Coincidence of embryonic growth and uterine protein in the ferret

Preliminary studies (Daniel & Krishnan, 1969) have shown that the fluids taken from the lumen of uteri of those animals having obligate delayed implantation (e.g. black bear, fur seal, mink and armadillo) are low in protein content and specifically lack any component comparable to the rabbit uterine protein, blastokinin (Krishnan & Daniel, 1967), during the period when the blastocysts are in diapause. In mink uterine fluids, taken near the end of the diapause period (immediately preceding embryo ‘reactivation’ and subsequent implantation), the amount of protein rises markedly and traces of a component similar to blastokinin can sometimes be detected (Daniel, 1968; Daniel & Krishnan, 1969). Thus, the reinitiation of blastocyst growth in mink parallels quantitative (and possibly qualitative) changes in the uterine lumen proteins. It is of interest to know whether uterine protein increases with embryonic growth in the ferret (Mustela furo), a mustelid that is closely related to the mink (Mustela vision) but does not have delayed implantation.

This paper reports the results of studies comparing the protein content of ferret uterine fluid with stages of embryonic growth throughout the preimplantation and early post-implantation periods.

Ferrets were made available to the author during the spring of 1969 from the colony of John Hammond, Jr. of Cambridge University. Animals were selected for breeding by the condition of the vulva. Duration of the pregnancy and age of the embryos were measured from copulation as time zero. Animals were killed by cervical dislocation on days 0, 1,3, 4, 5, 7, 8, 9, 10, 13, and 16 post coitum (± 1 h), and embryos and uterine fluids collected at each of these times. Uteri were removed from the animals. One horn was cut free at the cervical end, the cut end blotted dry of blood on filter paper, and the horn then flushed with 1 ml of Tyrode’s solution and the flushings collected into watch glasses. After the contained embryos were removed to fresh saline, the flushings were centrifuged to remove cellular debris and then frozen for future analysis of their protein content (presumably, representative of the soluble proteins present in the uterine lumen). The second horn was torn open with forceps, any obvious embryos removed and added to those from the first horn, and then filter-paper discs applied to the exposed endometrial surface to absorb the undiluted uterine fluids. These discs were frozen and later used as the sample source for acrylamide disc-gel electrophoresis. Exceptions to these methods were necessary for postimplantation stages where it was impossible to flush the embryos from the horn or to get complete total samples of the fluids. In these cases, those portions of the uteri that were free of a conceptus were cut away and serially flushed with the same volume of saline. These additional cut surfaces increased the probability of contamination of the sample with blood proteins but two animals, 13 and 16 days pseudopregnant, provided supplementary uncontaminated samples.

The isolated embryos were measured with an ocular micrometer and cell number approximations made of pre-morula stages (less than 5 days p.c). Acetic-orcein squash preparations were made of morula and blastocyst stages to facilitate cell counts and determination of mitotic indices.

Total protein was determined spectrophotometrically by the Lowry procedure (Lowry, Rosebrough, Farr & Randall, 1951) compared to known sample concentrations of bovine serum albumin. Two 0-5 ml aliquots of each sample were measured and estimated as micrograms of protein per uterine horn. Single determinations of the proteins were made of four animals that had only degenerate eggs in their uteri after being respectively 7, 8, 13, and 16 days pseudopregnant.

Uterine fluid samples collected on filter-paper discs were subjected to acrylamide gel electrophoresis in Tris-glycine buffer at pH 8·7. Runs were started at 60 V and increased gradually to 240 V with 5 mA/tube until the indicator dye was about 1 cm from the bottom of the tube. Gels were fixed in 7 % acetic acid and were stained in amido black for 3 h. They were then cleared in a series of washes of 7 % acetic acid.

For comparison, similar samples were taken from two mink during the delay period, namely 15 and 16 days post coitum, respectively, and measurements were made in the same way as described above.

Table 1 lists the essential data from these studies. These include the uterine protein determinations, embryo size, cell number and mitotic indices calculated from the pooled cell counts of each stage.

The relationship between embryo growth and uterine protein is shown graphically in Fig. 1.

The results of the acrylamide gel electrophoresis studies provided little useful information. Except for albumin, which was obvious in all of the samples, there was poor resolution of distinct banding in the patterns from uterine fluids, an observation which suggests that most of the proteins may be of low molecular weight and thus readily diffusing species. Samples taken on the ninth and tenth days show some serum-like components but at no time was it possible to demonstrate the band that had the mobility of rabbit blastokinin.

Fig. 2 plots cell replication of pre-implantation ferret embryos as a function of time; it seems to show that no period of growth arrest or embryonic dormancy exists in the ferret and provides the basis for explaining the relatively slow rate of embryonic growth. It includes measurements extracted from the data of other investigators.

According to Hamilton (1934) ‘incipient’ blastocysts are first found in the ferret on the sixth day post coitum, and Hammond & Walton (1934) in their analysis of Robinson’s (1918) data report blastula stages at 144 h. Marston & Kelly (1969) found both morulae and blastocysts at 168 h p.c., and in Chang’s (1968) observations blastocysts were not found until day 7, only morulae on day 6. Apparently, initial cavitation occurs sometime during the seventh day (i.e. 144–168 h p.c.). Implantation is on or about day 12 of the 40–43 day gestation period.

Fig. 1 demonstrates the close parallelism between blastocyst expansion and the amount of protein present in the fluid of the uterine lumen (the parallelism continues after implantation in relation to total embryo growth). During the cleavage period when the ovum is relatively self-sufficient and little tissue differentiation is occurring, the protein level is negligible, but with blastulation it increases and then rises dramatically with implantation. The variability in the measurements makes is impossible to say with certainty that there is not a new constant higher level of protein maintained for days 7 through 9 (as opposed to a continuously rising level), but there is no doubt that the level is two to three times higher than during the preblastocyst period. That the protein comes from the genital tract rather than the embryo is shown by the higher concentrations found in the pseudopregnant animals, where uteri do not contain viable embryos. In the post-implantation pseudopregnancies the protein level is significantly lower than in pregnancies of comparable stage. But this reflects a different reproductive state, uterine size and vascularity, placental development, etc.

Buchanan (1966) has described the general growth of the uterine luminal epithelium with steady increase in secretory products throughout the preimplantation period of the ferret, and the ‘copious secretions’ that accompany hypertrophy of the epithelium during the implantation and post-implantation periods. He concludes from histochemical evidence that the principal uterine secretions are mucoproteins and glycoproteins. That these events are directly related to the changes in the protein content of the uterine lumen seems very likely.

The relatively long pre-implantation period of the ferret and slow growth of its embryo lead one to suspect that an embryonic diapause of short duration could exist in this mustelid. However, the information presented here discounts that suspicion. In addition to the increasing protein content in the uterine fluids, it is obvious that ferret blastocysts continue to grow and maintain a high level of mitotic activity. Thus, there is no evidence of any obligate delay in implantation. (Recently, Buchanan (1969) has shown that a facultative delay might be induced in the ferret because blastocyst expansion can be prevented by ovariectomy on day 4 and implantation prevented by ovariectomy before day 10.)

This calculation shows that the ferret blastocyst cells have a mitotic duration of 80+ min. (For comparison, in the rabbit blastocyst where the mitotic index is about 4·4 %, the cells double in 8 h because the mitotic duration is only about 30 min.) Thus, prolonged mitosis seems to account for the somewhat slower growth of the pre-implantation ferret embryo.

When mink embryos, taken during the middle (15 and 16 days p.c) of an average diapause period of about 3 weeks duration, are compared to those of the ferret, they are seen to be of a size intermediate between 7-to 8-day p.c. ferret blastocysts, but exhibit essentially no mitotic activity. The protein content of the uterine fluids of these mink is very low and at a level comparable to that found in the ferret uteri between 0 and 5 days p.c. : previous to the time of blastocyst development! Earlier studies from this laboratory (Daniel & Krishnan, 1969), though not reporting absolute measurements of mink uterine protein, did nevertheless note that it was very low (in all animals tested with delayed implantation) and that it increased fourfold or more just before embryo activation and subsequent implantation. Such an increase would provide a level not unlike that reported here during implantation of the ferret. These observations support the hypothesis that the embryonic diapause accompanying delayed implantation results from a uterine condition where inadequate protein is available to the blastocyst.

Chang (1968) has demonstrated that mink blastocysts transplanted to ferret uteri will become ‘activated’ and grow, while ferret blastocysts become ‘dormant’ and eventually die in mink uteri. He has further shown that hybrid embryos, resulting from ferret ova fertilized by mink sperm, do not experience a diapause when they reach the blastocyst stage in the ferret uterus (Chang, 1965). These studies support the hypothesis expressed above, in that, although they do not relate directly to uterine protein, they clearly show that some condition of the uterus, and not some inherent quality of the embryo alone, is responsible for the arrested development of mustelid blastocysts in cases of delayed implantation.

Le contenu protémique de liquides extraits de la cavité utérine du furet a été mis en corrélation avec la croissance de l’embryon en pré-implantation. Des observations similaires ont été faites, à titre de comparaison, chez le vison. Il est conclu que chez Je furet:

Le contenu protémique s’accroît parallèlement avec l’expansion du blastocyste, s’élevant particulièrement lorsque l’embryon s’implante.

Les protéines présentes au début de la période sont de types qui diffusent facilement en électrophorèse; elles deviennent plus semblables à celles du sérum au moment de l’implantation.

Il n’y a aucune évidence d’une diapause embryonnaire. L’activité mitotique dans les cellules du blastocyste reste élevée (à peu près 5 %).

La période prolongée de pré-implantation résulte d’une longue durée de la mitose (80+ min) des cellules du blastocyste.

La teneur en protéines trouvée au cours de la période pré-blastocystaire est semblable à celle des liquides prélevés dans l’utérus de vison au cours de la période de diapause embryonnaire. Cette observation apporte de nouvelles preuves circonstanciées en faveur d’une relation entre la diapause accompagnent l’ovulation différée et la disponibilité limitée en protéines.

The author wishes to express his gratitude to John Hammond Jr. for supplying the animals used in this study, and for his constructive criticism of the manuscript, to Dr C. R. Austin and Dr D. A. T. New in whose laboratories the work was done, and to Dr John Marston, Mrs Pat Coppola and Mrs Venitha Dharmawardena for technical assistance. The work was supported in part by NSF Grant no. GB-6363, AEC Contract no. AT (11-1)-1597, and a faculty fellowship to the author from the University of Colorado.