Results 4b-Phorbol 12-myristate 13-acetate PMA protects cells from cadmium toxicity and inhibits delayed ERK activation Cell death resulting from low-dose cadmium poisoning is functional
Trang 1following cadmium intoxication is negatively regulated
by a protein kinase C-dependent pathway affecting
cadmium transport
Patrick Martin, Kim E Boulukos, Marie C Poggi and Philippe Pognonec
CNRS FRE3094, Universite´ de Nice Sophia Antipolis, France
Among the proteins known to play key roles in cell
physiology, extracellular signal-related kinase (ERK) is
one of the most widely studied Originally
character-ized as a protein responding to mitogenic stimulation
by phosphorylation on tyrosine residues [1], ERK is a
downstream substrate of the proto-oncogene Raf [2]
ERK rapidly became identified as a central kinase
involved in signal transduction pathways [3] ERK is
predominantly viewed as a kinase responsible for cell
growth [4] and⁄ or protection against apoptosis [5] Typical ERK activation is associated with strong and rapid phosphorylation within minutes following stimu-lation, and with a more modest but persistent activa-tion that can last for up to a few hours [6] More recently, a new type of ERK activity has been reported that surprisingly links ERK activation to cell death This novel function of ERK in cell death has been observed in certain cell types and organs, including
Keywords
cadmium; extracellular signal-related kinase
(ERK); protein kinase C (PKC); sustained
activation; ZIP8
Correspondence
P Pognonec, Transcriptional Regulation and
Differentiation, CNRS FRE3094, Universite´
de Nice, Parc Valrose, 06108 Nice cedex 2,
France
Fax ⁄ Tel: +33 492 07 64 13
E-mail: pognonec@unice.fr
(Received 16 September 2008, revised 12
December 2008, accepted 12 January 2009)
doi:10.1111/j.1742-4658.2009.06899.x
Extracellular signal-related kinase (ERK) is a well-known kinase taking part in a signal transduction cascade in response to extracellular stimuli ERK is generally viewed as a kinase that is rapidly and transiently phos-phorylated following mitogenic stimulation This activation results in ERK phosphorylating further downstream targets, thus transmitting and ampli-fying the original stimulus, and ultimately resulting in the onset of cellular proliferation and⁄ or protection against apoptosis More recently, several groups have identified a strikingly new type of ERK activation that results
in cell death This activation is very different from conventional ERK acti-vation, as it occurs several hours after the original stimulation, and results
in the sustained phosphorylation of ERK, which can be observed for up to several days One way of inducing this delayed ERK activation is by low-dose cadmium treatment We show here that sustained ERK activation induced by cadmium in human kidney-derived cells is inhibited following protein kinase C (PKC) activation, even when this activation occurs hours before intoxication Furthermore, PKC inhibition results in an enhanced ERK activation in response to cadmium, even when inhibition is induced hours before intoxication PKCe appears to be the most implicated isotype
in this phenomenon Finally, we present evidence suggesting that the ZIP8 transporter is involved in this process, as multiple small interfering RNAs against ZIP8 have a protective effect against cadmium treatment Our results indicate that PKC activation negatively affects ZIP8 transporter activity, thus protecting cells against cadmium poisoning
Abbreviations
eGFP, enhanced green fluorescent protein; ERK, extracellular signal-related kinase; GFP, green fluorescent protein; P-ERK, phosphorylated extracellular signal-related kinase; PKC, protein kinase C; PMA, 4b-phorbol 12-myristate 13-acetate; siRNA, small interfering RNA.
Trang 2neurons in various neurodegeneration models, brain
injury resulting from ischemia–reperfusion [7–9], and
myeloid leukemic cells following cisplatin
chemother-apy [10] An elegant study relying on a Raf-estrogen
receptor (Raf–ER) chimera also demonstrated that
induced sustained ERK activation indeed results in
apoptosis in vitro [11] The common point found in
ERK-associated cell death appears to be an
unconven-tional activation of ERK, which can remain
phosphor-ylated for up to several days We recently identified
another way of activating ERK-induced cell death by
treating cells with cadmium [12] Low concentrations of
cadmium result in a delayed but sustained
phosphory-lation of ERK that is ultimately accompanied by cell
death, suggesting that ERK functions differ depending
upon its kinetics of activation Supporting this
hypoth-esis are earlier studies concerning proliferation versus
neuronal differentiation of PC12 cells [13,14], in which
the classic transient activation is linked to cell
prolifera-tion, whereas a more sustained activation (lasting for a
few hours) is the trademark of differentiation
Rapid, transient activation is thus associated with
cell proliferation and protection against apoptosis,
whereas delayed, sustained activation is associated with
cell death The difference in signaling events leading to
these two distinct patterns of ERK activation remains
largely unexplored In this study, we began to address
the molecular differences involved in rapid, transient
activation versus sustained activation, following
cad-mium exposure To explore this, we treated cultured
cells with low doses of cadmium in order to induce
long-term ERK activation, as previously reported [12]
We then compared this activation with that obtained
after treatment with both cadmium and phorbol ester,
a well-known activator of conventional and novel
pro-tein kinase Cs (PKCs) that induces transient ERK
acti-vation We show that the delayed and sustained
cadmium-dependent ERK activation is strongly
inhib-ited by concomitant treatment or pretreatment of cells
with phorbol ester, and that this inhibition is
associ-ated with better cell survival Furthermore, we found
that PKCe is the isotype predominantly associated
with the delayed ERK activation in HEK cells
Sur-prisingly, we show that treatment of cells with a PKC
inhibitor results in an enhanced response of cells to
cadmium and increased phosphorylation of ERK
Thus, the sustained ERK activation observed after
cadmium treatment is under the negative control of
PKC We also show that this conventional PKC
pro-tective effect is probably due to a modification of the
ZIP8 cadmium transporter activity Our results led us
to propose a model comparing the traditional quick
and transient ERK activation with the specifics of this
unconventional cadmium-dependent activation, taking into account the involvement of the ZIP8 transporter
in this process, as demonstrated by the protective effect of different ZIP8 small interfering RNAs (siR-NAs) against cadmium intoxication
Results
4b-Phorbol 12-myristate 13-acetate (PMA) protects cells from cadmium toxicity and inhibits delayed ERK activation
Cell death resulting from low-dose cadmium poisoning
is functionally linked to sustained ERK activation [12]
As ERK is known to be efficiently and transiently acti-vated by PKC [15], we investigated the effects of the phorbol ester PMA, a diacylglycerol analog that acti-vates PKC isoforms, on the cell response to cadmium
We found that PMA treatment paradoxically rendered cells resistant to CdCl2(Fig 1A), as cells were dramat-ically protected from the toxic effects of cadmium This protective effect was readily seen 16–24 h after 0.5 and 1 lm exposure, when intoxicated cells became rounded and started to detach from the plate, while still metabolizing as seen by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide analysis (data not shown) In contrast, PMA-cotreated cells (MTT, data not shown) remained phenotypically similar to control cells, as presented here both visually (Fig 1A) and quantitatively (Fig 1C) More than 90% of the cells became rounded following 24 h of treatment with
1 lm CdCl2, whereas only 23% were affected in the presence of PMA When cadmium was applied in the presence of PMA, and as expected from the pheno-typic observation, the absolute number of cells also remained higher than when cadmium was applied alone (see legend to Fig 1) To determine the level of cell toxicity following the different cadmium treat-ments used throughout this study, we measured the loss of cell integrity under these different conditions
As illustrated in Fig 1C, 14% and 23% of the cells lost their integrity following 24 h of exposure to 0.5 and 1 lm cadmium, respectively This indicates that only a small fraction of the rounded cells visible in Fig 1A lost their integrity Under the highest cad-mium concentration used later in this study (20 lm cadmium for 4 h), 9% of the cells lost their integrity after 4 h of exposure This relatively moderate effect after 4 h rapidly increased, as 52% of the cells became affected after 8 h of exposure to 20 lm
We next investigated whether the protective effect of PMA could affect the ERK response to CdCl2 As we previously showed that ERK activation following
Trang 3cadmium treatment was clearly detected after 24 h
in HEK cells following exposure to 2 lm CdCl2 [12], all ERK activation experiments were performed under these conditions As depicted in Fig 1D, PMA resulted, as expected, in the conventional early activa-tion of ERK as seen within 15 min by strong phos-phorylation that progressively diminished over time to
be practically undetectable 24 h later On the other hand, CdCl2 treatment resulted in the previously reported delayed activation of ERK (24 h) [12] Inter-estingly, when both PMA and CdCl2 were simulta-neously applied to cells, early ERK activation occurred normally, but delayed ERK activation was strongly diminished This suggests that PMA treatment interferes with molecular mechanisms that participate in sustained ERK activation following CdCl2treatment
PKC isotypes implicated in cadmium-dependent delayed ERK activation are distinct from those responsible for ERK short-term activation
As PMA activates both conventional and novel PKCs,
we investigated which particular isozymes are involved
in the PMA protective effect reported here Three PKC isozymes are found in HEK cells: PKCa, PKCd and PKCe Using siRNA, we knocked down each of these isotypes, and followed the activation of ERK in response either to a 30 min PMA treatment or to 24 h
of exposure to 2 lm CdCl2 Two prevalidated siRNAs were used for each isozyme, and gave similar results
As seen in Fig 2A, we observed that PKCa knock-down had the most severe effect on the rapid and transient ERK response to PMA, whereas PKCd and PKCe knockdown had a less marked effect as com-pared to a control experiment using a green fluorescent protein (GFP) siRNA This result is in good agree-ment with a previous report [16], and demonstrates the validity of the siRNA used in our study In parallel,
we monitored the activation of PKC using a pan-PKC antibody directed against the Ser657 hydrophobic site found in the C-terminal part of PKCa, which detects the three isotypes present in HEK cells, with an appar-ent molecular mass in the 80 kDa range The level of activated PKC matched the level of phosphorylated ERK (P-ERK) detected Interestingly, we found that
in cells treated with CdCl2 (Fig 2B), PKCe had the most dramatic effect on delayed ERK activation, whereas PKCa and PKCd had more modest effects This indicates that the signal transduction cascade leading to ERK activation following CdCl2exposure is molecularly distinct from the one resulting in short-term ERK activation, and that PKCe is the most activated PKC isotype following cadmium exposure
A
D
Fig 1 PMA protects cells against cadmium toxicity and diminishes
cadmium-induced ERK activation (A) Phenotypic observation of a
PMA protective effect against cadmium toxicity HEK293 cells were
cultivated and treated with 0, 0.5 or 1 l M CdCl 2 in the presence of
0 or 100 ngÆmL)1PMA Pictures illustrating the phenotypic
modifi-cations of the cells were taken 24 h following treatment (B)
Deter-mination of cadmium cytotoxicity on HEK293 cells HEK293 cells
were treated for the indicated times with the indicated CdCl2
con-centrations Loss of cell integrity was then measured as indicated
in Experimental procedures The 100% value corresponds to the
measurement of loss of cell integrity following total lysis of the
cells with detergent (C) Graphical representation of the phenotypic
modification of cells from the experiment depicted in (A) Rounded
cells after PMA treatment are indicated by gray bars, and black
bars represent rounded cells in the absence of PMA treatment.
NS, no significant difference (P > 0.05); ***highly significant
(P < 0.001) Plotted values are from at least 400 counted cells for
each condition, with error bars corresponding to the standard
devia-tion from three independent counts As compared to control cells
(no PMA, no Cd) set to 100%, cell numbers in the different
condi-tions were: 0.5 l M , 81%; 1 l M , 65%; PMA ⁄ 0 l M , 96%;
PMA ⁄ 0.5 l M, 89%; PMA ⁄ 1 l M, 74% (D) Western blot analysis of
total ERK and P-ERK HEK293 cells were treated with either 2 l M
CdCl2(Cd), 100 ngÆmL)1PMA (PMA) or both (PMA + Cd) for the
indicated times [0 min, 15 min (0.25), 2 h (2), 8 h (8) and 1 day
(24)] Fifteen micrograms of protein lysates was loaded onto two
parallel gels, and the transferred proteins were analyzed for either
P-ERK or total ERK (ERK).
Trang 4Cadmium treatment results in delayed PKC
activation
We found a delayed and sustained activation of ERK
following CdCl2 treatment As PKC is an upstream
kinase in the mitogen-activated protein kinase
path-way, we investigated the possibility that CdCl2
treat-ment could directly affect PKC activation Lysates
from cells exposed to CdCl2, PMA or a combination
of both were analyzed by western blotting to investi-gate the status of ERK and PKC activation As indi-cated in Fig 2C, the PMA-induced PKC phosphorylation became practically undetectable after
24 h [17], and strong activation of PKC was observed
24 h after CdCl2 treatment There is a clear correlation between PKC activation and ERK activation Interest-ingly, at least two reports have already demonstrated that 24 h of treatment with cadmium activates PKC [18,19] Our work confirms these observations, and indicates that this activation is the likely cause for the ERK cadmium-dependent delayed activation reported earlier [12]
PMA-driven PKC activity inhibits cadmium-dependent delayed ERK activation
To determine whether the PMA protective effect requires PKC kinase function, we analyzed the effect
of GF 109203X, a specific, broad-spectrum PKC inhib-itor [20], on the ERK response to cadmium
GF 109203X is very effective in inhibiting the ERK response, as 0.3 lm significantly reduced ERK activa-tion following a 30 min exposure of cells to PMA, whereas 3 lm totally inhibited ERK activation (data not shown) Surprisingly, the GF 109203X effect on delayed ERK activation following cadmium exposure was a mirror image of its effect on ERK activation fol-lowing PMA treatment: increasing concentrations of
GF 109203X resulted in a specific and dose-dependent increase in ERK activation, which was already detect-able at 0.3 lm and was maximal at 3 lm, whereas
10 lm GF 109203X treatment in the absence of cad-mium did not result in any ERK activation (Fig 3A)
GF 109203X concentrations in the low micromolar range are known to be specific for PKC [20] Further-more, we found that the effect of GF 109203X on the cell response to cadmium dominated that of PMA, as cotreatment with PMA, GF 109203X and cadmium gave results similar to those obtained with
GF 109203X and cadmium (Fig 4 and data not shown), as expected from the GF 109203X mode of action on the catalytic site of PKC [20] Finally, we found that when GF 109203X was applied up to 24 h before cadmium (Fig 3B, ‘)24’), ERK activation was also stronger than in the absence of the inhibitor, whereas the addition of the inhibitor 4 h after cad-mium treatment no longer resulted in strong ERK activation Taken together, these results suggest an inhibitory role for PKC activation in delayed ERK activation, as blocking of PKC activity resulted in increased ERK activation in response to cadmium,
A
B
C
Fig 2 PKCe is the predominant PKC isotype associated with
cadmium response in HEK cells, in contrast to phorbol ester
activa-tion Western blot analysis of total ERK, P-ERK and activated PKCs.
HEK293 cells transfected 48 h earlier with siRNA against PKCa,
PKCb, PKCe or GFP were treated either with (A) 100 ngÆmL)1
phor-bol ester for 30 min (+) or nothing ( )), or with (B) 2 l M CdCl 2 for
24 h, 2 l M CdCl2 and 100 ngÆmL)1 PMA for 24 h, or nothing.
Fifteen micrograms of protein lysates was loaded onto two parallel
gels, and the transferred proteins were analyzed for either P-ERK
and phosphorylated PKC (P-PKC), or total ERK (ERK) (C) Delayed
PKC activation in response to cadmium treatment Cells treated
with either 2 l M CdCl 2 for 24 h, 100 ngÆmL)1phorbol ester for 24 h
or 30 min, or nothing After lysis, 15 lg of total protein was loaded
onto two parallel gels, and the transferred proteins were analyzed
for either P-ERK and activated PKC (P-PKC) or total ERK (ERK).
Trang 5whereas activating of PKC resulted in reduced ERK
activation in response to cadmium
PKC initiates a hit-and-run process resulting in
cell protection against cadmium toxicity
As GF 109203X enhanced ERK activation even when
applied several hours before cadmium, we investigated
whether PMA could also inhibit ERK activation
when applied several hours before cadmium Cells
were pretreated with PMA for 8 h, and then exposed
to cadmium for 24 h This resulted in strongly reduced activation of ERK as compared to treatment with cadmium alone, as shown in Fig 3C (compare
‘Cd’ to ‘PMA then Cd’) This inhibition was similar,
if not stronger, than that observed when PMA and cadmium were applied simultaneously (Fig 3C, ‘PMA and Cd’) This suggests that transient PKC activation following PMA addition can occur at least 8 h before cadmium addition, without affecting PMA inhibitory potential when CdCl2 is finally added several hours later
Although the above experiment clearly demonstrates that PKC activation can occur 8 h before cadmium addition and still inhibit delayed ERK activation, we wanted to determine whether PKC downregulation, known to occur following PMA treatment [21], partici-pates in the ERK response to cadmium To answer this question, we needed to perform a 16 h PMA pre-treatment of cells, which is known to completely downregulate PKC [21] However, our 2 lm cadmium condition required a 24 h delay before ERK activation could be detected, which could be long enough for PKC to regain its steady-state level after these com-bined 40 h In order to stay within a time frame in which PKC is completely downregulated, we used a higher cadmium concentration In the presence of
20 lm cadmium, faster intracellular cadmium accumu-lation resulted in faster ERK activation, which was already strong after 4 h (Fig 3D, left-hand panel), whereas cytotoxicity remained low (9%; Fig 1B) When this higher-concentration cadmium treatment was performed after a 16 h PMA pretreatment, the remaining PMA-driven ERK activation seen after 16 h was still easily detectable (Fig 3D, ‘0’ in the right-hand panel), but remained unaffected by the 4 h 20 lm cadmium treatment These results indicate that the strong inhibition of the ERK response to cadmium observed with an 8 h pretreatment with PMA was also present when PMA was applied up to 16 h before cad-mium, even with a 10-fold higher cadmium exposure
As, at this time, PKC was completely downregulated,
as seen by the total absence of immediate ERK activa-tion following a second treatment with PMA (data not shown), the PMA-driven inhibition of the ERK delayed response to cadmium is likely to reflect a PKC
‘hit-and-run’ effect, whose consequences will protect cells from subsequent cadmium treatment This is con-firmed by GF 109203X, which enhanced the ERK response to cadmium, even when the inhibitor was applied up to 24 h before cadmium In the PKC siRNA experiment presented in Fig 2, we nevertheless observed inhibition of ERK activation following
A
D
Fig 3 Pharmacological inhibition of PKC enhances delayed and
sustained ERK phosphorylation, whereas PMA pretreatment
reduces the ERK response to cadmium (A) Western blot analysis
of P-ERK and total ERK (ERK) in HEK293 cells, following a 24 h
treatment with 2 l M CdCl2(Cd, +) in the presence of the indicated
GF 109203X concentrations (GFX, l M ) Fifteen micrograms of
pro-tein lysate was loaded onto two parallel gels, and the transferred
proteins were analyzed for either P-ERK or total ERK (B) Western
blot analysis of P-ERK and total ERK in HEK293 cells in the
pres-ence (+) or abspres-ence ( )) of 2 l M CdCl 2 , and treated or not treated
with 5 l M GF 109203X 24 h before ( )24), simultaneously (0) or 4 h
(+4) after cadmium addition (C) Western blot analysis of P-ERK
and total ERK in HEK293 cells following no treatment (Ø), 24 h in
the presence of 2 l M CdCl2(Cd), pretreatment with 100 ngÆmL)1
PMA for 8 h followed by 2 l M CdCl2for 24 h (PMA then Cd), or
treatment with 100 ngÆmL)1PMA and 2 l M CdCl 2 for 24 h (PMA
and Cd) Fifteen micrograms of protein lysate was loaded on two
parallel gels, and the transferred proteins were analyzed for either
P-ERK or total ERK (D) Western blot analysis of P-ERK and total
ERK in HEK293 cells following treatment for 0 min, 15 min (0.25),
30 min (0.5), 1, 2 and 4 h with 20 l M CdCl2(20 l M Cd) or with
100 ngÆmL)1PMA (PMA 16 h then Cd).
Trang 6cadmium–PMA cotreatment, despite the substantial
knockdown of the different active PKC isozymes It is
likely that the remaining fractions of phosphorylated
PKC detected were sufficient to drive the hit-and-run
protective effect reported here
In other words, whereas PMA treatment, which
acti-vates PKC, protects cells from cadmium-induced cell
death and diminishes ERK activation, GF 109203X
treatment, which completely inhibits PKC activity,
results in an increased response of ERK to cadmium
treatment This rules out the possibility that the
PMA-mediated downregulation of PKC could be responsible
for the inhibition of the sustained ERK activation,
and suggests that a hit-and-run process initiated by
PKC activation is involved This is reinforced by the
observation that the siRNA against PKCe, which was
quite effective in diminishing the level of activated
ERK after 24 h in the presence of cadmium (Fig 2),
did not significantly protect cells from cadmium toxic-ity, unless PMA was applied (data not shown)
Pharmaceutical inhibition of PKC renders cells hypersensitive to cadmium treatment
As we have previously linked the cellular response to cadmium with sustained ERK activation, we wanted
to determine the effect of GF 109203X on the mor-phology of cells exposed to cadmium This is illus-trated in Fig 4A, where the phenotypic response of cells to cadmium was even more dramatic in the pres-ence of GF 109203X, whereas GF 109203X alone did not result in any significant morphological modifica-tion of the cells Quantificamodifica-tion of this experiment indi-cated that whereas a 24 h cadmium treatment left 12%
of the cells still adhering to the culture dishes, cotreat-ment with cadmium and GF 109203X resulted in
A
B
Fig 4 Inhibition of PKC sensitizes cells to cadmium (A) Phenotypic observation of increased cellular sensibility to cadmium after PKC inhibi-tion HEK293 cells were cultivated in DMEM and 10% fetal bovine serum (Ø), in the presence of: 2.5 l M GF 109203X (GFX); 2 l M CdCl 2
(Cd), 2 l M CdCl2and 2.5 l M GF 109203X (Cd GFX); 100 ngÆmL)1PMA (PMA); 2 l M CdCl2and 100 ngÆmL)1PMA (Cd P); 2 l M CdCl2and
100 ngÆmL)1PMA, and 2.5 l M GF 109203X (Cd P GFX) Pictures illustrating the phenotypic modifications of the cells were taken 24 h after treatment As compared to control cells (Ø) set to 100%, cell numbers in the different conditions were: Cd, 52%; Cd GFX, 47%; GFX, 82%; PMA, 94%; Cd PMA, 63%, Cd PMA GFX, 44% (B) Graphic representation of phenotypic modifications of the cells from the experiment depicted in (A) Percentages of rounded cells are represented by black bars Plotted values are from at least 300 counted cells from each condition, with error bars corresponding to standard deviation from three different counts ***Difference highly significant (P < 0.001) Legends are as in (A).
Trang 7virtually all of the cells floating The PMA protective
effect was also substantially reduced in the presence of
GF 109203X, as over 85% of the cells were rounded
when treated with cadmium, PMA and GF 109203X,
whereas only 25% of the cells were rounded in the
presence of cadmium and PMA (Fig 4B) This
indi-cates that PKC inhibition by GF 109203X is dominant
over PMA activation, as predicted by the GF 109203X
target on PKC, which is distinct from the
PMA-inter-acting domain [20] This confirms that PKC is indeed
the central actor in the observation reported here
PMA downregulation of delayed ERK activation
is not dependent upon protein synthesis
As transient PKC activation up to 16 h before
cad-mium treatment inhibited delayed ERK activation, we
investigated the possibility that protein synthesis could
be required in the follow-up of this hit-and-run
pro-cess To this end, we concomitantly treated cells with
cadmium and emetine, a protein synthesis inhibitor
As seen in Fig 5B, short-term ERK activation in
response to a 15 min PMA treatment alone was, as
expected, unaffected by emetine However, in the
presence of emetine, ERK activation in response to a
24 h cadmium treatment was substantially blunted as compared to cadmium alone (Fig 5A), as it was in the presence of PMA Interestingly, when PMA and eme-tine were added together, the overall ERK activation
in response to cadmium was further decreased as com-pared to the effect of each individually The emetine effect is thus likely to represent the natural degrada-tion of a protein involved in the cellular response to cadmium, whereas PMA treatment is still able to inhi-bit delayed ERK activation, but only through the remainder of the factor whose neosynthesis is blocked
by emetine This strongly suggests that the PKC hit-and-run action is technically independent of protein neosynthesis, but acts through a relatively unstable protein
Taken together, these results suggest that whereas PMA, which mimics diacylglycerol in activating PKC, protects cells from cadmium exposure and limits the associated activation of ERK, GF 109203X has the opposite effect As the protective effect of PMA was also observed when it was applied hours before cad-mium addition, we reasoned that PKC activity could participate in a process that results in the modification
of a pre-existing protein, leading to protection from cadmium toxicity
PKC activation results in a decrease in intracellular cadmium accumulation
To determine whether PMA could directly limit cad-mium entry into cells, we measured cadcad-mium accumu-lation using a radioactive tracer in the presence or the absence of PMA A reduction of cadmium entry was indeed observed, as in the presence of PMA, the inter-nal cadmium concentration was reduced by approxi-mately 30% (Fig 6A) No significant cadmium release from cells was detected, either in the presence or in the absence of PMA (data not shown) To determine whether this reduction in cadmium accumulation caused by PMA could be sufficient to explain the protective effect observed, we incubated cells with a higher concentration of cadmium in the medium (3 lm) in the presence of PMA, and measured cadmium accumulation Despite the higher extracel-lular concentration of cadmium, we still obtained a 27.5 ± 7.7% reduction in the intracellular cadmium accumulation as compared to what we observed after incubation in 2 lm cadmium alone (Fig 6B) We then compared ERK activation under these two conditions, and again found a strong correlation between a reduc-tion in cadmium entry and ERK phosphorylareduc-tion (Fig 6B, insert) Total ERK was unchanged (data not shown) This indicates that modulation of cadmium
A
B
Fig 5 Blocking protein synthesis negatively affects ERK sustained
activation in response to cadmium treatment, but does not affect
PMA inhibitory effect (A) Western blot analysis of P-ERK and total
ERK (ERK) in HEK293 cells, following a 24 h treatment with 2 l M
CdCl 2 (Cd, +) in the presence of 10 l M emetine (Em, +) and
100 ngÆmL)1PMA (PMA, +) Fifteen micrograms of protein lysate
was loaded onto two parallel gels, and the transferred proteins
were analyzed for either P-ERK or total ERK (B) Western blot
anal-ysis of P-ERK and total ERK in HEK293 cells, after a 10 min 10 l M
emetine treatment (Em, +), followed by a 15 min treatment with
100 ngÆmL)1PMA (PMA, +) Fifteen micrograms of protein lysate
was loaded onto two parallel gels, and the transferred proteins
were analyzed for either P-ERK or total ERK.
Trang 8entry following PMA treatment is likely to be the main
reason for the protective effect of PMA reported here
ZIP8 transporter knockdown protects cells from
cadmium toxicity
In parallel, and because we found that emetine also
limited ERK activation by cadmium, we investigated
the net effect of protein synthesis on cadmium entry
into cells As shown in Fig 6C, the presence of
eme-tine resulted in a close to 50% decrease in cadmium
accumulation after a 24 h treatment This decrease in
cadmium accumulation is very likely to be the direct
result of the cadmium transporter(s) turnover
Mathe-matical calculation, based on the comparison of
cad-mium entry measurements in the presence or in the
absence of protein synthesis inhibitor, allowed us to
determine that the half-life of this cadmium
trans-porter would be 11.5 ± 1.5 h (see Experimental
proce-dures) ZIP8 has recently been reported to be the
major cadmium transporter in cells [22], and its
char-acteristics are in perfect agreement with our previously
reported observations concerning cadmium entry into
HEK cells [23] Our mathematical calculation is also in
excellent agreement with preliminary data from the
Nebert laboratory (D W Nebert, Department of
Environmental Health, University of Cincinnati, USA,
personal communication), which estimates the ZIP8
half-life to be 12 ± 2 h In order to investigate
whether ZIP8 could actually be responsible for
cadmium entry, we transfected HEK cells with an
enhanced GFP (eGFP) bicistronic vector (PRIG [24]) expressing or not expressing (control) the ZIP8 trans-porter (ZIP8) Cells that were actually transfected were unambiguously traced by fluoromicroscopy (Fig 7A)
We observed that after 48 h of exposure to 120 nm CdCl2, cells remained unaffected, as expected for such
a low dose of cadmium, whereas cells expressing exog-enous ZIP8 transporters were all rounded and dis-played very weak GFP signals, reflecting the enhanced toxic effect of this very low cadmium concentration, owing to the increased number of ZIP8 transporters in these transfected cells CdCl2 at 120 nm was used in this experiment, because titration experiments indi-cated that it is the lowest cadmium concentration resulting in the complete phenotypic change in cells transfected with the ZIP8 expression vector (data not shown) This demonstrates that ZIP8 is indeed able to transport cadmium into cells In order to investigate the actual contribution of endogenous ZIP8 to cad-mium entry into cells, we performed ZIP8 knockdown
by siRNA transfection Our ZIP8 siRNA mix was first validated on cells transiently expressing a bicistronic RNA encoding ZIP8 on the first cistron, followed by GFP on the second The ZIP8 siRNA mix substan-tially knocked down the GFP signal, indicating that these siRNAs do result in the bicistronic RNA degra-dation (data not shown) As shown in Fig 7B, cells cotransfected with a GFP marker and the ZIP8 siRNA and exposed to 2 lm CdCl2 were markedly protected from cadmium intoxication as compared to cells transfected with control eGFP siRNA, or no siRNA
Fig 6 Quantification of cadmium accumulation in cells (A) Comparison of intracellular cadmium accumulation from medium containing
2 l M CdCl2between control cells (Ø) and 100 ngÆmL)1PMA-treated cells (PMA) Control cells are set as 100% (B) Comparison of intracellu-lar cadmium accumulation between 2 l M CdCl 2 -treated cells and 3 l M CdCl 2 + 100 ngÆmL)1PMA-treated cells Cells treated with 2 l M
CdCl 2 are set as 100% Upper part: corresponding western blot showing P-ERK activation status (C) Comparison of intracellular cadmium accumulation from medium containing 2 l M CdCl2between control cells (Ø), 100 ngÆmL)1-PMA treated cells (PMA), 10 l M emetine-treated cells (Em), and 100 ngÆmL)1PMA + 10 l M emetine-treated cells (PMA + Em) Control cells are set as 100% Cells were incubated for 24 h, lysed, and counted as described in Experimental procedures **Significant difference (P < 0.01), ***highly significant difference (P < 0.001).
Trang 9siRNA knockdown commonly leaves a small
propor-tion of the protein of interest in the cells We are thus
confident that our results demonstrate that ZIP8
par-ticipates in the entry of cadmium into HEK cells
Cadmium accumulation in cells concomitantly
trea-ted with PMA and emetine was also investigatrea-ted As
shown in Fig 6C, we found that these effects were
additive, as presented in Fig 5 for the inhibition of
ERK activation Whereas emetine resulted in close to
a 50% decrease in cadmium accumulation, a simulta-neous PMA treatment was still able to reduce cad-mium accumulation by an additional 30% The simplest explanation is that PMA has a direct negative effect on cadmium transporter function Thus, inde-pendent of the quantity of transporter present in the cell, PMA treatment would always result in a 30% decrease in cadmium accumulation as compared to conditions devoid of PMA, which is in good agree-ment with our results of Fig 5, suggesting that the PKC hit-and-run effect is independent of protein neo-synthesis
To summarize, we show here that PMA-driven PKC activation inhibits sustained ERK phosphorylation observed after cadmium exposure, and protects cells from cadmium toxicity Conversely, blockage of PKC activation enhances cadmium toxicity and ERK activation This PKC effect stems from a protein neosynthesis-independent hit-and-run phenomenon that ultimately results in decreased intracellular accumula-tion of cadmium Finally, ZIP8 knockdown by siRNA substantially reduces cadmium toxicity, demonstrating the pivotal role of ZIP8 in cadmium intoxication Taken together, these results suggest that a PKC-driven modification of ZIP8 decreases its transporter activity, thus reducing sustained ERK activation and conse-quently affecting cell response to cadmium poisoning
Discussion
Investigation of the molecular basis of cadmium toxic-ity during the last decade has resulted in the character-ization of multiple pathways involved in this poisoning [25] This reflects the large panel of targets affected by cadmium Kinases such as P38, ERK and Jun N-ter-minal kinase have been shown to play a role in this cadmium poisoning However, differences in experi-mental conditions, such as high cadmium concentra-tions applied for short periods versus low doses for extended periods, have resulted in data that are some-times hard to reconcile [26,27] Other cellular mecha-nisms are also affected by cadmium The homeostasis
of metals that are essential for diverse biological func-tions, such as calcium, zinc and iron, is disturbed by cadmium [28] Similarly, oxidative mechanisms are also perturbed by cadmium [29] The experiments presented
in this article, based on exposure to low cadmium con-centrations, are comparable to chronic intoxications Our results link the sustained activation of ERK following low-dose cadmium treatments to PKC func-tion, and ultimately to the activity of the ZIP8 zinc transporter Previous work has indicated that cadmium substitutes for zinc in the regulatory domain of PKC,
A
B
Fig 7 ZIP8 transports cadmium into cells and plays a role in
cadmium toxicity (A) HEK cells were transfected with either an
eGFP-expressing control vector (Control), or a bicistronic vector
expressing ZIP8 and eGFP Twenty-four hours after transfection,
cells were exposed to 0.12 l M CdCl2(0.12 l M Cd) or regular
med-ium (No Cd) Cells were observed by epifluorescence 48 h later,
and pictures were taken using the same exposure time for all
conditions Only cells expressing ZIP8 and exposed to a very low
cadmium concentration died, as visualized by their weak GFP signal
and their rounded phenotype As compared to control transfected
cell number, set to 100%, cell numbers in the different conditions
were: ZIP8, 92%; 0.12 l M , 107%; ZIP8 + 0.12 l M , 28% (and all
cells were rounded and weakly fluorescent) (B) Endogenous ZIP8
is involved in cadmium entry into cells Cells were transfected with
an eGFP-expressing plasmid together with no siRNA, an siRNA
directed against eGFP (eGFP siRNA), or an siRNA mix directed
against ZIP8 After 24 h in the presence of 2 l M CdCl 2 , cells were
dying in both eGFP siRNA and no siRNA conditions, whereas viable
cells were still present in the well containing the ZIP8 siRNA mix.
Exposure time was kept constant for all conditions As compared
to ZIP8 siRNA mix-transfected cells, set to 100%, cell numbers in
the different conditions were: eGFP siRNA, 47% (practically all cells
were rounded and not fluorescent); no siRNA, 39% (practically all
cells were rounded and weakly fluorescent).
Trang 10resulting in its activation [30,31] Furthermore, Cd2+,
which is structurally similar to Ca2+, may also
substi-tute for calcium in the activation of PKC isoforms
Calcium levels are highly and rapidly modulated in
cells, but those of cadmium are not, its levels
remain-ing stable We previously demonstrated that
extracellu-lar calcium cannot compete with cadmium for entry
into cells [23] We showed here that PKC isoforms are
indeed subjected to sustained activation following a
24 h cadmium treatment, leading to the delayed ERK
activation reported earlier [12] As PMA pretreatment
results in the rapid activation⁄ downregulation of PKC
and reduces ZIP8 activity, a cadmium effect on PKC
is no longer possible, resulting in the protective effect
reported here In the absence of early PKC
activa-tion⁄ downregulation, ZIP8 is fully active, and
unim-paired cadmium entry into the cell results in the
progressive displacement of calcium, leading to delayed
PKC activation, which itself results in sustained ERK
activation, and ultimately cell death Delayed PKC
activation will also diminish ZIP8 activity, but as the
cells have already been exposed to cadmium, they will
nevertheless undergo cell death, as cadmium does not
exit intact cells Further analysis will be required to
unveil the molecular mechanisms involved in the
cad-mium-dependent delayed PKC activation reported
here
The results presented here led us to propose an
inte-grative model, which is depicted in Fig 8 Two types
of ERK activation are shown On the right side of the
scheme is the early and transient ERK response to
the phorbol ester PMA that is efficiently abrogated
by the specific broad-spectrum PKC inhibitor GF
109203X This corresponds to the conventional
mito-gen-activated protein kinase cascade observed, for
example, following growth factor stimulation On the left side of the scheme is the unconventional and sus-tained ERK activation, like the one observed following low-dose cadmium treatment We show here that: (a) this long-term activation relies on a molecular process under the negative control of PKC, as pharmacological inhibition of PKC results in enhanced ERK activation, whereas a transient PKC activator results in reduced ERK activation; and (b) this PKC negative control does not rely on protein neosynthesis, since even though emetine treatment diminishes ERK activation following cadmium exposure, the remaining activation
is still under the negative control of PKC
The negative control of PKC on sustained ERK activation in response to cadmium is confirmed by PKC inhibitor treatment Indeed, GF 109203X actu-ally enhances ERK activation following cadmium treatment in a dose-dependent manner, whereas PKC activation by PMA diminishes this activation Interest-ingly, among the three predominant PKC isotypes found in HEK cells (PKCa, PKCd and PKCe), siRNA knockdown of PKCe is the most effective in diminish-ing sustained ERK activation followdiminish-ing cadmium treatment, whereas, as already known, siRNA knock-down of PKCa is the most effective in diminishing classic ERK activation by PMA The observation that PKC siRNAs do not increase ERK activation in response to cadmium as GF 109203X does is likely to reflect the fraction of PKC still present and functional within the cells, whereas GF 109203X completely blocks all PKC function
In this article we present evidence suggesting that the protective effect of PMA is probably due to mod-ulation of the activity of the recently identified cad-mium transporter ZIP8 We calculated the half-life of the factor targeted by PMA to be 11.5 ± 1.5 h, which is in perfect agreement with the 12 ± 2 h found for ZIP8 in the Nebert laboratory (D W Nebert, Department of Environmental Health, Uni-versity of Cincinnati, USA, personal communication)
We also demonstrate that ZIP8 exogenous expression
is capable of enhancing cadmium toxicity, whereas ZIP8 knockdown results in a diminution of cadmium toxicity PKC has already been shown to be able to affect different transporters, either by directly modify-ing their activities [32,33] or by affectmodify-ing their stabil-ity [34] In silico analysis of the ZIP8 structure reveals that there are three putative sites that could be phos-phorylated by PKC (94-SSK-96; 361-STR-363; 424-TGR-426) It will be of interest to determine whether these sites are actually phosphorylated in response to PKC activation, and whether their modification affects ZIP8 function
Fig 8 Scheme depicting the different pathways involved in the
control of sustained ERK activation Cd, CdCl2 stimulation; PMA,
PMA treatment; GFX, GF 109203X treatment; Em, emetine; ZIP8,
ZIP8 transporter Arrows with a minus sign indicate a repression
effect, and arrows with a plus sign represent an activation process.
The dashed line represents an indirect process.