Progressive and predictive markers of disease evolution: platelet transcriptome in chronic myeloproliferative neoplasms

Introduction:: Predicting disease progression remains a particularly challenging endeavor in several age-related disorders and chronic cancers, thus limiting early detection, risk stratification, and preventive interventions. Classical engineering principles, such as systems approaches, would significantly benefit these biological problems, but remain underutilized to date. Novel technologies developed using omic-data (e.g. disease-specific cell types), paired with bioinformatic and systems biology approaches will have transformative consequences for diagnosis, prognosis, and the development of new therapeutics, with increasing likelihood of deploying such approaches in clinical settings too.



Blood platelets are one of the few human cells that live and function without a nucleus. Likely because they are anucleate, platelets have been traditionally viewed as limited to hemostatic function. Emerging data point otherwise: platelets have emerged as versatile effectors of diverse and dynamic function, including in inflammatory and immune dysfunction, of critical significance in cancer.



Cancer is associated with dysregulated platelet function and known to effect associated cardiovascular outcomes (e.g. cancer-associated thrombosis and bleeding, or immuno-thrombosis). Cancer is also accompanied by a decreased tolerance to physiological stress, and an increase in local and systemic inflammatory responses. Dysregulated inflammation, in turn, contributes to an array of pathological processes, wherein again, emerging data point to platelets as cellular effectors through direct and/or indirect mechanisms. Together, we now understand that cancer is associated with platelet reprogramming, and an altered platelet molecular signature that essentially mirrors the underlying disease.


Materials and Methods:: Considering the critical role of inflammation in the vascular risk; and considering emerging evidence for platelet functional interplay between thrombosis, inflammation, and cancer, we studied the platelet transcriptome in a large clinical cohort of chronic progressive blood cancer patients, myeloproliferative neoplasms (MPN), and three phenotypic subtypes, essential thrombocythemia (ET, n=24), polycythemia vera (PV, n=33), and myelofibrosis (MF,n=42).



MPN patient blood samples were obtained with informed consent and approval by the Stanford Institutional Review Board. To minimize effects of long-term storage of platelets at room temperature and loss of platelet RNA quality and quantity, samples were processed within 6 hours after blood collection. Established methods were followed for platelet isolation, purification, and RNA isolation. Sequencing and analysis on platelets from independent human blood samples were performed at the Stanford Functional Genomics facility. Samples were sequenced for 36 cycles, paired-end, on the Illumina GAIIx sequencer. Sequence reads were processed, and all downstream informatics/statistical analyses performed with a combination of packages in the open-source programming language R, version 3.3.3 (https://cran.r-project.org/).



In addition, we leveraged machine learning algorithms (Lasso penalized regression) to predict the advanced subtype, MF from the either healthy controls, ET or PV. We evaluated platelet transcriptomic features with clinical features in two sets of training and validation cohorts. We also assessed the progressive and monotonic upward or downward trend in gene expression using Mann-Kendall trend test and identified statistically significant progressive genes across all three MPN subtypes.


Results, Conclusions, and Discussions:: Results: Using platelet transcriptome profiling, we observed dynamic shifts in MPN immune inflammatory profile and preferential expression of interferon-, proliferation-, and proteostasis-associated genes as a progressive gradient across the three MPN subtypes. We quantify how platelet gene expression, alternative splicing and the associated molecular pathways are altered in each of ET, PV, and MF with respect to healthy donors. We observe that the platelet transcriptome mirrors disease progression; and offers disease-specific signatures with the greatest distinction from controls in the most severe phenotype, MF. Differential markers in each of ET, PV and MF also highlight candidate genes as potential mediators of the pro-thrombotic and pro-fibrotic phenotypes in MPNs. Most importantly, platelet gene expression profiling in MPNs offers directions for prediction of myelofibrosis. Under two conditions of external validation temporal (using our two cohort design) and geographical (independently published datasets in MF), we uniquely discriminate MPN subtypes from each other, and healthy donors using three model types and predict MF at high accuracy. The highest performing model used a set of progressively differentiated MPN genes at an area under the (ROC) curve of 0.96 (temporal) and 0.97 (geographical); and rendered a core signature of < 5 candidate markers as top predictors of disease progression.



Conclusions: In summary, we present a comprehensive catalog of the platelet transcriptome in chronic progressive cancers (MPN) with immediate relevance to defining subtype-specific molecular differences and predicting the advanced phenotype, myelofibrosis. The platelet transcriptome is significantly reprogrammed in the MPN setting, with a wealth of transcript associations that may be missed in using conventional tissue sources such as serum, plasma, whole blood, or bulk bone marrow.



Discussion: While this work has been focused on the overarching progressive platelet transcriptome across MPNs, these data open an important avenue for utilizing platelet RNA signatures to better understand specific MPN complications such as the risk of thrombosis and bleeding, or fibrosis and disease transformation. Altogether, this study demonstrated in chronic MPNs provides a comprehensive framework for exploiting the platelet transcriptome and may inform future studies toward mechanistic understanding and therapeutic development in MPNs, and potentially other age-related disorders.


Acknowledgements (Optional): : Acknowledgements This work was funded by US National Institutes of Health grants 1K08HG010061-01A1 and 3UL1TR001085-04S1 (research re-entry award) to A.K, 1S10OD018220 and 1S10OD021763 shared instrumentation grants to the Stanford Functional Genomics Facility and the Stanford Research Computing Center, and the Charles and Ann Johnson Foundation to J.G. All authors thank the patients at the Stanford Cancer Center for their generous participation in this research, and the Stanford Functional Genomics Facility for genomic data storage.

References (Optional): :


1 Weyrich, A. S. Platelets: more than a sack of glue. Hematology Am Soc Hematol Educ Program 2014, 400-403 (2014).

2 Best, M. G., Vancura, A. & Wurdinger, T. Platelet RNA as a circulating biomarker trove for cancer diagnostics. J Thromb Haemost 15, 1295-1306 (2017).

3 Rondina, M. T., Weyrich, A. S. & Zimmerman, G. A. Platelets as cellular effectors of inflammation in vascular diseases. Circulation research 112, 1506-1519 (2013).

4 Best, M. G., Wesseling, P. & Wurdinger, T. Tumor-Educated Platelets as a Noninvasive Biomarker Source for Cancer Detection and Progression Monitoring. Cancer Research 78, 3407-3412 (2018).

5 Angelini, D. E. & Khorana, A. A. Building a CAT clinic - real-world systems approaches to prevention and treatment. Thromb Res 208, 173-175 (2021).

6 Mahajan, A., Brunson, A., Adesina, O., Keegan, T. H. M. & Wun, T. The incidence of cancer-associated thrombosis is increasing over time. Blood Adv 6, 307-320 (2022).

7 D'Ambrosi, S., Nilsson, R. J. & Wurdinger, T. Platelets and tumor-associated RNA transfer. Blood 137, 3181-3191 (2021).

8 In 't Veld, S. & Wurdinger, T. Tumor-educated platelets. Blood 133, 2359-2364 (2019).

9 Tjon-Kon-Fat, L.-A., Sol, N., Wurdinger, T. & Nilsson, R. J. A. Platelet RNA in Cancer Diagnostics. Semin Thromb Hemost 44, 135-141 (2018).

10 Krishnan, A. & Thomas, S. Toward platelet transcriptomics in cancer diagnosis, prognosis, and therapy. British Journal of Cancer In Press (2021).

11 Shen, Z., Du, W., Perkins, C., Fechter, L., Natu, V., Maecker, H. et al. Platelet transcriptome identifies progressive markers and potential therapeutic targets in chronic myeloproliferative neoplasms. Cell Reports Medicine 2, 100425 (2021).

12 Thomas, S. & Krishnan, A. Platelet Heterogeneity in Myeloproliferative Neoplasms. Arterioscler Thromb Vasc Biol 10.1161/ATVBAHA.121.316373, ATVBAHA121316373 (2021).