Growing at a healthy rate of
10-15% per year, the therapeutic proteins market is
presently estimated at more than $30 billion. Natural
proteins, such as recombinant insulin, erythropoietin (epo),
and GCSF, together with the ever-expanding assortment of
monoclonal antibody (mAb)-based drugs on the market,
account for much of that growth and commercial success.
The ability to exploit molecular biology tools and
high-speed computational analysis to modify, manipulate,
and model natural proteins — engineering them in very
directed ways to enhance their biological potency,
improve their metabolic and pharmacokinetic properties,
and increase their resistance to proteolytic degradation
— offers enormous opportunities for creating more
"druggable" disease-fighting proteins and more
powerful second-generation mAb drugs.
Bassil Dahiyat, president and CEO at Xencor, sees
"a lot of potential to engineer monoclonal
antibodies to be more active, more potent, and better
tolerated," by optimizing how they interact with
and recruit the immune system. Today’s mAbs "are
far from perfect," he says.
Building Better Proteins
The central challenge in engineering large, complex
protein molecules, including mAbs, is determining how to
optimize their innate biological function, while
enhancing their selectivity and potency and minimizing
their less desirable characteristics, such as their
tendency to aggregate. These properties are often in
conflict with one another, and molecular engineering
strategies require a careful balance between fine-tuning
a desirable feature without enhancing an undesirable
one.
Xencor relies on sophisticated computational tools,
computer modeling, and quantitative metrics to make
"virtual" changes to protein molecules and
rapidly and systematically analyze the effects of those
modifications on a protein’s overall structure,
predicted activity, and pharmacodynamics. This approach
yields a limited set of engineered molecules that can
then be tested in the laboratory against a disease
target.
For example, Xencor has applied these tools to
engineer the Fc (constant) region of monoclonal
antibodies to enhance their immune effector function and
has demonstrated a 100-fold improvement in an antibody’s
killing potency against a tumor cell. The company hopes
to take this compound into the clinic within the next 12
months. During the first half of 2007, Xencor plans to
file an IND for an engineered mAb to treat Hodgkin’s
disease.
Anticalins are engineered human proteins derived from
a lipocalin scaffold. Similar to antibodies, lipocalins
have a hypervariable region that Pieris, a German
biopharmaceutical company, is using as a protein
scaffold for engineering Anticalin drugs that tightly
bind their targets. Pieris has generated 3D crystal
structures of a target-specific Anticalin in complex
with its ligand, cytotoxic T lymphocyte antigen (CTLA)-4.
The findings support the predicted plasticity of the
human lipocalin scaffold.
PRS-010 and PRS-050, the company’s lead compounds,
target the CTLA-4 co-receptor (to inhibit T-cell
activation) and VEGF, respectively. Both have potential
implications for treating cancer. Pieris is also
exploring PRS-050’s potential for inhibiting retinal
neovascularization.
Protein engineering could enable a single dose of a
drug that might exert its effects at a target site for
only a few hours to be effective for several weeks.
Consider the recent example of an anti-inflammatory drug
injected directly into an arthritic joint to treat
osteoarthritis. Duke University researchers modified
interleukin-1 receptor antagonist (IL1RA) by attaching
to it a second protein, an elastin-like polypeptide (ELP).
ELPs clump together at body temperature. As the ELPs
aggregate they form a drug depot; their slow
disaggregation enables prolonged release of the IL1RA at
the site of inflammation (J. Controlled Release
2006;115(2):175-182).
Effective Engineering
At the core of protein engineering efforts is an
understanding of the basic structure of natural proteins
that have potential therapeutic applications. The
National Institutes of Health’s Protein Structure
Initiative (PSI) includes a planned Materials Repository
that will offer a centralized store of PSI-generated
clones available to researchers at a minimal fee. The
repository will be housed at Harvard Medical School’s
Institute of Proteomics, in Boston, Mass.
Another component of the 10-year initiative, begun in
2000, is a planned knowledge base, or information hub.
The knowledge base, slated for fall 2007, will compile
the structural information generated by participating
centers and provide database-mining capabilities based
on parameters of protein structure and function.
