PDF, English Download (6MB) | Terms of use

Abstract

In 2018 about 2 million people died from cancer in Europe alone, with 90% of cancer related deaths being due to metastatic spread of the disease. Despite the high impact of metastasis for patients` survival, the molecular mechanisms of metastatic progression are still poorly understood and therapeutic opportunities are limited. Metastasis can occur months to years after removal of the primary tumor, rendering potential therapeutic interventions even more difficult. This suggests that therapy resistant cancer cells can survive in the body, able to initiate metastasis after latency periods. Our knowledge of the molecular mechanisms enabling these cells to survive in unfavorable microenvironments is limited and specific therapeutic targeting of these cells is currently not possible. In this study breast cancer dissemination was investigated in orthotopic in vivo models using different human breast cancer cell lines as well as patient derived xenograft (PDX) models. Breast cancer cells were not only detected in lung, liver and bone, organs that are prone to develop breast cancer metastases, but also in kidney, pancreas and spleen. These organs rarely develop metastases in patients and show no metastatic growth in our models. The integrity and viability of those cells was confirmed by in vitro cultivation of cancer cells, isolated from different organs. Thus, it was hypothesized that disseminated cancer cells (DCCs) in kidney, pancreas and spleen may resemble a subpopulation of cancer cells in a metastatic latency period. To further investigate this hypothesis the in vivo location of DCCs in different organs, their survival capability as well as growth potential was further analyzed. It was confirmed that the majority of DCCs in the kidney had extravasated, but stayed non-proliferative and were located in close proximity to blood vessels. Furthermore, DCCs from PDX models, located in kidney, pancreas and spleen, survive for a prolonged time period without significant growth after resection of the mammary tumor. DCCs isolated from kidney, pancreas and spleen maintain their growth potential and were able to initiate metastatic growth in the lung upon intravenous injection. While being able to form lung metastases, injected DCCs still do not grow in their organ, originally isolated from. This suggests that not an aggressive subpopulation was selected but that the microenvironment influenced the growth pattern of these cells. Taken together, it was demonstrated that breast cancer DCCs survive in the non-supportive organ environments of kidney, pancreas and spleen without growth and maintain their metastatic potential. To identify survival mechanisms applied by DCCs, gene expression profiling of DCCs isolated from kidney and pancreas compared to cancer cells from the mammary tumor was performed, analyzing the transcriptional profile of DCCs being distinctly different from cells of the mammary tumor. However, substantial similarities were identified between DCCs in kidney and pancreas. These observations suggested that DCCs despite their different organ of isolation may use similar survival mechanisms in unfavorable microenvironments. Based on the similarities of DDCs from kidney and pancreas, it was hypothesized that mutual transcriptomic changes in kidney and pancreas DCCs may be regulated intrinsically. A second gene expression profile of cancer cells from kidney, pancreas, lung and mammary tumor, having been cultured in vitro for 48h, revealed that a significant percentage of the transcriptomic changes in DCCs were maintained upon short-term culture, supporting the hypothesis that these changes are regulated cancer cell intrinsically. Furthermore, DCCs from kidney and pancreas have a different transcriptomic profile compared to lung derived metastatic cells, suggesting that DCCs resemble a latent subpopulation in the metastatic cascade. As the aim was to identify survival cues of DCCs that may be used therapeutically, the common signaling pathways and functions in kidney and pancreas DCCs were analyzed in more detail. Using Gene Set Enrichment Analysis (GSEA) and Gene Ontology (GO) analysis molecular mechanisms and pathways were identified. Mechanisms of epithelial to mesenchymal transition (EMT) as well as cell death mechanisms were repressed in DCCs from both organs. A reduced rate of apoptotic events in DCCs of the kidney was confirmed by cleaved caspase 3 staining compared to lung metastasis and cells from the mammary tumor. Furthermore, based on the gene expression profiles, pronounced metabolic changes were observed in DCCs. Genes involved in metabolic pathways including energy, glucose and nucleotide metabolism were downregulated in DCCs. A closer look revealed that DCCs underwent a metabolic switch from glycolysis to oxidative phosphorylation. Genes involved in glucose import into the cell as well as enzymes of the glycolytic pathway were repressed at the transcriptional level. In contrast, the expression of genes, encoding for proteins of the electron transport chain in mitochondria, was upregulated in DCCs. These metabolic changes were accompanied by a downregulation of hypoxia response genes. This downregulation may happen in context of the metabolic switch from glycolysis to oxidative phosphorylation as hypoxia leads to induction of glycolysis as several of the glycolytic enzymes are direct target genes of the hypoxia induced transcription factor HIF1. Antigen presentation mediated by major histocompatibility complex II (MHCII) molecules was repressed on mRNA and on protein level in DCCs as well. The repression of antigen presentation is most likely caused by cancer cell autonomous mechanisms as the experiments were performed in immune compromised mouse models. Further investigations suggested that YAP1 regulated genes, which were upregulated in DCCs, may negatively regulate MHC II gene expression. In vitro experiments confirmed that YAP1 can regulate MHC II transcript as well as protein levels. Moreover, DNA repair mechanisms and cell cycle checkpoint genes were enriched in DCCs from kidney and pancreas. Lastly, the transcriptomic profile of DCCs from kidney and pancreas showed enrichment for chemotherapy resistance signatures. Following up on this enrichment, the therapy resistance of DCCs was further investigated in vivo using two different chemotherapeutic treatments, a combination of Doxorubicin and Cyclophosphamide as well as monotherapy with Paclitaxel. While cancer cells in the mammary tumor and metastatic cells in the lung responded well to the treatment, DCCs in kidney, pancreas and spleen were not affected and the number of cancer cell in these organs remained practically unchanged. Thus, in this model breast cancer DCCs in kidney, pancreas and spleen resemble a pool of cancer cells able to survive therapy. Due to the central role of chemotherapy resistance for the outcome of therapeutic interventions, the chemotherapy resistance mechanisms of DCCs were analyzed in more detail. The chemotherapy resistance signature, enriched in DCCs, was compared to the genes upregulated in kidney and pancreas DCCs to identify overlaps. The comparison revealed an overlap of the tetraspanin (TSPAN) gene family members TSPAN8 and TSPAN1 in all three datasets. To investigate the importance of the identified molecules, the fractions of TSPAN expressing cancer cells were analyzed in a mouse model as well as in patient samples. The TSPAN8 single as well as TSPAN8 and TSPAN1 double-positive cell population was increased in DCCs in kidney, pancreas and spleen compared to the mammary tumor. Furthermore, TSPAN8 single positive as well as TSPAN8 and TPSAN1 co-expressing cancer cells were detected in ascites and pleural effusion samples of three out of four breast cancer patients by FACS analysis. As survival without growth was identified as an essential characteristic of DCCs from kidney, pancreas and spleen, the role of TSPAN8 and TSPAN1 for survival was investigated. Under sphere forming conditions in vitro, knockdown of TSPAN8 and TSPAN1 resulted in an increased rate of apoptosis. Furthermore, TSPAN8 and TSPAN1 co-expressing cells were enriched in the label retaining, non-proliferative population of spheres. Due to the context dependency of TSPAN functions, the role of TSPAN8 and TSPAN1 was also analyzed in DCCs. TSPAN8 and TSPAN1 knockdown were injected in the mammary fat pad, DCC were isolated and transcriptomic analysis was performed compared to control cells. This experiment confirmed an association of TSPAN8 and TSPAN1 with chemotherapy resistance of DCCs as the therapy resistance signature enriched in DCCs was lost upon knockdown of TSPANs. The transcriptional data from TSPAN8 and TSPAN1 knockdown DCCs compared to control DCCs indicated further that TSPAN8 and TSPAN1 may also regulate stem cell properties in DCCs in vivo. Functionally, in vivo injections of TSPAN8 and TSPAN1 double knockdown cells confirmed the crucial role of these two molecules for survival and chemotherapy resistance of DCCs. The knockdown significantly reduced the number of cancer cells in pancreas and spleen. In addition, knockdown of TSPAN8 and TSPAN1 sensitized DCCs to chemotherapy. A combination of the knockdown with Doxorubicin and Cyclophosphamide treatment resulted in a significant reduction of the cancer cell numbers in kidney, pancreas and spleen in vivo compared to chemotherapy alone. In conclusion, this study provides insights into the biology of dormant disseminated breast cancer cells and may reveal novel opportunities to develop therapeutic strategies against dormant DCCs during metastatic latency periods. Depletion of TSPAN1 and TSPAN8 rendered DCCs sensitive to chemotherapy. Thus, the combination of TSPAN8 and TSPAN1 inhibition with chemotherapy might be a therapeutic strategy worth considering further investigations.