To the Editor:

We read with interest a report by Kashiwagi et al. (2003) in the April issue of the journal, in which the authors detailed their results on imprinting studies in the mouse Atp10a gene, previously called as Atp10c/pfatp, the mouse ortholog of the human ATP10C gene. They analyzed the imprinting status using a nucleotide polymorphism in 30-week-old adult F1 hybrid mice of crosses between two different strains (C57BL/6J x JF1) of Mus musculus. They demonstrated that Atp10a is imprinted in a tissue-specific manner, with predominant expression of the maternal allele in the hippocampus and olfactory bulb.

Contrary to their results, we have recently reported that the mouse Atp10a gene escapes genomic imprinting in all fetal (embryonic day 16) and adult (6 weeks) tissues examined, including the hippocampus (Kayashima et al. 2003). In our imprinting study, the PWK strain of Mus musculus was mated with C57BL/6 J strain to make F1 hybrid mice, and a polymorphic site, different from that reported by Kashiwagi et al. (2003), was used for intron-spanning RT-PCR assay. Differences in materials and methods between the two studies were mouse strains, ages of adult mice, and polymorphic sites used for the imprinting analysis.

To verify the discrepancy in the imprinting status of Atp10a in adult mice, three F1 hybrid mice at the age of 30 weeks were investigated. The allelic expression of Atp10a was analyzed in RT-PCR assay using two polymorphic sites; one in Atp10a ORF (Kayashima et al. 2003) and the other in the 3'UTR reported in JF1 strain (Kashiwagi et al. 2003), which was also polymorphic in PWK strain. The experiments were independently performed in each of three F1 hybrids and revealed bi-allelic expression of the Atp10a gene in both the hippocampus and olfactory bulb (Fig. 1). On the other hand, an imprinted gene, Ube3a, showed predominant expression from the maternal allele in the identical cDNAs from the frontal cortex, olfactory bulb, and hippocampus (Fig. 1). We also analyzed the imprinting status in the primary brain-cell culture system and found that Atp10a was not imprinted in cultured neurons (data not shown), where Ube3a was completely imprinted (Yamasaki et al. 2003). These results confirmed that Atp10a escapes genomic imprinting whereas Ube3a shows neuron-specific imprinting in the brain in our F1 hybrid (PWK x C57BL/6 J) mice system.

Fig. 1.
figure 1

Imprinting analysis of Ube3a and Atp10a in 30-week-old adult F1 hybrid mice from female PWK x male C57BL/6 J mice. Predominant Ube3a expression from the maternal allele was demonstrated in the identical cDNAs from the brain samples (upper panel) using polymorphism (Yamasaki et al. 2003; Chamberlain and Brannan 2001). Bi-allelic expression of Atp10a was detected using a polymorphism (middle and lower panels) reported in the Atp10a ORF (Kayashima et al. 2003) and 3'UTR (Kashiwagi et al. 2003), respectively. All experiments were performed under the same PCR conditions as described previously (Yamasaki et al. 2003; Kayashima et al. 2003; Kashiwagi et al. 2003). Plus and minus signs mean with and without Tsp509I, HpyCh4IV, and MspI digestions (upper, middle, and lower panels), respectively. M and P mean RT-PCR products originated from the maternal and paternal alleles, respectively

Full size image

How can we explain such conflicting results in the imprinting status of Atp10a between two studies? The most plausible explanation may be strain background-dependent imprinting. A good example is the mouse Kvlqt1 gene, of which imprinted expression is detected in fetal tissues from female CAST/Ei (CS) mice x male 129/SvEv (129) mice, whereas Kvlqt1 exhibits partial loss of imprinting (LOI) in fetal tissues from female 129 x male CS mice (Jiang et al. 1998).

We have revealed bi-allelic expression of Atp10a in C57BL/6 J and PWK strains, but our result does not absolutely rule out its imprinted expression in other strains, including JF1. However, strain background-dependent imprinting cannot easily explain the results, because C57BL/6 J strain was used for a common parent of F1 hybrid mice in both studies. Other factors, such as the strain-dependent total expression level of Atp10a, may also affect its allele expression in C57BL/6 J and JF1. The fact that imprinting of the human ATP10C gene was not seen in the brain from one normal individual (Meguro et al. 2001) may support strain-background-dependent imprinting of mouse Atp10a.

In addition to repeated experiments in the F1 hybrid mice, further investigations on the imprinted expression of Atp10a in their F1 hybrid mice (C57BL/6 J x JF1) will elucidate the molecular mechanism of brain-specific imprinting. It also remains to be investigated when Atp10a imprinting becomes prominent in the brain regions and whether it correlates with imprinted Ube3a expression under control of an imprinting center.