`THE JOURNAL OF INVESTIGATIVE DERMATOLOGY, 76:76-79, 1981
`Copyright© 1981 by The Williams & Wiikins Co.
`
`Vol. 76, No. 2
`Printed in U.S.A.
`
`Physiocochemical Characterization of the Human Nail: I. Pressure Sealed
`Apparatus for Measuring Nail Plate Permeabilities
`
`KENNETH A. WALTERS, PH.D., GORDON L. FLYNN, PH.D., AND JOHN R. MARVEL, PH.D.
`
`University of Michigan, College of Pharmacy (KW & GF), Ann Arbor, Michigan and Dermatology Division, Ortho Pharmaceutical
`Corporation (JM), Raritan, New Jersey, U.S.A.
`
`Diffusion characteristics of the nail plate are necessary
`in providing the baselines for rational topical manage-
`ment of nail infections. In order to develop such baselines
`a unique stainless steel diffusion cell has been designed.
`The cell permits the exposure of 0.38 cm2 of nail plate to
`a bathing medium which is stirred by small motors
`mounted above the cell. The diffusion of water, methanol
`and ethanol at constant temperature (37°0, has been
`examined over periods up to 4 h. Average permeability
`coefficients of water, methanol and ethanol were deter-
`mined as 16.5 _+ 5.9 x 10 3cmhr 1 5.6 + 1.2 × 10 3cm
`hr-1 and 5.8 _+ 3.1 × 10 ~ cm hr 1 respectively. Moreover
`rates of diffusion across the nail were inversely propor-
`tional to nail thickness. Based on methanol data, nail
`plate barrier property appears stable for long periods of
`aqueous immersion.
`
`Little is known concerning the permeability of the human
`nail plate yet such knowledge is important to the understanding
`of toxicities of topically applied chemicals [1,2], to structure/
`activity considerations in drug design, and to the formulation
`of topical drug delivery systems and cosmetic systems. A par-
`ticular problem is that nail is prone to attack by a host of fungi
`and micro-organisms, some of which invade the living tissues of
`the nail bed and nail fold and others invade the nail plate itself
`[3,4]. It is uncertain whether the nail is sufficiently permeable
`to allow succe~ful topical management of such conditions [6].
`In order to assess the feasibility of chemical penetration, a
`systematic study of the physicochemical factors which govern
`deposition of topically applied chemicals into nail plate, includ-
`ing definition of molecular size and lipophilicity requirements,
`seems necessary. There are only a few quantitative measure-
`ments of nail plate permeability in the literature [6-9]. This
`dearth of information may be attributed to the difficulty of
`working with a relatively rigid membrane of variable thickness.
`Problems attending construction of a suitable apparatus for the
`direct measurement of nail plate permeability have yet to be
`surmounted. The initial purpose of our research was thus to
`construct a unique diffusion cell to quantitatively assess nail
`plate permeability.
`
`MATERIALS
`
`Tritiated water* ~H-methanol,t and :4C-ethanolS were selected as
`permeants. Saline (0.9% sodium chloride irrigation solution [2]) was the
`medium used to prepare the permeant solutions, which were 1 × 10-3
`molar or less concentration and thus too dilute to have direct effects on
`the membrane barrier properties. Cadaver nails were obtained from the
`University of Michigan, School of Medicine [3].
`
`PERMEABILITY CELL AND METHOD
`
`Stainless steel diffusion cells, which expose 0.38 cmz of nail plate to
`the compartmental solutions, were constructed. The internal and ex-
`ternal diameter of the cells were 0.7 cm and 1.3 cm, respectively. The
`length of each cell half, taken from the edge of the seal, was 3.9 cm.
`The design allowed a tight seal between the nail and the half-cells using
`male-female interlocking flanges under pressure (Fig 1). The cell was
`placed in a clamping device operating as a "c-clamp" (Fig 2) and
`pressure was exerted about the circumference of a nail plate section cut
`to fit snugly within the wider flange’s diameter. Sample and stirrer
`ports were provided to each half-cell, the individual volumes of which
`varied from 1.4-1.6 cm:~. Cells were immersed in individual constant
`temperature water baths and the contents were stirred with small
`Teflon propellers attached to Teflon shafts exiting the stirring port.
`Small Teflon discs with set-pins were used to position the shaft with
`propeller in the center of the chamber. These also prevented evapora-
`tion from this port. The connection of the upper shaft end to the
`stirring motors (Hurst, 150 rmp§) was with a short section of flexible
`tubing (Fig 3).
`Nails were pre-immersed in normal saline for 24 h before use. Pre-
`immersion served to clean the nails and also to facilitate trimming to
`fit the diffusion cells. Nail thicknesses were measured with a micrometer
`prior to insertion in the diffusion ceils.
`After assembly of a cell, the 2 compartments were ffl]ed with either
`normal saline or citrate buffer and the stirring shafts were connected to
`the stirring motors. Citrate buffer was only used during some determi-
`nations of ethanol permeability. No differences in permeability rate
`were observed when the buffer was employed. The contents were mixed
`for 30 min and background samples were taken. The donor compart-
`ment was then charged with a concentrate of the radiochemical per-
`meants and the run commenced. Five minutes were allowed for uniform
`mixing and then the "zero time" donor sample was withdrawn. Samples
`were taken from the receiver compartments at predetermined intervals
`up to 4 h. At the end of a run a second donor sample was taken. The
`mean of the initial and final donor samples was used as the effective
`concentration differential across the nail over the course of the exper-
`iment. The late donor sample was never less than 15% of the "zero
`time" donor sample. Samples were taken from the water bath before
`and after the run to check for cell leakage.¶
`The samples were added to scintillation cocktail and counted on a
`Beckman LS 9000 liquid scintillation counter. The experiments were
`performed at 37°C. In all cases a dual label procedure was followed.
`The scintillation counter employed is equipped with energy windows
`which, when properly set, allow the separation of low energy emissions
`(predominantly 3H) from the higher energy emissions of carbon-14
`[10]. When concentrations of the radioisotopes are in certain propor-
`tions, it is possible to accurately factor the total disintegrations into its
`component tritium and carbon-14 contributions.
`The permeability of ~H-methanol was also studied as a function of
`duration of immersion of the nail in saline. Experiments were performed
`sequentially over a 49-day period. Between runs the nail was stored in
`saline solution.
`
`Manuscript received May 13, 1980; accepted for publication July 26,
`1980.
`Reprint requests to: Dr. Gordon L. Flynn, University of Michigan,
`College of Pharmacy, Ann Arbor, MI 48109.
`* New England Nuclear, Boston, Massachusetts.
`"~ Abbott Laboratories, North Chicago, Illinois.
`:~ Generously supplied by Dr. T.M. Oelrich, Department of Anatomy.
`
`§ Hurst, Mfg. Co. Ltd., Princeton, Ind.
`¶ The seal on the nail edges should be sufficient to preclude inter-
`compartmental leakage. However, leakage &round the membrane edges
`can also be discounted on the basis that large differences (20-fold) occur
`in the permeability rates of homologues with only slight variation in
`molecular volumes. For example, the permeability coefficients for pro-
`panol and octanol are 0.83 _+ 0.15 and 0.27 _+ 0.03 × 10-3 cm hx
`respectively. These data will be fully reported in a later article.
`
`76
`
`
`
`
`
`
`ARGENTUM EX1039
`
`Page 1
`
`
`
`simPl°P°rt
`
`Feb. 1981
`
`Stirrer Port
`
`[
`
`Stirrer Port
`
`!
`
`,(
`
`Membrane
`
`FIG l. Cross-section of the interlocking flange of the diffusion cell
`
`}
`
`HUMAN NAIL PLATE PERMEATION
`
`77
`
`V = volume of the half-ceil into which the permeants were
`collected
`dC = actual pseudo-steady-state slope of plots as in Fig 4 in
`cpm/hr/cm3.
`
`RESULTS
`
`The permeability coefficients for 3H-water, ~H-methanol and
`14C-ethanol are given in the Table. These are provided with
`standard deviations and the number of experiments with each
`compound. Also tabulated are thickness normalized permeabil-
`ity coefficients. Figure 5 shows the effect of nail hydration time
`on the permeability of methanol. The data are for a single nail
`placed in a diffusion cell at the indicated times and otherwise
`kept in fresh saline. Figure 6 indicates how the permeabilities
`of methanol and ethanol were influenced by nail thickness.
`
`DISCUSSION
`
`Several investigators have studied the water permeability of
`nail plate [6-9]. Butch and Windsor [8] report water fluxes of
`1.16-5.30 mg/cm~/hr, through toenails, depending upon tem-
`perature and humidity. They assessed weight loss of a brass
`cylinder, filled with saline and closed at one end by nail plate.
`Weight loss was assumed to be due to loss of water through the
`nail. Using a similar method Baden, Goldsmith, and Fleming
`[6] reported that the flux of water across the nail ranges
`between 2.0-3.0 mg/cm2/hr. Spruit [7] designed a moisture
`sensing device which noninvasively measures insensible per-
`spiration. He found the nail water vapour loss to average 2.40
`
`Fla 2. The diffusion cell and clamping device.
`
`DATA ANALYSIS
`
`After factoring the data into its tritium and carbon-14 com-
`ponents, the individual data were plotted as counts (amount of
`permeant) collected in the receptor compartment as a function
`of time (Fig 4). Correction was made for sampling, which in all
`cases was done with replacement using normal saline. The
`permeability coefficient for a given run was calculated from:
`
`dC
`J = P.A.~C = V--
`dt
`
`Eq. 1
`
`where:
`
`J = the total flux or amount penetrated vs time, cpm hr-~
`or cpm/hr, in the pseudo-steady-state regions of plots
`such as in Fig 4
`P = the permeability coefficient (cm/hr)
`A = the cross-sectional area of the membrane (cmz)
`hC = the concentration differential across the membrane as
`described above
`
`FIG 3. The diffusion cell, water bath and stirring motors.
`
`Page 2
`
`
`
`78
`
`WALTERS, FLYNN, AND MARVEL
`
`Vol. 76, No. 2
`
`~ r
`
`,
`
`,
`1.0
`
`m
`
`!
`2.0
`
`across the nail does not evidence systematic change over a
`period of up to 49 days (Fig 5) suggests that, after the initial 24
`h soaking, no further significant hydration occurs in this mem-
`brane. The upper limit of the nail plate’s capacity for water
`appears to ba about 25% [6] and is only 2-3 times greater than
`the normal water content of the nail, which is estimated to lie
`between 7 to 12% [6,11]. The approximately 5-fold greater flux
`of water reported here relative to that observed by others using
`a nail plate with a dry dorsal surface seems a reasonable
`consequence of full hydration. The relative constancy of meth-
`anols permeability coefficient over a 49-day period suggests
`that the hydrated nail is a stable barrier and is able to withstand
`the repeated application of pressure without severe detriment.
`The rate of diffusion of methanol and ethanol is clearly
`dependent on nail thickness (Fig 6). This is a general expecta-
`tion as, for a simple is¯tropic membrane, the permeability rate
`is related to membrane thickness by
`
`KD
`P = --
`h
`
`Eq. 2
`
`l’.O
`
`2’.0
`
`3".0
`
`4’.0
`
`providing boundary layers are insignificant. In this expression
`K is the membrane--solution partition coefficient of the per-
`
`o
`r~
`
`O
`v
`
`0
`4d
`o
`o
`
`4~
`
`4J
`0
`o
`
`~3
`
`>
`
`4a
`
`40
`
`2O
`
`0.4
`
`0.2
`
`7
`~ 10.c~
`
`o B.@
`x
`
`6,e
`2
`
`4.0
`
`Mean ¯
`
`¯
`
`Oo~
`
`¯
`
`¯
`
`(f
`
`HYDKATION TIME (days)
`FIG 5. The effect of nail hydration time on the permeation of
`methanol.
`
`TIME (hr)
`FIG 4. Typical nail permeation plots for water, methanol and
`ethanol.
`
`TABLE Permeability coefficients
`
`Permeant
`
`Permeability~ ,, No. of
`coefficient (x 10)
`observa-
`tions
`cm hr-’ s-_ SD
`
`Normalizedb
`permeability
`coefficient
`
`No. of
`observa-
`tions
`
`:~H-water
`3H-methanol
`’4C-ethanol
`
`16.5 +_ 5.9
`5.6 -4- 1.2
`5.8 ± 3.1
`
`22
`26
`26
`
`10.5 -+ 3.5
`5.1 _+ 1.1
`2.7 ± 0.8
`
`12
`18
`26
`
`P observed.
`b p observed × nail thickness (ram).
`
`12
`
`~ 10
`
`mg/cm~/hr, in close agreement with the earlier in vitro results.
`Moreover, Spruit found that when the thickness of the nail was
`accounted for, there was little variability in the permeability
`rate. The observed rate is about 10 times greater than insensible × 8
`perspiration measured across normal skin. This actually sug-
`gests that nail plate is up to 1000 times more intrinsically
`permeable to water than stratum corneum because the ratio of ~
`thicknesses of these tissues is roughly 100 (nail over stratum
`co~eum).
`In the present experiments a water flux of 12.6 + 5.8 mg!cm2/ 2
`hr was obtained (calculated from total water transported over
`the experimental run). This is about 5-fold greater than litera-
`ture values. However, the previous investigators worked with ~
`nails with a dry, dorsal surface. The present results appear to 2
`be the first assessements of rates across hydrated, totally im-
`mersed nails. Consideration must be given to the possibility
`that hydration of the dorsal surface increased the rate of water’s
`diffusion across this layer of the nail. Regardless, agreement
`with previous work is generally satisfactory.
`The observation that the permeability rate of methanol
`
`4
`
`I
`
`¯
`
`¯
`
`{ Methanol ¯
`
`I Ethanol ¯
`
`RI~CIPROCAL NAIL TIIICKNESS (turn-I)
`FIG 6. The relationship between nail thickness and permeation rates
`of methanol and ethanol.
`
`Page 3
`
`
`
`Feb. 1981
`
`HUMAN NAIL PLATE PERMEATION
`
`79
`
`meant, D is the diffusion constant of the permeant within the
`membrane, and h is the membrane thickness. The fact that the
`reciprocal dependency on thickness is not perfectly obeyed for
`either permeant suggests that either the substrata of the nail
`have somewhat different permeabilities and proportions or
`conditions which cause nails to thicken also lead to a less
`permeable plate on a per unit thickness basis. In this study a
`total of 14 nails were used ranging in thickness from 0.33 mm
`to 1.43 mm, and all were either t-mger or thumb nails, too few
`to make a firm conclusion as to the exact shape of the prot’fle
`and the causes of the nonlinearity. The data presented here are
`nevertheless indicative of a substantial thickness dependency
`(for methanol p < 0.01; for ethanol p < 0.001).
`In summary, a diffusion cell has been designed which is useful
`in the quantitative measurement of human nail plate permea-
`bilities. Preliminary studies using the diffusion cell indicate that
`the unique pressure-seal used in this apparatus is without
`detrimental effect on the barrier properties of the plate.
`
`We should like to thank Johnson and Johnson (Ortho) Inc. for
`financial support.
`
`REFERENCES
`
`1. Sulzberger MD, Rein CR, Fanburg SJ, Wolf M, Shair HM, Popkin
`GL: Allergic eczematous reactions of the nail bed: Persistent
`subungual & ungual changes based on contact with ’undercoats’
`containing artificial resins and rubber. J Invest Dermatol 12: 67-
`72, 1948
`2. Shelley WB: Onycholysis due to topical 5-fluorou~acil. Acta Der-
`matovener (Stockh) 52: 320-322, 1972
`3. English MP: Nails and Fungi. Br J Dermatol 94: 697-701, 1976
`4. Sagher F: Histological examinations of fungus infections of the
`nails. J Invest Dermatol 13: 338-357, 1949
`5. Baranov AF, Marochkina IA, Konopikhina GA, Kolokolova NV,
`Jakimenko DV, Panova PM: Experiences in treatment of ony-
`chomycoses with keratolytics and fungicidal plaster. Vestn Der-
`matol Vener 36: 65-67, 1962
`6. Baden HP, Goldsmith LA, Fleming B: A comparative study of the
`physicochemical properties of human keratinized tissues.
`Biochim Biophys Acta 322: 269-278, 1973
`7. Spruit D: Measurement of water vapor loss through human nail in
`vivo. J Invest Dermatol 56: 359-361, 1971
`8. Butch GE, Windsor T: Diffusion of water through dead plantar,
`palmer and torsal human skin and through toenails. Arch Der-
`matol Syph 51: 39-41, 1946
`9. Jacobi O: Die n~igel des lebenden menschen und die perspiratio
`insensibills. Arch Klin Expt Dermatol 214: 559-572, 1962
`10. LS 9000 Instruction Manual, Beckman Instruments, 1979.
`11. Forslind B: Biophysical studies of the normal nail. Acta Dermatov-
`ener (Stockh) 50: 161-168, 1970
`
`/ /
`
`Page 4
`
`