Study of proteins secreted by tumor cells points to new ways of combating cancer

Researchers at the Federal University of São Paulo set out to discover why proteins that should be found in cell nuclei ended up in cytoplasm or elsewhere in certain situations. This unexpected phenomenon may indicate a relevant pattern for diagnostic and treatment purposes.

By Janaína Simões  |  Agência FAPESP – A group of researchers at the Federal University of São Paulo (UNIFESP) in Brazil has set out to discover why under certain circumstances proteins that should be found in tumor cell nuclei end up elsewhere, in cytoplasm (the semifluid substance of a cell that is external to the nuclear membrane and internal to the cellular membrane) or even outside the cell. According to the scientists, this unexpected phenomenon could point to a pattern that is relevant to the diagnosis and prognosis of several types of cancer.

The main findings of the study are reported in an article published in the journal Traffic and featured on the cover of its November 2021 issue. In the study, Juliana A. de Morais and André Zelanis, researchers at the Functional Proteomics Laboratory in the Institute of Science and Technology (ICT-UNIFESP), re-analyzed public data posted to international repositories by scientists in various countries. The data used in the study refer to the “secretome”, the set of proteins expressed by the organism and secreted into the extracellular space.

A healthy cell has a repertoire of proteins required for basic functions, including histones, nuclear proteins around which DNA winds to form chromosomes. “Proteins are produced and sent to different parts of the cell. Some are sent outside the cell to perform functions such as creating the extracellular matrix,” Zelanis explained.

Information on the path to follow is in the structure of the molecule. “Proteins follow predefined secretion routes, which we call canonical pathways. Proteins needed outside the cell, for example, have signals pointing to this secretion pathway,” he said.

Tumor cells tend to be genomically unstable, containing many mutations in genes that regulate cell growth and proliferation. Protein secretion and other metabolic processes can become dysregulated.

Discarded data

Researchers often find proteins in unexpected places. While analyzing in vitro cultures in the laboratory, they typically separate cells from secreted molecules, and secretome databases are available for this reason.

The main hypothesis to explain protein secretion via non-canonical pathways was that the cell ruptured during collection. In other words, proteins that were in the “wrong” place were thought to have contaminated the area while the sample was being handled.

“Researchers have always considered such data to be due to the collection procedure and discarded it as irrelevant to the analysis. The originality of our study resides precisely in our decision to take a different approach to the investigation of proteins that aren’t secreted normally,” Zelanis said.

The first stage of the study entailed a proteomic analysis of public repository data on breast, colon and ovarian cancer cells, as well as melanoma and Ewing sarcoma, a rare type of cancer mainly seen in children and young adults, typically in bones and soft tissue. Proteins of interest were listed, and the researchers identified the proteins that had followed canonical pathways, eliminating them from the ensuing stages of the analysis.

A pattern was detected in the other group. Some proteins were sent to the cell nucleus via non-canonical pathways and were secreted in an “improper” manner. This pattern was observed in all tumor cell types analyzed.

The researchers also used the Human Protein Atlas, which contains images of histological sections for practically all proteins in normal and cancer tissues. 

“In the Atlas, we were able, for example, to look at a normal skin cell and a skin cancer cell to see where the proteins are located in both. Some of the proteins we found in cancer cells were in the cytoplasm [rather than the nucleus, as expected],” Zelanis said.

They identified a total of 6,092 unique proteins, 38% of which were not conventionally secreted (following non-canonical pathways) and did not originate in sample contamination. Of these, 19 were present in all the tumor cell types analyzed and were observed in cytoplasm, although their functions were associated with nuclei.

According to Zelanis, the same pattern has been found by researchers worldwide, not just in Brazil, so that biases due to the methods used by each researcher and laboratory to identify the proteins are highly unlikely.

“Next, we set out to see if there was any histological evidence that tumor proteins were in different locations. We found such evidence for some of the 19 proteins. This is another independent observation that validates our findings,” Zelanis said.

In his view, it is too soon to know whether the 19 proteins can be used as biological markers for diagnosis and treatment of tumors. “We only performed an analysis based on bioinformatics. Many other studies are needed to confirm the findings,” he said.

The results do not explain why the proteins followed non-canonical pathways and were secreted to cytoplasm or how the process relates to cancer, for example.

Next steps will consist of selecting proteins to investigate their functions in different types of cancer, starting with melanoma. “So far, we’ve studied proteins from cultured cell secretomes. Only one of the samples came from a patient. We now want to study plasma from melanoma patients to look for these proteins and find biomarkers,” Zelanis said. “We need to validate everything we’ve discovered so far by comparing samples from patients with and without cancer and seeing whether the process occurs only in tumor cells.”

The study was supported by FAPESP via two projects (19/07282-8 and 14/06579-3).

The article “Bioinformatic reanalysis of public proteomics data reveals that nuclear proteins are recurrent in cancer secretomes” is at: onlinelibrary.wiley.com/doi/10.1111/tra.12827.