Experimental Cell Research xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Experimental Cell Research

journal homepage: www.elsevier.com/locate/yexcr

Improved therapeutic efficacy of mammalian expressed-recombinant

interferon gamma against ovarian cancer cells
Ali Razaghia, Carina Villacrésb, Vincent Jungb, Narges Mashkoura, Michael Butlerb,
⁎
Leigh Owensa, Kirsten Heimanna,
a
Centre for Biodiscovery and Molecular Development of Therapeutics, James Cook University, Townsville QLD 4811, Australia
b
Department of Microbiology, University of Manitoba, Winnipeg, MB, Canada R3T 2N2

A R T I C L E I N F O A BS T RAC T

Keywords: Human interferon gamma (hIFNγ) affects tumour cells and modulates immune responses, showing promise as
Interferon gamma an anti-cancer biotherapeutic. This study investigated the effect of glycosylation and expression system of
Glycosylation recombinant hIFNγ in ovarian carcinoma cell lines, PEO1 and SKOV3. The efficacy of E. coli- and mammalian-
Ovarian cancer expressed hIFNγ (hIFNγ-CHO and HEK293, glycosylated/de-glycosylated) on cytostasis, cell death (MTT, and
SKOV3
Guava-ViaCount® flow-cytometry) and apoptotic signalling (Western blot of Cdk2, histone H3, procaspase-3,
FADD
Therapeutic efficacy
FADD, cleaved PARP, and caspase-3) was examined. Hydrophilic Interaction Liquid Chromatography
determined the structure of N-linked glycans present in HEK293-expressed hIFNγ (hIFNγ-HEK). PEO1 was
more sensitive to hIFNγ than SKOV3, but responses were dose-dependent and expression platform/glycosyla-
tion status-independent, whereas SKOV3 responded to mammalian-expressed hIFNγ in a dose-independent
manner, only. Complex-type oligosaccharides dominated the N-glycosylation pattern of hIFNγ-HEK with some
terminal sialylation and core fucosylation. Cleaved PARP and cleaved caspase-3 were not detected in either cell
line, but FADD was expressed in SKOV3 with levels increased following treatment. In conclusion, hIFNγ did not
induce apoptosis in either cell line. Mammalian- expressed hIFNγ increased cell death in the drug-resistant
SKOV3. The presence of FADD in SKOV3, which may inhibit apoptosis through activation of NF-κB, could serve
as a novel therapeutic target.

1. Introduction
Human interferon-gamma (hIFNγ) is a cytokine with immunomo-
dulatory properties, vital for innate and adaptive immunity against
viral/microbial infections and exhibits cytotoxic/cytostatic activity
against cancer cells. Native hIFNγ is naturally synthesised by CD4+ T
helper cell type 1 (Th1) lymphocytes, CD8+ cytotoxic lymphocytes and
natural killer (NK) cells [1,2].
The active form of native hIFNγ is a soluble homodimer with two N-
glycosylation sites on asparagines (N25 & N97) on the surface of each
dimer [1–3]. The glycans at N25 are fucosylated and are mainly
complex-type sialylated oligosaccharides with the sugar composition
of N-acetylneuraminic acid, galactose, mannose, N-acetylglucosamine,
and fucose. In contrast, the glycans at N97 are non-fucosylated hybrid
high-mannose structures with a sugar composition of N-acetylneur-
aminic acid, galactose, mannose, and N-acetylglucosamine [2,4].
Commercial hIFNγ is bacterially derived from Escherichia coli
(human interferon gamma-1b (hIFNγ-1b), tradename:
ACTIMMUNE®), approved for clinical treatment of chronic granulo-
matous disease and malignant osteopetrosis, with growing prospect for
cancer immunotherapy [2]. A short half-life stemming from a lack of
glycosylation limits the efficacy of hIFNγ-1b [5] and production is
costly due to the formation of inclusion bodies and endotoxin
contamination [6]. More cost-efficient production is being explored
in eukaryotic systems such as yeast (e.g. Pichia pastoris) [7], protozoa
(e.g. Leishmania sp.), and mammalian cells (e.g. Chinese hamster

Abbreviations: 2-AB, 2-aminobenzoic acid; 3D, three-dimensional structure; AU, arbitrary unit; BSA, bovine serum albumin; CHO, Chinese hamster ovary; hIFNγ-CHO, CHO-
expressed hIFNγ; deglyco-hIFNγ-HEK, deglycosylated HEK293-expressed hIFNγ; delgyco-hIFNγ-CHO, deglycosylated CHO-expressed hIFNγ; FBS, foetal bovine serum; hIFNγ-
HEK, HEK293-expressed hIFNγ; hIFNγ-R, hIFNγ receptor; HEK293, human embryonic kidney 293; hIFNγ-1b, human interferon gamma-1b;; hIFNγ, human interferon-gamma;
HILIC, hydrophilic interaction liquid chromatography; IRF-1, interferon regulatory factor-1; IL1β, interleukin 1 beta; NK, natural killer cells; NF-κB, nuclear factor kappa-light-
chain-enhancer of activated B cells; PNGase F, peptide N-glycosidase F; FADD, Fas-Associated death domain; PARP, poly-ADP-ribose polymerase; RIPK3, receptor-interacting
kinase 3; STAT1, signal transducer & activator of transcription-1; TNFα, tumour necrosis factor alpha
⁎
Corresponding author.
E-mail address: Kirsten.heimann@jcu.edu.au (K. Heimann).

http://dx.doi.org/10.1016/j.yexcr.2017.08.014
Received 17 May 2017; Received in revised form 5 August 2017; Accepted 8 August 2017
0014-4827/ © 2017 Elsevier Inc. All rights reserved.

Please cite this article as: Razaghi, A., Experimental Cell Research (2017), http://dx.doi.org/10.1016/j.yexcr.2017.08.014
A. Razaghi et al. Experimental Cell Research xxx (xxxx) xxx–xxx

ovary (CHO), human embryonic kidney 293 (HEK293), mice, rat) [2],
with best results to date in mammalian expression systems (higher
productivities, similarity to native hIFNγ in glycosylation and protein
folding) [2]. Differences exist in the N-glycan structure of hIFNγ
expressed in CHO; i.e. non-human-like sialylation potentially induces
immunogenicity and affects stability [2,8].
Glycan residues provide protease resistance, and the type of glycan
affects half-life (i.e. high mannose-type hIFNγ expressed in insect cells
had a shorter half-life) and pharmacokinetics of hIFNγ [3]. Detailed
knowledge of the effect of glycosylation on the therapeutic efficacy of
hIFNγ is limited, although glycosylation has been proven to be
essential for the therapeutic efficacy of protein drugs [9]. This study
explored the efficacy of glycosylation status of recombinant hIFNγ from
three different expression systems (E. coli, CHO and HEK293) against
the ovarian cancer cell lines, PEO1 and SKOV3, to evaluate whether or
not mammalian expressed recombinant hIFNγ can be a superior
substitute for the prokaryotic product.
Ovarian cancer is the 5th deadliest cancer in women and treatment
of ovarian cancer with hIFNγ-1b underwent phase I, II & III clinical
trials with mixed results. Some of the identified obstacles were tumour
insensitivity to hIFNγ and inability to deliver hIFNγ locally [10–12]. In
contrast, preclinical in vitro & in vivo studies on ovarian cancer cell
lines all showed a degree of sensitivity to hIFNγ-1b except for SKOV3
(Table 1).
Anti-cancer efficacy of hIFNγ in ovarian cancer cells was shown to
be due to cytostasis, through activation of p53 and p21 leading to cell
cycle arrest (Fig. 1) and/or cytotoxicity, causing cell death. The latter
was subdivided into pyroptosis through activation of interferon reg-
ulatory factor-1 (IRF-1) and caspase-1, and intrinsic apoptotic signal-
ling via activation of IRF-1, caspase-8, cytochrome-c release from
mitochondria, the caspase cascade and inhibition of poly-ADP-ribose
polymerase (PARP) (Fig. 1; Table 1). Treatment of OVCAR3 with
hIFNγ-1b led to both, apoptosis and cytostasis. Prior work on PEO1
also showed the possible engagement of both, cytostasis and apoptosis
[13,14], which had been suggested based on signal transducer &
activator of transcription-1 (STAT1) activation and p53 expression
[15]. Therefore, this research aimed to investigate other signalling
molecules (cleaved PARP and caspase-3) to provide more conclusive
evidence of hIFNγ-induced apoptosis. In contrast, high levels of
phosphorylated Fas-Associated Death Domain (FADD), a regulator of
cell cycle progression, proliferation, tumorigenesis and necroptosis
[16], increased activation of the anti-apoptotic transcription factor,
nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB),
a biomarker for aggressive phenotypes in lymphomas, lung and color-
ectal carcinomas (Fig. 1) [17,18]. hIFNγ, through induction of the Fas
pathway and export of FasL [19], can activate FADD (Fig. 1), which
may explain h-IFNγ resistance in cancer cell lines like SKOV3.
Therefore FADD levels following treatment with hIFNγ were also
investigated here in PEO1 and SKOV3.
2. Material & methods
2.1. Ovarian carcinoma cell lines & cultivation
Two ovarian carcinoma cell lines, PEO1 (passage № ≥ 40)
(Catalogue № 10032308) and SKOV3 (passage № ≥ 20) (Catalogue
№ 91091004), were purchased from the European Collection of
Authenticated Cell Cultures, UK. Cell lines were maintained at 37 °C
in a humidified growth chamber in 5% CO2 in RPMI-1640 medium
supplemented with L-glutamine and sodium bicarbonate (Sigma,
R8758, Castle Hill, NSW 1765, Australia) and addition of 10% foetal
bovine serum (FBS) (Sigma, F4135; USA origin); penicillins
(100 U mL−1) and streptomycin (100 µg mL−1). Cells were subcultured
when confluent. For the passage, cells were detached through addition
of Gibco® Trypsin-EDTA (0.25%) solution (ThermoFisher Scientific,
Newstead, QLD 4006, Australia). Subsequently, trypsin was deacti-
vated by addition of FBS-containing complete medium. Subcultures
were seeded with 50 ×103 cells per vented T-25 flask (25 cm2, Orange
Scientific, Sydney, NSW, Australia). Culture medium was changed
every three days.
2.2. Recombinant hIFNγ
Products from three different protein expression platforms were
purchased; hIFNγ-1b from Sigma (SRP3058), CHO-expressed (hIFNγ-
CHO) from SinoBiologicals (11725-HNAS, Beijing, China) (Purity: >
92%) and HEK293-expressed (hIFNγ-HEK) from Acrobiosystems
(IFG-H4211, Newark, DE 19711, USA). All recombinant hIFNγ
products were provided as a lyophilized powder from PBS (pH 7.4)
with a purity > 92%, and were tested for endotoxin content (< 0.1
EU μg−1).

Table 1
Summary of preclinical treatments of ovarian cancer cell lines with hIFNγ-1b.

Cell Line Mechanism of Action Signalling Molecules Detected Ref.

2774 Apoptosis Nitric oxidea [40]

Pyroptosis Induction of IRF-1 & activation of caspase-1 [41]
HOC7 Apoptosis Nitric oxidea [40]
OAW42 Apoptosis PARP inhibition [14]
OVCAR3 Apoptosis Nitric oxidea [40]
Less sensitive Induction of IRF-1 [41]
Apoptosis PARP inhibition [14]
Cytostasis/Apoptosis Cell cycle arrest; G1, G2, and S phases; [42]
Induction of p53, Bax, and caspase-3b
Cytostasis Increasing p21 mRNA [15]
OVCAR4, OVCAR5 Cytostasis/Apoptosis PARP inhibition [14]
PA1 Pyroptosis Induction of IRF-1 & activation of caspase-1 [41]
PEO1 Apoptosis/Cytostasis Depolarization of mitochondrial membrane, release of cytochrome C & activation of caspase-9 [13]
Induction of caspase-8 & 9 [14]
Cytostasis Increasing STAT1 & [15]
p21 mRNA
PEO14, PEO16 Cytostasis/Apoptosis PARP inhibition [15]
SKOV3 Insensitive Initial expression of p21, IRF-1 mRNA detected. However, later the pattern of expression was reduced to the [14,15,41]
level of untreated cells.
SW626 Cytostasis/Apoptosis PARP inhibition [14]

a
Polytherapy of hIFNγ, interleukin 1 beta (IL1β), and tumour necrosis factor alpha (TNFα).
b
Polytherapy of hIFNγ plus TNFα.

2
A. Razaghi et al. Experimental Cell Research xxx (xxxx) xxx–xxx

Fig. 1. hIFnγ-induced signal transduction in ovarian carcinoma cells. Involvement of the FADD pathway is hypothetical. The figure has been composed based on information obtained
from [13–15,17–19,25,29,30,34,36,41,43–47]. Bax, Bcl-2-associated X protein, Bid, BH3 interacting domain death agonist, CASP1, 3, 7, 8, 9, Caspase1, 3, 7, 8, 9; CYT-C,
Cytochrome-C; c-FLIP, Cellular FLICE (FADD-like IL-1β-converting enzyme)-inhibitory protein; FADD, Fas-Associated Death Domain Protein; Fas, Cell surface death receptor;
FasL, Fas Ligand; GAS, Gamma interferon-activated sequence; IRF-1, Interferon-regulated factor-1; Jak, Janus kinase; NF-κB, Nuclear factor kappa-light-chain-enhancer of
activated B cells; PARP, Poly-(ADP-ribose) polymerase; STAT1, Signal transducer & activator of transcription-1; TRAIL, TNF-related apoptosis-inducing ligand; DR, Death receptor.

2.3. Deglycosylation
N-glycans are positioned on the surface of the hIFNγ dimmer [1,3]
and are thus readily available targets for enzymatic removal. hIFNγ-
CHO and hIFNγ-HEK were deglycosylated (delgyco-hIFNγ-CHO;
deglyco-hIFNγ-HEK293) using the GlycoProfile™ II Enzymatic In-
solution N-Deglycosylation Kit (Sigma, PP0201) following the manu-
facturer's instructions with the following modification: β-mercap-
toethanol, a denaturant, was omitted due to potential cytotoxic effects
on cancer cells and heat-inactivation (100 °C) of hIFNγ, as temperature
of 40–64 °C inactivates hIFNγ.
2.4. In-vitro treatments
50 ×103 cells were inoculated per T25-flask (n = 3) and incubated
for 24 h. 5 mL of complete RPMI-1640 medium supplemented with
100 ng mL−1 of each six treatment conditions (hIFNγ-1b, hIFNγ-CHO,
deglyco-hIFNγ-CHO, hIFNγ-HEK and deglyco-hIFNγ-HEK) and in-
cubated for 72 h. Cells cultured in the same medium but in the absence
of recombinant hIFNγs served as negative controls.
leading to the determination of total cell counts. ii) The “viability dye”
stains dead cells.
The cytotoxic effect was determined as a percentage based on the
division of total dead cells to estimated absolute cell number. The
percentile of the cytostatic effect of each treatment was calculated as
per equation 1 (Eq. (1)):
(Control[total cells]−Treatment[total cells])
Cytostasis(%) = ×100
Control[total cells] (1)
2.5.2. Cell cycle analysis
The percentile of cells in G0/G1, S, and G2 phases was determined
based on DNA content. Before treatment, cells were synchronised in G0
by culturing for 24 h in RPMI medium without FBS. Following
treatment, the Guava® Cell Cycle Reagent (Merck, 4500-0220) was
used following the manufacturer's protocol.
2.6. Protein extraction, determination & analysis

2.5. Cytotoxic & cytostatic measurements
A Guava® easyCyte 8HT Benchtop Flow Cytometer (Merck, 0500–
4008) was used for conducting assays as follows;
2.5.1. Cell counting & viability
Absolute cell counts, identification of dead cells, and viability were
determined using the Guava® ViaCount® Reagent (Merck, 4000-0040,
USA) on cell suspensions. This assay distinguishes viable cells from
dead ones based on the differential permeability of cell membranes to
the DNA-binding dyes used. i) The “nuclear dye” stains nucleated cells
Following detachment at confluency after 72 h treatment, cells were
washed with PBS then lysed by adding 0.5 mL ice-cold lysis buffer
(10 mM Tris-HCl, 100 mM NaCl, pH 8, 25 mM EDTA, pH 8, 0.5% SDS,
1 Protease inhibitor cocktail tablet (Boehringer, 1697498, Mannheim,
Germany) per 10 mL) and vigorous pipetting. Afterwards, lysates were
frozen at −20 °C overnight, before thawing and centrifugation at
12,000g for 5 min at 4 °C. The total protein content of the cell lysate
supernatant was determined spectrophotometrically (Enspire®
Multimode Plate Reader (PerkinElmer, Glen Waverly, VIC,
Australia)) using the Pierce™ BCA Protein Assay Kit (ThermoFisher,
23225) using bovine serum albumin (BSA) as a standard.

3
A. Razaghi et al.
2.6.1. SDS-PAGE & Western blot analysis
To denature samples; 20 µg of total protein was added to Laemmli
sample buffer (200 mM Tris-Cl, (pH 6.8), 400 mM dithiothreitol, 8%
SDS, 0.4% bromophenol blue, 40% glycerol) followed by heating at
95 °C for 5 min. Samples were loaded on 12% Mini-PROTEAN® TGX™
Precast Gels (Biorad, 4561044, Gladesville, NSW 2111, Australia).
Electrophoresis was conducted at 200 V for 30 min using 1X running
buffer (25 mM Tris, 192 mM glycine, and 0.1% SDS). Precision Plus
Protein™ Standard (Biorad, 1610373) was used to determine product
size. Proteins were transferred to PVDF membranes using Trans-Blot®
Turbo™ Mini PVDF Transfer Packs (Biorad, 1704156) and a Trans-
Blot® Turbo™ Transfer System (Biorad) using the manufacturer's 3-
min protocol. Blots were incubated in blocking buffer (5% BSA in Tris-
buffered saline with Tween 20 (TBST) buffer; 20 mM Tris base, pH 7.5,
150 mM NaCl, 0.1% Tween 20) at room temperature with agitation for
1 h. Antibodies (see below) were diluted in blocking buffer. Blot
membranes were exposed to the primary antibody at 4 °C overnight
and to the secondary antibody with agitation at room temperature for
1 h. Clarity Max™ Western ECL Substrate (Biorad, 1705062) and the
Syngene G: Box XRQ chemiluminescent system were used for detection
and imaging, respectively.
Densitometry using the Gel Analyser function in ImageJ (version
1.51i, USA) was used to quantitate relative intensity, as per equation 2
(Eq. (2)):
Relative intensity
Blot of interest [arbitrary unit ]
=
Corresponding blot of the loading control (β − actin )[arbitrary unit ]
(2)
2.6.2. Antibodies
Apoptosis Western Blot Cocktail kit (pro/p17-caspase 3, cleaved-
PARP, β-actin) (Abcam, ab136812, Melbourne, VIC, Australia) was
used with dilution for primary antibody cocktail 1:250 and 1:100 for
goat HRP-conjugated secondary antibody cocktail. Cell Cycle and
Apoptosis WB Cocktail kit (Cdk2 pTyr15 (pCdk), Histone H3 pSer10
(pHH3) /β-actin/PARP) (Abcam, ab139417) was used with dilution for
primary antibody cocktail 1:250 and 1:2500 for secondary antibody
cocktail. Anti-FADD antibody (Abcam, ab24533) was used with dilu-
tion for primary antibody 1:1000 and 1:100 for goat HRP-conjugated
secondary antibody.
2.7. Dose-response assay
Dose-response was evaluated using the CellTiter 96® Non-
Radioactive Cell Proliferation (MTT) Assay (Promega, G4000,
Alexandria, NSW 2015, Australia). Approximately 3000 cells were
inoculated per 100 μL volume of complete RPMI-1640 medium per
well of a 96-well assay plate in dilution series of hIFNγ-1b and hIFNγ-
HEK (0–500 ng mL−1). Following the manufacturer's protocol, absor-
bance was recorded spectrophotometrically (EnSpire®).
2.8. Microscopy
Cultured cells were imaged with an inverted digital microscope
(Olympus IX53, Japan) in T25 flasks after 72 h of incubation with
hIFNγ and untreated controls using phase contrast at 100× magnifica-
tion.
2.9. Glycan analysis
The structure of N-linked glycans was determined following Royle,
Radcliffe, Dwek and Rudd [20]. The hIFNγ-HEK sample was recon-
stituted in 300 μL of Milli-Q water and denatured with 15 μL of beta-
mercaptoethanol for 5 min at 95 °C for analysis by gradient (8–16%)
4
Experimental Cell Research xxx (xxxx) xxx–xxx
SDS-PAGE gels (Biorad, Hercules, USA) stained with Coomassie
brilliant blue. Protein bands were finally chopped, washed and treated
overnight with peptide N-glycosidase F (PNGase F) (Roche
Diagnostics, Indianapolis, USA) to release glycans. The released N-
glycans were labelled with 2-aminobenzoic acid (2-AB).
2.9.1. HPLC profiling of 2-AB- labelled N-glycans
Previously 2-AB-labelled N-glycans were analysed by hydrophilic
interaction liquid chromatography (HILIC) with fluorescence detection
using an amide HPLC Waters X-Bridge column (4.6 × 250 mm;
3.5 µm) and a constant flow rate of 0.86 mL min−1. The buffers used
were 50 mM ammonium formate (pH 4.4, buffer A) and acetonitrile
(ACN, buffer B) with a linear gradient of 20 – 50% applied over 48 min,
followed by an increase from 50% to 100%. Glycan structures were
assigned using Empower Software (Waters™, USA) where GU values
were obtained by using a 2-AB labelled dextran ladder standard and
compared with NIBRT glycan database (https://glycobase.nibrt.ie).
2.10. Statistics
Data (n = 3 replicates) were analysed via one-way ANOVA using S-
PLUS® software, version 8.0 with α set to 0.01. Homogeneity of
variance and normality was ascertained using the Leven's and the
Shapiro-Wilks tests, respectively. Tukey-Kramer HSD posthoc tests
were performed to determine the source of significance at p < 0.01.
3. Results
3.1. Effects of recombinant hIFNγ on PEO1
Treatments with all recombinant hIFNγ products caused 60–65%
dead cells through cytotoxicity as determined by ViaCount® (Fig. 2A).
Statistical analyses confirmed a significant effect of treatment com-
pared to controls, but there was no statistically significant effect (p <
0.01) between treatments (i.e. expression source or glycosylation
status) (Fig. 2A). Neither cleaved PARP nor cleaved caspase-3 was
detected in Western blots, however, a slight reduction in relative
intensity of procaspase-3 expression (an inactive zymogen of caspase
3) was observed after treatment with all recombinant hIFNγ products,
as quantified by Image-J as a ratio of 0.9–1.1, compared to a ratio of
1.3 in controls (Fig. 3D). Expression of FADD was not detected in
PEO1 cells.
Based on the number of PEO1 cells, all recombinant hIFNγ
products caused 54–57% of cytostasis with no significant difference
between treatments; except, a slightly stronger and statistically sig-
nificant effect ~62% (p < 0.01) observed in cells treated with hIFNγ-
HEK (Fig. 2B).
Treatment with all recombinant hIFNγ products, irrespective of
source and glycosylation, affected the cell cycle of PEO1 shown by an
increase of cells in G2 phase, 28–30%, compared to controls, ~22% (p
< 0.01) (Fig. 2C). Western blot analysis confirmed reduced relative
intensities of Cdk2 expression, as quantified by Image-J as a ratio of
0.5–0.7, compared to a ratio of 1.1 in controls (Fig. 3E). Relative
intensities of Histone H3, the M-phase indicator, were comparable
between controls and treatments (Fig. 3E).
The growth of PEO1 was strongly inhibited in a dose-dependent but
source-independent manner following treatment with HIFNγ-1b and
hIFNγ-HEK (Fig. 2E). The calculated IC50 was ~200 ng mL−1 for both
(Fig. 2E).
3.2. Effects of recombinant hIFNγ on SKOV3
Effect of treatment on induction of cell death was significant (p <
0.01) with cytotoxic efficacy declining: hIFNγCHO (~28%) =
hIFNγHEK (~24%) = deglycol-hIFNγHEK (~23%) > deglycol-
hIFNγCHO (~17%) ≥ hIFNγ1b (~11%) ≈ control (~5%) (Fig. 2A).
A. Razaghi et al. Experimental Cell Research xxx (xxxx) xxx–xxx

Fig. 2. 72 h-treatment effects of recombinant hIFNγ from three different expression systems and their deglycosylated forms on PEO1 and SKOV3. (A) Cytotoxic effect (B) cytostatic
effect (calculated as per Eq. (1)). Cell cycle analysis of (C) PEO1 and (D) SKOV3. 48 h-dose-response to treatment with hIFNγ-1b and hIFNγ-HEK on the growth of (E) PEO1 and (F)
SOKV3. For dose-response effect; growth is expressed as a fraction of control values (Mean ± SD, n = 3), p < 0.01.

Procaspase-3 was detected, but treatments had no effect on relative
expression intensity (Fig. 3A), cleaved PARP and caspase-3 were not
detected. Western blotting detected FADD expression in all treatments
and controls. Notably, relative expression was increased after treat-
ment, as quantified by Image-J as a ratio of 0.9–2.1, compared to a
ratio of 0.7 in controls; with highest levels being observed after
treatment with hIFNγ-CHO (ratio of 2.1) (Fig. 3C).
Cell cycle analysis of SKOV 3 cells showed a significant percent of
cells arrested in S-phase following treatment with all recombinant
hIFNγ products, 68–70%, compared to controls (~31%) (Fig. 2D) (p <
0.01). Source of hIFNγ or glycosylation status had no significant
impact. The largest cytostatic effects were observed for hIFNγ-
HEK293 (glycosylated ~43% and deglycosylated ~44%), hIFNγ-CHO
(glycosylated ~44%), deglyco-hIFNγ-CHO (~31%) was less effective,

5
A. Razaghi et al. Experimental Cell Research xxx (xxxx) xxx–xxx

Fig. 3. Western blot analysis of recombinant hIFNγ-induced signalling molecules in SKOV3 and PEO1. (A, D) procaspase-3, (C) FADD, (B, E) Cdk2 (as an indication of G1/S phase),
Histone H3 (as a biomarker of M phase), Untreated cells (controls) were used to determine un-induced signalling molecule levels and β-actin, a housekeeping protein, was used as a
loading control to obtain relative intensity histograms with Image-J (Eq. (2)). * C had the same loading controls as B.

and hIFNγ-1b (~24%) was the least effective (Fig. 2B) (p < 0.01).
Western blot analysis confirmed an increase in relative expression of
Cdk2 after treatment, as quantified by Image-J as a ratio of 1–1.6,
compared to a ratio of 0.7 in controls. In contrast, relative expression of
Histone H3 was reduced in treatments, as quantified by Image-J as a
ratio of 0.2–0.4 compared to a ratio of 0.6 in controls (Fig. 3B).
The growth of SKOV3 was inhibited in a dose-independent but
source-dependent manner following treatment with HIFNγ-1b and
hIFNγ-HEK (Fig. 2F). Treatment with 15–500 ng mL−1 inhibited
growth by approximately 15% and 25% for hIFNγ-1b and hIFNγ-
HEK, respectively. Since no dose response was observed, an IC50 of
hIFNγ could not be calculated for this cell line.
Furthermore, SKOV3 cells exposed to hIFNγ-CHO, deglyco-hIFNγ-
CHO, hIFNγ-HEK, and deglyco-hIFNγ-HEK showed changes in cell
morphology (Fig. 4), with cells becoming thin and elongated more
frequently, while such alterations were not observed for PEO1.
3.3. Glycan analysis
The glycan analysis of hIFNγ-HEK showed that the protein is
heterogeneously glycosylated containing bi-, tri-, and tetra-antennary
structures dominated by complex-type oligosaccharides, many of which

6
A. Razaghi et al. Experimental Cell Research xxx (xxxx) xxx–xxx

Fig. 4. Recombinant hIFNγ-induced cell elongation in SKOV3 cells. Phase contrast micrographs of SKOV3 cells with elongated thin shapes after 72 h treatment. (A) Control (B)
hIFNγ-1b (C) hIFNγ-CHO (D) deglyco-hIFNγ-CHO (E) hIFNγ-HEK (F) deglyco-hIFNγ-HEK. White arrows point to elongated thin-shaped cells.

Fig. 5. NP-HPLC profile of 2-AB-labelled hIFNγ-HEK N-glycans.

contained terminal sialylation and core fucosylation. Mannose, galac- 4. Discussion

tose, fucose and sialic acid were major constituents of the sugar
composition (Fig. 5). High heterogeneity in cancer cells regarding genomic mutations,
DNA copy number, oncogene bypass, and epigenetic alterations is the

7
A. Razaghi et al.
leading cause of inconsistent responses to chemotherapy and drug-
resistance [21]. This can explain the observed differences in cytotoxi-
city, cytostasis, IC50 and therapeutic efficacy of the recombinant
hIFNγs used here in PEO1 and SKOV3.
4.1. hIFNγ efficacy in PEO1
PEO1 was highly susceptible to treatment with hIFNγ regardless of
expression platform and glycosylation. Although key apoptotic signal-
ling molecules (cleaved PARP and caspase-3) were not detected by
Western blot, levels of procaspase-3 decreased slightly. Caspase-3 is an
executioner of apoptosis through the proteolytic processing of down-
stream proteins such as PARP. Caspase-3 (active p17 and p12
subunits) is activated through cleavage of procaspase-3, which is a
zymogen (inactive precursor). Thus, monitoring the altered proportion
of procaspase-3 to activated/cleaved caspase-3, i.e. a decrease of the
procaspase-3 or an increase of caspase 3, can be used as a biomarker
for apoptosis [22]. These results seem to oppose apoptotic signal
transduction as a consequence of hIFNγ treatment in PEO1, contrary
to previous reports based on flow-cytometric measurements of induced
levels of caspase 8 and 9 and Western blot detection of cleaved PARP,
caspase-8 and 9 [13,14]. Some differences in cell numbers (not given
vs. 10,000 cells mL−1), hIFNγ concentrations used (≈ 250 vs.
100 ng mL−1), treatment duration (96 vs. 72 h) and particularly
confluence status of the culture (100% vs. 40% used here) and passage
number between their and our study limits comparability. Apoptosis in
in vitro culture conditions has been reported to be strongly cell density-
dependent e.g. high cell density induced apoptosis in an autocrine
manner following the release of an unknown low molecular weight
peptide-containing factor in leukaemia HL-60 cells [23]. In contrast, a
high passage number has been implicated in causing resistance to
apoptosis [24], which is a possible explanation why apoptosis was not
observed in this study. Since cytostatic effects and mechanisms of cell
death of hIFNγ were not comparable with previous publications on
PEO1, there is an urgent need for standardisation of toxicity assays
(e.g. confluency, passage number, duration of treatment and use of
bioactivity units), especially when aiming to unravel signalling cascades
and treatment-induced cell death mechanisms.
Necroptotic cell death in PEO1 appears unlikely since activation of
FADD is essential for induction of necroptosis [25]. Autophagy remains
a possibility for an hIFNγ-induced cell death mechanism in this high
passage number PEO1 cell line, as hIFNγ has been shown to induce
autophagy signalling pathways with growth inhibition and cell death in
other cancer cell lines like hepatocellular carcinoma [26]. Thus, it is
recommended to investigate the presence of key autophagic cell death
proteins e.g. LC3 (microtubule-associated protein 1 light chain 3),
Beclin-1 or Atg5 (Autophagy protein 5) proteins [26].
Treatment with recombinant hIFNγ led to cytostasis caused by cell
cycle arrest in G2 phase, correlating well with previous reports on
hIFNγ-1b-induced elevated levels of p21 mRNA and cytostasis [15].
Decreased levels of the Cdk2 protein in treatments correlate with lower
cell numbers in G1/S phases in comparison to controls [27]. Low
histone H3 levels in controls and treatments may indicate a very short
M phase.
hIFNγ-1b dose response experiments on growth and cell death in
ovarian cancers show high variability depending on cell lines used [14].
For example, reported IC50s were 10 units mL−1 ( = 0.5 ng mL−1) and
> 5000 units mL−1 (> 250 ng mL−1) for OVCAR4 and SW626, respec-
tively. This study is the first examining dose-responses for SKOV3 and
PEO1, which were dose-independent and dose-dependent, respec-
tively. Comparisons of IC50s between our study and previous studies
is hampered due to significant differences in experimental conditions
(e.g. cell density, the difference in cell lines, and duration of treatment,
see above).
8
Experimental Cell Research xxx (xxxx) xxx–xxx
4.2. hIFNγ efficacy in SKOV3
SKOV3 has been reported as insensitive to treatment with hIFNγ-
1b [14]. However, this study showed a degree of sensitivity to
mammalian expressed hIFNγ. Deglycosylation had a marginal impact
on the cytotoxic efficacy but a significant impact on the cytostatic
efficacy in CHO-, but not HEK-expressed hIFNγ. This suggests that
other factors like the proper folding of the native protein in mammalian
expression systems are as important as glycosylation for drug potency
and pharmacodynamics [28].
Signalling molecules typically involved in apoptosis (cleaved PARP
and caspase3) were not detected by Western blot analysis, ruling out
apoptosis as the principal mechanism of cell death. Although FADD
was present in SKOV3 at detectable levels in controls and treatments,
FADD is unlikely an initiator of the extrinsic apoptotic pathway, as no
cleaved PARP and caspase-3 were detected. Overall cell death was low
in SKOV3, which together with the presence of FADD in control cells,
may indicate that FADD inhibits apoptosis through the activation of
NF-κB (Fig. 1) [17] and autophagy [29]. FADD-induced necroptotic
cell death via mitochondrial dysfunction requires further investigation
on increased levels of proteins participating in the formation of
complex II (Fig. 1) [25]. Of all cell death mechanisms, necroptosis is
the most probable in SKOV 3, an interpretation supported by hIFNγ-
induced cell death mechanisms in other cancer types [16,30], but direct
evidence via detection of the core signalling molecule, receptor-
interacting kinase 3 (RIPK3) [25], is required.
FADD levels in SKOV3 cells could potentially correlate with
tumorigenesis, anti-apoptotic behaviour, and resistance against antic-
ancer drugs, which has also been reported for lymphomas, lung and
colorectal carcinomas [18]. This could make FADD a novel target to
overcome drug-resistance in this ovarian cancer cell line (Fig. 1) [17].
In SKOV3, all three types of hIFNγ caused cytostasis through cell cycle
arrest in S phase accompanied by an increase in Cdk2 and a decrease in
Histone 3. This result is supported by previous studies showing hIFNγ-1b-
induced elevated levels of p21 - and IRF-1 mRNA in SKOV3 [15]. Other
cancer types sensitive to hIFNγ-1b also showed a similar trend, e.g. the
induction of cell cycle arrest in human mesothelioma with hIFNγ-1b
seems to depend on cyclin regulation through p21WAF1/CIP1 and p27Kip1-
independent mechanisms [31] and relies on p53 and p21WAF1/clp1/Sdl1
pathways in hepatocytes [32] (Fig. 1).
4.3. Differences in responses of PEO1 & SKOV3 to treatment with
hIFNγ
In general, comparison of cytostasis following treatment with
recombinant hIFNγ showed that PEO1 was 10–30% more sensitive
than SKOV3 (Fig. 2B). Furthermore, cytotoxicity of hIFNγ was greater
(~70% dead cells) compared to SKOV3 (max. ~30% dead cells)
(Fig. 2A).
Treatment with mammalian-expressed hIFNγ induced higher levels
of cell death compared to untreated controls and hIFNγ-1b treatment
only in SKOV3. While glycosylation can affect the three-dimensional
structure (3D) and binding affinity of glycoproteins in general [33],
deglycosylation possibly did not influence the 3D structure of hIFNγ,
since the N-glycosylation sites are on the external surface of the
homodimer, explaining the no difference in efficacy in treatments with
glycosylated and deglycosylated hIFNγ-HEK in SKOV3. In contrast,
deglycosylation of hIFNγ-CHO decreased cell death. This contrasts
with the observation of hIFNγ-induced cell death in PEO1, where
source and glycosylation did not result in different outcomes. This
seems to indicate that the hIFNγ receptor (hIFNγ-R) in PEO1 and
SKOV3 may be different in their ability to bind hIFNγ. The SKOV3
hIFNγ-R could bind de-glycosylated/glycosylated hIFNγ better than
bacterial unglycosylated hIFNγ-1b, if the latter has a different 3D
structure, while the hIFNγ-R interaction in PEO1 cells could be more
tolerant of differences in 3D structure. This hypothesis is supported by
A. Razaghi et al.
studies that show mutations in hIFNγ-Rs affect binding to the receptor
[19,34–36]. However, it may be possible that deglycosylation efficiency
could have been different for the two sources of hIFNγ, which could be
validated via SDS-PAGE in future studies. Future detailed studies on
amino acid sequence differences in the receptors and crystallography to
unravel changes in the 3D structure of hIFNγ forms and binding sites
of the receptor are also required. Furthermore, induction of FADD in
SKOV3, but not PEO1, could indicate hIFNγ-initiated Fas-signalling in
general with increasing efficacy by the mammalian forms, especially the
CHO-derived glycosylated form of hIFNγ. This conclusion is supported
by regulation of Fas-ligand expression via expression of hIFNγ in
human CD4+ T-lymphocytes [19].
4.4. Glycan analysis
In CHO, N-glycosylation of hIFNγ has been reported to carry
terminal Gal-α1-3-Gal and to be highly sialylated [8]. Additionally,
the composition of the culture medium and production platform have
been shown to influence the heterogeneity of N-glycosylation (of both
N-, one N-site and non-glycosylated) [37,38]. While the N-glycan
oligosaccharide structure has been examined in hIFNγ-CHO, reporting
high levels of sialylation [39], our study is the first to report on the N-
glycan structure in hIFNγ-HEK, which validates earlier reports on the
absence of terminal Gal-α1-3-Gal and low levels of sialylation in
human-expressed hIFNγ.
In summary, hIFNγ is currently undergoing clinical trials, however
knowledge on how different sources of recombinant hIFNγ, with diver-
gent glycosylation profiles, affect cell proliferation, death and signalling is
limited. This study also investigated effects of differences in glycosylation
on the efficacy of recombinant hIFNγ by comparing the efficacy of
recombinant hIFNγ from two mammalian expression systems (CHO
and HEK293) in their glycosylated and deglycosylated forms against
hIFNγ-1b for the potential treatment of ovarian cancers, using a sensitive
cell line (PEO1) and a resistant cell line (SKOV3). As all recombinant
hIFNγ products had a purity > 92%, and were tested for endotoxin
content (< 0.1 EU μg−1), similar bioactivities were assumed; however, a
contribution to the observed effects due vendor-specific differences in
purification methods cannot be completely ruled out.
5. Conclusion
Through investigation of potential apoptotic signalling pathways, a
new potential drug target was discovered in SKOV3, FADD, being
expressed in SKOV3 but not in PEO1. Fas-associated signalling could
be responsible for treatment-resistance of this form of ovarian cancer.
Overall, based on cytostatic and cytotoxic results achieved in SKOV3,
hIFNγ expressed in HEK293, but not E. coli shows promise for the
treatment of tumorigenic and anti-cancer drug-resistant ovarian can-
cers. Based on these promising results, we suggest that a continuum of
preclinical research is undertaken to investigate responsivity of more
ovarian cancer cell lines, other cancer types and recombinant glyco-
proteins to generalise the idea that mammalian/human expression
platforms can enhance the therapeutic efficacy of anti-cancer biother-
apeutics.
Conflict of interest
All authors declare no conflict of interest. The funders had no role
in study design, data collection and analysis, decision to publish or
preparation of the manuscript.
Ethical approval
This article does not contain any studies with human participants or
animals performed by any of the authors.
9
Experimental Cell Research xxx (xxxx) xxx–xxx
Acknowledgment
The authors wish to acknowledge the provision of a James Cook
University Postgraduate Research Scholarship (JCUPRS) to Ali
Razaghi. The study was financed by the Higher Degree Research
Enhancement Scheme (HDR ES) grant allocated to AR. We acknowl-
edge A/Prof. Patrick Schaeffer for his assistance with protein analysis.
References
[1] Ha Young, K.J. Hardy, Role of interferon-y in immune cell regulation, J. Leukoc.
Biol. 58 (1995) 373–381.
[2] A. Razaghi, L. Owens, K. Heimann, Review of the recombinant human interferon
gamma as an immunotherapeutic: impacts of production platforms and glycosy-
lation, J. Biotechnol. 240 (2016) 48–60.
[3] T. Sareneva, K. Cantell, L. Pyhala, J. Pirhonen, I. Julkunen, Effect of carbohydrates
on the pharmacokinetics of human interferon-gamma, J. Interferon Res. 13 (1993)
267–269.
[4] T. Sareneva, J. Pirhonen, K. Cantell, I. Julkunen, N-glycosylation of human
interferon-y: glycans at Asn-25 are critical for protease resistance, Biochem. J. 308
(1995) 9–14.
[5] N. Miyakawa, M. Nishikawa, Y. Takahashi, M. Ando, M. Misaka, Y. Watanabe,
Y. Takakura, Prolonged circulation half-life of interferon gamma activity by gene
delivery of interferon gamma-serum albumin fusion protein in mice, J. Pharm. Sci.
100 (2011) 2350–2357.
[6] A. Razaghi, E. Tan, L.H. Lua, L. Owens, O.P. Karthikeyan, K. Heimann, Is Pichia
pastoris a realistic platform for industrial production of recombinant human
interferon gamma?, Biologicals 45 (2017) 52–60.
[7] A. Razaghi, R. Huerlimann, L. Owens, K. Heimann, Increased expression and
secretion of recombinant hIFNγ through amino acid starvation-induced selective
pressure on the adjacent HIS4 gene in Pichia pastoris, Eur. Pharm. J. 62 (2015)
43–50.
[8] D. Ghaderi, M. Zhang, N. Hurtado-Ziola, A. Varki, Production platforms for
biotherapeutic glycoproteins. Occurrence, impact, and challenges of non-human
sialylation, Biotechnol. Genet. Eng. Rev. 28 (2012) 147–176.
[9] R.J. Sola, K. Griebenow, Glycosylation of therapeutic proteins: an effective strategy
to optimize efficacy, BioDrugs 24 (2010) 9–21.
[10] G.P. Dunn, C.M. Koebel, R.D. Schreiber, Interferons, immunity and cancer
immunoediting, Nat. Rev. Immunol. 6 (2006) 836–848.
[11] G.H. Windbichler, H. Hausmaninger, W. Stummvoll, A.H. Graf, C. Kainz,
J. Lahodny, U. Denison, E. Muller-Holzner, C. Marth, Interferon-gamma in the
first-line therapy of ovarian cancer: a randomized phase III trial, Br. J. Cancer 82
(2000) 1138–1144.
[12] D.S. Alberts, C. Marth, R.D. Alvarez, G. Johnson, M. Bidzinski, D.R. Kardatzke,
W.Z. Bradford, J. Loutit, D.H. Kirn, M.C. Clouser, M. Markman,
G.C.T. Consortium, Randomized phase 3 trial of interferon gamma-1b plus
standard carboplatin/paclitaxel versus carboplatin/paclitaxel alone for first-line
treatment of advanced ovarian and primary peritoneal carcinomas: results from a
prospectively designed analysis of progression-free survival, Gynecol. Oncol. 109
(2008) 174–181.
[13] C. Barton, D. Davies, F. Balkwill, F. Burke, Involvement of both intrinsic and
extrinsic pathways in IFN-gamma-induced apoptosis that are enhanced with
cisplatin, Eur. J. Cancer 41 (2005) 1474–1486.
[14] L. Wall, F. Burke, C. Barton, J. Smyth, F. Balkwill, IFN-gamma induces apoptosis in
ovarian cancer cells in vivo and in vitro, Clin. Cancer Res. 9 (2003) 2487–2496.
[15] F. Burke, P.D. Smith, M.R. Crompton, C. Upton, F.R. Balkwill, Cytotoxic response
of ovarian cancer cell lines to IFN-gamma is associated with sustained induction of
IRF-1 and p21 mRNA, Br. J. Cancer 80 (1999) 1236–1244.
[16] L. Tourneur, G. Chiocchia, FADD: a regulator of life and death, Trends Immunol.
31 (2010) 260–269.
[17] K.A. Schinske, S. Nyati, A.P. Khan, T.M. Williams, T.D. Johnson, B.D. Ross,
R.P. Tomas, A. Rehemtulla, A novel kinase inhibitor of FADD phosphorylation
chemosensitizes through the inhibition of NF-kappaB, Mol. Cancer Ther. 10 (2011)
1807–1817.
[18] S. Patel, D. Murphy, E. Haralambieva, Z.A. Abdulla, K.K. Wong, H. Chen, E. Gould,
G. Roncador, C. Hatton, A.P. Anderson, A.H. Banham, K. Pulford, Increased
expression of phosphorylated FADD in anaplastic large cell and other T-Cell
lymphomas, Biomark. Insights 9 (2014) 77–84.
[19] D. Boselli, G. Losana, P. Bernabei, D. Bosisio, P. Drysdale, R. Kiessling, J.S. Gaston,
D. Lammas, J.L. Casanova, D.S. Kumararatne, F. Novelli, IFN-gamma regulates Fas
ligand expression in human CD4+ T lymphocytes and controls their anti-myco-
bacterial cytotoxic functions, Eur. J. Immunol. 37 (2007) 2196–2204.
[20] L. Royle, C.M. Radcliffe, R.A. Dwek, P.M. Rudd, Detailed structural analysis of N-
glycans released from glycoproteins in SDS-PAGE gel bands using HPLC combined
with exoglycosidase array digestions, Methods Mol. Biol. 347 (2006) 125–143.
[21] X. Hu, Z. Zhang, Understanding the genetic mechanisms of cancer drug resistance
using genomic approaches, Trends Genet. 32 (2016) 127–137.
[22] K.S. Putt, G.W. Chen, J.M. Pearson, J.S. Sandhorst, M.S. Hoagland, J.T. Kwon,
S.K. Hwang, H. Jin, M.I. Churchwell, M.H. Cho, D.R. Doerge, W.G. Helferich,
P.J. Hergenrother, Small-molecule activation of procaspase-3 to caspase-3 as a
personalized anticancer strategy, Nat. Chem. Biol. 2 (2006) 543–550.
[23] K. Saeki, A. Yuo, M. Kato, K. Miyazono, Y. Yazaki, F. Takaku, Cell density-
dependent apoptosis in HL-60 cells, which is mediated by an unknown soluble
A. Razaghi et al.
factor, is inhibited by transforming growth factor 1 and overexpression of Bcl-2, J.
Biol. Chem. 272 (1997) 20003–20010.
[24] L. Pronsato, A. La Colla, A.C. Ronda, L. Milanesi, R. Boland, A. Vasconsuelo, High
passage numbers induce resistance to apoptosis in C2C12 muscle cells, Biocell: Off.
J. Soc. Latinoam. De. Microsc. Electron. … et al. 37 (2013) 1–9.
[25] M. Pasparakis, P. Vandenabeele, Necroptosis and its role in inflammation, Nature
517 (2015) 311–320.
[26] P. Li, Q. Du, Z. Cao, Z. Guo, J. Evankovich, W. Yan, Y. Chang, L. Shao, D.B. Stolz,
A. Tsung, D.A. Geller, Interferon-gamma induces autophagy with growth inhibition
and cell death in human hepatocellular carcinoma (HCC) cells through interferon-
regulatory factor-1 (IRF-1), Cancer Lett. 314 (2012) 213–222.
[27] Q. Liu, X. Liu, J. Gao, X. Shi, X. Hu, S. Wang, Y. Luo, Overexpression of DOC-1R
inhibits cell cycle G1/S transition by repressing CDK2 expression and activation,
Int. J. Biol. Sci. 9 (2013) 541–549.
[28] S. Frokjaer, D.E. Otzen, Protein drug stability: a formulation challenge, Nat. Rev.
Drug Discov. 4 (2005) 298–306.
[29] B.D. Bell, S. Leverrier, B.M. Weist, R.H. Newton, A.F. Arechiga, K.A. Luhrs,
N.S. Morrissette, C.M. Walsh, FADD and caspase-8 control the outcome of
autophagic signaling in proliferating T cells, Proc. Natl. Acad. Sci. 105 (2008)
16677–16682.
[30] R.J. Thapa, S.H. Basagoudanavar, S. Nogusa, K. Irrinki, K. Mallilankaraman,
M.J. Slifker, A.A. Beg, M. Madesh, S. Balachandran, NF-kappaB protects cells from
gamma interferon-induced RIP1-dependent necroptosis, Mol. Cell Biol. 31 (2011)
2934–2946.
[31] C. Vivo, F. Levy, Y. Pilatte, J. Fleury-Feith, P. Chretien, I. Monnet, L. Kheuang,
M.C. Jaurand, Control of cell cycle progression in human mesothelioma cells
treated with gamma interferon, Oncogene 20 (2001) 1085–1093.
[32] A. Kano, Y. Watanabe, N. Takeda, T.S.-i. Aizawa, Analysis of IFN-γ-induced cell
cycle arrest and cell death in hepatocytes, J. Biochem. 121 (1997) 677–683.
[33] X. Huang, J.J. Barchi Jr., F.D. Lung, P.P. Roller, P.L. Nara, J. Muschik,
R.R. Garrity, Glycosylation affects both the three-dimensional structure and
antibody binding properties of the HIV-1IIIB GP120 peptide RP135, Biochemistry
36 (1997) 10846–10856.
[34] X. Xu, X.Y. Fu, J. Plate, A.S. Chong, IFN-gamma induces cell growth inhibition by
Fas-mediated apoptosis: requirement of STAT1 protein for up-regulation of Fas
and FasL expression, Cancer Res. 58 (1998) 2832–2837.
[35] D. Rosner, V. Stoneman, T. Littlewood, N. McCarthy, N. Figg, Y. Wang, G. Tellides,
M. Bennett, Interferon-gamma induces Fas trafficking and sensitization to apop-
tosis in vascular smooth muscle cells via a PI3K- and Akt-dependent mechanism,
10
Experimental Cell Research xxx (xxxx) xxx–xxx
Am. J. Pathol. 168 (2006) 2054–2063.
[36] M.D.S. Jean, Fi Brignole, G. Feldmann, A. Goguel, C. Baudouin, Interferon-γ
induces apoptosis and expression of inflammation-related proteins in Chang
conjunctival cells, Invest Ophthalmol. Vis. Sci. 40 (1999) 2199–2212.
[37] C. Pm, I. Ap, H. Pm, The macroheterogeneity of recombinant human interferon‐
gamma produced by Chinese‐hamster ovary cells is affected by the protein and lipid
content of the culture medium, Biotechnol. Appl. Biochem. 21 (1995) 87–100.
[38] N. Jenkins, P. Castro, S. Menon, A. Ison, A. Bull, Effect of lipid supplements on the
production and glycosylation of recombinant interferon-gamma expressed in CHO
cells, Cytotechnology 15 (1994) 209–215.
[39] D.C. James, R.B. Freedman, M. Hoare, O.W. Ogonah, B.C. Rooney, O.A. Larionov,
V.N. Dobrovolsky, O.V. Lagutin, N. Jenkins, N-glycosylation of recombinant human
interferon-y produced in different animal expression systems, Biotechnology 13
(1995) 64771.
[40] J. Rieder, R. Jahnke, M. Schloesser, M. Seibel, M. Czechowski, C. Marth,
G. Hoffmann, Nitric oxide-dependent apoptosis in ovarian carcinoma cell lines,
Gynecol. Oncol. 82 (2001) 172–176.
[41] E.J. Kim, J.M. Lee, S.E. Namkoong, S.J. Um, J.S. Park, Interferon regulatory
factor-1 mediates interferon-gamma-induced apoptosis in ovarian carcinoma cells,
J. Cell Biochem. 85 (2002) 369–380.
[42] Y.-Q. Guan, Z. Zheng, L. Liang, Z. Li, L. Zhang, J. Du, J.-M. Liu, The apoptosis of
OVCAR-3 induced by TNF-α plus IFN-γ co-immobilized polylactic acid copolymers,
J. Mater. Chem. 22 (2012) 14746.
[43] E.C. Alappat, C. Feig, B. Boyerinas, J. Volkland, M. Samuels, A.E. Murmann,
A. Thorburn, V.J. Kidd, C.A. Slaughter, S.L. Osborn, A. Winoto, W.J. Tang,
M.E. Peter, Phosphorylation of FADD at serine 194 by CKIalpha regulates its
nonapoptotic activities, Mol. Cell 19 (2005) 321–332.
[44] E.W. Lee, J. Seo, M. Jeong, S. Lee, J. Song, The roles of FADD in extrinsic apoptosis
and necroptosis, BMB Rep. 45 (2012) 496–508.
[45] L.C. Li, S. Jayaram, L. Ganesh, L. Qian, J. Rotmensch, A.V. Maker, B.S. Prabhakar,
Knockdown of MADD and c-FLIP overcomes resistance to TRAIL-induced apop-
tosis in ovarian cancer cells, Am. J. Obstet. Gynecol. 205 (362) (2011) e312–e325.
[46] J.O. Pyo, M.H. Jang, Y.K. Kwon, H.J. Lee, J.I. Jun, H.N. Woo, D.H. Cho, B. Choi,
H. Lee, J.H. Kim, N. Mizushima, Y. Oshumi, Y.K. Jung, Essential roles of Atg5 and
FADD in autophagic cell death: dissection of autophagic cell death into vacuole
formation and cell death, J. Biol. Chem. 280 (2005) 20722–20729.
[47] S.Y. Park, J.W. Seol, Y.J. Lee, J.H. Cho, H.S. Kang, I.S. Kim, S.H. Park, T.H. Kim,
J.H. Yim, M. Kim, T.R. Billiar, D.W. Seol, IFN-gamma enhances TRAIL-induced
apoptosis through IRF-1, Eur. J. Biochem. 271 (2004) 4222–4228.