Mevalonate reverses the developmental arrest of preimplantation mouse embryos by Compactin, an inhibitor of HMG Co A reductase

During preimplantation development in the mouse, two distinct tissues, the inner cell mass and the trophectoderm are formed at the blastocyst stage after six cleavage divisions (Gardner, 1971). Prior to this, a major Ca²⁺-dependent morphogenetic event called compaction occurs at the 8-cell stage when close membrane apposition and cell flattening are observed (Ducibella & Anderson, 1975; Lehtonen, 1980; Kimber, Surani & Barton, 1982), together with the regionalization of cell surface constituents (Handyside, 1980). At the 16-cell stage, a group of inner cells is established for the first time (Handyside, 1981; Johnson & Ziomek, 1981) which presumably go to form the inner cell mass whilst the outer cells give rise to trophectoderm (Surani & Handy side, 1983). Cell surface properties determine cell interactions and adhesiveness and influence cell position within the embryo (Kimber et al. 1982; Surani & Handyside, 1983) in which cell surface changes (Hyafil, Morello, Babinet & Jacob, 1980; Johnson & Calarco, 1980; Magnuson & Epstein, 1981; Kapadia, Feizi & Evans, 1981; Kimber & Surani, 1982) probably have a major influence on interactions between cells (see Surani, Kimber & Barton, 1981).

For the synthesis of N-glycosidically linked glycoproteins, the polyisoprenoid lipid, dolichol, is the major saccharide carrier (Hemming, 1977). We have previously shown that tunicamycin, a specific inhibitor of N-acetylglucosaminyl pyrophosphoryl dolichol phosphate (Takatsuki, Arima & Tamura, 1971) substantially inhibits glycosylation, disrupts compaction and alters the cell surface properties (Surani, 1979; Surani, Kimber & Handyside, 1981; Atienza-Samols, Pine & Sherman, 1981). The levels of dolichol compared to cholesterol are substantially higher in many developing systems examined (Potter, Millet, James & Kandutsch, 1981; Harford & Waechter, 1981; Carson & Lennarz, 1981). Hydroxymethyl-glutaryl Coenzyme A reductase (HMG Co A reductase) is the key regulatory enzyme in the formation of mevalonic acid which is a precursor of divergent biosynthetic pathways of both sterol and non-sterol isoprenes (Brown et al. 1978; James & Kandutsch, 1979). These products include cholesterol, dolichol, isopentenyl adenine and ubiquinone (see Fig. 6). In this study we have examined the consequences of inhibition of HMG Co A reductase by Compactin (or ML236B) (Brown et al. 1978) and an oxygenated sterol, Diosgenin (Mills & Adamany, 1978). This study demonstrates that Compactin disrupts early embryonic development and inhibits protein glycosylation and these effects can be completely reversed by exogenous mevalonic acid.

Embryos were obtained from 3 to 5-weeks old outbred strain of MFI mice (ARC colony established from OLAC stock) which were superovulated using 5i.u. pregnant mare’s serum followed 42–48h later by 5i.u. human chorionic gonadotrophin (HCG) (Intervet, Milton, UK). Each female was caged with an Fi (C57BL/CBA) male and checked the following morning for the vaginal plug; this was counted as day 1 of pregnancy.

Compactin (ML-236B) was a gift from Dr Akira Endo (Sankyo, Tokyo, Japan) and Dr R. Fears (Beecham Pharmaceuticals, Epsom, UK). The lactone form of Compactin was converted to the acid form by heating at 50 °C for 1 h in 0·1 M-NaOH (Kaneko, Hazama-Shimada & Endo, 1978). The solution was then adjusted to pH 8·0 with 1N-HC1 and the concentration of the compound to 1 mg/ml in 0·01 M-tris-HCl, pH 8·2. The sodium salt of Compactin thus obtained was divided into 10μ aliquots and stored at –20 °C. DL-4,4,10-β-trimethyl-trans-decal-3β-ol (TMD) was a gift from Drs J. A. Nelson and T. A. Spencer (Dartmouth College, NH, USA). DL-mevalonic acid lactone, cholesterol, dolichol, dolichol monophosphate, retinoic acid, coenzyme Q₁₀, isopentenyl adenine and Diosgenin (5,20 a,22 a,25o-spirosten-3β-ol) were all obtained from Sigma. Mevalonic acid was dissolved in the embryo culture medium at a concentration of lmgμd medium and stored at —20 °C. Dolichol, dolichol monophosphate and coenzyme Q₁₀ were dissolved in chloroform at a concentration of 10 mg/ml, 2 mg/ml and lmg/100μl, respectively. Isopentenyl adenine was dissolved at 0·5 mg/ml ethanol. Retinoic acid was first dissolved in ethanol (3 μg/1 ethanol) and then made up to 500/11 in the embryo culture medium. All these compounds were usually stored for 1 week and a maximum of 2 weeks at —20 °C under nitrogen. The rest of the compounds were freshly made on the day of the experiment. Diosgenin was made in ethanol (2 mg/ml). Cholesterol was first dissolved in chloroform (100 μg/μl) and 5 μl of this stock solution were then added to 1μ0 ml of culture medium with 4 mg/ml bovine serum albumin with vigorous vortexing. Lectins and their antibodies were obtained from Vector (California, USA), and protein A-Sepharose from Pharmacia (Uppsala, Sweden). All radiochemicals were from Amersham (UK).

Two-cell embryos were flushed from oviducts between 2 and 5 p.m. on day 2 of pregnancy at about 45–48 h post HCG. Embryos were cultured in Brinster’s medium (Brinster, 1970) supplemented with 4 mg/ml bovine serum albumin (BMOC-3).

Two-cell embryos were used immediately in experiments or cultured overnight in microdrops of BMOC-3+BSA in Sterilin Petri dishes under paraffin oil. The following morning, precompacted 8-cell embryos (approx. 65 h post HCG) were removed and washed 6x through fresh medium. The embryos were then divided into groups of 20–40 embryos. Subsequent cultures were carried out in microtitre plate wells containing 0·1–0·3ml of each test medium and with adhesive plate sealers. The drops were allowed to reach equilibrium at 37 °C in 5 % CO₂ in air for 30 min before groups of embryos were transferred to each of the test mediums.

Embryos were examined periodically and photographed using a Leitz Diavert inverted microscope under phase contrast and bright field.

Embryos were fixed in 2·5 % glutaraldehyde, 1 % paraformaldehyde in 0·075 M-sodium cacodylate buffer (pH 7·5) containing 2 mM-calcium, and 0·1 % potassium ferricyanide. Embryos were post fixed in 1 % osmium tetroxide, stained overnight in uranyl acetate, dehydrated and embedded in Epon. Thick (0–5 μm) and thin (50–100 nm) sections were stained with 1 % toluidine blue and viewed and photographed using a Zeiss microscope. Thin sections were stained with a saturated solution of uranyl acetate in 50 % ethanol followed by lead citrate (Reynolds, 1963) and examined in an AEI801 electron microscope.

The number of cells was determined by counting nuclei in air-dried preparations of embryos (Tarkowski, 1966).

For the estimation of incorporation of [³H]leucine and [³H]sugar precursors into embryos, they were cultured in the presence of Compactin for approximately 24 h commencing at the 8-cell stage. [³H]leucine (sp. act. 105Ci/mmol) and [³⁵S]methionine (sp. act. 900–1200 Ci/mmol) were added at 200μCi/ml in glucose-free medium as described previously (Surani, 1979). For labelling of embryos with sugar precursors, 50μl of a 2× concentrated glucose-free medium was used to which the following radioactive precursors were added; 25μCi [³H]glucosamine (sp. act. 38Ci/mmol), 25μCi [³H]mannose (sp. act. 2Ci/ mmol) and 25 pC\ [³H]galactose (sp. act. 18Ci/mmol). Embryos were labelled for approximately 6 h in groups of between 10–30 embryos with the amino acid precursors or in groups of 50–100 with the sugar precursors. At the end of the labelling period, embryos were washed and processed to determine incorporation of the precursors in embryos in the control group and compared with those cultured in the presence of Compactin essentially as described previously (Surani, 1979).

Embryos labelled with [³⁵S]methionine were also analysed on 8–15 % SDS-polyacrylamide gradient gels and the radiolabelled proteins visualized by fluorography exactly as described previously (Moor, Osborn, Cran & Walters, 1981). In some cases, galactosyl glycopeptides were immunoprecipitated after the addition of peanut lectin followed by an addition of antibody against peanut lectin (PL Biochemicals) and the immune complexes were extracted using Protein A-agarose gels. The method used was essentially as described elsewhere (Magnuson & Epstein, 1981). The extracted glycopeptides were also analysed by polyacrylamide gel electrophoresis as described above.

Control embryos and those cultured in Compactin were examined by fluorescence microscopy after treatment with FITC-Concanavalin A as described (Kimber et al. 1982) using a Zeiss epifluorescence microscope.

In preliminary experiments Compactin was used in the range of 0·5–8·0μg/ ml. Two-cell embryos developed to the 8-cell stage in all cases and underwent compaction and formed advanced morulae after 48 h in culture. When embryos were cultured in 0·5–1·0μg/ml Compactin, a small proportion (10-20%) formed into poorly developed blastocysts. Embryos cultured in 2·0μg/ml Compactin failed to develop into blastocysts and this was considered as the minimum concentration of Compactin necessary to completely disrupt development. Similar preliminary experiments with Diosgenin (1·0–5·0μg/ml) established that 5–0 μg/ml Diosgenin was necessary to disrupt development to the blastocyst stage.

Preliminary experiments were also carried out using Compactin and Diosgenin in medium containing normal or fatty-acid-free albumin (from Sigma). There was no marked difference between the effect of the compounds in medium with fatty-acid-free albumin and in medium containing fatty acids. Fatty-acid-free albumin also did not hinder development to the blastocyst stage in the absence of the inhibitors. In all cases except where indicated normal bovine serum albumin was present in the culture medium.

We first examined the influence of Compactin and Diosgenin on cell proliferation on 8-cell precompacted embryos. After 24 h in culture (89 h post HCG), the number of cells in the control group and in the presence of Compactin was similar but those cultured in Diosgenin had only about half the number of cells and these had not increased considerably after 31 h (96 h post HCG) in culture. After further 7h in culture, embryos in Compactin showed a slight increase in the number of cells to 32 cells compared with about 42 cells in the embryos in the control group (Table 1). Both of the inhibitors have been previously shown to specifically block HMG Co A reductase. Therefore the influence of the two major products of this pathway, mevalonic acid and cholesterol, on embryonic development in the presence of either Diosgenin or Compactin was examined.

The influence of mevalonic acid and cholesterol in the presence of Diosgenin is shown in Table 2. Mevalonic acid upto lmg/ml was unable to reverse the influence of Diosgenin. However, cholesterol at 50–100μg/ml was able to rescue approximately 50 % of the embryos (see Fig. 1). Many of these embryos which were apparently rescued formed blastocysts that appeared to be poorly developed. Similar experiments were carried out with Compactin (Table 3). As little as 10 μg/ml (0·08 mM) mevalonic acid in the presence of Compactin enabled nearly 90% of the embryos to develop to the blastocyst stage. Even 1μg/ml mevalonic acid was partially effective and rescued over 30 % of the embryos. By contrast, cholesterol at up to 100μg/ml was virtually ineffective in reversing the influence of Compactin. The embryos that were rescued by mevalonic acid appeared to be normal when examined under an inverted microscope (see Fig. 1). Furthermore, the number of cells in the embryos rescued by cholesterol in the presence of Diosgenin was substantially less than the number of cells in the embryos in the control group. However, the embryos rescued by mevalonic acid in the presence of Compactin had the normal number of cells.

The role of cholesterol in early embryonic development was examined further with a specific inhibitor of its synthesis. TMD inhibits cyclization of squalene and hence prevents synthesis of lanosterol and cholesterol (Chang et al. 1979). This inhibitor was also of interest since its site of action is distal to the formation of dolichol and therefore TMD would not affect the concentration of dolichol. TMD at a concentration of up to 25 μg/ml was virtually ineffective in blocking development to the blastocyst stage. At 50μg/ml TMD seemed highly toxic. Similar results were obtained with embryos cultured from 2-cell or 8-cell stage onwards in the presence of TMD. Culture of embryos in the presence of TMD and fatty-acid-free albumin also did not influence development. The results were therefore combined (Table 4).

The influence of Compactin on embryos was examined further. When 2- or 8-cell embryos were cultured in the presence of Compactin, they developed normally at first, undergoing normal cleavage divisions and compaction at the 8-cell stage and proceeded to the 32-cell stage. However the embryos started to decompact at the 16-cell stage and the majority of the embryos were decompacted after 30–35h in culture (Fig. 1). As shown in Fig. 1A, the embryos in the control group were fully compacted or at the blastocyst stage. Embryos in the presence of Compactin (Fig. IB) had rounded and distinct blastomeres.

Mevalonic acid prevented this effect of Compactin on embryos (Fig. 1C). Some of these embryos were sectioned and the semithin sections examined by light microscopy. Figures 2A and 2B show that the embryos in the control group were at the late morulae-early blastocyst stage. The outer blastomeres were characteristically flattened. Embryos cultured in Compactin (Fig. 2C and D) however, failed to show the flattened morphology of outer cells: the blastomeres generally tended to be rounder. Examination of the embryos by electron microscopy (Fig. 3) revealed much-reduced areas of membrane apposition when embryos were cultured in the presence of Compactin. However, the unapposed peripheral surfaces of the blastomeres had a high density of microvilli similar to the density on the outer surface of control morulae. Large intercellular spaces were present inside the morulae between the rounded blastomeres in contrast to the closely packed arrangement of cells in control morulae. Although microvilli were present on their inner membranes, they did not interdigitate closely with those of adjacent blastomeres over the inner membrane area as in the control compacted morulae. However, in some instances adherens junctions were observed at the apical regions of membrane contact between cells. Cytoplasmic blebs and large processes sometimes extended from the surface of Compactin-treated embryos (Fig. 3D) and there were regions of microvillar interdigitation. The mitochondrial population of both control late morulae and Compactin-treated embryos consisted partly of the vacuolate type found in early preimplantation embryos and partly of the non-vacuole form with clearly defined cristae found in the blastocyst? No differences could be found in the rough endoplasmic reticulum or in the distribution of microfilaments and microtubules between control and Compactin-treated embryos. One noticeable difference was found in the nucleus. In the control embryos the majority of the nuclei could be detected with distinct nuclear membranes, whereas in the experimental group the nuclear membranes were barely detectable and the chromatin was highly condensed.

The influence of Compactin on the synthesis of macromolecules was also examined. Eight-cell embryos were first cultured in the presence of Compactin for 24 h and then labelled with amino acids and sugar precursors for 6 h after this time. Figure 4 shows that the incorporation of amino acid precursors was not affected by Compactin, rather there was an increase in the incorporation of both [³⁵S]methionine (117 %) and [³H]leucine (135 %) relative to the control values. On the contrary, the incorporation of all of the sugar precursors tested was substantially reduced to between 34—48 % relative to the control values.

Both the polypeptides and glycopeptides were also analysed on 8–15 % polyacrylamide gels. No qualitative differences were detectable in polypeptides or glycopeptides of embryos cultured in the presence of Compactin (Fig. 5).

The binding of fluorescein-conjugated Concanavalin A to the embryonic cell surface was examined. Although occasionally there appeared to be a slight reduction in the binding of Concanavalin A to embryos cultured in Compactin, there was no marked quantitative difference compared with the controls (data not shown). Further detailed studies are needed to establish if Compactin has an effect on the cell surface constituents of embryos.

Finally, since the influence of Compactin appears to be due to specific inhibition of HMG Co A reductase, several key products of this pathway were used to determine whether the effect of Compactin could be overcome by these compounds (Table 5). As mentioned previously, mevalonic acid was highly effective in reversing the influence of Compactin. Cholesterol however was ineffective. Other compounds tested were dolichol, dolichol monophosphate, retinoic acid, coenzyme Qio and isopentenyl adenine. All of these were ineffective in reversing the influence of Compactin. The compounds were also tested in various concentrations and in a large variety of combinations. So far none of the combinations employed have been successful in reversing the influence of Compactin (data not shown).

Compactin, a competitive inhibitor of HMG Co A reductase (Brown et al. 1978) interrupts preimplantation development of mouse embryos after the 16-cell stage. Mevalonic acid (1–10μg/ml), the product of this enzyme activity (see Fig. 6) can totally reverse the influence of Compactin. In contrast, Diosgenin acts relatively quickly and prevents compaction at the 8-cell stage but the effect is not reversible by mevalonic acid, although Diosgenin is also suggested to inhibit HMG Co A reductase (Mills & Adamany, 1978). However, Diosgenin and other oxygenated sterols can sometimes act non-specifically when they become inserted in the plasma membrane (Gordon, Bass & Yachnin, 1980). Such nonspecific effect is partly reversible by exogenous cholesterol which presumably displaces the compound from the plasma membrane. Similar results were previously obtained with other oxygenated sterols (Pratt, Keith & Chakraborty, 1980), but such influence is not necessarily a reflection of de novo synthesis of sterols during early development in the mouse. Indeed there is little synthesis of sterols up to the blastocyst stage (Pratt, 1982; Carson, Hsu & Lennarz, 1982). Furthermore, cholesterol, the major product of mevalonic acid did not reverse the effects of Compactin on embryos; these embryos had approximately the same number of cells (30) as in the blastocysts formed when cholesterol was present along with Diosgenin. In addition, TMD, a specific inhibitor of squalence cyclization (Chang et al. 1979) had no effect on development to the blastocyst stage. The combined observations indicate that products of mevalonate other than cholesterol, such as the non-sterol isoprenes dolichol and isopentenyl adenine, may be equally important if not crucial for the rescue of embryos by mevalonic acid in the presence of Compactin after the 16-cell stage.

Only 0·03–0·11 % of acetate is generally incorporated into dolichol compared with sterols, but their levels can increase by 10- to 100-fold in developing systems where cell differentiation is underway (James & Kandutsch, 1980; Potter et al. 1981; Harford & Waechter, 1980; Carson & Lennarz, 1981). This increase in dolichol, the key lipid saccharide carrier in the glycosylation of N-glycosidically-linked glycoproteins (Hemming, 1977), and mannosylphosphoryl dolichol (Harford & Waechter, 1980; Carson & Lennarz, 1981) is probably essential for rapid changes in the cell surface glycoproteins during cell interactions and morphogenesis. In this respect it was interesting to observe the effect of Compactin on the incorporation of sugar precursors into embryos which was markedly reduced compared with the incorporation of amino acids. Since overall protein synthesis, both quantitative and qualitative remains unaffected, this suggests an inhibition of protein glycosylation probably resulting from the lack of dolichol. We cannot entirely rule out the effect of Compactin on polypeptides since qualitative differences may be detected if they are analysed by the twodimensional gel-electrophoretic system. The decrease in the incorporation of the sugars was not caused by the influence of Compactin on intracellular membranes since the endoplasmic reticulum and other organelles such as mitochondria developed normally. The influence of Compactin first becomes detectable at the 16-cell stage when some of the interstitial glycoproteins such as laminin are detected between cells (Leivo, Vaheri, Timpl & Wartiovaara, 1980). However, further work is necessary to demonstrate that the decompaction of embryos caused by Compactin is as a result of changes in the cell surface glycoproteins.

There are similarities between the effects of Compactin and tunicamycin on mouse embryos. Tunicamycin, a specific inhibitor of synthesis of dolichol-linked saccharides (Takatsuki et al. 1981) also caused decompaction of embryos and inhibited protein glycosylation (Surani, 1979; Ateinza-Samols et al. 1981) and caused changes in the cell surface properties of blastomeres (Surani et al. 1981). In sea urchin embryos, dolichol alone can overcome the inhibitory effects of Compactin on protein glycosylation as well as development (Carson & Lennarz, 1979,1981). Although dolichol alone (or in combination with other products of mevalonic acid) failed to reverse the influence of Compactin on mouse embryos, we have not ruled out the possibility that in the mouse embryos many of the compounds only enter the plasma membrane and not the cytoplasm.

Isopentenyl adenine, another product of mevalonic acid may be crucial during development for DNA replication and cell growth. It appears that in embryos cultured in Compactin, the cells ceased division at a specific point in the cell cycle, since the majority of them had highly condensed chromatin and lacked clear nuclear membrane. This finding may be significant since isopentenyl adenine is implicated in DNA-polymerase-dependent DNA replication (Habenicht, Glomset & Ross, 1980; Quesney-Huneeus, Wiley & Siperstein, 1979, 1980). This aspect requires further investigation.

Recent studies indicate that mevalonate itself has an influence on cell shape since in its absence, cells tend to round up (Schmidt et al. 1982; Cohen, Massoglia & Gospodarowicz, 1982). This role of mevalonate is independent of its effect via dolichol-mediated protein glycosylation and suggests that mevalonate itself may be required during morphogenesis and embryonic development when alterations in cell shape occur.

We thank Mrs S. C. Barton for expert assistance and the research workers cited in the paper for their generous gifts of various compounds.