Livestock cloning and in vitro embryo culture have been adversely affected by the exceptional size of some resulting lambs and calves 1 ; multiple abnormali- ties associated with ‘large offspring syn- drome’ (LOS) currently limit application of these technologies 1 . Similar fetal overgrowth in humans and mice can result from altered expression of several imprinted genes 2 (for example, H19 (ref. 3), Igf2 (ref. 4), Igf2r (refs. 5,6)) that are only expressed from one parental allele. Cloning or non-physiological embryo culture environments may result in inappropriate epigenetic modification of imprinted genes during early embryo-genesis 2 , when many allele-spe- cific imprints are established or main- tained. Epigenetic changes in pre-implantation embryos may affect gene expression during later fetal development, in a manner similar to those in mouse embryonic stem (ES) cells 7 . This also has implications for the safe development of therapeutic embryo-derived stem cell appli- cations, as epigenetic alterations in imprinted genes are associated with tumorigenesis. Although a variety of cloning and in vitro embryo culture procedures have been asso- ciated with LOS, the incidence has been neither predictable nor reproducible 1 . In sheep, we established reliable models for LOS in which in vivo fertilized eggs are cul- tured in vitro for five days with co-cultured granulosa cells and/or serum before trans- fer into recipient ewes 8 . Of 48 fetuses recov- ered at day 125 of gestation (term=147 d), 12 weighed 5.5 kg or more (the largest was 8.2 kg) and were defined as large offspring (LO) for gene expression analyses (the largest fetus derived from a non-cultured control embryo was 4.5 kg, n=22). Using relative RT–PCR (Fig. 1a; ref. 9), we observed no difference in the steady-state tran- script levels of IGF2 mRNA in liver, kidney, forelimb extensor carpi radialis muscle and heart between control and LO fetuses (Table 1). Thus, the mechanism underlying ovine LOS is not analagous to the fetal overgrowth typical of Beckwith Weidemann syndrome in humans, which is induced by loss of imprint- ing and overexpression of IGF2 (ref. 4). By contrast, expression of IGF2R (Fig. 1a, b) in LO fetuses was reduced by 30–60% relative to the control group (Table 1). Expression of IGF2R was not significantly different between the tissues of control fetuses and those fetuses derived from cultured embryos within the weight range of the control group (non-large cultured (NLC); data not shown). Moreover, tissue IGF2R expression was not correlated with fetal weight for the control and/or NLC groups, although significant negative correlations were obtained when the LO fetuses were included in the analyses (liver –0.571, P=0.001; heart –0.675, P<0.001; kid- ney –0.492, P=0.001; muscle –0.635, P<0.001). By probing western ligand blots of tissue protein extracts and plasma samples with 125 I-IGF2 (ref. 10; Fig. 1c,d), we found 61% and 81% reductions in muscle and liver IGF2R protein levels, respectively (P<0.001; Table 1), and a 67% reduction in the circulating form (P<0.001; Table 1). In view of the 30% increased birth weight in mice with disrupted Igf2r expression 4–6 , our observations indicate a causative role for IGF2R in LOS in sheep. To investigate whether reduced IGF2R expression was due to an epigenetic effect, we cloned the sheep second intron differentially methylated region (DMR2) from a λ phage genomic library, using ovine cDNA probes cloned by RT–PCR of bovine exons 2 and 3 (Fig. 1e). Restriction analysis of the 2.2-kb CpG island revealed approximately 70% methylation at a Mbo1/Sau3AI site probed with a 726-bp SacI DMR2 fragment in control fetal heart and a complete loss of methylation in 9 of 12 LO individuals Fig. 1 Validation of relative RT–PCR assay and ovine IGF2R/DMR2 identification. a, Rep- resentative validation of RT–PCR assay to quantify levels of steady-state transcript (in this case liver IGF2R) relative to the 18S ribo- somal subunit. Both IGF2R (primers J03527: 5´ 530–551, 3´ 769–750) and 18S were amplified in a single tube, with an appropriate cycle number and ratio of 18S primer (attenuating Competimer (Ambion) 9 to ensure linear amplifi- cation over at least a 10- fold dilution of cDNA). b, RT–PCR amplifications representing IGF2R and 18S in control and LO liver samples. c, Immuno-blot- ting of sheep fetal plasma using an IGF2R polyclonal antibody revealed a single band that co-migrates with purified bovine IGF2R and the ≅200-kD band identified by western lig- and blot analysis of fetal plasma. d, IGF2 western ligand blot representing control and LO plasma samples identify IGF2R protein and the IGF- binding proteins. e, NotI (N) fragment map of ovine IGF2R DMR2. f, Representative South- ern blots of control and LO heart genomic DNA digested with SacI (S1) alone or plus methyla- tion-insensitive (MboI; m) or methylation-sensitive (Sau3AI; s) isochizomers and probed with a 726-bp SacI fragment of DMR2. brief communications nature genetics • volume 27 • february 2001 153 Epigenetic change in IGF2R is associated with fetal overgrowth after sheep embryo culture Manipulation or non-physiological embryo culture environments can lead to defec- tive fetal programming in livestock. Our demonstration of reduced fetal methylation and expression of ovine IGF2R suggests pre-implantation embryo procedures may be vulnerable to epigenetic alterations in imprinted genes. This highlights the potential benefits of epigenetic diagnostic screening in developing embryo procedures. 200ng100ng 50ng 20ng - 18S - IGF2R CONT LO - 18S - IGF2R immunoblot ligand blot bovine IGF2R sheep plasma bovine IGF2R sheep plasma CONT LO - IGF2R IGFBP2 IGFBP1 IGFBP4 18.5kb 2kb IGFBP3 doublet CONT LO DMR2 DMR2 N s m s m S1 s m s m N S1 m/s ex 3 ex 2 726bp 726bp 653bp S1 a b c d e f © 2001 Nature Publishing Group http://genetics.nature.com © 2001 Nature Publishing Group http://genetics.nature.com