Effect of Li⁺ on preimplantation mouse embryos

The vegetalization induced by Li⁺ on early embryos of several species has been ascribed to its interference with a dorsoventral morphogenetic pattern, although the underlying molecular mechanism still remains obscure (see Stanisstreet & Osborn, 1976). No pre-existing morphogenetic pattern has been convincingly described nor experimentally demonstrated in early mammalian embryos (see Izquierdo, 1977) and for the differentiation of morula to blastocyst the current explanation is the so called inside-outside model (Tarkowski & Wroblewska, 1967). It states that peripheral cells of an advanced morula become trophectoderm and interior cells become inner cell mass as a result of their different microenvironment. However, the accumulated evidence on behalf of this model does not exclude the possibility of an inside-outside gradient being set up during cleavage. Even assuming that all blastomeres of a late morula are similar, if each one produces an equivalent amount of a substance that diffuses to the space around the embryo, one would expect the establishment of a concentration gradient (Crick, 1970). Would Li⁺ interfere with such a gradient ? If this is so, may the effect be compared with the vegetalization induced by Li⁺ in early echinoderm and amphibian embryos? Although these were the main reasons for this inquiry, we were also interested in knowing whether Li⁺ might affect fertility or development in humans, in view of its frequent utilization for the treatment of manic-depressive illness (see Baldes-sarini & Lipinski, 1975).

Two-cell embryos were collected from spontaneously ovulating mice of the Swiss Rockefeller strain after 40 h of development; counting began at 0 h of the day in which the vaginal plug was detected. The embryos were cultured in Falcon plastic Petri dishes in microdrops of Bigger’s medium (Biggers, Whitten & Whittingham, 1971) supplemented with 4mg/ml Bovine Serum Albumin (Calbiochem) under mineral oil in an atmosphere of 5 % (v/v) CO₂ in humid air at 37 °C. NaCl in the medium was partially replaced by LiCl in concentrations ranging from 1 to 30 mM which means, at most, reducing NaCl by less than one third. After different periods in culture, all embryos which had not developed beyond a 4-cell stage were eliminated. The rest were described under a dissecting microscope and the cell nuclei were counted by Tarkowski’s method (1966). Some embryos were fixed in 3 % glutaraldehyde in 0 ·1 M cacodylate buffer pH 7 ·4 for 1 h, post fixed with OsO₄ 1 %, dehydrated in graded acetone and embedded in Spurr’s resin (Polyscience). Semithin sections (0 ·5 μm) were stained with Toluidine blue in borax and observed with a light microscope.

In each experiment the embryos obtained from two or more females were pooled and cultured in two microdrops, one with and the other without LiCl. After 48 h in culture, the cells in each embryo were counted and the relative cell number expressed as the ratio of the mean number of cells in treated embryos over the mean number of cells in control embryos. This procedure takes care of wide variations that might occur in the timing of ovulation and fertilization. Figure 1 shows that the relative cell number diminishes as a function of Li⁺ concentration. Each point represents one experiment with an average of 10 embryos cultured with Li⁺ and 6 control embryos. The total number of embryos examined in this experimental series is 566.

The cumulative effect of Li⁺ during a 48 h culture provides no information on whether early or late cleavage might be preferentially affected; therefore, in another series of experiments, 170 2-cell embryos were cultured for two successive 24 h periods separated by a brief rinse. When the embryos were cultured with 10 mM-LiCl during the first period and without the salt during the second period, the relative cell number was 0 ·91 ±0 ·05 while inverting the sequence of cultures the ratio was reduced to 0 ·78 ± 0 ·06. These results show that late cleavage is more affected than early cleavage and additionally, that rinsing removes Li⁺ from its site of action. Shorter culture periods were not tried because the different phases of the cell cycle, which in this material normally takes 11 h (Luthardt & Donahue 1975), may introduce some confusion in the interpretation of the results.

The onset of blastulation in embryos which were cultured in different LiCl concentrations is shown in Fig. 2. The abscissa represents hours of development in situ plus hours of development in vitro. The points represent the percentage of blastulae in microdrops containing 9 or 10 embryos. Each microdrop was observed two or three times during the 30 h long experiments. The total number of embryos observed in control cultures was 214 and in cultures with 5, 10, 15 and 20 mM-LiCl, the number of embryos was 121, 87, 88 and 75, respectively. Figure 2 shows that control embryos, as well as embryos cultured in different Li⁺ concentrations, begin to blastulate at 87 –88 h of development and that all those which blastulate do so within a 20 h interval. After 108 h of development, 100 % of the embryos have blastulated in cultures with up to 10 mM-LiCl whereas only 85 % blastulated in 15 mM and 75 % in 20 mM-LiCl, even though these cultures weie observed for 24 additional hours.

Values: for controls: x(l) = 0 · 175, x(2) = 86 ·3, x(3) = 100; for LiCl 5mM: x(l) = 0 ·137, x(2) = 85 ·8, x(3) = 100; for LiCl 10MM: X(1) = 0133, x(2) = 86 · 0, x(3) = 100; for LiCl 15 MM: X(1) = 0 ·110, x(2) = 87 ·3, x(3) = 85; for LiCl 20 mM: x(l) = 0 · 097, x(2) = 86 ·8, x(3) = 75.

Our observations on the timing of compaction are somewhat incomplete but they suggest that compaction, similarly to blastulation, is not retarded by Li⁺.

Since the precise recognition of nascent blastulae lends itself to subjective appraisals we took as the number of cells at blastulation the mean number of cells per embryo in populations which included only late morulae and early blastulae. That is, excluded from these populations were all morulae with less cells than the evident blastula with less cells, and all blastulae with more cells than the evident morula with more cells. For example: 122 embryos cultured during 48 h without Li⁺ attained different stages, from a 15-cell morula up to a 30-cell blastocyst; however, since the latest evident morula had 25 cells and the earliest evident blastocyst had 21 cells, we calculated the mean number of cells per embryo in a population including only the 31 embryos which had more than 21 and less than 25 cells. The mean number was 23 ·32 ± 28 (Fig. 3). Each point in Fig. 3 represents the mean number of cells per embryo in populations comprising an average of 17 ·5 embryos. The total number of embryos considered in these series was 141. Figure 3 shows that cell number at blastulation decreases as LiCl concentration increases.

The experimental series were designed so as to test the effect of different LiCl concentrations on relative cell number, on timing of blastulation, and on number of cells at blastulation.

The blastulae obtained were either normal blastocysts or blastocysts with a reduced or even absent inner cell mass. The latter, that is trophoblastic vesicles, account for 6, 21 and 43 % of the embryos which blastulated in cultures containing 10, 15 and 20 mM-LiCl respectively. Figure 4 shows sections through the meridian plane of a blastocyst developed in control medium, a blastocyst with reduced inner cell mass developed in a medium with 15 mM-LiCl and a trophoblastic vesicle developed in a medium with 20 mM-LiCl.

It has been shown in mammalian nervous tissue that Li⁺, at concentration similar to those used in this research, inhibits the synthesis of cyclic nucleotides probably by interfering with the role of divalent cations on the activity of adenyl cyclase and guanyl cyclase (Forn & Valdecasas, 1971; Bunney et al. 1979; Zatz, 1979). Since the levels of cyclic nucleotides have been related to cell proliferation in several differentiated tissues (Berridge, 1975) one might assume that this is the mechanism by which Li⁺ affects mammalian cleavage. However, in sea-urchin embryos, no variations in adenyl cyclase activity have been detected during cleavage nor are variations in cleavage induced by exogenous cAMP, though its intracellular level increases greatly (Amy & Rebhun, 1977).

As to the onset of blastulation, since it was almost simultaneous in all experimental series, our results support the idea that its timing in the mouse depends neither on the total number of cells in the embryo nor on the number of cell cycles elapsed since fertilization (Smith & McLaren, 1977; Fernández & Izquierdo, 1980). This conclusion does not imply that mammalian blastulation depends on a process which is triggered by the activation of the oocyte and then proceeds unchecked by embryonic metabolism; actually, blastulation in the mouse has been supressed by α-amanitin (Braude, 1979), reversibly retarded by an inhibitor of polyamine biosynthesis (Alexandre, 1979) and by rabbit antiserum to a mouse embryonal carcinoma cell line (Johnson et al. 1979).

The increasing percentage of trophoblastic vesicles which developed in cultures with increasing concentration of LiCl finds a suitable interpretation in the inside-outside model. Since cleavage is retarded by Li⁺ while blastulation is not, when the latter occurs all cells may lie outside and therefore, according to the model, would differentiate into trophectoderm. However convincing this explanation may be in the case of mammals it may not account for vegetalization induced by Li⁺ in early sea-urchin and amphibian embryos.

Cleavage of amphibian ectoderm is retarded by Li⁺ (Flickinger, Lauth & Stambrook, 1970; Osborn & Stanisstreet, 1977) although at concentrations which are about ten times higher than those used in our research. However, if Li⁺ were to retard cleavage without delaying blastulation in echinoderms and amphibia as it does in mammals, one might conceive a general interpretation of vegetalization assuming that cells of each presumptive germ layer must divide a certain number of times before blastulation (and/or gastrulation) in order to differentiate normally. This conjecture, in our opinion, deserves a test. Information available at present does not refute nor support the idea that trophoblastic vesicle formation is a vegetalization effect induced by Li⁺. Therefore, it would be p emature to suggest that during cleavage of the mouse embryo emerges a Li⁺-sensitive gradient.

Now, as to the effect that Li⁺ may have on humans, since the effective concentrations of this ion in our experiments are at least 2 or 3 times higher than therapeutical doses (Baldessarini & Lipinski, 1975), our results do not support reservations about unwanted effects on fertility or preimplantation development.

This research was partially financed by Grants from PLAMIRH, PNUD/UNESCO Regional Program and the Ford Foundation.