Laboratory for In Situ Technologies and Morpho-Proteomics
Main Field(s) of Research, Abstract
Adequate cancer management depends on accurate tumor staging and grading together with relevant predictive and prognostic molecular markers.
We are using an image-based computational method of pan-cytokeratin immunohistochemistry (pan-CK IHC) staining to characterize morphological features of lung squamous cell carcinoma (LSCC) with prognostic relevance. Chemotherapy-naïve and neo-adjuvant chemotherapy-treated LSCC patients are analysed. We identified tumor fragmentation (TF) as a new morphologic feature significantly associated with blood vessel infiltration and poor outcome in two independent clinical cohorts by analysing tissue microarrays and whole tumor sections. TF was defined as the total number of epithelial tumor sheets separated from each other by desmoplastic stroma per given 2D surface. The prognostic value of TF was confirmed by a similar human-based scoring of haematoxylin-eosin (HE) stained tissues from The Cancer Genome Atlas (TCGA) and was independent from tumor stage. Protein extracts were prepared from formalin-fixed paraffin-embedded (FFPE) tumor blocks by the filter-aided sample preparation (FASP) column and tryptic peptides analysed by shotgun mass spectrometry (LC-MS/MS) at Functional Genomics Centre Zurich (FGCZ). Integrated proteomic and genomic analyses revealed an increase of extracellular matrix remodeling and focal adhesion processes in tumors with high TF. TF also correlated with markers of epithelial-mesenchymal transition such as periostin. Wet and fixed LSCC samples were analysed by X-ray microtomography at the TOMCAT beamline of the Paul Scherrer Institute (PSI), demonstrating that 2D tumor fragments correspond to coherent sprout-like tumor outgrowths of variable thickness in 3D. In conclusion, we think that TF is useful for improved grading of lung squamous cell carcinoma.
Currently, we are evaluating the histologic changes after neo-adjuvant chemotherapy of LSCC patients by quantifying vital, necrotic and stromal tumor compartments on the pan-CK IHC. Histologic features are correlated with pre-surgical positron emission tomography (PET) as a measure of tumor metabolic activity. To detect novel prognostic and diagnostic morphologic features unrecognizable to the human eye, we have started a deep learning project for computerized classification of NSCLC into squamous cell or adenocarcinoma on a convolutional neural network (CNN) at Zurich University of Applied Sciences (ZHAW).
For implementation of micro-immunohistochemistry as key technology for in-situ proteomics, we are using a prototype device of the Microfluidic Tissue Processor (MTP) developed at Lunaphore SA, EPF Lausanne, for whole section IHC following laser microdissection, DNA extraction and next generation sequencing. We used MTP for identification of the predictive BRAF V600E mutation in lung adenocarcinoma and for PTEN IHC in the European Thoracic Oncology Platform (ETOP) Lungscape PTEN 002 project. Such oncogenic markers interfere with peri-tumoral immunity and we have thus started to correlate them with corresponding immune biomarkers. Thereby, we treat tissue slice cultures of solid NSCLC and liquid cultures of malignant pleural effusions with chemo- and immunotherapy, respectively.
Main Fields of Research, Keywords
Lung cancer, computerized image analysis, microfluidic immunohistochemistry, tissue slice culture, liquid effusion culture, localized molecular analysis, laser microdissection, in-situ proteomics.
Special Techniques and Equipment
Automated IHC platforms Ventana Ultra and Leica Bond RX, fluorescent in-situ hybridization (FISH), chromogenic/silver in-situ hybridization (CISH/SISH), Palm-Zeiss laser capture microdissection (LCM), microfluidic tissue processor (MTP).
Education and Training
We hold regular research seminars and provide histopathological and immunohistochemical expertise for human and animal tissue samples.