Table 1.
Tissue (n - TCGA) | Master transcription factor | Alteration frequency in TCGA database | Uniprot predicted coactivators | Alteration frequency in TCGA database | Uniprot predicted corepressors (http://www.uniprot.org/) | Alteration frequency in TCGA database |
---|---|---|---|---|---|---|
Liver (n = 442) | GATA4 6, 119 | 67% Hetloss | ARID1A | 44% Hetloss, fs* | KDM1B | 42% Amp, Gain |
ARID2 | 17% Hetloss, fs* | BAZ1B | ||||
FOXA1 120 | 3% Amp Gain | KMT2A | 28% Hetloss, fs* | SUZ12 | 32% Amp, Gain | |
SMARCA4 | 24% Hetloss, fs* | DNMT1 | ||||
FOXA2 120 | 30% Amp Gain | SMARCAD1 | 46% Hetloss | BAZ2A | 25% Amp, Gain | |
15% Amp, Gain | ||||||
14% Amp, Gain | ||||||
Pancreas (n = 109) | GATA4 121 | 49% Hetloss | ARID1A | 49% Hetloss, fs* | BAZ1B | 57% Amp, Gain |
ARID1B | 61% Hetloss | DNMT1 | ||||
GATA6 121 | 23% Hetloss | ARID3C | 43% Hetloss | UHRF1 | 50% Amp, Gain | |
SMARCD1 | 50% Hetloss | SUZ12 | ||||
PTF1A 122 | 20% Hetloss | SMARCB1 | 34% Hetloss | BAZ2A | 45% Amp, Gain | |
FOXA2 123 | 43% Amp Gain | 40% Amp, Gain | ||||
PDX1 | 58% Hetloss | 39% Amp, Gain | ||||
Ovary (n = 302) | GATA4 83 | 69% Hetloss | ARID3A | 91% Hetloss | EZH2 | 38% Gain amp |
ARID3B | 50% Hetloss | DNMT1 | 39% Gain amp | |||
FOXL1/2 82, 87 | 67% Amp, Gain | ARID3C | 37% Hetloss | BAZ1A | 18% Gain amp | |
SMARCAD1 | 68% Hetloss | EED | 39% Gain amp | |||
FOXO1 124 | 63% Hetloss | ARID1B | 67% Hetloss | |||
ARID1A | 44% Hetloss | |||||
SMARCA1 | 47% Hetloss |
We analyzed TCGA data in cBioPortal to determine genetic alterations in genes mediating differentiation pathways. Lineage specific transcription factors were identified using lineage tracing studies. Cofactors interacting with lineage specific transcription factors were determined using data deposited in UniProt database (http://www.uniprot.org/). Master transcription factors are lineage specific and they recruit various coactivators to cooperate and turn on differentiation genes. While heterozygous loss of GATA4 and inactivation by frameshift mutations of GATA4 coactivators are frequent in hepatocellular carcinoma, other master transcription factors such as FOXA1 are available to mediate differentiation pathways. However, corepressors such as KDM1B, which are also recruited by these TFs, are aberrant in HCC by copy number gains and amplification. Such alterations impair ability for differentiation to ensue in HCC through epigenetic suppression of target genes5, 88, 89. These forms of alterations are commonly observed also in PDAC, and OVC. Since corepressors are either gained or amplified in cancer but not inactivated by frameshift mutations, inhibition of these enzymes may serve as logical molecular targets of therapy (Fig. 5c, d)
*Frameshift mutation; hetloss, heterozygous deletion