Metastatic breast cancer (MBC) is the most common cancer-related cause of death in women worldwide. Despite considerable advances in cancer therapy and modern targeted treatment approaches, many metastatic tumors are still incurable. One reason for this is the complex heterogeneity of tumors. Scientists have been limited in their ability to investigate this variety, partly due to methodological challenges. In a comprehensive study, an international team co-led by LMU biomedical scientist Johanna Klughammer in conjunction with Daniel Abravanel, Aviv Regev and Nikhil Wagle, Physician-Scientists and Researchers at Dana-Farber Cancer Institute and the Broad Institute of MIT and Harvard, has now employed diverse methods to analyze the diversity of metastatic cells and their interactions with their cellular environment. The results form a valuable basis for future studies and afford new insights into the biology of metastatic breast cancer.

Within the framework of the Human Tumor Atlas Network (HTAN), the researchers analyzed the totality of RNA molecules from 67 biopsies taken from 60 patients with breast cancer metastases. These so-called transcriptomes show which genes are active at a certain point in time. The researchers used two different approaches, measuring either the entire transcriptome of the cell or only that of the cell nucleus. However, these methods cause information about the spatial organization of cells to be lost. To capture this information, the researchers additionally created spatial expression profiles for 15 samples and analyzed serial tissue sections with up to four different spatially resolved methods.

Comprehensive dataset

“With this comprehensive dataset, we were able to compare the various methods and reveal commonalities and differences. Our results can help scientists in the future to select the most suitable method for their research question,” says Klughammer, who lead the computational analysis. The metastases studied by the team came from nine different locations in the body — including from the brain, liver, and bones — and represented different clinically relevant subtypes. “Although the broad approach somewhat limits the statistical power for individual variables, we found interesting clues about the biology of the metastatic cells nonetheless,” says Klughammer. This includes information about which cell types occur in the metastases, how certain genes are activated in these cells, and how the cells are spatially arranged in the metastasis.

One thing that surprised the researchers, for example, was that the malignant cells in the body of a patient exhibited very stable expression profiles. “From some patients, we investigated more than one biopsy and found very similar expression profiles, even when the metastases were located in different parts of the body or when the samples were taken up to 200 days apart,” says Klughammer. Conversely, the researchers found big differences between patients.

Cell profiles characterized

The researchers also found connections between certain clinical characteristics and different malignant expression phenotypes: Most malignant cells displayed the typical features of epithelial cells. In some samples, however, the researchers observed the expression of genes associated with stem cells, neurons, or cartilage. Interestingly, the sample with stem-cell-like behavior came from a patient who had the shortest survival time in the group despite early diagnosis and proper treatment. The sample with “cartilage-like” behavior was the only one with a histology typical for a rare and often difficult-to-treat form of breast cancer.

Metastases generally consist of cancer cells, the natural cells of the host tissue, and migrated immune cells that can potentially fight cancer cells. “An exciting discovery was that the expression profile of the tumor cells differed according to whether they had immune cells (so called T or NK cells) in their close vicinity or not,” says Klughammer. For example, the researchers found that tumor cells which seemed to keep T/NK cells out of their close vicinity were more likely to express the SOX4 gene compared to tumor cells of the same biopsy that had T/NK cells in their close vicinity. This provides further evidence that immune escape processes also happen in local neighborhoods and not only on a tumor-wide scale.

“Overall, our study offers new insights into the biology of metastatic breast cancer and shows the potential of spatial expression profiling,” explains Klughammer. “In the long term, our results could contribute toa more nuanced classification of patients and more targeted therapeutic approaches.”