**1. Introduction**

*"I believe there is little reason to question the presence of innate systems that are able to restructure a genome. It is now necessary to learn of these systems and to determine why many of them are quiescent and remain so over very long periods of time only to be triggered into action by forms of stress, the consequences of which vary according to the nature of the challenge to be met". Barbara Mc Clintock, (1978), as cited in (Jorgensen, 2004).* 

Attempts to link specific neurobehavioural phenotypes with causative genes, a necessary step preceding the development of highly specific neuroimaging biomarkers for these phenotypes, have been spectacularly unsuccessful, although numerous significant gene candidates emerged during the process. Various family study approaches, including twin studies, sib pairs, 'trios' and so-called 'pure multiplex pedigrees' excluding comorbid disorders have been employed in various linkage and association study designs to determine the genes and proteins underlying the relevant disorders. Some genomewide association studies surprised investigators with the information that, when initially promising genomic coding region hotspots were supersaturated with markers at increasingly closer map distances, the results showed weakening of signals in targeted exonic areas, possibly indicating an influence outside these coding areas. The importance of epigenetics (including increasingly complex regulatory region and RNA metabolism gene expression-modifying mechanisms), gradually emerged during the course of molecular genetic 'brain and behaviour' studies. "Epigenetics is defined as mitotically or meiotically stable molecular processes that regulate genome activity independent of DNA sequence. The term 'heritable' has been included in the definition, but has been omitted recently since this implies generational inheritance by definition and therefore does not include all elements of epigenetics (Skinner, 2011). The brain represents a particular area of interest with regard to epigenetics, where it has been demonstrated that epigenetic modifications are not static, but dynamically change in response to external stimuli including synaptic activity (Crepaldi & Riccio, 2009). The increasing likelihood of the role played by transgenerational epigenetic influences in neurogenetics studies now also impacts on the tracking of neuropsychiatric disorder gene candidates in families, implying a requirement for analyses of both coding as well as noncoding polymorphisms and the manner in which these interact,

Stress Shaping Brains: Higher Order DNA/Chromosome

Campayo et al., 2009).

more comprehensive 'interactome'.

enabling both mass screening and targeted imaging.

**3. Epigenetics and the brain – foetal programming** 

strong emerging research theme (Nelson & Monteggia, 2011).

Mechanisms Underlying Epigenetic Programming of the Brain Transcriptome 351

better definition and stratification of preclinical study groups, and for providing direct biological measures of response (Waerzeggers et al., 2010). Not all biomarker molecules are suitable for radioimaging approaches, though many potentially interesting molecular biomarkers also appear to be suitable molecular imaging agents. Activatable molecular probes are designed to elicit a detectable change in signal upon enzymatic activity or in response to specific biomolecular interactions. In many cases, these unique characteristics allow for very high signal-to-background ratios compared with conventional targeted contrast agents and they open up the possibility of imaging intracellular targets (Garcia-

Molecular biology now offers system-wide insights which have to be incorporated in the derivation of appropriate imaging biomarkers: **'***Omics'* is a general term for a broad discipline of science for analysing interactions of biological information. These include studies of the genome, fragilome, transcriptome, proteome, metabolome, expressome, interactome etc. The main focus in these endeavours is on mapping information objects such as genes, proteins, and ligands, finding interaction relationships among the objects and engineering the networks to understand and manipulate the regulatory mechanisms. *Systems biology* integrates information from the various 'omics' subfields, and generates a

Combined with the disciplines of molecular imaging and molecular medicine, systems biology approaches to understanding disease complexity promises to provide predictive, preventive and personalized medicine that are expected by many to be able to transform healthcare in the future. Continued development of these technologies and applications requires collaboration transcending traditional boundaries between disciplines – e.g. for suitable biomarker consideration, clinical molecular geneticists and radiochemists have to liaise to decide on the choice and molecular characteristics of biomarkers suitable for both laboratory assays as well as being able to bind radioligands. Using disorder specific biomarkers for both laboratory assays and radioimaging allows a two-pronged approach

The observed foetal basis of some adult onset diseases requires both epigenetic and genetic factors to be involved in regulating developmental biology outcomes, as emphasized by Skinner (2011), who cites the now classical example of insulin resistance and obesity. Developmental studies of metabolic 'programming' suggest that insulin resistance may appear during early development in individuals born small for gestational age. Insulin resistance can promote obesity, which in turn, could sustain the state of insulin resistance in later life. Skinner et al. (2011) stresses that "the current paradigm of DNA mutational events promoting evolution is accurate, but the inclusion of epigenetics allows for a much higher degree of variability in the biological system to facilitate an adaptation event and epigenetic transgenerational inheritance is a novel concept with considerable experimental support in plant and mammalian studies. This insight, therefore, does not modify the fundamental Darwinian evolutionary paradigm, but adds a neo-Lamarckian component allowing a more diverse molecular mechanism" (Skinner 2011). The role of epigenetics in controlling neuronal functions that may ultimately underlie behavioural adaptations represents a

as well as distinguishing between context dependent and germline dependent epigenetic changes (Crews, 2008).

The extensive comorbidity recorded in neuropsychiatric disorders and the acknowledgement of their role when devising candidate gene approaches has been a major challenge. Even though the importance of epistasis and gene pleiotropy are generally acknowledged, there is still no clear understanding of the mechanisms of comorbidity, and biomarker investigative choices thus often remain too simplistic.

The increasing application of sophisticated network dynamics analyses may provide the means to resolve these issues, and the problems posed by variable overlapping of disparate conditions in complex syndromes may soon be better understood by a 'diseasome' approach (Potkin et al., 2010; Barabási et al., 2011), as outlined below. While seemingly introducing even more variables, the aim of network medicine concepts is to actually reduce the noise arising from experiments producing vast amounts of data and organise the information existing in published and newly executed research, ranging from studies of single markers to massively parallel array sequencing experiments. Applying genetic data from genome-wide association studies in a gene network analytic approach, using brain imaging as a quantitative trait phenotype, can increase the statistical power to identify the molecular pathways in which risk genes participate (Potkin et al., 2010). These authors demonstrated the utility of a regulatory network approach by measuring correlations among transcript levels in the mouse and human postmortem tissue which allowed the derivation of an enriched gene set that identified several microRNA's that could be associated with negative symptom schizophrenia. Last but not least, there may be a belated resurgence of interest in the value of 'clinical experience' when discussing which phenotypes to analyse and how to identify the most appropriate research subjects by using experienced investigators within a specific field, rather than relying on massive amounts of data (and biospecimens) accrued by means of research questionnaires dealt with by inexperienced but willing junior research assistants.

This review deals only with the first step in the development of radioligands for imaging in neuropathology (i.e. understanding the clinical genetic context before the selection of disorder-relevant biomarkers). Once a suitable molecular biomarker candidate can be demonstrated, this needs to be followed by selection of leading compounds, radionuclide, labeled position, and synthesis methods; in vitro and in vivo evaluation including probability for imaging, selectivity, specificity, and species differences, and finally an evaluation of factors impinging on safety such as acute toxicity, mutagenicity and radiation dosimetry (Ishiwata, 2009). This does not mean that simpler biomarker identification approaches may not be effective, but it is proposed that a total understanding of the full scope of specific neurobehavioural problems and the mechanisms according to which they overlap can only be developed in this manner. An attempt will be made to integrate the fundamentals of an interface system in the brain with an evolutionary understanding of the mechanism that could give rise to a cluster of seemingly unrelated disorders. This approach may be useful as a model according to which some currently important facets regarding comorbidity can be placed in context.

#### **2. Combining in vivo neuroimaging technologies with 'omics' biomarkers**

Noninvasiveness of molecular imaging offers a potent advantage for monitoring endpoints of molecular medicine interventions. In particular, pharmaceutically-relevant neuroimaging endpoints based on disorder-specific biomarkers have great potential for

as well as distinguishing between context dependent and germline dependent epigenetic

The extensive comorbidity recorded in neuropsychiatric disorders and the acknowledgement of their role when devising candidate gene approaches has been a major challenge. Even though the importance of epistasis and gene pleiotropy are generally acknowledged, there is still no clear understanding of the mechanisms of comorbidity, and

The increasing application of sophisticated network dynamics analyses may provide the means to resolve these issues, and the problems posed by variable overlapping of disparate conditions in complex syndromes may soon be better understood by a 'diseasome' approach (Potkin et al., 2010; Barabási et al., 2011), as outlined below. While seemingly introducing even more variables, the aim of network medicine concepts is to actually reduce the noise arising from experiments producing vast amounts of data and organise the information existing in published and newly executed research, ranging from studies of single markers to massively parallel array sequencing experiments. Applying genetic data from genome-wide association studies in a gene network analytic approach, using brain imaging as a quantitative trait phenotype, can increase the statistical power to identify the molecular pathways in which risk genes participate (Potkin et al., 2010). These authors demonstrated the utility of a regulatory network approach by measuring correlations among transcript levels in the mouse and human postmortem tissue which allowed the derivation of an enriched gene set that identified several microRNA's that could be associated with negative symptom schizophrenia. Last but not least, there may be a belated resurgence of interest in the value of 'clinical experience' when discussing which phenotypes to analyse and how to identify the most appropriate research subjects by using experienced investigators within a specific field, rather than relying on massive amounts of data (and biospecimens) accrued by means of research questionnaires

This review deals only with the first step in the development of radioligands for imaging in neuropathology (i.e. understanding the clinical genetic context before the selection of disorder-relevant biomarkers). Once a suitable molecular biomarker candidate can be demonstrated, this needs to be followed by selection of leading compounds, radionuclide, labeled position, and synthesis methods; in vitro and in vivo evaluation including probability for imaging, selectivity, specificity, and species differences, and finally an evaluation of factors impinging on safety such as acute toxicity, mutagenicity and radiation dosimetry (Ishiwata, 2009). This does not mean that simpler biomarker identification approaches may not be effective, but it is proposed that a total understanding of the full scope of specific neurobehavioural problems and the mechanisms according to which they overlap can only be developed in this manner. An attempt will be made to integrate the fundamentals of an interface system in the brain with an evolutionary understanding of the mechanism that could give rise to a cluster of seemingly unrelated disorders. This approach may be useful as a model according to which some currently important facets regarding

**2. Combining in vivo neuroimaging technologies with 'omics' biomarkers** 

Noninvasiveness of molecular imaging offers a potent advantage for monitoring endpoints of molecular medicine interventions. In particular, pharmaceutically-relevant neuroimaging endpoints based on disorder-specific biomarkers have great potential for

biomarker investigative choices thus often remain too simplistic.

dealt with by inexperienced but willing junior research assistants.

comorbidity can be placed in context.

changes (Crews, 2008).

better definition and stratification of preclinical study groups, and for providing direct biological measures of response (Waerzeggers et al., 2010). Not all biomarker molecules are suitable for radioimaging approaches, though many potentially interesting molecular biomarkers also appear to be suitable molecular imaging agents. Activatable molecular probes are designed to elicit a detectable change in signal upon enzymatic activity or in response to specific biomolecular interactions. In many cases, these unique characteristics allow for very high signal-to-background ratios compared with conventional targeted contrast agents and they open up the possibility of imaging intracellular targets (Garcia-Campayo et al., 2009).

Molecular biology now offers system-wide insights which have to be incorporated in the derivation of appropriate imaging biomarkers: **'***Omics'* is a general term for a broad discipline of science for analysing interactions of biological information. These include studies of the genome, fragilome, transcriptome, proteome, metabolome, expressome, interactome etc. The main focus in these endeavours is on mapping information objects such as genes, proteins, and ligands, finding interaction relationships among the objects and engineering the networks to understand and manipulate the regulatory mechanisms. *Systems biology* integrates information from the various 'omics' subfields, and generates a more comprehensive 'interactome'.

Combined with the disciplines of molecular imaging and molecular medicine, systems biology approaches to understanding disease complexity promises to provide predictive, preventive and personalized medicine that are expected by many to be able to transform healthcare in the future. Continued development of these technologies and applications requires collaboration transcending traditional boundaries between disciplines – e.g. for suitable biomarker consideration, clinical molecular geneticists and radiochemists have to liaise to decide on the choice and molecular characteristics of biomarkers suitable for both laboratory assays as well as being able to bind radioligands. Using disorder specific biomarkers for both laboratory assays and radioimaging allows a two-pronged approach enabling both mass screening and targeted imaging.
