Several features of the overgrowth phenotype observed in Day 80 SCNT and IVF fetuses of the present study, e.g., disproportionate body and organ growth, normal number of placentomes but increased placentome size, have been reported previously but at later stages of gestation. Bovine IVF fetuses at Day 70 of gestation, in contrast, were not heavier than fetuses resulting from embryos produced in vivo. Recent data on embryonic and fetal development have indicated an initial growth retardation of in vitro-derived concepti, which disappeared toward the end of the first trimester of pregnancy, suggesting that disturbances in early development may cause a compensatory growth or a disruption in the regulation of fetal growth. The point of inflexion in growth between in vitro-produced fetuses and controls appeared to occur between Days 65 and 72 of development. The age of investigated fetuses, and differences in in vitro protocols that might affect the severity of the overgrowth phenotype, could thus explain different results.

Epigenetic perturbations that could affect the expression of genes involved in embryonic development and fetal growth regulation have been suggested as possible causes of the LOS syndrome and epigenetic change at a single locus has been associated with LOS in ovine fetuses. Several groups have observed significant hy-permethylation of DNA in bovine preimplantation SCNT embryos on a sequence-specific, chromosomal and global scale, and the level of methylated cytosine has been correlated with aberrant histone modification and the developmental potential of cloned embryos.

The significant hypermethylation (9.2% vs. AI control fetuses) of liver DNA from Day 80 SCNT fetuses investigated in the present study extends these observations to viable cloned fetuses and suggests that aberrant epigenetic modifications due to incomplete reprogramming in the early bovine embryo persist at least until the end of the first trimester of pregnancy. This is consistent with the general stability of established DNA methylation patterns in somatic cells, but it is presently unknown if hypermethylation would still be detectable at term. However, DNA methylation data reported for preimplantation embryos and our present data obtained from samples of viable fetuses are in striking contrast with the recently reported hypomethylation of DNA obtained from skin biopsies of viable and aborted SCNT fetuses aged 53-62 days and 58-210 days, respectively. A variety of protocols and nuclear donor cells are employed to generate SCNT embryos, and the general difference (hypermethylation vs. hypomethylation) between our data and the data presented by Cezar et al. could therefore be due to differences in SCNT protocols and/or nuclear donor cells. The analysis of overgrown SCNT fetuses in the present study and the different types of tissues used for DNA isolation, liver and placenta in the present study vs. skin in the study of Cezar et al., could also contribute to the observed differences. Methylation of DNA in some samples from the latter study might also have been affected by degradation of DNA in aborted fetuses. The percentage of 5mC reported by Cezar et al. for DNA from individual viable SCNT or IVF fetuses and adult SCNT or AI cattle ranged from 10% to 40%, although it is generally established that less than 6% of cytosine residues in mammalian DNA are methylated (e.g., ). Similar values have also been reported for various bovine tissues. While we cannot presently explain the absolute differences between our data and the data published by Cezar et al., our results have been derived with a highly accurate method that has been designed to quantify DNA methyla-tion levels only from intact, chemically derivatizable DNA. In addition, the overall methylation levels detected in our samples ranged from 2.98% to 3.71% 5mC for liver DNA samples and from 1.18% to 2.39% 5mC for cotyledon DNA samples. This is in excellent agreement with published data on a variety of differentiated mammalian tissues. Hypomethylation of placental vs. embryonic DNA has been previously described in the mouse.

An additional and interesting result of the present study is the apparent hypermethylation of liver DNA from IVF fetuses. Although the difference in 5mC content between IVF samples and controls was not significant (P < 0.18), it is striking that this value is ~50% of the significant difference between SCNT samples and controls, and thus corresponds to the intermediate phenotypic values observed for IVF fetuses in relation to the phenotypic differences observed between AI controls and SCNT fetuses. Above all, this includes the reduced crown-rump length:thorax ratio that was found to be a highly indicative external hallmark of the overgrowth phenotype observed in both SCNT and IVF fetuses (Figs. 1 and 3A). It has been proposed that demethylation of somatic donor nuclei after nuclear transfer could occur by the same active demethylation mechanism by which the paternal genome is demethylated after fertilization, suggesting that an important trigger for demethyl-ation is remodeling of chromatin by factors present in the oocyte cytoplasm. This could provide a link between fetal overgrowth in bovine SCNT and IVF fetuses, dependent on the in vitro production protocol of IVF embryos. The IVF fetuses investigated in the present study were generated according to a protocol that consistently produces overgrown fetuses, based on differences in oocyte maturation and culture conditions, as compared with a protocol that yields phenotypically normal IVF fetuses (unpublished data). A further indication for such a common mechanism is provided by the positive regressions of phenotypic parameters reflecting the overgrowth phenotype on liver 5mC content, which approached significance at P < 0.0730.079, and, most important, the significant negative regression observed for crown-rump length:thorax ratio that was characteristic of the disproportionate overgrowth phenotype observed in SCNT and IVF fetuses, on liver 5mC (Fig. 5D).

The developmental potential and phenotypic outcome of embryos derived by in vitro techniques is so far rather unpredictable. Accumulating evidence suggests that epigenetic mechanisms may be disturbed by manipulation and/or culture conditions. We show for the first time that hypermethylation of fetal DNA is associated with disproportionate overgrowth of SCNT fetuses and is, to a lesser extent, also observed in IVF fetuses. Quantitative analysis of overall DNA methylation further documents the limitations of current embryo technologies and may be a useful parameter for the evaluation and improvement of IVF and SCNT procedures.