To me, HPLC is not a very sensitive method for detecting global hypomethylation. I remember in a paper, when DNMT is disrupted, the loss of methylation detected by HPLC is very subtle. Since HPLC measures directly methyl cytosine in DNA and given there is also regional hypermethylation, the reduction in the amount of methyl cytosines in cancer cells may not reach a level detectable by HPLC. There is recent review article on this topic which may be helpful.
Global DNA hypomethylation in cancer: review of validated methods and clinical significance. Clin Chem Lab Med. 2012 Oct 1;50(10):1733-42. doi: 10.1515/cclm-2011-0902.