Limits...
Promoter DNA methylation pattern identifies prognostic subgroups in childhood T-cell acute lymphoblastic leukemia.

Borssén M, Palmqvist L, Karrman K, Abrahamsson J, Behrendtz M, Heldrup J, Forestier E, Roos G, Degerman S - PLoS ONE (2013)

Bottom Line: CIMP- T-ALL patients had significantly worse overall and event free survival (p = 0.02 and p = 0.001, respectively) compared to CIMP+ cases.CIMP status was an independent factor for survival in multivariate analysis including age, gender and white blood cell count.Analysis of differently methylated genes in the CIMP subgroups showed an overrepresentation of transcription factors, ligands and polycomb target genes.

View Article: PubMed Central - PubMed

Affiliation: Department of Medical Biosciences, Pathology, Umeå University, Umeå, Sweden.

ABSTRACT

Background: Treatment of pediatric T-cell acute lymphoblastic leukemia (T-ALL) has improved, but there is a considerable fraction of patients experiencing a poor outcome. There is a need for better prognostic markers and aberrant DNA methylation is a candidate in other malignancies, but its potential prognostic significance in T-ALL is hitherto undecided.

Design and methods: Genome wide promoter DNA methylation analysis was performed in pediatric T-ALL samples (n = 43) using arrays covering >27000 CpG sites. Clinical outcome was evaluated in relation to methylation status and compared with a contemporary T-ALL group not tested for methylation (n = 32).

Results: Based on CpG island methylator phenotype (CIMP), T-ALL samples were subgrouped as CIMP+ (high methylation) and CIMP- (low methylation). CIMP- T-ALL patients had significantly worse overall and event free survival (p = 0.02 and p = 0.001, respectively) compared to CIMP+ cases. CIMP status was an independent factor for survival in multivariate analysis including age, gender and white blood cell count. Analysis of differently methylated genes in the CIMP subgroups showed an overrepresentation of transcription factors, ligands and polycomb target genes.

Conclusions: We identified global promoter methylation profiling as being of relevance for subgrouping and prognostication of pediatric T-ALL.

Show MeSH

Related in: MedlinePlus

Differently methylated genes in T-ALL are overrepresented by transcription factors and ligands and enriched for polycomb target genes.A) Lists of polycomb target genes identified in human embryonic fibroblasts and embryonic stem cells were compared with the DMGs in T-ALL in a Venn diagram. 260 genes were commonly present in these lists. B) The most variable CpG sites in T-ALL (1347 CpGs/1038 genes) and the 260 common polycomb genes identified in Figure 4A were evaluated for protein function, and compared with the protein function distribution of genes within the GeneGO database (23868 genes) and the Illumina methylation array (15077 genes). Transcription factors and ligands were overrepresented in the DMG in T-ALL compared with the distribution within the database and the entire methylation array.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3675104&req=5

pone-0065373-g004: Differently methylated genes in T-ALL are overrepresented by transcription factors and ligands and enriched for polycomb target genes.A) Lists of polycomb target genes identified in human embryonic fibroblasts and embryonic stem cells were compared with the DMGs in T-ALL in a Venn diagram. 260 genes were commonly present in these lists. B) The most variable CpG sites in T-ALL (1347 CpGs/1038 genes) and the 260 common polycomb genes identified in Figure 4A were evaluated for protein function, and compared with the protein function distribution of genes within the GeneGO database (23868 genes) and the Illumina methylation array (15077 genes). Transcription factors and ligands were overrepresented in the DMG in T-ALL compared with the distribution within the database and the entire methylation array.

Mentions: Our initial GeneGO Metacore analysis of DMGs indicated an overrepresentation of genes associated with the polycomb repressive complexes (PRC) 1 and 2. This is in line with recent reports on AML as well as solid tumor types where hypermethylation of PRC target genes has been observed [14], [26], [27]. To further investigate this we compared previously published lists of PRC target genes identified in human embryonic stem cells (Lee et al.) [25] and human embryonic fibroblasts (Bracken et al.) [24] with our list of most variable gene promoters in T-ALL (n = 1038) (Figure 4A). A high proportion of genes (62%) in our list were identified as polycomb target genes by Lee et al. and/or Bracken et al. The number of PRC target genes in our list was much higher than expected from a random selection (p<0.0001), indicating a preferential methylation of these genes (Figure 4A). We analyzed if the HumMeth27K methylation array was biased regarding the number of CpG sites measured per gene. The Lee and Bracken polycomb target gene lists were combined and compared with the entire HumMeth27K array for distribution of 1, 2 and ≥3 CpG sites per gene. The distribution was comparable, but with a slightly higher number of CpGs in polycomb target genes on the array. However, there was a highly significant overrepresentation of polycomb target genes among the differently methylated CpG sites in T-ALL that cannot be explained by the small array bias (Figure S3 A and S3 B).


Promoter DNA methylation pattern identifies prognostic subgroups in childhood T-cell acute lymphoblastic leukemia.

Borssén M, Palmqvist L, Karrman K, Abrahamsson J, Behrendtz M, Heldrup J, Forestier E, Roos G, Degerman S - PLoS ONE (2013)

Differently methylated genes in T-ALL are overrepresented by transcription factors and ligands and enriched for polycomb target genes.A) Lists of polycomb target genes identified in human embryonic fibroblasts and embryonic stem cells were compared with the DMGs in T-ALL in a Venn diagram. 260 genes were commonly present in these lists. B) The most variable CpG sites in T-ALL (1347 CpGs/1038 genes) and the 260 common polycomb genes identified in Figure 4A were evaluated for protein function, and compared with the protein function distribution of genes within the GeneGO database (23868 genes) and the Illumina methylation array (15077 genes). Transcription factors and ligands were overrepresented in the DMG in T-ALL compared with the distribution within the database and the entire methylation array.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3675104&req=5

pone-0065373-g004: Differently methylated genes in T-ALL are overrepresented by transcription factors and ligands and enriched for polycomb target genes.A) Lists of polycomb target genes identified in human embryonic fibroblasts and embryonic stem cells were compared with the DMGs in T-ALL in a Venn diagram. 260 genes were commonly present in these lists. B) The most variable CpG sites in T-ALL (1347 CpGs/1038 genes) and the 260 common polycomb genes identified in Figure 4A were evaluated for protein function, and compared with the protein function distribution of genes within the GeneGO database (23868 genes) and the Illumina methylation array (15077 genes). Transcription factors and ligands were overrepresented in the DMG in T-ALL compared with the distribution within the database and the entire methylation array.
Mentions: Our initial GeneGO Metacore analysis of DMGs indicated an overrepresentation of genes associated with the polycomb repressive complexes (PRC) 1 and 2. This is in line with recent reports on AML as well as solid tumor types where hypermethylation of PRC target genes has been observed [14], [26], [27]. To further investigate this we compared previously published lists of PRC target genes identified in human embryonic stem cells (Lee et al.) [25] and human embryonic fibroblasts (Bracken et al.) [24] with our list of most variable gene promoters in T-ALL (n = 1038) (Figure 4A). A high proportion of genes (62%) in our list were identified as polycomb target genes by Lee et al. and/or Bracken et al. The number of PRC target genes in our list was much higher than expected from a random selection (p<0.0001), indicating a preferential methylation of these genes (Figure 4A). We analyzed if the HumMeth27K methylation array was biased regarding the number of CpG sites measured per gene. The Lee and Bracken polycomb target gene lists were combined and compared with the entire HumMeth27K array for distribution of 1, 2 and ≥3 CpG sites per gene. The distribution was comparable, but with a slightly higher number of CpGs in polycomb target genes on the array. However, there was a highly significant overrepresentation of polycomb target genes among the differently methylated CpG sites in T-ALL that cannot be explained by the small array bias (Figure S3 A and S3 B).

Bottom Line: CIMP- T-ALL patients had significantly worse overall and event free survival (p = 0.02 and p = 0.001, respectively) compared to CIMP+ cases.CIMP status was an independent factor for survival in multivariate analysis including age, gender and white blood cell count.Analysis of differently methylated genes in the CIMP subgroups showed an overrepresentation of transcription factors, ligands and polycomb target genes.

View Article: PubMed Central - PubMed

Affiliation: Department of Medical Biosciences, Pathology, Umeå University, Umeå, Sweden.

ABSTRACT

Background: Treatment of pediatric T-cell acute lymphoblastic leukemia (T-ALL) has improved, but there is a considerable fraction of patients experiencing a poor outcome. There is a need for better prognostic markers and aberrant DNA methylation is a candidate in other malignancies, but its potential prognostic significance in T-ALL is hitherto undecided.

Design and methods: Genome wide promoter DNA methylation analysis was performed in pediatric T-ALL samples (n = 43) using arrays covering >27000 CpG sites. Clinical outcome was evaluated in relation to methylation status and compared with a contemporary T-ALL group not tested for methylation (n = 32).

Results: Based on CpG island methylator phenotype (CIMP), T-ALL samples were subgrouped as CIMP+ (high methylation) and CIMP- (low methylation). CIMP- T-ALL patients had significantly worse overall and event free survival (p = 0.02 and p = 0.001, respectively) compared to CIMP+ cases. CIMP status was an independent factor for survival in multivariate analysis including age, gender and white blood cell count. Analysis of differently methylated genes in the CIMP subgroups showed an overrepresentation of transcription factors, ligands and polycomb target genes.

Conclusions: We identified global promoter methylation profiling as being of relevance for subgrouping and prognostication of pediatric T-ALL.

Show MeSH
Related in: MedlinePlus