| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Molecular Biology and Genetics |
Department of Cell and Developmental Biology, Oregon Health Sciences University, Portland, Oregon 97201 [M. S. I., O. P. R., J. W., T-H. D., B. E. M.], and Intramural Research Support Program, Science Applications International Corporation Frederick [D. L. N.] and Laboratory of Drug Discovery, Research and Development, Developmental Therapeutics Program, Division of Cancer Treatment and Diagnosis [S. M. R.], National Cancer Institute-Frederick Cancer Research and Development Center, Frederick, Maryland 21702
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Based on results obtained in recent knockout models, the apoptotic program in response to DNA damage appears to involve the mitochondrial apoptotic pathway mediated by caspase-9 (23, 24, 25) . However, in certain Fas-expressing cells, anticancer drugs and radiation activate the transcriptional induction of the gene for FasL that leads to increased FasL expression. FasL further engages Fas and triggers Fas-dependent, caspase-8-mediated cell death (26, 27, 28) . Furthermore, the absence of a functional wild-type p53 tumor suppressor protein generally impedes the induction of apoptosis by DNA-damaging agents such as genotoxic drugs and ionizing or UV radiation (for reviews see Refs. 29 and 30 ), thus establishing an important role for p53 in the DNA damage-induced apoptosis. However, the relationships between p53 and either of the apoptotic pathways is poorly understood. Despite the incompletely understood mechanisms of apoptosis induced by DNA-damaging chemicals and ionizing radiation, these agents constitute the predominant part of the available anticancer therapy.
An entirely novel approach to induce cytotoxicity in target cancer cells is based on the ability of the amphibian endoribonuclease (RNase) onconase (P-30 protein) to kill rapidly proliferating cells (with a certain preference for cancer cells) both in tissue culture and in the mouse (31, 32, 33, 34, 35) . Onconase, originally isolated from oocytes of Rana pipiens (36) , is a member of a growing family of extracellular cytotoxic RNases (for a review, see Ref. 37 ). The cytotoxic properties of onconase ultimately depend on its ability to enter target cells and on its RNA-hydrolyzing capacity (33) . Recently, onconase, in combination with doxorubicin, has been found to dramatically increase the life span of nude mice bearing human breast carcinoma cells (38) . Interestingly, pancreatic RNase A, which is strongly homologous to onconase, is not cytotoxic (39) . The cellular protein RI, which is an inhibitor of ribonucleases, appears to possess a much greater affinity for RNase A than for onconase (40) , thus providing one possible explanation for the different cytotoxic abilities of the two RNases. An onconase-based immunotoxin with increased tumor cell specificity is approved for clinical trials.4 In the present work, we undertook to elucidate the intracellular mechanisms responsible for the cytotoxic effects of onconase. We used a lipofection-mediated delivery of enzyme into mammalian cells and asked whether it could induce characteristic features of apoptosis in these cells. We demonstrate here that onconase treatment caused degradation of cellular tRNA but left rRNA and mRNA apparently undamaged. Onconase induced a characteristic apoptotic pattern of cell death involving chromatin degradation, nuclear pyknosis and fragmentation, cell membrane blebbing, and activation of caspases-9, -3, and -7. The proapoptotic effects of onconase did not require the presence of wild-type p53 and of the FasL/Fas/FADD/caspase-8 proapoptotic cascade. Although onconase-induced cytotoxicity correlated with onconase-induced inhibition of protein synthesis, we present evidence for the existence of an onconase-triggered apoptotic mechanism that is likely to be independent of the inhibition of protein synthesis. Finally, we demonstrate that onconase-induced cell death and caspase-9 activation are accompanied by unusually little release of cytochrome c from mitochondria and a lack of detectable translocation of Bax from the cytosol onto the mitochondria, suggesting that Bax- and cytochrome c-independent mechanisms of caspase-9 activation might be involved in mediating onconase cytotoxicity.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cell Culture.
HeLa tk- cells were maintained in DMEM supplemented with
10% calf serum (HyClone, Logan, UT). All experiments presented here
were performed with logarithmically growing cultures that had not
reached more than 50% confluence. For this purpose, cells were plated
18–24 h before treatment at a density of 2.6 x 106, 5 x 105, or 2 x 105 cells/plate (or well) in 10-cm plates, 6-well
plates, or 12-well plates, respectively. The p530/0
fibroblast cell line was derived from p530/0 mouse primary
embryonic fibroblasts (a generous gift from Dr. Markus Grompe) through
spontaneous immortalization. The absence of p53 in these cells at both
allele and protein levels was verified by PCR and Western blot
analyses, respectively, and it will be described
elsewhere.5
Onconase Preparation.
Native onconase was purified from Rana pipiens oocytes
(Nasco, Fort Atkinson, WI) following the published protocol
(36)
as described previously (42)
.
Lipofectin-mediated Delivery of Onconase.
Delivery of onconase with Lipofectin was performed in DMEM containing
0.5% calf serum. Lipofectin/onconase mixes were prepared following the
procedure outlined for the delivery of diphtheria toxin and 
-sarcin
in Ref. 41
. Before the application of the
Lipofectin/onconase mixes, the cells were washed once with serum-free
DMEM. Four h after the addition of the Lipofectin/onconase mix, the
cells were fed calf serum to a final concentration of 10%.
Antibodies.
Antibodies against PARP (H-250, sc-7150, and A-20, sc-1562), cytochrome
c (H-104, sc-7159), Bax (N-20, sc-493HRP), rabbit IgG
(sc-2004), and mouse IgG (sc-2005) were from Santa Cruz Biotechnology.
Antibodies against caspase-9 (#66571A), caspase-8 (#66231A), caspase-7
(#66871A), caspase-3 (#65906E), and caspase-1 (#66441A) were from
PharMingen. The activating (CH11, #05-201) and blocking (ZB4, #05-338)
antibodies against human Fas were from Upstate Biotechnology.
Measurement of Protein Synthesis via [3H]Leucine
Incorporation.
Incorporation of [3H]leucine was performed as described
previously (40)
, except that the cells were not subjected
to leucine deprivation before onconase treatment and
[3H]leucine pulse labeling. Two µCi of
[3H]leucine/ml medium were used for labeling.
Detection of Apoptosis.
For detection of apoptosis using TUNEL, cells were plated at low
density on Thermanox slides (Nunc) and treated as indicated in the
text, and apoptotic cells were detected using the In Situ
Cell Death Detection kit from Boehringer Mannheim following the
instructions of the manufacturer. For detection of apoptosis by
morphological criteria, cells were plated on 6-well plates and treated
with onconase as described in the text or in the figure legends. At the
indicated times after the treatment, the cells were washed twice with
PBS to remove dead cells. Cells that remained attached to the plates
were fixed in a solution containing 95% ethanol and 5% acetic acid
for 15 minutes at RT. The cells were then rehydrated with several
washes in a sodium phosphate buffer [10 mM (pH 7.9)] for
1 h. The cells were stained by applying a 10 µg/ml acridine
orange solution in sodium phosphate buffer (pH 7.9) for 1 h,
followed by extensive washing with sodium phosphate buffer (pH 7.9).
Visualization of the incorporated acridine orange into nuclear DNA was
achieved in fluorescent microscopy using an ARC Lamp UV source (Ludl
Electronic Products Ltd., Hawthorne, NY) and an IM35 Zeiss microscope.
Phase-contrast viewing using the same microscope was done with the
in-built visual light source. Photographs of cell were taken with a
CONTAX 167MT camera (Kyocera Co., Tokyo, Japan) and Kodak 100 Elite
Chrome film (Eastman-Kodak). For DNA fragmentation analysis, cells were
treated, harvested, and lysed, and nuclei were sedimented as described
below for preparation of nuclear extracts. The nuclear pellets were
resuspended in 50 mM Tris-HCl (pH 7.8), 10 mM
EDTA, and 0.5% (w/v) sodium N-lauroyl sarcosinate.
DNase-free RNase A (Sigma) was added (f.c. 0.5 mg/ml) for 20 min at
50°C. Proteinase K (Life Technologies, Inc.) was added (f.c. 0.5
mg/ml) for an additional 50 min at 50°C. The samples were then
resolved in 2% nondenaturing agarose gels. DNA was visualized using
ethidium bromide and UV transillumination.
Measurement of Cell Death/Survival.
For the 24-h death assay, a modification of the technique of Goillot
et al. (43)
was used. Briefly, HeLa cells were
plated in 12-well plates. Eighteen to 24 h later, the cells were
treated as described in the text or in the figure legends. Four h after
the treatment, the cells were supplemented with calf serum to 10%.
Twenty-four h after the treatments, the cells were washed twice with
PBS to remove dead cells. Cells that remained attached to the plates
were simultaneously fixed and stained for 10 min by the addition of a
solution containing 0.05% crystal violet, 20% (v/v) ethanol, 0.37%
(v/v) formaldehyde, 80% (v/v) H2O. The wells were then
rinsed extensively with water and dried. Crystal violet was extracted
for 1 h from the cells by addition of 2 ml methanol/well and
vigorous shaking. The absorbance of the extract was measured at 570 nm.
After subtracting the background staining from plates containing only
growth medium but not cells and processed as described above, the
absorbance of control cells was taken to indicate a 100% survival, and
the percentage cell survival of the treated cells was calculated
accordingly. Percentage cell death was calculated as follows:
100 - percent survival.
Preparation of RNA.
Total RNA containing the fraction of tRNA was prepared as follows.
Cells were treated as described in the text or in the figure legends.
At the indicated times after treatment, the cells were washed twice in
ice-cold PBS, scraped in ice-cold PBS, and centrifuged at
16,000 x g for 1 min at 4°C. The cellular
pellet was resuspended in 375 µl of a solution containing 10
mM Tris-HCl (pH 7.0), 150 mM NaCl, and 1
mM EDTA; thereafter, 26 µl of 10% (v/v) NP40 were added
to lyse the cells for 2 min on ice. Nuclei were sedimented at
16,000 x g for 30 s at 4°C, and 375
µl of a solution containing 20 mM Tris-HCl (pH 7.8), 350
mM NaCl, 20 mM EDTA, and 1% (v/v) SDS were
added to the supernatant. One extraction with 750 µl of
phenol/chloroform/isoamyl alcohol (25:24:1) and one extraction with
chloroform/isoamyl alcohol (24:1) were performed and RNA from the
aqueous phase was precipitated with 2 volumes of ethanol. RNA
precipitate was dissolved in RNase-free water, and an equal volume of a
2x RNA-loading solution [50% (v/v) formamide, 6% (v/v)
formaldehyde, 20% (v/v) glycerol, 2 mM sodium phosphate
buffer (pH 7.0), and 10 µg/ml ethidium bromide without bromphenol
blue] was added. The RNA samples were denatured by heating for 10 min
at 85°C.
Northern Blot Analysis.
RNA samples were resolved electrophoretically in either 1% or 4%
denaturing agarose gels (44)
and transferred onto Hybond-N
membrane (Amersham Life Science). Hybridization with radioactively
labeled DNA probes was done using the ExpressHyb hybridization solution
(Clontech) following the manufacturer’s instructions. The
hybridization probe for tRNALys was the oligonucleotide
5'-CTGAGATTAAGAGTCTCATGCTCTACCGACTGAGCTAGCC-3' (Life Technologies,
Inc.). The hybridization probe for cyclophilin is described elsewhere
(44)
.
Reverse Transcription of rRNA by Primer Extension.
Probing for damage to 28S rRNA by reverse transcriptase-mediated primer
extension was done as described previously (40)
. The
following two primers (45)
were used (a)
5'-CCCACAGATGGTAGCT-3'; and (b) 5'-CGACATCGAAGGATCA-3'.
Preparation of Nuclear and Cytosolic Extracts for Western Blot
Analysis.
The following protocol applies to the experiments presented in Figs. 4
, 6B
, 7A
, and 7
B. Cells were treated as
described in the text and in the figure legends. At the indicated
times, both attached and detached cells were collected by scraping
directly in the medium, sedimented at 16,000 x g for 1 min at 4°C, washed by resuspending in ice-cold
PBS, and resedimented. Cellular pellets were resuspended in 100 µl of
a solution containing 10 mM HEPES-KOH (pH 7.9), 60
mM KCl, 1 mM EDTA, 0.5% (v/v) NP40, 1
mM DTT, and Complete Protease Inhibitor mixture (Roche
Molecular Biochemicals, Indianapolis, IN) and lysed for 5 min on ice.
Nuclei were sedimented at 735 x g for 5 min
and resuspended in a solution containing 250 mM Tris-HCl
(pH 7.8), 60 mM KCl, 1 mM DTT, Complete
Protease Inhibitor mixture, whereas the post-nuclear supernatants were
designated "cytosolic extracts" after removal of cellular debris at
16,000 x g for 10 min at 4°C. For
extraction of nuclear proteins, the salt concentration was elevated to
500 mM KCl, and the proteins were extracted at 4°C for 30
min with vigorous agitation, followed by removal of nuclear envelopes
at 16,000 x g for 10 min at 4°C.
|
|
|
Preparation of Mitochondria.
For the detection of cytochrome c presented in Fig. 7C
, mitochondria were prepared and analyzed as described in
Bossy-Wetzel et al. (46)
. For the combined
analysis of Bax and cytochrome c shown in Fig. 8A
, the procedure described in Saikumar et al.
(19)
was applied, with modifications. Control and
appropriately treated cells were harvested by scraping directly in the
medium, sedimented, washed once in ice-cold PBS and once in isotonic
sucrose buffer [SU buffer-250 mM sucrose, 10
mM HEPES, 10 mM KCl, 1.5 mM
MgCl2, 1 mM EDTA, 1 mM EGTA (pH
7.1)] and then permeabilized for 1 min at room temperature in SU
buffer containing 0.025% digitonin and Complete Protease Inhibitor
mixture. The efficiency of permeabilization was nearly 100% as
measured by trypan blue exclusion. After sedimenting the cellular
pellet, the supernatant was designated "cytosolic fraction." The
pellet of permeabilized cells was resuspended in SU buffer containing
0.5% Triton X-100 and Complete Protease Inhibitor mixture and kept for
10 min on ice, after which the Triton X-100-soluble and -insoluble
fractions were separated by centrifugation. All three fractions were
resolved in 15% SDS-PAGE, transferred onto polyvinylidene difluoride
membrane, and probed with appropriate antibodies as shown in Fig. 8A.
|
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
75% cell killing within 24 h of
treatment, see below). Immediately after, 2 h after, or 4 h
after the treatment with onconase, the cells were harvested, and total
cellular RNA was prepared and resolved electrophoretically in either
1% or 4% denaturing agarose gels. Hydrolysis of 28S or 18S rRNA was
undetectable by ethidium bromide visualization in the 1% gel (Fig. 1
,Lanes 4–6). We also applied a highly sensitive method (see
"Materials and Methods") based on the lesion-induced arrest of a
reverse transcriptase-driven primer extension to detect
single-nucleotide damages to domains V and VI of 28S rRNA (41
, 45)
. These domains are directly involved in translational
elongation and are the most common targets for fungal antibiotics and
polypeptide ribotoxins (see Ref. 41
and references
therein). Using two primers that allow the scanning of an approximately
1-kb region within these domains, we failed to detect any sites that
were susceptible to cleavage by onconase (data not shown). We therefore
concluded that rRNA is an unlikely candidate for an onconase target
in vivo.
|
, Lanes 4–6). The same band was further
resolved in a 4% gel into three distinct RNA species, of which only
the species with the highest mobility was significantly affected by
onconase treatment (Fig. 1
, Lanes 7–9). Based on the
apparent mobilities of these three bands, we concluded that they
represent the 5.8S rRNA, the 5S rRNA, and the population of tRNAs,
respectively. To confirm that the band that decreased in abundance
after onconase treatment of cells corresponded to tRNA, we performed
Northern blot analysis of the same RNA preparations using a
radioactively labeled 40-mer DNA oligonucleotide probe complementary to
the human tRNALys. Indeed, at 2 and 4 h after exposure
of cells to onconase, the relative abundance of tRNALys was
decreased to 68% and 55% of the control cells, respectively (Fig. 1
,
Lanes 1–3). Furthermore, probing of the same membrane with
a probe specific for cyclophilin mRNA revealed a single hybridizing
band and an absence of degradation products corresponding to
these mRNAs (Fig. 1
, Lanes 10–12). The abundance and
integrity of three other nonrelated mRNAs (c-jun,
c-fos, and surviving mRNAs) was assessed; these mRNAs also
displayed a single hybridizing band and an absence of degradation
products (data not shown).
Onconase-induced Cytotoxicity.
Although onconase is a cytotoxin with intrinsic internalization
capacity, its translocation across the cellular membrane is the
rate-limiting step of cytotoxic action (33)
. We avoided
this potential complication by Lipofectin-mediated delivery of onconase
into all cell types used in this work. To determine the ability of
Lipofectin-delivered onconase to kill HeLa cells, the cells were
treated with Lipofectin alone or with onconase (0.01, 0.1, or 1
µg/ml) and Lipofectin, and the cell survival was measured 24 h
later using crystal violet staining as described in "Materials and
Methods." Fig. 2
demonstrates that onconase caused cell death that correlated strongly
with the logarithm of onconase concentration; the half-maximal killing
concentration (IC50) for onconase in HeLa cells was 0.05
µg/ml (Fig. 2
; Table 1
, column C). In contrast, onconase added without Lipofectin to the medium
of cells at concentrations of up to 30 µg/ml was not cytotoxic (data
not shown).
|
|
0.1 µg/ml (Fig. 2
; Table 1
, column A). Correlation
analysis (percentage of inhibition of protein synthesis
versus percentage of killing) revealed that at
IC50 for inhibition of protein synthesis, onconase kills
65 ± 5% of the cells in 24 h (Table 1
, column E).
Can the cytotoxic potential of onconase result entirely from the
inhibition of protein synthesis? We reasoned that if this were the
case, a similar correlation between cytotoxicity and protein synthesis
should be observed for other inhibitors of protein synthesis. To test
this hypothesis, we performed the same correlation analysis
(translation versus survival) for emetine, a potent and
irreversible inhibitor of ribosomal translocation (see Ref.
41
and references therein). We observed reproducibly that
at similar levels of inhibition of protein synthesis, onconase is
significantly more cytotoxic. For instance, at IC50 for
inhibition of protein synthesis (0.04 µg/ml; Table 1
, column A),
emetine kills merely 15 ± 5% of the cells in 24 h
(Table 1
, column E). Although not as pronounced, similar results were
obtained using another antibiotic inhibitor of protein synthesis,
cycloheximide (Table 1
, column E). These results are consistent with
the notion that a part of the cytotoxic action of onconase results from
onconase-triggered death mechanisms other than inhibition of protein
synthesis (see below).
Onconase Triggers an Apoptotic Cell Death Program.
To identify and characterize the mode of cell death that is triggered
by onconase treatment, we sought to determine whether onconase-treated
cells display characteristic features of apoptosis. Twenty-four h after
the onconase treatment, HeLa cells displayed a characteristic
internucleosomal chromatin cleavage (Fig. 3A
,Lane 2). Furthermore, treatment of HeLa cells with onconase
induced a distinct pattern of nuclear pyknosis and fragmentation
(karyorrhexis) as viewed by fluorescence microscopy after acridine
orange staining (Fig. 3B)
. Treatment with onconase also
induced a characteristic blebbing of the cellular membrane. Membrane
blebbing in HeLa cells was detectable as early as 4 h after
treatment and continued to be evident at 16 h after the treatment
(data not shown for HeLa cells, but see Fig. 3C
for
p530/0 mouse fibroblasts). We next investigated whether the
mode of onconase-induced cell death involves the activation of an
apoptosis-specific DEVDase activity. To this end, we tested for the
appearance in the nuclear fractions of HeLa cells of specific cleavage
products of the enzyme PARP, a target for caspase-3 (47
, 48)
. Fig. 4A
(top) demonstrates that onconase treatment caused a
dose-dependent disappearance of the Mr
116,000 full-size PARP and a concomitant accumulation of the
Mr 89,000 cleavage product. In conclusion, the
ability of onconase to induce internucleosomal chromatin fragmentation,
karyorrhexis, membrane blebbing, and activation of caspases supported
the notion that onconase actively engages the cellular apoptotic
machinery and triggers apoptotic death.
|
98%), instantaneously (within the first minute after addition), and
irreversibly (41)
. Therefore, emetine-induced apoptosis
(Fig. 4B)
represents, in its purest form, the mode of
apoptosis triggered by inhibition of protein synthesis. When studied
using TUNEL, emetine-treated HeLa cells did not display TUNEL-positive
(apoptotic) cells at 6 or 8 h after treatment; TUNEL-positive
cells were detected only at 10 h after treatment (Fig. 5A)
. In contrast, a substantial portion of onconase-treated cells were
TUNEL-positive as early as 6 h after treatment (Fig. 5A)
. The observed early TUNEL positivity of onconase-treated
cells was not caused by a deoxyribonuclease (DNase) activity (or
contamination) of the onconase preparation because we failed to detect
such activity (or contamination) under four different experimental
conditions in vitro (Fig. 5B)
. Because the
ability of onconase to inhibit protein synthesis has neither the
potency nor the immediacy so characteristic for emetine, the results
shown in Fig. 5A
are consistent with the notion that the
early onconase-induced apoptosis (e.g., at 6 h after
treatment) does not result entirely from inhibition of translation (see
"Discussion").
|
, nuclear pyknosis, DEVD-specific caspase
activity, DNA fragmentation, and massive cell death (Fig. 3C
and data not shown). Importantly, we did not observe significant
differences between the responsiveness of p530/0 and
p53+/+ mouse fibroblasts to onconase (data not shown).
These experiments conclusively demonstrated that the presence of a
functional p53 protein is not required for onconase-induced apoptosis.
Partial Resistance of Onconase-induced Apoptosis and Caspase
Activity to Inhibition by the Caspase Inhibitor zVADfmk.
When treated with the anti-Fas agonistic antibody CH11, HeLa
cells undergo a massive apoptosis within a 24-h period after treatment
(Fig. 6A)
. This cell death is accompanied by a marked increase in DEVDase activity
as measured by the specific cleavage of the p116 PARP, generating the
two fragments p89 and p27 (Fig. 6B
, compare Lanes
1 and 3). Both cell death and caspase activity induced
by CH11 could be completely prevented by pretreatment of cells with the
nonspecific caspase inhibitor zVADfmk (Fig. 6, A and B
, compare Lanes 3 and 4). Because the
CH11 dose used for these experiments (0.5 µg/ml) was approximately 10
times higher than the one required for maximum killing and PARP
cleavage (data not shown), a conclusion can be safely reached that
zVADfmk (at the concentration used) possesses a complete ability to
prevent Fas-mediated apoptosis and caspase activity. However, when
applied to pretreat cells before onconase treatment, the effect of
zVADfmk was markedly different. zVADfmk was not able to completely
prevent cell death but caused a shift to the right in the
IC50 of onconase by more than 1 log (Table 1
, compare
columns C and D). As expected, this IC50 shift was not
caused by an effect of zVADfmk on the ability of onconase to inhibit
protein synthesis (Table 1
, compare columns A and B). Rather, the
IC50 shift correlated well with a partial inhibition in
onconase-induced DEVDase activity (Fig. 4A
, compare
top and bottom). Correlation analysis revealed
that at IC50 for inhibition of protein synthesis, onconase
kills only 35% of the zVADfmk-pretreated cells, compared with 65%
cell killing in the absence of zVADfmk (Table 1
, compare columns E and
F).
Does zVADfmk interfere with the part of onconase-induced
cytotoxicity that results from inhibition of protein synthesis? To
address this question, we investigated the effect of zVADfmk on
emetine-induced cell death. First, we observed that a concentration of
emetine that caused 70% killing (0.3 µg/ml; data not shown) also
caused DEVDase activity, as measured by PARP cleavage (Fig. 4B
, top, Lane 4). This DEVDase
activity was due to specific caspase activation because zVADfmk almost
completely abolished the cleavage of PARP in response to emetine (Fig. 4B
, bottom, Lane 4). However, in
contrast to the effect of zVADfmk on onconase-induced cytotoxicity, the
caspase inhibitor caused only a modest (4.5-fold) increase in the
killing IC50 for emetine (Table 1
, columns C and D) and
failed to significantly affect the cell killing induced by emetine at
IC50 for translation (Table 1
, columns E and F). Taken
together, the results presented in Figs. 2
3
4
and Table 1
are
consistent with the following two notions: (a) cell death
caused by classical inhibitors of protein synthesis (emetine and
cycloheximide) requires a significant inhibition of the cellular
translation machinery, and, despite the appearance of active caspases,
this cell death can proceed in the absence of caspase activity as well;
and (b) onconase-induced cell death appears to involve a
component that is probably independent of inhibition of protein
synthesis and that is zVADfmk sensitive (see Fig. 9
and
"Discussion").
|
shows
that preincubation of HeLa cells with ZB4 abrogated the cell death
induced by CH11, whereas Fig. 6D
demonstrates that the same
pretreatment completely failed to inhibit onconase-induced apoptosis.
We therefore concluded that onconase-induced apoptosis does not utilize
the FasL/Fas death signaling pathway.
Procaspases-9, -3, and -7, But Not Procaspase-1 or -8, Are
Processed in Onconase-treated Cells.
To determine the caspases that are activated by onconase, we
prepared cytosolic extracts from HeLa cells that were treated with
Lipofectin alone or with Lipofectin and onconase for 0, 24, or 48 h and studied the processing of procaspase-1, -8, -9, -3, and -7. The
use of caspase-1 as a negative control was determined by the fact that
this caspase does not seem to be involved in mediating apoptosis
(53, 54, 55)
. The (auto)processing of caspase-8 and -9 was
detected by the decreased intensity in Western blot analysis of the
respective band corresponding to the full-size procaspases. As
expected, 24 h after onconase treatment, no significant caspase-1
cleavage was observed (Fig. 7A
, compare Lanes 1 and 2 with Lanes 3 and
4). Similarly, both 24 h (data not shown) and 48 h
after onconase treatment, the majority of caspase-8 was in the
procaspase form (Fig. 7B
, Lanes 3 and
4). In contrast, there was a massive processing of
procaspase-8 in cells treated for 48 h with CH11 (Fig. 7B
, Lanes 1 and 2). Procaspase-9,
however, displayed onconase-induced processing that was detectable both
24 and 48 h after the treatment (Fig. 7A
, compare
Lanes 1–3 with Lane 4 and Fig. 7B
,
compare Lanes 3 and 4). Procaspase-9 was also
processed in response to treatment of cells with CH11 (Fig. 7B
, Lanes 1 and 2). Active caspase-9
can directly process two DEVDases, caspase-3 and its closest relative,
caspase-7 (11)
; therefore, if the processing of
procaspase-9 observed in onconase-treated cells leads to the generation
of active caspase-9, then a subsequent processing of caspase-3 and -7
should be expected. Indeed, cells treated with onconase displayed
processing of the procaspase forms and accumulation of the
characteristic COOH-terminal Mr 17,000 fragments
that result from the specific processing of the
Mr 32,000 procaspase-3 and the
Mr 35,000 procaspase-7, respectively (Fig. 7A
, Lane 4; data not shown for caspase-7). This
proteolytic processing of procaspase-3 and -7 is likely to generate the
active caspase-3 and -7 because DEVDase activities were detected by
means of PARP cleavage in onconase-treated cells (Fig. 4A)
.
Thus, procaspase-9, -3, and -7, but not procaspase-1 and -8, appeared
to be activated by onconase.
Lack of Massive Release of Cytochrome c from
Mitochondria into the Cytosol and of Translocation of Bax from the
Cytosol onto Mitochondria in Response to Onconase.
Release of cytochrome c from mitochondria is a key event for
initiating a caspase-9/caspase-3 cascade and for amplifying the
effectiveness of Fas ligation through a caspase-8-dependent activation
of caspase-9 (56, 57, 58)
. Most notably, cytochrome
c is a cofactor for caspase-9 activation (12)
.
We therefore investigated whether cytochrome c (normally
entirely sequestered in the mitochondria) is released into the cytosol
on onconase treatment. Twenty-four h after treatment of HeLa cells with
onconase, there was a detectable amount of cytochrome c in
the cytosolic fraction compared to the absence of detectable cytochrome
c in the Lipofectin-treated cells (Fig. 7C
,
top, compare Lanes 3 and 4).
Consistently, however, significantly less cytochrome c was
released from mitochondria into the cytosol in response to onconase
than in response to Fas ligation (Fig. 7C
, top,
compare Lanes 2 and 4) and other proapoptotic
stimuli including actinomycin D or UV radiation (both not shown; all
agents were applied at doses that triggered similar levels of cell
death). To find a possible explanation for the low level of cytochrome
c release from mitochondria in onconase-treated cells, we
studied the subcellular distribution of the proapoptotic protein Bax in
control and onconase-treated cells. Recent studies have demonstrated
that the transition of Bax from a soluble (cytosolic) form to a
mitochondrial membrane-inserted form is sufficient and possibly
required for the release of cytochrome c from apoptotic
mitochondria, both in in vitro reconstitution systems and
in vivo (14, 15, 16, 17, 18, 19, 20)
. In search of a positive
control, we investigated the subcellular distribution of Bax in cells
induced to undergo apoptosis in response to treatment with polyI · polyC, a mimic of double-stranded RNA. In HeLa cells, polyI · polyC treatment induces classical features of apoptosis
(e.g., caspase
activation).6
Importantly, polyI · polyC (under treatment conditions described
in Fig. 8A
), in HeLa cells, causes inhibition of protein synthesis with kinetics and
amplitude similar to those induced by onconase and triggers a similar
degree of cell death within 24 h after treatment.6 For
instance, in the experiment shown in Fig. 8A
, polyI ·
polyC caused 70% cell death, and onconase caused >65% cell death
24 h after the treatment. Fig. 8
shows that in the control cells,
the majority of Bax protein was found in the cytosol, with a minor
fraction found attached to sedimentable structures that were soluble in
Triton X-100 (Fig. 8A
, top, compare Lanes
1 and 4). Six h after polyI · polyC treatment,
there was a dramatic redistribution of Bax: the protein levels were
almost undetectable in the cytosol and were greatly increased in the
Triton X-100-soluble organellar fraction (Fig. 8A
,
top, compare Lanes 2 and 5). In
contrast, in onconase-treated cells, the levels of Bax protein were
substantially decreased in both the cytosolic and the organellar
fractions, and there were no indications of a subcellular
redistribution of Bax (Fig. 8A
, top, compare
Lanes 3 and 6). The disappearance of Bax from
onconase-treated cells was almost complete 24 h after the
treatment (data not shown). The differences in the subcellular location
of Bax in polyI · polyC- and onconase-treated cells correlated
with even more dramatic differences in the location of cytochrome
c. In the control cells, the majority of cytochrome
c was found in the Triton X-100-insoluble organellar
fraction (Fig. 8A
, bottom, compare Lanes
1 and 7). After polyI · polyC treatment, the
Triton X-100-insoluble organellar fraction was substantially depleted
of cytochrome c, and the levels of cytochrome c
were dramatically increased in the cytosol (Fig. 8A
,
bottom, compare Lanes 2 and 8). In
contrast, in onconase-treated cells, there was no detectable release of
cytochrome c 6 h after the treatment (Fig. 8A
, bottom, compare Lanes 3 and
9). The results shown in Figs. 7
8
A raised
the possibility that a yet-to-be-identified cytochrome
c-independent mechanism for caspase-9 activation is
triggered in cells in response to onconase. Recently, it was discovered
that a substantial portion of procaspase-9 is sequestered in
mitochondria (59
, 60)
. On apoptotic stimulation,
mitochondrial procaspase-9 was found to translocate to the cytosol and
undergo a subsequent proteolytic activation there (59
, 60)
. Therefore, using the approach described in Fig. 8A
, we investigated whether procaspase-9 undergoes
translocation to the cytosol from an organellar (presumably
mitochondrial) storage site before its proteolytic activation in
response to onconase treatment (Fig. 7, A and B)
.
Indeed, the organellar fraction of onconase-treated HeLa cells
displayed decreased levels of procaspase-9 6 h after the treatment
(Fig. 8B
, top panel, compare Lanes 5
and 6), concomitant with a detectable increase in the levels
of soluble cytosolic procaspase-9 (Fig. 8B
, top
panel, compare Lanes 1 and 2). In contrast,
procaspase-8 was found to be entirely cytosolic (Fig. 8B
,
bottom panel, Lanes 1 and 2) and, as
described above (Fig. 7, A and B)
, unchanged in
response to onconase treatment.
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Apoptotic Mode of Cell Death Induced by Onconase.
Adverse toxins are known to trigger cytolysis via nonapoptotic
mechanisms. One such example is the drug capsaicin, the pungent
ingredient in chili peppers, that has a therapeutic potential for
targeted killing of primary afferent neurons (61
, 62)
.
However, a nonapoptotic mode of cell death could hamper the potential
application of onconase in the therapy of cancer because nonapoptotic
cell death could produce severe inflammatory and immune complications
in patients. Although previously reported as apoptotic (63
, 64)
, we present, for the first time, a detailed characterization
of the mode of cell death induced by onconase, demonstrating that
onconase-treated cells display classical hallmarks of apoptosis such as
chromatin fragmentation, nuclear pyknosis and karyorrhexis, plasma
membrane blebbing, and activation of caspases.
The Nature of the Onconase-induced Death Signal(s): Possible
Involvement of tRNA?
Previous studies have demonstrated that the cytotoxicity of onconase
invariably depends on its catalytic capacity as a ribonuclease
(33
, 40) . The data presented in this work prompt a
speculation that the major determinant of onconase-induced cytotoxicity
is a death signal that is generated in response to induced RNA
hydrolysis. This signal is generated even when concentrations of
onconase are applied that are otherwise insufficient for severe
inhibition of translation. The only RNA population significantly
affected by onconase in HeLa and mouse fibroblast cells appeared to be
tRNA (Fig. 1
; data not shown for fibroblasts). This is in agreement
with our previous findings that onconase preferentially degrades tRNA
in vitro in reticulocyte lysates and when injected into
Xenopus laevis oocytes (34)
. Therefore, the
most likely candidate for an apoptosis-signaling intermediate is
onconase-damaged tRNA. Although we cannot exclude the existence of
certain onconase-sensitive mRNAs, we consider this possibility
unlikely. The coding region of cyclophilin mRNA contains 67 potential
sites for hydrolysis by onconase (5'-UpG-3'), yet apparent hydrolysis
of this highly abundant mRNA was undetectable in onconase-treated cells
(Fig. 1)
. Curiously, onconase easily cleaves rRNA and diverse mRNAs
(including cyclophilin mRNA) in vitro (data not shown). This
finding demonstrates that intracellular conditions determine the
specificity of onconase in vivo. Among the possible
candidates for such specific conditions, we consider the proper
subcellular localization of both target RNAs and onconase and/or
specific postinternalization modification(s) of onconase. These
possibilities are currently being explored.
The Complex Nature of Onconase-induced Cell Death.
A hypothetical model of the cytotoxic pathways triggered by onconase is
presented in Fig. 9
and is discussed below. One of the first effects of onconase on target
cells is the degradation of cellular tRNA. An inevitable effect of
sustained tRNA degradation is inhibition of protein synthesis.
Prolonged inhibition of translation, in turn, will induce cell death
that results, in part, from caspase-mediated apoptosis (see, for
instance, the example with emetine, Figs. 4B
and 5A
). However, three lines of evidence suggest that
onconase-induced apoptosis does not entirely result from inhibition of
protein synthesis. First, whereas antibiotic inhibitors of translation
(as exemplified in Fig. 5A
by emetine) induce apoptotic cell
death with relatively slow kinetics, onconase triggered a markedly
early apoptosis (as measured by the appearance of TUNEL-positive cells;
Fig. 5A
). Second, emetine and cycloheximide required
significantly higher levels of inhibition of translation to induce cell
death than onconase (Table 1
; data not shown for cycloheximide). Third,
zVADfmk, although almost entirely abolishing DEVDase activity induced
by emetine (Fig. 4B)
, could not significantly protect cells
from emetine-induced cell death (Table 1
, compare columns E and F). It
is likely that in the presence of zVADfmk, the emetine-treated cells
die from secondary necrosis, a form of cell death that has been
recently postulated for cells in which certain aspects of apoptosis are
impaired or proceed in an atypical way (65)
.
Onconase-induced DEVDase activity was, overall, similarly susceptible
to zVADfmk (Fig. 4)
, but the inhibitor increased the killing
IC50 for onconase by more than 1 log (Table 1
, compare
columns C and D) and efficiently reduced the cytotoxic potential of
onconase at IC50 for translation (Table 1
, compare columns
E and F). In contrast to both onconase and antibiotic inhibitors of
protein synthesis, cell death and DEVDase activation in response to Fas
ligation are completely inhibited by zVADfmk (Fig. 6, A and B)
. The most obvious mechanistic difference between Fas- and
onconase-induced cell death is the inability of Fas ligation to inhibit
protein synthesis in HeLa cells (data not shown). Our results are
therefore consistent with the notion that in addition to the mode of
cell death that is dependent on protein synthesis inhibition (and
shared by any inhibitor of translation in HeLa cells), onconase
triggers a complementary proapoptotic mechanism that is zVADfmk
sensitive. This complementary proapoptotic mechanism activates
caspase-9 and -3/-7 but is distinct from the classical Fas-induced
pathway (FasL/Fas 
FADD caspase-8 Bid Bax cytochrome
c caspase-9 caspase-3). Furthermore, our results
(Figs. 6C
and 8)
also raised the possibility of a Bax- and
cytochrome c-independent mechanism of caspase-9 activation
in cells treated with onconase. Alternatively, it is conceivable that
feedback self-amplifying caspase "loops" are involved in executing
the onconase-induced cell death program. Subsequent to its processing
and activation by caspase-9, caspase-3 can process and activate
additional molecules of caspase-9, thus leading to the
self-amplification of the caspase cascade (11)
. We have
detected caspase-3/-7 (DEVDase) activity as early as 1–2 h after
onconase application (data not shown). At these time points (and up to
6 h, see Fig. 8A
), we were unable to detect a
measurable release of cytochrome c. However, it is possible
that low levels of cytochrome c release and active caspase-9
(below the sensitivity of the assays used here) lead to the initial
increase in DEVDase activity and that these DEVDases, in turn, process
more procaspase-9, thereby amplifying the caspase-9/caspase-3(-7)
cascade without further involvement of Bax or cytochrome c.
Onconase in Comparison with Other Cytotoxic Enzymes Used for
Designing Anticancer Immunotoxins.
Experimental evidence from different laboratories shows that the
mechanisms of cytotoxic actions of natural toxins used to design novel
therapeutic agents with anticancer properties (such as ricin A chain
and Pseudomonas exotoxin A) are more complex than mere of
inhibition of protein synthesis, a property which these agents have in
common (see Ref. 41
and references therein). For instance,
Keppler-Hafkemeyer et al. (66)
have
investigated the mode of cell death triggered by a genetically
engineered immunotoxin containing Pseudomonas exotoxin A as
a cytotoxic agent. Pseudomonas exotoxin A causes inhibition
of protein synthesis through inactivation of translation elongation
factor EF-2 (see Ref. 41
and references therein). Similar
to our findings using onconase, Keppler-Hafkemeyer et al.
have found that the antibody-Pseudomonas exotoxin A fusion
protein also triggered apoptosis and activated caspases
(66)
. However, zVADfmk has been inefficient in preventing
Pseudomonas exotoxin A-induced cell death, despite the
ablation of apoptotic morphology (66)
. In this respect,
Pseudomonas exotoxin A displayed a behavior more reminiscent
of emetine (as described in this work) than of onconase (see Table 1
).
Therefore, in light of the possible use of both cytotoxic enzymes for
immunotoxin design, future parallel studies are required to investigate
whether onconase and Pseudomonas exotoxin A use the same or
different cytotoxic mechanisms.
Lack of p53 Dependence of the Onconase-induced Apoptotic Program:
Possible Implications in Therapy.
We demonstrate in this work that tRNA damage might be a physiologically
relevant death signal in mammalian cells. Similar to DNA damage, this
signal is ultimately communicated to caspases involved in inducing an
apoptotic phenotype. However, unlike DNA damage, the ability of RNA
damage to induce cell death is not affected by the presence or the
absence of a functional p53 protein (Fig. 3C
and text). The
vast majority of human tumors have inactivated the function of p53,
either through acquired or inherited mutation in one allele, followed
by loss of heterozygosity or by altering the expression or function of
critical p53 regulators, such as the human homologue of
mdm-2. Increased resistance of some tumors with inactivated
p53 to conventional DNA-damaging therapy has been reported (67
, 68)
. Therefore, both the p53 independence and the
nonmutagenicity of RNA damage make RNA an attractive target
for developing novel proapoptotic anticancer strategies.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Supported by USPHS Grants CA-39360 and ES-08456
(to B. E. M.) and by an N. L. Tartar Research
Fund Fellowship (to M. S. I.). 
2 To whom requests for reprints should be
addressed, at the Department of Cell and Developmental Biology, Oregon
Health Sciences University, 3181 S.W. Sam Jackson Park Road, Mail Code
L215, Portland, OR 97201. Phone: (503) 494-7811; Fax: (503) 494-4253;
E-mail: magunb{at}OHSU.edu
3 The abbreviations used are: FasL, Fas ligand;
zVADfmk, benzyloxycarbonyl-Val-Ala-Asp(OMe) fluoromethyl ketone;
polyI · polyC, polyinosinic · polycytidylic acid;
DEVDase, caspase-3 or another caspase with caspase-3-like proteolytic
specificity; PARP, poly(ADP)ribose polymerase; TUNEL, terminal
deoxynucleotidyl transferase-mediated dUTP nick end labeling; TBS,
Tris-buffered saline; FADD, Fas-associating protein with death
domain.
4 D. L. Newton and S. M. Rybak,
unpublished results.
5 M. S. Iordanov and B. E. Magun,
manuscript in preparation.
6 M. S. Iordanov and B. E. Magun,
unpublished results.
Received 9/30/99. Accepted 2/ 2/00.
|
REFERENCES |
|---|
