*Assayed by immunocytochemistry. ^Assayed by electron microscopy.

*Assayed by immunocytochemistry. ^Assayed by electron microscopy.

microscopic analyses show that heat shock can lead to increased density of pericen-triolar material (PCM) [6-8] or complete disintegration of centrosomes [9].

The loss of centrosome-associated protein localization after heat shock, as detected by immunolabeling, could reflect protein dissociation from a centrosome scaffold, protein degradation, protein denaturation or aggregation, or a combination of these factors. For the most part, the mechanism of heat shock-induced loss of centrosome protein labeling is not understood. Localization of the centroso-mal protein CP190 is reduced after heat shock, but protein levels on Western blots do not change [4]. Therefore, at least in this instance, diminished labeling does not appear to reflect protein degradation.

Recovery of centrosome function following heat shock, by contrast, may be directly linked to changes in heat shock protein expression or function. A universal response to heat shock is the rapid and transient increase in expression of a small number of heat shock proteins or Hsps. Hsps belong to a large family of both constitutive and stress-induced "chaperones" that play an important role in mediating protein folding, transport and assembly-disassembly of polypeptide complexes (for reviews see [10-12]). Cells exposed to a relatively short heat treatment become "thermotolerant", and show increased survival in response to a second heat shock. The capacity of the centrosome to re-grow microtubules and the efficiency of centrosomal staining by anti-pericentrin antibody recover more efficiently in thermotolerant cells. Furthermore, the inducible form of Hsp72 accumulates at centrosomes following heat shock [13]. Heat shock proteins thus appear to play a role in restoring centrosome function following heat stress, perhaps by promoting assembly of multi-component centrosome protein complexes.

The ability to rapidly recover centrosome function following heat shock correlates with cell survival, and this may reflect the importance of centrosome function during mitosis. Thermotolerant cells recover centrosome function more efficiently than naïve cells. These cells also show enhanced clonogenic survival and less frequent centrosome and spindle abnormalities after heat shock [8]. By contrast, following heat shock, naïve cells often progress through an aborted division and die by a non-apoptotic pathway. Mitotic cells with multiple centrosomes and abnormal spindles are more frequently observed in heat-shocked tumor cells than in heat-shocked non-tumor controls, and the increase in abnormal spindles correlates with increased cell death by a non-apoptotic mechanism [14]. Cell death following an aborted mitotic division, termed "mitotic catastrophe", is also commonly observed following DNA damage (for a review see [15]). Cell death by mitotic catastrophe eliminates cells from the population. Centrosome disruption may therefore function to eliminate heat shock-damaged cells from a normal cell population.

Several lines of evidence suggest that heat shock proteins are also required for normal centrosome assembly. A number of heat shock proteins co-localize with centrosomes under normal growth conditions, including Hsp90, TCP-1, Hsp73, and Hsp70 [13, 16, 17]. Furthermore, isolated Drosophila centrosomes contain Hsp90, as determined by MALDI mass spectrometry, and Hsp90 localizes to cen-trosomes throughout the cell cycle, and at different stages of development. Significantly, Hsp90 mutations disrupt centrosome organization and lead to assembly of aberrant spindles and impaired chromosome segregation [18]. Hsp90 is required for the stability of Polo kinase, which regulates several aspects of cell division, including centrosome maturation and function. Hsp90 may therefore promote centrosome function by maintaining Polo kinase activity [19]. Moreover, in Xenopus oocyte extracts an Hsp70/Hsp90 complex appears to directly sequester centrin, which is released when the centrosome assembles on oocyte activation [20]. Heat shock proteins may therefore promote centrosome function by directly interacting with structural components of this organelle and by stabilizing key regulators of centro-some function.

Heat shock appears to have differential effects on centrosome organization, depending on cell-cycle phase. In HeLa cells, heat shock-induced Hsp72 localization to centrosomes is most pronounced during mitosis [13], and we have recently found that heat shock-induced loss of y-tubulin-GFP from the centrosome is also more dramatic during mitosis (O. C. M. Sibon, unpublished data). Furthermore, in Drosophila cultured cells, heat shock does not alter the ultrastructure of interphase centrioles, but leads to severe defects in centriole organization in mitotic cells [9]. Mitotic centrosomes thus appear to be particularly sensitive to heat stress. This may function to trigger mitotic catastrophe, which then disposes of cells that sustain irreparable heat damage.

Centrosomes and the Unfolded Protein Response

A potentially related form of stress is triggered by accumulation of unfolded proteins, which leads to formation of dense protein aggregates, called "aggresomes", which localize to the vicinity of centrosomes. Under normal physiological condi tions, quality control pathways catalyze proper folding of nascent proteins and degrade improperly-folded polypeptides and protein aggregates [21, 22]. However, when the quality control mechanisms are disrupted or overloaded, through a pathological condition or experimental manipulation, improperly-folded proteins accumulate in the cytoplasm and aggregates often form in the pericentrosomal area [23] (for reviews see [24, 25]). Johnston and colleagues [23] first described these structures in cells over-expressing the cystic fibrosis transmembrane conductance regulator (CFTR). In these cells, CFTR accumulated in a pericentriolar structure surrounded by a cage ofvimentin filaments. Inhibition of proteasome function induces deposition of other proteins in similar aggresomes, which are also associated with centrosomes. Aggresome formation is dependent on intact micro-tubules, suggesting that clustering around the centrosome is an active process. Johnston and coworkers proposed that aggresome formation is a general response to accumulation of unfolded proteins, which occurs when proteasome capacity is exceeded [23].

Consistent with this hypothesis, aggresome-like clusters of insoluble/misfolded proteins that co-localize with the centrosomal marker y-tubulin have now been observed in a number of situations, including human diseases and disease models. Intra-cytoplasmic protein aggregates (or Lewy bodies) that are found in neurodegenerative disorders such as Parkinson's disease and dementia co-localize with y-tubulin-containing structures [26]. A mutant form of the prion protein associated with transmissible spongiform encephalopathies also forms aggresome-like clusters that co-localizes with y-tubulin [27]. A missense mutation in aB-crystallin (aB), the cause of a desmin-related myopathy, causes aggregates of desmin and aB in aggresomes concentrating at and around centrosomes as determined using y-tubulin as a centrosomal marker [28]. A mutation in the FERM domain of schwannomin, the product of the NF2 tumor suppressor gene, also causes mis-folding and accumulation in aggresome-like structures that again co-localize with structures that are immunoreactive for y-tubulin [29]. It is unclear if these structures contribute to disease progression or are a secondary consequence of the disease state. However, aggresome formation appears to be a common consequence of accumulation of insoluble proteins.

Active proteasomal complexes localize to centrosomes under basal conditions [30, 31]. The centrosome may therefore represent a primary site for degradation of misfolded proteins. Consistent with this speculation, centrosome-associated ag-gresomes enlarge when proteosome activity is inhibited by drugs [23, 26, 27, 32].

Under some condition, the distribution of y-tubulin appears to be altered when aggresomes form. y-Tubulin and pericentrin are present at aggresomes in cells from patients with Parkinson's disease and other dementias that produce Lewy bodies [26]. However, in some cells containing mutant schwannomin aggregates y-tubulin labeling is absent [29]. Garcia-Mata et al. reported disruption of astral microtubule organization around the aggresomes, consistent with defects in the y-TuRC [33]. However, at the electron microscopic level, aggresomes appear electron dense and are surrounded by a cage of intermediate filaments [23, 32-34], and the interior of the aggresome is not generally immunolabeled by antibodies

[23]. It is therefore possible that y-TuRC is at the centrosome and biochemically active, but is sterically prevented from functioning in microtubule nucleation when dense aggresomes are present.

The link between aggresomes and human disease appears to be quite strong. By nucleating and organizing microtubules, the centrosome could promote aggre-some assembly, and defects in centrosome function due to aggresome formation could contribute to pathogenesis. A molecular understanding of the link between centrosomes and the unfolded protein response could shed light on a clinically intractable group of neurodegenerative diseases.

Centrosome Disruption in Response to Genotoxic Stress

Maintenance of genomic integrity is critical to normal development and disease prevention, and conserved pathways promote damage repair or eliminate mutant cells from the population. DNA damage and replication checkpoints delay the cell cycle to allow repair of genetic lesions or completion of DNA replication. In systems ranging from mammalian tumors to early Drosophila embryos, checkpoint failures that allow DNA damage or incomplete replication to persist into mitosis trigger "mitotic catastrophe", a poorly understood process characterized by delays in metaphase, chromosome segregation failures, and cell death by non-apoptotic mechanisms[15, 35-37]. The molecular mechanism of mitotic catastrophe has not been analyzed in detail, but this process appears to be a significant cause of chemotherapy-induced cell death in tumors and may serve an important genome maintenance function (reviewed in[15]) Studies in Drosophila embryos demonstrate that checkpoint failures and "mitotic catastrophe" are linked to mitosis-specific defects in centrosome structure and function, anastral mitotic spindle assembly, and chromosome segregation failures on mitotic exit [36]. The Drosophila homolog of the Chk2 tumor suppressor kinase is essential to mitotic catastrophe in early embryos, demonstrating that this is a genetically programmed response to genotoxic lesions [38]. The mitotic response to DNA damage in early Drosophila embryos is reviewed below, followed by a discussion of recent studies in cultured mammalian cells.

Was this article helpful?

0 0

Post a comment