Oncogene (2006) 25, 2601–2614
& 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00
ORIGINAL ARTICLE
The UVB-induced gene expression profile of human epidermis in vivo is
different from that of cultured keratinocytes
CD Enk1, J Jacob-Hirsch2, H Gal3, I Verbovetski4, N Amariglio2, D Mevorach4, A Ingber1,
D Givol3, G Rechavi2 and M Hochberg1
1
Department of Dermatology, The Hadassah-Hebrew University Medical Center, Jerusalem, Israel; 2Department of Pediatric
Hemato-Oncology and Functional Genomics, Safra Children’s Hospital, Sheba Medical Center and Sackler School of Medicine,
Tel-Aviv University,Tel Aviv, Israel; 3Department of Molecular Cell Biology, Weizmann Institute of Science, Rehovot, Israel
and 4The Laboratory for Cellular and Molecular Immunology, Department of Medicine, The Hadassah-Hebrew University Medical
Center, Jerusalem, Israel
In order to obtain a comprehensive picture of the
molecular events regulating cutaneous photodamage of
intact human epidermis, suction blister roofs obtained
after a single dose of in vivo ultraviolet (UV)B exposure
were used for microarray profiling. We found a changed
expression of 619 genes. Half of the UVB-regulated genes
had returned to pre-exposure baseline levels at 72 h,
underscoring the transient character of the molecular
cutaneous UVB response. Of special interest was our
finding that several of the central p53 target genes
remained unaffected following UVB exposure in spite of
p53 protein accumulation. We next compared the in vivo
expression profiles of epidermal sheets to that of cultured
human epidermal keratinocytes exposed to UVB in vitro.
We found 1931 genes that differed in their expression
profiles between the two groups. The expression profile in
intact epidemis was geared mainly towards DNA repair,
whereas cultured keratinocytes responded predominantly
by activating genes associated with cell-cycle arrest and
apoptosis. These differences in expression profiles might
reflect differences between mature differentiating kerati-
nocytes in the suprabasal epidermal layers versus
exponentially proliferating keratinocytes in cell culture.
Our findings show that extreme care should be taken when
extrapolating from findings based on keratinocyte cultures
to changes in intact epidermis.
Oncogene (2006) 25, 2601–2614. doi:10.1038/sj.onc.1209292;
published online 23 January 2006
Keywords: UVB; epidermis; p53; in vivo; in vitro;
microarrays
Introduction
The skin constitutes the outer barrier of the human
organism towards external stimuli and ultraviolet (UV)
Correspondence: Dr CD Enk, Department of Dermatology, Hadassah
Medical Center, P.O. Box 12000, IL 91010, Israel.
E-mail: enk@md.huji.ac.il
Received 11 July 2005; revised 6 October 2005; accepted 28 October
2005; published online 23 January 2006
www.nature.com/onc
radiation. UVB, with a wavelength range between 290
and 320 nm, represents one of the most important
environmental hazards affecting human skin (Hahn
and Weinberg, 2002). To protect itself against the
DNA-damaging effects of sunlight, the skin disposes
over highly complicated cellular programs, including
cell-cycle arrest, DNA repair and apoptosis (Brash et al.,
1996). Failure in selected elements of these defensive
strategies may result in propagation of cutaneous
cancers. Handling of photodamage by the skin constitu-
tes a crucial survival mechanism, and the regulation and
activity of cutaneous UVB-regulated genes represents an
attractive model for the functioning of central protective
cellular responses under physiological conditions.
The advent of microarray technology allows the
generation of global expression profiles and differential
expression of thousands of genes (Park et al., 2002; Joos
et al., 2003; Sellheyer and Belbin, 2004). Several recent
studies have employed microarray profiling to study
UVB-regulated gene expression in cultured human
keratinocytes (Becker et al., 2001; Li et al., 2001;
Murakami et al., 2001; Sesto et al., 2002; Takao et al.,
2002; Dazard et al., 2003; Pisarchik et al., 2004; Lee
et al., 2005), revealing that a wide range of genes
regulating transcription factors, cytoskeletal proteins,
secreted signaling molecules and controlling cellular
functions such as signal transduction, terminal differ-
entiation and apoptosis, are induced by UVB. However,
since these studies differ considerably in terms of time
kinetics, UVB dose, light sources and hybridization
techniques, it is difficult to generate a comprehensive
picture of the complex changes in gene expression taking
place in UVB-irradiated keratinocytes.
An additional drawback common to all these studies
is the fact that they were performed on cultured
keratinocytes under in vitro conditions. Skin tissue
cultures differ substantially from the intact tissues they
are supposed to imitate: Keratinocytes are usually
cultured in the presence of hormones and growth factors
that might affect cell-cycle regulation, stress response
and apoptosis (Gibbs et al., 1998), thereby modifying
the response to UVB. Furthermore, confluence
in keratinocyte cultures affects the regulation of
 UVB-induced gene expression profile of human epidermis
CD Enk et al
2602
programmed cell death, and a significant difference in
sensitivity to UVB-induced apoptosis between subcon-
fluent cultures and exponentially growing cells have
been reported (Gniadecki et al., 1997). Finally, keratino-
cyte cultures constitute a homogeneous population of
cells lacking the additional cellular constituents of intact
epidermis such as Langerhans cells and intraepidermal
inflammatory cells that together with the keratinocytes
play important roles in the response of human epidermis
to UV exposure (Baadsgaard, 1991).
Using a unique model system based on UV-exposed
human epidermis followed by isolation of epidermal
cells from suction blister roofs, we have recently studied
the global expression profile in intact human epidermis
using oligonucleotide microarrays (Enk et al., 2004).
With this in vivo setup that overcomes several of the
limitations outlined above, we identified more than 800
UVB-regulated genes, some of which not previously
known to be UVB sensitive. In the present study, we
expand this investigation by monitoring the differen- Figure 1 Number up- and downregulated genes at different time-
tially expressed genes over a 72 h period following UVB points following UVB irradiation of intact human epidermis.
exposure. We found that the majority of the in vivo
UVB-regulated genes changed their expression at the
24 h time-point, but had returned to background levels
within 72 h. We next compared the expression profile of sion profiles (Kannan et al., 2001). Applying this
the in vivo exposed epidermal cells to that of cultured method to genes that were at least twofold changed in
UVB-exposed normal human epidermal keratinocytes all volunteers at least at one time-point, triggered
(NHEK), thus providing the first direct comparison distinct, well-ordered gene expression and partitioned
between the UV response of intact epidermis and the samples according to interval after UV exposure
keratinocyte cultures. We here report that the expression rather than according to individuals (Figure 2). This
profile in intact epidermis was geared mainly towards profiling pattern testifies to the robustness of the data set
DNA repair, whereas cultured keratinocytes responded and indicates that our results reflect an overall UVB
predominantly by activating genes associated with cell- response. The algorithm showed that the most closely
cycle arrest and apoptosis. related expression profiles were those of the 0, 2 and 72
time-points, whereas the 24 h expression profile dis-
played greatest dissimilarity from the other time-points.
Results and Discussion This suggests that the majority of UVB-induced changes
in epidermal gene expression is transient and has
UVB-induced gene expression reversed to background expression within 72 h. The
To study the global gene expression profile in intact finding that the early UVB response involves a relatively
human epidermis following a single physiological UVB small number of genes, whereas the 24 h response is
exposure, we exposed the inner forearms of three more complex and expresses a larger number of genes
volunteers to 4 MED of UVB in vivo. Suction blisters corresponds well with the recent findings of Lee et al.
were raised at 2, 24 and 72 h following exposure, and (2005) who examined the differentially expressed genes
mRNA extracted from the blister roofs underwent gene of UVB-irradiated immortalized HaCat keratinocytes at
profiling using Affymetrix Human Focus oligonucleotide 0.5, 6 and 24 h.
arrays as described in Methods. Applying an arbitrary The dendogram revealed eight major, stable clusters
filter level of twofold changes in the ratios of gene containing different expression kinetics of up- and
expression in all volunteers at least at one time-point, we downregulated genes at 2, 24 and 72 h after UV
found 619 genes whose expression was modified by UVB exposure (Figure 2). Interestingly, in five of the eight
(Figure 1), of which 246 genes were upregulated and 373 clusters (cluster 1, 2, 5, 7 and 8), representing 316 of the
were downregulated, representing approximately 12% of 619 UVB-regulated genes (51%), the expression pattern
the ‘valid’ genes (1.5–2% of the human genome). In our had returned to pre-exposure baseline levels a 72 h,
list, the majority, 53% of the differentially expressed again underscoring the transient character of the
genes, were found at the 24 h time-point, whereas 21 and genomic changes taking place in human skin following
26% of the UVB-regulated genes were found 2 and 72 h, a single, low-dose, UVB exposure.
respectively, following exposure.

Gene clustering Functional groups

Clustering the DNA chip data by an unsupervised UVB-responsive genes were categorized into clusters
approach organizes genes by similarities in their expres- based on known or presumed functions. Table 1 depicts
Oncogene
 UVB-induced gene expression profile of human epidermis
CD Enk et al
2603

Figure 2 Cluster algorithm showing clustering of the 619 genes with at least twofold expression level changes in all volunteers in at
least one time-point following in vivo UVB exposure of intact human epidermis. The normalized expression level for each gene (rows)
in each sample (columns) is indicated by a color code (green downregulated and red upregulated). When applying an unsupervised
algorithm, the samples were partitioned according to interval after UV exposure and showed greatest similarities among the 0, 2 and
72 h expression patterns. Z-score analysis grouped the genes into eight kinetically distinguishable clusters. The number of genes in each
cluster is indicated. In five of the clusters, representing half of the UVB-modulated genes, the expression pattern had returned to
baseline patterns at 72 h.

selected genes of interest (http://www.hadassah.org.il/
Departments/photobiology; for the full list). The 619
differentially expressed genes were found in a wide range
of functional categories at one or more time-points after
UVB irradiation.
S100. Several skin disorders have been associated with
the S100 family of calcium-binding proteins, including
disorders of keratinization, proliferation and differentia-
tion. The S100A7, S100A8 and S100A9 protein genes are
absent or minimally expressed in normal epidermis, but
are markedly overexpressed in psoriatic lesions, in
keratinocytes with abnormal differentiation and in cuta-
neous tumors (Boni et al., 1997; Alowami et al., 2003;
Broome et al., 2003). S100A8 and S100A9 gene expression
is upregulated after injury of murine and human epidermis
(Thorey et al., 2001), and Grimbaldeston et al. (2003)
demonstrated that S100A8, but not S100A9, was over-
expressed in the epidermis of UVA-irradiated BALB/c
mice in response to reactive oxygen species (ROS). We
here report upregulation of S100A6, S100A7, S100A8,
S100A9 and S100P protein genes in human epidermis
following UVB radiation. All S100 protein genes showed
the same pattern of expression with upregulation at 24 h
and sustained activity at 72 h. S100A2 is expressed in
benign nevi but not in metastatic melanomas, and loss of
S100A2 activity might play a role in tumorigenesis and
constitute an early event in melanoma development
(Maelandsmo et al., 1997). In normal human keratino-
cytes, S100A2 nuclear staining disappears following
treatment with H2O2, and S100A2 protein might be
involved in keratinocyte responses to ROS (Zhang et al.,
2002). The downregulation of S100A2 expression follow-
ing UVB radiation reported in our study could thus result
from the activity of free radicals on the epidermal cells.
Serine protease inhibitors. Our data show that a
number of serine proteinase inhibitors were induced
following UVB irradiation. The squamous cell carci-
noma antigen (SCCA1 or SERPINB3) plays a role in
differentiation of squamous epithelium (Crombach
et al., 1989; Kato, 1996) and significantly inhibits certain
forms of apoptosis (Suminami et al., 2000). In human
skin, SCCA is expressed in epidermis overlying inflam-
matory lesions such as psoriasis and dermatitis (Hor-
iuchi et al., 1994; Takeda et al., 2002). There is no
consensus regarding the expression in normal epidermis
(Duk et al., 1989; Cataltepe et al., 2000; Takeda et al.,
2002). Our data show that SCCA1 is dramatically
increased 24 and 72 h after UVB exposure. Of specific
interest is our finding of moderate but significant
upregulation of serpinB13 (hurpin/headpin/PI13). This
new member of the human ov-serpin family has
previously been demonstrated in the human immor-
talized HaCaT keratinocyte cell line, in cultured normal
human keratinocytes, in lesional skin from psoriatic
Oncogene
 UVB-induced gene expression profile of human epidermis
CD Enk et al
2604
Table 1 Gene expression changes in intact human epidermis following in vivo UVB irradiation
GenBank Symbol Gene/protein Fold change from
nonirradiated
2h 24 h 72 h

Epidermal differentiation complex, calcium-binding proteins

NM_005978.2 S100A2 S100 calcium-binding protein A2 0.34 1.75 1.76
NM_014624.2 S100A6 S100 calcium-binding protein A6 (calcyclin) 0.79 1.64 4.24
NM_002963.2 S100A7 S100 calcium-binding protein A7 (psoriasin 1) 0.69 22.43 12.02
NM_002964.2 S100A8 S100 calcium-binding protein A8 (calgranulin A) 1.32 14.57 12.52
NM_002965.2 S100A9 S100 calcium-binding protein A9 (calgranulin B) 0.73 22.04 15.19
NM_005980.1 S100P S100 calcium-binding protein P 1.30 3.61 3.53

Serpins (serine proteases inhibitors)

AF119873.1 SERPINA1 Serine (or cysteine) proteinase inhibitor, clade A (alpha-1 antiproteinase, 3.21 7.22 2.19
antitrypsin), member 1
AI554300 SERPINB1 Serine (or cysteine) proteinase inhibitor, clade B (ovalbumin), member 1 0.45 3.95 0.99
AF169949.1 SERPINB13 Serine (or cysteine) proteinase inhibitor, clade B (ovalbumin), member 13 1.03 1.97 2.69
BC005224.1 SERPINB3 Serine (or cysteine) proteinase inhibitor, clade B (ovalbumin), member 3 0.89 13.11 9.91

Extracellular matrix affecting factors

NM_006665.1 HPSE Heparanase 0.79 5.12 3.95
NM_002421.2 MMP1 Matrix metalloproteinase 1 (interstitial collagenase) 0.18 15.56 2.57
NM_002422.2 MMP3 Matrix metalloproteinase 3 (stromelysin 1, progelatinase) 0.10 4.61 1.03
NM_003254.1 TIMP1 Tissue inhibitor of metalloproteinase 1 (erythroid potentiating activity, 1.22 5.43 3.41
collagenase inhibitor)
L10343 SKALP/PI3 Elafin: neutrophil and pancreatic elastase-specific inhibitor of skin 0.52 21.55 14.19

Cell cycle
NM_001211.2 BUB1B BUB1 budding uninhibited by benzimidazoles 1 homolog beta (yeast) 1.01 0.63 2.24
NM_001237.1 CCNA2 Cyclin A2 2.31 1.67 8.50
NM_004701.2 CCNB2 Cyclin B2 1.46 0.92 5.90
NM_001759.1 CCND2 Cyclin D2 0.93 1.49 4.00
AW134535 CCNG2 Cyclin G2 0.92 2.70 1.05
AL524035 CDC2 Cell division cycle 2, G1 to S and G2 to M 0.92 0.46 4.19
NM_001255.1 CDC20 CDC20 cell division cycle 20 homolog (Saccharomyces cerevisiae) 2.29 1.17 6.12
NM_021873.1 CDC25B Cell division cycle 25B 1.71 1.36 3.61
AF213033.1 CDKN3 Cyclin-dependent kinase inhibitor 3 (CDK2-associated dual specificity 1.70 0.94 5.40
phosphatase)
NM_001826.1 CKS1B CDC28 protein kinase regulatory subunit 1B 1.42 0.89 2.02
NM_001953.2 ECGF1 Endothelial cell growth factor 1 (platelet-derived) 1.27 4.40 4.85
NM_022346.1 HCAP-G Chromosome condensation protein G 1.05 0.96 2.52
NM_006101.1 HEC Highly expressed in cancer, rich in leucine heptad repeats 4.15 3.77 21.36
NM_002358.2 MAD2L1 MAD2 mitotic arrest deficient-like 1 (yeast) 1.66 0.97 3.86
NM_004526.1 MCM2 MCM2 minichromosome maintenance deficient 2, mitotin (S. cerevisiae) 1.38 1.07 3.13
NM_002388.2 MCM3 MCM3 minichromosome maintenance deficient 3 (S. cerevisiae) 1.34 0.85 2.26
AA807529 MCM5 MCM5 minichromosome maintenance deficient 5, cell division cycle 46 1.72 0.72 3.22
(S. cerevisiae)
NM_005915.2 MCM6 MCM6 minichromosome maintenance deficient 6 (MIS5 homolog, S. pombe) 1.52 1.56 3.13
(S. cerevisiae)
NM_002592.1 PCNA Proliferating cell nuclear antigen 1.77 1.78 3.38
BC001422.1 PGF Placental growth factor, vascular endothelial growth factor-related protein 0.23 13.02 9.11
NM_005030.1 PLK Polo-like kinase (Drosophila) 1.44 1.52 3.89
NM_000946.1 PRIM1 Primase, polypeptide 1, 49 kDa 2.08 0.91 2.43
NM_021000.1 PTTG1 Pituitary tumor-transforming 1 1.29 0.79 3.76
BC003186.1 Pfs2 DNA replication complex GINS protein PSF2 1.93 0.82 4.69
BC001866.1 RFC5 Replication factor C (activator 1) 5, 36.5 kDa 1.78 1.20 2.29
NM_006397.1 RNASEH2A Ribonuclease H2, large subunit 2.11 0.87 2.83
BC005264.1 RPA3 Replication protein A3, 14 kDa 1.71 1.15 3.05
BC001886.1 RRM2 Ribonucleotide reductase M2 polypeptide 1.81 0.77 8.55
NM_014264.1 STK18 Serine/threonine kinase 18 0.80 0.67 2.49
NM_003236.1 TGFA Transforming growth factor, alpha 0.54 2.80 1.54
AF098158.1 TPX2 TPX2, microtubule-associated protein homolog (Xenopus laevis) 1.26 0.41 2.57
NM_001952.2 E2F6 E2F transcription factor 6 1.01 0.63 2.24
NM_003620.1 PPM1D Protein phosphatase 1D magnesium-dependent, delta isoform 2.31 1.67 8.50
AF022375.1 VEGF Vascular endothelial growth factor 1.46 0.92 5.90
NM_002094.1 GSPT1 G1 to S phase transition 1 0.93 1.49 4.00
NM_014574.1 STRN3 Striatin, calmodulin binding protein 3 0.92 2.70 1.05
NM_002748.1 MAPK6 Mitogen-activated protein kinase 6 0.92 0.46 4.19
NM_002467.1 MYC v-myc myelocytomatosis viral oncogene homolog (avian) 2.29 1.17 6.12
NM_002266.1 KPNA2 Karyopherin alpha 2 (RAG cohort 1, importin alpha 1) 1.71 1.36 3.61
NM_006732.1 FOSB FBJ murine osteosarcoma viral oncogene homolog B 1.70 0.94 5.40

Oncogene
 UVB-induced gene expression profile of human epidermis
CD Enk et al
2605
Table 1 (continued )
GenBank Symbol Gene/protein Fold change from
nonirradiated
2h 24 h 72 h

NM_005197.1 CHES1 Checkpoint suppressor 1 1.42 0.89 2.02

N33167 CDKN1C Cyclin-dependent kinase inhibitor 1C (p57, Kip2) 1.27 4.40 4.85
M73554.1 CCND1 Cyclin D1 (PRAD1: parathyroid adenomatosis 1) 1.05 0.96 2.52
Z24459 MTCP1 Mature T-cell proliferation 1 4.15 3.77 21.36
NM_001253.1 CDC5L CDC5 cell division cycle 5-like (S. pombe) 1.66 0.97 3.86
L20320.1 CDK7 Cyclin-dependent kinase 7 (MO15 homolog, Xenopus laevis, cdk-activating kinase) 1.38 1.07 3.13

Repair
D21089.1 XPC Xeroderma pigmentosum, complementation group C 2.52 2.39 1.81
U64315.1 ERCC4 Excision repair cross-complementing rodent repair deficiency, complementation 0.71 0.80 1.42
(XPF) group F

Histones
BC002649.1 HIST1H1C Histone 1, H1c 2.31 1.22 2.57
NM_021052.1 HIST1H2AE Histone 1, H2ae 6.86 2.41 1.76
NM_003523.1 HIST1H2BE Histone 1, H2be 4.09 3.67 2.48
NM_003522.1 HIST1H2BF Histone 1, H2bf 4.67 2.98 1.42
NM_003525.1 HIST1H2BI Histone 1, H2bi 3.43 2.76 2.14
BC000893.1 HIST1H2BK Histone 1, H2bk 1.32 2.62 1.34
AA451996 HIST2H2AA Histone 2, H2aa 3.83 2.00 0.94
NM_003528.1 HIST2H2BE Histone 2, H2be 6.16 1.90 0.71

Apoptosis
NM_001168.1 BIRC5 Baculoviral IAP repeat-containing 5 (survivin) 2.22 0.47 7.54
NM_000633 BCL2 B-cell CLL/lymphoma 2 1.24 0.49 0.66
U66879 BAD BCL2-antagonist of cell death 1.39 0.71 1.37
NM_004281 BAG3 BCL2-associated athanogene 3 0.49 0.66 0.75
NM_001188 BAK1 BCL2-antagonist/killer 1 0.88 1.87 1.37
AA457021 BAG5 BCL2-associated athanogene 5 0.76 0.63 0.45
AF060922 BNIP3L BCL2/adenovirus E1B 19 kDa interacting protein 3-like 0.85 0.81 2.06
NM_004346 CASP3 Caspase 3, apoptosis-related cysteine protease 0.59 0.88 0.64
U25804 CASP4 Caspase 4, apoptosis-related cysteine protease 1.31 1.64 1.48
BF439983 CASP8 caspase 8, apoptosis-related cysteine protease 1.33 1.97 1.28
NM_003655.1 CBX4 Chromobox homolog 4 (Pc class homolog, Drosophila) 1.85 0.40 0.66
NM_015322.1 FEM1B fem-1 homolog b (Caenorhabditis elegans) 0.24 0.99 0.81
NM_002015.2 FOXO1A Forkhead box O1A (rhabdomyosarcoma) 0.40 0.53 0.70
NM_001562.1 IL18 Interleukin 18 (interferon-gamma-inducing factor) 1.16 0.35 1.55
NM_007350.1 PHLDA1 Pleckstrin homology-like domain, family A, member 1 0.20 2.60 2.55
NM_014456.1 PDCD4 Programmed cell death 4 (neoplastic transformation inhibitor) 1.15 0.21 0.78
NM_006378.1 SEMA4D Sema domain, immunoglobulin domain (Ig), transmembrane domain (TM) and 0.93 0.36 0.83
short cytoplasmic domain, (semaphorin) 4D
NM_004760.1 STK17A Serine/threonine kinase 17a (apoptosis-inducing) 0.18 1.26 0.53
NM_004226.1 STK17B Serine/threonine kinase 17b (apoptosis-inducing) 0.27 0.42 0.50
BF224259 SPF30 Splicing factor 30, survival of motor neuron-related 0.35 0.90 0.54
NM_006290.1 TNFAIP3 Tumor necrosis factor, alpha-induced protein 3 0.41 0.62 0.49
NM_006048.1 UBE4B ubiquitination factor E4B (UFD2 homolog, yeast) 1.32 0.45 0.66
NM_002467.1 MYC v-myc myelocytomatosis viral oncogene homolog (avian) 0.43 0.16 0.31
NM_004052.2 BNIP3 BCL2/adenovirus E1B 19 kDa interacting protein 3 1.31 1.86 3.27
NM_001279.1 CIDEA Cell death-inducing DFFA-like effector a 0.89 1.36 2.99
AF053641.1 CSE1L CSE1 chromosome segregation 1-like (yeast) 1.43 2.23 2.54
NM_004938.1 DAPK1 Death-associated protein kinase 1 1.14 5.41 2.45
NM_004944.1 DNASE1L3 Deoxyribonuclease I-like 3 0.92 7.81 3.13
NM_002462.1 MX1 Myxovirus (influenza virus) resistance 1, interferon-inducible protein p78 (mouse) 1.61 4.07 5.14
BC001808.1 NME6 Nonmetastatic cells 6, protein expressed in (nucleoside-diphosphate kinase) 1.17 2.99 2.22
NM_007350.1 PHLDA1 Pleckstrin homology-like domain, family A, member 1 0.20 2.60 2.55
AF016266.1 TNFRSF10B TRAIL, tumor necrosis factor receptor superfamily, member 10b 0.34 2.33 0.78

Cytokines and chemokines

NM_004887.1 CXCL14 Chemokine (C-X-C motif) ligand 14 1.26 0.21 0.87
NM_006729.1 DIAPH2 Diaphanous homolog 2 (Drosophila) 0.98 0.16 0.58
AF196186.1 PARD3 Par-3 partitioning defective 3 homolog (C. elegans) 1.14 0.34 1.01
NM_020299.1 AKR1B10 Aldo-keto reductase family 1, member B10 (aldose reductase) 1.90 10.23 11.95
NM_001175.1 ARHGDIB Rho GDP dissociation inhibitor (GDI) beta 1.03 2.94 1.83
BG327863 CD24 CD24 antigen (small-cell lung carcinoma cluster 4 antigen) 0.68 2.82 1.84
M92934.1 CTGF Connective tissue growth factor 1.60 0.99 4.65
NM_004942.2 DEFB4 Defensin, beta 4 1.18 59.75 30.81
NM_001935.1 DPP4 Dipeptidylpeptidase 4 (CD26, adenosine deaminase complexing protein 2) 1.60 1.68 3.58
NM_000505.2 F12 Coagulation factor XII (Hageman factor) 2.01 3.68 3.80

Oncogene
 UVB-induced gene expression profile of human epidermis
CD Enk et al
2606
Table 1 (continued )
GenBank Symbol Gene/protein Fold change from
nonirradiated
2h 24 h 72 h

NM_005532.1 IFI27 Interferon, alpha-inducible protein 27 2.87 3.63 12.10

NM_006435.1 IFITM2 Interferon-induced transmembrane protein 2 (1-8D) 1.77 2.62 2.40
BF338947 IFITM3 Interferon-induced transmembrane protein 3 (1-8U) 1.92 4.60 5.28
NM_001560.1 IL13RA1 Interleukin 13 receptor, alpha 1 1.12 3.58 1.07
NM_003856.1 IL1RL1 interleukin 1 receptor-like 1 0.32 8.19 6.01
NM_001562.1 IL18 Interleukin 18 (interferon-gamma-inducing factor) 1.16 0.35 1.55
NM_004633.1 IL1R2 Interleukin 1 receptor, type II 0.24 1.05 0.44
NM_003856.1 IL1RL1 Interleukin 1 receptor-like 1 0.32 8.19 6.01
NM_014432.1 IL20RA Interleukin 20 receptor, alpha 1.37 0.36 0.67
NM_006850.1 IL24 Interleukin 24 0.61 4.15 1.34
NM_000565.1 IL6R Interleukin 6 receptor 0.15 0.50 0.41
AF043337.1 IL8 Interleukin 8 2.15 1.12 0.27
NM_001557.1 IL8RB Interleukin 8 receptor, beta 0.96 2.83 1.78
NM_004030.1 IRF7 Interferon regulatory factor 7 2.66 1.78 1.09
NM_006674.1 MICA MHC class I polypeptide-related sequence A 3.78 1.79 0.97
NM_006187.1 OAS3 20 -50 -oligoadenylate synthetase 3, 100 kDa 1.34 2.00 2.84
NM_003999.1 OSMR Oncostatin M receptor 0.84 2.73 0.91
AI346350 PMSCL1 Polymyositis/scleroderma autoantigen 1, 75 kDa 1.44 0.97 5.78
NM_006404.1 PROCR Protein C receptor, endothelial (EPCR) 1.01 3.86 3.62
AB005043.1 SOCS1 Suppressor of cytokine signaling 1 3.02 3.04 1.89

Oxidatave stress related

NM_001752.1 CAT Catalase 1.15 0.24 0.89
NM_002085.1 GPX4 Glutathione peroxidase 4 (phospholipid hydroperoxidase) 1.76 0.51 1.43
NM_005109.1 OSR1 Oxidative-stress responsive 1 0.29 1.41 0.74
L19185 PRDX2 Peroxiredoxin 2 1.34 0.39 1.29
NM_000963.1 PTGS2 Prostaglandin-endoperoxide synthase 2 (prostaglandin G/H synthase and 0.76 0.18 0.71
cyclooxygenase)
NM_005410.1 SEPP1 Selenoprotein P, plasma, 1 1.29 0.26 0.79

Genes with at least twofold changes in expression in all volunteers at least at one time-point are listed.

patients and only weakly expressed in normal skin (Abts
et al., 1999). Hurpin overexpression in HaCaT cells
confers resistance to UV-induced apoptosis (Welss et al.,
2003). Our finding of overexpression of hurpin in UVB-
irradiated normal human epidermis suggests a role for
hurpin in the hyperproliferative state conferred to UV-
irradiated epidermis (Lee et al., 2002).
Extracellular matrix. The matrix metalloproteinases
(MMP) form a family of structurally and functionally
related zinc endopeptidases capable of degrading
structural proteins in connective tissue with implications
on tissue plasticity and migration of tumor cells
(Birkedal-Hansen et al., 1993; Kahari and Saarialho-
Kere, 1997). MMPs are tightly regulated by a special
class of tissue inhibitors of matrix protease, TIMP-1.
MMP induction by UV has been implicated in the
qualitative and quantitative alterations in the composi-
tion of the dermal extracellular matrix, a hallmark of
photoaged skin (Scharffetter et al., 1991; Fisher et al.,
1996). In our list, MMP-1 and MMP-3 expression was
dramatically induced after 24 h and remained upregu-
lated though to a lower degree at 72 h. These findings
are in accord with those of Fisher et al. (1996)
who demonstrated induction of MMP-1 and MMP-3
at 16–24 h in human skin biopsies following 2 MED
of UVB radiation, and those of Lahmann et al. (2001)
and Sudel et al. (2003) who found a peak of MMP-1
Oncogene
expression at 24 h and sustained upregulation at 48 h
after 2 MED of solar simulator radiation. In our study,
TIMP-1 was upregulated already at 2 h and remained
high for 72 h with a distinct peak at 24 h. It is intriguing
that a single low dose of UVB induces long-lasting
activity of MMP-1 with deleterious effects on dermal
extracellular matrix. However, the ratio of MMP-1 and
TIMP-1 is considered to be relevant for determining the
amount of active MMP-1 (Sudel et al., 2003). Calculat-
ing these ratios with our data (0.15, 3.38 and 0.75 for
the 2, 24 and 72 h time-points) indicates that MMP-1
activity following a single UVB dose is short lived and is
abolished at 72 h. SKALP/PI3, a recently described
proteinase inhibitor absent from normal epidermis but
found in hyperproliferative skin such as psoriasis and
healing wounds, and regulated by UVB and proinflam-
matory cytokines (Pfundt et al., 2000; Tanaka et al.,
2000b), was strongly upregulated at 24 h and remained
high at 72 h in our list.
Cell cycle. Cell-cycle progression and proliferation is a
tightly controlled process induced by a wide range of
exogenous events including UVB radiation that is
regulated by a complex series of endogenous processes
(Celis et al., 1987). We found that most of the cell-cycle
regulator genes were strongly upregulated at 72 h
but only partially at earlier time-points, indicating a
late burst of cell-cycle progression and proliferation.

Interestingly, cyclins A2, B2, D2 and G2 as well as
PCNA were upregulated, consistent with the lack of
GADD45 and WAF1/CIP1/p21 upregulation (Table 2).
Repair. UVB causes DNA lesions in epidermal cells,
predominantly cyclobutane pyrimidine dimers (CPD)
and pyrimidine–pyrimidone [6–4] photoproducts (Sage,
1993). The biological significance of these DNA lesions
depends upon the capacity of the cell to repair the
damage before it can be incorporated permanently into
the genome. Typically, DNA damage is rapidly repaired
at a relative high rate by inducing upregulation of
nucleotide excision repair (Young et al., 1996). In our
list, XPC, which is involved in DNA excision repair with
special affinity to DNA UV photoproducts (Kosmoski
et al., 2001), was clearly upregulated already 2 h after
UVB-induced gene expression profile of human epidermis
CD Enk et al
2607
exposure. The interaction of photoproducts with his-
tones results in modulation of the nucleosome structure
causing altered cromatin structure with increased
accessibility to repair enzymes (Brown et al., 1993; Liu
et al., 2000). Interestingly, an early and strong upregu-
lation of eight different histone genes was included in
our list.
Apoptosis. UV-induced apoptosis is a highly complex
process in which different molecular pathways such as
DNA damage, activation of the tumor suppressor gene
p53, triggering of cell death receptors, and ROS are
involved (Aragane et al., 1998; Wehrli et al., 2000). In
our list, genes belonging to the death receptor pathway
such as TRAIL-R2 and DAPK-1 were upregulated at
the 24 h time-point, and were shut down at 72 h,

Table 2 p53 target gene expression changes in intact human epidermis following in vivo UVB irradiation
Symbol Gene product 2h 24 h 72 h

Patient Patient Patient

I II III I II III I II III

APAF1 Apoptotic protease-activating factor 0.9 0.7 0.7 0.3 0.9 1.6 0.5 1.1 0.3
BAI1 Brain-specific angiogenesis inhibitor 1 1.2 0.5 0.6 0.5 0.6 0.6 0.3 0.2 0.4
BAX BCL2-associated X protein 1.3 1.0 1.5 1.9 2.0 2.0 1.9 0.2 0.4
BID BH3 interacting domain death agonist 2.5 1.1 0.6 6.5 1.9 0.7 4.0 1.1 0.8
BTG2 BTG family, member 2 1.4 1.0 1.4 1.0 0.5 0.9 0.8 0.6 0.8
CASP6 Caspase 6, apoptosis-related cysteine protease 1.7 3.2 1.6 0.5 1.1 0.6 1.0 1.9 0.9
CCNG1 Cyclin G1 0.6 1.0 0.8 1.6 2.6 1.5 1.4 2.0 1.6
CDKN1A Cyclin-dependent kinase inhibitor 1A (p21, Cip1) 0.5 0.6 0.8 1.0 0.9 1.1 0.7 0.7 0.7
GADD45A Growth arrest and DNA-damage-inducible, alpha 1.1 0.9 0.9 1.1 0.6 0.8 1.0 0.5 0.5
GADD45B growth arrest and DNA-damage-inducible, beta 3.2 1.4 1.1 0.6 0.6 0.7 0.8 0.7 0.9
GADD45G Growth arrest and DNA-damage-inducible, gamma 1.1 1.9 3.0 0.6 2.6 2.3 1.4 0.4 2.1
GGA1 Golgi associated, gamma adaptin ear-containing, 1.3 1.1 1.1 1.1 1.1 1.1 1.1 1.3 1.1
ARF-binding protein 1
GML GPI anchored molecule-like protein 10.6 6.5 0.4 4.9 12.1 2.0 8.0 3.0 0.9
HERC1 Hect (homologous to the E6-AP (UBE3A) carboxyl 1.2 0.9 1.1 0.6 0.4 0.4 0.6 0.7 0.9
terminus) domain and RCC1 (CHC1)-like domain (RLD) 1
IGFBP3 Insulin-like growth factor-binding protein 3 0.5 0.3 0.6 1.0 0.6 0.5 1.5 0.6 0.5
MDM2 Mdm2, transformed 3T3 cell double minute 2, p53-binding 0.5 1.4 0.8 1.1 1.4 1.9 1.4 2.3 0.7
protein (mouse)
MDM4 Mdm4, transformed 3T3 cell double minute 4, p53-binding 1.7 3.7 0.3 1.2 0.5 0.3 16.0 0.7 0.8
protein (mouse)
NOX1 NADPH oxidase 1 8.0 2.3 0.9 1.0 1.4 2.3 2.3 0.8 0.4
NOX3 NADPH oxidase 3 1.3 0.7 1.7 0.8 0.5 1.2 1.3 1.1 1.2
NOX4 NADPH oxidase 4 1.0 1.5 2.0 0.4 1.5 4.6 1.3 1.1 2.3
NOX5 NADPH oxidase, EF hand calcium-binding domain 5 1.9 5.7 0.5 0.9 9.2 1.1 1.2 6.5 0.8
P53AIP1 p53-regulated apoptosis-inducing protein 1 1.5 2.3 1.9 0.4 0.4 0.3 1.1 2.0 3.0
PERP p53-induced protein PIGPC1 1.1 1.2 1.1 0.8 0.9 1.1 0.9 1.1 1.0
PLAGL1 Pleiomorphic adenoma gene-like 1 1.1 1.2 1.0 0.7 0.4 0.8 0.6 1.1 1.1
REPRIMO Candidate mediator of the p53-dependent G2 arrest 0.9 0.9 1.1 0.8 0.4 0.8 0.1 0.3 0.7
TNFRSF6 Tumor necrosis factor receptor superfamily, member 6 1.1 0.7 0.7 1.4 1.2 1.5 1.1 0.9 1.0
TP53 Tumor protein p53 (Li–Fraumeni syndrome) 1.6 1.5 1.3 0.7 0.6 0.5 0.9 0.9 0.8
TP53BP1 Tumor protein p53 binding protein, 1 2.0 1.7 1.3 1.2 0.8 0.1 1.3 1.1 1.4
TP53BP2 Tumor protein p53-binding protein, 2 0.4 0.3 0.6 1.1 1.2 1.1 0.7 0.8 0.7
WIG1 p53 target zinc finger protein 1.1 0.8 0.8 1.1 1.5 1.0 0.9 0.7 0.9
ERCC3 Excision repair cross-complementing (xeroderma 1.5 1.6 0.9 1.1 1.1 0.9 0.9 1.1 0.9
(XPB) pigmentosum group B complementing)
XPA Excision repair cross-complementing (xeroderma 1.9 1.5 1.5 0.9 0.8 0.9 0.6 0.9 1.2
pigmentosum group A complementing)
XPC Excision repair cross-complementing (xeroderma 2.5 1.6 1.7 1.9 1.6 2.5 1.0 0.9 2.1
pigmentosum group C complementing)
ERCC5 Excision repair cross-complementing (xeroderma 1.2 2.1 1.7 0.5 0.7 0.5 0.7 1.4 1.4
(XPG) pigmentosum group G complementing)

Numbers show fold change of expression.

Oncogene
 UVB-induced gene expression profile of human epidermis
CD Enk et al
2608
underscoring the transient character of the apoptotic
response. Supporting the picture of an early burst of
proapoptotic activity, the antiapoptotic gene, Bcl-2, was
downregulated at 24 h, whereas the apoptosis inhibitor,
BIRC5 (Survivin), was strongly upregulated at 72 h.
Evidence of involvement of the ROS pathway is found
in the downregulation of catalase, glutathione perox-
idase, oxidative-stress responsive 1, peroxiredoxin 2 and
prostaglandin-endoperoxide synthase.
Recently, death-receptor-mediated, transcription-in-
dependent signaling modulation has emerged as an
important determinant of cell survival during develop-
ment and cellular homeostasis. These mechanisms
involve key post-translational modifications that affect
the activities of proteins at different levels in the
signaling pathways (Tran et al., 2004). Thus, UVB-
induced apoptosis, here evidenced by the presence of
sunburn cells (Figure 3a) and by a diffuse, epidermal
annexin V staining (Figure 3b), might be mediated by
both gene activation and by transcription-independent
mechanisms.
p53. The ‘guardian of the genome’, p53, is a tumor
suppressor protein that plays a key role in the protective
responses against exogenous injuries such as UV by
initiating a cascade of fine-tuned events through
transactivation of its target genes. Target genes of p53
include genes involved in cell-cycle arrest such the
cyclin-dependent kinase inhibitor WAF1/CIP1/p21 (el
Deiry et al., 1993), the proto-oncogene Mdm2 (Barak
et al., 1993), a negative regulator of p53, the growth
arrest and DNA damage-inducible GADD45 (Kastan
et al., 1992), Bax (Miyashita et al., 1994), Cyclin G
(Okamoto and Beach, 1994), PCNA (Shivakumar et al.,
1995), IGF-BP3 (Buckbinder et al., 1995), B99 (Utrera
et al., 1998), p53R2 (Tanaka et al., 2000a) and 14-3-3
Figure 3 Apoptosis staining of human epidermis 24 h after
exposure to UVB. (a) Sunburn cells stained by H&E; (b) Frozen
sections stained for annexin V.
Oncogene
sigma (Hermeking et al., 1997). p53 target genes present
on the Affymetrix Human Focus oligonucleotide arrays
are listed in Table 2. Surprisingly, applying the filter
used throughout this paper, that is, twofold change in all
three volunteers, only GML was upregulated and BAI1
downregulated at 24 and 72 h, respectively. Applying a
less strict filter of 1.5-fold changes, additional p53 target
genes were changed, including BAX (up at 24 h), CASP6
(up at 2 h), CCNG1 (up at 24 h), HERC1 (down at 24 h),
IGFBP3 (down at 2 h), TP53BP2 (down at 2 h), XPA
(up at 2 h) and XPC (up at 2 and 24 h). However,
even with these less restrictive cutoff criteria, key p53
targets such as WAF1/CIP1/p21, GADD45 and Mdm2
remained unaffected following UVB exposure.
To verify the validity of these profiling data, we first
ascertained the accumulation of p53 protein in human
epidermis following exposure to 4 MED of UVB in vivo
(Figure 4). Our results confirmed that p53 has a very low
basal expression in human skin and accumulates upon
UV irradiation (Hall et al., 1993; Liu et al., 1994;
Courtois et al., 1997). We next confirmed the microarray
data by semiquantitative RT–PCR using primers for key
p53 target. As can be seen in Figure 5, neither WAF1/
CIP1/p21, GADD45, Bax or Mdm2 gene expression was
upregulated in spite of p53 protein accumulation 24 h
following UVB exposure, whereas SCCA1, included as a
positive control, was clearly increased. Thus, our data
show that in spite of p53 protein accumulation in
epidermal cells following a single in vivo UVB exposure,
p53 did not cause transactivation of its known target
genes.
Our current understanding of the regulation of p53 in
human skin following UV damage is mostly based on
keratinocyte cultures. To which degree such models
reflect genuine physiological processes is largely un-
known. Thus, the p53 repertoire and signaling pathways
of the cell is determined by several biological parameters
such as cell type, oncogenic composition, the nature and
the intensity of the extracellular stimulus, and the level
of p53 (Lassus et al., 1996; Li and Ho, 1998; Sionov and
Haupt, 1999; Zhao et al., 2000). Furthermore, the
resistance of keratinocytes to apoptosis depends on the
differentiation status and the location of the cells within
the epidermal layers (Qin et al., 2002). Many of these
parameters are not well controlled in the available cell
culture models. Our unexpected finding that only few of
the p53 target genes were activated following a single
in vivo UVB exposure might reflect crucial differences in
Figure 4 p53 staining of human epidermis 24 h after exposure to
UVB. (a) Unexposed; (b) UVB-exposed skin.

Figure 5 Reverse transcriptase–polymerase chain reaction (RT–
PCR) verification of selected p53 target genes 24 h after in vivo
UVB irradiation of intact human epidermis. Nonirradiated
epidermis served as a control. RNA was obtained from four
volunteers different from those tested by microarray analysis.
G3PDH was used as housekeeping gene. SCCA1 served as a
positive control. The PCR product was stained with ethydium
bromide (a), and signal intensities were normalized for G3PDH by
densitometry (b).
terms of cellular composition, differentiation, intra-
cellular p53 levels and UV sensitivity existing between
cell cultures and in vivo models like the one we have
used. It is of interest to note that in a recent paper,
Marionnet et al. (2003) examined the expression profile
of human epidermis following in vivo low-dose solar
simulator UV exposure and found GADD45 and Bax to
be moderately upregulated. Although a different light
source was used in this study and verification of the
upregulation of the two genes was not performed, the
data appear to contradict our findings. However,
whereas the epidermal sheets in our study originated
from chronically sun-exposed forearm skin, the derma-
tome specimens in the Marionnet et al. (2003) paper
were obtained from buttock skin, not previously
exposed to sunlight. This might partially explain the
apparent discrepancy in GADD45 and Bax regulation,
since pre-exposure of human keratinocytes to a low dose
of UVB significantly alters the p53 signaling pathways
and apoptotic response compared to cells that were not
pre-exposed (Decraene et al., 2004).
Possible mechanisms resulting in lack of activity of
p53 transcription targets despite increase in p53 protein
levels include inhibitory modification of the p53 proteins
such as dephosphorylation, mono-ubiquitination, ned-
dylation (Shieh et al., 1997; Siliciano et al., 1997; Wang
et al., 2004; Xirodimas et al., 2004) or the expression of
inhibitory DNp63 (Yang et al., 1998). Elucidation of the
mechanisms involved in the silencing of the p53
signaling pathways following in vivo UVB irradiation
of human skin might provide clues for understanding
the processes allowing the proliferation of transformed
cells in human skin.
UVB-induced gene expression profile of human epidermis
CD Enk et al
2609
Figure 6 Venn diagram representing up- and downregulated genes
in in vivo UVB-irradiated intact human epidermis and in vitro
exposed NHEK. The diagram shows the number of genes with the
indicated expression patterns.
Comparison of UVB-regulated genes in intact epidermis
and cultured keratinocytes
Intact human epidermis is a well-ordered multilayer
organ in which the majority of the cells, that is, cells of
the suprabasal layers, are keratinocytes undergoing
various stages of differentiation. In contrast, subcon-
fluent cultured NHEK is a monolayer system consisting
of exponentially proliferating keratinocytes. In order to
investigate whether these differences in maturity influ-
ence the molecular repertoire of epidermal cells in
response to UVB, we compared the in vivo expression
profiles of epidermal sheets from the present study to
that of NHEK exposed to UVB in vitro. The profiling
results of the NHEK were previously published (Dazard
et al., 2003) and the full list of UVB-regulated genes can
be found at http://www.weizmann.ac.il/home/ligivol/
UVB_project /table2.xls. In total, we identified 1931
genes that differed in their expression profiles among in
vivo and in vitro irradiated epidermal cells, for which the
overlapping and nonoverlapping genes among the
groups are illustrated in Figure 6. When directly
comparing the UVB response of individual genes
(Table 3), we found multiple differences in genes
governing crucial cellular functions such as p53, cell-
cycle arrest, DNA repair and apoptosis. The overall
picture exhibited by these differences in expression fits a
pattern of exponentially proliferating cells reacting to
UVB damage by initiating cellular processes associated
with cell-cycle arrest and apoptosis, whereas differen-
tiating cells respond by DNA repair (http://www.
hadassah.org.il/Departments/photobiology; for the full
list) (Qin et al., 2002; Chaturvedi et al., 2004). Thus, key
Oncogene
 UVB-induced gene expression profile of human epidermis
CD Enk et al
2610
Table 3 Comparison of UVB-induced expression changes of in vivo exposed intact epidermis and in vitro irradiated NHEK
GenBank Symbol Gene/protein Epidermis NHEK

Cell cycle
NM_003914.1 CCNA1 Cyclin A1 ¼ m
NM_001759.1 CCND2 Cyclin D2 m k
NM_004701.2 CCNB2 Cyclin B2 m k
AW134535 CCNG2 Cyclin G2 m k
NM_000946 PRIM1 DNA primase, G1 to S cell-cycle reactome m ¼
NM_001269 CHC1 Chromosome condensation 1, (G1/S transition of m ¼
mitotic cell cycle DNA packaging)
N33167 CDKN1C Cyclin-dependent kinase inhibitor 1C (p57, Kip2) k m
NM_005197 CHES1 Checkpoint suppressor 1 k ¼
M68520 CDK2 Cyclin-dependent kinase 2 k ¼
U49844 ATR Ataxia telangiectasia and Rad3 related k k

Repair
NM_000380.1 XPA Xeroderma pigmentosum, complementation group A ¼ k
D21089.1 XPC Xeroderma pigmentosum, complementation group C m ¼
NM_000107 DDB2 Damage-specific DNA-binding protein 2, nucleotide-excision repair m ¼
NM_002691 POLD1 Polymerase delta 1, DNA replication DNA repair response to UV m ¼
NM_006297 XRCC1 X-ray repair complementing defective repair m ¼
AL080203 POLE Polymerase (DNA directed), epsilon; DNA replication and DNA repair m ¼
AJ243797 TREX1 Three prime repair exonuclease 1 m
NM_002690 POLB Polymerase (DNA directed), beta (DNA replication and DNA repair) ¼ m
NM_001983 ERCC1 Excision repair cross-complementing rodent repair deficiency, ¼ m
complementation group 1
NM_002412 MGMT O-6-methylguanine-DNA methyltransferase; DNA ligation, DNA repair ¼ m

Apoptosis
M14745 Bcl-2 ¼ m
NM_004052 BNIP3 BCL2/adenovirus E1B-interacting protein 3 (antiapoptosis) m ¼
AL117381 BCL2L1 BCL2-like 1 (antiapoptosis) m ¼
M59465 TNFAIP3 TNFa- induced protein 3 k m
NM_001066 TNFRSF1B Tumor necrosis factor receptor superfamily, member 1B ¼ m
BF439983 CASP8 Caspase 8, apoptosis-related cysteine protease m k
U13699 CASP1 Caspase 1, apoptosis-related cysteine protease ¼ m
AK024029 MOAP1 Modulator of apoptosis 1 ¼ m
U83981 PPP1R15A Protein phosphatase 1, apoptosis response to DNA damage stimulus ¼ m
BF686824 DAPK3 Death-associated protein kinase 3 k ¼
NM_001166 BIRC2 Baculoviral IAP repeat-containing 2 (antiapoptosis) ¼ k

p53 target genes

NM_015675.1 GADD45B Growth arrest and DNA-damage-inducible, beta ¼ m
NM_000389.1 CDKN1A Cyclin-dependent kinase inhibitor 1A (p21, Cip1) ¼ m
NM_001924.2 GADD45A Growth arrest and DNA-damage-inducible, alpha ¼ m
M31159.1 IGFBP3 Insulin-like growth factor-binding protein 3 k ¼
NM_005426.1 TP53BP2 Tumor protein p53-binding protein, 2 k ¼
NM_003922.1 HERC1 Hect (homologous to the E6-AP (UBE3A) carboxyl terminus) k ¼
domain and RCC1 (CHC1)-like domain (RLD) 1
NM_002656.1 PLAGL1 Pleiomorphic adenoma gene-like 1 k ¼
NM_005657.1 TP53BP1 Tumor protein p53-binding protein, 1 ¼ k

Growth factors and oncogenes

NM_002467.1 MYC v-Myc myelocytomatosis viral oncogene homolog (avian) k k
NM_001953.2 ECGF1 Endothelial cell growth factor 1 (platelet-derived) m m

Epidermal differentiation complex, calcium-binding proteins

NM_002965.2 S100A9 S100 calcium-binding protein A9 (calgranulin B) m m

Histones
NM_017445.1 H2BFT H2B histone family, member T m m
NM_003528.1 H2BFQ H2B histone family, member Q m m
AA451996 H2AFO H2A histone family, member O m m
NM_003522.1 H2BFG H2B histone family, member G m m
BC000893.1 H2BFT H2B histone family, member T m m
NM_021052.1 H2AFA H2A histone family, member A m m
BC002649.1 H1F2 H1 histone family, member 2 m m
NM_006026.1 H1FX H1 histone family, member X m m

Serpins (Serine proteases inhibitors)

AF119873.1 SERPINA1 Serine (or cysteine) proteinase inhibitor, clade A (alpha-1 antiproteinase, m ¼
antitrypsin), member 1

Oncogene
 UVB-induced gene expression profile of human epidermis
CD Enk et al
2611
Table 3 (continued )
GenBank Symbol Gene/protein Epidermis NHEK

AI554300 SERPINB1 Serine (or cysteine) proteinase inhibitor, clade B (ovalbumin), member 1 m ¼
BC005224.1 SERPINB3 SCCA1 Serine (or cysteine) proteinase inhibitor, clade B (ovalbumin), member 3 m ¼

Cytokines and chemokines

NM_005532.1 IFI27 Interferon, alpha-inducible protein 27 m m
NM_003856.1 IL1RL1 Interleukin 1 receptor-like 1 m m
NM_004591.1 CCL20 Chemokine (C-C motif) ligand 20 m m
NM_001560.1 IL13RA1 Interleukin 13 receptor, alpha 1 m m
AF043337.1 IL8 Interleukin 8 m m
NM_001562.1 IL18 Interleukin 18 k k

Extracellular matrix affecting factors

NM_002421.2 MMP1 Matrix metalloproteinase 1 (interstitial collagenase) m m
NM_001814.1 CTSC Cathepsin C m m
NM_003254.1 TIMP1 Tissue inhibitor of metalloproteinase 1 m m

Oxidative stress related

NM_002084.2 GPX3 Glutathione peroxidase 3 (plasma) m m
NM_000581.1 GPX1 Glutathione peroxidase 1 m m
NM_002083.1 GPX2 Glutathione peroxidase 2 (gastrointestinal) m m
NM_001752.1 CAT Catalase k k

Heat shock
X51757 HSPA6 Heat shock 70 kDa protein 6 (HSP70B0 ) k k

m, Upregulated; k, downregulated; ¼ , no change.

cell-cycle progression markers such as cyclins B2, D2
and G2 were all downregulated and p57/KIP2 upregu-
lated in NHEK but not in intact epidermis, where
an opposite picture indicating cell-cycle progression
was found. Also among apoptosis-related genes, a
functional dichotomy was evident, since the proapop-
titic genes, TNFAIP3, TNFRSF1B, CASP1, MOAP1
and PPP1R15A, were upregulated only in NHEK,
whereas antiapoptotic genes such as BNIP3 and
BCL2L1 were upregulated only in intact epidermis.
Likewise, in the DNA repair group, XPC, DDB2,
POLD1, XRCC1, POLE and TREX1 were only
upregulated in intact epidermis. And in the p53 group
where key target genes such as GADD45A, GADD45B
and WAF1/CIP1/p21 were unaffected in intact epider-
mis, the same genes were upregulated in NHEK.
The fact that the comparison between intact epidermis
and cultured keratinocytes was not performed in parallel
raises a number of issues: (1) Although both microarray
profilings were performed in the same laboratory,
different Affymetrix microarrays were used in the two
experiments. To allow a direct comparison between the
two experiments, a series of compensatory bioinformatic
procedures were performed as detailed in Methods to
compensate for potential pitfalls due to differences in
construction of the microarrays,; (2) UV light impinging
on intact human epidermis is absorbed or scattered by
pigment cells and by chromophores in the stratum
corneum, and only 20–30% of the incoming light
reaches the keratinocytes of the deeper layers of the
epidermis. None of these physiological structures
diminish the amount of UV light reaching the mono-
layer keratinocytes in culture. To compensate for these
differences in UV accessibility, the in vivo UVB exposure
in the present study was performed with 150–250 mJ/cm2,
whereas the keratinocyte cultures were exposed to
40 mJ/cm2, using light sources with identical UVB spectra.
Keeping in mind that approximately 100 mJ/cm2 of
UVB can be encountered upon exposure to 60–90 min
of midday Summer sun, it is worth noticing that the
UVB exposure delivered to both intact epidermis and
to the keratinocyte cultures is environmentally relevant;
(3) The hypothesis-generating nature of microarray
profiling appears to justify our approach of comparing
two bioinformatic analyses despite the fact that they
were not performed in parallel. However, future
parallel, hypothesis-driven studies are required to
compare selected candidate genes for their function
and relevance under in vivo and in vitro conditions.
Methods
Array processing
All experiments were performed using Affymetrix Human
Focus oligonucleotide arrays, as described (http://www.
affymetrix.com/support/technical/datasheets/human datasheet.
pdf). Total RNA from each sample was used to prepare
biotinylated target RNA, with minor modifications from
the manufacturer’s recommendations (http://www.affymetrix.
com/support/technical/manual/expression_manual.affx).
Briefly, 5 mg of mRNA was used to generate first-strand
cDNA by using a T7-linked oligo(dT) primer. After second-
strand synthesis, in vitro transcription was performed with
biotinylated UTP and CTP (Enzo Diagnostics), resulting in
approximately 300-fold amplification of RNA. The target
cDNA generated from each sample was processed as per the
manufacturer’s recommendation using an Affymetrix Gene-
Chip Instrument System (http://www.affymetrix.com/support/
Oncogene
 UVB-induced gene expression profile of human epidermis
CD Enk et al
2612
technical/manual/expression_manual.affx). Briefly, spike con-
trols were added to 15 mg fragmented cRNA before overnight
hybridization. Arrays were then washed and stained with
streptavidin–phycoerythrin, before being scanned on an
Affymetrix GeneChip scanner. Additionally, quality and
amount of starting RNA was confirmed using an agarose
gel. After scanning, array images were assessed by eye to
confirm scanner alignment and the absence of significant
bubbles or scratches on the chip surface. 30 /50 ratios for
GAPDH and beta-actin were confirmed to be within
acceptable limits (1.01–1.59 and 1.04–1.82), and BioB spike
controls were found to be present on all chips, with BioC,
BioD and CreX also present in increasing intensity. When
scaled to a target intensity of 150 (using Affymetrix MAS 5.0
array analysis software), scaling factors for all arrays were
within acceptable limits (1.07–2.08), as were background,
Q values and mean intensities. Details of quality control
measures can be found at http://www.ncbi.nlm.nih.gov/geo/
and at http://eng.sheba.co.il/genomics.
Data analysis
The 8757 probe sets contained in the Affymetrix Human Focus
oligonucleotide array were filtered using Mas 5 algorithm
(http://www.hadassah.org.il/Departments/photobiology). A
list of 5380 ‘valid’ probe sets, representing probe sets with
signals higher than 20 and detected as present (P) in at least
one sample, was obtained. Treated and control samples were
compared in each time-point. The comparison generated a list
of 619 ‘active genes’ representing probe sets changed by at least
twofold (as calculated from the MAS 5 Log Ratio values)
(LRX1 or LRp1) and detected as ‘Increased’ or as as
‘Decrease’ (I or D, P-value 0.0025) or ‘Marginal Increased’ or
as or Marginal Decrease (MI or MD, P-value 0.003 ) in all
treated sample as compared to all the control samples in at
least one time-point. This list excluded upregulated genes in all
treated samples with signals lower than 20 or detected as
absent, and downregulated gene with base line signals lower
than 20 and detected as absent in the control samples.
Hierarchical clustering was performed using Spotfire Deci-
sionSite for Functional Genomics (Somerville, MA, USA).
Genes were classified into functional groups using the GO
annotation tool (Dennis Jr et al., 2003). Over-representation
calculations were carried out using Ease (Hosack et al., 2003).
Functional classifications with an ‘Ease score’ lower than 0.05
were marked as over-represented.
Data comparison of in vivo and in vitro UVB-regulated genes
The list of UVB-regulated genes following in vivo exposure
of intact epidermis obtained from the present study was
compared to a previously published list of differentially
regulated genes in NHEK following in vitro UVB irradiation
(Dazard et al., 2003).
In order to compare the two experiments, initial compensa-
tion had to be made for the use of different arrays (Affymetrix
Human Focus for the in vivo and U95 Affymetrix arrays for
the in vitro study). To select the modulated genes in the in vivo
study, we applied a filter of at least twofold change at 2 or 24 h
or at both time-points in at least two out of the three
volunteers. In total, 381 and 642 genes passed this filter for the
up- and downregulated genes, respectively. In the in vitro
study, the original list of 273 genes was first intersected with
the list of probe sets present on the Focus array, leaving 247
genes that were then compared to the list of 381 upregulated
genes from the in vivo study. Similarly, the list of 363
downregulated genes in the in vitro study was first intersected
with the list of genes present on Focus array, leaving 202 genes
Oncogene
that were compared to the list of 642 downregulated genes
from the in vivo study.
Volunteers
Three healthy volunteers were recruited after informed consent
and prior approval from the Ethical Committee on Experi-
ments on Humans (Helsinki Committee). Inclusion criteria
were skin type II–III, age 18-30 years, non-smokers, no regular
or concomitant medication.
UVB irradiation
Areas of 9 cm2 on previously unexposed skin of the inner
forearms were irradiated with UVB from a bank of four FS40
fluorescent lamps (Philips) that emit wavelengths between
280 and 320 nm with a peak at 313 nm. Light intensity, deter-
mined using a Waldmann (Waldmann GBH, Schwenningen,
Germany) UV radiometer, was 0.75 mW/cm2 at a distance of
20 cm from the light source. Single doses of 150–250 mJ/cm2,
equivalent to 4 MED, were used.
p53 immunohistochemistry
Formalin-fixed, paraffin-embedded tissue sections of punch-
biopsies taken from nonirradiated and 24 h after 4 MED
UVB-irradiated human skin, and a positive control section
(carcinoma of the colon) were cut at 4–5 mm, deparaffinized in
a clearing solution and rehydrated in alcohol. Endogenous
peroxidase was blocked with 3% hydrogen peroxidase in
distilled water for 5 min. Antigen retrieval was performed by
microwave heating at 921C for 20 min in 1 mmol EDTA,
pH 8.0. Slides were then immunostained in a Ventana Nexes
automated staining system with an AEC staining kit. Primary
antibodies were applied (monoclonal mouse antihuman p53
protein, clone D0-7; DAKO, Glostrup, Denmark) in dilution
1/50. Incubation time was 32 min at 371C.
Staining for apoptosis
Punch biopsies obtained 24 h after UVB irradiation were
stained for apoptosis by H&E (sunburn cells) and annexin V.
Samples were immediately frozen in OCT in liquid nitrogen
and were stained with AnV-FITC (Roche Diagnostics,
Mannheim, Germany).
Epidermal sheets and RNA extraction
Suction blisters (2.5 cm in diameter) were induced 2, 24, and
72 h after irradiation, and RNA was extracted from the blister
roofs (Enk and Katz, 1994). Blisters from adjacent nonirra-
diated skin served as control. Total RNA was extracted from
the blister roofs using the SV Total RNA Isolation System
(Promega, Madison, WI, USA).
Semiquantitiative RT–PCR
RT–PCR was performed with the Access RT–PCR system
(Promega). Total RNA (1 mg) was used as templates in
reaction mixtures containing 2 ml 25 mM MgSO4, 1 ml 10 mM
dNTP mix, 20 pmols of each of the primers for all genes, 5 U
AMV reverse transcriptase, 5 U Tfl DNA polymerase and
reaction buffer in a total volume of 50 ml. RT was performed at
481C for 45 min and 951C for 2 min, followed by PCR. Cycling
conditions and mRNA concentrations for each gene were
determined by prior titration experiments. The following
cycling times and temperatures were used: 31cycles at 941 for
60 s, 601 for 60 s and 721 for 90 s (p21); 28 cycles at 941 for 60 s,
601 for 60 s and 721 for 90 s (GAAD45, Bax and mdm2);
28 cycles at 941 for 60 s, 551 for 60 s and 721 for 90 s (G3PDH).
In total, 10 ml of the PCR products were run on 2% agarose
ethidium bromide gel electrophoresis and visualized under
UV. Semiquantitative estimations were obtained by

correcting for the G3PDH housekeeping gene. Band intensities
were determined by densitometry measurements using Multi-
Analystt/PC Version1.1 (Bio-Rad, CA, USA). Reactions were
regularly performed in the absence of reverse transcriptase or
template to control for contamination of genomic DNA or
carryover of cDNA.
Primers sequences were:
p21 sense : 50 -AGGATCCATGTCAGAACCGGCTGG-30 ;
antisense: 50 -CAGGATCCTGTGGGCGGATTAGGGCT-30
GAAD45 sense: 50 -AGA ACG ACA TCA ACA TCC TGC-30 ;
antisense: 50 -AAT GTG GAT TCG TCA CCA GA-30
Bax sense: 50 -GGGGACGAACTGGACAGTAA-30 ;
antisense: 50 -CAGTTGAAGTTGCCGTCAGA-30
Mdm2 sense: 50 -CAGCTTCGGAACAAGAGACC-30 ;
antisense: 50 -GTCCGATGATTCCTGCTGAT-30 ;
SCCA1 sense: 50 -CCTGAAGGTAATATTGGCAGC-30 ;
antisense: 50 -GGGATGAGAATCTGCCATAG-30 ;
G3PDH sense: 50 -TGAAGGTCGGAGTCAACGGATTT
GGT-30 ;
antisense: 50 -CATGTGGGCCATGAGGTCCACCAC-30
Conclusions
Using global microarray profiling to investigate the molecular
response of intact human epidermis to a single, physiological
dose of in vivo UVB irradiation, we found that genes
controlling numerous cellular functions were regulated by
UVB. Although the majority of the UVB-modulated genes
were found at the 24 h time-point, differentially expressed
genes could be found already after 2 h. The transient character
of the UVB response was demonstrated by the finding that half
of the up- and downregulated genes had returned to back-
ground levels within 72 h following exposure.
Based on the assumption that keratinocyte cultures closely
mirror changes taking place in intact epidermis, our present
understanding of the influence of UVB on human skin is
mainly based on in vitro models. Our study is the first to
directly compare the molecular changes in intact human
epidermal cells with those of in vitro exposed NHEK. We
found that close to 2000 genes differed in their expression
References
Abts HF, Welss T, Mirmohammadsadegh A, Kohrer K,
Michel G, Ruzicka T. (1999). J Mol Biol 293: 29–39.
Alowami S, Qing G, Emberley E, Snell L, Watson PH. (2003).
BMC Dermatol 3: 1.
Aragane Y, Kulms D, Metze D, Wilkes G, Poppelmann B,
Luger TA et al. (1998). J Cell Biol 140: 171–182.
Baadsgaard O. (1991). Arch Dermatol 127: 99–109.
Barak Y, Juven T, Haffner R, Oren M. (1993). EMBO J 12:
461–468.
Becker B, Vogt T, Landthaler M, Stolz W. (2001). J Invest
Dermatol 116: 983–988.
Birkedal-Hansen H, Moore WG, Bodden MK, Windsor LJ,
Birkedal-Hansen B, DeCarlo A et al. (1993). Crit Rev Oral
Biol Med 4: 197–250.
Boni R, Heizmann CW, Doguoglu A, Ilg EC, Schafer BW,
Dummer R et al. (1997). J Cutan Pathol 24: 76–80.
Brash DE, Ziegler A, Jonason AS, Simon JA, Kunala S, Leffell
DJ. (1996). J Investig Dermatol Symp Proc 1: 136–142.
Broome AM, Ryan D, Eckert RL. (2003). J Histochem
Cytochem 51: 675–685.
Brown DW, Libertini LJ, Suquet C, Small EW, Smerdon MJ.
(1993). Biochemistry 32: 10527–10531.
Buckbinder L, Talbott R, Velasco-Miguel S, Takenaka I, Faha
B, Seizinger BR et al. (1995). Nature 377: 646–649.
UVB-induced gene expression profile of human epidermis
CD Enk et al
2613
profiles between the two groups: The expression profile of in
vivo exposed intact epidermis was geared towards DNA repair,
whereas in vitro irradiated NHEK responded by activating
genes associated with cell-cycle arrest and apoptosis. These
differences in expression profiles might reflect differences
between mature differentiating keratinocytes in the suprabasal
epidermal layers versus exponentially proliferating NHEK in
cell culture. Of special interest is the p53 profile: In contrast to
the paradigm that key target genes such as GADD45, WAF1/
CIP1/p21, Bax and Mdm2 are upregulated in a properly
functioning p53 tumor suppressor cascade, we found that in
spite of p53 protein accumulation in epidermal cells, several
central downstream genes were unaffected by a single in vivo
UVB exposure. Assuming that under physiological conditions,
a single UVB exposure has little carcinogenic potential, the
lack of activation of genes involved in cell-cycle arrest and
p53 target genes suggests that signaling pathways that can
cause DNA repair without prior cell-cycle arrest, or that
transcription-independent mechanisms are activated in intact
epidermis. It should be noted that we have limited our in depth
comparison of UVB-induced expression profiles to the p53
tumor suppressor gene, since it plays a central role in the
regulation of the UV response. It is indeed possible that other
conclusions might have been reached had we broadened our
investigation also to include additional molecular pathways
such as the NF-kB family of transcription factors, known as
‘the central mediator of the human immune response’. Future
studies should be carried out to elucidate such possibilities.
Our findings show that extreme care should be taken when
extrapolating from findings based on keratinocyte cultures to
changes in intact epidermis.
Acknowledgements
This study was supported by grants from The Israel Cancer
Association through the donation from The Brown Founda-
tion, Florida (grant no. 20030004-B) and The Hadassah
Women Zionist Organization of America (HWZOA) Com-
pensatory Research Fund.
Cataltepe S, Gornstein ER, Schick C, Kamachi Y, Chatson K,
Fries J et al. (2000). J Histochem Cytochem 48: 113–122.
Celis JE, Madsen P, Celis A, Nielsen HV, Gesser B. (1987).
FEBS Lett 220: 1–7.
Chaturvedi V, Bacon P, Bodner B, Nickoloff BJ. (2004).
J Invest Dermatol 123: 1200–1203.
Courtois SJ, Woodworth CD, Degreef H, Garmyn M. (1997).
Exp Cell Res 233: 135–144.
Crombach G, Scharl A, Vierbuchen M, Wurz H, Bolte A.
(1989). Cancer 63: 1337–1342.
Dazard JE, Gal H, Amariglio N, Rechavi G, Domany E,
Givol D. (2003). Oncogene 22: 2993–3006.
Decraene D, Gan D, Mammone T, Maes D, Declereq L,
Germyn M. (2004). J Investig Dermatol 122: A140.
Dennis Jr G, Sherman BT, Hosack DA, Yang J, Gao W, Lane
HC et al. (2003). Genome Biol 4: 3.
Duk JM, Voorst Vader PC, ten Hoor KA, Hollema H,
Doeglas HM, de Bruijn HW. (1989). Cancer 64: 1652–1656.
el Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R,
Trent JM et al. (1993). Cell 75: 817–825.
Enk CD, Katz SI. (1994). Arch Dermatol Res 287: 72–77.
Enk CD, Shahar I, Amariglio N, Rechavi G, Kaminski N,
Hochberg M. (2004). Photodermatol Photoimmunol Photo-
med 20: 129–137.
Oncogene
 UVB-induced gene expression profile of human epidermis
CD Enk et al
2614
Fisher GJ, Datta SC, Talwar HS, Wang ZQ, Varani J, Kang S
et al. (1996). Nature 379: 335–339.
Gibbs S, Boelsma E, Kempenaar J, Ponec M. (1998). Cell
Tissue Res 292: 107–114.
Gniadecki R, Hansen M, Wulf HC. (1997). J Invest Dermatol
109: 163–169.
Grimbaldeston MA, Geczy CL, Tedla N, Finlay-Jones JJ,
Hart PH. (2003). J Invest Dermatol 121: 1168–1174.
Hahn WC, Weinberg RA. (2002). Nat Rev Cancer 2: 331–341.
Hall PA, McKee PH, Menage HD, Dover R, Lane DP. (1993).
Oncogene 8: 203–207.
Hermeking H, Lengauer C, Polyak K, He TC, Zhang L,
Thiagalingam S et al. (1997). Mol Cell 1: 3–11.
Horiuchi Y, Tsukahara T, Otoyama K. (1994). J Dermatol 21:
67–72.
Hosack DA, Dennis Jr G, Sherman BT, Lane HC, Lempicki
RA. (2003). Genome Biol 4: R70.
Joos L, Eryuksel E, Brutsche MH. (2003). Swiss Med Wkly
133: 31–38.
Kahari VM, Saarialho-Kere U. (1997). Exp Dermatol 6:
199–213.
Kannan K, Amariglio N, Rechavi G, Jakob-Hirsch J, Kela I,
Kaminski N et al. (2001). Oncogene 20: 2225–2234.
Kastan MB, Zhan Q, el Deiry WS, Carrier F, Jacks T, Walsh
WV et al. (1992). Cell 71: 587–597.
Kato H. (1996). Anticancer Res 16: 2149–2153.
Kosmoski JV, Ackerman EJ, Smerdon MJ. (2001). Proc Natl
Acad Sci USA 98: 10113–10118.
Lahmann C, Young AR, Wittern KP, Bergemann J. (2001).
Photochem Photobiol 73: 657–663.
Lassus P, Ferlin M, Piette J, Hibner U. (1996). EMBO J 15:
4566–4573.
Lee JH, An HT, Chung JH, Kim KH, Eun HC, Cho KH.
(2002). Photodermatol Photoimmunol Photomed 18: 253–261.
Lee KM, Lee JG, Seo EY, Lee WH, Nam YH, Yang JM et al.
(2005). Br J Dermatol 152: 52–59.
Li D, Turi TG, Schuck A, Freedberg IM, Khitrov G,
Blumenberg M. (2001). FASEB J 15: 2533–2535.
Li G, Ho VC. (1998). Br J Dermatol 139: 3–10.
Liu M, Dhanwada KR, Birt DF, Hecht S, Pelling JC. (1994).
Carcinogenesis 15: 1089–1092.
Liu X, Mann DB, Suquet C, Springer DL, Smerdon MJ.
(2000). Biochemistry 39: 557–566.
Maelandsmo GM, Florenes VA, Mellingsaeter T, Hovig E,
Kerbel RS, Fodstad O. (1997). Int J Cancer 74: 464–469.
Marionnet C, Bernerd F, Dumas A, Verrecchia F, Mollier K,
Compan D et al. (2003). J Invest Dermatol 121: 1447–1458.
Miyashita T, Krajewski S, Krajewska M, Wang HG, Lin HK,
Liebermann DA et al. (1994). Oncogene 9: 1799–1805.
Murakami T, Fujimoto M, Ohtsuki M, Nakagawa H. (2001).
J Dermatol Sci 27: 121–129.
Okamoto K, Beach D. (1994). EMBO J 13: 4816–4822.
Park WY, Hwang CI, Im CN, Kang MJ, Woo JH, Kim JH
et al. (2002). Oncogene 21: 8521–8528.
Supplementary Information accompanies the paper on Oncogene website (http://www.nature.com/onc)
Oncogene
Pfundt R, Wingens M, Bergers M, Zweers M, Frenken M,
Schalkwijk J. (2000). Arch Dermatol Res 292: 180–187.
Pisarchik A, Wortsman J, Slominski A. (2004). Gene 341:
199–207.
Qin JZ, Chaturvedi V, Denning MF, Bacon P, Panella J,
Choubey D et al. (2002). Oncogene 21: 2991–3002.
Sage E. (1993). Photochem Photobiol 57: 163–174.
Scharffetter K, Wlaschek M, Hogg A, Bolsen K, Schothorst A,
Goerz G et al. (1991). Arch Dermatol Res 283: 506–511.
Sellheyer K, Belbin TJ. (2004). J Am Acad Dermatol 51:
681–692.
Sesto A, Navarro M, Burslem F, Jorcano JL. (2002). Proc Natl
Acad Sci USA 99: 2965–2970.
Shieh SY, Ikeda M, Taya Y, Prives C. (1997). Cell 91: 325–334.
Shivakumar CV, Brown DR, Deb S, Deb SP. (1995). Mol Cell
Biol 15: 6785–6793.
Siliciano JD, Canman CE, Taya Y, Sakaguchi K, Appella E,
Kastan MB. (1997). Genes Dev 11: 3471–3481.
Sionov RV, Haupt Y. (1999). Oncogene 18: 6145–6157.
Sudel KM, Venzke K, Knussmann-Hartig E, Moll I,
Stab F, Wenck H et al. (2003). Photochem Photobiol 78:
355–360.
Suminami Y, Nagashima S, Vujanovic NL, Hirabayashi K,
Kato H, Whiteside TL. (2000). Br J Cancer 82: 981–989.
Takao J, Ariizumi K, Dougherty II, Cruz Jr PD. (2002).
Photodermatol Photoimmunol Photomed 18: 5–13.
Takeda A, Higuchi D, Takahashi T, Ogo M, Baciu P,
Goetinck PF et al. (2002). J Invest Dermatol 118: 147–154.
Tanaka H, Arakawa H, Yamaguchi T, Shiraishi K, Fukuda S,
Matsui K et al. (2000a). Nature 404: 42–49.
Tanaka N, Fujioka A, Tajima S, Ishibashi A, Hirose S.
(2000b). Br J Dermatol 143: 728–732.
Thorey IS, Roth J, Regenbogen J, Halle JP, Bittner M, Vogl T
et al. (2001). J Biol Chem 276: 35818–35825.
Tran SE, Meinander A, Eriksson JE. (2004). Trends Biochem
Sci 29: 601–608.
Utrera R, Collavin L, Lazarevic D, Delia D, Schneider C.
(1998). EMBO J 17: 5015–5025.
Wang X, Taplick J, Geva N, Oren M. (2004). FEBS Lett 561:
195–201.
Wehrli P, Viard I, Bullani R, Tschopp J, French LE. (2000).
J Invest Dermatol 115: 141–148.
Welss T, Sun J, Irving JA, Blum R, Smith AI, Whisstock JC
et al. (2003). Biochemistry 42: 7381–7389.
Xirodimas DP, Saville MK, Bourdon JC, Hay RT, Lane DP.
(2004). Cell 118: 83–97.
Yang A, Kaghad M, Wang Y, Gillett E, Fleming MD, Dotsch
V et al. (1998). Mol Cell 2: 305–316.
Young AR, Chadwick CA, Harrison GI, Hawk JL, Nikaido
O, Potten CS. (1996). J Invest Dermatol 106: 1307–1313.
Zhang T, Woods TL, Elder JT. (2002). J Invest Dermatol 119:
1196–1201.
Zhao R, Gish K, Murphy M, Yin Y, Notterman D, Hoffman
WH et al. (2000). Genes Dev 14: 981–993.