Evolving technologies for gene expression
analysis offer a compromise between high cost microarrays restrictive
single-endpoint screens.
| Sep
1, 2005 |
| By:
Joseph
Monforte |
| Pharmaceutical
Discovery |
|
Joseph Monforte, PhD
|
The rate at which the genetics of complex diseases are being deciphered
has been accelerating at an astounding pace. Using the latest genomic
tools, researchers have begun to characterize in great detail multiple
genetic mechanisms that give rise to such diseases as cancer,
cardiovascular diseases, neurodegenerative disorders and diabetes, as well
as numerous functional pathways associated with disease response such as
stress and immune response, cell cycle regulation, cell differentiation,
cell proliferation and cell death. This exponential growth in our
knowledgebase of disease genetics has led to the identification of a large
array of genes, proteins and pathways that potentially play a central role
in disease development and may be potential targets for therapeutic
intervention.
Experimentally delving deeper into how these genes may function
pathologically and protectively in disease is an area of intense study in
the research community. More challenging is figuring out the most
efficient means for translating these discoveries into well-characterized
compounds and compound classes that can influence these genes and, as a
consequence, offer the potential for new therapeutic classes.
Cost Limitation
Steve Casey
|
Many of the newest analytical methods, such as multiplexed gene expression
analysis using DNA microarrays, have not been applied to compound
screening and discovery. A major reason for this roadblock is cost. DNA
microarrays continue to cost hundreds of dollars per sample to run, while
the high-throughput screening strategy, involving brute force scans
through hundreds of thousands or even millions of compounds, can only be
done if the cost of a single assay runs around a few cents. Because of
this limitation, the predominant screening assay formats used today fall
into two categories: target-specific and phenotypic.
Target-specific screens, such as protein binding, enzyme activity, and
reporter gene assays focus on capturing the effects of a given compound on
a single gene or protein endpoint, while phenotypic screens typically
capture gross cellular changes such as viability, apoptosis, proliferation
or ion flux. Both of these screening approaches have significant value,
but they fall short when it comes to assessing the effectiveness of new
compounds targeting complex, multigenic diseases like cancer and diabetes.
Target-specific screens are too focused and cannot observe multigenic
responses to perturbations. Cell-based phenotypic screens are too broad
and cannot be used to differentiate the multiple pathways that can be
altered to produce a phenotypic response nor can they effectively be used
to optimize and direct compound development toward specific mechanisms of
action.
Quantitative Multiplexed PCR
Fortunately, newer methods for gene expression analysis are evolving
that offer a compromise between the high cost of microarray-based gene
expression analysis and single endpoint screens. As an example,
quantitative multiplexed PCR techniques such as the Beckman GeXP system
offer the ability to analyze 20-30 genes at a relatively modest cost of
running a single PCR reaction. Other multiplexed processes based on flow
cytometry or using low density microarrays also offer some promise for
achieving lower cost.
In addition to a high level of multiplexing, the PCR-based approach has
the added advantage of flexibility. Because of the simple nature of
nucleic acids it is relatively straightforward to use a particular gene
expression analysis method to analyze virtually any gene. This is a
central tenet to the creation and use of microarrays. And unlike
protein-targeted and phenotypic screens, which can take upwards of 6
months to develop, the simple polymeric nature of mRNA means that assay
development can be very fast, e.g. a few days, and that the methods
established to develop and run one particular assay can be readily and
simply applied to any other gene-expression-based screen as needed. The
ultimate value of such PCR-based screens may not be in the primary,
high-throughput screen but as a method for quickly and cost-effectively
following up in the characterization of any hits from these screens for a
variety of responses related to efficacy, potency and toxicity.
To effectively select and prioritize
lead compounds with improved odds of utility, one must characterize
compounds for both their potency in affecting a desired target (or, more
frequently these days, multiple targets) and also their secondary,
off-target activities, such as toxicity. Early recognition of unwanted
side-effects of pharmacologically active substances is vital to avoiding
costly failures in the clinical arena. For example, liver damage is still
the leading reason for withdrawing drugs from the market and for
abandoning lead candidates during development [1-3]. Unfortunately, a full
assessment of all lead compounds using current toxicological methods is
impractical because of the high costs of such tests. And other problems
complicate matters, such as the long time required to conduct studies,
high compound requirements and the relevance of animal findings to humans.
Gene Sets
Interestingly, classical toxicology
and high throughput screening suffer similar limitations: the primary
tests, both in vitro and in vivo, generally cannot assess the effects of
compounds at the multigenic, gene response level. The tests are either too
focused (e.g. the Ames test for DNA mutation) or too broad (e.g.
gross animal tissue pathology). To address this deficit, researchers in
the field of toxicogenomics have made significant progress in applying
gene expression methods to develop sets of predictive gene markers for a
variety of toxicological endpoints. For example, Thomas et.al. [4]
used microarray studies to identify 12 key transcripts out of 1200 that
can predictively track 5 major toxicological responses. While Burczynski et.al.
[5] identified a 20 gene set for quantifying cytotoxicity and DNA damage
in HepG2 cells, and Amin, et.al [6] identified a 29-gene set
indicative of renal toxicity. In fact, dozens of genes have been
identified that are able to track the effects of compounds in terms of
triggering specific responses, such as metabolic induction, oxidative
stress, DNA damage response, proliferation and the like.
As can be seen in the work in
toxicogenomics, the assessment of specific compound activities can be
tracked with relatively modest sets of genes. This is a common theme in
gene expression analysis. While it is important to observe more than one
gene, there are no 30,000-gene diseases, nor are there compounds that
affect 30,000 genes. It is now often possible through preliminary
microarray studies and literature review to quickly develop a set of
genes, often 50 or fewer, that can be used to characterize a particular
set of candidate compounds for a variety of gene and pathway-specific
responses, both specific to the desired target pathways and undesirable
toxicological endpoints. Moreover, these gene expression responses are
proving to be better predictors than current methods for how the different
compounds will behave as they are moved from in vitro models to in vivo
studies and ultimately on to humans.
Joseph Monforte, PhD is Vice
President and Chief Scientific Officer at Althea Technologies, Inc.,
11040, Roselle St., San Diego, CA 92121, jmonforte@altheatech.com
Please address all correspondence to Steve Casey, founder and COO
of Expression Analysis Inc., a provider of regulatory compliant genomic
processing services, who can be reached at scasey@expressionanalysis.com
References
1) Chitturi, S. and George, J. (2002)
Hepatotoxicity of commonly used drugs: nonsteroidal anti-inflammatory
drugs, antihypertensives, antidiabetic agents, anticonvulsants,
lipid-lowering agents, psychotropic drugs. Semin Liver Dis, 22:
169-183.
2) Farrell, G. C. and Liddle, C.
(2002) Drugs and the liver updated. Semin Liver Dis, 22:
109-113.
3) Lasser, K. E., et.al. (2002)
Timing of new black box warnings and withdrawals for prescription
medications. Journal of the American Medical Association, 287:
2215-2220.
4) Thomas, R.S., et.al. (2001)
Identification of Toxicologically Predictive Gene Sets Using cDNA
Microarrays. Molecular Pharmacology 60: 1189-1194
5) Burczynski, M.E. et.al. (2000)
Toxicogenomics-Based Discrimination of Toxic Mechanism in HepG2 Human
Hepatoma Cells, Tox. Sci. 58,399-415.
6) Amin, R.P. et.al. (2004)
Identification of putative gene based markers of renal toxicity, Environ. Health
Perspect. 112(4), 465-79
|
