Risk factors for non-melanoma skin cancer (NMSC)
1. Skin carcinoma progression analyzed by genome-wide gene expression profiling and mimicked by validated organotypic in vitro human skin cancer models; towards an integrative system-biology approach and an in vitro screening method for targeted skin cancer therapies.
2. Genetic and environmental risk factors for the development of skin cancer in organ-transplant recipients.

Skin carcinoma is the most common cancer in white populations. The clinical
problem of cutaneous squamous cell carcinomas (SCC) is particularly dramatic
among immunosuppressed patients, e.g. renal-transplant recipients. The most
important risk factor for development of SCC in these patients is lifelong
immunosuppression. In addition, environmental factors including sun (UV)
exposure and human papilloma virus (HPV) infection, and genetic factors
including fair skin and genetic polymorphisms are suspected contributors to
development of the early stages of SCC. For a better understanding of skin
cancer oncogenesis and, ultimately, to reduce the excessive burden of
carcinomas in organ-transplant patients, we plan to initiate two studies.

The first study focuses on the understanding of the oncogenic alterations in
skin cells leading to preneoplastic lesions and tumors as well as on the
establishment and validation of organotypic in vitro skin cultures mimicking
the various tumor stages. A basic understanding of these pathogenic mechanisms
is anticipated to open up the possibility of well-targeted therapies
circumventing the immunocompromised state of the transplant patients. Skin
tumorigenesis is a multistep process and many gene mutations and other genetic
alterations have been reported. The immediate goal of this first study is to
identify pathways (instead of individual genetic changes) that are consistently
involved in the formation of SCCs and their precursor lesions, actinic
keratoses (AK). In addition, we aim to develop robust and validated in vitro
human skin models mimicking SCC in various stages of development, enabling
(high-throughput) studies leading to molecular targets for clinical
intervention. The analysis of skin tumor progression and corresponding genetic
alterations brings together the research interests and expertise of four LUMC
departments: Dermatology, Nephrology, Medical Microbiology and Toxicogenetics.
The unique material consisting of SCC and AK from the same patient allows for
genetic analysis and in vitro modeling of the process of skin tumorigenesis in
isogenic backgrounds.

Secondly, since organ transplantation is complicated by a highly increased risk
of skin cancer after the transplantation, we plan to initiate a case-control
study to identify genetic and environmental risk factors for the development of
skin cancer in organ-transplant recipients. In addition, we plan to focus on
potential associations of genetic polymorphisms and HPV infection with skin
cancer. We plan to focus on genotyping of important candidate genes for skin
cancer so that we will be able to assign persons at risk for skin cancer
already before organ transplantation.

Altogether, firm establishment of consistent shifts in expression profiles of
ensembles of genes related to certain signaling pathways will provide insight
in the multistep SCC carcinogenesis. In addition, validated organotypic in
vitro human skin cancer models will be used for evaluation of novel therapeutic
options, significantly reducing the use of animals for experimentation.
Furthermore, we aim to identify genes, which are associated with an increased
risk of skin cancer and/or HPV infection. This will enable us to identify
organ-transplant recipients with an increased risk of skin cancer and/or HPV
infection. Data of all studies will be integrated, so as to ultimately reduce
the burden of skin cancer in organ-transplant recipients.

Study 1:
The ultimate goal is to abrogate or reduce the excessive burden of carcinomas
in organ transplant patients. We aim to elucidate and establish consistent
changes in genomic and expression (both mRNA and miRNA) profiles in the course
of tumor development from normal skin through AK to SCC. Such results may
reveal consistent changes in signaling pathways and allow identification of
molecular targets for effective intervention. In addition, to address the
clinical problem, we aim to develop robust and validated in vitro human skin
models mimicking SCC in various stages of development, enabling
(high-throughput) studies leading to novel opportunities for clinical
intervention.

Innovative Aspects

The possibility to follow AK and SCC development in an isogenic background
forms a robust approach for the identification of affected key pathways. It
will generate superior data since filtering to reject inter-individual
variability is not required and fewer patients will be needed to obtain
statistically relevant data. Instead of zooming in on single genes, we will
identify affected cellular pathways, which will provide an improved and more
consistent data interpretation. In addition, analysis of differential miRNA
expression allows for possible identification of miRNAs associated with skin
tumorigenesis. By complementary analysis of chromosomal rearrangements (CGH
array analyses, SNP arrays, PCC) and gene silencing (methylation array
analyses) in fresh skin (tumor) material we will identify the molecular causes
for gene expression modulation. The establishment of organotypic in vitro human
skin models mimicking skin carcinoma in various stages allows for mutual
confirmation of gene and miRNA expression data. In addition to the obvious
clinical goal, we intend to introduce a methodically innovative and versatile
approach to experimental in vitro modeling of human cancer, ultimately
eradicating the need for animal models as surrogates.

Study 2:

The current molecular-epidemiological project will study the interaction
between genetic and environmental risk factors for squamous-cell carcinoma
(SCC) in organ-transplant recipients: (a) which associations can be found
between genetic polymorphisms and SCC; (b) how do HPV infection and sun
exposure modulate these associations and (c) which associations can be found
between genetic polymorphisms and HPV infection.

Study 1:

Organ-transplant patients with skin carcinomas are regularly seen in the
outpatient dermatological clinic and have entered a long-term monitoring
programme at this clinic. Biopsies of normal unexposed skin (eg. upper inner
arm), AK, and SCC from the same patient (isogenic) will be collected and either
subjected directly to whole genome and proteome analysis or to establish skin
carcinoma models which will be subsequently analyzed.

In first instance, we aim to include a cohort of 15 patients. On average, one
patient per week may be subjected to the removal of SCC and concurrent removal
of two AKs and two biopsies from normal unexposed skin (inner arm). A central
biopsy will be taken from both the SCCs and AKs before submitting the tumor to
the LUMC Department of Pathology for proper diagnosis and creation of a uniform
group of patients. Blood samples will be drawn as reference material for
genetic analyses. Except for the fresh material required for premature
chromosome condensation (PCC) analyses (see below), half of the biopsy material
will be snap frozen and it is expected that the same material can be used for
m(i)RNA microarrays, CGH and methylation analysis. The other half of the
material will be used for establishment of the in vitro organotypic human skin
carcinoma models. After the first phase, processing of patient material and
interpretation of the results will be evaluated before a second cohort of 20
patients will enter this combined study.

Organotypic in vitro human skin carcinoma models

For the generation of skin carcinoma models we will use the skin explant
technique developed in our laboratory. This technique, initially pioneered by
Boxman et al. for growth of SCC cells, and later further optimized by El
Ghalbzouri et al, allows the expansion of keratinocytes as well as an increased
life span of the skin explant within a 3D culture model. Skin models generated
from normal skin, AK and SCC will be sampled in time (e.g. 2 and 8 weeks) and
the keratinocytes will be subjected to a whole genome and targeted proteome
analyses. All data arising from these skin models will be compared to
expression data obtained directly from the epidermal cells from skin biopsies
of transplant patients. By combining these results we aim to deduce a
(pre)malignant molecular signature of the different stages of skin cancer and
determine whether it is conserved and maintained in time within the respective
human skin carcinoma models. Moreover, this combined approach will deliver
biological endpoints which might be potential targets for the development of
new treatment strategies of these tumors.

Laboratory material processing for direct analysis

For direct analysis of gene expression, part of each half biopsy will be
sectioned for (immuno) histochemical analyses to check for admixtures to the
tumor cells of stroma and infiltrating lymphocytes. About 25-50 sections of 20
µm thickness will be cut and trimmed to remove surrounding dermal tissue. These
sections should at least contain 70% tumor cells. Normal skin biopsies will be
treated according to the split skin method to separate the epidermis from the
dermis. These isolated epidermal sections will be used as reference material
for the array analyses of AKs and SCCs.

RNA and DNA isolation

If a biopsy contains too much admixture of stroma or infiltrating cells, it
will not be processed for subsequent extraction of m(i)RNA and DNA. Recently,
several commercially available kits have been introduced that allow
simultaneous isolation of RNA and DNA from the same tissue section. This allows
us to isolate sufficient amounts of high quality RNA and DNA for further
analyses from biopsy material. However, if too many SCC biopsies have to be
rejected because of admixtures, we will start microdissecting the material to
purify the tumor material. Evidently, the RNA and DNA then needs to be
amplified before further analyses.

Genome wide analysis of gene expression and genomic alterations

Genome-wide assessment of gene expression is carried out by m(i)RNA microarray
analysis. A pair-wise comparison of obtained expression data will discriminate
between biological (individual primary samples) and experimental (organotypic
skin models) variation within the group of test samples (e.g. unexposed skin,
AK or SCC) and statistically significant differences between groups. Results of
interest will be verified using quantitative RT-PCR and in situ hybridization.
The outcomes of microarray analysis will be correlated to other genome-wide
analyses, including SNP arrays or CGH arrays to detect gene amplifications or
losses, and CpG islands methylation arrays. The challenge with this wealth of
data lies in the extraction of a coherent and robust model. The presently
available techniques can readily be employed to analyze tumor progression, and
outline characteristic changes that evolve in the transition from normal
(naïve) skin to AK, and from AK to SCC, and possibly further to metastatic SCC.

Bio-informatical analyses

Computational analysis is essential to transform the masses of generated data
into a mechanistic understanding of disease. We will use up to date approaches
in microarray expression analysis that aim at uncovering the modular
organization and function of transcriptional networks and responses in cancer
instead of zooming in on changes of single genes.

Targeted proteome analysis

First, genes of interest selected on the basis of their altered gene expression
pattern (e.g. signal transduction pathways/ oncogenic pathways) in
(pre)malignant lesions will be analyzed by Western blot and immunohistochemical
techniques (depending on antibody availability) in the HSE generated from
normal skin, AK and carcinoma. This will provide insight as to whether
transcriptional differences indeed result in differences in protein expression.
Second, we will use commercially available antibodies against activated
(phosphorylated) signaling proteins demonstrated to play a role in early stages
of carcinogenesis of skin cells or other epithelial tumors (e.g. MAP-kinase or
PI3K/Akt signaling). Third, unbiased expression analysis of the epidermal
proteins of skin models will be performed using quantitative proteomics
approaches, including nanoHPLC and 2D-PAGE followed by mass spectrometry as
described. Preliminary studies show that the use of fluorescent dyes greatly
enhances the sensitivity of the 2D-approach and predicts that multiple samples
(e.g. HSE generated from normal skin, AK and carcinoma or HSE generated from
carcinoma and sampled in time) can be separately labeled, combined and
subjected to a single analysis in a 2D run. This will demonstrate quantitative
differences of proteins between the different stages of tumor development. Such
differentially expressed proteins will be subjected to mass spectrometric
analysis for identification as described and subsequently validated by
immunohistochemistry and/or Western blotting.

Mutational and HPV analyses

Tumors will be screened for the presence of known oncogenic factors (e.g. P53
mutations) and HPV to be correlated to results from the gene profiling analyses.

Premature chromosome condensation (PCC)

Chemical induction of premature chromosome condensation (PCC) will be used to
analyze the chromosomal constitution of SCCs and AKs. The protein phosphatase
inhibitor calyculin A is capable to induce PCC in cells isolated from freshly
dissected adenomatous polyps of a patient (as small as 2-4 mm2) with hereditary
colorectal cancer (Bezrookove et al, 2003). In fresh tumor specimens, a regimen
of 80 nM calyculin A for 75 min after only 2 days of culturing, resulted in a
PCC index of 2-5%. pq-COBRA-FISH (Combined Binary Ratio labeling-fluorescence
in situ hybridization) will be used for molecular karyotyping of numerical and
structural chromosomal abnormalities.

Model validation and proof of principle

We aim to stimulate the identified carcinogenic events in the skin by
engineering HSEs that harbor characteristics of (pre)malignant skin tumors.
Since future (high throughput) screening for therapeutics requires production
of in vitro models on demand, we will start with exploring whether storage and
reseeding of keratinocytes obtained from growing skin explants is a feasible
approach, providing that the molecular (pre)malignant signature is maintained
in skin explants. Generated 3D models will be validated for the presence of the
(pre)malignant characteristics using the m(i)RNA and protein markers identified
above. Finally, as proof of principle, we aim to assess whether metabolic
blockers known to be effective in suppressing skin cancer outgrowth in both
mice and human, are also effective in the treatment of skin models with
(pre)malignant phenotype. Effects of treatment will be determined by analysis
of gene and protein expression profiles.

Study 2:

Organ-transplant patients with a high tumor burden have entered a long-term
monitoring program run by Dr. Jan Nico Bouwes Bavinck at the Dermatological
clinic of the LUMC. We would like to include 150 organ-transplant recipients
with a history of SCC and 250 recipients without a history of skin cancer,
matched for age, sex and time period after transplantation. Questionnaires will
be used to collect data about sun exposure, smoking, etc. Hair bulbs supposedly
represent the reservoir of skin HPV types. HPV infection will be measured by
the presence of high-risk HPV types in plucked eyebrow hairs. Blood will be
collected for the isolation of genomic DNA for genetic typing. We already
collected these materials from 966 immunocompetent persons. The latter group,
consisting of 161 patients with SCC, 426 with basal cell carcinoma or malignant
melanoma and 386 persons without any skin cancer will be used as a control
group representative for the non-immunosuppressed population.

HPV and genetic analyses

The presence of HPV in the plucked eyebrow hairs of the patients will be
determined by established PCR-based assays as well as antibodies against viral
early and late proteins. Genetic analyses will start with validation by
confirming the association between SCC and IL-10 polymorphisms, which are
reported to be associated with SCC in transplant recipients. Next additional,
cytokine polymorphisms will be characterized, e.g. TNF*, IL-1*, IL-2, IL-4, and
IL-6. Time and money permitting, more candidate genes (XP, EVER, GST, etc.)
will be tested. Genotyping will be performed using the Sequenom MassArray
system, a well-established platform for high-throughput genotyping using
MALDI-TOF mass spectrometry that is present in the laboratory of department of
Molecular Epidemiology of our Centre. The association between the genetic
polymorphisms and SCC and the modulation of this association by HPV infection
will be statistically analyzed by multivariate techniques.

Not applicable.