The effect of microtubule- and microfilament-disrupting drugs on preimplantation mouse embryos

The cell cytoskeleton, consists of a complex cytoplasmic arrangement of microtubules and microfilaments which are believed to be involved in many different cell functions, such as cell movement, cell division, endo- and exo-cytosis, surface receptor mobility, etc. (see Allison, 1973). It has recently been suggested that it might also play a role in the maintenance of the meiotic block in the unfertilized mouse oocyte (Siracusa, Whittingham, Codonesu & De Felici, 1978). Drugs affecting the cytoskeleton are currently being used to investigate the role of microtubules and microfilaments in these various cellular activities. However, at high concentrations the drugs can produce non-specific effects such as the inhibition of membrane transport (Mizel & Wilson, 1972; Wilson, 1975). Thus it is important to determine the lowest effective concentration of the drugs which will produce cytoskeletal disruption with the minimum of side effects and to know whether normal cytoskeletal functions are resumed when the drugs are removed. Experiments were designed to study the effects of microfilament- and microtubule-disrupting drugs on the cleavage of preimplantation mouse embryos in vitro. Six compounds were examined; four which interfere with microtubule assembly (colchicine, colcemid, vinblastine and griseofulvin) and two which impair microfilament function (cytochalasin B and D). The reversibility of action of the drugs was also studied.

Adult female albino mice (MFI, Olac, U.K.) were superovulated with intraperitoneal injections of 5 i.u. pregnant mares’ serum gonadotrophin (PMSG Folligon, Intervet) given at 10.00 to 13.00 h and 5 i.u. of human chorionic gonadotrophin (hCG-Chorulon, Intervet) given approximately 48 h later. The females were mated with albino males of the same strain. One-cell embryos at the pronuclear stage and 2-cell embryos were flushed from the oviducts of mated females approximately 28 h and 51 h after the injection of hCG. Further details on the experimental treatments are given in the Results section. The embryos were cultured in 10 μl droplets of medium (Whittingham, 1971) under paraffin oil, at 37 °C, in a humidified atmosphere of 5 % CO₂ in air.

Vinblastine sulphate, colchicine, demecolcine (colcemid), griseofulvin, cyto-chalasin B (CB) and cytochalasin D (CD) were purchased from Sigma. Concentrated stock solutions of colchicine, colcemid and vinblastine were prepared in triple distilled water, and kept frozen in small aliquots in the dark; each aliquot was thawed and used only once. Griseofulvin and the cytochalasins were dissolved in dimethyl sulphoxide (DMSO) (BDH Analar) at 0 · 1 M and 1 mg/ml respectively; the stock solutions were also kept frozen in the dark, but were used several times after freezing and thawing without any apparent loss of potency. DMSO had no detrimental effect on cleavage when added to control groups in concentrations of up to 1·5 %.

Lumicolchicine was prepared from colchicine by a modification of a method described by Mizel & Wilson (1972). An aqueous solution of colchicine at 2 × 10⁻⁵ M was irradiated in a quartz cuvette with two long wave u.v. (366 mm) lamps (Birchover Instruments Ltd.). The transformation of colchicine into its analog lumicolchicine was monitored every 10 min with a spectrophotometer by measuring the absorbance at 350 nm (colchicine) and 261 nm (lumicolchicine). The ultraviolet irradiation was stopped when the 350 nm peak disappeared and the 261 nm peak ceased to increase.

Zygotes were removed and incubated in various drugs approxi-mately 28 h post hCG; the following morning the number of 1- and 2-cell embryos was recorded. Vinblastine appeared to be the most powerful of the four antimicrotubular agents tested (Fig. 1c): a concentration of 10^{− 8} M completely prevented the zygotes from undergoing the first cleavage. Colcemid and colchicine were tenfold less effective than vinblastine.

Fragmentation occurred in some of the embryos exposed to concentrations that only partially inhibited cleavage. This phenomenon was also observed with the cytochalasins. The fragmented embryos were excluded from the calculations because in this aberrant form of cytokinesis the embryos were neither blocked at the 1-cell stage nor had they progressed to the 2-cell stage.

Lumicolchicine (2 × 10⁻⁷M) had no effect on the first cleavage division (Table 1).

Concentrations of griseofulvin up to 3 × 10⁻⁵ M did not affect the first mitotic division (Table 2). In the presence of 1 × 10⁻⁴ M and 3 × 10⁻⁴ M griseofulvin all the zygotes completed cytokinesis, but in 9 and 31 % of the cases respectively, the mitosis yielded three daughter cells of unequal size. This probably resulted from the formation of tripolar spindles. Higher concentrations could not be used because the griseofulvin slowly precipitated out of the medium during incubation.

The effect of cytochalasin B and cytochalasin D is shown in Fig. 2. CD prevented the first cleavage at a concentration (3 × 10⁻⁷ M) approximately 35-fold lower than CB (1 × 10^{− 5}M).

Two-cell embryos removed approximately 51 h post hCG were incubated in the various drugs. At approximately 72 h post hCG, the number of embryos that had reached at least the 4-cell stage was recorded. After an additional 48 h in the continuous presence of the drug the number of embryos at the blastocyst stage was recorded. The results obtained for the second cleavage (Tables 1–4) were similar to those for the first cleavage. Concentrations of the drugs that were ineffective in preventing the first and second cleavage divisions did not significantly affect the development of 2-cell embryos to the blastocyst stage, with the exception of colchicine and griseofulvin. Colchicine 1 × 10⁻⁸ M had no effect on the first two cleavage divisions (Fig. 1 and Table 3) but prevented blastocyst formation (only 65 % reached the 8-cell stage); Imuicolchicine had no effect on cleavage and blasto-cyst formation. Most of the 2-cell embryos which cleaved in the presence of 1 × 10⁻⁴ and 3 × 10⁻⁴M griseofulvin (86 and 42% respectively, Table 2) showed blastomeres of unequal size, No blastocysts were formed in the presence of 3 × 10⁻⁴ M griseofulvin, and only 19 % in 1 × 10⁻⁴ M.

Two-cell embryos removed approximately 51 h post hCG were treated for 6 h with effective inhibitory concentrations of the drugs, and then washed and transferred to control medium to observe further development. At transfer, 88 % (range 67–100 %) of the control embryos had cleaved to 4-cell stage, while all the treated embryos were at the 2-cell stage. The next morning the treated embryos still lagged one division behind the control embryos (4 cells v. 8 cells). At approximately 120 h post hCG the number of blastocysts was scored.

Embryos treated with CB or CD appeared to recover completely, and were able to develop into blastocysts in similar proportion to the controls (Table 5). Development after a 6 h treatment with the microtubule-disruptive drugs was less satisfactory. Only 68% of the embryos treated with 3 × 10⁻⁷ M colcemid and 18 % of those treated with 3 × 10⁻⁸ M vinblastine reached the blastocyst stage v. 90 % of the controls (Table 5). Colchicine treatment was irreversible, the 2-cell embryos treated with 3 × 10⁻⁷M colchicine for 2h or more failed to resume cleavage (Table 6).

In this study we investigated the sensitivity of early preimplantation mouse embryos to microfilament- and microtubule-disruptive drugs, and the reversibility of such treatments. Two drugs which interfere with the function of microfilaments, cytochalasin B and D, were examined. The mechanism of action of the drugs has been only partially clarified. Apparently they cause a state of hypercontraction, and displacement of the cortical mesh of microfilaments, and this probably entails detachment of the connexions between microfilaments and the cell membrane (Miranda, Godman, Deitch & Tanebaum, 1974 a). This postulated detachment may explain many of the effects of the cytochalasins, including the inhibition of cell division. The effect of the cytochalasins appears to be rapidly reversible, since the microfilament web is reconstituted in the cortex within 1 h after withdrawal of CD (Miranda, Godman & Tanebaum, 1974b). CD is more powerful than CB, and unlike CB it does not interfere with hexose up-take in most cell lines (Miranda et al. Our experiments have confirmed that mole for mole CD is more potent than CB in preventing cell division (Fig. 2 and Table 4). Complete inhibition of the first cleavage in mouse embryos was obtained with a concentration of CD approximately 35-fold lower than CB (3 × 10⁻⁷ M v. 1 × 10⁻⁵ M). A similar concentration of CB (4 μg/ml) was found to be the lowest that totally prevented the morula to blastocyst transition of mouse embryos (Granholm & Brenner, 1976). As already mentioned, at concentrations immediately below the minimally effective ones, a proportion of the embryos underwent fragmentation. This phenomenon was also observed with microtubule-disruptive drugs, and it is probably due to the disorderly activity of a partly disrupted cytoskeleton.

It has been shown previously (Snow, 1973, 1975; Niermerko, 1975; Tarkowski, Witkowska & Opas, 1977) that preimplantation mouse embryos treated for 6 – 12 h with 5 – 10 μg/ml CB at the 1- or 2-cell stage, can develop to the blastocyst stage and beyond; the treatment induces polyploidy in most of the embryos. Similar findings of complete reversibility of CB effects have also been reported for a variety of cell types (Yamada, Spooner & Wessels, 1971 ; Spooner, Yamada & Wessels, 1971; Schaeffer, Schaeffer & Brick, 1973, etc.). We confirm that a 6h treatment with 1 × 10^{− 5}M CB (≃ 5μg/ml) is well tolerated by the 2-cell mouse embryos (Table 5) and have extended the observation to show that CD is equally effective at a concentration of 1 × 10⁻⁶ M (≃ 0 · 5 μg/ml). No attempt was made to study whether CD causes any of the latent effects which become visible during post-blastocyst development after CB treatment of preimplantation mouse embryos (Granholm & Brenner, 1976). The greater specificity of the action of CD and the lower concentrations needed to obtain the same effect on cell cleavage as CB, makes it a more suitable choice for experiments in which minimal damage to the embryo is essential, such as diploidization after removal of one pronucleus (Hoppe & Illmensee, 1977).

The mechanism of action of three of the drugs which interfere with micro-tubular function (colchicine, colcemid and vinblastine) is well established : they bind to tubulin and prevent the polymerization of tubulin to form micro-tubules. Since most microtubules are in equilibrium with a soluble pool of tubulin, this results in the dissolution of microtubules (see review by Wilson, Anderson, Grisham & Chin, 1975). Vinblastine is the most potent drug in the group in preventing cleavage. The minimal effective concentration of vin-biastine (1 × 10^{− 8} M) is tenfold lower than the concentration of colchicine or colcemid required to obtain the same effect (Fig. 1 and Table 3). However, the effect of colcemid was far more reversible than that of vinblastine or colchicine (Table 5). After a 6 h treatment with effective inhibitory concentrations of the drugs, the development of 2-cell embryos to blastocysts was reduced from 75 % of controls with colcemid 3 × 10⁻⁷ M ( ≃ 04 μg/ml), to 20% with vinblastine 3 × 10⁻M (≃ 0 · 03μg/ml), to complete developmental arrest with colchicine 3 × 10^{− 7} M (≃ 0 · 1 μg/ml). A short treatment with colcemid (1 · 5 h) was found to be non toxic for rabbit eggs undergoing in vitro fertilization (Bomsel-Helmreich, 1965); the treatment produced triploid embryos by suppressing the extrusion of the second polar body.

Colchicine is known to be more ‘toxic’ than its derivative colcemid, and the action of colcemid on cultured cells can be reversed more easily (see for instance Kleinfeld & Sisken, 1966; Daniels, 1975). To examine whether the irreversibility of action of colchicine on mouse embryos was due to some non-specific toxic effect, preimplantation embryos were also treated with lumicolchicine. Lumi-colchicine is a structural isomer of colchicine (obtained by u.v. irradiation of colchicine) which does not bind to tubulin or disrupt microtubules (Wilson et al. 1974), but retains the non-specific properties of colchicine, such as the interaction with cell membranes (Stadler & Franke, 1974) and the inhibition of nucleoside transport (Mizel & Wilson, 1972). By comparing the effects of colchicine and lumicolchicine, specific (i.e. due to microtubule depolymerization) and non-specific effects of colchicine can therefore be discriminated (Mizel & Wilson, 1972; Obika et al. 1978).

The present studies show that zygotes can cleave to 2-cell, and 2-cell embryos can develop normally to blastocysts in the continuous presence of 2 × 10⁻⁷ M lumicolchicine, whereas embryos treated similarly with colchicine (2 × 10⁻⁷M) failed to divide or develop to the blastocyst stage (Table 1). Thus, the irreversibility of colchicine is related to its direct action on microtubules and not to any non-specific side effects. This finding agrees with the observation that the binding of tritium-labelled colchicine to purified tubulin at 37 °C is almost irreversible (Wilson et al. 1974); the colchicine-tubulin complex dissociation half-life is 36 h (Garland & Teller, 1973). Our results also show that the minimum period of exposure to 3 × 10^{− 7} M colchicine, which allows the alkaloid to bind to a sufficient number of tubulin molecules to prevent cleavage, is 2 h (Table 6). This agrees with the biochemical findings which show a similar slow rate of colchicine binding to purified tubulin (Wilson et al. 1975).

The fourth drug examined for its action on microtubular function was griseofulvin (a mould metabolite). Unlike the other microtubular-disruptive drugs tested, its mechanism of action is not completely understood. Although 1- and 2-cell mouse embryos appeared to be quite resistant to the action of griseofulvin, the second cleavage division was more sensitive than the first for cleavage inhibition and the incidence of unequal cytoplasmic division. Concentrations up to 3 × 10^{− 4}M did not prevent the first cleavage division but cleavage was inhibited in 50 % of 2-cell embryos treated with concentrations of griseofulvin between 1 × 10⁻⁴ and 3 × 10⁻⁴ M. Only a few of the embryos undergoing first cleavage and the majority of those undergoing second cleavage showed unequal cytoplasmic division when treated with 1 × 10⁻⁴ M or higher concentrations of griseosulvin. The unequal cytoplasmic divisions were probably caused by the formation of multipolar mitoses similar to those described in other cells treated with griseofulvin (Grisham, Wilson & Bensch, 1973; Adair, 1974; Weber, Wehland & Herzog, 1976). Although the 2-cell embryos showed an apparent increase in sensitivity to griseofulvin, they were unaffected by the continuous presence of 3 × 10⁻⁵ M griseofulvin during culture to the blastocyst stage (Table 2).

Other types of cells appear to be much more sensitive to the action of griseo-fulvin. Concentrations of 2 –4 × 10^{− 5} M griseofulvin cause 50 % mitotic arrest in mouse 3T3 cells (Weber et al. 1976) and in HeLa cells (Grisham et al. 1973) and complete metaphase block in Chinese hamster V79 cells (Adair, 1974). At 1 × 10^{− 5} M griseofulvin there is a rapid but reversible disappearance of the meiotic spindle in Pectinaria oocytes (Malawista, Sato & Bensch, 1968). More recently the depolymerization of microtubules by griseofulvin has been shown by immunofluorescence microscopy in 3T3 cells at 5 × 10^{− 5} M (Weber et al. 1976) and human fibroblasts at 5 × 10^{− 4} M (Spiegelman, Lopata & Kirschner, 1979). The relative resistance of early mouse embryos to griseofulvin may be due to its low permeability. This will now be investigated with the use of labelled griseo-fulvin.

Apart from the possible permeability barrier to the drug, a further explanation for the low sensitivity of mouse embryos to griseofulvin can be postulated, based upon recent observations on the regeneration of the microtubular system after the removal of the drug. According to Spiegelman et al. (1979) there appear to be two types of microtubule-organizing centres : a single primary initiation site, which is probably the centriole, and multiple secondary sites. Since the secondary sites recover from griseofulvin treatment more slowly than the primary sites Spiegelman et al. (1979) suggest that griseofulvin has a differential effect on the two types of initiation centres. Thus it is tempting to speculate that a connexion exists between the relative resistance of mouse embryos to griseofulvin and the lack of centrioles in mouse embryos up to the morula stage (Szollosi, 1972). Such speculation is further supported by the observation that plant cells, which also lack centrioles, are much less sensitive of griseofulvin than animal cells (Deysson, 1964).

This work was performed under C.N.R. Research Project ‘Biology of Reproduction’ (Grant, no. 79.01175.85), C.N.R. Grant CT 79.01010.04, and was also supported in part by the Ford Foundation (Grant no. 790.0659).