This essay addresses inherent problems within most
pharmacological and biological assays that all scientists should know. A basic scientific
principle is that scientific theories are overturned by experimental evidence
that doesn’t support the current theory, but this is only true if the
experiments are not only accurate and repeatable, but also accurately represent
the natural reactions that happen within the whole animal or organism. This is
especially important in the biological and pharmacological sciences, because
there are numerous variables that influence the experimental outcomes that
receive little or no mention in the mainstream scientific literature. In order
to develop new drugs, we need to increase our awareness of the critical
variables that are seldom mentioned in most experimental assays. Many of these
factors may be why assays are often difficult or impossible to reproduce. When
we look at the bigger picture, it is a wonder that we place so much trust in our
assays for the important roles that they play in everything from drug
development to medicine and research. To step back and see this bigger picture,
I will briefly discuss three general factors that often alter the outcomes of these
assays.
Three Factors Complicating Most
Pharmacological and Biological Assays
1) Confounding Effects:
An overlooked subject that greatly influences
pharmacological and biological assays is desensitization or substrate
inhibition. This is observed when higher doses of drugs, ligands or substrate
molecules show a decrease in the response of the receptor or activity of an
enzyme. Most receptor and enzymatic reactions display some negative feedback
such as desensitization or substrate inhibition. However, this often goes
unnoticed because many assays measure the cumulative response, which is the
area under the actual dose-response curve (or the area under the kinetic
reaction curve) for different drug, ligand or substrate concentrations. Assays
measuring a cumulative response mask the underlying desensitization or
substrate inhibition within these systems.
It is often, wrongly, thought that by keeping the drug or
substrate molecules at lower levels the effects of desensitization or substrate
inhibition will be reduced or prevented.
This isn’t accurate, because the phenomena of desensitization and
substrate inhibition is inherent within the chemical equilibria of these
systems. Many scientists and medical doctors aren’t aware of the ubiquity of
drug desensitization or enzyme substrate inhibition, because many experimental
assays aren’t capable of clearly demonstrating these phenomena. These issues
present a large problem in drug development and the proper pharmaceutical
treatment of patients.
Interpretation of assay results is perhaps another one of
the most confounding factors. Our interpretations are largely model dependent. We
tend to see what our models allow us to see. Models, such as curve fitting, often
create confounding factors that are very much dependent on the assumptions underlying
these models. Any two-dimensional curve can be fitted by an arbitrary polynomial
to arbitrarily high powers; however, this type of fitting does little to
elucidate the underlying biophysical mechanisms of these systems. Statistically
we can get a good fit with a bad model. Statistics alone cannot determine the
quality of our research. A good model should be like a good pair of glasses
that helps us see more clearly the underlying biophysical and physicochemical mechanisms
of our assays. Bad models blur our understanding and create confusion.
Current trends using extremely simplified simulations to model
and understand the reactions of enzymes or receptors may be hindering our
understanding of the underlying basic principles. It seems unlikely that such relatively
small differences of 2-5 kcal/mole in binding energies of drugs, ligands or
substrates will produce signal-specific wiggles against the background thermal
noise in these much larger molecules. Because of our limitations, our current
models cannot accommodate many of the other important molecules included within
these complex systems, such as the solvent, lipid, other proteins, counter-ions,
redox, energy, and cofactor molecules that are necessary components for the activation
of these systems in their natural state. Because these complex systems are simplified
and run in a vacuum, we’re stuck trying to make sense of rather meaningless
wiggles. These observations often serve to obfuscate rather than clarify our
understanding of the underlying biophysical principles.
2) Past Histories:
In general, assays can be divided into those using whole
organisms (in vivo) or isolated preparations (in vitro), but
there’s also the more general and less controlled variables of past histories
and current environmental conditions surrounding these assays, which applies to
both in vivo and in vitro assays. Past histories include
considerations such as how the animals were handled, caged, fed, type of
bedding, etc., or how the isolated preparation was prepared, such as type and
amounts of buffers used, fluctuations in temperatures (and the duration and
order of these fluctuations), osmotic pressures, pH, solvents; exposures to
atmospheric (oxidation) conditions at what stages of the preparation and for
what length of time, etc. How pure were our preparations, chemicals, buffers,
solvent(s), cofactors, etc? Did the type of containment vessels (glass/plastic)
alter the preparations?
Past exposures to various exogenous (xenobiotics,
pollutants) and endogenous (steroids, fatty acids) chemicals may induce
metabolic pathways such as the Cytochromes P450 enzyme systems (CPYs) that
handles many exogenous and endogenous molecules. These induced pathways can
profoundly alter our assays in ways that we don’t currently see or understand
fully.
These factors are usually not reported in the scientific
literature to the detailed extent that is necessarily suggested here. In some
cases, scientists have recognized these confounding facts and tried to account
for them, but in general we should all be aware of these serious problems in
interpreting any experimental results. Only by recognizing these problems and
the limitations that they place on our current assays can we make future
progress toward a better scientific understanding of the experimentally
observed responses of receptor and enzyme systems.
3) Assay Conditions:
In general, the more procedures that required to isolate any
biological sample, the more errors accumulate such that easy replication
becomes very difficult or impossible.
Very little or no attention is given to the REDOX
(Reduction-Oxidation reactions) environment of receptors or enzymes when
measuring their activities in vitro. In vitro, there is little or
no regard to the possible effects of light, high oxygen (including REDOX
status), electromagnetic gradients, etc. The processes of isolating receptor or
enzyme molecules often exposes them to oxidation conditions that generally go
unaccounted for the possible effects on their redox sensitive groups.
The requirements of cofactors for enzyme reactions has been
previously discovered, but there may be additional requirements such as the
requirement for an energy source from molecules such as ATP or GTP or membrane
energy gradients. Additional molecules may be necessary to sustain and
regenerate these energy molecules or gradients across membranes. There may also
be requirements for essential regional molecules such as gases (CO2, H2S, NO,
etc.) that may have far greater tissue concentrations in vivo than in our
in vitro assays.
Another problem that may seem simple, but is quite
complicated, is simply determining the pH-dependence of a receptor response, or
an enzymatic reaction. Often these reactions are done at various pHs to find
the optimum pH for that specific reaction under the conditions of specific
temperature, pressure, osmotic pressure, etc. To complicate matters, the
binding drug, ligand, or substrate molecules may have their own titratable
groups that are pH-dependent that differ from the receptor or enzyme molecules,
which often have multiple titratable groups that may also act to influence each
other. Other problems arise because the receptor or enzyme molecules are often
membrane bound in vivo whereas the assays are performed in vitro.
Other significant problems, which are very difficult to accommodate into in
vitro assays, is that biological membranes often separate regions with
different pHs, counter-ion concentrations, osmolarities, etc. Even the simpler
problems such as determining the proper buffer(s) to use, the concentration,
temperature correction(s), unwanted effects on other molecules, such as
solvent(s), cofactors, pH-detector(s), etc. are daunting.
In general, it is also very difficult to perform assays
under strictly in vivo conditions. The confounding problems with using the
whole organism entails many additional factors that affect the experimentally
measured responses, which include the pharmacokinetic factors such as the ADME
(Absorption, Distribution, Metabolism and Excretion). Each of the ADME factors have multiple
complexities that can confound experimental observations. Just considering the distribution
factor alone is often complicated by a drug, ligand, or substrate molecule
having to cross one or more membrane barriers, and by the differing tortuosity
of the route to tissue-embedded groups of receptor or enzyme molecules. These
barriers can greatly delay or alter the drug or substrate molecules from reaching
their target receptors or enzymes. The surrounding microenvironments often
determine the further metabolism and replenishment rates of the drug, ligand,
or substrate molecules, which are also dependent on the lipid compositions of
membranes as important considerations. The correct biochemical and biophysical
tensions across these membranes are also vitally important to ensure that these
assays accurately reflect their natural biological activities. The correct
ionic, osmotic, electrochemical, and pH gradients are the most obviously
important ones, but there are many others.
Assays that test experimental drugs for potential further
drug development are perhaps one of the most critical components of the drug
discovery process; yet they remain poorly characterized for this as well as for
other biological purposes. This is a very general description of several
problems with experimental assays that I’ve noticed over the four plus decades
of my career covering experimental, computational and theoretical approaches to
many scientific problems. Some of these problems may seem simple but continue
to remain largely marginalized and go unnoticed. They need to be recognized and
openly discussed so that further progress can be made. Other problems are much more complex than
current experimental techniques can handle, but knowledge of these problems may
spur improved assays or at least make us aware of the many problems inherent
within our current pharmacological and biological assays.
Richard G. Lanzara, MPH, Ph.D.
President and Principal Scientific Officer
Bio Balance, Inc.
