ARTICLE
UV-A radiation (320-400 nm) is involved in skin carcinogenesis and photoaging.
Reactive oxygen species are deleterious to DNA, membranes, and proteins
but their exact role in mutagenesis and lethality remains unclear in the
different skin cell types [1]. Furthermore, defense mechanisms and repair
capacities can be very different from one cellular type to another. Both
epidermal and dermal cells are targets for UV-A oxidative stress and their
antioxidant defenses can be overcome. Based on their anatomical site and
their functional and metabolic particularities, keratinocytes and fibroblasts
may have different responses to ultraviolet radiation. Cultured cell models
have helped to characterize the cytotoxic action of UV-A and the role
of reactive oxygen species in UV-A-induced cellular damage [1]. In fact,
the majority of studies have used cell lines and few studies focus on
human diploid cells [2]. Moreover, a comparison of UV-A susceptibility
of cell lines and normal cells has not been investigated in detail with
respect to pathways regulating reactive oxygen species.
The purpose of our study was first to evaluate, under basal conditions,
total glutathione levels (GSH), antioxidant enzyme activities [selenium-glutathione
peroxidase (GSH-Px), copper-zinc and manganese superoxide dismutases (CuZn-SOD
and Mn-SOD)] and lipid peroxidation levels in both normal cutaneous keratinocytes
and fibroblasts. Secondly, we tested the cytotoxicity of UV-A1 radiation
on diploid keratinocytes and fibroblasts derived from normal skin by examining
the capacity of cells to adhere and proliferate after the stress, using
the MTT colorimetric method. Finally, we compared the antioxidant content
and the cytotoxicity of UV-A1 radiation on two established keratinocyte
cell lines, the NCTC2544 and the HaCaT keratinocytes, and on an established
fibroblast cell line, the MRC5 fibroblasts.
Materials and methods
Media and reagents. Culture media, supplements and buffers for
keratinocyte cultures, RPMI-1640, fetal calf serum (FCS), L-glutamine
and PBS buffer were purchased from Gibco (Grand Island, USA). Trypsin-EDTA,
penicillin, streptomycin and kanamycin were obtained from Boehringer (Mannheim,
Germany) and fungizone from Squibb (Princetown, USA). Other chemicals
were obtained from Prolabo (Paris, France).
Tissue culture. Cultures of human basal keratinocytes were established
from foreskins (age 1-6 years). Monolayer cultures of non-differentiated
keratinocytes were grown in low calcium concentration (0.09 mM) growth
medium (keratinocyte serum-free medium, SFM) supplemented with epidermal
growth factor (5 ng/mL) and bovine pituitary extract (50 mg/mL). The cells
were used in the fourth transfer. NCTC2544 keratinocytes (human skin cell
line) were purchased from Flow laboratories (Virginia, USA) and cultured
in monolayer in 90% NCTC-135 medium + 10% fetal calf serum (FCS). The
spontaneously immortalized human keratinocyte cell line HaCaT was obtained
from the German Cancer Research Center (Heidelberg, Germany) and was cultured
in RPMI-1640 with NaHCO3, penicillin 50,000 UI/L, streptomycin
50 mg/L, L-glutamine 4 mM, added with 10% FCS. Normal human fibroblasts
from non-exposed areas were obtained from skin biopsies and cultured as
previously described [3]. In all experiments, fibroblast strains were
used in the same subculture (before the fifth transfer). MRC5 fibroblasts
(human fetal lung cell line) were purchased from Flow laboratories (Virginia,
USA) and cultured under the same conditions as the normal fibroblasts.
All cells were incubated at 37° C in a CO2-enriched atmosphere
(94% air/6% CO2) (Forma Scientific incubator, Marietta, USA).
UV-A1 irradiation. At confluence, cells were rinsed twice and
irradiated in PBS buffer without calcium-magnesium (pH 7.4) with a Uvasun
3000 apparatus (Mutzhas, Munich, Germany), marketed in France by Uvasun-France.
The spectrum is from 340 to 420 nm with a maximum intensity at 375 nm.
A compensated Kipp and Zonen thermopile coupled to a digital voltmeter
was used to measure the UV-A energy effectively received by the cells
through the culture dishes. Sham-irradiated control cells were placed
under aluminium foil in PBS buffer while irradiation was being carried
out. Preliminary work allowed us to define the radiation dose for each
cell type. The radiation doses used were from 3.6 to 8 J/cm2
for normal cutaneous fibroblasts, from 7.2 to 46.8 J/cm2 for
MRC5 fibroblasts, from 8 to 96 J/cm2 for basal cutaneous keratinocytes,
NCTC2544 and HaCaT keratinocytes.
Determination of total glutathione levels,
enzymatic activities and lipid peroxidation. Under basal conditions,
without irradiation, keratinocytes and fibroblasts were harvested and
washed three times in isotonic 400 mM Tris-HCl (pH 7.3). After washing,
cells were homogenized at 4° C in a Potter tissue homogenizer in
hypotonic 20 mM Tris-HCl. Immediately after homogenization, lipid peroxidation
was assayed in the lysate with the fluorescent technique using thiobarbituric
acid. Thiobarbituric acid reactant (TBAR) values were obtained using a
calibration curve of malondialdehyde (standard: 1,1,3,3-tetraethoxypropane)(MDA-kit,
Sobioda®, Grenoble, France) [4]. In order to evaluate total
glutathione, 100 ml of the lysate were transferred immediately into a
tube containing 400 ml of 6% (W/W) metaphosphoric acid in water. The solution
was mixed and centrifuged for 10 min at 4,000 rpm at 4° C. The acidic
protein-free supernatants were stored at 80° C until analysis.
Total glutathione was determined according to a slight modification of
the procedure previously described by Akerboom and Sies [5]. Briefly,
an aliquot of the deproteinized extract was neutralized with a solution
containing 0.4 M N-morpholoinopropanesulfonic acid, 2 mM EDTA and adjusted
to pH 6.75 with 1 M KOH. Glutathione was evaluated using enzymatic cycling
of GSH by means of NADPH and glutathione reductase coupled with 5,5'-dithiobis(2-nitrobenzoic
acid)(DTNB). After centrifuging the cell lysate at 2,665 g for 10 min,
GSH-Px was assayed by the method of Gunzler et al. [6], and total
SOD activity was determined using the method of Marklund and Marklund
[7]. Manganese SOD activity was determined after inhibition of the copper-zinc
SOD using potassium cyanide 9 mM as previously described [8].
Evaluation of UV-A1 cytotoxicity. Survival and proliferation
capacities were determined using two different methods. After UV-A irradiation,
the cells were rinsed twice with PBS buffer. Then, they were replated
in fresh medium and placed in an incubator for 18 hrs. The culture dishes
were then rinsed twice vigorously with isotonic saline to eliminate dead
cells. Percentage survival was determined as previously described [3,
9]. Each assay was run in duplicate. The toxic effect of UV-A radiation
was evaluated with the formula [(control cells irradiated cells)/control
cells] x 100. Results are expressed as a percentage of cytotoxicity compared
to sham-irradiated control cells. Cell survival was also quantified using
the modified MTT (3-(4,5-dimethylthiazol-yl)-2,5-diphenyltetrazolium bromide)
colorimetric method previously described [9].
Statistics. The results are presented as mean ± standard
deviation. All data were processed statistically by analysis of variance
using the Newman-Keuls test. The differences were considered to be significant
when p < 0.05.
Results
Antioxidant defenses and lipid peroxidation under basal conditions
in keratinocytes and fibroblasts. Under basal conditions, total glutathione
levels, GSH-Px and SOD activities (Table
I) were significantly higher in normal basal keratinocytes than
in normal fibroblasts by approximately 50-100% (p < 0.0001) while cellular
lipid peroxidation was far less. NCTC2544 and HaCaT keratinocytes demonstrated
significantly lower GSH-Px activities (p < 0.001) and higher GSH levels
(p < 0.0001) than the corresponding normal cells (Table
I). Similarily, MRC5 fibroblasts had significantly lower GSH-Px
activities (p < 0.001) and higher GSH levels (p < 0.0001) than normal
fibroblasts. The total SOD activity was slightly lower in NCTC2544 and
MRC5 cell lines compared to their normal counterparts with a significant
decrease of Mn-SOD in NCTC2544 keratinocytes compared to normal keratinocytes
(p < 0.0001). In the HaCaT cell line, all SOD enzymes were significantly
lower compared to the normal keratinocytes (p < 0.0001). Under the
same conditions, lipid peroxidation was 10-fold higher in NCTC2544 keratinocytes
and about 4-fold higher in HaCaT keratinocytes, compared to basal keratinocytes,
but 75% lower in MRC5 fibroblasts than in normal fibroblasts (p < 0.0001).
UV-A1-induced cytotoxicity on keratinocytes and fibroblasts.
UVA-1 cytotoxicity evaluated by examining the capacity of cells to adhere
and proliferate is reported in Figure
1. These results have been reproduced and confirmed using the
MTT method (data not shown). In all cells tested, the cytotoxic action
of UV-A1 was dose-dependent (Fig.
1). Fibroblasts were far more sensitive than basal keratinocytes
with a toxic dose 50 (TD50) of 5.8 J/cm2 for fibroblasts vs
48 J/cm2 for keratinocytes. Thus, epidermal cells demonstrate
a UV-A1 irradiation resistance factor of approximately 8 compared to fibroblasts.
NCTC2544 keratinocytes were more sensitive to UV-A1 stress than normal
keratinocytes, with a TD50 of 24 J/cm2 vs 48 J/cm2
(Fig. 1). In contrast,
HaCaT keratinocytes were more resistant than normal keratinocytes with
a TD50 of 60 J/cm2 (results not shown in Figure
1). On the other hand, MRC5 fibroblasts appeared more resistant
than normal fibroblasts with a resistance factor around 4: the TD50 was
23 J/cm2 for MRC5 fibroblasts vs 5.8 J/cm2
for normal fibroblasts.
Discussion
It is now well known that UV-A radiation creates an oxidative stress
for skin cells via reactive oxygen species involved in the cytotoxic
and mutagenic processes [1, 10]. Thus, it is of primary interest to better
characterize cellular oxidative damage and the activity of antioxidant
molecules in order to understand their role in UV-A-induced injury. However,
cellular chromophores and antioxidant defenses differ from one type of
cell to another, and probably between normal cells and immortalized cell
lines. Yet, many studies on UV-A cytotoxicity and molecular damage do
not define clearly the cellular type used.
In this study, we characterized the cytotoxic action of UV-A1 radiation
on normal diploid cutaneous keratinocytes and fibroblasts derived from
non-sunexposed areas from different donors and on two established cell
lines. Because cell membranes are an important target of reactive oxygen
species, we analyzed damage to membrane lipids under basal conditions.
In all cells tested, UV-A1-induced cytotoxicity was dose-dependent. However,
UV-A1 cytotoxicity differs with each cellular type. Thus, under our conditions,
normal cutaneous keratinocytes were more resistant than fibroblasts by
a factor of eight and demonstrated significantly lower lipid peroxidation
under basal conditions. The lethal action of monochromatic irradiation
at 325, 334 and 365 nm was first shown using cutaneous diploid fibroblasts
[11, 12]. Tyrrell and cowokers compared the UV-A cytotoxicity on both
fibroblasts and keratinocytes, as measured by loss of clone-forming ability,
and found similar sensitivity of both cell types at 365 nm, whereas fibroblasts
were more sensitive at 313 and 405 nm [13, 14]. Using the trypan blue
exclusion test, Moysan also found that fibroblasts were more sensitive
to UV-A than keratinocytes [15]. Our studies suggest that this difference
in sensitivity could be explained in part by the antioxidant defense mechanisms
since GSH, GSH-Px and SOD are significantly higher in keratinocytes compared
to fibroblasts. Indeed, previous works have shown a direct relationship
between UV-A cytotoxicity and the level of antioxidant molecules such
as GSH and GSH-Px [3, 14]. Our results agree with those of the study of
antioxidant molecules in crude extracts of human epidermis and dermis
[16]. Applegate et al. also found lower cytotoxicity and membrane
peroxidation in keratinocytes compared to fibroblasts with a constitutive
expression of heme oxygenase 2 in epidermal cells, known to be an antioxidant
enzyme [17]. However, other authors have found more antioxidant molecules
in fibroblasts than in keratinocytes. Yohn et al., using cells
from different donors, found increased GSH-Px, SOD and catalase in fibroblasts
compared to keratinocytes, and in keratinocytes compared to melanocytes
[18]. Moysan et al., using cells from the same biopsy, found no
link between UV-A cytotoxicity and antioxidant capacity since SOD, catalase
and GSH were identical in both cells and GSH-Px was higher in fibroblasts
[15]. Furthermore, Vessey et al. recently showed that as human
keratinocytes differentiate, they develop higher levels of the components
of the GSH pathway and an increased resistance to the toxic effects of
peroxides [19]. Thus, futher studies are needed to specify the exact role
of antioxidant defenses in cutaneous cell protection against UV-A radiation.
An interesting point in our results is represented by the wide range of
values for SOD activities in normal keratinocytes from different donors
which could explain the variability of the data on UV-A cytotoxicity in
the literature. In our work, two limiting factors are represented by the
different donors and the difference in the medium used for keratinocytes
and fibroblasts, which could influence the results. However, the factor
of eight difference in UV-A1 cytotoxicity between normal keratinocytes
and fibroblasts was found for all cells tested and did not appear to be
a coincidence. Moreover, the results concerning the HaCat cell line showed
a different UV-A1 cytotoxicity compared to normal fibroblasts while both
cells were cultured in the same medium.
Established epidermal and dermal cell lines are often used to study
UV-induced cellular damages in vitro. Our data suggest that antioxidant
status and behaviour of such cell lines is different from that of normal
cells. NCTC2544, HaCaT keratinocytes and MRC5 fibroblasts present quite
a different sensitivity toward UV-A1 radiation compared to normal cells:
NCTC2544 keratinocytes are more sensitive than normal basal keratinocytes,
while HaCaT keratinocytes are more resistant, and MRC5 fibroblasts are
less sensitive than normal fibroblasts. Furthermore, we found that the
antioxidant enzyme activities studied (GSH-Px and SOD) are decreased in
cell lines compared to normal cells while the GSH levels are higher. The
different sensivity to UV-A1 radiation between normal cells and cell lines
cannot be readily explained by variations of the antioxidant systems studied
in this work. Differences in endogenous chromophores, other antioxidant
molecule expression or in membrane polyunsaturated fatty acid composition
may explain the different responses to UV-A1 irradiation. Other factors,
such as number of subcultures, media and buffer composition (in particular
calcium concentration), and confluence may also modify UV-A1 cytotoxicity.
Taken together, results from all cells show that there is no relationship
between UV-A1 cytotoxicity and a particular antioxidant molecule and emphasize
the idea of a cooperative and synergistic antioxidant defense in cutaneous
cells. Furthermore, we found in all cells, apart from HaCaT keratinocytes,
a positive relationship between UV-A1-cytotoxicity and basal lipid peroxidation
evaluated by TBAR. These observations enhance the role of reactive oxygen
species in UV-A1-induced cytotoxicity showing a low basal lipid peroxidation
and the best resistance to UV-A1 stress in cells with a high antioxidant
molecule content.
Abbreviations:
CuZn-SOD, copper and zinc superoxide dismutase; GSH, glutathione; GSH-Px,
glutathione peroxidase; Mn-SOD, manganese superoxide dismutase; SOD, superoxide
dismutases; TBAR, substance reacting with thiobarbituric acid; UV-A, ultraviolet
A radiation (320-400 nm); UV-A1, UV-A wavelengths from 340 to 400 nm.
CONCLUSION
It is known that UV-A radiation participates in skin carcinogenesis and
photoaging. However, cutaneous cell susceptibility to UV-A radiation is
not clearly established. The solar energy received by skin cells is quite
different in the epidermis and the dermis implicating that damage to epidermal
and dermal cells is probably very different. This work highlights, under
our culture conditions, the different antioxidant status between keratinocytes
and fibroblasts affording better protection to epidermal cells. The epidermis
is the primary target for oxidative stress generated not only by solar
irradiation but also by other physical and chemical agents in the environment.
Whereas fibroblasts are protected by the overlying epidermis, keratinocytes
need constitutively high levels of antioxidant molecules to protect themselves.
Therefore, although cultured cells are good models for studying UV-A-induced
molecular cellular damage and potential antioxidant compounds, it appears
essential to take into account the cells used and to define and standardize
culture and irradiation conditions.
Acknowledgments
This work was supported in part by a grant from Labcatal laboratories
and from "Region Rhône-Alpes" (France). We are grateful to Pr. B.A.
Gilchrest and Dr. H.R. Byers, Boston University, for critically reading
the manuscript. We thank A. Favier and P. Amblard for helpful discussions,
and A.M. Monjo and J. Meo for technical assistance.
REFERENCES
1. Tyrrell RM, Keyse SM. New trends in photobiology. The interactions
of UV-A radiation with cultured cells. J Photochem Photobiol 1990;
4: 349-61.
2. Coohill TP. The effects of the ultraviolet wavelengths of radiation
present in sunlight on human cells in vitro. Photochem Photobiol
1987; 46: 1043-50.
3. Leccia MT, Richard MJ, Béani JC, Faure H, Monjo AM, Cadet
J, Amblard P, Favier A. Protective effects of selenium and zinc on UV-A
damage in human skin fibroblasts. Photochem Photobiol 1993; 58:
548-53.
4. Richard MJ, Portal B, Meo J, Coudray C, Hadjian A, Favier A. Malondialdehyde
kit evaluated for determining plasma and lipoprotein fractions that react
with thiobarbituric acid. Clin Chem 1992; 38: 704-9.
5. Akerboom TP, Sies H. Assay of glutathione, glutathione disulfide
and glutathione mixed disulfides in biological samples. In: Methods
in enzymology Vol. 77. Academic Press, New-York, eds. 1981: 373-82.
6. Gunzler WA, Kremers H, Flohe L. An improved coupled test procedure
for glutathione peroxidase in blood. Z Klin Chem Klin Biochem 1974;
12: 444-8.
7. Marklund SL, Marklund G. Involvement of the superoxide anion radical
in the autooxidation of pyrogallol and a convenient assay of superoxide
dismutase. Eur J Biochem 1974; 47: 469-74.
8. Parat MO, Richard MJ, Leccia MT, Amblard P, Favier A, Béani
JC. Does manganese protect cultured human skin fibroblasts against oxidative
injury by UVA, dithranol and hydrogen peroxide? Free Rad Res 1995;
23: 339-51.
9. Richard MJ, Guiraud P, Monjo AM, Favier A. Development of a simple
antioxidant screening assay using human skin fibroblasts. Free Rad
Res Comms 1992; 16: 303-14.
10. Tyrrell RM, Pidoux M. Singlet oxygen involvement in the inactivation
of cultured human fibroblasts by UV-A (334 nm, 365 nm) and near-visible
(405 nm) radiations. Photochem Photobiol 1989; 49: 407-12.
11. Smith PJ, Paterson MC. Lethality and the induction and repair of
DNA damage in far, mid and near UV-irradiated human fibroblasts: comparison
of effects in normal, xeroderma pigmentosum and Bloom's syndrome cells.
Photochem Photobiol 1982; 36: 333-43.
12. Keyse SM, Moss SH, Davies DJ. Action spectra for inactivation of
normal and xeroderma pigmentosum human skin fibroblasts by ultraviolet
radiations. Photochem Photobiol 1983; 37: 307-12.
13. Tyrrell RM, Pidoux M. Action spectra for human skin cells: estimates
of the relative cytotoxicity of the middle ultraviolet, near ultraviolet,
and violet regions of sunlight on epidermal keratinocytes. Cancer Res
1987; 47: 1825-9.
14. Tyrrell RM, Pidoux M. Correlation between endogenous glutathione
content and sensitivity of cultured human skin cells to radiation at defined
wavelengths in the solar ultraviolet range. Photochem Photobiol
1988; 47: 405-12.
15. Moysan A, Clement-Lacroix PM, Dubertret L, Morliere P. Effects of
ultraviolet A and antioxidant defense in cultured fibroblasts and keratinocytes.
Photodermatol Photoimmunol Photomed 1995; 11: 192-7.
16. Shindo Y, Witt E, Han D, Epstein W, Packer L. Enzymic and non-enzymic
antioxidants in epidermis and dermis of human skin. J Invest Dermatol
1994; 102: 122-4.
17. Applegate LA, Noel A, Vile G, Frenk E, Tyrrell RM. Two genes contribute
to different extents to the heme oxygenase enzyme activity measured in
cultured human skin fibroblasts and keratinocytes: implications for protection
against oxidant stress. Photochem Photobiol 1995; 61: 285-91.
18. Yohn JJ, Norris DA, Yrastorza DG, Buno IJ, Leff JA, Hake SS, Repine
JE. Disparate antioxidant enzyme activities in cultured human cutaneous
fibroblasts, keratinocytes and melanocytes. J Invest Dermatol 1991;
97: 405-9.
19. Vessey DA, Lee K, Boyer TD. Differentiation-induced enhancement
of the ability of cultured human keratinocytes to suppress oxidative stress.
J Invest Dermatol 1995; 104: 355-8.
|
Version imprimable
