In PCa. Although IGFBP3 also showed a greater underexpression in PCa oETS+ when compared to PCa ERG+, there was a non-significant increase in IGFBP3 expression after DAC treatment of the ETV1positive LNCaP cell line, precluding a consistent relationship between higher IGFBP3 methylation levels and ETV1 rearrangements. We also confirmed that TGFBR2 expression is reduced in PCa [50?1], which is compatible with the tumor suppressor role of TGFBR2 in PCa cells described by others [52], but MedChemExpress ��-Sitosterol ��-D-glucoside promoter methylation 1676428 does not seem to be involved. On the other hand, we found that NR0B1 was poorly expressed in PCa and in NPT, so our data do not support the previously reported immunoreactivity of DAX1 (protein encoded by the NR0B1 gene) in a significant proportion of PCa [53]. Based on our microarray findings of differential expression of ECRG4, LDOC1, 
HIST1H4L and KCNN2 between PCa harboring ERG rearrangements and those without ETS fusions, we decided to validate these data in an independent series of tumors. Among the five genes downregulated in PCa ERG+, we choose ECRG4 and LDOC1 for further study based on their tumor suppressor activity in other cancer models (see below; [43?4]). We also evaluated the expression of these genes in ESFT and ARMS in order to verify if there was any significant difference in their expression that might be attributable to EWSR1-ETS rearrangements. We here report for the first time that expression of both ECRG4 and LDOC1 is significantly decreased in PCa when compared to NPT. However, this was independent of the ETS status, contrarily to our initial microarray data suggesting a specific underexpression in PCa with ERG fusion genes. Consistent with a recent study that has associated CpG island hypermethylation of ECRG4 with recurrence in prostate carcinoma [54], our MSP analysis showed a significantly higher methylation frequency in PCa comparing with NPT, thus representing a mechanism of gene silencing that might be involved in all order KDM5A-IN-1 molecular subgroups of PCa. In LNCaP and 22Rv1 cell lines, however, DAC treatment was not sufficient to allow de novo ECRG4 expression, thus suggesting that other regulatory mechanisms may act in ECRG4 underexpression. The mechanism of LDOC1 downregulation is currently unknown, but because we did not find aberrant promoter methylation at this locus, other epigenetic or genetic alterations are probably causally involved. Finally, although the chimeric EWSR1-FLI1 protein has been found to bind the promoter of both LDOC1 and ECRGin vitro [20], we here show that their expression is not significantly different between ESFT and ARMS, thus suggesting that either the expression of these genes is not regulated by that chimeric protein in ESFT or that a different regulatory mechanism in ARMS is regulating the expression of LDOC1 and ECRG4 to similar levels. Our microarray findings of differential expression of HIST1H4L and KCNN2 in different molecular subsets of PCa were confirmed by qRT-PCR in an independent series. HIST1H4L is a gene that encodes a histone, which is a basic nuclear protein responsible for the nucleosome structure of the chromosomal fiber in eukaryotes. We here show for the first time that HIST1H4L expression is specifically and significantly increased in PCa harboring ERG fusion genes, both when compared to other PCa molecular subtypes and with NPT. These findings indicate that HIST1H4L is a potential target of ERG fusion genes, as also illustrated by our demonstration of direct bin.In PCa. Although IGFBP3 also showed a greater underexpression in PCa oETS+ when compared to PCa ERG+, there was a non-significant increase in IGFBP3 expression after DAC treatment of the ETV1positive LNCaP cell line, precluding a consistent relationship between higher IGFBP3 methylation levels and ETV1 rearrangements. We also confirmed that TGFBR2 expression is reduced in PCa [50?1], which is compatible with the tumor suppressor role of TGFBR2 in PCa cells described by others [52], but promoter methylation 1676428 does not seem to be involved. On the other hand, we found that NR0B1 was poorly expressed in PCa and in NPT, so our data do not support the previously reported immunoreactivity of DAX1 (protein encoded by the NR0B1 gene) in a significant proportion of PCa [53]. Based on our microarray findings of differential expression of ECRG4, LDOC1, HIST1H4L and KCNN2 between PCa harboring ERG rearrangements and those without ETS fusions, we decided to validate these data in an independent series of tumors. Among the five genes downregulated in PCa ERG+, we choose ECRG4 and LDOC1 for further study based on their tumor suppressor activity in other cancer models (see below; [43?4]). We also evaluated the expression of these genes in ESFT and ARMS in order to verify if there was any significant difference in their expression that might be attributable to EWSR1-ETS rearrangements. We here report for the first time that expression of both ECRG4 and LDOC1 is significantly decreased in PCa when compared to NPT. However, this was independent of the ETS status, contrarily to our initial microarray data suggesting a specific underexpression in PCa with ERG fusion genes. Consistent with a recent study that has associated CpG island hypermethylation of ECRG4 with recurrence in prostate carcinoma [54], our MSP analysis showed a significantly higher methylation frequency in PCa comparing with NPT, thus representing a mechanism of gene silencing that might be involved in all molecular subgroups of PCa. In LNCaP and 22Rv1 cell lines, however, DAC treatment was not sufficient to allow de novo ECRG4 expression, thus suggesting that other regulatory mechanisms may act in ECRG4 underexpression. The mechanism of LDOC1 downregulation is currently unknown, but because we did not find aberrant promoter methylation at this locus, other epigenetic or genetic alterations are probably causally involved. Finally, although the chimeric EWSR1-FLI1 protein has been found to bind the promoter of both LDOC1 and ECRGin vitro [20], we here show that their expression is not significantly different between ESFT and ARMS, thus suggesting that either the expression of these genes is not regulated by that chimeric protein in ESFT or that a different regulatory mechanism in ARMS is regulating the expression of LDOC1 and ECRG4 to similar levels. Our microarray findings of differential expression of HIST1H4L and KCNN2 in different molecular subsets of PCa were confirmed by qRT-PCR in an independent series. HIST1H4L is a gene that encodes a histone, which is a basic nuclear protein responsible for the nucleosome structure of the chromosomal fiber in eukaryotes. We here show for the first time that HIST1H4L expression is specifically and significantly increased in PCa harboring ERG fusion genes, both when compared to other PCa 
molecular subtypes and with NPT. These findings indicate that HIST1H4L is a potential target of ERG fusion genes, as also illustrated by our demonstration of direct bin.
