| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Drug Research and Evaluation, Istituto Superiore di Sanita’, Rome, Italy
1Correspondence: Department of Drug Research and Evaluation, Istituto Superiore di Sanita’, viale Regina Elena 299, 00161, Rome, Italy. E-mail: malorni{at}iss.it
ABSTRACT
The term mitotic catastrophe (MC) was coined to describe the mammalian cell death caused by aberrant mitosis. MC occurs with features that are fundamentally different from those typifying other forms of cell death, including apoptosis. We report here for the first time that the Rac-activating toxin CNF1 interferes with the occurrence of MC and leads to aneuploidy and multinucleation. This seems to be in line with the anti-apoptotic activity of the toxin and consistent with the hypothesis that points at CNF1 as a toxin bearing a carcinogenic potential.—Malorni, W., Fiorentini, C. Is the Rac GTPase-activating toxin CNF1 a smart hijacker of host cell fate?
Key Words: mitotic catastrophe • multipolar mitosis • Rac GTPase • CNF1 • autophagy
BACKGROUND: MITOTIC CATASTROPHE AND CNF1
THE TERM MITOTIC CATASTROPHE (MC) was coined to describe the lethal fate of Schizosaccharomyces pombe cells forced to prematurely enter mitosis by overexpression of Cdc2, the analog of mammalian CDK1 molecule (1)
. More recently, MC has assumed a broader definition and has been used to indicate the type of mammalian cell death caused by aberrant mitosis. In a physiological condition, the M phase, characterized by sister chromatid alignment, is followed by cytokinesis in which the cytoplasm and its contents, including chromosomes, are partitioned into two daughter cells. Entry into M phase from G2 phase is driven by the activation of the CDK1-cyclin complex, normally maintained in an inactive state by phosphorylation. A defect in CDK1 inactivation has been indicated as a possible key event in the occurrence of MC (1)
. This cell death process occurs with features that are fundamentally different from those typifying necrosis ("oncosis"), apoptosis (type 1 cell death), or autophagy (type 2 cell death) (1
, 2)
and can therefore represent a fourth form of cell death. MC has also been associated with the formation of multinucleated, giant cells containing uncondensed chromosomes (3
4
5)
. The fact that caspase inhibitors, such as z-VAD.fmk, are unable to prevent the appearance of dying giant multinucleated cells (as those induced, for instance, by treatment with spindle poisons) led us to conclude that mitotic catastrophe is unconnected to apoptosis (6
, 7)
. In contrast to this conclusion, however, manipulations that disable the apoptotic machinery, such as overexpression of Bcl-2 (6)
, have been shown to reduce apoptosis and enhance the frequency of catastrophic mitoses. Activation of caspase 2 and caspase 3 has been reported to occasionally occur in cells undergoing MC (3)
. Hence, the discrepancies among results from different groups render the scenario depicting the MC pathway puzzling.
MC can be induced by premature induction of mitosis, before the completion of S or G2 phases (3)
, by the overduplication of centrosomes leading to multipolar mitosis (8)
or by a failure in centrosome duplication that consequently hinders chromosome segregation (9)
. According to common wisdom, MC can be promoted by agents that perturb the cell cytoskeleton, specifically those that damage microtubules and disrupt the mitotic spindle. For example, the drug paclitaxel, an agent originally derived from taxol plants and capable of permanently stabilize microtubules by blocking their depolymerization, induces an abnormal metaphase in which the sister chromatids fail to segregate properly. Both microtubule-hyperpolymerizing (taxanes, elutherobins, epothilones, laulimalide, sarcodictyins, docodermolide) and microtubule-depolymerizing agents (Vinca alkaloids, cryptophyscins, halichondrins, estramustine, and colchicine) can provoke MC (1)
. It is still unknown whether the actin microfilaments, in addition to microtubules, can play a role in setting off MC.
Cytoskeletal activity, including actin assembly and microtubule dynamics, is chiefly controlled by proteins belonging to the Rho GTPase family (encompassing the Rho, Rac, and Cdc42 subfamilies) (10)
. These regulatory G-proteins represent a preferential target for an increasing number of bacterial protein toxins, including the cytotoxic necrotizing factor 1 (CNF1) from Escherichia coli (11)
. In HEp-2 cells, CNF1 has been reported to permanently activate the Rac GTPase, leading to a series of cell responses that include NF-
B activation (12)
and protection against apoptosis (13)
, thus hijacking the host cell fate toward survival.
HOW DOES CNF1 HANDLE THE HOST CELL FATE?
It is largely established that CNF1 administration to different epithelial cells in culture results in the formation of giant multinucleated cells bearing prominent ruffling membranes (14)
, phenomena that become particularly evident after 24–48 h. A deeper investigation of the processes switched on by CNF1, which lead to multinucleation in epithelial cells, allowed us to depict a complex scenario. In fact, by different morphological approaches we observed nuclear constrictions and budding that probably gave raise to small nuclear bodies, all features suggesting the occurrence of an "amitotic nuclear division" (Fig. 1
A, B, inset). Strikingly, together with normal interphase nuclei, dispersed chromosomes could also be evidenced in multinucleated cells (Fig. 1C
). This last feature was suggestive of a defect in the mechanism that controls chromosome condensation. Furthermore, the presence of normally shaped and segmented nuclei in cells challenged with CNF1 was accompanied by the occurrence, at later times (starting from 48 h of toxin incubation), of a number of multipolar metaphases, as evidenced by microtubule labeling by immunofluorescence microscopy (Fig. 1D, F
). Consistent with this, semi-thin sections of mitotic cells confirm the ability of the toxin to alter the mitotic process, metaphasic chromosomes being directed toward asymmetrical positioned points of attraction (Fig. 1E, G
, hypothetically corresponding to the asters shown in Fig. 1D, F
). Multinucleation, multipolar mitoses, as well as nuclear budding have also been quantitatively evaluated after 72 h of exposure to the toxin (Fig. 1H
). The results clearly showed a significant increase in altered cells that resulted from the prolonged incubation with CNF1. In accordance with the above results, analysis of isolated chromosomes evidenced the toxin capacity of inducing an imbalance in chromosomal partition between daughter cells (Fig. 1I-L
). In fact, although in chromosomal preparations from untreated cells only normal diploid chromosomal patterns, e.g., n = 22 chromosomes in CHO hamster cells (Fig. 1I
) and occasional tetraploid patterns (not shown) were detected, CNF1 treatment led to aberrant chromosomal aggregations, consisting of supernumerary chromosomes (Fig. 1J
, n=34 chromosomes) and asymmetric chromosomal partitioning, e.g., n = 6 or n = 16 chromosomes (Fig. 1K, L
, respectively). This could be suggestive of a derangement in normal mitotic spindle formation and subsequent asymmetric chromosomal segregation in daughter cells. Surprisingly, CNF1-treated cells could undergo the final steps of cell division (i.e., cytokinesis and fragmoplast formation) as viewed by tubulin staining that evidenced what remained of the mitotic spindles at the cleavage furrow (Fig. 1M, N
). Transmission electron microscopy analysis showed that such an asymmetrical division occurred between cells that, at least morphologically, remained healthy (Fig. 1O
). Although the final destiny of cells exposed to the toxin is still undefined, a clue comes from the observation of typical signs of autophagy (i.e., the presence of autophagic vacuoles in the cell cytoplasm of multinucleated cells) (Fig. 1P, Q
) after prolonged exposure to CNF1 (4–6 days). The same area was rich in the lysosomal-associated protein Lamp-1 (Fig. 1R
), a typical recognized marker of acidic vacuoles in autophagic cells (15)
.
|
UNRAVELING THE COMPLEX CELL RESPONSES TO CNF1: A HYPOTHESIS
Far from being a mere end point, cell death is an essential and highly orchestrated process whose defective execution is at the basis of many major human health foes, including cancer. CNF1 has been indicated as a possible transforming agent (16)
since, despite its name, it is a survival factor able to 1) block apoptosis (17)
, 2) promote cell processes associated with anoikis resistance (13)
, and 3) provoke the appearance of multipolar mitoses that, however, did not end in MC (this paper). As hypothesized in Fig. 2
, the long-lasting activation of the Rac GTPase by CNF1 can play a role in the disturbance of mitotic spindle movements as well as in the alteration of cytokinesis, in accordance with evidence that implicates both microfilaments and microtubules as targets of the toxin (18
, 19)
. Since Rac (and Rho) molecules contribute to chromosomal segregation and migration at the cell poles, and also to control the contractile ring dynamics during cytokinesis (10)
, it is conceivable that CNF1 activity could result in a defective supervision of the physiological mitosis-associated cellular changes. Moreover, in the context of small GTPases, an effect of CNF1 on the Ran GTPase involved in the nuclear changes occurring in mitosis (20)
cannot be ruled out. It remains unclear, however, how the sustained CNF1-induced activation of Rac that deceptively instructs the complex machinery leading to mitosis can subsequently pass over cell death, resulting in an abortive MC. This failure is consistent with the assertion that MC suppression can somehow be linked to the onset of aneuploidy (3
, 4
, 21)
. Fittingly, we show here that the prolonged incubation with CNF1 resulted in an altered chromosome distribution as well as in an aberrant asymmetric cell division into daughter cells, bearing a severe inequality in the number of nuclei. These findings are doubtless linked to aneuploidy, which is known to participate in oncogenesis (22)
, and also to published data attributing a role to bacterial protein toxins in cell neoplastic transformation and cancer onset (16)
. If aneuploidy and cell transformation can indicate a propensity to an aberrant survival, the impeded priming to death by apoptosis or MC raises the question of how cells challenged with CNF1 can undergo death in the long run. A part of the long-lasting CNF1-treated cells shows hallmarks of autophagy. The role of autophagy in cell destiny, however, is still controversial (23)
. In fact, besides representing a peculiar form of cell death, autophagy also appears crucial in the mechanisms leading to cancer cell survival, tumor formation, and progression (24
, 25)
. Hence, CNF1 influences cell fate apparently by a multifaceted activity that consists in the 1) hindrance of apoptosis (13
, 17)
, 2) impairment of the "normal" mitotic process (resulting in multinucleated cell formation and mitotic catastrophe inhibition; ref 14
and this work, respectively), and 3) the promotion of autophagy (this work). In a word, CNF1 acts by hijacking cells toward a "dangerous survival."
|
CONCLUSIONS
Hence, the complex scenario depicted herein that tentatively connects multinucleation, MC, autophagy, and apoptosis in cells challenged with a Rac-activating toxin is in keeping with the hypothesis (16
, 26)
that certain toxins from bacterial pathogens can contribute to the scrambling of cell life and death in a tissue, incidentally playing a role in cancer onset (Fig. 2)
.
Received for publication October 9, 2005. Accepted for publication November 29, 2005.
REFERENCES
Balpha to the ruffling membranes. Mol. Biol. Cell 15,1124-1133
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
