Sir,

We read with interest the article of Morimoto et al (2004) that investigated the relation of epigenetic inactivation of CIITA with levels of HLA-DR, one of the MHC class II genes, in haematopoietic tumour cells.

Recent DNA microarray studies have showed that decreased levels of MHC class II molecules are associated with high metastatic potential and/or poor prognosis, not only in haematopoietic tumour (Rimsza et al, 2004) but also in several adenocarcinomas (Ramaswamy et al, 2003) and hepatocellular carcinoma (HCC) (Iizuka et al, 2003), suggesting the common role of MHC class II molecules in tumour progression. In this regard, elegant work by Morimoto et al (2004) could provide a therapeutic strategy against various malignancies. The authors stated that, by suppressing the expression of MHC class II molecules, epigenetic inactivation of CIITA provided a survival advantage to a subset of haematopoietic tumours (Morimoto et al, 2004); however, they did not investigate the relation between levels of CIITA and HLA-DR and tumour progression.

Our recent DNA microarray study showed that decreased levels of HLA-DRA gene were related to recurrence of HCC (Iizuka et al, 2003). We also confirmed that levels of HLA-DR protein by tumour cells were related to recurrence of HCC (unpublished data). Using the DNA microarray data set of HCC (available at http://surgery2.med.yamaguchi-u.ac.jp/research/DNAchip/hcc-recurrence/index.html), we investigated the relation between levels of CIITA (three probes, U18288, U18259, and X74301) and those of HLA-DRA (probe X00274) and HLA-DRB1 (probe M33600). Given that CIITA levels in HCC were markedly low, it is reasonable to assume that this phenomenon may be due in part to its epigenetic inactivation. However, in our data, there were no correlations between levels of CIITA and those of the two HLA-DR genes. Thus, in the context of human cancer tissues, the transcriptional regulation of HLA-DR is likely to be much complex. From this standpoint, we recommend that the authors will clarify the relation of CIITA methylation status to constitutive expression of HLA-DR in a larger cohort of haematopoietic tumour. On the basis of their finding (Morimoto et al, 2004) and our present finding, in our opinion, methylation status of CIITA could be a therapeutic target combined with interferon-gamma for preventing the progression of haematopoietic tumour and HLA-DR may be superior to CIITA as a marker for progression of haematopoietic tumour.

Additionally, tumour HLA-DR might have a function other than the antigen-presenting function. Recently, Altomonte et al (2003) showed a possible involvement of HLA-DR in a signalling pathway linked to cell adhesion in melanoma cells. Thus, we expect that further work by Morimoto et al would provide a clue to incorporate MHC class II antigens into the mainstream of molecular basis underlying tumour progression.