicidal effect. If the "free" concentration (Cfree) required for
the germicidal effect of preservatfve P , the formulation% liposome-
water partition coefficient (Kp) and the total lipid concentration are
known, the totaI concentration of preservative (Ctotal) required in
the formulation can be cdculated from a modification of Bean's
equation (26.) :
E. Stability
Formulation stability e m be s u b d i ~ d e dinto chemical, phy sicat, and
f ipossome- speddc characteristics.
1. Chemical
For ehemica3. stability of phospholipid-based liposomes, lipid permi-
dation and fatty acyl-ester hydrolysis are the primav concerns.
Phospholipid peroxidation is accelerated by oxygen, transition metal
ions, elevated temperature, and light. ft is usually controlled by
the addition of chelating agents to bind iron salts (271, antioxidants
such as a-tocopherol ( 2 8 ) , by using phospholipid that has saturated
or mono-unsaturated fatty acids, a s combinations, Cholesterol ,
anather primary component of lipasome formulations, is mininndly af-
fected by oxidation in liposomes because the phoaphatidylchslines
are very effective inhibitors ( 29) .
Fatty acid ester hydralysis is a time-, temperature-, and pH-
dependent process, Hydrolysis results in the formation of lysopkos-
phatidylcholine and free fatty add, Depending on the active ingre-
dient, changes in liposorne size, efflux rate, parcutaneous a b m q -
Lion, and in d v s performance may be affected. Fmkjer et ale (30)
have published a v e v excellent, systematic study of the chemiczil,
phgsicd , and release rate stabifity of plnosphatibylehollne-choles-
terol Eposomes. The pseudo-first-order rate constants for phos-
phatidylclholine hydrolysis at 10°G showed a pH minimum at 6.5.
Only 6% fiydroly sis of distearylpkosphatidg1~holinc!to ly saphaspha-
tidylcholine was observed aAer 300 days at pH 6.5 and 2 s 0 6 ,
Thus, ester hydrolysis can be minimized by formufating in the vi-
cinity of pH 6,5.
2, Physical
Changes in the bulk physical properties of a liposorne preparation
usually reflect changes in vesicle agpegation or hereasing fiposome
size.
Cross aggregation of liposomes can be observed naerascoplcafly
as "creaming?' or floccula~onof the ori@na"lly homogeneous disper-
sion. Occasionally, controlled flocculation of liposomes can be put
to good purpose in eontrolling drug suspensions f 31), but, sFor the
most past, aggregation is a nuisance from the rriewpoint of cosmetic
acceptabiility., and it may have undesirable effects on formulation
properlies.
Vesicle aggregation can be controlled by imparting an electro-
.
static charge to the liposomes Typicdly , negatively charged lipids,
such 8s pho~phatidyfglycemfand phosphatidflsefine, are added in
small quantities (about 2%- 10%by weight of total, phospholipid).
The addition of amphipaths, such as steary-larnine or eetgltrimethyl-
ammonium bromide, impart a positive surface charge to liposomes ,
but cationic lipid toxicity usually precludes use of positively
charged lipids in. pharmaeeutica?l products, A chelating agent, such
a s ethylenediaminetetraacetic acid (EBTA) ( 0.01%-O, 1% w / v) should
d s o be included to prevent divalent metal cation induced a g o e g a -
tion.
Growth of mean Xiposome size can &SO be inhibited by imparting
an efectrostatic charge, the inclusion of trace amounts of chelating
agent and by avoiding formulations based on SUVs, Because the
high radius of curvature in StlVs cause them to be thermodgnamical-
ly sts&ned, such. preparations show net vesicle p o w t h , which is
particularly rapid when the formulation is incubated a t , o r thermally
cycled through, the clcystdline-liquid c q s t d f i n e phase transition
of the lipid (32).
3, Liposome-Speeiffc Parameters
Changes in the chemical or physic& stability of a liposome h m u l a -
tion wB1 oftentimes be refieeted by changes in the liposome-specific
parameters. Appreciable changes in mean particle size may &feet
the release rate of active fngredient ( 2 3 ) . The captured volume
and the percentage entrapment will be changed if vesicle dkrneter
increases or decreases.
Thermodynamic equilib~umwill drive the flux of entrapped ma-
te&& into the continuous aqueous phase of a liposome vehicle.
Therefore, the desired percent entrapment shelf life of a liposorns
formulation may be difficult to achieve if unentrapped materid must
be removed and the release rate in the storage buffer is not mark-
edly different from that of the in viva release rate, By example,
consider a Eiposorne 6i3rmulatlion af a water-soluble d m g that is de-
s i p e d to release half its contents within 2 days of application to an
ophthdmic surface. A cleaned up formulation must have a storage
buffer release rate at least three orders of mamitude slower than
the in viva rate to mhieve 2 years of percent entrapment stabaitg.
Liposorne-Based Vehicles f"or Topical Deltver~y
A. Ophthafmie Studies
In 1981, Smolfn et aJ, (33) used free and liposorne-atrapped idoxu-
ridine, an antiurird agent, to study the treatment of herpes sirnpXex:
ksratitis in rabbits, An aqueous solutlian of 0.1% (wIv) idoxu~dine,
0. I % idoxuddine in phosjphatidylcholine-phosphatfdylglyeerol --eho-
lesterol LUV-OLV Xiposornes (35%entrapped dmg) , or the liposome
placebo, were applied three times d a y to rabbits that had been in-
fected p r e ~ ~ ~by s l corned
y abrasion and application sf a titer of
d r u s . Eiposome -entrapped idoxtl~dinereduced the size of viral-
induced corneal defects sipificantly, compared with free drug or
liposome placebo.
Lee and colleagues (3$,35) investigated the delivery of free and
liposorne-entrappa inulin and epinephrine XICI , bath, water-soluble
model compounds of substantially different permeabaity, In d t r o
release expeAments revealed that epinephrine HCI crossed liposome
baayers readily; 50% af osgi-nally- entrapped epinephrine HCX was
released from MLVs within 1 h r , whereas more than 90%of the en-
trapped inulin remained entrapped after 4 hr, In vivo studies In a
rabbit model Indicated that liposomes did not slow down ep-inephdne
HCl removal from tear f'fufd, Fufihermore, 45 min aAer instalation
of the liposome vehicle, epinephdne concentrations in eye tissues
(conjunctiva, cornea, aqueous humor, and iris-cniaw body) were
significantly- lower than aqueous solutions of epinephrine HC1.
However, when using Xiposome-entrapped fnulin in viva, marker
levels were increased many-fold over aqueous soXut.jlans In the con-
junctiva, cornea, and iris-cniary body. Inugn could not be de-
tected art; dl in the aqueous humor when entrapped in liposomes,
Thus, depending on the drug, Iiposome entrapment changed marker
.
dist ribution s i ~ i f i c a n t l y Liposonres were able ta increase localized
cancentrations of a water-soluble marker with low membrane perme-
abnity, perhaps by adhering to ocular tissues.
These results were vefibed by Ahrned and Patton (36) who &so
found that transcorneal u p t k e of MLV liposome-entrapped inuIin
liposomes is suppressed and noncorneal uptake was enhanced when
compared with an aqueous inuBn sdution, Aliquols of free or Xip-
osornal-entrapped tritiated hulin were instnled in rabbit eyes and
assayed for radiolabel 20 min later. Increased inulin levels were
found in the cornea and irls-c3iaw body when Eposorne-entrapped
inulin was instailled, Whereas free inulin was able to penetrate into
the aqueous humor to some smd1 d e p e e , liposome-entrapped inulin
was not detectable in the aquews humor. The authors suggested
that liposonle dosage forms can be used to promote d r u g absorption
selectively to noncomed oeular tissues,
In subsequent studies, Lee and co-workers found that Iiposome-
entrapped inulin u p t a e into intraocular tissues was indeed depend -
ent on adhesion of Bposomes to corneal epithelium ( 3 7 ) . The reten-
tion of MLV (phospatidylckroline-cholesterol) liposomes in tears, and
their interaction with corned and conjunctival, sur.flaees was studied
using tritietted-inulin (5%, w l v ) as a marker for entrtapped contents.
The rate constant for dearing these neutral MLVs was only 30%
slower than that compared with free inulin, and this rate constant
was insensitive to volume of instzed dose* Inulin-loaded liposornes
competed with empty gposomes for binding sites in the corneal and
eonjunetival surfaces, Pretreatment with empty MLVs reduced in-
ulin-loaded liposome binding by greater than 75% for both surfaces.
Neve&heless, liposomes did not kind 8~d;X-yto these surfaces; dm-
ing removed them effectively, The authors suggest that dssorption
into tear fluid is the major pathway of clearance,
Another factor inETuencing the reduced concentrations of inulin,
in various eye tissues was the slow liiposomzrl release rate of inulin
(about I% per hour in this study]. Inulin is a large-moleeular-
weight (about 5000 MW), water-soluble compound and is not expected
to have a rapid e f f l u x rate.
.
Lee et al f 37) concluded that topicdly applied, neutral MLV
liposomes are removed from the corned surface by tear drainage.
The extent of entrapped inulin absorption seemed to be controlled
by the number of adsorbed liposomes 8s opposed to the number of
instilled liposornes, Adsorption was not sufficiently strong, and ef-
flux from liposomes was too slow, La sustdn inulin concentration in
intraacular tissues. Thus, improved ocular cLte1ive;rzy. of water-soluble
entrapped materials requires enhanced liposorne-binding affinity and
a rapid release rate,
K r o h and ca-workers attacked the problem of i m p r o ~ n goeular
retention of liposomes by pretreating isolated rabbit corneas with
wheat germ agglutinin (WGA) and instnling p'bosphatidylcholine-
m h e d b r d n ganglioside MLV or SUV liposomes (38). Were, lectin-
mediated adhesion of liposomes enhanced binding to rabbit corneal
epithelium 2.5-fold. Transcorned ilux of radiolabeled carbachol en-
trapped in such ganglios-ide-containing liposomes was studied in an
in e t r o model using excised corneas with simulated "teefar Bow."
Unentrapped earbachol was not removed from the Iipossme formufa-
tion. Thirty minutes after instnlatlian there were no significant dif-
ferences in carbachol actidt y in cornea or transcorned receiver
.
compartment fir liposome-entrapped or free carbachol However,
90 min after instillation the liposome vehicle was able to m&nl;&n
significantly higher earbachol concentrations in the cornea and in
the transcorned receiver compartment. The authors propose that
the liposome vehiclet@aabjlity to sustain drug levels at 90 m h in the
receiver compartment was due -lo the continuing presence of drug at
the corneal surface.
Schaeffer and Krohn ( 3 9 ) studied in d t r o lipasorne-corned
binding mediated by electrostatic adsorption. Stearylarnine or di-
cetyl phosphate was used to modify the net charge of phosphaedyl-
ehoEne-cholesterol Iiposornes. Rank ordedng of liposome binding
to excirsed rabbit corneas showed significantly greater affinity
fabout twofold) of positively charged liposomes compared with neg-
atively charged Iiposornes, In turn, negatively charged lipasomes
bound twofold greater than neutrtzf, liposomes.
Recently, Gua and co-workers (40) have been able to demon-
strate that the in vivo retention of radioiodinated liposomes in rab-
bit eyes is increased many-fold by using MLVs containing positivr?ly
charged cholesteryl esters. They observed that liposome binding in
the eye was saturable, and the half-life of Xiposorne retention cauld
be manipulated by varying the charged cholesteryl es"tr/phosphdi-
pid ratio, the length of the spacer arm separating amino function
from the lipid head group, and lipid phase transition (fiuidi_Ly).
Schaeffer and Krohn (39) investigated the in d t r o transearned
flux of the water-soluble drug penienlin G in MLV or LUV lfposome
vehicles of quditatively different electrostatic charge, A11 liposome
vehicles tested showed significantly improved, drug delivery over an
aqueous penicalln G solution. Penicalin G flux: wross; rabbit cor-
neas J h r after inst3lation exhibited the same charge-dependent
rank ordering as lipid binding. Free d m g added to preformed,
empty liposomes did not show increased. d m g flux, suggesting that
d m g encap;rsulatian and liposome binding was a prerequisite for i m -
proved delivery,
The transcorneal Elux of the lipaphnie dmg hdozrole incorporat-
ed in neutral phosatidylcholine liposomes or in a palysarbale 86 die-
pession was evaluated in 6 v o wMh rats (39). A t f. h r after instiXla-
t.ion, a 1.Q rng indoxole per miXX9iter liposome vehicle preparation
prodded. the same drrrg concentrations in rat aqueous hurnor (about
20 vglml) as a 10.0 mg indoxole per mnlniter polysorbale 80 vehicle,
Furthermore, Schaeffer and Krohn discovered that the Iiposome ve-
hicle caused none of the histoia@cal changes induced by polysorbate
80 ocular toxicity (unpublished results). Thus;, in addition to im-
pmving the solubility of the anti-inflammatory agent indoxale, ve-
hicle safety was improved.
Shiota et &I , reported a sever-&-fold improvement of lfpo-
(41)
phnic dtamin A deliveq in viva. Increased specific a c t i ~ t y in
. var-
ious rabbit eye tissues persisted for up to 8 h r after instalation of
dtamin A in positively charged liposomes, when compared with 14-
tamin A in a simple vegetaible oil vehicle.
The consistent conclusion from ophthallmic studies is that lipo-
somes provlide improved local concentrations of active ingredients
proTajlded the compound is lipophaic, or it is water-soluble but not
very membrane-permeabXe, The indoxole study &so suggests that a
liposome vehicle may improve safety by replacing other more toxic
excipients used for drug solub;Ilizatian, The foregoing studks ali~so
indicate that liposornss improve transcornetrl Rux of lipophnic mole-
cules and for, at least, some water-soluble moleales such as peni-
cillin G.
8. Dermal Studies
The first results of topicdly applied liposomes were reported by
Mezei and Gulasekharam (42) in which tdamclindone acetonide en-
trapped within, ~pdmitoylpE.losphatidylch01i~~.e-cholestero1 MLVs re-
sulted in sipificantly Increased d m g levels in rabbit dermis and
epidermis, reduced blmd and reduced u ~ n concentrations,
e Xn a
series of reports (43-45) Mezei and co-workers used MLVs to dem-
onstrate the in viva transdermal delivergr of cortieosteroid drugs,
such as triamcinolone acetonide, tAamcino3lone acetanide 2 I-pdmitate ,
and progesterone, in rabbits and guinea pigs. Controls were con-
ventionail dosage forms such as ointments, although it is not clear
that control and liposome preparations had similar thermodynamic ae-
t i d t y to drive steroid partitioning out of the vehicle. Prepariltions
were compared tissue by tissue, For tdantcindone acetonide and
its more lipophilie pdmitate ester deAvative, liposorne preparations
were reported to consistently @ve simificantly higher epidermd and
dermal d m g concentrations, no significant differences in systemic
tissues (except for blood, which was significantly lower), and re-
duced u ~ n w yexcretion. Mezei a r m e s that this is indicative of
selective tsiarncinolone acetonide delirrev to skin tissues.
This generalization was tilsa reported for econazole base and
econazofe nitrlrLe (44). For this active ingredient, all skin tissues
showed increased concentrations of radiolabeled. drrxg, The pathrn
of in viva disposition of seven different ecanazole-liposorne prepa-
rations is simBar to triamcinolone acetonide; but; different lipid corn-
positions delivered different quantities, Multiple dosing wfth control
and econazole-loaded Iiposome preparations increased econazole ae-
cumulatfon silf"ni;Flcantly when lipssome vehicles were applied,
However, progesterone-loacled Xiposorne preparations skawed in-
creased concentrations in all skin tissues, and with the exception of
Liposome-Based Vehicles for Topical Delivery 341
freeze-frac-
ma~n.eutcresonance
mixtures as in a tran-
of pure phOSI>h~l
of pure
Nishihata's
ab-
0. Conclusions
The results of many studies may be summarized the
servations that
V. FUTURE DIRECTIONS
A. Basic Research
Tapiedly applied, dmg-loaded fiposomes can substantially imprave
d m g loading, drug deliveq, and sustained release, thereby offer-
ing cfearcut advantages over traditional dosage forms. These ad-
vantages are particularly eddemt for the more lipaphi3ic aetive in-
gredients, in which elevated locd concentrations are conaistently
demonstrated and that for, at least, some dmgs reduces the sys-
temic concentrations, Because there is no evcidence for liposome
penetration through intact topical surfaces, such as the stratum
corneum, it is perfectly lo@cal that water-soluble, membrane-imper-
meant active inpedients face an additional barder to increased bio-
avaSabi2ity when entrapped within liposomes , For topical surfaces,
the stmeturd integrity of which Is not compromised, a second gen-
eration of lipssome vehicles vvnl have to be developed,
However, water-soluble drugs entrapped in convention& lipo-
somes may play an extremely useful mle in treating conditions ;In
which the i n t e e t y of the stratum comeurn or other topied surface
has been damaged. The study by Smctlin and co-workers on vciral
infection. of abraded rabbit corneas is espedslly encoura@ng (33).
Liposome-Based Vehicles for Topical Delivery 345
ACKNOWLEDGMENTS
I would like to thank Drs. R, Abra, L. Guo, and IM. Woodle for a
critical reading of this manuscript and for their many useful sug-
gestions.
REFERENCES
.
F , Szoka , J r , and P) , Papahadjopoulos , Proe, Matl, Aead,
Sei, , U S A , 75: 4184, 19.78.
9, M. N, Kremer, IM, W. J, v. d , Esker, 6 , Pathmamanohoran,
and P. H. Wiersema, Biaehemistry, 17:3932, 1971,
W . G. Enoeh, and P, Stdttmatter, Proe, Na21. Acad, Set, ,
USA, 76: 145, 1919.
3 . Pkriffppot , S . Mutaftschiev, and J. P. Liautard, Biochim.
B t o p k y ~ ,Acea, 734:137, 1983.
R, 111, Sehwendener , Cancer Drug B e l i v e v , 3: 123, 1986,
Q , IZurnBuehl, and H. C. Weder, Blochim, B h p h y s . Aeta, 640:
252, 1981.
R. 9. XClimchak, and R . P , Leuk, Biopharm. Manuf"act, , 1: 1 8 ,
1988.
R , H Guy, S , Wadgraft, M e J. Taylor, and I , W, KelXaway,
J, Pharm, Pharmaeot, , 35: 12, 1983.
Z. T. Chowhan, T. Yotsuyanae, and W. L, Higuchi, Biaefiim,
Biophys. Aeta, 266: 320, 1972.
M. Kornatsu, K . Hig8ki, W e Okamoto, M, Miyagawa, M. Hashida,
and H, S e z e i , Chem, Pharm. Bull, 34: 3415, 1986,
N, S . Bean, and S M, Hernan-tickahl, J , Pharm. Pharmacol, ,
I
f6:58t, 1964.
3, EM. C. Gutteddge, FEBS Lett., 17%:245, 1984,
6 . A , Hunt, and S , Tssang, fnt. 6 , Pkarm., 8: 101, 1981.
N, I>. Weiner, W. C . Bmning, and A . Felmeister, J, Pharm,
Sci, , 62: 1202, 1973,
S. Frokjes, E . L. Hjosth, and 0. Worts, in Optimization of
Drug DeZilaeq, Alfred Benzon Symposium 17 (HeBundgailrd,
A. B. Hanson, and W. Kofod, eds.). Munkagaard, Coperr-
h a p n , p. 384, 1982,
8.-L. Law, W. -YeLo, and G,-W, Teh, J . Pharm, S C ~,, 76: 545,
1987.
.
F , 2. Martin, and R , C MaeDonaXd, cSr"oche?mistv, 15: 321,
1976.
G , Smolin, M. Qkurnoto, S. F d e r , and E), Condon, Am, d .
Qphlhalma!, 91: 220, 19.81.
.
It. E. Stratford, J r , , B, C Yang , M , A. Redell, and V, H,
L e Lee, Int, J, Pharm, , 13: 263, 1983.
E . S t r a t f o ~ d ,J r . , B. 43. Yang, M. A , Redell, and V . W ,
L , Lee, Curr, Eye Res, , 2~317, 1983,
I , Ahmed, and T. F. Patton, I n t , J. PPlarm., 343163, 1986,
V, H. L. Lee, K, A . Takernoto, and D. S. limoto, Curr, Eye
Res. , 3: 585, 1984,
H. E . Schaeffer, J, M, Breitfeller, and D L L . Krohn, Invest*
OphlhaEmsE, Vise Set, , 23: 530, 1982.
EX. E . Sehaeffer, and D. E , KroEm, Invest, OphthaZmol, V i s ,
Sci, , 22: 220, 1982,
L. 8. S , Guo, 6 . T, Redemann, and R . Radhakritshnan, In-
varst, Qphthafmol, V i s , Set. , 28: 72a, 1987.
R . Shiiola, A. Sarris, R. Radkit&rlshnan, and L. S. S, Guo,
Invest, OphthalmoE, Vis. Sei, , 28: 159a, 1981,
M, Mezei, and 17. Gulasekharlam , L i f e Scr". , 26: 1473, 1980,
M , Mezei, and V. Gulasekharam , J. Pharm, PharmacoZ, , 34: $73,
1982,
M, Mezei, in Topics in Pharmaceutical Sciences (D, D *
.
Breimer and P Speiiser , eds .I, Elseder Science Putbliahers ,
Amsterdam, p. 345, 1985.
El. B. Harsanyi, 3, C . HBcf-rie, and M, Mezei, J. D e n t , Res,,
65: 1133, 1986,
M. G . Ganesan, N, D, Wetner, C, L. PXynn, and N. F", H,
No, lnt, J. Pharm., 20: 139, 1984,
N . F. H. N o , M. G. Canesarr, M. D. Weiner, and G. L. Flynn,
J. ConlmlZed Release, 2~61,1885.
K . 13, Ertel, and 3. T, Carstensen, Int. I , P h e w , , 3 4 ~ 1 7 9 ,
1986.
TI, M, Knepp, fit. S , Hbz, F. C . Szoka, Jr., and XE. R . Guy,
J , Controlled Release, 5: 211, 1981,
3, Lasch, and W , Wohlrab, Biomed, Btoclzim, Acla, 45: 1295,
1988,
W , Woltfrcab, and J, Lasch, Demalo2ogica, 174: 18, 1987,
A . J , M, flermorken, M. W e A , C. Nukkelhoven, A . M . G ,
Vermeesch-Markslag, G . M. A. A. Gms, P . Wir-tz, and 2,
Zlegenmeyer, J. P h a m , PharmacoE, , 35: 334, 1984.
H, Kornatsu, 15. Olramoto, K , Miyagawa, M, Hashida, and W.
Sezaki, Chem. Pharm, Bull. , 34: 3423, 1986,
.
T, Niahihata, R Kotera, V. N&ans, and M. Pamazakf , Ghem,
Plzamm, Bull, , 35: 380q 12987,
S , W . Nuf, T , P , Stewart, P. E . Ueaglle, and A , E). Albert,
A r c h , Bioehem, Biophys. , 20rt 227, 1981,
A , KaPo, Y, Ishibashi, and Y , Miyaki, J. Pharm, Pharmacol,,
39:399, 1986.
SurtaGtant Association Coiloids
as Topical Drug Delivery Vehicles
Figure 17.2 Phase diagram according to Ekwafl (12) Tor the system
water-octanol-sodium octmoate at 25OC. The single-phase re@ons
are denoted by L1 for the normal micellar, E for the hexagonal Xiq-
uid c q s t d , D for the lamellar Equid crystal, and L2 for the inverse
mieelfar, This phase behavior is typical. for a single-tajil: anionic
surfactant in combination with a medium chain alcohol. (From Ref.
11, used with permission of Aeadernie Press),
where v and lc are appraximately the volume and the length of the
hydr0cam;bon portion of the amphiphae, and acl, is the optimum area
354 O s b o m e , Ward, and OVJeilt
where n is slightly less than the number of carbon atoms per am-
phiphiIe chdn. Determination of a. is accomplished by x-ray dif-
fraction measurements on the lamellar liquid crystal.
Defined this way, the values for packing ratios related to the
different stmctures are:
Structure Ratio
Qral LDZ0 in
Eye Skin
Alcohol rat (mgikg) irritation irritation
aOral LDs0 for mouse, oral LDS0 for humans 0.5 to 5 glkg.
ble, and topicd dosage forms. Studies that have been completed to
identify a microemulsion formulation that will be pharmaceutica1:y ae-
ceptable, that i s , nontoxic, nonirritating, and so on, will be the
focus of this section.
The formation of micmemulsions has traditionally been dependent
upon the addition of a medium-chain length alcohol to function as a
cosurfactant. Because alcohols of this chain length range tend to
be skin o r eye irritants (Table 13, their use In topical. formulations
is very limited. An early study that attempted to find a more suit-
able topical microemulsion compared the solubilization of h y d m o r t i -
sone in a traditional microemulsion with the solubilization in a micro-
emulsion contetining pharmaceutically acceptable surfactants (22)-
The traditiona1 mieroemulsion consisted of sodium stearate-sodium
myristate surfactants and moderate-chain l e n e h alcohols (2.e.. n-
butanol t h r w g h n-heptanol) whereas the pharmaceutical microemul-
sion was a mixture of Brij 35 and Arlacel 186 combined with iso-
propanol. For both the traditional and pharmaceutical microeml-
sions, solubility of hydrocartisone was found to be independent of
the waterlofi ratio studied (0.2-0.5) and independent of the oil
chain length (n-alkanes Cg, (214, Cf5, C16). The pharmaceutical
formulation speciEed in this study consisted of 10 ml n-alkane (oil) ,
4 ml isopmganol, 4 g Arlacel 186, 2 g Brij 35, and 1 to 5 ml water,
A more recent study by this research group (23) indicated that
the combination of the anionic surfactant docusate sodium (dioctyl
sodium sulfosuccinate , which is the USP grade of the substance so-
dium 1.4-bis(2-ethylhexy1)-sulfosuccinate and is marketed as Aero-
sol-OT, most commonly abbreviated as AOT in the chemicall litera-
Surfactant Association Colloids as Vehicles
ARLACEL 2 0
w AUI
WATER
WATER
WATER AQT
75 until less than 40% AOT-75 is present (i.e, , greater than 60%
Arlacel 20). Thus, the mklng problems characteristic o f the w a x y
200% AQT are not encountered with A0T-75, Fudfiermare, the rhe-
slam studies indic&e that the most readily combined binary surfac-
tant mixtures will contain p e a t e r than 411% AOT-75. Because micro-
emulsions are thermsdynamica.lIy stable and their formation is revers-
ible with temperature, the mixing of 100% AQT could be facilitated by
heating.. However, studies indicating hydrolysis of AQT upon long-
term storage ( 29) diseaurages this approach,
The extent of the mieraemulsion re@on between 100%AOT- 75
and 50: 411 AQT-751Artacel 22 is shown in Elwre 9. This one-phase
re@on grows in area smoothly. A11 surfactant ratios except the
Qsborne, W a r d , and QtNeill
0 Spindle Q 1
A Spindle $2
0 Spindle a3
HEXADEGANE
HEXADECANE
bile salt mixtures. These systems have been studied for use as in-
jectable~ (34,35) and togicals (36). Additional studies on the phase
behavior of these systems have been completed in efforts to better
understand dissolution-solub1lization of cholesterol @llstones (37,
38). Evaluation of this system for topical drug delivery included
comparison of totally hydrogenated lecithin with natural lecithin that
had been "purifiedt' but not hydrogenated, and replacement of wa-
ter in part by other polar solvents (39).
The results of this study are shown in Figure 13. The striking
feature is the difference bet ween the totally hydmgenated lecithin
and the unsaturated purified lecithin. The distinct decrease in the
ability of the sodium cholate micelles to incorporate the totally hy-
drogenated lecithin can be attributed to the stiffness of the leci-
thin's saturated chains. This stiffness would be expected to raise
the transition temperature of the bilayer, thus pmventing the for-
Surfaetcmt Association Colloids as Vehicles
WEXABECANE
WATER AQT-75IARLAGEL 20
Figure 17.12 Pseudoternary phase diagram for the water-hexadec-
ane-AOT-75-ArlaceX 2Q system at 24 2 2%. The AOT-"151ArlaceI
20 ratios are as follows: 40: 60, AOT-751Arlacel 20;
30: 70 AOT-751Arlacel 20; 0: 80 AOT-751ArXacel 20;
90 AQT-751Arlact31 20; and ------ - 100% Arlacel 20,
water system, compared with having water as the solvent. For the
water system, a lamdlar Xquid cqstalline surface film immediately
formed around the added ledthin, severely h-indedng solubilfzation,
The bilayer destabsidrxg action of the pmpylene glycol prevented.
this film formation resulting in rapid disfsolutisn of the lecithin in
the sodium eholate- water mieellar system. Replacement of half the
water with butanedial had an effect simnar to the use of propylene
glycolI
skin barrier effects associated with the stratum corneum, the diffu-
sion of water within the vehicle was measured using a pulsed field
gradient fourier transform NMR technique (53). The unitless value
D/Dw is the ratio of the self-diffusion coefficient for water in the
microemulsion vehicle divided by the self-diffusian coefficient for
water in water. Therefore, a DlDw value of 0.100 means than, on
the average, water in the 45%water micmemulsian environment dif-
fuses at one-tenth the rate that a water molewle would diffuse in a
totally aqueous environment. The initially low values for DlDw for
microemulsions of low water content, and subsequent increase with
addition of water, can be attributed to binding of the water mole-
cules to the surfactmt headgroup. Thus, for the 15%water micro-
emulsion, most of the water in the micrmmulsion is bound to the
surfactant headgroup and is not available for transport across the
skin. The result is that transport of water across the skin from
this particular microemulsion is less than the transport of water
from neat water, that is, normalized flux is less than unity. Pre-
treatment studies using octanol, AOT , and AOT loetano1 (58: 42 ra-
tio; Table 3) further indicated that the enhancement in water pene-
tration from the high water content microemulsions was a result of
a synergistic effect between AOT and octanol and was not due to
the vehicle's microemulsion strmcture.
Additional studies that evaluated the in vitro transdermal trans-
port of labeled glucose from the microemulsions across cadaver skin
(511 showed that the glucose essentially crossed with the water.
Thus, microemulsion systems that showed high water transport, also
demonstrated high glucos@transport when an infinitely small amount
of labeled glucose was spiked into the microemulsion.
3 74 Osborne, Ward, and OtNeill
Octanol
AOT (from EtOH)
AOT foetan01 58: 42
8, Liquid Crystals
The close proximity of the liquid crystalline regions to the micro-
emulsion region allows comparison of drug transport from the struc-
turally different, but compositionally similar, surfactant association
colloids. A summation of this comparimn is given in Table 4. Two
different study designs were used to evaluate in vitro transdermd
flux of water across human skin after application of a micmmulsion
or Iiquid crystal. Study design A utilized surgical skin obtained
from radical mastectamy eases, with the skin being used within 24
to 48 h r after procurement* This skin was full thickness, stored on
wet ice, and used after removal of the subcutaneous fat (50). Re-
sulting fIux values were normalized against flux values for tritiated
water using skin from the same donor in an attempt to eliminate in-
dividual variation. For study design A, the vehicles were prepared
by dilution of tritium-labeled water with distiUed water, and then
addition of the appropriate amount of AOT dissolved in oetanol. A
single-label liquid scintmation-counting study was used to assay the
receiver phase of the Bow through transdermsl cell. In contrast,
study design B used dermatomed cadaver skin that was shipped on
dry ice, and that remained frozen until the time of use. Five rep-
licates of each formulation were tested using skin from a single in-
a ~ d u s l , Although normalization of flux values was not required to
minimize variation, the necessary transport data was determined to
allow direct comparison with the results from study design A . Sam-
ples were prepared with tritiattld water as described in design A
with the exception that a 4-pCi spike of glucose D - E ~ ~ C ( U(5.7
)I
mCi/mmol) was dissolved in the water before addition of the AQT-
octanol solution. For study design B, a dual-label liquid scintilla-
tion-counting study was used to assay the receiver phase of the
flow through transdermal cell.
Considering the differences in experimental design bet ween
these two studies, the normalizd flux of water across the skin after
Surfactant Association Cotloids as Vehicles 375
V. COhtCLLlDtNG REMARKS
REFERENCES
101:269, 1885,
d. Rodgers, and P. A , Winsor, J. Colloid Interface Sei, , 30;
247, 1969..
. .
E 1 , Frances, and T S , Hart, J . Colloid Interface Sci, , 94:
2, 1983.
B. W . Osborne, C A . Middletan, and R . L. Rogers, J . B ~ s -
I
1. THEORY
I I. FORMULATION
0.
U
NEUTRALIZED
CARBOMER n
D
0 9
O 10 20 30 40 50 60
TEMPERATURE (%)
II1. PROCESSING
REFERENCES
I. INTRODUCTION
A. What Are Silicones?
The term silicone refers to a class of ~yntheticpolymers that are
based on dternating silicon-oxygen, or siloxane (-Si-0-1 units.
The silicones that are most commonly used in topical pmducts are
polydiorganosiloxanes in which two organic groups are bonded to
each silicon atom. For commercial polydiorganosiloxanes , the organic
groups are neady always methyl groups, and such materials are re-
ferred to as polydimsthylsiloxanes (PDMS) . Other types of silicones
that are used in topical products can be thought of as derivatives
of PDMS in which some of the metfiyl groups have been replaced
with other organic groups,
Polymethylsiloxanes are produced in one of two forms: linear or
cyclic, Both are widely used in topical formulations and together
account for most of the volume of silicones umd in these applications,
Linear PDMS is a ooIorless, odorless oil that is available in a wide
range of viscosities, The viscosity of linear PDMS is directly related
to molecular weight, and can range from less that 1 centistoke to
over 1 million centistokes (cs), Cyclic PDMS is an odorless, color-
less, low-visoosity , volattle oil. Commercially produced cyclic PDMS
is available in a fairly narmw range of molecular weights, corre-
spanding to the cyclic species with four, five, or six dimethylsilox-
ane units in the ring. A s expected, volatility of cyclic PDMS is in-
versely related to molecular weight. Unlike linear PDMS , cyclic
PDMS is a fairly good solvent, and this is one reason for its popu-
larity among formulators.
The chemical and physical properties of silicones that contain
organic substituents other than methyl vary according to the nature
of the organic substituent. For example, substitution of some meth-
yl groups on PDMS with polyethylene oxide chains dramatically in-
creases the viscosity of the material because of the polarity of the
polyethylene oxide. The chemical properties of the organic substit-
uents are generally not affected by the siloxane backbone. Conse-
quently, polydiorganosiloxanes exhibit the chemical reactivity that
would be associated with the organic substituents, The two types
of silicones with nonmethyl substituents that are used in topical
products are discussed in Sections 1I.C and I1 .D.
C, Phenyltrimethylsiloxane
PhenyE trimethicone is the name that the CTFA (2) has assigned to
silicones that conform to the structure shown in Figure 3. Commer-
cially available phenyl trimethicone is a &ture of species corre-
sponding to the structures indicated by Figure 3 for which the value
of n ranges from one to three. The increased number of organic
substituents on the silicon In phenyl trimethicone greatly increases
its solubility in organic ingredients relative to dimethyl silicones.
Phenyl trimethicone is the only silicone that is miscible in aU pro-
portions with 95% ethanol. It is also miscible with all commonly used
nonpolar organic ingredients in topical formulations.
Because of its solubility in other ingredients, phenyl trimethi-
cone is easy to incorporate into existing topical formulations. Typi-
caUy it provides the same benefits a s other low-viscosity silicones.
Phenyl trimethicone has the additional benefit of a somewhat higher
refractive index than dimethyl silicones ( 1.46 vs. 1.39). Hence, it
is used in topical formulations for which gloss is desired, such as
in a hair dressing.
Use of Silicones in Topical Products
Ingredient ~o~ub~ity
CH3
Figure 19.3, Phenyl trimethicone.
Starch
Oil phase
dimethicone ( 500 cs) 3.0
mineral oil 1.5 Rlearol I Witco Chemical.
petrolatum 1.0 Sono-Jell No. 9 I' Witco ChemicQ
stearic acid (50%min.) 3.0
cetyl alcohol 1.0
Water phase
water 89.3
triethanolamine (99%) 1.2
Oil phase
cyclomethicone 4.0 345 Fluid / Dow Corning
dimethicone ( 1000 cs) 1.0
synthetic beeswax 2.0 Syncrowax BB4 / Croda
ozokerite wax 2.0
glycepyl oleate (and)
propylene glycol 2.0 Arlacel 186 / XCI Amerieas
mineral oil 15.0
Water phaae
water
glycerin
methicone base. The salt will settle to the bottom, but is easily re-
dispersed by shaking. This type of formulation has improved aes-
thetics relative to antiperspirant shown in Table 4 because the elim-
ination of water gives the formulation a pleasant dry feel when it is
applied to the skin. Other active ingredients that are availabIe in
the form of finely divided powders could be substituted for the alu-
minun cfiiorohydrate to produce a variety of different formulations.
Water 59.8
Carbomer 934 0.5 Carbopol 934 / B .I?. Goadrich
Triethanolamine (99%) 1.2
Glycerin 34.2
Propylene glycol 2.0
Dimethimne copoXyol 2.3 193 Surfactant I Ilow Corning
Propylene glycol
Ethanol (SDA 40)
PPC 10 cetyl ether 10.0 Procetyl 16 I Cmda
lsostearyl alcohol 19.8 Adol 66 1 Sherex
Cyclomethicone 39.8 345 Fluid / Dow Corning
Sodium stearate 8.0
hol m e heated to about 80°C, and the aluminum chlorohydrate is
stirred In. After the aluminum chlorohydrate is dispersed, the cy-
clomethicone is added, and the mixture is stirred untn it Q uniform.
The formulation should be allowed to cool, with mixing, until just
above the mlidification temperature (50-60°C) and then poured into
molds.
Oil phase
cyclomethicone (and)
32256 Formulation Aid /
dimethicone ccrpolyol 7.2 Dow Corning
cyclomethicone 9.6
mineral oil 7.4
pareth- 15-3 0.4 Tergitol 15-S-3 /
Union Carbide
Water phase
glycerin 20.2
water 53.2
sodium chloride 2.0
Oil phase
eyelomethicone (and)
32256 Formulation Aid /
dimethicone copolyol 6.0 Dow Corning
cyclomethicone 27.0
polymrbate-20 1.0 Tween 20 / XCI Amedcas
Water phase
aluminum chlorohydrate
(powder) 20.0
water 46.0
needed for emulsion stabBity, and sodium chloflde has been used jn
this example, The nonionic surfactant pareth- 15-3 is included to
improve emulsion stability. The viscosity of the formulation can be
controlled by adjusting the ofi-phaselwater-phase ratio, Preparation
of the formulation is sirnirar to other topical emulsion formulathns,
except that no heating is needed. The oll phase and water phase
are mixed in sepamte containers and the emulsion is made by slowly
addbg the water phase to the oil phase with rapid mking. A high-
shear mher, such 8s an Eppenbach or Ssversan, is desirable for
makhg this type of emulsion, but a stirrer equipped with a Codes
blade can be used. For large quantities of emulsion, the best pm-
cedure is to prepare the batch with a conventional stirrer and then
pass it through a colloid ma1 to ensure that the entire batch i s ade-
quately mhed,
Table 14 lists a starting f"srmulation for an mtiperspirant emul-
sion.. in contrast with the formulation given in Table 8, the anti-
perspirant salt will not settle out because it is dissolved in the wa-
ter phase. No additional electrolyte is required in the formulation
because of the high level of antiperspirant salt used. The formula-
tian is prepared using the same method given for the predous ex-
ample.
REFERENCES
APPENDIX
Suppliers
Witco G hemicd Corporation
Sonneborn Didsion
520 Madison Avenue
Mew York, Mew Uork 10022
(212) 605-3908
Use of Silicones in Tapfcat Products