Finally we would congratulate @jonstewart with his return to the Daily Show!

Our results establish ZNF143/ZFP143 as a conserved transcriptional regulator of cell proliferation and differentiation by safeguarding mitochondrial activity.

Using an in vitro embryo model revealed that ZNF143 is key for proper multicellular organism development.

But wait, there’s more!

Are these mitochondria still functional? Hell, no! ZNF143 loss causes mitochondrial membrane depolarization, mimicking known mitochondrial dysfunction phenotypes.

When we did live-cell imaging of mitochondria, we found that ZNF143 depletion changes mitochondrial morphology.

What are the consequences for the cells? Well, their proliferation and differentiation capabilities are dramatically reduced.

We found that ZNF143 binding sites are very conserved across human and mouse. ZNF143 binds to the promoters of many nuclear-encoded mitochondrial genes and its depletion results in their rapid and specific down-regulation at transcriptional and protein levels.

So, ZNF143 was incorrectly associated with chromatin looping for almost a decade! What is ZNF143 doing then?

How was it then picked up as a looping factor? By re-analysis of published ZNF143 occupancy data, we found that a commonly used antibody cross-reacts with CTCF.

Nothing! Our results show that ZNF143 has no function in chromatin looping nor on CTCF binding.

A decade ago, ZNF143 was found to bind with CTCF and suggested to be a looping factor. To study its direct role in chromatin looping, we generated a ZNF143 degron line in mouse embryonic stem cells.

In conclusion, we find that in MRT MYC is epigenetically upregulated by distal elements that we call RhOMEs. We show that these elements are used in a patient-specific manner, adding another layer of complexity to tumorigenesis, namely intertumoral epigenetic heterogeneity.

To our surprise the inactivation of BRD9 had very similar effects as reexpressing SMARCB1, suggesting that the tumorigenic phenotype is driven primarily by the non-canonical complex.

But what about the RhOMEs? Well, we find that also in the scATACseq data, patient-specific usage of RhOME1-3 is found. These results show that we can find recurrent patient-specific enhancers that activate MYC and likely drive proliferation.

Furthermore, we performed combined scATACseq and scRNAseq on patient tumors. While all the healthy cells cluster together, the tumor cells cluster by patient. Note that these tumors are karyotypically normal apart from the SMARCB1 loss!

When we investigated two other MRT organoid lines with 4C we were in for a surprise. In the first one we found the same super enhancer interaction. But in the second one we found two different enhancers interacting with the MYC promoter.

When we zoom into this region, we find a stretch of accessible chromatin that resembles a super enhancer which is lost when SMARCB1 is reconstituted (E+1.1Mb). We have additional analyses in the paper that this region is indeed a MYC super enhancer in MRTs.

When we zoom into a specific locus, i.e. the well-known MYC oncogene, we find that an interaction between the MYC promoter and a 1.1MB downstream region is lost when we re-express SMARCB1.

We then performed ATACseq with and without SMARCB1 reexpression (CTRL vs. SMARCB1+). We find two classes of accessible sites that are lost: 1) putative enhancer regions marked by H3K27ac and cohesin & 2) CTCF binding sites.