People

 Members
 Bioinformatics
 Facilities

Resources

 Publications
 Recent Grants
 Available Positions
 Where We Are
 Links

Recent Grant

The adult human organism contains heterogeneous reservoirs of multipotent stem cells characterized by a diversified differentiation potential. Understanding their biology at a system level would advance our ability to selectively activate and control their differentiation potential.
Aside from the basic implications this would represent a substantial progress in regenerative medicine by providing a rational framework for using small molecules to control cell trans-determination and reprogramming.
Here we propose a combined experimental and modelling approach to assemble a predictive model of mesoderm stem cell differentiation. Different cell states will be identified by a vector in the differentiation hyperspace, the coordinates of the vector being the activation levels of a large number of nodes of a logic model linking the cell signalling network to the transcription regulatory network.
The premise of this proposal is that differentiation is equivalent to rewiring the cell regulatory network as a consequence of induced perturbation of the gene expression program. This process can be rationally controlled by perturbing specific nodes of the signalling network that in turn control transcription factor activation.
We will develop this novel strategy using the mesoangioblast ex vivo differentiation system. Mesoangioblasts are one of the many different types of mesoderm stem/progenitor cells that exhibit myogenic potential. Ex vivo, they readily differentiate into striated muscle. However, under appropriate conditions they can also differentiate, into smooth muscle and adipocytes, albeit less efficiently. We will start by assembling, training and optimizing different predictive models for the undifferentiated mesoangioblast. Next by a combination of experiments and modelling approaches we will learn how, by perturbing the signalling models with different inhibitors and activators, we can rewire the cell networks to induce trans-determination or reprogramming.

According to The World Health Organization (WHO), overweight and obesity are the most important known avoidable causes of cancer after tobacco. Indeed, a number of epidemiological studies indicate a direct correlation between body-mass index and incidence of cancer.
On the same line, evidence has been accumulating over the past 100 years that lifestyle can influence cancer initiation and progression and that a diet rich in calories can increase the risk of cancer incidence and relapse. Autophagy has been proposed as one of the main cell processes that are affected by diet while, at the same time, affecting via complex mechanisms the risk of developing cancer. We have formed a multidisciplinary team to ask whether the link between caloric intake and cancer could be mediated by modulation of autophagy and whether or not this regulation occurs via epigenetic mechanisms.
The project stems from and takes advantage of a unique bank of human samples (ORDET and DIANA) collected over the years by members of this consortium and annotated for dietary habits. The individuals participating in this trial were followed for twenty years in order to be able to correlate caloric intake, and other variables, to cancer development. We propose to apply a number of high throughput sequencing approaches to characterize the transcription and epigenetic profile of leukocyte samples from individuals participating in this trial. The approach will be complemented by studies in animal and cell models to investigate the molecular mechanisms underlying the organism response to low caloric diet and their relation to cancer development.
We aim at gathering information that once integrated in a general model will help interpreting the results of the human studies while at the same time forming a molecular framework for rationally designing new combined diet and pharmacological regimens to help decreasing risk of cancer incidence.

Background
Cancers are heterogeneous both at the genotypic and the phenotypic level. In early days this statement was only based on histological observations. More recently the development of technologies for genome wide sequencing of single cells has provided evidence for a genetic heterogeneity that is even higher than originally anticipated. This phenotypic heterogeneity, aside from providing evidence for mechanisms of tumor evolution and clonal selection, can be linked to crucially important clinical aspects such as prognosis, resistance to therapy and ability to seed metastases.
Hypothesis
Over recent years evidence has been accumulating that solid tumors comprise diverse cell subpopulations with distinct phenotypic traits that have genetic and non-genetic origin. In the era of targeted treatment, evidence for a relationship between intra-tumor heterogeneity, resistance to therapy and clinical outcome is emerging. The premise of this proposal is that, in order to achieve a rational approach to treatment, one should fully describe the diverse cell populations in a tumor and understand their response to environmental perturbations, including therapeutic treatments.
Aims
The aim of the project is to develop a strategy to fully characterize the molecular phenotypes of diverse cell populations in tumors and understand their dynamic behavior when the system is perturbed. The proposed approach is based on a new technology, mass cytometry, that permits to measure up to approximately 40 molecular parameters of each cell in a large (millions) mixed cell population. Compared to traditional cytometry based on fluorescent probes this new hardware permits to distinguish even minority populations with little phenotypic differences.
Experimental design
We propose to use a mouse model of myosarcoma to define and compare the different cell populations in the healthy muscle and in tumor tissues. This will be achieved by measuring the expression levels and the activation state of a large number of markers (cell surface proteins and phospho-peptides) at the single cell level in cell suspensions. By applying barcoding multiplexing strategies this basic experiment will be repeated under a large number of experimental perturbations both in vivo and ex vivo. The resulting large dataset describing the network of molecular relationships will be used to train cell-population specific logic models that may be used to infer the behavior of the system under conditions that have not been tested experimentally.
Expected results
- A molecular profile of the phenotypic diversity of cell populations in a tumor in comparison with that observed in the corresponding normal tissue.
- A predictive logic model describing how each cell population responds to genetic, environmental and chemical perturbations.
- A rational framework to design and interpret therapeutic protocols.