Over the years, we have learned that drugs act in very complex ways and cause a combination of wanted and unwanted effects, many—if not most—of which are still poorly understood. Attaining a deeper knowledge of how drugs act in the body and their connections to therapeutic and toxicological outcomes now requires a systems-level approach.
At this time, NIGMS has a substantial grant investment in pharmacology and in systems biology, but we have not seen a great deal of activity integrating pharmacology with systems biology to benefit drug discovery and the understanding of drug action.
With this in mind, NIGMS hosted the first Quantitative and Systems Pharmacology Workshop last September. We brought together researchers from systems biology, pharmacology and pharmacokinetic/pharmacodynamic modeling to figure out how these fields can come together to advance drug design and discovery. The group addressed the topic from the standpoint of both horizontal integration (various networks in various cell systems) and vertical integration (connections between pathways at different levels of organization, tissues, organs, etc.). You can read the workshop report for a summary of the discussions.
An important outcome from the meeting was the feedback we got from participants about how far apart their disciplines presently are yet how much they have in common. The participants encouraged us to create more opportunities for them to interact and help bridge their disciplines.
To help promote and facilitate these interactions, we are now planning our second Quantitative and Systems Pharmacology meeting for fall 2010. We want your input to help shape the program. What do you think are the cutting-edge topics? What are the biggest challenges? What advances are needed to develop a systems approach to therapeutics?
You can comment here, or send an e-mail to me or any of the other meeting organizers. They include Sarah Dunsmore, Richard Okita and Peter Lyster from NIGMS and Grace Peng from the National Institute of Biomedical Imaging and Bioengineering.
Mike, this is a great and timely effort. Our Metabolomics Network for Drug Response Phenotype would be happy to participate and share experiences on how metabolomics, the global science of biochemistry, enables mapping of changes in networks and pathways upon drug treatment. Studies with important classes of therapies such as statins and SSRIs demonstrate that effects of these drugs on metabolism is comprehensive and includes changes in many pathways, some related to therapeutic benefit and others to side effects. Metabolomics data combined with genetic data is allowing us to start to link genotype to phenotype and mapping of pathways implicated in variations in response to therapy. Additionally, metabolomics data is highlighting an important role of the microbiome and how we respond to therapy. Please let us know how we can help as you start to plan for this meeting. Best wishes, Rima
Mike – I also agree that this is a very important effort. We in the CDP Center (cdpcenter.org) continue to pursue methods to measure and model drug-cell interactions. For example, our recent analysis of TRAIL demonstrates a significant role for non-genetic cell-to-cell heterogeneity in determining killing fraction (Spencer, Sorger et al, Nature 2009 – PMID: 19363473). We are expanding these studies to small molecule therapeutics targeting ErbB receptors and to intravital analysis of signaling in mice. We look forward to working much more closely with the pharmacology and drug discovery communities you have identified. We will post links to some recent reviews and papers as the come out.
Mike,
Ultimately the path to translational research from systems biology is systems pathology and pharmacology. So this is an excellent idea. Peter Lyster and I brainstormed last year about this and came up with a conceptual framework. You may already have this, but thought it is worth posting to stimulate ideas for next meeting.
The goal of systems pharmacology is to identify:
(1) What is difference between normal physiology and pathology, and what are the systems that are involved?
(2) What are best targets at the “systems level” for pharmacological intervention?
(3) What is the normal function that can be reestablished following therapy (includes metabolism)?
(4) What is a good combinatorial therapy (drug cocktail) that is likely to address multiple facets of the pathology?
For e.g. in the case of, say, insulin resistance we want to identify:
(1) What networks and physiology are altered; not just one molecule, but downstream to metabolism? Is there an independent well-defined processes? Can we divide the process into subsystems? Invite pathophysiology speakers.
(2) What are points of intervention? What perturbations of systems are optimal to achieve the change to non-pathological phenotype? Then to reverse the question, look at current drugs or drug attempts and ask why the do or don’t work.
(3) When a drug works, what is minimal physiology and metabolism in terms of phenotypic effect, i.e., what are drug pathways ?
(4) What is the rational process for obtaining non-interfering therapeutic target combinations so that drug cocktails can be designed?
Where do we have understanding for how to address these issues? What is needed is:
(1) High throughput genotype and phenotype measurements in normal and disease systems (includes polymorphism data, genotype and phenotype data,…)
(2) What are the known instances where there has been drug analysis of interventions at level of systems?
(3) Can we analyze and make use of drug success and failures at the systems level?
To put the analysis on firm footing, and elaborate on what we mean by ‘horizontal’ and ‘vertical’ integration, we gather examples that are amenable to systems analysis, and where there is lot known in terms of pathways. This will constitute the first draft of the ‘state of play’:
(1) Insulin resistance diabetes, obesity, and atherosclerosis, e.g., there are drugs such as thiazolidinediones (TZD) used for insulin-based approaches.
(2) Hypertension; there are good drugs affecting hypertensive pathways and good therapeutics, e.g., diuretics.
(3) Cancer, e.g., breast and ovarian cancer, e.g. tamoxifen.
(4) Inflammation, e.g., there are anti-inflammatory drugs, e.g. cox2 inhibitors
We will address the issue of what clinical measurements can be used in diagnosis and treatment. This should extend beyond, say, urine and blood assays, e.g., we need tissues and tissue-specific mapping of cell-tissue-organ phenotype. We should be able to identify markers (progression of disease) and the progression to downstream metabolite. Examples for the above list are:
(1) For diabetes we can sample adipose tissue and muscle tissue.
(2) For hypertension, we can measure EEG, blood pressure, hematocrit, red blood cell count, and oxygen levels.
(3) For cancer we may acquire tissue/cell biopsy.
(4) For inflammation we can obtain macrophages and measure cytokines/chemokines.
Of course, more difficult cases are not included in this list but should be noted for completeness, e.g., liver which is difficult to sample, migraine which is difficult to measure and quantitate, and schizophrenia which is extremely complicated. We should try to form a complete list and explain why they represent difficult cases.
The following is a list of micro-level questions which arise out of the coarse-grained discussion above. We need to ask and hopefully begin to answer:
(1) What do we know about biochemical networks and disease? What can we understand about individual tissue from high throughput studies?
(2) How much is normal physiology for a given disease different from patholophysiology from the network point of view?
(3) How much variability is there in a disease based on origin, onset, progression, response to drug(s)?
(4) What are current treatments and how do they influence pathways that are highlighted in question (1), i.e., what do they address and what do they not address?
(5) What are good drug targets? Why are they good? How do you identify good metrics to identify drug targets? Can we come up with good rules to say this identifies good drug target?
(6) What is the relevance and impact of metabolism and downstream effects? How would you deal with toxicity?
(7) What are physiological pathways in drug targeting and delivery to targeted tissue/organ/cell, e.g., PK/PD questions?
(8) What are technologies necessarily for Multiscale genotyping and phenotyping at systems level (microfluidics, computational data analysis, …), and what technologies are missing for systems level disease level analysis and pharmacology?
(9) What are requirements for data standards, e.g., data format, that enable appropriate data acquisition, diagnostics, and research?
[POSSIBLE TOPICS TO COVER IN SESSIONS ]
Quantitative and Systems Pharmacology
1. Cellular pathways in normal and pathological function
a. Insulin resistance and pathways
b. Cancer pathways
c. Metabolic disorders
2. Pharmacological perturbations of pathways
a. Statins and cholesterol pathways
b. ACE inhibitor pathways
c. Hypertension drugs and interventions
3. Quantitative models of pathways and pharmacological and genetic interventions
a. GPCR/Calcium pathway modeling
b. Kinase pathways
c. PharmGKB models
4. Drug Metabolism and efficacy
a. Transport in the kidney
b. P450 metabolism of drugs
c. Modeling toxicity
Sorry, it is rather long!
Shankar