Figure 1

Altered histological structure of the liver in pathological perinatal clones.
(A) Liver sections stained with haemalin-eosin-saffron are shown for one AI control foetus, one cloned foetus, one control calf and one cloned calf. Fibrotic areas are revealed by their yellow appearance following saffron staining. The arrows indicate disorganization of the rows of hepatocytes and the asterisks highlight examples of intracytoplasmic lipid vesicles corresponding to mild steatosis. Scale bar: 25 μm. (B) Thirty animals were graded in categories according to the severity of seven types of lesions (Tables 1 and 2), and MCA was performed. Categories containing fewer than four individuals were merged, generating G1G2, S1S2, A2A3, and I1I2. Upper left panel: barplot of the contribution of each dimension to total variance, showing a marked drop between the first and second dimensions and then smaller drops. Upper right panel: variable factor map, illustrating the contribution of each variable to dimensions 1 and 2. Lower panel: individual and category factor map, with 95% confidence ellipses shown for each group. The proximity of individuals to a category indicates that this category is more represented among these individuals than in the remaining population. (C,D) For each animal, the area and a shape factor representing circularity (1: perfect circles) were measured on at least 200 hepatocyte nuclei and the mean (C) and CV (D) were calculated. The CV is indicative of cell heterogeneity in the tissue. Each dot represents one animal. Median values are indicated by horizontal lines. Open dots: perinatal AI controls; filled dots: perinatal clones. Prenatal (circles) and postnatal (squares) animals were considered separately. Significant differences between AI and clones and between prenatal and postnatal animals are indicated (p < 0.05, permutation test). (E) For perinatal animals with clearly observable cell limits, the cell area and shape factor were measured on 35 hepatocytes, and the mean was calculated. The differences between AI controls and clones were not significant.