Researchers
at the Stanford University School of Medicine and Intel Corp. have
collaborated to synthesize and study a grid-like array of short pieces
of a disease-associated protein on silicon chips normally used in
computer microprocessors. They used this chip, which was created through
a process used to make semiconductors, to identify patients with a
particularly severe form of the autoimmune disease lupus.
Although
the new technology is focused on research applications, it has the
potential to eventually improve diagnoses of a multitude of diseases, as
well as to determine more quickly what drugs may be most effective for a
particular patient. It may also speed drug development by enabling
researchers to better understand how proteins interact in the body.
“When
I see patients in the clinic right now, I may know they have arthritis,
but I don’t know which of the 20 or 30 types of the disease they have,”
said associate professor of medicine Paul (P.J.) Utz, MD, noting that
existing methods can take days or even weeks to answer such questions.
“Now we can measure thousands of protein interactions at a time,
integrate this information to diagnose the disease and even determine
how severe it may be. We may soon be able to do this routinely while the
patient is still in the physician’s office.”
Utz is a co-senior author of the research, published online Aug. 19 in Nature Medicine.
Postdoctoral scholar Chih Long Liu, PhD, and Madoo Varma, PhD, director
and head of life science research operations and business strategy for
Intel’s Integrated Biosystems Laboratory, are the other senior authors.
Graduate student Jordan Price is the first author. The research was
funded in part by Intel Corp., and Intel scientists created the protein
array on the silicon chips for the Stanford researchers to study.
Within
each of our cells, proteins enter into and disband physical
relationships in dizzying succession—outdoing even our most-freewheeling
Facebook friends. This intricate dance forms the machinery responsible
for driving cell growth, sparking immune reactions and even causing
disease. But understanding the microscopic minutiae of their fleeting
attractions (why exactly is protein X hooking up with protein Y?), and
the subsequent biological repercussions, has been difficult and
time-consuming.
To
better understand these interactions, researchers at Intel synthesized
short segments of biological proteins, called peptides, on silicon
wafers. To do so, they turned to the same process used to make
semiconductors, employing a method using sequential steps of light
exposure and chemical reactions called photolithography. The Stanford
researchers then used the chip, which they’ve termed an Intel array, to
analyze thousands of protein interactions simultaneously to diagnose
disease, assess therapies and even design more-effective drugs.
The
researchers hope to eventually embed an integrated semiconductor
circuit within the microprocessor-ready silicon chip to create a sort of
minicomputer that could take the guesswork and decision-making out of
many clinical processes. It could perhaps spell out patient-specific
diagnoses with letters of the alphabet, or identify which potential
treatments are most likely to be effective.
The
technology described in the study echoes that of DNA microarrays, in
which thousands of unique nucleotide sequences are dotted on a glass
slide in a grid-like pattern to identify patterns of gene expression in
cells and tissues. Prior to the collaboration with Intel, Utz and his
colleagues were using a similar technique for peptides—affixing them in
defined patterns to glass or other substrates and then washing them with
solutions of cellular or blood-borne proteins. A binding event between a
protein in the solution, such as an antibody, and its slide-bound
partner is indicated by a fluorescent signal, which is developed through
a meticulous and lengthy series of detection steps.
About
four years ago, however, researchers at Intel approached Utz and his
colleagues with the idea of using silicon as a microarray platform to
synthesize the peptides directly on the chip, rather than making the
peptides separately and spotting them on the array using a robot.
“Honestly,
we thought it wouldn’t work,” said Utz. But it did, and it had several
advantages. For one thing, silicon is much less sticky to proteins than
glass. As a result, researchers can skip some experimental steps meant
to block random binding of peptides to the substrate. Silicon also
allows the researchers to arrange the individual peptides more closely
together, using the space much more efficiently. Finally, unlike glass,
silicon alone does not fluoresce, making signal detection easier.
There’s also the promise of devising new, faster detection methods on the more-versatile silicon chip.
“If
we couple these Intel arrays with an electronic detection method, for
example, we could have real-time sensing over a period of minutes,” said
Utz.
In
the study, the researchers tested whether their array could help
categorize patients with lupus—an autoimmune disease in which patients
make antibodies that attack a type of protein in their cells called a
histone (in addition to other proteins).
“Lupus
is highly variable, and in some cases is quite severe,” said Utz.
“About half of patients are likely to require more intensive therapy. We
wanted to see whether we could identify these patients with our
arrays.”
Using
the new silicon chips, the researchers were able to identify patients
with lupus who expressed high levels of antibodies against a particular
histone called 2B. They then confirmed that these patients were
precisely the ones struggling with a more severe form of the disease.
“Companies
developing therapies to block the pathway responsible for this binding
are now accepting patients with lupus for clinical trials without
knowing which subset of disease they are in,” said Utz. “This method
could potentially be used to identify only those patients likely to
benefit, and aid in the identification of effective drugs.”
To
make the discovery, the researchers made a microarray using the last 21
amino acids of histone 2B. Histones keep DNA packaged tightly within a
cell’s nucleus; binding of various proteins to the exposed end of the
histone selectively grants or excludes access to the packed genes. The
global importance of the binding events, and the fact that autoimmune
diseases like lupus arise when the body makes antibodies against the
histone’s end, led the researcher to choose it for their first test of
the technology.
In
making the array, they synthesized every possible overlapping sequence
of every length from the short string of amino acids: 1-21 (the
full-length sequence) to 17-20 (four amino acids) to 2-20 (19 amino
acids) and all other possible variations. Three of the amino acids are
also sometimes modified to carry extra chemical groups that can enhance
or impede protein binding. Including every possible length and
combination of modified and unmodified amino acids gave nearly 9,000
unique peptide dots on the array. They then washed the chip with
solutions of antibodies known to bind the sequence.
The
pattern of binding showed that one antibody could recognize and bind to
a sequence composed of only two amino acids of the original 21. Another
required at least four amino acids, one of them modified, for binding.
Analyzing the binding of solutions of other antibodies in each case
delineated specific binding regions, or epitopes, within the original
short sequence.
Understanding
the binding at such levels of detail will allow researchers to tinker
with drugs meant to disrupt, enhance or mimic biological reactions
within our cells to create better therapies, or to understand how and
why natural processes sometimes go awry.
The
researchers are now exploring the use of the technique to help design
influenza vaccines that elicit a strong immune response, as well as ways
to incorporate the three-dimensional folding involved in most protein
interactions.
Other
Stanford researchers involved in the study include former research
assistant Stephanie Tangsombatvisit; postdoctoral scholar Dan Levy, PhD;
and associate professor of biology Or Gozani, MD, PhD.
In
addition to Intel, the research was supported by the National Science
Foundation; the Stanford Genome Training Program; the Donald E. and
Delia B. Baxter Foundation; the Ellison Medical Foundation; the National
Heart, Lung and Blood Institute; the Canadian Institute of Health; the
Ben May Trust; the Floren Family Trust; the National Institutes of
Health; and the European Union Seventh Framework Programme.
Source: Stanford University
