The fact that MAGE I protein co-localizes with KAP1 to Ki67 supports this model. It is important to note that the effects of MAGE-A3 and MAGEC2 are thus far always similar for a given gene and do not appear to be random, indicating a true biologic effect. Since both MAGE-A3 and MAGE-C2 have similar effects on each of the genes studied the reason for the opposite effects does not appear likely to reside in the MAGE I proteins and is more likely a function of the particular KZNF. KZNFs bind KAP1 by their KRAB domains, which are heterogeneous and may contain members of one or both of several KRAB A and KRAB B motifs. We speculate that different KRAB domains may bind to KAP1 with differing affinities or at slightly different sites, or may interact with the ubiquitin ligase complexes in different ways, leading either to increased KZNF ubiquitination and decreased binding seen with ID1, or to increased binding and co-localization with MAGE I seen with Ki67. Except for sites of DNA damage, where KAP1 binds DNA in a sequence independent fashion, KAP1 normally requires binding to KZNFs to guide it to specific genes. KAP1 binds to,7,000 sites in the human genome and there are about 270 KZNFs, indicating that some KZNFs must bind to more than one site. Unfortunately, antibodies are not available for most KZNFs and it is not known where most of them bind, making it problematic and beyond the scope of this work to completely analyze their regulation and function, and thus identify the KZNFs that bind to Ki67. While the results with Ki67 seem to indicate an anti- proliferative affect of MAGE I proteins, it is important to consider the role of MAGE I in the DNA damage response, which requires a temporary pause in the cell cycle for DNA repair. Interestingly, when looked at over the course of 72 h we saw an initial drop in Ki67 expression followed by slow partial recovery, compatible with a temporary cell cycle halt. While it remains to be determined how MAGE I regulation of Ki67 may affect overall cell proliferation or DNA damage repair, when these data are combined with our studies of ZNF382 they show unequivocally that MAGE I proteins may have differential effects on different KAP1 KZNF targets. BKM120 because transmission of signals in biologic systems may be complex and intertwined, it is often problematic to assign a direct causal relationship to regulation of specific genes by an individual molecule or sets of molecules. Genes in areas of chromatin relaxation may still be negatively regulated by other mechanisms, so removal of KAP1 mediated repression by MAGE I expression does not guarantee protein expression. However, the ability of KZNF targeted KAP1 to suppress specific genes by causing localized chromatin condensation trumps other forms of regulation because if the gene is condensed, it cannot be transcribed. Therefore, our finding that MAGE I expression affects KAP1 binding and chromatin condensation unequivocally defines direct gene regulation by MAGE I proteins. Glaucoma is a leading cause of blindness worldwide, characterized by specific visual field defects due to the loss of retinal ganglion cells and damage to the optic nerve head. The result is a patchy loss of vision, generally in a peripheral to central manner.