Nicolas Young Lab Projects

Epigenetic inhibitors are an emerging and often very successful class of cancer drugs that target chromatin modifying enzymes, or sometime chromatin reading proteins to manipulate the chromatin regulatory system. We are currently working with several of these inhibitors, including some novel inhibitors developed here at Baylor College of Medicine, to understand in far greater detail that ever before what biochemical effects can be observed in response to treatment at the proteoform level. Our focus is to understand how specific combinations of PTMs respond differentially to these inhibitors and on what time scales. This will inform what types of cancer may respond to each inhibitor and how combinations of therapies may be effectively applied synergistically.

Histones are also relatively unique in that there are multiple copies of the gene for each class distributed to different genomic locations. These sometimes express the same protein product, as with histone H4, but often express slight variations on the theme, called variants. Some variants are known to contain different functions than the other members of the class. For example, the histone variant H3.3 is strongly correlated to transcriptional activity where the other H3 variants: H3.1 and H3.2 are more neutral. Our recent work has shown that histone H1 variants behave very differently and are not interchangeable. Histone H1.4 becomes hyper acetylated during mitosis where the other variants become only mono- or di-phosphorylated, despite having many of the CDK consensus site conserved. Previous work has mostly not distinguished between the variants when analyzing the PTMs, due specifically to this high level of conservation. In many cases, only top down proteomics can successfully analyze PTMs and variants simultaneously. We have developed methods to analyze histone proteoforms with variant specificity and continue to probe their biology.

Much of our current effort is focused on histone proteoforms; however, we are also developing efforts that encompass transcription factors, nuclear receptors and co-activator/co-repressor complexes. These are essential (and classical) components of the chromatin regulatory system that are intimately connected to histones and histone PTMs. Although some of these PTMs are connected to function, many more have proven ambiguous or enigmatic, since like histone PTMs often in concert and proteoform level biological function remains to be elucidated. The transcription factor on which we are focused are deeply connected to cancer (overexpressed or dysregulated in cancer) and this work may elucidate additional mechanisms that could be prone to therapeutic intervention.

Many cancers develop resistance to cancer therapies and other cancers are inherently not prone to certain therapies for reasons not fully understood. We are working to understand the role of chromatin regulatory mechanisms in drug resistance. Our initial results show that specific histone proteoforms are affected by treatment and that drug resistant models exhibit different responses. On top of this variations in the baseline epigenetic background appear to affect the observed response.

Complementary to our other projects we are also developing novel methods to pharmacologically profile the molecular level activities of cancer drugs. We are able to rapidly determine the proteins/enzymes affected the precise site of activity. This is achieved as a profile of all of the targets (on- and off-target) so that the wider picture of the potential beneficial and detrimental effects may be observed early in development. We will be working with collaborators to integrate this approach into an innovative drug discovery pipeline. Our major interests here also revolve around chromatin PTMs as many of the classes of drugs we are focused on inhibit chromatin modifying enzymes or chromatin reading proteins. This also allows us to compare molecular targeting data to the downstream effects that we measure on histone and TF proteoforms.

DNA damage response and repair has intrinsic connections to chromatin signaling; however, many of the details of the connections thereof are yet to be fully understood. The classical early signal for DNA damage that is widely used to confirm the presence of damage is gamma-H2A.X, which is a variant of histone H2A that has been phosphorylated at serine 139. The repair process itself requires remodeling of chromatin to access the underlying DNA. There have previously been specific combinations of histone PTMs that have been directly connected mechanistically to the DNA damage repair process. DNA damage agents are also front line cancer drugs that have been shown to sometimes work more effectively in combination of epigenetic inhibitors.

As an emerging field the development of computational tools to effectively analyze our data is essential, especially for the groundbreaking problems that we are addressing for the first time. Top down proteomics data is extraordinarily complex to begin and when minor variations in sequence in introduced it can sometimes lead to incorrect conclusions without careful data analysis. Often a degenerate set of proteoforms that are all structural isomers of one another (with only slight differences in the placement of PTMs) can confound analysis. We also take a very quantitative approach to top down proteomics which is less common in the already small community of top down researchers. This often requires custom algorithm development. To these ends we have developed a top down proteomic data analysis platform from the ground up with broad ranging capabilities.

We also are working toward a more through statistical measure of proteoform confidence. The standard approaches to estimate false discovery rate used in bottom up proteomics for protein identification are mostly irrelevant in proteoform identification. They inform us about correct protein identification and usually results in extraordinarily high confidence metrics with the extensive sequence information we obtain; however, it does not mean that the specific proteoform is correct. There are some scoring algorithms specific to proteoforms; however, it is still a developing field requiring continued development.

Once the proteoforms are measured, we also need to understand how they are related to each other, which ones are correlated in their behavior in time, dose, etc. and across multiple experiments. We also have built a set of custom data analysis tools and bioinformatics tools for addressing these questions.