SIGNAL CONVERTER PROTEINS - Fusion Protein Technologies for Biopharmaceuticals: Applications and Challenges - page 445

that are being actively developed for the clinic, design is
streamlined. TRAIL-containing SCPs exemplify this nicely.
In effect, this constitutes a new category of personalized
onco-medicine, which looks to higher order tumor cell
surface protein co-expression profiles: CTLA-4

once a dimer of FasL trimers and a trimer of CTLA-4
dimers. This is functionally significant since hexamerization
affords optimal functionality for FasL, and perhaps other
TNF family members [41,105].
The TNF superfamily can be readily mined to generate
other type I

FasL for

CD80/86 þ CD95 þ B-cell malignancies, CD40

FasL for II SCP of interest. However, this fusion formula

unduly narrows the scope of potential SCP. In some instan-
ces, the desired SCP components happen to be both type II
membrane proteins. To circumvent type II

CD154 þ CD95 þ T-cell malignancies, Fn14

TRAIL for
TWEAK þ TRAIL-R þ solid tumors, and so on. Thus, kinase
inhibitors and SCP can be thought of as complementary
personalized onco-medicine therapeutics, attacking cancer
cells from within and without, respectively.

II fusions in such
cases, one can substitute derivatives from agonistic anti-
bodies (by definition, type I) at the amino termini of the
chimeric proteins, thereby allowing for a type I

II SCP.
While such fusion proteins lack the simplicity of type I II

30.8 STRUCTURAL CONSTRAINTS
IN SCP DESIGN

ligand ligand SCP, they have proven functional in our hands.

Production of antibody TNF ligand fusions necessitates

special steps to enrich for the active trimeric forms.
Where the SCP components are to come from two type I
proteins, one can similarly consider antibody

Just as functional requirements have guided SCP design, so
too have structural considerations. The type I (extracellular
amino-terminus) versus type II (extracellular carboxyl-ter-
minus) status of the parental membrane proteins constrains
their placement within fusion proteins. When it came to GPI
protein paints, the predecessors for SCP in the protein
transfer developmental line, the focus was exclusively on
type I membrane proteins, which were modified by append-
ing a GPI modification signal sequence to their carboxyl
termini. The Fc fusion protein paints that followed accom-
modated both type I and II proteins, albeit calling for
alternative positioning of the Fc domain in each. That is,
for type I and II Fc fusions, the Fc domain was positioned at
the carboxyl or amino termini of the fusion proteins, respec-
tively. This served to free up the business ends of the
respective molecules and minimize chances that a covalently
linked Fc appendage would interfere with their activity.
In transitioning to SCP, the most elegant approach was to
generate type I

ligand fusions.

However, in this instance, type I ligand fusions may

also be acceptable, given ample precedents for the function-
ality of such fusions, as evidenced by scFv proteins.

I ligand

30.9 CODING SCP FUNCTIONAL REPERTOIRES

The multifunctionality of individual SCP invites new ways
of codifying their respective functional repertoires. A func-
tional bar code could be devised to concisely capture each
SCP's documented functionalities. For instance, a coding
schema might be structured around an “SCP

fx” format,
wherein fx conveys the full set of known functional capa-
bilities attributed to a particular SCP. A starting set of
primary SCP fx designations might include

j

t2c converts trans signals between two cells

II fusions, an orientation that leaves both
business ends of the chimeric protein free. In fact, all of the
SCP set forth thus far have a type I

tr redirects converted trans signals to third-party cells

t/bd signals bidirectionally in trans

II structural configura-
tion. By accommodating both type I and II proteins within
SCP, there is the added advantage of encompassing at once
members of the immunoglobulin (predominantly type I) and
TNF (predominantly type II) superfamilies, which together
dominate the cell membrane protein repertoire, within the
immune system and beyond.
Interesting stoichiometries and fortuitous higher order
multimer configurations arise when one couples trimeric
TNF superfamily members to nontrimeric type I membrane
proteins. This is exemplified by Fn14

c/ud signals uni-directionally in cis

c/bd signals bidirectionally in cis

neo/t elicits neoexpression of a trans signaling ligand or

receptor

neo/c elicits neoexpression of a cis signaling ligand or

receptor

As an example, the coding schema for Fn14 TRAIL

would be:

tweak þ inflammø act Teff

TRAIL, where our
three-dimensional model suggests that trimerization,
enforced by trimeric TWEAK on the otherwise monomeric
Fn14 component, yields a neo-Fn14 trimer, which can in
turn stabilize the active, trimeric state of the fused TRAIL
component [80]. In the case of CTLA-4

Fn14 TRAIL : SCP j tr f gþ

c = ud f tumor > liver ; pancreas ; breast ; brain ; colon g

TRAIL is
an SCP that has two known functional modes: (1) a TSRP
mode, in which it redirects trans signals from TWEAK-
positive inflammatory cells to activated, effector T cells
and (2) a CLBP mode, in which it directs loop-back

Decoded, this designation conveys that Fn14

FasL, our modeling
suggests that interplay of dimeric CTLA-4 with trimeric
FasL allows for a hexameric configuration, representing at