Elsevier

Drug Discovery Today

Volume 8, Issue 24, 15 December 2003, Pages 1121-1127
Drug Discovery Today

Review
In silico multicellular systems biology and minimal genomes

https://doi.org/10.1016/S1359-6446(03)02918-0Get rights and content

Abstract

The in vivo and in silico understanding of genomes and networks in cellular and multicellular systems is essential for drug discovery for multicellular diseases. In silico methodologies, when integrated with in vivo engineering methods, lay the groundwork for understanding multicellular organisms and their genomes. The quest to construct a minimal cell can be followed by designed, minimal multicellular organisms. In silico multicellular systems biology will be essential in the design and construction of minimal genomes for minimal multicellular organisms. Advanced methodologies come to light that can aid drug discovery. These novel approaches include multicellular pharmacodynamics and networked multicellular pharmacodynamics.

Section snippets

Genome semantics or how does a chimpanzee differ from a human?

The genes of a chimpanzee and a human are virtually identical, therefore, it must be the regulatory genome architecture where the genomes and humans and chimpanzees will differ sufficiently to account for the differences in development, morphology and behavior.

The central problem of post-genomic biology and medicine is to understand the meaning of genomes. To understand genomes we need to view them in their biological context. This includes the context of the cell and the dynamic context of a

The systems biology hierarchy

Computational systems biology 8, 9, 10, 11, 12, 13, 14 can be described as a set of areas of research and modeling that fall into a hierarchy (Figure 1). In this hierarchy, all levels have an equal contributing value; for example, while the higher levels tend to be supported by the lower levels, a higher level might also be used to organize and give meaning to a lower level. To illustrate this, mutations in the genome, at a low level, are described in terms of their effects on the morphology at

Properties of systems

Ideally, in an analogous way to the relationship between phenomenological and statistical thermodynamics, systems biology will permit the modeling of properties at different levels of abstraction and ontology. The information required depends on the level of information that we are modeling. Each level of information has its own ontology of objects and relations. Thus, we can have models that are incomplete and yet accurate at a given level. For example, one might model the effects of homeotic

The semantic hierarchy and the meaning of genomes

To the hierarchy of levels modeling in systems biology, there corresponds a semantic hierarchy of levels of information and ontology. Furthermore, there is an information flow between these levels. The specification of the semantics of genomes requires that we understand how the cell interprets the genome in different contexts and levels. At the lowest level, we have the transcription into RNA and then translation of mRNA into proteins. At a higher level, proteins have particular functions

Problems with the bottom-up approach

Although there is a lot of data being generated in laboratories about the expression of genes, RNA and proteins, we can not interpret the data unless we have a semantics of the given genome. In other words, we need to understand how the genome functions in the multicellular system. How do we get the semantics? The dominant approach is bottom-up; we create mutations and observe the effects. However, genomes and their organisms are highly complex and, as a result, using a direct bottom-up search

In silico genomes

One way of reducing the search space for the semantics of genomes is to conduct in silico simulations of genomes that lead to multicellular phenomena that correspond to natural multicellular phenomena. In such an approach, the semantics of genomes is constrained by what we know from research in molecular biology, cell biology and genetics, as well as a century of experimentation in developmental biology. Modeling multicellular or cellular processes is, in essence, a process of theory

Minimal genomes and minimal cells

Complementary to a virtual, in silico approach that models the semantics of single cell and multicellular genomes, is the in vivo construction of genomic networks that regulate cell processes. The work by Venter and others to construct a minimal cell with a minimal genome is part of this effort 34, 35, 36, 37, 38, 39. Whereas Venter's group is trying to construct a minimal cell bottom-up, others 37, 38, 39, 40 are reducing simple bacterial genomes to find a minimal set of essential genes. The

Minimal multicellular genomes

If the construction of a minimal cell can be achieved, what happens next? I believe that the next step is to investigate the regulatory properties of minimal cells. After that it becomes possible to investigate minimal multicellular genomes (mMCG) and their organisms. An mMCG for a multicellular system is the simplest genome that is capable of generating that system. In other words, mMCGs generate mMCSs. An mMCS is a multicellular structure that develops from a single cell using a minimum of

Engineered mMCOs and drug discovery

This forward engineering process will open up new areas of biotechnology as well as multicellular pharmacodynamics (MCPD; see subsequent sections). In particular, networked MCPD might see great advantages in that an in silico model of an mMCO can then be tested and corrected by how the corresponding natural mMCO responds to a drug. One problem with the minimalist approach is that because the cell is minimal it might lack some of the drug responses of the normal cell. However, this is outweighed

Systems biology and drug discovery

As described previously, the genome can and should be interpreted at different levels of information and ontology. In the drug discovery process, there is also a realization by scientists that an ontological view restricted to the molecular level might hinder the drug discovery process. There is a movement to complement the molecular view with a systemic view in drug design 44. Regulatory networks require a systemic understanding. Although it is certainly true that there are powerful and

Multi-cellular pharmacodynamics

Pharmacokinetics is the study of the distribution of drugs in organs and tissue. Pharmacodynamics (PD) goes a step further and attempts to get at the causes and workings of ADMET properties 44. We propose to extend PD one step further.

Multicellular modeling of PD uses multicellular models where cells can be in diversely differentiated states. It uses multicellular tissue to model and dynamically simulate and display the effect of drug distribution and other ADMET properties and is a hybrid

Networked multicellular pharmacodynamics

Furthermore, MCPD models can be extended to model regulatory genomic networks together with signal transduction pathways, as part of a complex of interacting components in the cell; these are known as networked-MCPD (Net-MCPD) models. In this way, drug interactions with the cell, the genome, the cell signaling dynamics and the multicellular system can also become accessible to modeling. In this approach, many levels of the systems biology hierarchy are involved in modeling and simulation to the

In silico cancer modeling and simulation

For example, in cancer the regulatory networks in the genome and cell signaling dynamics can have a key role in the etiology, ontogenesis and dynamics of the disease. In this case, drug ADMET properties must be supplemented by additional properties that influence the dynamics of relevant cellular disease states. MCPD models enhanced with genome and cell signaling components might give us a deeper insight into the dynamics of cancers and their response to drugs in a dynamic multicellular context.

Conclusion

At present, drug discovery is still dominated by a bottom-up approach that mimics the flow of information dictated by the Central Dogma 47. However, there are inherent limitations with this approach because of the NP-complexity of the search space. Here, an alternative, multileveled approach has been proposed that includes in silico multicellular systems biology in tandem with in vivo forward and reverse engineering methods to analyze and design minimal genomes for mMCSs. A test bed for drug

References (47)

  • H. Bolouri et al.

    Modeling transcriptional regulatory networks

    Bioessays

    (2002)
  • E.H. Davidson

    Genomic Regulatory Systems: Development and Evolution

    (2001)
  • Werner, E. (1996) Logical foundations of distributed artificial intelligence. In Foundations of Distributed Artificial...
  • Werner, E. (1992) The design of multi-agent systems. In Decentralized AI (Vol 3) (Werner, E. and Demazeau, Y., eds),...
  • H. Kitano

    Systems biology: a brief overview

    Science

    (2002)
  • H. Kitano

    Computational systems biology

    Nature

    (2002)
  • Werner, E. (2003) In silico cell signaling underground. Sci. STKE 170,...
  • J.M. Carlson et al.

    Complexity and robustness

    Proc. Natl. Acad. Sci. U. S. A.

    (2002)
  • M.E. Csete et al.

    Reverse engineering of biological complexity

    Science

    (2002)
  • B. Palsson

    The challenges of in silico biology

    Nat. Biotechnol.

    (2000)
  • B.O. Palsson

    What lies beyond bioinformatics?

    Nat. Biotechnol.

    (1997)
  • T. Werner

    Cluster analysis and promoter modelling as bioinformatics tools for the identification of target genes from expression array data

    Pharmacogenomics

    (2001)
  • E.P. van Someren

    Genetic network modeling

    Pharmacogenomics

    (2002)
  • Cited by (15)

    • Structure and dynamics of molecular networks: A novel paradigm of drug discovery: A comprehensive review

      2013, Pharmacology and Therapeutics
      Citation Excerpt :

      cytoskeletal networks or membrane organelle networks (Michaelis et al., 2005; Escribá et al., 2008; Gombos et al., 2011); inter-neuronal, inter-lymphocyte and other intercellular networks including extracellular matrix, cytokine, endocrine or paracrine networks (Jerne, 1974, 1984; Cohen, 1992; Werner, 2003, 2005; Small, 2007; Jung et al., 2011; Acharyya et al., 2012; Margineanu, 2012); the ecological networks of the microorganisms living in human gut, oral cavity, skin, etc. and their interconnected networks with human cells (Ben-Jacob et al., 2012; Clemente et al., 2012; Mueller et al., 2012);

    • History of the seminar on the conceptual foundations of systems biology

      2013, Progress in Biophysics and Molecular Biology
    • The influence of traditional herbal formulas on cytokine activity

      2010, Toxicology
      Citation Excerpt :

      The partial perturbations of medicinal plants on a pharmacological network mimic physiological scenarios where hundreds of different enzyme systems and receptor types and subtypes act in concert (Ágoston et al., 2005). Recent research into cellular networks demonstrates that pharmacological agents that provide multiple signals can impact the complex equilibrium of whole cellular networks more favorably than drugs that act on a single target (Ágoston et al., 2005; Briskin, 2000; Csermely, 2004; Csermely et al., 2005; Keith et al., 2005; Keith and Zimmermann, 2004; Morphy et al., 2004; Werner, 2003). Additionally, multi-target agents need affect their targets only partially, which corresponds well with the presumed low-affinity interactions of medicinal plants (Ágoston et al., 2005).

    • The promise of a virtual lab in drug discovery

      2006, Drug Discovery Today
      Citation Excerpt :

      Nevertheless, industrial drug discovery data are usually of high quality, because of ample resources and laboratory standardization, good regulation, a well-defined pipeline and a distinct final goal, all of which form an extremely good basis for application of an omics approach in drug discovery. Recent literature about omics' usefulness in drug discovery provides a wide range of potential solutions, from outsourcing to adopting a systems or integrative biology approach [19–26]. However, although none of these solutions will provide the ultimate answer to today's drug discovery problems [27], they all share elements of collaborated use of distributed resources, reuse of data and methods and computer support.

    • Systems-ADME/Tox: Resources and network approaches

      2006, Journal of Pharmacological and Toxicological Methods
    View all citing articles on Scopus
    View full text