Supplementary Information
SUPPLEMENTARY RESULTS
PCI of rGel/EGF on A-431 xenografts
A superficial wound could be observed over or in close vicinity to the tumor in
the PCI of 2.5 µg rGel/EGF-treated animal experiencing complete response and in
one of the PCI of 5 µg rGel/EGF-treated animals. The wounds healed, however,
efficiently and no wound or scarring was detectable within 20 days after treatment
(Fig. 4E).
PCI of rGel/EGF cytotoxicity in HNSCC cell lines with diverse differentiation
and EGFR expression
For maximal treatment specificity, rGel/EGF should be minimally toxic when
given as monotherapy and exert high targeted cytotoxicity when activated by PCI.
Estimating the ratio between the targeting index assessed with and without PCI
revealed a ≥ 5-fold larger targeting index when rGel/EGF was combined with PCI in
all the HNSCC cell lines (Fig. 5C). The TIPCI among the HNSCC cell lines was found
to correlate with the TI without PCI (Suppl. Fig. S3D), except in the SCC-026 cells
where PCI augmented the effect of rGel/EGF to the largest extent, increasing the
targeting index of rGel/EGF by a factor of 40.6 (Fig. 5C and Suppl. Fig. S3D).
Establishment and characterization of HNSCC xenografts
The four HNSCC cell lines were inoculated s.c. into athymic nude mice in
order to establish HNSCC xenografts. Two of the cell lines, SCC-026 and -040,
formed measurable tumors five days after inoculation (Suppl. Fig. S4A). Both
xenograft tumor models were found to be medium aggressive and all mice were
sacrificed within 35 or 40 days after inoculation for SCC-026 and SCC-040,
respectively. Tumors from both SCC-026 and SCC-040 xenograft models were found
to become darker and softer after reaching a size of approx. 900 and 400 mm3,
respectively (3-4 weeks after injection) (Suppl. Fig. S4B). Dissection showed that the
tumors were filled with fluid and histological examination of SCC-040 tumor
harvested 38 days after inoculation (tumor volume ~ 1000 mm3) confirmed a cystic
structure (Suppl. Fig. S4C). Our findings are in agreement with previous reports
describing cystic metastasis from HNSCC in patients. Histological evaluations 1, 2
and 3 weeks after inoculation identified the SCC-026 tumors as well differentiated,
keratinized squamous cell carcinoma (Suppl. Fig. S4D). No major cystic formations
were, however, detected by H&E staining of paraffin sections in SCC-026 xenografts
the first two weeks after inoculation. Antitumor effects following PCI of rGel/EGF on
SCC-026 xenografts were, therefore, measured 3-8 days post treatment (i.e. 7-12 days
after inoculation) to avoid any interference from cystic formation. Eight days post
treatment, initiating cyst formation was observed in all tumors, including the
untreated tumors, indicating that this is a characteristic of the tumor model and not a
result of the treatment (Fig. 6C). SCC-074 and SCC-099 did not form tumors in any
of the inoculated animals (Suppl. Fig. S4A).
SUPPLEMENTARY DISCUSSION
The rGel/EGF production
The objectives of this study were adapted according to the low yield and
concentration of rGel/EGF in the primary production. The rGel/EGF product was also
found to have reduced N-glycosidic activity compared to rGel itself. Several
approaches must therefore be considered in future studies to optimize the rGel/EGF
production. Firstly, a re-optimization of the codon choice to improve protein
expression in E. coli, as well as modifications of the experimental conditions during
protein expression (such as temperature, OD and IPTG concentration) should be
explored to increase the rGel/EGF yield. Secondly, the current configuration of
rGel/EGF may not be the optimal one, as the orientation may impact on the ratio of
inclusion bodies relative to soluble proteins formed in the production as well as on the
cytotoxicity profile. Also, the possibility that rGel domains responsible for enzymatic
activity are masked in the current orientation cannot be ruled out. Comparisons should
therefore be made with EGF fused to the N-terminal of rGel (EGF/rGel). Orientation
of the rGel/EGF construct is further expected to influence on the ability to form
trimers. No rGel/EGF trimers were identified after incorporating an isoleucine-zipper
trimer, which indicated misfolding of the protein, in accordance with the presence of
inclusion bodies in the rGel/EGF solution. The low molecular weight of the 42 kDa
monomer might present an obstacle for systemic delivery due to renal clearance and
future work should therefore include attempts to optimize the size (e.g. by
incorporating a peptide/protein linker), as well as to increase the purity, ribosome
inactivating activity and stock concentration of the final product.