| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,§
*
Laboratory of Transplantation Immunology, Louis-Charles Simard Research Center, and
Nephrology Service, Notre-Dame Hospital, University of Montreal; and
Montreal General Hospital and
§
Department of Surgery, McGill University, Montreal, Canada
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
B becomes
active after the enzymatic cleavage of its precursor by the proteasome
(12); IB
, the inhibitor of NF-B, and c-Jun protein are
degraded via the proteasome pathway (12, 13).
According to sedimentation rates, the proteasome could be purified as
26S and 20S complexes. The 20S proteasome is a cylindrical proteolytic
core composed of multiple and ß subunits. Each subunit is coded
by a different gene in high eukaryotic cells, and the total number of
subunits varies among different species (14). In vitro, the purified
20S proteasomes can digest small peptides in an ATP-independent
fashion, but they are inactive on intact folded proteins (15). The 20S
proteasome can bind at its ends a 19S regulator and forms the 26S
proteasome, which degrades ubiquitinated protein in an ATP-dependent
fashion (1). The 20S proteasome can also complex with an 11S activator
called PA28 (14). PA28 is a ring-like hexamer or heptamer composed of
and ß subunits (PA28 and PA28ß), both of which are about 29
kDa in size (16, 17). It is not clear whether the 20S proteasome can
associate with both 19S and 11S regulators at the same time.
There are two better characterized mechanisms regulating the protein
degradation via the proteasome pathway. The first is that of substrate
selection. This process is controlled by a cascade of enzymes called
the ubiquitin-activating enzyme, the ubiquitin-conjugating enzyme, and
the ubiquitin ligase (1). In addition, the 19S regulator controls the
entry of the ubiquitinated protein into the 20S catalytic core. The
second mechanism is the activity of the 20S proteasome, which is
enhanced by the 11S PA28 (16). It is not clear whether and how the 11S
PA28 exerts its effect on the 26S proteasome, since it and the 19S
regulator do not seem to associate with the 20S at the same time.
Moreover, whether the 20S complex exists in parallel to the 26S complex
in vivo is still an open question. Nevertheless, it has been shown that
overexpression of PA28 could indeed significantly augment Ag
processing by the proteasome in vivo (18).
Our recent work has revealed that PA28 expression is up-regulated during T cell activation, and probably as a result, the ex vivo proteasome activity is fourfold higher in the activated T cells than that in the resting T cells (19). Such an augmented activity probably reflects the increased need to destroy short-lived regulatory proteins and other types of proteins during T cell activation and proliferation. Consequently, it is logical to hypothesize that blocking the proteasome activity will interfere with the activation and proliferation of T cells.
Certain peptide aldehydes, such as N-acetyl-L-leucinyl-L-leucinal-L-norleucinal (LLnL) and N-carbobenzyoxyl-L-leucinyl-L-leucinyl-L-norvalinal (MG115) are competitive inhibitors of chymotrypsin (20, 21). These agents could effectively block the chymotrypsin-like activity and, to a lesser extent, the trypsin-like and PGPH activities of the proteasome (5). They have been employed to study the function of the proteasome in various cellular processes. A caveat of such studies is that these peptide aldehydes are not specific to the proteasome peptidases, and other cellular cysteine protease, such as calpain and cathepsin B (5, 22), are also potently inhibited. This makes some interpretations less assuring.
In 1991, Omura et al. reported the discovery of lactacystin (LAC), which could induce neurite outgrowth (23, 24). Fenteany et al. subsequently found that LAC is a proteasome-specific protease inhibitor (25). It inhibits the three major peptidase activities (i.e., chymotrypsin-like, trypsin-like, and PGPH activities) of the proteasome, and the inhibition of the first two is irreversible in in vitro assays. LAC does not affect other proteases, such as calpain, cathepsin B, chymotrypsin, trypsin, and papain. Currently, LAC is the only proteasome-specific protease inhibitor available. Taking advantage of LAC’s specificity and potency, in this study we used this compound to investigate the roles of proteasomes in T lymphocyte activation and proliferation. We have demonstrated that the proteasome is essential for progression of T cells from the G0 to S phase. Probably as a result of blockage of cycling, the activated, but not the resting, T cells underwent apoptosis when treated with LAC. We have also shown that the proteasome controls the protein levels of p21Cip1 and p27Kip1 as well as CDK2 activity in the G1 phase, and such a control mechanism might be essential in cell cycle progression.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
RPMI 1640, FCS, penicillin/streptomycin, and
L-glutamine were purchased from Life Technologies
(Burlington, Canada). Lymphoprep was purchased from Nycomed (Oslo,
Norway). PHA, hydroxyurea, nocodazole, and histone H1 were obtained
from Sigma Chemical Co. (St. Louis, MO). Staphylococcus
aureus Cowan I (SAC) were obtained from Calbiochem (La Jolla, CA),
and lactacystin was obtained from Dr. E. J. Corey (25). Human
rIL-2 was obtained from Hoffmann-La Roche (Nutley, NJ), and
anti-CD3 mAb OKT3 was obtained from American Type Culture
Collection (Rockville, MD). FITC-conjugated anti-CD3 mAb (clone
SFCIRW2–8C8) and PE-conjugated anti-CD25 mAb (clone IHT44H3) were
purchased from Coulter (Miami, FL). Anti-CD28 mAb (clone 9.3) was a
gift from Dr. P. Linsley (26). A fluorogenic chymotrypsin substrate,
sLLVY-MCA, was obtained from Peninsula Laboratories (Belmont, CA).
Rabbit antisera against cyclin A, cyclin E, p27Kip1,
p21Cip1, CDK2, and CDK4 were purchased from Santa Cruz
Biotechnology (Santa Cruz, CA). [
-32P]ATP (3000
µCi/mmol) and [125I]protein A (30 mCi/mg protein) were
obtained from Amersham (Oakville, Canada), and
[methyl-3H]thymidine (2 Ci/mmol) was obtained from ICN
(Irvine, CA).
Cell culture
PBMC and tonsillar T cells were prepared as described previously (27, 28). The cells were cultured in RPMI 1640 supplemented with 10% FCS, L-glutamine, and antibiotics. [3H]thymidine uptake was conducted as described previously (27, 28).
DNA fragmentation assay
The assay was performed according to a protocol described by Liu et al. (29) with some modifications. Briefly, 2 to 6 million cells were resuspended in 50 µl of PBS followed by 300 µl of lysis buffer (100 mM Tris-HCl (pH 8.0), 5 mM EDTA, 0.2 M NaCl. 0.2% (w/v) SDS, and 0.2 mg/ml proteinase K). After overnight incubation at 37°C, 350 µl of 3 M NaCl was added to the mixture, and cell debris was removed by centrifugation at 13,000 x g for 20 min at room temperature. DNA in the supernatant was precipitated with an equal volume of 100% ethanol. The pellet was washed with cold 70% ethanol and then dissolved in 20 µl of Tris-EDTA buffer containing 0.2 mg/ml RNase A. After incubation at 37°C for 2 h, the DNA was resolved on 2% agarose gel and visualized with ethidium bromide staining.
Electron microscopy
T cells and Jurkat cells were examined by electron microscopy as described by Tsao and Duguid (30).
Flow cytometry for IL-2R
Two-color staining with FITC-anti-CD3 and PE-anti-CD25 was performed on tonsillar T cells. The method was described previously (28).
Proteinase assay
Jurkat cells were cultured with various treatments and were harvested and sonicated in 300 µl of PBS on ice for 40 s. Twenty micrograms of protein per sample from the cleared lysates was supplemented to 100 µl with 0.1 M Tris buffer (pH 8.2). The fluorogenic chymotrypsin substrate sLLVY-MCA was added at a final concentration of 10 nM. The samples were incubated at 37°C for 15 min, and the reaction was terminated by adding 4 µl of 2.5 M HCl. The samples were then diluted to 2 ml with 0.1 M Tris, pH 8.2, and their fluorescence intensity was measured using a PTI fluorometer (Photo Technology International, South Brunswick, NJ). The excitation wavelength was 380 nm, and the emission wavelength was 440 nm.
Cell cycle synchronization of T cells and Jurkat cells
Tonsillar T cells were cultured in the presence of 2 µg/ml PHA and 1 mM hydroxyurea for 40 h. The cells thus treated were synchronized at the G1/S phase. The synchronization was released by washing out hydroxyurea, and the cells were cultured in medium for additional 6 to 22 h according to the requirements of the individual experiments. The synchronization of Jurkat cells was described in our previous publication (31). Briefly, the Jurkat cells were starved in isoleucine-deficient medium for 24 h, followed by 16-h treatment with 2 mM hydroxyurea (HU). Cells thus treated were synchronized at the G1/S boundary. For synchronization at the G2/M boundary, the G1/S synchronized cells were released from hydroxyurea and cultured in regular medium for 6 h, and then treated with 0.1 µg/ml nocodazole for 16 h. The cells were thus synchronized at the G2/M boundary.
Cell cycle analysis
Flow cytometry was employed for cell cycle analysis for T cells and Jurkat cells as described previously (31), using propidium iodide staining.
Immunoblotting
Immunoblotting was employed to evaluate the levels of cyclin E, cyclin A, p21Cip1, and p27Kip1. The general protocol was described in our previous publication (32). Briefly, lymphocytes were lysed in the presence of proteinase inhibitors. The cleared lysates were quantitated for protein concentrations. An equal amount of lysate proteins (40 µg) of each sample was resolved by 10% SDS-PAGE and was transferred to polyvinylidene difluoride (PVDF) membranes (Millipore Corp., Bedford, MA). The membranes were then blocked with 5% milk and hybridized with rabbit antisera against cyclin E, cyclin A, p27Kip1, and p21Cip1 at the dilutions suggested by the manufacturer. The signals on the membrane were detected by [125I]protein A followed by autoradiography.
Immunoprecipitation and the kinase assay
Lymphocytes were lysed by the lysis buffer used for
immunoblotting (32), and cleared lysates were quantitated for their
protein content. For immunoprecipitation, 50 µl of rabbit antisera
against CDK2, CDK4, or cyclin E was added to the lysates equivalent to
20 or 40 µg of protein depending on the experiment. After 2-h
incubation at 4°C, the immune complexes were recovered by protein
A-conjugated Sepharose (Pharmacia Biotech, Montreal, Canada). The
immune complexes bound to protein A-Sepharose were extensively washed
in lysis buffer without detergents or EDTA and resuspended in 50 µl
of kinase reaction buffer (100 mM NaCl, 20 mM HEPES (pH 7.5), 5 mM
MnCl2, 5 mM MgCl2, 25 µM cold ATP, 2.5
µCi of [-32P]ATP, and 3 µg of histone H1 as a
substrate). The reaction was conducted for 10 min at room temperature
and was stopped by adding SDS-PAGE loading buffer. After boiling for 3
min, the samples were subjected to 10% SDS-PAGE. The proteins were
then transferred to PVDF membranes, and the signals were detected by
autoradiography.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The role of proteasome in T cell activation and proliferation was
first examined in PBMC, using the proteasome-specific inhibitor LAC.
PBMC were activated with various stimulants. LAC was added to the cells
at the beginning of culture (0 h) along with the stimulants.
[3H]thymidine uptake between 48 and 64 h of
64-h cultures was used as a parameter for cell proliferation. As shown
in Figure 1
, LAC strongly and
dose-dependently inhibited the T cell proliferation induced by the T
cell mitogen PHA (Fig. 1A), by cross-linking TCR with
anti-CD3
(Fig. 1B), or by Ca2+
ionophore plus cross-linking of the T cell costimulating molecule CD28
(Fig. 1C). The T cell-independent B cell
proliferation induced with SAC plus IL-2 in tonsillar B cells was also
potently inhibited by LAC (Fig. 1D). In all four systems employed, LAC
at 5 µM exerted near-maximal inhibition. The results suggest that
LAC’s effect is not lymphocyte type (T or B cells) or stimulant
specific. Rather, it probably affects certain downstream events
governing a more general process(es) in lymphocyte activation and
proliferation.
|
Since LAC has been reported to induce apoptosis in U937 cells (32), it is crucial to examine whether the LAC-induced inhibition of cell proliferation is due to cell death, be it apoptosis or necrosis.
The viability of T cells and Jurkat cells after LAC treatment was first
evaluated with trypan blue exclusion. Resting T cells (T cells in
medium) or PHA-stimulated T cells were cultured with 10 µM LAC (LAC
added at the beginning of the culture). As shown in Figure 2A, after 16-h culture the
viability of the cells showed only a minor decrease (<12%) in
LAC-treated cells compared with that in cells without LAC (97 vs 92%
for cells in medium, 94 vs 83% for cells with PHA). After a prolonged
culture for 64 h, the decrease was more prominent, although it was
still less than 27% (97 vs 79% for cells in medium, 90 vs 63% for
cells with PHA).
|
B). The viability of the activated T cells dropped
from 94 to 46% after an additional 24-h culture, although 9-h culture
did not change the viability significantly according to trypan blue
exclusion. On the other hand, the viability of the resting T cells in
medium had only a small decrease (from 98% of the control to 87% of
the LAC-treated cells) after 24 h in the presence of LAC.
Why did LAC added at 0 h along with PHA cause less cell death than
LAC added 40 h after PHA stimulation (Fig. 2, A vs
B)? We will demonstrate later in this paper that LAC
is rapidly degraded in the cell culture. After 24 h in culture
medium, LAC lost its activity, and at 40 h, when the T cells were
fully activated and became more susceptible, there was no biologically
active LAC in the culture. This could explain the observed difference
in terms of viability between the 0 and 40 h addition of LAC to
PHA-activated T cells.
The effect of LAC on Jurkat cells was quite similar to that on
activated T cells. Less than 8-h exposure to 6 mM LAC did not induce
apparent Jurkat cell death, while about 60% of the Jurkat cells
were trypan blue positive after 24-h culture with LAC (Fig. 2C).
We next employed DNA laddering to study the mode of cell death caused
by LAC, and the result of this experiment also reflected the degree of
cell death after different treatments. As shown in Figure 2D, resting T cells treated with 10 µM LAC for 24 h
showed no apparent DNA breakdown (lanes 1 and
2). This correlated to the good cell viability, as
shown in Figure 2B. On the other hand, clear DNA ladders
could be observed in activated T cells (40 h post-PHA stimulation)
treated with LAC for additional 9 h (lanes 3 and
4). After 24 h of LAC treatment, the ladders
became less discrete, and this probably reflected further DNA
breakdown. For Jurkat cells, DNA fragmentation could be detected as
early as 6 h after LAC treatment, and after 16 h, the
fragmentation became more prominent (Fig. 2E).
Electron microscopy was also employed to examine the mode of cell death
induced by LAC. The resting T cells (cells cultured in medium; Fig. 3A), activated T cells
(40 h after PHA activation; Fig. 3C), and Jurkat
cells (Fig. 3E) were all healthy looking. Occasional
condensed nuclei were observed in medium-cultured T cells (Fig. 3A), and this is not unusual. The resting T cells
treated with LAC (10 µM) for 24 h were still healthy (Fig. 3B). However, nuclear condensation, which is a
hallmark of apoptosis, was frequently observed in activated T cells and
Jurkat cells after they were exposed to LAC (10 and 6 µM,
respectively) for 24 h (Fig. 3, D and
F).
|
The data in this section indicate the following. 1) LAC’s differential
effect on the viability of resting vs cycling cells suggests that it is
not simply nonspecific cytotoxicity, but relates to the status of the
cell cycle. 2) Cell death, without a doubt, contributes to but cannot
solely account for the observed inhibition of proliferation by LAC,
since there is still a significant percentage (
60%) of live cells
at the end of the culture, as determined by trypan blue exclusion.
Moreover, we will elaborate later that the cell death is a consequence
of blockage of cell cycle progress. 3) Admittedly the trypan
blue-negative cells include some early apoptotic cells, as
evidenced by the fact that DNA laddering could be detected in a largely
trypan blue-negative population. However, it does not necessarily mean
that the whole population is dead. We will later demonstrate that most
Jurkat cells treated with LAC for 6 to 8 h could still progress
normally in the cell cycle despite the fact that a certain degree of
apoptosis could be detected in these cells. 4) LAC can be used to study
the role of proteasomes in lymphocyte activation and proliferation as
long as the compound is applied only once at the beginning of
activation of resting T cells, and the experiment is conducted in 24 to
40 h, or LAC is present for <8 h in the case of cycling cells,
since such treatments do not drastically affect the viability of the
cells.
Effect of LAC was rapid and reversible
We next investigated how fast and how long LAC could exert its
effects on the lymphocytes, since this information is necessary to
assess the requirement for the proteasome activity in events related to
cell activation and proliferation. PBMC were pretreated with LAC (10
µM) or medium for the period indicated in Figure 4A. The cells were then washed
and recultured in the presence of PHA. Thymidine uptake was measured 3
days later. It was clearly demonstrated that 3-h preincubation with LAC
was sufficient to cause significant inhibition of the subsequent
mitogen-stimulated proliferation in T cells, although 16-h
preincubation with LAC was more effective. This result indicates that
LAC can enter the cells rapidly within 3 h.
|
B, the proteasome
activity in Jurkat cells was almost completely inhibited by 3-h
preincubation with LAC at 6 µM. Five hours after the LAC was washed
out, the proteasome activity in the cells was still significantly
inhibited, but the inhibition was reduced compared with that at 0
h. By 21 h, the proteasome activity returned to a near-normal
level. It should be noted that the short 3-h treatment with LAC did not
affect the viability of the Jurkat cells, and this is also reflected by
the normal proteasome activity of the treated cells at 21 h. The
result shows that LAC is not stable and loses its activity within
21 h in the cells.
We also investigated whether LAC was stable in the culture supernatant.
LAC (6 µM) was added to Jurkat cells culture for 4, 6, 16, or 24
h. The conditioned medium was harvested and used to treat fresh Jurkat
cells for 3 h, then the proteasome activity in the lysates of the
fresh Jurkat cells was assayed. As shown in Figure 4C, 4- to
24-h conditioned medium without LAC did not affect the proteasome
activity of the fresh Jurkat cells. The medium conditioned with LAC up
to 6 h could still actively inhibit the enzymatic activity, but
after 16 h, the LAC-conditioned medium lost its inhibitory effect.
The loss of LAC activity in the 16- and 24-h conditioned medium is
unlikely to be due to trapping of LAC by proteasomes released by dead
Jurkat cells, because LAC could rapidly enter the live cells, and the
equilibrium of the LAC concentration between both sides of the
cytoplasmic membrane should be established very fast. Thus, the
proteasomes, whether released or not, should not make a difference in
terms of trapping LAC. Besides, we have also noticed that LAC kept in
cell-free culture medium at 4°C lost its activity within 24 h
(data not shown). These results indicate that LAC is not only unstable
within the cells, but is also unstable in the supernatant.
LAC’s capability to enter the cells to inhibit the proteasome activity
rapidly (<3 h) and its short active duration within the cell and in
the culture medium (16 h) make the compound a useful reagent to
evaluate the requirement for proteasome activity in various events
during cell activation and proliferation, since we could pinpoint the
period when proteasome activity is critical.
Proteasome activity is required for IL-2R up-regulation
In the four systems of T and B cell activation and proliferation
studied in the first section, the growth-promoting activity of IL-2 is
indirectly (for stimulation by PHA, anti-CD3, and anti-CD28
plus ionomycin) or directly (for SAC plus IL-2) involved. We then
investigated the role of proteasome in IL-2R expression and IL-2
production. As shown in Figure 5, CD25
was up-regulated in CD3+ T cells 40 h after
stimulation with PHA. When LAC (10 µM) was added at the beginning of
the culture, the up-regulation was significantly inhibited. On the
other hand, IL-2 production by PBMC 2 to 4 days after PHA stimulation
in the absence or the presence of LAC (10 µM, added at the beginning
of the culture) was also examined, but no consistent difference was
found (data not shown). Under the experimental condition used, the
viability of the LAC-treated cell was reasonable (>80% at 40 h),
as described in the previous section, because LAC was added only once
initially. Moreover, the absence of a consistent change in IL-2
production in LAC-treated cells was a functional indication that the
cell viability was reasonable and is not of a concern in interpreting
the data. The results from this section indicate that IL-2R
up-regulation, but not IL-2 production, is proteasome dependent, and
the suppressed IL-2R expression probably contributes to LAC’s
inhibitory effect on T cell activation and proliferation.
|
Like that of normal T cells, the proliferation of Jurkat cells was potently inhibited by LAC (data not shown). We used synchronized Jurkat cells to identify the LAC-sensitive phase(s) of the cell cycle.
Jurkat cells were first synchronized at the G2/M
boundary by nocodazole (Fig. 6A). The cells were
released from the blockage by washing out nocodazole. In the control
sample, more than half the cells traversed through the M phase and
arrived at the G1 phase within 4 h. In the test
sample, LAC (6 µM) was added to the culture 3 h before the
release, so the compound could have enough time to enter the cells. LAC
was also added to the culture after the release. However, the Jurkat
cells treated with LAC traversed through the M phase to the
G1 phase at a similar pace as the control cells. Since the
total duration of the assay was around 7 h (3-h preincubation plus
4 h after the release), LAC was certainly active during this
period. The fact that most of the synchronized Jurkat cells could
traverse through G2/M to G1 in the presence of
LAC for 7 h again suggests that the viability of the cells thus
treated is not a matter of concern. This result shows that the
G2 to G1 progression is not proteasome
dependent.
|
B). When LAC was added
to the culture immediately after cell release, it slowed, but did not
block, cell cycle progression from the G1/S boundary to the
G2/M phase, as evidenced by the histograms 6 and 9 h
after the release. It should be noted that although the percentage of
cells in the S/G2/M phase in the LAC-treated sample was
similar to that in controls (the inset table of Fig. 6B), the peak of fluorescence lagged behind
(histogram array). Beyond 9 h, the cells gradually lost their
synchronization, the viability of the cells started to decline, and LAC
gradually lost its activity, so the data became difficult to interpret.
The results from this part of the study suggest that proteasome
activity is required for optimal progression from the G1/S
boundary to the G2/M phase, because the progression could
still proceed, albeit at a slower pace, when proteasome activity is
inhibited. These results also imply that the absolutely
proteasome-dependent window during the cell cycle, as evidenced by the
near-total inhibition of S phase entry in LAC-treated
mitogen-stimulated lymphocytes according to the proliferation data,
must be in the G1 phase before the target point of HU,
which inhibits ribonucleotide reductase in the G1/S
boundary (33). The cycling Jurkat cells are obviously not the best model to study the events in the G1 phase, since the G2/M synchronization becomes desynchronized by the time the cells re-enter the S phase, and there is no appropriate method to synchronize the Jurkat cells at the early G1 phase. We therefore decided to use mitogen-stimulated normal T cells to study the role of the proteasome in the G1 phase.
T cells from PBMC were at G0 when isolated. After 16-h
stimulation with PHA, they remained at the pre-S phase (Fig. 6C). At 40 h, about 20% of the cells were in
the S and G2/M phases. The peak of
[3H]thymidine uptake according to a 16-h pulse was
between 48 and 64 h (data not shown), although at 64 h, the
percentage of cells in the S and G2/M phases was still
about 20% (Fig. 6C). The lack of an increase in the
percentage of cells in the S and G2/M phases at 64 h
compared with that at 40 h was probably due to the exit of the
cells from the S and G2/M phases. It should be noted that
the cycling T cells in this system never reached 100%, because about
15% of the cells were non-T cells, and an additional 20% were
nonresponsive T cells. Taking the cell proliferation and cell cycle
analysis together, the G1/S boundary of the cycling T cells
should be between about 35 and 48 h after PHA stimulation. The
boundary was broad because the synchronization was not ideal.
In this model the role of the proteasome in the S phase entry was
examined. As shown in Figure 6C, LAC added once at 16 h
could totally block S phase entry when examined at 40 h. We have
noticed that when the cell viability was evaluated at 40 h, there
was an increase in cell death, comparing the 16 h addition of LAC
with the 0 h addition (25 vs 17%; data not shown). The
increased cell death was also reflected in the cells with <2 N DNA in
the 40-h histogram. However, such a viability was still reasonable and
would not invalidate our conclusion. According to
[3H]thymidine uptake, LAC was strongly inhibitory even
when added as late as 40 h (Fig. 6D). However,
no difference in the percentage of the population in the S and
G2/M phases was observed at 64 h regardless of whether
LAC was added at 40 h according to flow cytometry (Fig. 6C). The discrepancy could be explained by the fact
that 20% of the cells were already in the S and G2/M
phases at 40 h when LAC was added. LAC prevented additional cells
from entering the S phase, thus explaining the lack of
[3H]thymidine uptake. At the same time, the drug slowed
cell cycle progression from the G1/S boundary to the
G2/M phase, hence the finding of a lingering population in
the S and G2/M phases according to flow cytometry.
It is worth mentioning the inhibition of proliferation by LAC was
probably a combined effect of blocking cell cycle progress and cell
death, with the latter possibly being the consequence of the former.
The later the compound was added when more T cells are activated, the
larger a proportion of the effect should be attributed to cell death
caused by LAC. The extensive cell death for the sample treated with LAC
at 40 h was not fully reflected in the flow cytometry (Fig. 6C) in the form of cells with <2 N DNA. This was due
to the fact that the histogram was gated on a region of largely viable
cells.
The results from this section indicate that the proteasome activity is not required from the G2/M to the G1 phase. It optimizes the progression from the G1/S boundary (as defined by the HU target point) to the G2/M phases, and it is absolutely required for the progression from the G0 to the S phase.
Proteasome activity is essential for CDK2, but not for CDK4, function
CDKs are critical for cell proliferation. CDK4 is essential in the early to mid-G1 phase to facilitate S phase entry (34, 35), and CDK2 is critical in the late G1 as well as throughout the S phase for cell cycle progression (36). We therefore examined the role of the proteasome in CDK4 and CDK2 activities in mitogen-stimulated T cells. In all the models used in this section, LAC was added only once at the beginning of culture. Consequently, the viability of the LAC-treated cells was good for the first 16 h and was reasonable at 40 h, but was not a factor that might interfere with the interpretation of the results.
As shown in Figure 7A, the
resting T cells had some CDK4 activity, and the activity reached a
plateau within 16 h of activation. This was in agreement with the
critical role of CDK4 in the early G1 phase. Inhibition of
the proteasome activity by LAC from 0 to 16 h (LAC added once at
0 h) did not affect CDK4 activity when examined at 16 and 40
h (Fig. 7A). This indicates that the induction and
maintenance of CDK4 activity during the G1 phase are not
proteasome dependent.
|
B), and this
reflected its essential role starting from the late G1
phase and extending to the early S phase. The presence of LAC from 0 to
16 h (LAC added once at 0 h) significantly inhibited CDK2
activity at 16 h and more so at 40 h. Therefore, proteasome
activity during the early activation stage (0–16 h) is essential for
activation of the kinase at the G1 phase and early S phase.
The unchanged CDK4 activity in the LAC-treated cells at 40 h
served as an internal control for the repressed CDK2 activity and
indicated that the latter was not due to the viability problem.
Since at the late G1 phase cyclin E is the predominant
partner of CDK2 (37), we next examined the cyclin E-associated CDK
activity. As shown in Figure 7C, despite the fact that the
level of cyclin E protein was increased after LAC treatment (LAC added
once at 0 h), cyclin E-associated kinase activity was almost
completely inhibited by LAC. These results indicate that CDK2 activity
and most likely cyclin E-associated CDK2 activity in the late
G1 phase are proteasome dependent. The results also suggest
that the inhibition of CDK2 activity is probably an important mechanism
accountable for LAC’s effect in blocking S phase entry.
Degradation of cyclin E, but not cyclin A, is proteasome dependent
Oscillation of cyclins during the cell cycle is a mode of
regulation for CDK activities. Since the CDK2 activity is proteasome
dependent, and CDK2 associates predominantly with cyclin E and cyclin A
at the G1/S boundary and during the S phase,
respectively (38, 39), we studied the role of the proteasome in
degradation of these two cyclins. As shown in Figure 8A, the cyclin E level was
apparently increased around 40 h after PHA stimulation of T cells,
which were then at the G1/S boundary. If the activated
cells were treated with HU, the cyclin E level was significantly
enhanced compared with that in cells treated with PHA alone (Fig. 8A). This reflects a better synchronization at the
G1/S boundary by HU and was consistent with our knowledge
that the cyclin E level peaked at the boundary. After the boundary, the
cyclin E level started to decline, and the decline was prevented by LAC
(Fig. 8A). This clearly demonstrates that the
degradation of cyclin E is a proteasome-dependent process, although
whether the increased cyclin E level contributes to LAC’s effect on
the cell cycle is a matter of debate (to be elaborated in
Discussion).
|
B.
The blockage of the cycle at the G1/S boundary with HU did
not further increase the cyclin A level. However, when the cycle passed
the boundary and entered the S phase, the cyclin A level was
significantly augmented (Fig. 8B), consistent with
the idea that cyclin A is mainly an S phase cyclin. Unlike that of
cyclin E, the level of cyclin A did not decline during the S phase, and
LAC did not affect its level during this period. This suggests that the
proteasome is not involved in cyclin A degradation, at least in the
G1 and S phases, and that LAC’s effect on inhibiting cell
proliferation is unlikely to be mediated via cyclin A levels. The G1/S phase synchronized T cells represented activated cells, and prolonged exposure to LAC would cause significant cell death. However, 6-h treatment of LAC did not apparently affect cell viability, while the blockage of cyclin E degradation, but not cyclin A degradation, was obvious at that time point. Moreover, cyclin A could be considered an internal control for cyclin E, indicating that the LAC-induced cell death should not affect the conclusion in this section.
Role of the proteasome in regulating levels of CDK inhibitors p27Kip1 and p21Cip1
In addition to the cyclin levels, the CDK activities were
controlled by several low m.w. inhibitors. We have examined in this
study the effect of the proteasome on the CDK inhibitors
p27Kip1 (39) and p21Cip1 (40). As shown in
Figure 9A, the resting T cells
had a high level of p27Kip1, and the level decreased
gradually when the cells moved to the G1/S boundary 40
h after mitogen stimulation. This is in agreement with previous reports
(41, 42). The presence of LAC (added once at 0 h) significantly
blocked the decrease when assayed at 16 h, showing that the
degradation is a proteasome-dependent process. The blockage was less
obvious when assayed at 40 h, probably because of the gradual loss
of LAC activity during the 40-h culture. The result suggests that the
blocking of p27Kip1 degradation is a mechanism that
contributes to the inhibitory effect of LAC on CDK2 activity.
|
B).
Such an induction suggests that p21Cip1 might be required
in the G1 phase for roles other than a CDK inhibitor (to be
elaborated in Discussion). Interestingly, LAC
strongly suppressed the up-regulation of p21Cip1 in the
G1 phase, indicating that the expression of
p21Cip1 is proteasome dependent and suggesting that the
proteasome might facilitate cell proliferation via its role in
p21Cip1 up-regulation during the G1 phase. In
this experiment, LAC was only added once at the beginning of the
culture, and the viability of the treated cells at 16 h was good
(83%) and should not be a concern in drawing conclusions. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Some other interesting findings are also reported in this study.
IL-2R CD25 up-regulation, but not IL-2 production, is proteasome
dependent; this was previously undocumented. The degradation of cyclin
E, but not that of cyclin A, depends on the proteasome. The induction
of CDK inhibitor p21Cip1 relies on the proteasome; this is
also a new finding.
LAC is the most selective proteasome inhibitor presently known. It has been reported that LAC irreversibly inhibits the chymotrypsin-like and trypsin-like activities of the proteasome when tested in vitro using the purified 20S complex (25). Interestingly, using the chymotrypsin-like activity of the proteasome as an indicator, we found that the inhibition of proteasome activity in the cells was rapidly reversible within <16 h, and the activity of LAC in the culture supernatants disappeared equally fast. This suggests that LAC is quickly degraded by certain enzymes present in the cells and culture supernatants. Alternatively, its spontaneous hydrolysis at high pH, as reported by Dick et al. (43), might be accelerated by certain enzymes from the cells at physiologic pH.
We have shown that LAC did not cause significant death in resting T cells or T cells in their early stage of activation. A short period of exposure (<8 h) of cycling T cells or Jurkat cells to LAC did not significantly affect the viability of these cells. This is consistent with the report by Grimm et al. (44). We, therefore, restricted analysis of our results within these limits to avoid possible complications in data interpretation due to extensive cell death.
We, however, have found that prolonged exposure (overnight culture) of cycling T cells to LAC, be they T cells 40 h after PHA stimulation or Jurkat cells, resulted in massive cell death in the form of apoptosis according to DNA fragmentation assay and electron microscopy. LAC has been reported both to inhibit and to promote apoptosis depending on the models used. LAC has a protective role in the apoptosis of thymocytes (44) and nerve growth factor-deprived sympathetic neurons (45), while it induces apoptosis in monoblast U937 cells (46). These results together with ours suggest that the role of the proteasome in cell death is cell type dependent. The different susceptibilities of resting vs cycling T cells to LAC in terms of cell death as described in our report also suggest that the role of the proteasome in cell death is cell cycle status dependent. Considering the important roles of the proteasome in cell cycle progression (Refs. 47 and 48 and our present report), it is not too surprising, since blocking the cell cycle progression of actively dividing cells results in cell death in many instances, while quiescent cells are not affected. We are currently investigating the mechanism by which LAC induces apoptosis of cycling T cells.
We have shown that the major effect of LAC in cell cycle progress was in the G1 phase. We then examined two critical CDKs in this phase of the cell cycle. CDK4 is a major cyclin D-dependent CDK in the G1 phase (37, 49, 50) and is necessary for cell cycle progress into the S phase, because microinjection of anti-cyclin D1 Ab into G1 fibroblasts blocks them at the G1/S boundary (34, 35). Our results showed that CDK4 activity was not affected by LAC, indicating that the activation of CDK4 is not a proteasome-dependent process. Another cyclin-dependent kinase, CDK2, is also critical for S phase entry. Expression of nonfunctional CDK2 blocks the cells at the G1 phase (36). The CDK2 activity is augmented in the late G1 phase compared with that in the G0 phase, and it selectively associates cyclin E in the G1 phase (37). CDK2 maintains elevated activity in the S phase, but it switches to cyclin A as its major partner. In this study we have for the first time shown that CDK2 activity in both the G1 and early S phases depends on the proteasome. The suppressed CDK2 activity and most likely the cyclin E-associated CDK2 activity in the G1 phase are probably contributing mechanisms for LAC’s effect in blocking S phase entry. Also, the inhibition of CDK2 activity in the S phase might be a cause of the slower progress from S to G2/M phases in LAC-treated Jurkat cells.
Since CDK2 activity mainly depends on cyclin E and cyclin A in G1 and S phases, respectively, we evaluated the role of the proteasome in the protein levels of these two cyclins. We have demonstrated that the degradation of cyclin E in both G1 and S phases is proteasome dependent, because the proteasome-specific inhibitor LAC could increase the cyclin E level in the G1 phase and block the decrease in cyclin E in the S phase. This has confirmed and extended a previous observation that the degradation of cyclin E is ubiquitin dependent and can be inhibited by a less specific proteasome inhibitor, LLnL (51). However, it is unlikely that the inhibition of S phase entry by LAC is due to the block of cyclin E degradation for the following reasons. 1) The major inhibition of cyclin E degradation occurs in the S phase, and it is already irrelevant to S phase entry. 2) Guadagno et al. have reported that overexpression of cyclin E does not affect cell cycle progress (52). Thus, the moderate increase in the cyclin E level during the G1 phase in LAC-treated cells is unlikely to be a cause of the inhibition of CDK2 activity. The level of cyclin A in the G1 and S phase was not affected by LAC. This shows that the degradation of cyclin A in the G1 and S phases, if it occurs, is not via the proteasome pathway, and that the inhibited CDK2 activity in the G1 and S phases is not due to the altered cyclin A level in LAC-treated T cells.
Six low m.w. CDK inhibitors have been recently identified in mammalian cells, and they play critical roles in regulating CDK activity. These inhibitors can be classified into two groups. The first group comprises four members of Ink4 family inhibitors (p16Ink4a, p15Ink4b, p18Ink4c, and p19Ink4d) (53, 54, 55). They specifically inhibit the activities of CDK4 and CDK6 by directly binding to these CDKs, and they are not known to affect other CDKs. Using polyclonal rabbit Abs (Santa Cruz Biotech) against these inhibitors, we could marginally detect the p16Ink4a signal in immunoblots of the T cell and Jurkat samples, but LAC treatment did not cause noticeable changes in the p16Ink4a level (data not shown). On the other hand, signals of p15Ink4b, p18Ink4c, and p19Ink4d in these cells were not detectable (data not shown). We have also shown that the CDK4 activity was not modulated by LAC. Together, these results suggest that the regulation of the CDK4 activity is not a proteasome-dependent process, and the observed antiproliferative effect of LAC is unlikely to be mediated through the protein levels of the Ink4 family CDK inhibitors.
The second group of inhibitors consists of p27Kip1 and p21Cip1 (56). They only interact with cyclin-CDK complexes, not with CDKs alone. p27Kip1 inhibits the activities of cyclin D-CDK2 and cyclin E-CDk2 (39). p27Kip1 is present at the highest level in the quiescent (G0) and early G1 phases, and the level decreases as the cells enter the S phase (42, 57, 58). The degradation of p27Kip1 is ubiquitin dependent and can be blocked by the chymotrypsin inhibitor LLnL (47). We have confirmed and extended these observations using the proteasome-specific inhibitor LAC. We have shown that in lymphocytes the decline of p27Kip1 from the G0 phase to the S phase could be suppressed by LAC, and it was thus proteasome dependent. The lingering high level of p27Kip1 at the end of the G1 phase in LAC-treated cells could well be a contributing mechanism for the LAC-caused CDK2 inhibition, which could, in turn, contribute to the block of S phase entry. In this regard, it is worth mentioning that rapamycin inhibits p27Kip1 degradation (42) and CDK2 activity (59), and that our study shows that rapamycin inhibits the mitogen-induced proteasome activity (19). Therefore, it is plausible that the proteasome/p27Kip1/CDK2 pathway might mediate to some extent the antiproliferative effect of rapamycin in lymphocytes.
The finding that LAC inhibited p21Cip1 induction in the G1 phase was quite intriguing. p21Cip1 is reportedly a pan-CDK inhibitor, and it inhibits the activities of cyclin D-CDK4/6, cyclin E-CDK2, cyclin A-CDK2, and cyclin B-CDK1 (40). However, we have found that the p21Cip1 level was significantly augmented from the G0 to the late G1 phase, and that inhibition of p21Cip1 by LAC correlated with the blocking of S phase entry. These results suggest that p21Cip1 might have positive functions in cell cycle progression during the G1 phase in addition to its inhibitory effect on CDK2 activity, which was actually enhanced in the late G1 phase despite the increase in p21Cip1 protein. Indeed, p21Cip1 protein has a domain that binds to the proliferation cell nuclear Ag, and a different domain that binds to the cyclin E-CDK2 (60). Such a structure suggests that it might serve as an adaptor to connect CDK2 with proliferation cell nuclear Ag, which, in turn, associates with a group of other proteins involved in DNA replication control (61). The adaptor function might explain the required induction of p21Cip1 in the G1 phase and might be responsible for its putative positive role. There are several recent articles describing the positive role of p21Cip1 in cell proliferation. Zhang et al. have reported that p21Cip1-containing cyclin kinases exist in both active and inactive states (62). Mentel et al. showed that mice lacking p21Cip1 have decreased cycling and absolute number of marrow and spleen hemopoietic progenitors (63). Most interestingly, LaBaer et al. have shown that at a low concentration, p21Cip1 promotes the assembly of the active CDK4/cyclin D complex, whereas at higher concentrations, it inhibits this activity (64). Of course, in our model, CDK4 activity is probably not important for LAC’s effect on cell proliferation, but the positive role of p21Cip1 has been supported by this finding in principle.
Inasmuch as the proteasome degrades a broad spectrum of proteins, its
regulatory role in T cell proliferation probably involves multiple
mechanisms and is not restricted to our findings on the CDKs and CDK
inhibitors, as alluded to above. As an example, inhibition of CD25
up-regulation, as shown in this study, could be one of many other
mechanisms. It should be noted that there is an NF-B binding site in
the 5' flanking sequence of the CD25 gene (65), and NF-B’s
activation is achieved by the proteasome-dependent process of NF-B
precursors and degradation of the NF-B inhibitor I-B (12).
However, we believe that factors other than NF-B are also
responsible for the inhibition of CD25 up-regulation, because the
IL-2 promoter also contains an NF-B binding site (65), but IL-2
production was not affected by LAC.
Our study has revealed that the proteasome has a pivotal function in immune responses, and its importance in the immune system has been underestimated in areas other than Ag processing. Much attention has been devoted to regulation of the anabolic process or protein modification to achieve immune regulation. Our study has proved that the other side of the coin, i.e., protein degradation via the proteasome, is an equally important process in this regard.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Jiangping Wu, Laboratory of Transplantation Immunology, Pavilion A. de Sève, Louis-Charles Simard Research Center, Room Y-5616, Notre-Dame Hospital, 1560 Sherbrooke St. East, Montreal, Quebec, Canada H2L 4M1.
3 Abbreviations used in this paper: PGPH, peptidylglutamyl peptide-hydrolyzing; NF-B, nuclear factor-B; LLnL, N-acetyl-L-leucinyl-L-leucinal-L-norleucinal; LAC, lactacystin; CDK2, cyclin-dependent kinase-2; SAC, Staphylococcus aureus Cowan I; PE, phycoerythrin; HU, hydroxyurea; PVDF, polyvinylidene difluoride.
Received for publication May 13, 1997. Accepted for publication October 7, 1997.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-interferon-inducible activator of the multicatalytic protease. J. Biol. Chem. 269:20727.-interferon-inducible protein activator of the 20S proteasome. FEBS Letters 366:37.[Medline]
in antigen presentation. Nature 381:166.[Medline]
