Involvement of Allele-Specific Methylation of Asparagine Synthetase Gene in Asparaginase Sensitivity of BCP-ALL

Asparaginase is one of the most important components for the treatment of ALL. ALL cells are supposed to be unable to synthesize adequate amounts of Asparagine (Asn), and, therefore, depend on extracellular source of Asn to survive. Asparaginase therapy induces the depletion of serum Asn by catalyzing the deamination of Asn and leads to cell death of ALL cells. Asparagine synthetase (ASNS) is an enzyme that produces Asn from Aspartic acid. Thus, silencing of the ASNS gene in ALL cells could be crucial for complete starving ALL cells of the Asn. Considering that the ASNS gene has a CpG island in its promotor, aberrant methylation of CpG island could be one of epigenetic mechanisms for silencing of ASNS gene in ALL cells. Previous qualitative analysis of ALL samples using methylation-specific restriction enzyme revealed frequent methylation of CpG island in the ASNS gene. However, associations of methylation status of ASNS gene with its expression level and sensitivity to asparaginase in ALL cells remain unknown. Moreover, little is known about mechanisms for leukemia-specific ASNS gene silencing by methylation. To shed light on these issues, we analyzed a large panel of BCP-ALL cell lines. We quantified ASNS gene expression level by real time RT-PCR in 79 BCP-ALL cell lines cultured in the presence or the absence of L-asparaginase (L-asp), and determined IC₅₀ values of L-asp using alamar blue assay. In the majority of cell lines, although degree of the induction was highly variable, ASNS gene expression level was upregulated in the presence of L-asp. IC₅₀ value of L-asp showed significant correlation with ASNS gene expression level cultured in the presence of L-asp (r=0.222, p=0.049) rather than that in the absence of L-asp (r=0.193, p=0.089). We next analyzed methylation status of the ASNS gene in 79 BCP-ALL cell lines by bisulfite PCR sequencing using a next-generation sequencer (NGS). Strong correlation was confirmed between mean % methylation by NGS and Sanger sequencing in representative cell lines. Of importance, mean % methylation in 79 BCP-ALL cell lines showed significant negative correlation with ASNS gene expression level cultured in the presence of L-asp (r=-0.482, p=6.73x10^-6) and, subsequently, IC₅₀ value of L-asp (r=-0.39, p=3.86x10^-4). Unexpectedly, % methylation of 79 cell lines distributed in three clusters; 15 cell lines (19%) were highly methylated (>66%, median; 89%), 26 cell lines (32.9%) were moderately methylated (33-66%, median; 40%), and 38 cell lines (48.1%) were weakly methylated (<33%, median; 3.7%). In the majority of moderately methylated cell lines, histograms of % methylation in each read of NGS showed two peaks of high and low methylation, suggesting an allele-specific methylation. In the middle of CpG island, tandem repeat polymorphism of 14bp nucleotides is located adjacent to methylation-specific restriction enzyme site of Aor13HI. Of note, in 7 out of 8 moderately methylated cell lines with heterozygous tandem repeat genotype, only single PCR product was detectable when PCR was performed after Aor13HI treatment, whereas two PCR products derived from two- and three-repeat alleles was detectable when PCR was performed without treatment, indicating an allele-specific methylation. We next analyzed a possible one-allele-loss of the ASNS gene in highly methylated (>66%; 8 cell lines) and weakly methylated (<20%; 12 cell lines) cell lines. We directly sequenced genotype in a portion of introns 2 and 4 and exon 5 based on the imputated SNP genotypes, and confirmed heterozygous genotype in every cell lines at least in one of eight SNPs analyzed, demonstrating that loss-of-heterozygosity is not the mechanism for high or low methylation of the ASNS gene. Similar pattern of methylation was observed in 52 BCP-ALL samples. Taken together, these observations indicate that stepwise allele-specific methylation of ASNSgene is critically involved in the sensitivity to L-asp of BCP-ALL.