SOLUBLE T-CELL ANTIGEN RECEPTORS - Fusion Protein Technologies for Biopharmaceuticals: Applications and Challenges - page 460

displayed on the cell surface, thus is not accessible to
antibody-based agents. As with other intracellular proteins,
p53 is processed and p53 peptides are presented at the cell
surface in the context of HLA molecules. We and others
have found that a peptide epitope (aa264-272) of p53
presented by HLA-A 0201 is displayed at high levels on
the surface of different human tumor cells and tissues,
whereas normal tissues do not present detectable levels of
this complex [15,33,34]. Since this epitope is within a region
of p53 that is rarely mutated, its cell-surface display serves
as an attractive broad-based target for tumors that over-
express p53.
Cytotoxic T-lymphocytes (CTLs) that specifically rec-
ognize human p53 (aa264-272) peptide/HLA-A 0201
complex have been generated from HLA-A 0201/K b -
transgenic mice that were immunized with the p53 peptide
[35]. The a / b TCR genes were isolated from a clonal line
of these cells and constructed in a single-chain format
genetically fused to human IL-2 to generate the soluble
264scTCR/IL-2 fusion protein [5]. The ALT-801 fusion
protein is a chimeric form of 264scTCR/IL-2 wherein the
mouse TCR b -constant domain of 264scTCR/IL-2 was
replaced with the human counterpart [17]. This replace-
ment is expected to significantly lower the immunogenicity
of the TCR portion of the molecule in humans. As
described in the following section, ALT-801 has been
thoroughly characterized in preclinical efficacy and toxi-
cology studies and in a Phase I human clinical study in
patients with p53-overexpressing (p53 þ ) metastatic malig-
nancies. This fusion protein is currently being evaluated in
Phase II clinical studies in patients with melanoma and
muscle invasive or metastatic urothelial cancer with addi-
tion indications in the planning stage. Using a similar
strategy, we have linked the p53 peptide-specific scTCR
domain to the human IgG1 domain to create the antibody-
like 264scTCR/IgG1 and corresponding chimeric ALT-802
fusion proteins [14], the characterization of which is
described in the following section.
We have also extended this approach to generate tar-
geted fusion proteins with scTCR domains that recognize
peptides specific to the melanoma differentiation antigens,
gp100 and MART-1, and the cancer testis antigen, MAGE-
A3, that are expressed at high frequency by human tumors
including melanoma, nonsmall cell lung carcinoma
(NSCLC), head and neck cancer, hepatocellular cancer,
and multiple myeloma [36,37]. In each case, these scTCR
domains are functionally active in recognizing their cog-
nate peptide/HLA-A 0201 complexes presented on the cell
surface. Overall, these fusion proteins have the potential to
treat a broad range of cancers in HLA-A 0201-positive
individuals, who represent

readily implemented based on current knowledge of tumor
antigen presentation and tumor-specific T-cell responses
[38].

31.3.2 Viral Therapeutics

Based on the success of our cancer-specific STAR program
and due to the lack of effective preventive vaccines or viral
therapies, we are taking a similar approach toward devel-
oping STAR molecules recognizing viral antigens from
HIV, hepatitis C (HCV), and cytomegalovirus (CVM)
antigens. These STAR molecules are being engineered
into various formats as fusion proteins and drug or radio-
isotope conjugates to deliver antiviral therapies to diseased
cells. To create the scTCR domains, we have established
collaborations with leading viral immunologists who have
provided Altor with virus-reactive human CTLs or TCR
genes obtained from infected donors. From these reagents,
we generated over 20 functional scTCRs reactive to dif-
ferent immunodominant viral peptide/HLA complexes. In
some cases, these scTCRs have been further engineered
though site-specific mutagenesis to generate variants with
improved binding affinity to the cognate viral peptide
antigen as well as known escape mutants that arise during
the course of viral infection. Thus, our collection of
antiviral STAR molecules provides a broad range of spe-
cific reactivities against infected cells. In the case of HIV,
we have developed a panel of STAR molecules capable of
recognizing epitopes that are restricted by different HLA
molecules including those associated with disease protec-
tion (HLA-B 5701 and HLA-B 2705) or found at high
frequency (i.e., HLA-A 0201 in the U.S./European popu-
lations and HLA-A24 in the Asian population). Through
our collaborations, we have demonstrated that these fusion
proteins are capable of recognizing viral peptide presented
on HIV-infected cells and directing cytotoxicity via con-
jugated IgG Fc domains or radioisotopes against cells
displaying HIV antigens. These agents could form the
basis of a new treatment strategy of HIV, whereby the
STAR fusion proteins target the reservoir of HIV-infected
cells in patients currently on approved antiviral therapies.
Potentially, such a strategy could result in the elimination
of viral infection as opposed to suppression of viral
replication with continued residual disease as observed
with current treatments. A similar approach is being taken
for HCV and CVM specific STAR fusion proteins where
our preclinical testing is focused on a STAR-Ck fusion
between the approved therapeutic cytokine IFN- a and a
high-affinity scTCR domain specific recognizing a peptide
derived from the HCV protein NS5 presented in the context
of HLA-A 0201. Testing is underway to determine if this
fusion protein capable of delivering IFN- a to HCV-
infected cells, thereby improving its efficacy while reduc-
ing systemic side effects.

50% of the U.S./European
population. Expansion of this strategy to tumor peptide
antigens displayed by other common HLA class I haplo-
types, such as HLA-A24 in the Asian population, could be