Downregulation but lack of promoter hypermethylation or somatic mutations of the potential tumor suppressor CXXC5 in MDS and AML with deletion 5q

To the Editor: During recent years, mutations in epigenetic modulators have been identified in several human cancers, including acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) (1). CXXC5 has been found to be necessary for retinoic acidinduced differentiation of myelocytic leukemia cells, identifying CXXC5 as a candidate tumor suppressor for myeloid transformation (2). CXXC5 belongs to the CXXC-zinc finger domain family comprising 13 proteins. Hitherto, most of these have been shown to be important epigenetic regulators (3–7). Three of the proteins (MLL, MLL2 and TET1) are translocated/rearranged in leukemia, and notably the CXXC domain is preserved in all known MLL fusion proteins and essential for transformation. It is therefore tempting to speculate that CXXC5 might affect myelopoiesis through epigenetic regulation. Moreover, CXXC5 has been identified as an inhibitor of the WNT pathway (8,9). Chronic activation of this pathway is observed in several human cancers including AML (10). CXXC5 is located on chromosome 5q31.2, 20 kb downstream of the distal telomeric marker (D5S594) for the critically deleted region (CDR) for higher-risk MDS and AML (11,12). The close proximity to the CDR may affect CXXC5 expression by loss of regulatory sequences, or through larger deletions stretching into the 5q31.2 region, commonly observed in MDS/AML (2). Finally, CXXC5 could be disrupted either by promoter hypermethylation or by somatic mutations. Of note in this context, the transcription start site for CXXC5 is embedded in a CpG island, indicating that CXXC5 expression may be regulated by promoter methylation. We analyzed the expression level of CXXC5 mRNA in CD34+ cells from 22 MDS/AML patients all with large deletions on chromosome 5q not only covering the CDR, but also regions including CXXC5. In around half of the patients (13), the expression level of CXXC5 was lower than 50% of that of normal controls, indicating that the transcription was downregulated by others means (Fig. 1). Subsequently, we screened 102 AML and 80 MDS samples (32 with deletion of 5q, 10 AML and 22 MDS) for promoter methylation of CXXC5 by methylation-specific melting curves; however, all samples were unmethylated. Next, 80 MDS and 95 AML samples (32 with deletion of 5q) were screened for somatic mutations in the CXXC5 coding region by high-resolution melting. One potential missense mutation was observed in one low-risk MDS patient. The mutation results in a substitution of amino acid 13 from glycine to serine. The substitution is not located in the CXXC domain of CXXC5 (2). As normal, uninvolved tissue from the patient hasn’t been examined, we cannot confirm its somatic origin. However, this alteration was only observed in 1 of the 175 patients, so even if this is a somatic mutation with impact on the function of CXXC5, point mutations of CXXC5 are not important players in pathogenesis of MDS or AML. Haploinsufficiency alone has been suggested to influence the functionality of other genes in this region (13–17), and we also observed a >50% downregulation of CXXC5 mRNA in almost half of the 5q deleted cases. Thus, although targeted disruption of CXXC5 is not an important mechanism, haploinsufficiency or indirect downregulation of CXXC5 may still play a role in malignant myelopoiesis. This, however, is a subject for further study.