The
Potential of Proteomics
Proteomics is a relatively new field in biology,
having only had its very name coined in 1994. However,
as its name suggests, proteomics shares some common
characteristics with an earlier project now well
known as genomics, the study and mapping of
the human genome, as well as the subsequent
application of the patterns thereby established.
Proteomics attempts to do much the same thing, but
with proteins rather than DNA. Indeed, the
terminology of genomics carries over to analogues in
proteomics; for example, whereas a genome is a
map of the genetic identity of an organism, a proteome
is a map of the protein identities of an
organism’s diverse cells and tissues. Moreover,
some of the key techniques, used to categorize and
study the vast amounts of information in genomics,
are finding new, analogous applications in processing
the still vaster volume of information indigenous to
proteomics.
So what is proteomics? What does it do? What
purpose does it serve? Specific definitions
vary, but in general distilled form, proteomics might
be described as the identification of all the
proteins manufactured by the cells in each human
tissue, the determination of each protein’s
precise three-dimensional molecular structure, and
the study of how the various proteins function and
interact. But considering that proteins
exhibit characteristic differences, not only from
person to person (as do genes), but even from tissue
to tissue in the same individual, we might well ask
why we should want to undertake such a colossal task.
Aside from natural curiosity, the reasons range from
the lofty—innovation in the treatment of
diseases and disorders—to the mundane—plain
old dollars-and-cents. By identifying and
targeting specific key proteins, proteomics offers a
way simultaneously to fine-tune and to accelerate the
development, testing, and approval of new drugs,
featuring greater effectiveness and fewer
side-effects, while at the same time keeping costs
tolerable.
With the continuing escalation in drug development
costs, recovery from failures, and efforts to recoup
costs before patents expire, the industry has come to
view the prohibitive pricing of new drugs as the only
way to maintain its profitable attractiveness to
investors. Indeed, the situation has evolved as
a dual crisis, with the health of patients pitted
against the health of the pharmaceutical industry.
As Jessica Cutter of Morgan Stanley observes in a Scientific
American article, “Pharmaceutical companies
are dependent on proteomics and like technologies to
overhaul their entire drug development process, or
they will not survive (Ezzell, 42).” Proteomics
is seen as a promising way to ameliorate this
unfortunate state of affairs, if not to overcome it
entirely.
Proteomics does not furnish us a free ride, however.
For one thing, the enormous job of cataloguing the
human proteome has been characterized as dwarfing the
earlier genome project—which took several years
to complete. Indeed, the expression,
“Genes were easy!” has become a cliché in
the field. The magnitude of the proteome
project becomes apparent only when we understand that
the number of distinct proteins generated by each
cell is many times the number of genes in the
entire genome. Even so, some researchers have
expressed hope of having the human proteome mapped in
a relatively short period of three years, in part
with the help of equipment and techniques adapted
from the genome project and other fields.
Besides the sheer magnitude of the task, other
obstacles include the difficulty of some processes
and the cost and availability of necessary equipment.
Proteomic analysis relies heavily upon a difficult
and delicate process known as two-dimensional gel
electrophoresis. To make this
slow-throughput 2-D gel process both speedier and
more reliable, and to accommodate the vast volume of
material and data, at least one research facility is
making use of high-precision robotic technology
adapted from the automotive industry. An
alternative (and in some cases an adjunct) is mass
spectrometry; however, this requires a machine with a
price tag of half a million dollars, which even then
has a tendency to miss certain types of proteins.
But despite its drawbacks, mass spectrometry is so
far responsible for the lion’s share of raw data
for proteomics. In just one example, according
to Bio-IT World, “Caprion's eight
instruments run around the clock, and generate 60GB
of data per machine, per hour (Branca).” Yet
another crucial tool of proteomics is x-ray
crystallography, used to determine the
three-dimensional configurations of protein
molecules. Until recently, this process
required access to dedicated x-ray beams, available
only at a few research sites equipped with huge and
costly synchrotrons. However, recently
developed x-ray lasers now enable the process to be
performed in ordinary laboratories. Now, to
gather and organize all of this information—let
alone track the myriad interactions—would have
been a hopelessly overwhelming task just a few years
ago. It demands immense computer resources and
comprehensive routines, capable of evaluating and
integrating multi-dimensional arrays of data. To
this end, highly effective microarray software
employed in the genomics project is being enhanced
and adapted to the exponentially more complex
requirements of proteomics.
Once this mountain of information has been neatly
organized, catalogued, and cross-referenced on
multiple levels, how can we use it? One of the
primary goals is to identify and target proteins
specifically characteristic of foreign agents or
disordered cells, turning them either “off”
(as to terminate growth of cancerous tumors) or
“on” (as to restore pancreatic function in
diabetics). In this way, a proteomically
tailored drug can home in on specific cells or
agents, without affecting healthy cells or other body
systems. This would represent a gigantic stride
forward from radiative and chemical therapy
techniques of today, which tend to produce severe
“collateral damage” to neighboring healthy
tissue or even throughout the body. Indeed, it
is conceivable that protein-targeted drugs might be
developed for the non-invasive treatment of many
systemic problems, such as digestive, endocrine, or
neural disorders.
Of course, treatment is only one side of the coin.
While traditional “shotgun” treatment of a
general class of agents and disorders often produces
tolerable results—usually with some side
effects—the precise identification of a problem
permits more effective treatment with less guesswork.
In addition to its therapeutic rewards, proteomics
offers the opportunity to diagnose existing disorders
precisely. Yet even this is not the whole of
it: Through general screening, proteomics might
even allow us to identify potential problems
in individuals—susceptibility to breast cancer
or renal disorder, for example—years in advance,
thus allowing early, preventive treatment of such
potential conditions. Indeed, the remarkably
precise diagnostic potential of proteomics has
already been demonstrated. According to a Scientific
American article (citing results of a study
published in Lancet), “…in 116
samples, that protein “fingerprint” picked
out every woman with ovarian cancer, including 18
early cases, and designated 63 out of 66 healthy
women as disease-free (McCook, 16).”
Questions arise, though, about whether it is
desirable to pre-diagnose and treat possible
problems, many of which may never actually manifest
themselves. Should we go to the effort and
expense of “immunizing” people against all
sorts of potential disorders, many of which will
never become actual problems? Or if not, is it
ethical to burden people with the knowledge that they
have, say, a slight tendency toward Alzheimer’s
disease, thus perhaps inflicting upon them and their
loved ones inordinate long-term anxiety—indeed,
the attendant disorders produced by anxiety itself?
Would we simply be trading one set of problems for
another?
The “right” answers to such questions may
vary from individual to individual. Many, no
doubt, would welcome the chance to ward off, in
advance, as many potential health problems as
possible, freeing them to address life’s more
pleasant or stimulating challenges with less
distraction. Yet there are others who are
averse to “knowing too much” about their
own futures, and who would prefer to play life’s
cards as they are dealt. Another question, one
that hits us squarely in the pocket book, is whether
medical insurance should pay for prognostication and
(possibly unneeded) preventive treatment. The
answer to this may well distill from practical
experience, of whether it turns out to be most
cost-effective to provide blanket preventive
treatment, selective prevention for high-risk cases,
or case-by-case corrective care. These are
answers we simply cannot know until we have all the
necessary facts. Nevertheless, as in other
fields of scientific inquiry dealing with fundamental
issues, it is important that we proceed with
research, so that whatever decisions we must make can
be informed.
A related
issue is that of exploiting short-term advances
during the course of the long-term project—a
bread-and-butter aspect somewhat overlooked during
the genome project, to the dismay of some corporate
boards. In contrast, in proteomics there is a
conscious effort to allow participating companies and
institutions to make profitable use of information as
it is obtained and processed, so that the project can
sustain and support itself to some extent, even as it
runs its course.