Microenvironmental control of malignant growth

The tumor microenvironment (TME) comprises a complex milieu of different cell types, including cancer associated fibroblasts (CAFs) and immune cells, blood vessels, and the extracellular matrix. Through its interaction with cancer cells, it plays an essential role in cancer invasion and metastasis. The inherent complexity of the TME presents a challenge to study it within experimental model systems. It underscores the importance of complementing such research with observation from human tumor tissues, wherein this intricate complexity is preserved.

In Paper IV, we introduce a new software designed to explore the Human Protein Atlas, an online database that includes image data on the protein expression across normal and cancerous tissues from immunohistochemically (IHC) stained tissues.

In Paper I, we use this software to identify 12 novel proteins expressed in cancerassociated fibroblasts, four revealing connections to Rho-kinase signaling. We contrast their expression across various tumors and against normal tissue fibroblasts, uncovering expression variability among cancer types and confirm their similarities with the myofibroblastic phenotype.

In Paper II, we explore the expression of the proteoglycan Decorin, abundantly present in normal connective tissue and having tumor inhibitory properties, showing its downregulation in the connective tissue surrounding tumors.

In Paper III, based on our observations in Paper I of the connection of Rhosignaling in CAFs, we study the effects of knocking out the related RhoA in fibroblasts both in vitro and in vivo models. We demonstrate that the knockout fibroblasts compromise their tumor inhibitory capacity, enhancing cancer cell growth, migration, and metastasis.

In Paper VI, we develop a new method for analyzing the extensive data within the Human Protein Atlas by developing a deep-learning-based image classifier. Utilizing a limited training image set, we classify all images available for the prostate, identifying 44 new markers of prostate basal cells.

In Paper IV, we explore the influence of the TME on cancer cells by systematically analyzing 20 pancreatic cancer patient samples utilizing an IHC panel. We define shifts in cancer cell phenotype relative to tissue localization, including a transition to a more indolent cancer phenotype, an effect on cancer cell proliferation, and a tendency to normalize the cancer cell phenotype.

In conclusion, we developed two new methods that enable us to study protein expression in normal and cancerous tissues by enhancing the capabilities of the HPA. We identified new markers of CAFs and revealed a connection to Rhosignaling. Knocking out the related RhoA in experimental systems resulted in the fibroblasts losing their cancer inhibitory capacity. Finally, we show the remarkable plasticity of cancer cells, demonstrating that their phenotype undergoes significant alterations based on their spatial localization within normal tissue.

List of scientific papers

I. Novel Signatures of Cancer-associated Fibroblasts. International Journal of Cancer. 133, no. 2 (2013): 286-293. Bozóky, B., Savchenko, A., Csermely, P., Korcsmáros, T., Dúl, Z., Pontén, F., Székely, L., & Klein, G.
https://doi.org/10.1002/ijc.28035

II. Decreased Decorin Expression in the Tumor Microenvironment. Cancer Medicine. 3, no. 3 (2014): 485-491. Bozoky, B., Savchenko, A., Guven, H., Ponten, F., Klein, G,. & Szekely, L.
https://doi.org/10.1002/cam4.231

III. RhoA Knockout Fibroblasts Lose Tumor-inhibitory Capacity in Vitro and Promote Tumor Growth in Vivo. Proceedings of the National Academy of Sciences. 114, no. 8 (2017): E1413-E1421. Alkasalias, T., Alexeyenko, A., Hennig, K., Danielsson, F., Lebbink, R. J., Fielden, M., Turunen, S. P., Lehti, K., Kashuba, V., Madapura, H. S., Bozoky, B., Lundberg, E., Balland, M., Guvén, H., Klein, G., Gad, A. K., & Pavlova, T.
https://doi.org/10.1073/pnas.1621161114

IV. Stabilization of the classical phenotype upon integration of pancreatic cancer cells into the duodenal epithelium. Neoplasia. 23(12), (2021): 1300-1306. Bozoky, B., Fernández Moro, C., Strell, C., Geyer, N., Heuchel, R. L., Löhr, J. M., Ernberg, I., Szekely, L., Gerling, M., & Bozóky, B.
https://doi.org/10.1016/j.neo.2021.11.007

V. AtlasGrabber: a software facilitating the high throughput analysis of the human protein atlas online database. BMC Bioinformatics, 23(1), (2022): 546. Bozoky, B., Szekely, L., Ernberg, I., & Savchenko, A.
https://doi.org/10.1186/s12859-022-05097-9

VI. Identification of novel protein markers of prostate basal cells by application of deep learning to images from the Human Protein Atlas. Bozoky, B., Szekely, L., Alexeyenko, A., Ernberg, I., Petrov, I. [Manuscript]