Seminário "Towards Predictive Medicine"

Risk factors in the framework of strongly coupled, non-linear biological systems

---

23/07/2012

Orador: Andrzej K. Drukier (BioTraces Inc. and George Mason University)

Abstract:

In preliminary studies with 1,200 clinical samples, we used ultrasensitive proteomics to study roles of cytokines/growth factors during inflammation invasion, tumor progression and metastasis. New risk factors have been elucidated. We argue that the main role of the immune system is in the oncogenesis stage, and plays a key role in suppression in the very early stages of cancer. Thus, we describe a program on a border between Predictive Oncology and Diagnostic Oncology, and focuses on reliable measurement and system analysis of low abundance signaling proteins.

This program includes novel methods of evaluating an individual's risk for cancers with focus on breast cancer. Currently, predictive medicine is based on epidemiological evidences, on correlation between different activities and diseases, e.g. smoking being a strong risk factor for lung cancer and diabetes for cardio-vascular disease. The concept of predictive biomarkers is seen as a next generation of predictive medicine. A large number of biomarkers have been proposed as there are about 10,000 different proteins in the human organism (including post-translational modification). About 20% of them may be involved in pathways involved in cancer. However, such description is appropriate when the biologic system is a weakly interacting, semi-linear system: the majority of important biological systems are however strongly coupled, non-linear systems. In such systems the notion of "independent" biomarkers is not a proper descriptor; the correlations of biomarkers are more informative than any individual biomarker.

We have studied how panels of ultrasensitive immunoassays targeting a new class of biomarkers may be used in diagnosis. When synergistically used, this permits permit a 10-20% increase in predictive power as compared with prior-art blood-based assays. For example, using significant cohorts HW(N=562) and BC(N=575), we have demonstrated about 95% sensitivity/specificity in blood-based detection of BC. Especially interesting are panels, which include low abundance signaling proteins such as cytokines and angiogenesis factors. New classes of panels were studied in which some bi-markers and tri-markers provide information about a local process at the site of the disease, while other small panels of bio-markers provide information about the organism's response to the disease, including the immune and angiogenesis responses.

The results of a pilot study are very promising: we have elucidated a few immune sub-types (overall 15-20% of general cohort) in which the risk for breast cancer is 5-10 fold higher. We have also found other immune sub-cohorts (20-25% of general population) in which the risk for breast cancer is lower. However, there is an evident weakness of our data: in some sub-cohorts the statistics of one of the groups (either HW or BC) is very low, whereas in other groups the statistics are significant.

Organização

Centro de Física Nuclear da Universidade de Lisboa

Contacto

Prof. Thomas Abbott Girard - Email: criodets@cii.fc.ul.pt – Telf.: 217 904 935

Website

http://cfnul.cii.fc.ul.pt/

Localização

Anfiteatro, Instituto para a Investigação Interdisciplinar da UL, Lisboa

Horário

14h00