The Human Protein Atlas – Organisation in Uppsala

Dr Cecilia Lindskog Bergström PI and Site Director HPA Uppsala
Prof Fredrik Ponten Clinical Director
Tissue Microarray Production, Immunohisto-
chemistry, and Scanning
Ing-Marie Olsson
Antibody approval, Protein Profiling, and
Antibody Destiny
Dr Cecilia Lindskog Bergström
Clinical Pathology       Prof Fredrik Ponten

Tissue Microarray Production, Immunohistochemistry, and Scanning

Team leader: Ing-Marie Olsson (research engineer)

Personnel: Anna-Maria Dénes (research engineer), Jonas Gustafsson (research engineer), Dennis Kesti (biomedical analyst), Lillemor Källström (biomedical analyst), Rutger Schutten (research engineer)

Responsibility, TMA production: (i) Production of tissue microarrays (TMAs) and cell microarrays (CMAs), (ii) handling of tissues (biobank material), cells and cell lines for TMA and CMA production and protein extraction, (iii) sectioning of tissue and TMA blocks, (iv) quality control of TMA-blocks, (v) maintenance and cultivation of cell lines and (vi) protein extraction for Western blot analysis.

Description, TMA production: Tissue microarrays are used in the basic protein profiling for the Human Protein Atlas. Each protein expression profile is based on 8 TMAs including normal tissues from 144 different individuals (triplicate samples of 44 different tissue types) and cancer tissues from 216 different patients (duplicate samples of tumor tissues representing the 20 most common forms of human cancer). In total, 576 tissue cores are immunostained and analyzed for each antibody. The CMAs are used for antibody validation and include duplicate samples of 44 cell lines.

Formalin fixed, paraffin embedded tissue specimens are collected from the Department of Pathology, Uppsala University Hospital. Representative areas to include on the TMAs are defined for each tissue specimen by visual inspection of a corresponding hematoxylin-eosin stained section under a microscope. Tissue cores from these selected areas are then taken and used for TMA production. Cell lines are cultured in vitro, harvested, fixed in formalin and dispersed in agarose prior to embedding in paraffin blocks. These cell-blocks are used for the construction of CMAs.

Responsibility, Immunohistochemistry: (i) Handling and storage of antisera, (ii) test and titration of antibodies for immunohistochemistry, and (iii) immunostaining of TMAs and CMAs.

Description, Immunohistochemistry: Antibodies are titrated using a specially designed test TMA, representing a limited selection of tissues and cells. The titration is performed based on stringent criteria, according to which different antibody dilutions are tested. The primary working dilution is based on the protein concentration for each antibody. When an optimal antibody dilution is found automated immunostaining is performed on sections from eight TMAs for basic protein profiling in the Human Protein Atlas project. Instruments and commercially available detection kits are used to ensure standardization and reproducibility of immunohistochemistry. In addition, protein extractions are prepared from fresh frozen tissues and cell lines for the purpose of running Western blot.

Responsibility, Scanning: (i) Scanning of immunohistochemically stained TMA slides, (ii) export and processing of scanned digital images from TMAs, and (iii) administration of TMAx for automated image analysis.

Description, Scanning: Immunostained TMA slides are scanned to generate high-resolution digital images, using 20x (tissue) or 40x (cells) magnification. Images from scanned TMAs are separated into spot images, representing immunostained tissue (1 mm diameter) or cell sample (0.6 mm diameter), and exported as individual TIFF files. Manual, pathology-based evaluation of images and annotation of protein expression in tissue is performed using an in-house built web based annotation software. Automated image-analysis algorithms using the TMAx system are used to generate protein expression data from images of immunostained CMAs.

Antibody approval, Protein Profiling, and Antibody Destiny

Team leader: Dr Cecilia Lindskog Bergström

Personnel: Anna-Maria Dénes (research engineer), Jonas Gustafsson (research engineer), Borbala Katona (research engineer), Feria Hikmet Noraddin (research engineer), Emil Lindström (research engineer), Rutger Schutten (research engineer), Charlotte Soläng (research engineer), Josefin Thelander (research engineer), Jimmy Vuu (research engineer)

Responsibility: (i) Validation of antibody target specificity (ii), maintenance of immunostaining reproducibility and quality, (iii) coordination and evaluation of antibodies submitted by commercial vendors and academic scientists, (iv) annotation and final approval of immunohistochemically stained normal and cancer tissues (v) generation of knowledge-based protein expression profiles, and (vi) continuous curation and quality assurance of protein profiles.

Description: Optimal antibody dilution and target specificity is assessed by microscopical examination. Available information in public databases on gene, RNA and protein level, internal and external RNA-seq data, as well as internal technical validation, including independent antibodies, protein arrays and Western blots, are considered in the decision process. For each approved antibody, a final immunostaining protocol is defined and subsequently applied to the full-scale TMAs. The images are then scanned for generation of high-resolution digital images. Manual annotation of immunohistochemistry is performed by specially educated personnel using a web-based in-house developed software to record the intensity and fraction of immunoreactive cells for each given cell population, and to determine the subcellular localization of immunoreactivity. A text comment summarizing the characteristics for each antibody is added to the annotation. The results are visualized in a summary view as bars corresponding to the protein expression level in each given cell type. In total 76 normal cell types from 144 individuals and 20 different cancer cell types from 216 different tumors are annotated for each antibody. An independent second observer curates all finished annotations to ensure uniform annotations of high quality. After annotation and curation, protein profiles are evaluated and the generated data is compared with gene, RNA and protein characterization data. Antibodies that are approved at this stage are assigned a reliability score and a knowledge-based protein expression profile and released for publication in the next version of the Human Protein Atlas.

Clinical Pathology

Team leader: Prof Fredrik Pontén

Personnel: Anna-Maria Dénes (research engineer)

Responsibility: (i) Develop strategies to identify potential biomarkers based on the HPA database and other efforts, (ii) validate proteins that can be used as clinical biomarkers for disease, (iii) participate in clinical studies, collect tumor material and clinical data to generate specific cancer TMAs coupled to clinical databases, (iv) perform statistical analysis and validate the clinical usefulness of identified biomarkers.

Description: The HPA database is actively mined for potential biomarkers with the aim to identify protein expression patterns that could be of medical or biological significance. Projects include various forms of human disease with an emphasis on cancer. Most projects are focused on identification and validation of biomarker candidates that can fulfill currently unmet clinical needs related to diagnostics, prognostics and treatment prediction. To address such questions, patient cohorts representing different cancers are defined and tumor material as well as clinical data is collected. These specifically designed cancer TMAs are produced and used for extended analysis of protein expression patterns to test and validate candidate proteins as useful biomarkers. The biomarker discovery and validation efforts include both internal projects and external collaborative projects.