Machine learning application identifies novel gene signatures from transcriptomic data of spontaneous canine hemangiosarcoma

Hemangiosarcoma Cells Promote Conserved Host-derived Hematopoietic Expansion

Hemangiosarcoma and angiosarcoma are soft-tissue sarcomas of blood vessel-forming cells in dogs and humans, respectively. These vasoformative sarcomas are aggressive and highly metastatic, with disorganized, irregular blood-filled vascular spaces. Our objective was to define molecular programs which support the niche that enables progression of canine hemangiosarcoma and human angiosarcoma. Dog-in-mouse hemangiosarcoma xenografts recapitulated the vasoformative and highly angiogenic morphology and molecular characteristics of primary tumors. Blood vessels in the tumors were complex and disorganized, and they were lined by both donor and host cells. In a series of xenografts, we observed that the transplanted hemangiosarcoma cells created exuberant myeloid hyperplasia and gave rise to lymphoproliferative tumors of mouse origin. Our functional analyses indicate that hemangiosarcoma cells generate a microenvironment that supports expansion and differentiation of hematopoietic progenitor populations. Furthermore, gene expression profiling data revealed hemangiosarcoma cells expressed a repertoire of hematopoietic cytokines capable of regulating the surrounding stromal cells. We conclude that canine hemangiosarcomas, and possibly human angiosarcomas, maintain molecular properties that provide hematopoietic support and facilitate stromal reactions, suggesting their potential involvement in promoting the growth of hematopoietic tumors.

SIGNIFICANCE

We demonstrate that hemangiosarcomas regulate molecular programs supporting hematopoietic expansion and differentiation, providing insights into their potential roles in creating a permissive stromal-immune environment for tumor progression.

Blocking tri-methylguanosine synthase 1 (TGS1) stops anchorage-independent growth of canine sarcomas

Tri methylguanosine synthase 1 (TGS1) is the enzyme that hyper methylates the hallmark 7-methyl-guanosine cap (m7G-cap) appended to the transcription start site of RNAs. The m7G-cap and the eIF4E-cap binding protein guide canonical cap-dependent translation of mRNAs, whereas hyper methylated cap, m2,2,7G-cap (TMG) lacks adequate eIF4E affinity and licenses entry into a different translation initiation pathway. The potential role for TGS1 and TMG-capped mRNA in neoplastic growth is unknown. Canine sarcoma has high translational value to the human disease. Cumulative downregulation of protein synthesis in osteosarcoma OSCA-40 was achieved cooperatively by siTGS1 and Torin-1. Torin-1 inhibited the proliferation of three canine sarcoma explants in a reversible manner that was eliminated by siRNA-downregulation of TGS1. TGS1 failure prevented the anchorage-independent growth of osteo- and hemangio-sarcomas and curtailed sarcoma recovery from mTOR inhibition. RNA immunoprecipitation studies identified TMG-capped mRNAs encoding TGS1, DHX9 and JUND. TMG-tgs1 transcripts were downregulated by leptomycin B and TGS1 failure was compensated by eIF4E mRNP-dependent tgs1 mRNA translation affected by mTOR. The evidence documents TMG-capped mRNAs are hallmarks of the investigated neoplasms and synergy between TGS1 specialized translation and canonical translation is involved in sarcoma recovery from mTOR inhibition. Therapeutic targeting of TGS1 activity in cancer is ripe for future exploration.

Multimetric feature selection for analyzing multicategory outcomes of colorectal cancer: random forest and multinomial logistic regression models

Colorectal cancer (CRC) is one of the most common cancers worldwide, and a leading cause of cancer deaths. Better classifying multicategory outcomes of CRC with clinical and omic data may help adjust treatment regimens based on individual's risk. Here, we selected the features that were useful for classifying four-category survival outcome of CRC using the clinical and transcriptomic data, or clinical, transcriptomic, microsatellite instability and selected oncogenic-driver data (all data) of TCGA. We also optimized multimetric feature selection to develop the best multinomial logistic regression (MLR) and random forest (RF) models that had the highest accuracy, precision, recall and F1 score, respectively. We identified 2073 differentially expressed genes of the TCGA RNASeq dataset. MLR overall outperformed RF in the multimetric feature selection. In both RF and MLR models, precision, recall and F1 score increased as the feature number increased and peaked at the feature number of 600-1000, while the models' accuracy remained stable. The best model was the MLR one with 825 features based on sum of squared coefficients using all data, and attained the best accuracy of 0.855, F1 of 0.738 and precision of 0.832, which were higher than those using clinical and transcriptomic data. The top-ranked features in the MLR model of the best performance using clinical and transcriptomic data were different from those using all data. However, pathologic staging, HBS1L, TSPYL4, and TP53TG3B were the overlapping top-20 ranked features in the best models using clinical and transcriptomic, or all data. Thus, we developed a multimetric feature-selection based MLR model that outperformed RF models in classifying four-category outcome of CRC patients. Interestingly, adding microsatellite instability and oncogenic-driver data to clinical and transcriptomic data improved models' performances. Precision and recall of tuned algorithms may change significantly as the feature number changes, but accuracy appears not sensitive to these changes.