An Unexpected
BMP in the Road
How the Human Genome Project May Help Diabetes
Evan David Rosen, M.D., Ph.D.
Assistant Professor of Medicine,
Harvard Medical School
When
the decisions was made more than a decade ago to sequence the
human genome, there was a tremendous amount of debate within
the scientific community about whether the project would be
worth the effort and expense. Proponents claimed that numerous
health benefits would result, such as the identification of
disease-causing genes and the development of new treatments.
Ultimately, these arguments carried the day.
We’ve now had the human
genome sequence in hand for more than a year. The mouse genome
was mapped recently, and several more genomes are likely in
the near future. Has the promise of genomics been fulfilled?
Well, the answer is yes and no. Clearly, the hunt for disease-causing
genes has been aided by the complete sequence, especially for
diseases caused by a single aberrant gene. For more common diseases
(like diabetes and obesity) that involve multiple genes, progress
has been a bit slower, although most people in the field feel
that with enough time and elbow grease, we will eventually find
these genes as well.
Where the promise of genomics
has been least fulfilled, however, is in the area of treatment.
Simply put, the vast amount of information that we have accumulated
through DNA sequencing of the human genome has not yet been
translated into useful drugs for people in need of them.
A new report in the journal Nature Biotechnology, however, introduced
a potentially exciting new anti-diabetes drug that was discovered
through genomic approaches—exactly the sort of discovery
that genome groupies have been promising for years.
Here’s how it happened.
Researchers at a biotechnology company called Human Genome Sciences
searched computer databases filled with DNA sequences, looking
for genes encoding proteins that are secreted from cells. The
idea was that these "secreted" proteins might be important
hormones or other messengers involved in the regulation of blood
sugar levels. This task would have been impossible except for
the fact that secreted proteins usually have a small identifying
tag on them, a so-called "signal sequence" that tells
the cell, "Hey, I’m supposed to be secreted."
The scientists looked at a staggeringly large number of protein
sequences (roughly 3 million) in tissues throughout the body,
and unearthed 8,000 that contained a signal sequence.
The next step was equally daunting.
Each of these 8,000 genes was studied individually. First they
were inserted into cells grown in the laboratory. Next the "media"—the
secretions of these cells—was collected. Each batch of
medium, which should contain one of the 8,000 secreted proteins,
was then injected onto several different types of cells relevant
to diabetes. For example, a sample from each of the 8,000 different
media was injected onto pancreatic beta cells, to see if they
triggered the secretion of insulin. Similarly, tests on liver
cells were performed to see if the samples could shut down the
liver's sugar production. Other types of cells were also studied,
such as muscle and fat cells, to see if they could be stimulated
to absorb glucose. This "high-throughput screen" yielded
a few candidate proteins with anti-diabetic potential. The most
promising one turned out to be a hormone called bone morphogenic
protein-9 (BMP-9).
The authors then collected purified
BMP-9, and repeated their assays. They found that BMP-9 could
suppress the ability of liver cells to make sugar nearly as
well as insulin. BMP-9 could also regulate key enzymes in muscle
cells known to be involved in glucose uptake. The next step
was to test the effects of BMP-9 in mice. In normal, non-diabetic
mice, BMP-9 caused a reduction in glucose levels comparable
to that seen with an injection of insulin. Insulin, however,
had its peak effect within thirty minutes, while BMP-9 didn't
kick in until 24-48 hours had passed. The reason for this delay
isn't clear, but it raises exciting questions about the possible
use of BMP-9 as a long-acting drug for diabetes. When BMP-9
was given to obese mice with type 2 diabetes, it also reduced
blood sugar levels effectively for close to 50 hours after the
injection. In addition to mimicking insulin's action in liver
and muscle, BMP-9 had other beneficial effects in the mice,
such as reducing food intake and stimulating insulin secretion.
Despite its name, BMP-9 is made
by liver cells, not bone. While these studies were designed
to look at the effects of BMP-9 at high doses, it may yet turn
out that BMP-9 is an important player in normal glucose regulation.
If so, it would be the first such factor produced by the liver
to be identified, although the existence of such a protein has
long been speculated.
BMP-9 is one of a family of secreted
proteins made by different tissues. Members of this family are
known to be involved in such processes as cellular growth and
organ development. Amazingly, over 4,000 papers have been published
on BMP proteins, with nary a clue that BMP-9 might be involved
in regulating glucose levels. While it remains to be seen if
BMP-9 will turn out to be a safe and effective diabetes therapy,
it is clear that we would never had any reason to study it in
diabetes without this sort of high-tech screen. And this screen
could not have been done without the knowledge created by the
sequencing of the complete genome.
References:
Cecil Chen, Krzysztof J. Grzegorzewski, Steve Barash, Qinghai
Zhao, Helmut Schneider, Qi Wang, Mallika Singh, Laurie Pukac,
Adam C. Bell, Roxanne Duan, Tim Coleman, Alokesh Duttaroy, Susan
Cheng, Jon Hirsch, Linyi Zhang, Yanick Lazard, Carrie Fischer,
Melisa Carey Barber, Zhi-Dong Ma, Ya-Qin Zhang, Peter Reavey,
Lilin Zhong, Baiqin Teng, Indra Sanyal, Steve M. Ruben, Olivier
Blondel, Charles E. Birse. An integrated functional genomics
screening program reveals a role for BMP-9 in glucose homeostasis.
Nature Biotechnology 21, 294 - 301 Mar 2003
Print
This Article
Dr.
Rosens Archives
