of the Max - MDC

Surgical Oncology
Peter M. Schlag
Our group established predictive and prognostic gene signatures and expression profiles for analysis of metastases
and drug resistance in colorectal cancer. The data provide the basis for novel therapeutic concepts in tumor
treatment that are applicable and validated in the clinic. We were successful in identifying new metastases-associated
genes, their importance for cancer metastases and their function in signaling pathways. We isolated and characterized
the new gene MACC1, revealed S100A4/metastasin as novel target gene of the β-catenin pathway and analyzed
the interplay of BMP-4 and Bambi in the context of β-catenin signaling. This provided insights into the tight
association between early and late events of the metastasis process particularly in colorectal cancer. In collaboration
with the groups of W. Birchmeier and M. Lipp achievements were made in the identification of patient-individualized
gene expression profiles for improved diagnosis and prediction regarding metastases formation and
patient survival. In addition to the progress made in molecular biological research, we were able to significantly
advance in the development of the nonviral in vivo gene transfer for effective local cancer gene therapy. The establishment
of the applicable jet-injection based gene transfer technology led to the initiation of the clinical testing
in a “proof of principle“ phase I trial at the Robert-Rössle-Klinik. Moreover, we developed and tested efficient therapy-controllable
vector systems for conditional transgene expression. These vectors will be employed in a multimodal
therapeutic setting, combining nonviral gene therapy with chemotherapy or hyperthermia.
Prognostic GeneChip for Colorectal Cancer
Wolfgang Kemmner, Marion Fehlker, Ulrike Stein,
Johannes Fritzmann, Peter M. Schlag
Colorectal cancer is one of the leading causes of death in
developed countries. Although prognosis for colorectal
cancer has improved during the last 20 years, nearly onehalf
of patients succumb as a result of metastasis. In an
initial study expression profiles of early colon cancer have
been studied by whole human genome oligonucleotide
microarrays. GeneChip analysis led to the identification of
some candidate genes showing a differential expression
between epithelial cells of healthy and tumor patients, but
also between non- and metastasizing carcinomas. Moreover,
several genes showing a strong over- or underexpression in
carcinoma cells of patients with poor survival were identified.
The next step was the development of a low-density
microarray with the aim to provide a cheap and easy to use
chip for the clinical routine. Oligonucleotides (70-mers) for
300 candidate genes identified in the initial genome wide
studies were spotted onto glass slides. The prognosis chip
was completed by the addition of probes for 90 stably
expressed genes, so-called housekeepers, and some common
prognostic molecules and tissue markers. Colorectal
carcinoma samples and normal mucosa were available from
the tumorbank of the Robert-Rössle-Klinik. Cryosections
were prepared of these samples, and after pathological survey
and sample processing of sections containing at least
60 % of carcinoma tissue, single color labelled cDNA was
hybridized to the chip. Meanwhile, in a retrospective study
70 colorectal cancer samples with a long-term follow-up of
more than 5 years were examined. Bioinformatic analysis
proved the robustness of the profiling. Using unsupervised
random-forest clustering some genes which show an expression
correlating well with disease-free survival was identified.
Our results show that by using the prognosis chip a
better prediction of patient survival were achieved than by
using histopathological parameters such as invasiveness (T)
and grading (G) alone. In an ongoing prospective study the
prognostic significance of the microarray for colorectal
cancer patients will be validated.
Identification of genes predictive for human
colon cancer metastasis and their functions in
cellular signaling pathways
Ulrike Stein, Wolfgang Walther, Franziska Arlt, Janice Smith,
Peter M. Schlag. In cooperation with Walter Birchmeier,
Iduna Fichtner (MDC), Erik D. Harris, Susan D. Mertins,
and Robert H. Shoemaker (National Cancer Institute,
Frederick, MD), Todd Waldman (Georgetown University,
Washington, DC), Claus Heizmann (University Zürich,
Switzerland), David Allard (University Liverpool, UK)
We identified an EST without any significant homology to
known sequences by comparing the gene expression pat-
Cancer Research 85