Int Arch Occup Environ Health (1999) 72: 528±532
ORIGINAL ARTICLE
P. Sartorelli á A. Cenni á G. Matteucci á L. Montomoli
M. T. Novelli á S. Palmi
Dermal exposure assessment of polycyclic aromatic hydrocarbons:
in vitro percutaneous penetration from lubricating oil
Received: 7 January 1999 / Accepted: 10 July 1999
Abstract Objectives: Percutaneous penetration of poly-
cyclic aromatic hydrocarbons (PAHs) is a€ected by
various factors connected to exposure conditions. The
nature of the matrix, such as that of oil, can strongly
a€ect their percutaneous penetration. Risk assessment
should consider these e€ects. We examined the e€ect of
matrix on percutaneous penetration of PAHs, particu-
larly that of lubricating oil. Methods: The test apparatus
consisted of an in vitro static di€usion cell system using
full-thickness monkey (Cercopithecus aetiops) skin as the
membrane and saline solution with gentamycin sulfate
and 4% bovine serum albumin as receptor ¯uid.
Chemical analysis of PAHs in the samples obtained
from cells was carried out by inverse-phase HPCL, and
the results were read by spectro¯uorimetry. Results:
Comparing the penetration of 13 PAHs from a lubri-
cating oil and from acetone solution with arti®cial sweat
resulted in a signi®cantly slower passage from the oil
matrix for acenaphthene, anthracene, phenanthrene,
¯uoranthene, naphthalene, pyrene, ¯uorene (Mann-
Whitney U test, P < 0.05). No signi®cant di€erences in
the passage were found for chrysene because, in the test
with oil, its concentration was very often below the de-
tection limit. For benzo[a]anthracene, benzo[b]¯uoran-
thene, benzo[k]¯uoranthene, and benzo[a]pyrene it was
possible to demonstrate a passage through the skin only
when compounds were applied in acetone solution with
arti®cial sweat. Conclusions: The results of the study
P. Sartorelli (&) á A. Cenni á L. Montomoli á M.T. Novelli
Istituto di Medicina del Lavoro,
UniversitaÁ degli Studi di Siena, Via dei Tu® 1,
I-53100 Siena, Italy
e-mail: sartorelli@unisi.it
Tel.: +39-577-586755; Fax: +39-577-586159
G. Matteucci
Biocine Sclavo, Via Fiorentina 1, I-53100 Siena, Italy
S. Palmi
Dipartimento di Medicina del Lavoro, ISPESL,
Via Urbana 167, I-00184 Roma, Italy
Ó Springer-Verlag 1999
suggest the necessity of dermal penetration data relevant
for risk assessment, obtained under experimental con-
ditions similar to the real exposure conditions.
Key words Dermal exposure á Percutaneous
penetration á Polycyclic aromatic hydrocarbons á
Risk assessment
Introduction
It is well known that percutaneous penetration of
chemicals is strongly a€ected by the vehicle [2, 10].
Dermal exposure to polycyclic aromatic hydrocarbons
(PAHs) can occur by deposition of vapors and particles
or by splashing of oils. These are very common occur-
rences at workplaces. The nature of the matrix can a€ect
percutaneous absorption of PAHs. Risk assessment
should consider these e€ects. A di€erent dermal bio-
availability of benzo[a]pyrene (BaP) from that of ace-
tone or mineral oils of varying viscosity has been found
[4]. When BaP was applied to mouse skin in acetone, the
degree of epidermal DNA and protein adduct formation
was about 15±20 times greater than that observed when
a low-viscosity oil was used as vehicle. However, when
applied in oils of varying viscosity only a twofold dif-
ference was seen across the whole viscosity range. We
studied the e€ect of matrix on percutaneous penetration
of PAHs and in particular from lubricating oil, using an
in vitro di€usion system.
Methods
The test apparatus consisted of a static di€usion cell system (cat-
alogue no. FDC 400, Crown Glass, N.J., USA) kept at a constant
temperature of 37 °C so that the skin surface temperature was
32 °C. The exposure area was 1.77 cm2 (diameter 1.5 cm). Full-
thickness monkey (Cercopithecus aetiops) skin from the abdomen
was used as the membrane. Comparative in vivo and in vitro
studies have demonstrated that percutaneous absorption through
monkey skin is similar to that in man [5, 11, 12]. Monkeys were
 529

Table 1 Applied concentration and physicochemical properties of sweat to simulate real conditions of exposure at the workplace.
tested polycyclic aromatic hydrocarbons Arti®cial sweat is a liquid with characteristics similar to those of
human sweat. Its composition is the following: 2.5 g sodium acid
Applied Molecular Log P Melting phosphate, 0.2 g triolein, 2 drops of Tween 85, water up to 1 l,
nmol/cm2 weight point pH 5.2 with hydrogen chloride [9]. As oil we used a commercial
synthetic lubricator for engines. Chemical analysis demonstrated
Naphthalene 160.0 128.2 3.40 81.0 that it does not contain PAHs over the detection limit. Barrier
Acenaphthene 120.0 154.2 3.92 95.0 integrity after applying 30 ll acetone solution was checked using
Fluorene 23.1 166.2 4.18 115.5 the [3H]H2O passage through human cadaver skin in two cells.
Anthracene 15.1 178.2 4.50 216.4 Using this technique previously described in literature [1], we
Phenanthrene 12.1 178.2 4.60 100.5 measured a normal percentage of absorption of water (£0.29%)
Pyrene 9.3 202.3 5.18 150.4 after 20 min, ®nding no apparent e€ects of this dose of acetone on
Benzo[a]anthracene 8.5 228.3 5.61 160.7 skin permeability.
Chrysene 6.5 228.3 5.91 253.8 Analysis of PAHs in the samples obtained from cells was car-
Benzo[b]¯uoranthene 16.5 252.3 6.12 168.3 ried out by puri®cation with acetonitrile in the presence of NaCl,
Benzo[k]¯uoranthene 6.3 252.3 6.84 215.7 followed by inverse-phase HPCL on a LC-PAH Supelchem column
Benzo[a]pyrene 6.1 252.3 6.50 178.1 (25 cm long, 4.6 mm i.d., 5 lm grain size) eluted with an aceto-
Dibenzo[a,h]anthracene 7.5 278.4 6.50 266.6 nitrile-water gradient. The results were read with a Shimadzu RF
Benzo[g,h,i]perilene 6.9 276.3 7.10 278.3 551 programmable excitation and emission wavelength spectro-
¯uorimeter. The coecient of variation of the analytical method
was 10.8% and the detection limits of the various substances were
(in nanomoles per liter): naphthalene 25.8, acenaphthene 1.67,
previously killed to produce poliovirus vaccine. Skin samples were ¯uorene 7.5, phenanthrene 2.80, anthracene 0.56, pyrene 0.49,
frozen and stored for few days. The receiving liquid was a saline benzo[a]anthracene 1.32, and chrysene 0.45.
solution with gentamycin sulfate and 4% bovine serum albumin. All statistical analysis was performed with the SPSS software
Liposoluble substances such as PAHs have also previously been package (SPSS, Chicago, Ill., USA).
found to di€use well with this receptor [7]. Ten 1-ml samples in
48 h were drawn for each cell.
A mixture of 13 PAHs (Table 1) was applied without occlusion Results
at the same concentration in 30 ll acetone solution (six cells) and
dispersed in a lubricating oil (seven cells). Using a small volume of
acetone, the volatile solvent evaporates leaving a ®lm of chemical The results were expressed as ¯ux at steady state Kp
on the surface if the skin. Then we added some drops of arti®cial (obtained dividing absorption ¯ux values at steady state

Table 2 Kp values (cm/h),

steady-state ¯uxes Kp Flux Lag time
(nmol cm2 h)1) and lag time )3
(h) of the tested compounds Naphthalene (1.87 ‹ 1.31) ´ 10 0.2740 ‹ 0.2189 4.86 ‹ 7.99
(mean ‹ SD) applied in a lu- Acenaphthene (1.72 ‹ 1.76) ´ 10)3 0.2255 ‹ 0.2432 8.37 ‹ 3.44
bricating oil Fluorene (1.64 ‹ 1.66) ´ 10)3 0.0363 ‹ 0.0355 5.70 ‹ 3.02
Anthracene (0.93 ‹ 0.98) ´ 10)3 0.0120 ‹ 0.0112 17.55 ‹ 4.73
Phenanthrene (0.50 ‹ 0.28) ´ 10)3 0.0060 ‹ 0.0035 15.15 ‹ 3.10
Pyrene (0.17 ‹ 0.04) ´ 10)3 0.0015 ‹ 0.0003 13.38 ‹ 8.91
Benzo[a]anthracene ±a ±a ±a
Chrysene (0.22 ‹ 0.12) ´ 10)3 0.0015 ‹ 0.0008 26.12 ‹ 3.34
Benzo[b]¯uoranthene ±a ±a ±a
Benzo[k]¯uoranthene ±a ±a ±a
Benzo[a]pyrene ±a ±a ±a
Dibenzo[a,h]anthracene ±a ±a ±a
Benzo[g,h,i]perilene ±a ±a ±a
a
Below the detection limit

Table 3 Kp values (cm/h),

steady state ¯uxes Kp Flux Lag time
(nmol cm2 h)1) and lag time )3
(h) of the tested compounds Naphthalene (6.31 ‹ 2.49) ´ 10 1.0107 ‹ 0.3981 1.18 ‹ 0.01
(mean ‹ SD) applied in acet- Acenaphthene (7.80 ‹ 4.10) ´ 10)3 0.9684 ‹ 0.5996 2.34 ‹ 2.31
one solution with arti®cial Fluorene (6.56 ‹ 5.33) ´ 10)3 0.1455 ‹ 0.1138 4.23 ‹ 3.99
sweat Anthracene (3.97 ‹ 2.82) ´ 10)3 0.0526 ‹ 0.0297 12.85 ‹ 7.18
Phenanthrene (2.63 ‹ 0.74) ´ 10)3 0.0319 ‹ 0.0089 10.95 ‹ 7.62
Pyrene (4.13 ‹ 4.36) ´ 10)3 0.0387 ‹ 0.0416 24.46 ‹ 2.68
Benzo[a]anthracene (1.72 ‹ 2.60) ´ 10)3 0.0142 ‹ 0.0215 27.14 ‹ 8.28
Chrysene (0.57 ‹ 0.43) ´ 10)3 0.0035 ‹ 0.0025 23.79 ‹ 2.25
Benzo[b]¯uoranthene (0.09 ‹ 0.04) ´ 10)3 0.0014 ‹ 0.0006 22.46 ‹ 21.12
Benzo[k]¯uoranthene (0.09 ‹ 0.04) ´ 10)3 0.0006 ‹ 0.0002 23.80 ‹ 25.70
Benzo[a]pyrene (0.23 ‹ 0.20) ´ 10)3 0.0014 ‹ 0.0012 31.21 ‹ 10.81
Dibenzo[a,h]anthracene ±a ±a ±a
Benzo[g,h,i]perilene ±a ±a ±a
a
Below the detection limit
530
Fig. 1 Cumulative percutaneous penetration of acenaphthene applied
in two di€erent vehicles
by the concentration of substance applied) and lag-time
mean values ‹ SD.
Comparing Kp values of compounds tested under the
two di€erent conditions (Table 2) resulted in a signi®-
cantly slower passage from the oil matrix for acenaph-
thene, anthracene, phenanthrene, ¯uoranthene,
naphthalene, pyrene, ¯uorene (Mann-Whitney U test,
P < 0.05). No signi®cant di€erences in the passage were
found for chrysene because, in the test with oil, very
often its concentration was below the detection limit.
For benzo[a]anthracene, benzo[b]¯uoranthene, ben-
zo[k]¯uoranthene, benzo[a]pyrene it was possible to
demonstrate a passage through the skin only when com-
pounds were applied in acetone solution with arti®cial
Fig. 2 Cumulative percutaneous penetration of anthracene applied in
two di€erent vehicles
Fig. 3 Cumulative percutaneous penetration of chrysene applied in
two di€erent vehicles
sweat. Figures 1±7 compare percutaneous penetration of
7 PAHs applied in the two di€erent vehicles.
Discussion
Our experiment demonstrated a slower penetration
through the skin of PAHs included in the oil matrix than
the same compounds at the same concentrations in ac-
etone solution with arti®cial sweat. This should be due
to the liposolubility of PAHs and their consequent af-
®nity with oily liquids. This could lead to an overesti-
mation of dermal exposure in workers exposed to
lubricating oils.
The results of the study suggest the necessity of der-
mal penetration data relevant for risk assessment,
Fig. 4 Cumulative percutaneous penetration of phenanthrene applied
in two di€erent vehicles

Fig. 5 Cumulative percutaneous penetration of ¯uorene applied in
two di€erent vehicles
obtained under experimental conditions similar to the
real exposure conditions. Factors such as short exposures
with non-steady-state conditions, chemicals included in
various vehicles (matrix and solvents) or solid can a€ect
percutaneous penetration.
According to literature, the penetration from solids is
very low [13, 14]. The percutaneous penetration of seven
PAHs, contained in two di€erent types of carbon and at
the same concentration in acetone solution with arti®cial
sweat, has been compared using static ambered cells, full
thickness monkey skin, and human plasma as the re-
ceptor and chemical analysis [3]. Results show no per-
cutaneous penetration of PAHs contained in the two
types of carbon (their concentration in the receptor ¯uid
was under detection limits) with a passage through the
skin of the compounds when they were in di€erent
vehicles (acetone solution with arti®cial sweat). Even the
presence of solvents in commercial formulations can
change penetration rates of chemicals [8].
Fig. 6 Cumulative percutaneous penetration of naphthalene applied
in two di€erent vehicles
531
Fig. 7 Cumulative percutaneous penetration of pyrene applied in two
di€erent vehicles
It is well known that properly conducted in vitro
experiments can predict the in vivo absorption rate with
reasonable approximation, but in vitro experimental
conditions should be carefully controlled. To under-
stand the dermal uptake of chemicals bound to soil,
dust, oil, water, and solvents, etc. information on the
pure substance is helpful, but additional factors must
also be considered. Those factors can be easily repro-
duced and controlled in in vitro experiments.
An attempt to establish international guidelines for
in vitro percutaneous penetration studies was made in
1996 by the OECD [6]. However, these guidelines have
not been accepted in all places, and now the OECD is
reconsidering them. On the other hand, the Percutane-
ous Penetration Subgroup of the Dermal Exposure
Network in the European Union is elaborating an in-
ternational research proposal in this ®eld. The aim of
this research is to produce in vitro penetration data that
are suciently reliable and standardized as to be
acceptable for regulatory purposes.
Acknowledgements This research was carried out in collaboration
with Istituto Superiore per la Prevenzione e la Sicurezza del Lavoro
(ISPESL), Rome, Italy.
References
1. Bronaugh RL, Stewart RF, Simon M (1986) Methods for
in vitro percutaneous absorption studies. VII. Use of excised
human skin. J Pharm Sci 75: 1094±1097
2. Bronaugh RL, Franz TJ (1986) Vehicle e€ects on percutaneous
absorption: in vivo and in vitro comparisons with human skin.
Br J Dermatol 115: 1±11
3. FoaÁ V, Elia G, Schiavulli N, Sartorelli P, Sciarra G, Cenni A,
Novelli MT, Mangani F, Cecchetti G, Iachetta R (1998) The
non bioavailability of carcinogenic polycyclic aromatic
hydrocarbons contained in coal dust. Medicina del Lavoro
(Milano) 89: 68±77
532
4. Ingram AJ, Phillips JC (1993) The dermal bioavailability of
radiolabelled benzo(a)pyrene from acetone or from oils of
di€ering viscosity, assessed by DNA and protein binding.
J Appl Toxicol 13: 25±32
5. Moody RP, Franklin CA, Ritter L, Maibach HI (1990) Dermal
absorption of the phenoxy herbicides 2,4-D, 2,4-D amine, 2,4-
D isooctyl, and 2,4,5-t in rabbits, rats, rhesus monkeys, and
humans: a cross-species comparison. J Toxicol Environ Health
29: 237±245
6. OECD (1996) Dermal delivery and percutaneous absorption: in
vitro method. OECD guideline for the testing of chemicals ±
proposal for a new guideline. OECD, Paris
7. Sartorelli P, Aprea C, Cenni A, Matteucci G, Novelli MT,
Sciarra G (1995) In vitro percutaneous absorption of phena-
nthrene through monkey skin. Medicina del Lavoro (Milano)
86: 34±39
8. Sartorelli P, Aprea C, Bussani R, Novelli MT, Orsi D, Sciarra
G (1997) In vitro percutaneous penetration of methyl-para-
thion from a commercial formulation through the human skin.
Occup Environ Med 54: 524±525
9. Sciarra G, Angotzi G (1984) Studio della cessazione del di
(2-etilesil) ftalato da ®nte pelli in PVC (abstract in
English). Giornale Italiano Medicina del Lavoro (Pavia) 6:
235±238
10. Twist JN, Zatz JL (1989) Interaction of vehicles with model
skin membranes in the permeation process. In: Bronaugh RL,
Maibach HI (eds) Percutaneous absorption. Dekker, New
York, pp 147±173
11. Wester RC, Maibach HI (1975) Percutaneous absorption in the
rhesus monkey compared to man. Toxicol Appl Pharmacol 32:
394±447
12. Wester RC, Maibach HI (1989) In vivo animal models for
percutaneous absorption. In Bronaugh RL, Maibach
HI (eds) Percutaneous absorption. Dekker, New York,
pp 221±238
13. Wester R, Bucks DAW, Maibach HI (1993) Percutaneous ab-
sorption of contaminants from soil. In Wang RGM, Knaak JB,
Maibach HI (eds) Health risk assessment. CRC, London,
pp 147±148
14. Yang JJ, Roy TA, Krueger AJ, Neil W, Mackerer CR (1989) In
vitro and in vivo percutaneous absorption of benzo(a)pyrene
from petroleum crude-forti®ed soil in the rat. Bull Environ
Contam Toxicol 43: 207±214