IMMUNORNASE FUSIONS - Fusion Protein Technologies for Biopharmaceuticals: Applications and Challenges - page 332

RNase 1 [108,109]. The latter was constructed by several
amino acid replacements. Two of them introduced Cys
residues in positions analogous to those in naturally dimeric
BsRNase. They formed two intermolecular disulfide bonds.
Other replacements either favored dimerization by creating
a hydrophobic interaction or enhanced cytotoxicity. The
novel immunoRNase was obtained by fusing the scFv
(with a modified linker) to the engineered enzyme through
a previously developed spacer [107]. This single-chain
construct dimerized by forming intermolecular disulfide
bonds in its enzyme moiety (molecular mass 92- kDa).
The dimeric form did not interact with the ribonuclease
inhibitor and was 2.5-fold more cytotoxic than the monomer
[107].
Another Erbicin-derived immunoagent was a compact,
fully functional anti-ErbB2 antibody [110]. In this construct,
Erbicin was fused to the Fc domain of human IgG. The
resulting antibody-like molecule (100 kDa) showed anti-
cancer activity in vitro and in vivo (see Reference 66 for
further discussion). Logical consequence of this project was
very recent fusion of RNase 1 to this compact antibody
[111]. The novel immunoRNase (140 kDa) was endowed
with RNase- and complement-dependent toxicities as well
as antibody-dependent cytostatic activity against ErbB2
positive tumor cells. Its catalytic activity was slightly lower
than that of the monomeric human immunoRNase devel-
oped earlier by the authors [103]; however, the mammalian
RNase inhibitor affected the novel construct to a signifi-
cantly lesser degree. Both effects might be due to steric
hindrances within the larger molecule. The novel immu-
noRNase was also more potent, both in vitro (toward ErbB2
positive SKBR3 cells) and in vivo (mice TUBO tumors) than
the parental monomeric compound.
The Erbicin-based immunoRNases described in this sec-
tion were found to interact with different ErbB2 epitope than
those targeted by Herceptin (trastuzumab) [112]. Trastuzu-
mab is the only ErbB2 specific monoclonal antibody
approved for cancer treatment. Therefore, these Erbicin-
based immunoRNases are active against Herceptin-resistant
tumors but, unlike Herceptin, do not elicit cardiotoxicity
[112-114].
Human pancreatic RNase was also targeted to the CD30
antigen overexpressed in lymphomas like Hodgkin- and
anaplastic large cell lymphomas, but not in resting hema-
tolymphoid cells [81]. An anti-CD30 scFv of murine Ber-H2
monoclonal antibodywas fused to theN-terminus of RNase 1;
the fusion protein (calculated mass 46.3 kDa) was expressed
in Drosophila S2 cells and purified from the cell culture
supernatant by cationic exchange chromatography [115]. It
was highly toxic to Karpas 299 and two other CD30-positive
cell lines and only slightly active against cells that do not
express this antigen. It also significantly reduced the growth of
CD30 positive tumors in the BALB/c mice model. Interest-
ingly,

inhibitor (interaction of immunoRNases with RNase inhibitor
is discussed later in this chapter).
More recently, Menzel et al. [116] designed the first
entirely human immunoRNase targeting CD30-positive
lymphomas. Human anti-CD30 scFv (4E4) generated by a
phage display methodology was fused to the Fc fragment of
human IG1 and then to HpRNase through the AAASSG
spacer. The V H and V L parts of the scFv were connected by
the (G 4 S) 3 linker. The fusion protein (75 kDa) was expressed
in human embryonic kidney cells (HEK 293T) with a good
yield of 4 mg/L. Purification from the supernatant by protein
A column chromatography predominantly yielded a homo-
dimer of about 150 kDa molecular mass. The product was
stable in human serum and strongly inhibited the growth of
CD30-positive Karpas-299 cells (Table 22.1).

22.2.3 Immuno-Fusion Proteins of Onconase 1 or Rana
Pipiens Liver RNase
Onconase 1 (ranpirnase), already discussed in Section 1.2,
was isolated from extracts of Rana pipiens (Leopard frog)
oocytes and early embryos; it was the major cytotoxic
component. Onconase 1 was identified as an RNase only
after amino acid sequencing [33]. Rana pipiens Liver RNase
(RapLR) is a close variant of Onconase 1 that was identified
by cloning from genomic DNA [117]. The two proteins
differ in four (of 104) amino acid residues. In RapLR, Ile11
of Onconase 1 is replaced by Leu, Asp20 by Asn, Lys 85 by
Thr, and Ser103 by His in RapLR. Since three other natural
variants of Onconase 1 were found in R. pipiens oocytes
[18], it seems that the frog genome contains at least five
genes encoding Onconase 1 variants with replacements at
positions 11, 20, 25, 85, and 103.
N-terminal pyroglutamate of Onconase 1 and other
known frog RNases, with the exception of amphinase
[18,118], is an integral part of the active sites of these
enzymes and is critical for their catalytic and antiprolifer-
ative/cytotoxic activities [119-122]. Unexpectedly, RapLR
was found to be fully active without the N-terminal pyro-
glutamic acid residue [123]. This finding is difficult to
explain, as the 3-D structure of this enzyme has not been
refined. The lack of N-terminal pyroglutamic acid has an
important practical implication. Unlike Onconase 1 , this
RNase may be fused to targeting molecules through its
N-terminus. However, fusion proteins of RapLR obtained
to date have been designed in the RNase-scFv format. The
enzyme was targeted to MUC1 (DF3) antigen, which a high
molecular weight, transmembrane glycoprotein, overex-
pressed in adenocarcinomas and hematological malignances
[124-126]. Single-chain Fv derived from humanized anti-
MUC1 HMFG1 antibody was engineered to a stable form by
an Arg Ala replacement at position 71 in V H fragment

[127]. The product was fused to the C-terminus of RapLRI
through the G 4 S spacer, the construct was expressed in

!

this construct was insensitive to placental RNase