Induction of Manganese Superoxide Dismutase in Human Dermal Fibroblasts: Title and subTitle BreakA UV-B–Mediated Paracrine Mechanism With the Release of Epidermal Interleukin 1α, Interleukin 1β, and Tumor Necrosis Factor α

CHRONOLOGICAL (intrinsic) aging affects the skin in a manner similar to other organs.¹ Superimposed on this innate process, extrinsic aging is related to environmental factors, mainly UV-induced photo-oxidative damage of the dermal connective tissue by reactive oxygen species (ROS). There is evidence that these processes, intrinsic and extrinsic aging, have at least in part overlapping biological, biochemical, and molecular mechanisms.² Ultraviolet B (280-320 nm) and UV-A (320-400 nm) radiation are components of terrestrial sunlight that can generate severe oxidative stress in skin cells via interaction with intracellular chromophores and photosensitizers and result in transient and permanent genetic damage and in the activation of cytoplasmic signal transduction pathways that are related to growth, senescence, and degradation of connective tissue. In this regard, UV-generated ROS have been shown to enhance the synthesis and activity of matrix-degrading metalloproteases³ and to contribute to telomere shortening.⁴

The term ROS collectively includes oxygen-centered radicals such as the superoxide anion (O₂⁻) and the hydroxyl radical (HO·) as well as nonradical species such as hydrogen peroxide (H₂O₂) and singlet oxygen (¹O₂), all of which are produced in the skin under UV irradiation. The hypothesis that ROS drive the aging process was forwarded by Harman⁵ and is based on the observation that about 2% of the oxygen taken up by the cell is chemically reduced in the mitochondria by the addition of single electrons and sequentially converted into ROS. In this context, in vitro experiments with fibroblast cultures have shown that α-phenyl-t-butyl nitrone, an antioxidant agent, substantially delays the onset of growth cessation in spontaneously aging fibroblasts.⁶ Furthermore, overexpression of superoxide anion– and hydrogen peroxide–detoxifying antioxidant enzymes (or mutations enhancing the ROS-detoxifying antioxidant enzymes) increased the organismic lifespan in aging models of Drosophila and Caenorhabditis elegans.⁷

Manganese (Mn) superoxide dismutase (SOD) is a nuclear-encoded antioxidant enzyme belonging to a complex interdependent antioxidant enzymatic network that protects cells from oxidative damage by rapidly converting superoxide anions to hydrogen peroxide and molecular oxygen.⁸ Hydrogen peroxide is subsequently detoxified by catalase and glutathione peroxidase.⁹ Manganese SOD is found in prokaryotes as well as eukaryotes and is localized in the mitochondrial matrix. Although MnSOD makes up only 25% of total SOD activity,¹⁰ it stands in the first line of defense against superoxide radicals generated as a by-product of mitochondrial oxidative phosphorylation, UV irradiation, and inflammatory processes. By regulating the homeostasis of oxygen radicals within the mitochondria, MnSOD participates in the control of tumor generation. Recently, MnSOD was shown to be a tumor suppressor,¹¹ and polymorphisms of the MnSOD gene have been found to cause susceptibility for breast cancer^{11 - 12} and colorectal carcinoma.¹³ Manganese SOD has also been identified as a major gerontogene controlling cellular senescence.^{14 - 15} Thus, MnSOD takes a pivotal role in counterbalancing tumorigenesis vs cellular aging.

A functional regulation of MnSOD independent of genetic polymorphisms was also observed. Manganese SOD messenger RNA (mRNA) and activity levels were found to be specifically up-regulated in human fibroblasts under gamma and UV-A irradiation.^{16 - 18} Expression of MnSOD can be induced by its substrate (the superoxide anion), by different cytokines such as interleukin (IL) 1, tumor necrosis factor (TNF) α, and lymphotoxins, or by lipopolysaccharides.¹⁹

We herein determine whether MnSOD is induced in fibroblasts or epidermal cells (HaCaT cells, primary keratinocytes) under UV-B irradiation within the first 24 hours and whether repeated UV-B exposure, as practiced for light hardening during phototherapy of various photodermatoses, may enhance the antioxidant response in the skin. Our results show that human dermal fibroblasts or epidermal cells do not respond with an increase in MnSOD activity directly to single or repeated UV-B irradiation. However, fibroblasts incubated with supernatants from UV-B–irradiated epidermal cells reveal a substantial increase in specific MnSOD mRNA and activity. This increase in fibroblast MnSOD is, at least in part, mediated by the paracrine effects of IL-1α, IL-1β, and TNF-α, which are released from the epidermal cells as soluble mediators in response to UV-B irradiation.

Primary human dermal fibroblasts and keratinocytes were established by outgrowth from foreskin biopsy specimens from healthy human donors.²⁰ Fibroblasts and HaCaT cells were cultured in Dulbecco modified Eagle medium (Gibco BRL, Eggenstein, Germany) with glutamine (2mM), penicillin (400 U/mL), streptomycin (50 µg/mL), and 10% fetal calf serum; keratinocytes were cultured in Keratinocyte-SFM medium (Gibco); each on plastic Petri dishes (6 cm in diameter) in a humified atmosphere at 5% carbon dioxide and 95% air at 37°C.

Fibroblasts, HaCaT cells, and primary keratinocytes were washed with 0.9% sodium chloride solution, collected with a cell scraper, and homogenized by sonification. Manganese SOD activity was subsequently detected using a modified nitroblue tetrazolium (NBT) reduction method.²¹ A xanthine–xanthine oxidase system, generating superoxide anions to reduce NBT at a constant rate of between 0.015 and 0.025 absorbance units per minute, competed with MnSOD superoxide anion dismutation activity. The NBT reduction was monitored spectrophotometrically at 560 nm, with 1 U of MnSOD corresponding to a 50% decrease in NBT reduction. Manganese SOD activity was differentiated from copper/zinc SOD by its resistance to cyanide.

Total RNA was isolated and analyzed by hybridization with a specific 737–base pair complementary DNA or oligonucleotide probes for human MnSOD and 18S ribosomal RNA.^{22 - 23} Two specific 1.0- and 4.0-kilobase (kb) MnSOD mRNA species exist. Whereas other investigators have found that the half-life of the MnSOD-specific transcripts substantially differed,²⁴ members of our group earlier,¹⁷ and herein, did not detect any difference in steady-state mRNA levels of both species under UV-A or UV-B irradiation. Therefore, the present Northern blot analyses show the 4.0-kb MnSOD mRNA species.

A 1000-W xenon high-pressure UV source was used in conjunction with a monochromator with holographic grating (Dermolum UMW; Müller, Moosinning, Germany), as described previously.^{22 ,25} For the experiments, the total UV-B spectrum (280-320 nm) was used, whereas UV-A und UV-C (WG 305/1; Dermolum UMW) wavelengths were cut off using appropriate filtering (Figure 1). The spectral distribution was measured by spectroradiometry with an OL-754 UV/visible light spectroradiometer (Optronic, Orlando, Fla). The fluence rate on the cell surface was 0.41 mW/cm² for the total UV-B spectrum. Fibroblasts, HaCaT cells, or primary keratinocytes were cultured in medium without fetal calf serum for 16 hours prior to irradiation to synchronize cell-cycle phases of the cells.²⁶ Monolayer cultures were rinsed twice with phosphate-buffered saline. Irradiations were performed under a thin layer of phosphate-buffered saline.²² Following UV-B irradiation, fibroblasts and HaCaT cells were cultured in Dulbecco modified Eagle medium without fetal calf serum for measured time periods.

Viability of dermal fibroblasts and HaCaT cells was measured 24 hours after UV-B irradiation: 3-(4,5-dimethylthiazol-2-yl)-2,5,diphenyltetrazolium bromide was used for quantification of metabolically active, viable cells.²⁷ Cytotoxicity was calculated as the percentage of formazan formation in cells 24 hours after UV-B irradiation compared with nonirradiated cells.

HaCaT cells and human primary keratinocytes were irradiated once, twice, or 3 times with 40 mJ/cm² of UV-B at intervals of 24 hours and maintained in unchanged medium between the irradiation cycles. Supernatants of HaCaT cells and human primary keratinocytes were harvested 24 hours after the last irradiation procedure and subjected to enzyme-linked immunosorbent assay (ELISA) for cytokine detection. Concentrations of human IL-1α, IL-1β, and TNF-α were determined using commercially available ELISA kits (R&D Systems, Wiesbaden, Germany) according to the manufacturer's protocols.

Antihuman IL-1α, antihuman IL-1β, and antihuman TNF-α antibodies (R&D Systems) were present at concentrations of 100 µg/mL each during irradiation and collection of HaCaT cell supernatants. Subsequently, the immunoprecipitated cytokines were depleted by centrifuging (600g) in the presence of 25 µL/mL of protein G agarose (Boehringer, Mannheim, Germany) for 15 minutes.²⁸ All supernatants depleted from the above-mentioned cytokines were subjected to ELISA with the corresponding specificity. No IL-1α, IL-1β, or TNF-α could be detected, which suggests that all cytokines had been completely removed (data not shown).

To study the effect of UV-B irradiation on MnSOD activity in human dermal fibroblasts and epidermal cells (HaCaT cells and primary keratinocytes), cells were irradiated with 30 mJ/cm² of UV-B, and subsequently cell lysates from the separately cultured cells were prepared at different times within the first 24 hours after irradiation. Since we intended to avoid cytotoxicity from UV-B irradiation, UV-B was used at nontoxic doses, which was the case at a range of 30 to 40 mJ/cm², yielding a viability of more than 75% as detected by 3-(4,5-dimethylthiazol-2-yl)-2,5,diphenyltetrazolium bromide assays. As a result, no increase in MnSOD activity could be detected in human dermal fibroblasts at any time after UV-B irradiation with 30 mJ/cm² (Figure 2A). Similarly, no induction of MnSOD activity was observed at any time after irradiation in HaCaT cells or primary keratinocytes (Figure 2B and C).

Since we could not detect any direct MnSOD induction under UV-B irradiation in human dermal fibroblasts or epidermal cells, which indicates an immediate adaptive antioxidant response within the first 24 hours, we investigated the role of soluble mediators possibly released by the epidermal cells under UV-B irradiation that could indirectly lead to an up-regulation of MnSOD in human dermal fibroblasts. Therefore, we incubated nonirradiated human dermal fibroblasts with the supernatants of once, twice, or 3-times UV-B–irradiated HaCaT cells (40 mJ/cm²). In one group, HaCaT cells were cultured in fresh medium after each irradiation; in the other group, HaCaT cells were maintained in the same medium for the whole irradiation period. Twenty-four hours after the last UV-B irradiation, supernatants were harvested. As a control, supernatants of nonirradiated HaCaT cells were also collected. Nonirradiated human dermal fibroblasts were then incubated with the collected supernatants. Here, an incubation period of 2 hours was chosen because initial experiments with recombinant proinflammatory cytokines (IL-1α, IL-1β, and TNF-α) had revealed a marked up-regulation of MnSOD mRNA and activity already after an intermittent cytokine supplementation lasting 2 hours. Subsequently, total RNA was isolated from the human dermal fibroblasts, and specific MnSOD mRNA levels were assessed. In comparison with the control, where human dermal fibroblasts were incubated with the supernatants of nonirradiated HaCaT cells, a significant increase in MnSOD mRNA was found in human dermal fibroblasts after incubation with the supernatants of UV-B-irradiated HaCaT cells. Since MnSOD expression in human dermal fibroblasts was higher in proportion to the number of times the HaCaT cells were irradiated and maintained in unchanged medium, this pointed toward an accumulation of soluble MnSOD-inducing mediators in the medium (Figure 3).

To investigate whether MnSOD activity in fibroblasts would also increase after incubation with the supernatants of irradiated HaCaT cells, the following experiment was performed. As shown above, MnSOD mRNA levels were increased most markedly after being irradiated 3 times when HaCaT medium was not changed. Therefore, an identical experimental setting was used also for the assessment of MnSOD activity levels. The HaCaT cells were irradiated 3 times with 40 mJ/cm² of UV-B and the supernatants were collected. Nonirradiated fibroblasts were then incubated with these supernatants for 1, 2, 4, 8, 12, and 24 hours. As controls, fibroblasts were incubated with the supernatants of nonirradiated HaCaT cells for 24 hours or were kept in their own medium without any transfer of supernatants. Manganese SOD activity was assessed as described above. Compared with control fibroblasts, fibroblasts incubated with the supernatants of repetitively irradiated HaCaT cells showed a significant increase in MnSOD activity, which was dependent on the incubation period of human dermal fibroblasts with the supernatants (Figure 4).

Initial experiments showed that repetitive UV-B irradiation of the epidermal HaCaT cells as well as of primary human keratinocytes led to a markedly increased release of IL-1α, IL-1β, and TNF-α, which have been shown to induce MnSOD. To investigate whether these cytokines were responsible for the increase in MnSOD mRNA and activity in our experimental setting, we repetitively irradiated HaCaT cells with UV-B. After each irradiation procedure, cells were put back on the same conditioned medium they were grown in before the irradiation to allow accumulation of cytokines. Supernatants of nonirradiated HaCaT cells served as negative controls. Interleukin 1α, IL-1β, and TNF-α were depleted from the supernatants of HaCaT cells with neutralizing antibodies. Subsequently, human dermal fibroblasts were incubated with the cytokine-depleted supernatants for 2 hours. As a result, MnSOD mRNA levels were significantly decreased after the simultaneous depletion of IL-1α, IL-1β, and TNF-α (Figure 5), which indicates that these cytokines were, at least in part, responsible for the induction of MnSOD expression and activity in fibroblasts.

The increasing incidence of skin malignancies and photoaging is considered to be at least in part due to UV-generated ROS. Both UV-A and UV-B irradiation in conjunction with intracellular photosensitizers have been shown to generate high levels of ROS, even hours and days after irradiation.^{4 ,29} This is particularly relevant today because UV irradiation is predicted to increase on the surface of the earth with depletion of the ozone layer.^{30 - 31} Therefore, an understanding of the antioxidant enzymes within resident skin cells and their regulation of UV-generated ROS or other signaling pathways is essential for our understanding of UV-protective mechanisms in the skin, which is urgently required for proper individual risk assessment and improved strategies for protection of sun-exposed skin.

The overall aim of this study was therefore to define adaptive antioxidant enzymatic mechanisms that confer immediate protection to resident cells against the prooxidant attack following UV-B irradiation. We found that, unlike UV-A irradiation,¹⁷ different doses of UV-B irradiation of fibroblasts, epidermal HaCaT cells, and primary human keratinocytes did not result in an up-regulation of MnSOD activity within the first 24 hours. By contrast, UV-B irradiation of HaCaT cells or keratinocytes and subsequent transfer of the corresponding supernatants to fibroblast monolayer cultures resulted in a significant increase in MnSOD mRNA and activity levels in fibroblasts. Furthermore, we found that incubation of fibroblasts with supernatants from repetitively low-dose UV-irradiated epidermal cells could even more dramatically enhance the synthesis and activity of MnSOD. This may be particularly relevant, as the MnSOD induction in fibroblasts confers a substantial protection against the cytotoxic effect of a subsequent high-dose UV-A insult.¹⁷

Epidermally released IL-1α, IL-1β, and TNF-α were identified to mediate, at least in part, the MnSOD-inducing effect in human dermal fibroblasts, suggesting a paracrine mechanism in the epidermal-dermal interaction to be responsible for an adaptive antioxidant response. Even though UV-A is about 10 times as prevalent as UV-B in terrestrial sunlight, UV-B is far more biologically active, by a ratio of approximately 1000:1. Thus, besides the previously discovered adaptive antioxidant response directly conferred by UV-A irradiation¹⁷ that protects from cytotoxic effects of subsequent high-dose UV-A irradiation, we hypothesized that UV-B irradiation may also contribute to an immediate antioxidant protective mechanism in human dermal fibroblasts. The UV-B doses administered in our experiments are physiologically relevant and can be easily acquired when exposed to sunlight,^{32 - 33} thus underlining the relevance of our findings.

However, recent observations by Leccia and coworkers³⁴ have demonstrated that UV-B also was a potent direct inducer of MnSOD activity in cultured human dermal fibroblasts, though with a delay of approximately 48 hours after UV-B irradiation, an observation confirmed by our preliminary data. However, we herein have focused on the immediate antioxidant effects that occur no later than 24 hours after UV-B irradiation.

It has been noted that long-term exposure of hairless mice for 2 hours per day to a source emitting mostly UV-A with 2% UV-B led to a substantial increase in SOD activity. However, after continued irradiation for 24 weeks, SOD activity decreased below the level of mock-treated animals,³⁵ whereas glutathione peroxidase activity remained elevated. These results suggest that long-term UV exposure of animals for months, even at suberythemal doses, may compromise the SOD-dependent antioxidant defense. It remains to be determined whether short-term repetitive UV-B irradiation—as observed in our in vitro experiments—may also result in an increase in MnSOD activity in fibroblasts in human skin in vivo and whether repetitive irradiation over months—similar to the findings in mice—will also compromise SOD activity in human skin.

A failure of proper detoxification of enhanced ROS levels by antioxidant enzymes drives intrinsic and extrinsic aging.^{4 ,14} Recently, low MnSOD activity has been shown to control the age-related decline of mitochondrial function.¹⁵ Accordingly, mitochondria from mice with a heterozygous deficiency in MnSOD showed evidence for an early and rapid accumulation of mitochondrial oxidative damage. By contrast, mice with normal MnSOD levels show the same age-related mitochondrial decline as the heterozygotes, but occurring later in life.¹⁵ Thus, adaptive up-regulation of MnSOD, as has previously been reported for UV-A irradiation and as we herein show for UV-B irradiation, effectively counterbalances increased superoxide anion fluxes and thus may reduce aging processes.

As to the regulation of the induced synthesis and activity of MnSOD after transfer of supernatants from UV-B–irradiated epidermal cells on dermal fibroblasts, it is most likely that IL-1α, IL-1β, and TNF-α, all released following UV-B irradiation of epidermal HaCaT cells and human primary keratinocytes, are required to mediate the observed paracrine UV-B effects.^{36 - 37} In fact, IL-1α, IL-1β, and TNF-α are well-known stimuli of the MnSOD.^{16 ,38} In addition, our data show that neutralizing antibodies against IL-1α, IL-1β, and TNF-α can significantly suppress the induction of MnSOD activity in fibroblasts after transfer of supernatants from previously UV-B–irradiated epidermal cells. It is not yet clear, however, whether superoxide anions and IL-1 represent completely independent mechanisms or act as causally related sequential events. There is some evidence that IL-1 can induce the generation of ROS in fibroblasts.^{39 - 40}

Strong differences exist in the spontaneous activity and inducibility of MnSOD under stimulation with IL-1.¹⁷ In conjunction with the present findings, this suggests a model where the paracrine inducibility of MnSOD is involved as a molecular determinant conferring differences in interindividual susceptibility for the development of photosensitivity, premature aging, and skin cancer. Because increased ROS fluxes play a major role in aging by driving gene expression pathways related to collagen degradation and elastin accumulation,⁴¹ the adaptive antioxidant response with autocrine or paracrine up-regulation of distinct antioxidant enzymes like MnSOD in human dermal fibroblasts may effectively counteract cellular and connective tissue aging. Furthermore, understanding the paracrine mechanism of MnSOD induction advances our understanding of phototherapeutic light hardening, which is commonly used for the treatment of photodermatoses such as polymorphic light eruptions and solar urticaria. Seeing that UV-A irradiation leads to a direct up-regulation of protective antioxidant enzymes in dermal fibroblasts¹⁷ while UV-B indirectly mediates an immediate induction of such enzymes in a paracrine fashion may help to further fine-tune and optimize phototherapeutic protocols.

Accepted for publication July 30, 2002.

This work was supported by grants DFG SCHA411/10-2 and DFG SCHA411/10-2 from the German Research Foundation, Bonn, Germany (Dr Scharffetter-Kochanek), and grant BMBF 07UVB55B/1 from the Bundesministerium für Bildung und Forschung, Berlin, Germany (Dr Scharffetter-Kochanek).

We wish to thank Irene Smith, PhD (Gentech, San Francisco, Calif), for supplying the complementary DNA clone for human MnSOD.

Corresponding author and reprints: Karin Scharffetter-Kochanek, MD, Department of Dermatology and Allergology, University of Ulm, Maienweg 12, 89081 Ulm, Germany (e-mail: karin.scharffetter-kochanek@medizin.uni-ulm.de).