Long overdue… sorry; American Society of Hematology is a very clinically oriented conference. This doesn’t mean that some truly inspirational basic research isn’t happening, because it is, but simply that the translational side of moving basic science to the patient is really pronounced. Maybe that’s why we enjoyed ourselves so much. On to the highlights for us bioinformaticians;
Identification of Leukemia Associated Minor Histocompatibility Antigens (mHA) through Computational Prediction and Targeted Mass Spectrometry:
Jefferson L. Lansford, BS, University of North Carolina at Chapel Hill
Obviously, immunotherapy is hot as well as using GWAS to find selective targets; but the combination of the two has been thus far, lackluster. This is due to the relatively unsuccessful approach of using GWAS to confirm mHAs. It’s hard to say the least. But, it’s kinda important to identify mHAs with tight binding affinity, narrow tissue restriction and high leukemia cell prevalence for immunotherapy methods to work. Well, this lab took a good stab at this hard problem. First, they in silico predicted mHA peptides from genotype data in HLA-matched allogeneic stem cell transplant (SCT) donor-recipient pairs. Then, they estimated the peptide/HLA binding with NetMHCpan (PMC3319061) and Human Protein Atlas combined with in-house RNAseq data. We love that NetMHCpan is based on two open sources of data: EMBL-EBI’s IMGT/HLA database and the Immune Epitope Database and Analysis Resource (IEDB). Anyhow, this lab then biologically validates their in silico predictions by purifying their in silico designed and engineered peptide epitopes from cell cultures and then measuring the predicted the selected mHA presentation and binding affinities of the epitopes with targeted ion-mobility tandem MS. Super cool combination of genome, proteome and chem-informatics and then validating with some newer MS instrumentation methods and applied to a very hard problem.
RNA Methylation and Its Role in the Hematopoietic System:
Frank Lyko, PhD, German Cancer Research Center
Ok, change gears.. Epigenetics. We’ve done a little work around DNA methylation; but mind explosion… RNA is methylated as well with apparently distinct regulatory roles. And what ?!?!?.. tRNA is differentially methylated. Expands the control of the central dogma, eh? I guess I shouldn’t be surprised. Same players however, an old friend… DNMT2. But these guys show some pretty convincing evidence in mice that our friend is also a tRNA methyltransferase and active in hematopoiesis. I smell a new therapy modality.
The next snippet is close to A2IDEA’s heart.. It’s actually our main mission. Shelley Herbrich, a talented graduate student, over at UT-MD Anderson Cancer Center surveyed a bunch public data to nominate novel targets in AML leukemic stems cells.
Robust Bioinformatics Approach for Identifying Novel AML LSC Targets: Putative Role of Galectin-1 in the Immune-Microenvironment.
Of note is, rather than doing the standard multiple hypothesis testing statistical tests after integrating all the studies used; she actually used the intersection of findings from each of the individual studies after statistical analysis to put forth some candidates. We’ve personally heard arguments for and against such methods. Anyhow, she then confirmed the putative targets in TCGA data and then, found a selective small molecule inhibitor (OTX008) for one of the putative targets, LGALS1 (cell surface and extracellular player for modulating cell-cell interactions). With biological validation, she ran into some roadblocks b/c a simple knockdown or shRNA inhibition of the target didn’t do much in vitro, but when they specifically looked at the secreted form, i.e., took growth media from shLGALS1 cells, they saw increased cell death of activated T cells and suppression of proliferation; which appear to be CD4+ / CD8+ T cell specific when testing in PBMCs. LGALS1 may just be player in the immune microenvironment of leukemic cells and uncovered from the re-analysis of public data.
Next up; how could ASH be ASH without some Long-Non-Coding RNA (lncRNA).
Implication of the Long Non-Coding RNA Crnde in Multiple Myeloma (MM)
Bertrand Arnulf, MD, PhD, APHP
Again, a project that starts with public data mining to reveal ~ 100 lncRNAs. From CD138+ plasmocytes healthy and MM patients. From which, the authors nominate a candidate thought to effect the physical association of polycomb group proteins. They used CRISPER/Cas9 to delete their candidate (Colorectal Neoplasia Differentially Expressed lncRNA) or CRNDE to show a dose-dependent repressive effect on myeloma cell growth. Again, another therapeutic modality to investigate.. Inhibition of lncRNA all put forth, at least in this case, by re-analyzing / mining public data.
Then there was a nice talk integrating proteomics data with metabolomics data to further understand my old friend HSP90.
HSP90 Facilitates Oncogene-Induced Metabolic Reprogramming in B-Cell Lymphomas.
Maria Nieves Calvo Vidal, PhD, Weill Cornell Medicine
Using the HSP90 inhibitor, PU-H71, in cell lines, this lab initially did untargeted proteomics and metabolomics data generation and overlaid the two datasets with the aid of KEGG to determine coordinated regulatory hubs. They then went on to measure compounds in a few enriched pathways with targeted methodologies to suggest a few enzyme assemblies that correlate with common features found in DLBCL and Burkitt lymphoma which are affected by HSP90 inhibition.
Another unbelievably satisfying ASH. Scientifically one of the best, disease focused, conferences we attend. It’s always so rewarding to not only see our personal bioinformatics work displayed, but to see all the other great ideas, methodologies and uses of data analyses helping patients.