Leaping beyond the genome—what lies ahead?
July 2002 William B. Neaves, PhD
(Editor’s note: William B. Neaves, PhD, delivered this talk at
the Association for Pathology Informatics’ Clinical Information System/Life
Science Roundtable, which is generously sponsored by Cerner Corp.
Dr. Neaves is president and CEO of Stowers Institute for Medical Research
and professor, University of Missouri at Kansas City School of Medicine.)
Within a 12-month period at the end of the old millennium
and the beginning of the new, Science and Nature published
the first sequences of insect,1
plant,2 and human genomes.3,4
Although these papers do not represent finished work for the genome
sequences of any of the three species, they rank among the most
influential contributions to human knowledge. They ushered humankind
into the beginning of the Postgenome Era, a historic moment when
comparative genomics changed forever how people think about themselves
in relationship to other species.
The Human Genome Project showed that the instructions for assembling
the molecular machinery of the human body are encoded in approximately
30,000 to 40,000 distinct genes.3,4
Human beings share 10 percent of these genes with the fruit fly,
a model organism at the forefront of genetics research for the last
100 years. Already nearly 200 genes implicated in the etiology of
human disease have been found in the fly genome. These include genes
associated with numerous forms of cancer as well as Alzheimer’s
disease, Parkinson’s disease, cystic fibrosis, muscular dystrophy,
and fragile X syndrome.5
Prior to the sequencing of the Drosophila genome, few would
have expected an animal without kidneys to hold much interest for
nephrologists. Several human genes involved in renal disease have
counterparts in the fly genome.5
These genes produce proteins that play crucial roles in transporting
fluid and electrolytes across epithelia. Like humans, flies must
excrete metabolic byproducts to maintain homeostasis. They rely
on malpighian tubules rather than kidneys, but the movement of ions
across membranes is common to both systems. At the level of molecular
machinery, effective means of accomplishing a task are shared across
a broad diversity of animal species.
Physicians should be prepared for astounding insights into the
cause and prevention of human disease resulting from studies of
fly genetics. In addition to many of the fly’s own genes that are
homologues of human disease-causing genes, it is possible to insert
actual human genes into flies. For example, a humanized fly model
of neurodegeneration caused by an introduced human ataxia gene was
recently created.6 These flies
express the human gene responsible for causing Machado-Joseph disease,
also known as spinocerebellar ataxia type 3. This is the most common
dominantly inherited progressive ataxia in humans that results from
polyglutamine expansion. The abnormal protein produced by the mutant
gene has a greatly expanded polyglutamine domain, and it is prone
to form harmful aggregates in the cytoplasm of neurons. Neuronal
degeneration in these flies follows the formation of cytoplasmic
protein aggregates just as it does in the human disease.
This humanized fly model is being used to develop new therapies
that suppress the onset of the disease. One of the most promising
approaches thus far has been the introduction and overexpression
of a human heat-shock gene in these flies.7
This gene produces a protein that functions as a molecular chaperone
to prevent misfolding and aggregation of the mutant polyglutamine-enriched
protein of Machado-Joseph disease. Flies doomed by the introduction
of the human disease-causing gene are rescued by the introduction
and expression of a second human gene that produces a protective
protein. Flies will clearly play an important role in understanding
the genetic etiology of human disease and in exploring ways to rectify
disease phenotypes.
But if the relevance of fly genetics to human disease
was surprising, physicians faced an even more unlikely revelation.
In December 2000, the first sequenced plant genome was published,
and the mustard weed was found to have more than 25,000 genes, nearly
twice as many as the fly and almost as many as humans.2
Astoundingly, more than 100 mustard weed genes are homologues of
human genes that are involved in diseases such as xeroderma pigmentosum,
hyperinsulinism, Wilson’s disease, ataxia-telangiectasia, cystic
fibrosis, and breast cancer.2
Obviously, the mustard weed lacks lungs and breasts, but the same
genes that control fundamental processes such as ion movement across
cell membranes, repair of damaged DNA, and division of cells are
present in this plant as well as in us. When these genes are expressed
inappropriately or undergo deleterious mutations, diseases characteristic
of the host organism appear.
Perhaps the most significant similarities between the genomes
of mustard weeds and humans reside in the family of DNA repair genes.
These are the genes involved in diseases such as xeroderma pigmentosum
and ataxia-telangiectasia.
Upon reflection, this makes sense. As different as plants and
humans are, they both must maintain and repair an information-encoding
system based on the integrity and fidelity of sequences of DNA bases
over extended periods of time. This is why the mustard weed genome
contains homologues of many DNA repair genes that are defective
in human diseases such as hereditary breast cancer, nonpolyposis
colon cancer, and xeroderma pigmentosum.2
Suddenly, scientists who study plant genetics are publishing papers
that cannot be ignored by the medical profession. This too is a
portent of things to come in the Postgenome Era. Physicians must
prepare for an exponential increase in the mass of medically relevant
new information that patients will expect their doctors to master.
And patients’ expectations grow every day as they rely on the Internet
to educate themselves about the latest medical advances. Physicians
in the Postgenome Era confront a challenge similar to that experienced
by priests during the Reformation.
Prior to the Reformation, priests in medieval Europe enjoyed a
monopoly on religious knowledge. Dissemination of the holy word
relied on an inefficient, labor-intensive technology. Generating
a single copy of the Bible consumed at least a full year of effort
by a well-trained and highly skilled scribe wielding a goose quill
over vellum sheets. Only the Church possessed enough wealth to sustain
the transmission of biblical knowledge from one generation of the
priesthood to the next. Ordinary people were ignorant of the written
word residing in scarce and inaccessible Bibles. Only ordained priests
had direct access to a Bible, and laypeople depended on them to
reveal its contents. Priests occupied positions of trust, prestige,
and power. In pre-Reformation Europe, it was good to be a priest.
In the second half of the 15th century, a technological breakthrough
sowed the seeds of a revolution in the way people acquired information.
Gutenberg invented the moveable-type printing press and mass-produced
Bibles, an act that triggered many unintended and unimagined consequences.
One of them was the Reformation, which shook the foundations of
the Church and changed forever the relationship between the priesthood
and laypeople.
By the end of the 16th century, ordinary people could own a Bible
and read for themselves what was in it. For example, Proverbs 24:5,
"vir sapiens fortis est et vir doctus robustus et validus," articulates
the concept that knowledge is power. The knowledge monopoly of the
priesthood disintegrated, and survivors had to find new ways to
add value to the lives of their parishioners. In post-Reformation
Europe, it became much harder for priests to know more than their
congregations.
Fast-forward 500 years to the second half of the 20th century
and consider the parallel between the priesthood during the Reformation
and the medical profession as it enters the Postgenome Era. Formerly
ignorant patients are being empowered by a new technology that gives
ordinary people access to the latest medical information. This time
it is the Internet rather than the printing press, but the threat
posed to professional hegemony is the same.
Physicians must learn to survive in the Postgenome Era by practicing
informatic medicine, the next level in the evolution of scientific
medicine. Armed with a good medical education and a mental prosthesis
(a portable computer connected to an interactive network), physicians
in the 21st century will add value to the lives of their patients
in ways that previous generations of physicians never dreamed possible.
Only the best information management tools will enable physicians
to keep up with the avalanche of new information relevant to human
medicine.
As the Postgenome Era begins, privately endowed medical research
institutes, universities, and hospitals are adding their efforts
to those funded by the $20 billion-plus annual budget of the National
Institutes of Health. Pouring forth from this collective endeavor
is a growing avalanche of new information that threatens to bury
physicians. Every week sees the publication of breakthroughs in
our understanding of the genetic causes of disease. And the rate
at which these breakthroughs are coming increases each month. The
burgeoning productivity of this research fuels the revolution in
biomedicine and illustrates the overwhelming task faced by physicians,
who must assimilate and use this new knowledge.
The low-hanging fruit of the Postgenome Era are diseases
in which single gene mutations play a decisive pathogenic role.
Many recent examples represent subsets of a larger disease category.
Each traditional disease category—each distinctive constellation
of symptoms—may be divided into many different subsets involving
one or more genes. Clearly, some subsets and some entire categories
of disease will involve more complex causes. Some will involve tens
or even hundreds of different genes. In all cases, interactions
among genes as well as those between genes and the environment must
be considered.
Although much low-hanging fruit will be harvested in the next
decade, many multigenic diseases will keep the research community
busy for several decades. Along the way, how we think about disease
will change dramatically. If diagnosing disease is the first step
to effective therapy, physicians will increasingly think of diseases
more in terms of causes than symptoms. Knowing that someone has
hypertension or muscular dystrophy or leukemia means relatively
little in the new age of personalized medicine. To treat such diseases
effectively, physicians need to know if the problem resides in a
mutated serine-threonine kinase gene, as in the case of pseudohypoaldosteronism
type II hypertension from work by Wilson and colleagues at Yale8;
a mutated ZNF9 gene in myotonic dystrophy type II from work by Liquori
and colleagues at Minnesota9;
or a mutated tyrosine kinase gene in chronic myelogenous leukemia
from work by Druker and colleagues at Oregon.10
For almost a century after Rudolph Virchow described a patient
with "white blood" in 1845,11
leukemia was thought to be a single disease. In much of the 20th
century, leukemia was classified according to the course of the
disease (acute or chronic) and the type of cell giving rise to the
malignancy (lymphocytic or myelogenous). Leukemia is now known to
result from a multitude of pathogenic mechanisms, many involving
chromosomal translocations and gene fusions.12
Leukemia illustrates how diseases in the Postgenome Era will be
defined less by signs and symptoms and more by molecular pathogenesis.13
The ability to target specifically each of the many fusion proteins
that cause various forms of lymphocytic and myelogenous leukemia
holds great promise for personalized treatment of patients with
blood cancer.
One category of leukemia, recognized in the Pregenome Era as chronic
myelogenous leukemia, results predominantly from a chromosomal translocation
that disrupts the normal DNA sequence of the gene for a growth-promoting
enzyme known as tyrosine kinase.
Fusion of the c-abl gene of chromosome 9 with the bcr gene of
chromosome 22 creates a novel chimeric oncogene. White blood cells
carrying this chromosomal translocation produce a fusion protein,
Bcr/Abl tyrosine kinase, that has five- to 10-fold greater enzymatic
activity than c-Abl tyrosine kinase. The constitutively expressed
and overly active enzyme causes cancerous proliferation of the affected
cells.14
The Novartis drug Gleevec binds to the active site of the mutant
tyrosine kinase and blocks its ability to promote abnormal cellular
growth.10 Although Gleevec can also bind normal tyrosine
kinase found in white blood cells lacking the chromosomal translocation,
it does no apparent harm to healthy cells and avoids the devastating
side effects associated with non-specific chemotherapeutic agents
traditionally used in cancer therapy. This new drug exhibits remarkable
efficacy in patients with chronic myelogenous leukemia. Gleevec
has quickly come to exemplify the Holy Grail of the pharmaceutical
industry, knocking out cancer cells while leaving healthy cells
alone.
But all is not perfect with this magic bullet. Despite taking
Gleevec, some patients progress from the chronic phase of myelogenous
leukemia to the acute, blast phase of the disease. Gleevec’s failure
to suppress leukemia in these patients yields fascinating insight
into the molecular pathology of the disease.15
In some of the resistant patients, amplification of the mutant
tyrosine kinase gene outpaced the ability of the inhibitory drug
to bind enough of the target enzymes and allowed uncontrolled growth
to occur. In most of the resistant patients, an additional single
amino acid substitution in the mutant tyrosine kinase had occurred.
Substituting an isoleucine for a threonine at the 315th amino acid
in the enzyme caused steric hinderance that blocked the binding
site of Gleevec. The doubly mutated tyrosine kinase retained its
cancer-producing activity, but Gleevec could no longer inhibit it.
Hence, it remains a rational drug target, and high-throughput screening
of candidate drugs may yield a new molecule capable of specifically
inhibiting the altered enzyme that Gleevec no longer affects.
Gleevec offers insight into the kind of personalized therapies
promised by gene-based molecular medicine. It is a dramatic departure
from the one-size-fits-all pharmacology of the 20th century and
opens the way for pharmacogenomics, one of many manifestations of
the Postgenome Era.
Pharmacogenomics owes its existence to genetic variation among
members of the human species. Among the individuals whose DNA was
studied, the Celera and the public sequencing projects revealed
fewer than three million single nucleotide differences out of nearly
three billion base pairs in the human genome.3,4
However, this one-base pair difference out of every thousand in
the human genome is enough to provide the theoretical basis for
identifying prospectively patients who will benefit from a therapeutic
drug and those who will develop harmful side effects. Many patients
will soon be genotyped, with their personal array of single nucleotide
differences sorted into two categories: those that are also found
in at least one percent of the population (single nucleotide polymorphisms,
or SNPs) and those that occur less frequently (mutations). One can
hardly imagine what this will mean in terms of storing and retrieving
relevant information. It will present enormous challenges to developers
and managers of clinical databases and patient records.
The underlying concept of pharmacogenomics is not new. Karl Landsteiner
introduced personalized medicine nearly a century ago with his classification
of humans into four phenotypes based on blood antigens of the A,
B, AB, and O groups.16 We now
know that these phenotypes result from a diploid combination in
each of us of three different alleles at the ABO gene locus.17
If an individual with only A alleles at the ABO locus receives a
transfusion of blood from an individual with only B alleles at the
ABO locus, that recipient experiences devastating side effects.
Knowing the genotype of a blood transfusion recipient permits informed
selection of the therapeutic agent (i.e. blood from a compatible
donor). Similarly, knowing the SNP profile of a patient will allow
a physician to avoid certain drugs in favor of those that are known
to produce only desirable outcomes in people with that genotype.
But as Gleevec illustrates, pharmacogenomics offers additional
opportunities beyond the ability to deploy existing drugs more intelligently.
In this age of personalized medicine, pharmaceutical companies will
develop entirely new drugs that target the abnormal proteins produced
by SNPs, mutations, and other individual genetic variants (including
chromosomal rearrangements) that alter gene sequences. Pharmaceutical
companies will also focus on protein targets that turn genes on
or off. There will soon emerge a new pharmacology based on drugs
that target protein transcription factors to suppress expression
of a disease-causing gene or to promote expression of a disease-suppressing
gene.
We now know that the human body has many more proteins than genes.
Although we have a few tens of thousands of genes in our genome,
those genes are capable of producing many hundreds of thousands
of distinct proteins. Add to this the potential of our personal
SNPs and mutations to yield thousands more variant proteins, and
one begins to appreciate the potential magnitude of the new proteomic
pharmacopoeia and the database required to deploy it intelligently.
Doctors must prepare for a Physician’s Desk Reference that
will rival the Oxford English Dictionary in its bulk. Like
the OED, the expanded PDR will be user-friendly only in an electronically
searchable format.
Gleevec also points to the financial dilemma society faces as
the new pharmacology of personalized medicine gains momentum. Highly
specific drugs tailored to a small subset of patients who will benefit
from custom-designed therapy will cost much more than traditional
drugs. A year’s supply of Gleevec for a patient known to have chronic
myelogenous leukemia costs $40,000.18
It is already known that a small number of patients relapse while
taking Gleevec and become nonresponders due to a new mutation in
the abnormal tyrosine kinase that originally caused their disease.15
Suppose a pharmaceutical company rises to the challenge of finding,
developing, and marketing a drug that will specifically target and
inactivate the doubly aberrant tyrosine kinase in these Gleevec-resistant
patients. How much must the company charge to recover its investment
in such a personalized drug?
Raising this question is not intended to discourage pursuit of
personalized therapies. But it does acknowledge an economic consequence
of pharmacogenomics. Society must decide how much of the economy
can be devoted to implementing the technologies resulting from the
revolution in biomedicine. And the health care industry must decide
how many resources can be devoted to managing the massive amounts
of patient data that pharmacogenomics portends.
Gleevec also shows how the Internet is changing the environment
in which medicine is practiced. Novartis has a Web site that tells
physicians how to prescribe Gleevec and how to help their patients
pay for it.19 This official Gleevec
site is intended for physicians, but patients can also visit it.
There is an unofficial Gleevec site, conscientiously maintained
by a grateful patient with chronic myelogenous leukemia whose disease
has been suppressed by the new drug.20
This site provides information to other CML patients and was created
specifically for them. In the equal-opportunity Internet age, physicians
may also visit this site.
This unofficial site is maintained by a man who spent 21 years
as a trooper with the Louisiana State Police, was diagnosed with
CML three years ago, and started Gleevec therapy a year later. He
operates the site as a labor of love for other CML victims and directs
all donations to the Leukemia and Lymphoma Society. He exemplifies
the contrast between the new age of smart clients versus the old
days of "helpless" patients.
Gleevec illustrates how the Internet will make the knowledge that
empowers physicians equally accessible to patients in the same way
the printing press and moveable type made the knowledge contained
in the Bible accessible to parishioners in 16th-century Europe.
Consider this statement from a CML patient whose spouse learned
about Gleevec on the Internet.21
"My doctors were very supportive, but sadly doctors are the last
to know about some developments. Get on the Internet, look at medical
progress made. I would not have (been in the Gleevec trials) if
it weren’t for my husband’s research (on the Internet)."
So how should physicians cope with continuous subclassification
and redefinition of diseases afflicting patients who aggressively
seek news of the latest breakthroughs in their areas of interest?
The situation will only get much worse, since research-intensive
biomedical institutions will generate new knowledge faster than
individual physicians can assimilate it.
First and foremost, physicians in training and physicians in practice
must learn to understand, appreciate, and use the best tools information
technology can offer. The platform of this tool kit is a mental
prosthesis.
Many of us need visual prostheses (eyeglasses, contact lenses)
to improve our ability to see. Eventually, many of us will need
auditory prostheses (hearing aids) to improve our ability to hear.
Already, all of us need mental prostheses to improve our ability
to acquire, organize, and interpret information. Physicians cannot
survive without a mental prosthesis. The massive amount of information
that must be retrieved and used each day is far beyond the capacity
of human memory. It can only be accomplished with the help of a
portable computer connected to an interactive network and equipped
with data-mining software that exercises logic to locate and assemble
information specific to the individual physician’s professional
requirements.
For those who deliver health care in the 21st century, the challenge
is to bring all the relevant information together at the same time
and place so physicians can make informed decisions about what is
best for the patient. Just as medical education had to be transformed
in the early 20th century to enable physicians to practice scientific
medicine, the training of physicians in the 21st century will have
to change to allow them to practice informatic medicine. Physicians
must prepare for lifelong, everyday, just-in-time acquisition of
knowledge from afar, and they will need new and better information
management tools.
The world of free enterprise will address this need. It represents
a substantial business opportunity. Being among the first to know,
or at least knowing as soon as smart clients, is crucial to maintaining
the physician’s role in health care. Innovative software developers
will create the new tools, and physicians will pay for them, since
effectiveness and status in a knowledge-based profession utterly
depend on it.
References
1. Adams M., et al. Science. 2000;287: 2185-2195.
2. The Arabidopsis Genome Initiative. Nature. 2000;408:796-815.
3. International Human Genome Sequencing Consortium. Nature. 2001;409: 860-921.
4. Venter J., et al. Science. 2001;291: 1304-1351.
5. Rubin G., et al. Science. 2000;287: 2204-2215.
6. Warrick J., et al. Cell. 1998;93:939-949.
7. Warrick J., et al. Nature Genetics. 1999; 23: 425-428.
8. Wilson F., et al. Science. 2001;293: 1107-1112.
9. Liquori C., et al. Science. 2001;293: 864-867.
10. Druker B., et al. N Eng J Med. 2001; 344:1031-1037.
11. Major R. Classic Descriptions of Disease. Charles C. Thomas; 1945:510-513.
12. Look A. Science. 1997;278:1059-1064.
13. Temple L., et al. Science. 2001;293: 807-808.
14. Lewin B. Genes VI. Oxford University Press; 1996:1147-1148.
15. Gorre M., et al. Science. 2001;293: 876-880.
16. www.nobel.se/medicine/laureates/1930/landsteiner-bio.html
17. Lewin B. Genes VI. Oxford University Press; 1996:69.
18. Fuhrmans V, Carroll J. The Wall Street Journal. May 11, 2001.
19. www.gleevec.com/
20. www.newcmldrug.com
21. Cox News Service (June 26, 2001) @ www.intellihealth.com
|