Microtubule-interacting drugs induce moderate and reversible damage to human bone marrow mesenchymal stem cells.

OBJECTIVES: This study aimed to investigate molecular and cellular changes induced in human bone marrow mesenchymal stem cells (hMSCs) after treatment with microtubule-interacting agents and to estimate damage to the bone marrow microenvironment caused by chemotherapy. MATERIALS AND METHODS: Using an in vitro hMSC culture system and biochemical and morphological approaches, we studied the effect of nocodazole and taxol(R) on microtubule and nuclear envelope organization, tubulin and p53 synthesis, cell cycle progression and proliferation and death of hMSCs isolated from healthy donors. RESULTS AND CONCLUSIONS: Both nocodazole and taxol reduced hMSC proliferation and induced changes in the microtubular network and nuclear envelope morphology and organization. However, they exhibited only a moderate effect on cell death and partial arrest of hMSCs at G(2) but not at M phase of the cell cycle. Both agents induced expression of p53, exclusively localized in abnormally shaped nuclei, while taxol, but not nocodazole, increased synthesis of beta-tubulin isoforms. Cell growth rates and microtubule and nuclear envelope organization gradually normalized after transfer, in drug-free medium. Our data indicate that microtubule-interacting drugs reversibly inhibit proliferation of hMSCs; additionally, their cytotoxic action and effect on microtubule and nuclear envelope organization are moderate and reversible. We conclude that alterations in human bone marrow cells of patients under taxol chemotherapy are transient and reversible.

Intraperitoneal chemotherapy with taxanes for ovarian cancer with peritoneal dissemination.

Paclitaxel and docetaxel are currently the two clinically available taxanes. The combination of a taxane and a platinum compound has become the systemic chemotherapy of choice for primary ovarian cancer. Despite the high activity of these drugs in systemic chemotherapy, the majority of patients with advanced ovarian cancer will develop recurrent disease and ultimately decease of this disease. Therefore, more effective systemic chemotherapy regimens or alternative treatment modalities are warranted. Intraperitoneal chemotherapy is such an alternative treatment option. Pharmacokinetic studies on intraperitoneal administration of paclitaxel and docetaxel demonstrated very high locoregional drug concentrations and exposure. Their activity and response seem to be dose-dependent and hence higher efficacy with limited systemic toxicity is to be expected. Intraperitoneal chemotherapy may be combined intraoperatively with hyperthermia, which enhances tissue penetration and cytotoxic activity of many drugs. The data concerning thermal enhancement of taxanes are inconsistent, but at the high locoregional concentrations provided by intraperitoneal drug administration such a thermal enhancement seems to exist. Clinical studies have clearly demonstrated the feasibility and efficacy of intraperitoneal instillation chemotherapy with taxanes in patients with ovarian cancer. Preliminary results of a phase III study demonstrated improved outcome with the addition of intraperitoneal instillation chemotherapy to systemic chemotherapy after optimal primary cytoreductive surgery. Intraoperative hyperthermic intraperitoneal chemotherapy with docetaxel has been performed in a single study, in which promising results were observed. Further clinical investigations with an adequate follow-up period are needed to confirm the promising initial results and to determine the exact efficacy of intraperitoneal chemotherapy with these drugs.

Histones H3/H4 form a tight complex with the inner nuclear membrane protein LBR and heterochromatin protein 1.

We have recently shown that heterochromatin protein 1 (HP1) interacts with the nuclear envelope in an acetylation-dependent manner. Using purified components and in vitro assays, we now demonstrate that HP1 forms a quaternary complex with the inner nuclear membrane protein LBR and a sub-set of core histones. This complex involves histone H3/H4 oligomers, which mediate binding of LBR to HP1 and cross-link these two proteins that do not interact directly with each other. Consistent with previous observations, HP1 and LBR binding to core histones is strongly inhibited when H3/H4 are modified by recombinant CREB-binding protein, revealing a new mechanism for anchoring domains of under-acetylated chromatin to the inner nuclear membrane.

Binding of heterochromatin protein 1 to the nuclear envelope is regulated by a soluble form of tubulin.

We have previously shown that the mouse heterochromatin protein 1 homologue M31 interacts dynamically with the nuclear envelope. Using quantitative in vitro assays, we now demonstrate that this interaction is potently inhibited by soluble factors present in mitotic and interphase cytosol. As indicated by depletion and order-of-addition experiments, the inhibitory activity co-isolates with a 55-kDa protein, which binds avidly to the nuclear envelope and presumably blocks M31-binding sites. Purification of this protein and microsequencing of tryptic peptides identify it as alpha2/6:beta2-tubulin. Consistent with this observation, bona fide tubulin, isolated from rat brain and maintained in a nonpolymerized state, abolishes binding of M31 to the nuclear envelope and aborts M31-mediated nuclear envelope reassembly in an in vitro system. These observations provide a new example of "moonlighting," a process whereby multimeric proteins switch function when their aggregation state or localization is altered.

Dynamic associations of heterochromatin protein 1 with the nuclear envelope.

To study the dynamics of mammalian HP1 proteins we have microinjected recombinant forms of mHP1alpha, M31 and M32 into the cytoplasm of living cells. As could be expected from previous studies, the three fusion proteins were efficiently transported into the nucleus and targeted specific chromatin areas. However, before incorporation into these areas the exogenous proteins accumulated in a peripheral zone and associated closely with the nuclear envelope. This transient association did not occur when the cells were treated with deacetylase inhibitors, indicating an acetylation-inhibited interaction. In line with these observations, recombinant HP1 proteins exhibited saturable binding to purified nuclear envelopes and stained the nuclei of detergent-permeabilized cells in a rim-like fashion. Competition experiments with various M31 mutants allowed mapping of the nuclear envelope-binding site within an N-terminal region that includes the chromodomain. A His(6)-tagged peptide representing this region inhibited recruitment of LAP2beta and B-type lamins around the surfaces of condensed chromosomes, suggesting involvement of HP1 proteins in nuclear envelope reassembly.

Taxol affects nuclear lamina and pore complex organization and inhibits import of karyophilic proteins into the cell nucleus.

Treatment of human carcinoma cells with Taxol induces focal unraveling of the nuclear lamina and extensive clustering or ectopic localization of the nuclear pore complexes. These striking aberrations develop when the cells are transferred to drug-free medium and are allowed to complete mitosis. As could be confirmed by terminal deoxynucleotidyl transferase-mediated nick end labeling assays, 4,6-diamidino-2-phenylindole staining, 5-bromo-2-deoxyuridine incorporation, and examination of the nuclear lamins by Western blotting, the malformation of the nuclear envelope is not a consequence of apoptosis or G1 arrest. In fact, Taxol-treated cells possessing a defective nuclear envelope remain alive and replication-competent for at least 24 h, undergoing programmed death 72 h after removal of the drug. While still in the nonapoptotic state, these cells lose the ability to import karyophilic proteins into the nucleus. Diminished nucleocytoplasmic transport through the nuclear pore complex can be readily demonstrated by in vitro assays involving digitonin-permeabilized cells or in vivo monitoring of nuclear factor-kappaB translocation upon stimulation with tumor necrosis factor-alpha. These observations reveal novel cellular targets of antimicrotubule drugs and may pave the way for improved schemes of anticancer treatment.

PBC68: a nuclear pore complex protein that associates reversibly with the mitotic spindle.

Using autoimmune antibodies from a patient with primary biliary cirrhosis we have identified a 68 kDa nuclear envelope protein, termed PBC68. This protein is co-precipitated with a 98 kDa and a 250 kDa polypeptide and is distinct from the nuclear lamins. Immunostaining of digitonin-permeabilized cells indicates that PBC68 is restricted to the inner (nucleoplasmic) face of the nuclear envelope, while indirect immunofluorescence and immunoelectron microscopy show that PBC68 is located on fibrillar structures emanating from the nuclear pore complex. The autoantigen is modified at early prophase and disassembles at prometaphase concurrently with the breakdown of the nuclear envelope. The disassembled material, instead of diffusing throughout the cytoplasm as other nucleoporins, is targeted to the mitotic spindle and remains stably bound to it until anaphase. At telophase PBC68 is released from the mitotic apparatus and reassembles late, after incorporation of LAP2B and B-type lamins, onto the reforming nuclear envelope. The partitioning of PBC68 in dividing cells supports the notion that subsets of nuclear envelope proteins are actively sorted during mitosis by transiently anchoring to spindle microtubules. Furthermore, the data suggest that specific constituents of pore complex are released in a stepwise fashion from their anchorage sites before becoming available for nuclear reassembly.

Rules to remodel by: what drives nuclear envelope disassembly and reassembly during mitosis?

In higher eukaryotic cells the nuclear envelope is reversibly disassembled during mitosis. Under in vivo conditions this process occurs in a sequential, stepwise fashion and involves a variety of structural intermediates. Here we discuss the topological features of these intermediates and their transient interactions with chromatin and the cytoskeleton. As it becomes apparent, nuclear envelope disassembly and reassembly are regulated at multiple levels by modulating the affinity of protein-protein interactions, limiting the availability of structural subunits in different areas of the mitotic cytoplasm, and redirecting mechanical forces exerted by the microtubules.

Nuclear envelope breakdown in mammalian cells involves stepwise lamina disassembly and microtubule-drive deformation of the nuclear membrane.

We have studied nuclear envelope disassembly in mammalian cells by morphological methods. The first signs of nuclear lamina depolymerization become evident in early prophase as A-type lamins start dissociating from the nuclear lamina and diffuse into the nucleoplasm. While B-type lamins are still associated with the inner nuclear membrane, two symmetrical indentations develop on antidiametric sites of the nuclear envelope. These indentations accommodate the sister centrosomes and associated astral microtubules. At mid- to late prophase, elongating microtubules apparently push on the nuclear surface and eventually penetrate the nucleus. At this point the nuclear envelope becomes freely permeable to large ligands, as indicated by experiments with digitonin-treated cells and by the massive release of solubilized A-type lamins into the cytoplasm. At the prophase/prometaphase transition, the B-type lamina is fragmented, but 'islands' of lamin B polymer can still be discerned on the tips of congressing chromosomes. Finally, at metaphase, the lamin B polymer breaks down into small pieces, which tend to concentrate in the area of the mitotic spindle. Nuclear envelope breakdown is not prevented when the microtubules are depolymerized by nocodazole; however, the mode of nuclear lamina fragmentation in the absence of microtubules is markedly different from the normal one and involves multiple raffles and gaps, which develop rapidly along the entire surface of the nuclear envelope. These data suggest that nuclear envelope disassembly is a stepwise process in which the microtubules play an important part.

The inner nuclear membrane protein LAP1 forms a native complex with B-type lamins and partitions with spindle-associated mitotic vesicles.

We have examined the in situ organization and nearest neighbours of the 'lamina-associated polypeptide-1' (LAP1), a type II membrane protein and a major constituent of the mammalian nuclear envelope. We show here that, during interphase, LAP1 forms multimeric assemblies which are suspended in the inner nuclear membrane and are specifically associated with B-type lamins. The LAP1-lamin B complex is distinct from analogous complexes formed by the 'lamina-associated polypeptide-2' (LAP2), another inner nuclear membrane protein, and includes a protein kinase. Upon nuclear envelope breakdown, LAP1 partitions with mitotic vesicles which carry nuclear lamin B. The LAP1 vesicles can be distinguished from fragments of the nuclear envelope containing LAP2 and exhibit a striking co-alignment with spindle microtubules. These observations suggest that the inner nuclear membrane comprises discrete territories which accommodate specific integral membrane proteins and are differentially disassembled during mitosis.

Self-inflicted traumatic macroglossia.

Macroglossia is a potentially life-threatening problem. The multiple aetiologies responsible for this condition include cysts, tumours, amyloidosis, angioedema, and infections. Although injuries to the oral cavity are common, self-inflicted traumatic macroglossia is not common. We report a case of a comatose patient with limb contractures and severe trismus who developed entrapment of the tongue with progressive macroglossia and near auto-amputation of the tongue. The importance of early recognition and treatment of tongue entrapment in comatose patients is emphasised since this is a potentially reversible condition.

Altered actin polymerization dynamics in various malignant cell types: evidence for differential sensitivity to cytochalasin B.

Using the DNase I inhibition assay, fluorimetric measurements, and immunoblot analysis, we studied quantitatively changes in the actin polymerization dynamics in primary cultures of normal and malignant human lymphocytes, normal human endometrial cells, and in various leukemic and endometrial adenocarcinoma cell lines. The G/total-actin ratio of malignant cells was found to be 1.37 to 1.81-fold higher compared to normal cells, indicating that malignant cells express reduced amounts of polymerized actin. The above findings were corroborated by fluorescence measurements of the amounts of rhodamine-phalloidin-labeled F-actin in normal and neoplastic cells, which showed significantly lower F-actin content in malignant cell preparations. Moreover, the total actin content, as quantitated by the DNase I inhibition assay and by immunoblot analysis, was found to be significantly decreased in the primary cultures of malignant human lymphocytes and endometrial cells when compared to the total actin levels in corresponding normal cells. Proliferation and viability measurements of normal and neoplastic cells in culture, treated equally with cytochalasin B (CB), revealed an increased susceptibility of malignant cells to this anticytoskeletal agent. This was not due to increased CB incorporation in neoplastic cells, as indicated by 3H-CB uptake experiments. In addition, fluorescence microscopy, in the presence of graded concentrations of CB, showed destabilization of microfilaments in the poorly differentiated endometrial adenocarcinoma HEC-50 cells, compared to the well-differentiated Ishikawa cells. In conclusion, all investigated malignant cells are characterized by: (a) higher G/total-actin ratio; (b) decreased F- and total-actin content; and (c) lower resistance to CB treatment. These quantitatively determined parameters may represent potential biochemical indicators reflecting malignant transformation. Moreover, it seems worthwhile to explore whether or not the differential sensitivity of malignant cells to anticytoskeletal drugs may provide a valuable approach to the manipulation of malignant cells.

Dexamethasone alters rapidly actin polymerization dynamics in human endometrial cells: evidence for nongenomic actions involving cAMP turnover.

Glucocorticoids, in addition to their well characterized effects on the genome, may affect cell function in a manner not involving genomic pathways. The mechanisms by which the latter is achieved are not yet clear. A possible means for this action may involve the actin cytoskeleton, since the dynamic equilibrium of actin polymerization changes rapidly following exposure to several stimuli, including hormones. The aim of the present work was to find out if glucocorticoids exert rapid, nongenomic effects on actin polymerization in Ishikawa human endometrial cells, which represent a well characterized in vitro cell model expressing functional glucocorticoid receptors. Short term exposure of the cells to the synthetic glucocorticoid dexamethasone resulted in an overall decrease of the G/total-actin ratio in a time- and dose-dependent manner. Specifically, in untreated Ishikawa cells the G/total-actin ratio was 0.48 +/- 0.01 (n = 26). It became 0.35 +/- 0.01 (n = 13, P < 0.01) following exposure to 10(-7) M dexamethasone for 15 min. This was induced by a significant decrease of the cellular G-actin level, without affecting the total actin content, indicating a rapid actin polymerization. This conclusion was fully confirmed by direct fluorimetry measurements, that showed a significant increase of the F-actin content by 44% (n = 6, P < 0.001) in cells treated with dexamethasone (10(-7)M, 15 min). The rapid dexamethasone-induced alterations of the state of actin polymerization were further supported by fluorescence microscopy. The latter studies showed that the microfilaments of cells pretreated with 10(-7)M dexamethasone for 15 min were more resistant to various concentrations of the antimicrofilament drug cytochalasin B, compared to untreated cells, implying microfilament stabilization. The action of dexamethasone on actin polymerization seems to be mediated via specific glucocorticoid binding sites, since the addition of the glucocorticoid antagonist RU486 completely abolished its effect. Moreover, it appears to act via non-transcriptional pathways, since actinomycin D did not block the dexamethasone-induced actin polymerization. In addition, cell treatment with 10(-7)M dexamethasone for 15 min fully reversed the forskolin-, but not the 8-bromo-cAMP-induced actin depolymerization. In line with these findings, the cAMP content of Ishikawa cells was decreased by 29.2% after a 15 min treatment with 10(-7)M dexamethasone (n = 4, P < 0.01). In conclusion, our results showed that dexamethasone induces rapid, time-, and dose-dependent changes in actin polymerization dynamics in Ishikawa cells. This action seems to be mediated via cAMP, involving probably nongenomic pathways. The above findings offer new perspectives for the understanding of the early cellular responses to glucocorticoids.

Differences in the G/total actin ratio and microfilament stability between normal and malignant human keratinocytes.

The state of polymerization of actin and the organization of actin filaments is widely believed to be related to cellular transformation. Since the intracellular monomer (G) and filamentous (F) actin content reflects the state of microfilament polymerization, we measured the G/total actin ratio in primary cultures of normal and malignant human keratinocytes. In normal keratinocytes the mean value of this ratio was 0.30 +/- 0.03 (mean +/- SE, n = 15), while in basal cell carcinoma (BCC) keratinocytes it was 0.49 +/- 0.03 (n = 8) and in squamous cell carcinoma keratinocytes (SCC) 0.5 +/- 0.07 (n = 4), indicating a 1.7-fold increase of the G/total actin ratio in malignant cells. These results imply that the proportion of polymerized actin is decreased markedly in malignant keratinocytes, suggesting alterations of microfilament structures which probably occur during the transformation process. This was supported by the morphological changes of microfilament structures as assessed by fluorescence microscopy. A different distribution of actin filaments in normal and malignant cells became evident; stress-fibres were converging in patches at several points in SCC cells, when compared to normal keratinocytes. Furthermore, incubation of normal and malignant keratinocytes with cytochalasin B indicated differences in the resistance of their microfilament networks. After 1 h exposure to 10(-6) and 10(-5) M cytochalasin B, microfilaments in normal cells appeared to be less affected than their counterparts in neoplastic cells. Even in a high excess of cytochalasin B (10(-4) M), normal keratinocytes preserved their shape, while both basal cell and SCC were totally disrupted.(ABSTRACT TRUNCATED AT 250 WORDS)

Cytochalasin B may shorten actin filaments by a mechanism independent of barbed end capping.

It is generally accepted that cytochalasin B (CB), as well as other cytochalasins, shorten actin filaments by blocking monomer addition at the fast-growing ("barbed") end of these polymers. Despite the predominance of this mechanism, recent evidence suggests that other interactions may also occur between CB and F-actin. To investigate this possibility further we have employed an actin derivative, prepared by substitution at Cys374 by a glutathionyl residue. We demonstrate here that CB did not significantly bind to glutathionyl F-actin under several ionic conditions. We further show that in the presence of CB the glutathionyl-F-actin exhibits a significantly higher ATPase activity than the non-modified F-actin. These data argue that the incorporation of glutathionyl groups prevents the high-affinity binding of CB to the barbed end of actin filaments, probably due to a decreased hydrophobicity of the CB binding site by the introduction of the hydrophilic glutathionyl residue. Despite the lack of substantial binding at equilibrium, we have found that the addition of CB to glutathionyl-F-actin results in extensive fragmentation of the filaments, as demonstrated by electron microscopy and by a significant reduction of the relative viscosity of actin solutions. These results are consistent with the idea that CB shortens glutathionyl-actin filaments by a mechanism distinct from barbed end capping. Glutathionyl F-actin offers an interesting model to study the complex mechanism of interaction of actin filaments with cytochalasins and with the physiologically important actin capping/severing proteins.

Effect of hepatocyte swelling on microtubule stability and tubulin mRNA levels.

Incubation of isolated rat hepatocytes under conditions known to induce cell swelling caused several alterations in microtubule physiology. As shown by immunofluorescence microscopy experiments in the absence and presence of triethyllead or colchicine (two well-established microtubule inhibitors), an apparent stabilization of the microtubule network became evident in hepatocytes exposed to hypotonic (190 mosmol/L) conditions. A similar stabilizing effect was also observed upon cell swelling induced by addition of insulin (100 nmol/L) or glutamine (10 mmol/L). The differential microtubule stabilities were not attributed to a differential incorporation of the antimicrotubular agents into hepatocytes as shown by [3H]colchicine-uptake experiments. The swelling-induced alterations of microtubules may contribute to the swelling-induced changes of liver cell function: in perfused rat liver it was found that the established inhibitory effect of hypotonic cell swelling on hepatic proteolysis was largely abolished in presence of colchicine. Tubulin mRNA levels increased by 1.9-, 2.1- and 2.7-fold in isolated hepatocytes being exposed for 120 min to hypotonic medium, insulin, or glutamine, respectively. The results suggest an involvement of microtubular structures in the regulation of liver metabolism in response to alterations of the cellular hydration state.

Normal and Ha-ras-1 oncogene transformed Buffalo rat liver (BRL) cells show differential resistance to cytoskeletal protein inhibitors.

In the present study, using immunofluorescence microscopy, we have demonstrated that normal and Ha-ras-1 transformed Buffalo rat liver (BRL) cells which were exposed to cytoskeletal protein inhibitors, showed a differential resistance of their microfilament and microtubule networks. One hour exposure of normal BRL cells to 10(-5) M cytochalasin B provoked a clear and already total breakdown of actin filaments. However, at this concentration of cytochalasin B, the microfilaments of transformed BRLHO6T1-1 cells were not seriously affected; a higher cytochalasin B concentration (> or = 2 x 10(-5) M) was required to induce a significant breakdown of microfilaments in these transformed cells. The two cell lines also demonstrated differential microtubule stability when they were treated with either colchicine or triethyllead. Three hours exposure to 10(-6) M of either antimicrotubule agents was sufficient to disrupt the microtubules of normal BRL cells, without affecting their counterparts in the transformed BRLHO6T1-1 cells. A 10-fold higher drug concentration (10(-5) M) was required to induce microtubular breakdown in the transformed BRL cells. The differential stability of microfilaments and microtubules in normal and transformed BRL cells that was observed could not be attributed to a differential internalization of the agents, as shown by experiments on the uptake of [3H]-cytochalasin B and triethyllead. In addition, the transformed BRLHO6T1-1 cells did not express altered actin and tubulin isoforms, as demonstrated by isoelectric focusing followed by immunoblotting analysis. We conclude that the transformation of BRL cells with the Ha-ras-1 oncogene results in a greater stability of microfilaments and microtubules, leading to a structurally firmer cell shape.(ABSTRACT TRUNCATED AT 250 WORDS)

Hepatocyte swelling leads to rapid decrease of the G-/total actin ratio and increases actin mRNA levels.

Exposure of isolated rat hepatocytes to hypotonic (190 mosmol/l) incubation media lowered the cellular G-actin level without affecting the total actin content: here the G-/total actin ratio decreased by 15.5 +/- 1.4% (n = 7). Similar effects were observed following isotonic cell swelling by either addition of glutamine (10 mM) or insulin (100 nM), resulting in a decrease of the G-/total actin ratios by 13.5 +/- 2.1% (n = 5) and 14.1 +/- 1.1% (n = 11), respectively. The effects of hypotonic exposure, glutamine and insulin on the G-/total actin ratio largely occurred within 1 min and persisted for at least 2 h in presence of the respective effectors. After a 120 min exposure to hypotonic media, glutamine or insulin the actin mRNA levels were increased 2.4-, 2.0- and 3.6-fold, respectively. Hypertonic exposure lowered the G-/total actin ratio by only 4.9 +/- 2.5% (n = 4) and increased actin mRNA levels only 1.2-fold. There was a close relationship between glutamine- and hypotonicity-induced cell swelling and the decrease of G-/total actin ratios. The data suggest that cell swelling exerts rapid and marked effects on the state of actin polymerization and increases actin mRNA levels. Thus, cytoskeletal alterations in response to cell swelling may be involved in the regulation of hepatic metabolism by cell volume.

Heat shock protein HSP90 and its association with the cytoskeleton: a morphological study.

To investigate the cellular localization of the 90-kilodalton heat shock protein (HSP90) and its interaction with the cytoskeleton, we performed single- and double-staining immunofluorescence microscopy of cytoskeletal proteins and HSP90 in the absence and presence of cytoskeletal inhibitors. As a model, we used a human endometrial adenocarcinoma cell line (Ishikawa cells), which expresses HSP90. We confirmed the recently reported colocalization of HSP90 with microtubules. However, Ishikawa cells treated with 10(-5) M of the antimicrotubule agents colchicine or triethyl lead showed residual filamentous structures stained with anti-HSP90 antibodies, while no microtubules were visualized with anti-tubulin antibodies. In the presence of 10(-5) M cytochalasin B, the microfilament staining of the cells disappeared, while residual filamentous structures were labeled with anti-HSP90 antibodies. Furthermore, Ishikawa cells treated with 10(-5) M triethyl lead and stained with anti-HSP90 antibodies demonstrated residual filamentous structures, clearly different from those of reorganized vimentin intermediate filaments. Conversely, similar reorganized morphology of filamentous structures stained with both anti-HSP90 and anti-cytokeratins antibodies were observed when Ishikawa cells were treated with 2 x 10(-5) M cytochalasin B and 2 x 10(-5) M colchicine. HSP90 was also present in Ishikawa cell preparations of the Triton X-100 insoluble cytoskeleton. In addition, Triton-insoluble cytoskeleton treated with 10(-5). M triethyl lead and double stained with anti-HSP90 and anti-vimentin antibodies demonstrated clearly different filamentous patterns, when exposed on the same photographic plaque.(ABSTRACT TRUNCATED AT 250 WORDS)