Final Report Summary - METABBC (Metabolic switch during mutant-PI3K-induced breast tumorigenesis and metastasis)
Each year 1.1 million new cases of breast cancer occur among women worldwide and ~50% sucumb to metastatic disease. Although the biology of tumor initiation, progression and metastasis has been intensively studied, the cellular/biochemical mechanisms orchestrating its steps remain largely elusive. During tumorigenesis and metastatic progression, tumor cells have to adapt to stringent microenvironments by switching on and off specific metabolic pathways (metabolic rewiring). Such metabolic rewiring can be controlled by several oncogenic pathways, such as the PI3K/AKT pathway, which is frequently activated in breast cancer. Data generated from our and other labs suggest that transformation and metastasis in these models is accompanied by metabolic changes, which present a unique window for therapeutic targeting. In this proposal, I identified novel metabolic enzymes that can regulate fundamental processes of breast cancer biology: from tumor initiation and cancer stem cell-like activity to metastasis. Furthermore, I explored the role, molecular details and confirmed at least one of these metabolic regulators as a powerful marker for poor overall survival in patients with breast cancer.
In this project, three strategies were used:
1) Analysis of gene expression databases to identify mutant PIK3CA -associated metabolic genes which are highly expressed in normal and neoplastic breast cell populations with stem-cell activity.
2) An inducible pooled shRNA screen to identify additional regulators of self-renewal and identified two candidate enzymes whose expression is lost in cells with high self-renewal ability (tumor suppressor activity).
3) Through a collaboration with an external lab with expertise in metabolomics analysis, I characterized the metabolome of primary breast tumors and respective organotropic metastasis using MS/LC methods and identified a metabolic enzyme that is upregulated and highly active in metastases compared with primary tumors, which I am currently validating.
To pursue Aim 1, we pre-selected a list of metabolic regulators that we found to be highly expressed in mutant PI3K breast cancer and overlapped it with datasets of genes whose expression is particularly high in normal breast cell populations enriched for stem cell activity, as well as in breast cancer cells with self-renewal activity. By overlapping our mutant PI3K gene list with three different datasets, we selected one candidate gene, which I later found to be specifically expressed in basal breast cancer and absent in luminal, estrogen receptor alpha (ERa) positive breast cancer. I used loss and gain of function approaches to dissect the role of this candidate. Accordingly, I knockdown its expression in basal human breast cancer cell lines. Interestingly, we found that knockdown of this enzyme leads to a decrease in the cancer stem cell- like population as assessed by ALDH activity, stem cell surface markers and growth in serum-free, ultra low attachment conditions (suspension). Furthermore, we confirmed a decreased tumorsphere forming potential and a concomitant decrease in tumor initiation capability using limiting dilution experiments in vivo. We also verified that the overexpression of this enzyme in luminal breast cancer cells accelerates tumor growth. Interestingly, we noticed that such overexpression caused a dampened response to estrogen treatment in MCF7 cells. We then investigated whether the dowregulation of this enzyme in the basal subtype could cause any change in the expression of cell differentiation markers and that was the case. Not only we noticed a loss in the expression for basal/mesenchymal markers, such as CK14, TGFb and FN1 but we also observed an upregulation of the luminal markers such as CK18 and several ERa- downstream target, such as GATA3, GREB1 and ERa itself. Such effect was also observed in primary normal human breast epithelial cells, where the downregulation of this enzyme in basal cells generated cells with pure luminal differentiation.
Altogether, our data indicate that we may have found a novel metabolic regulator of the stem cell population and of cell differentiation in the normal and neoplastic mammary gland.
We have also studied the molecular mechanisms underlying the alterations observed above and have identified a series of gene expression changes that are likely associated with a role for this enzyme in epigenetic regulation. Since cell fate can be regulated by epigenetic modifications of key linage-determining genes, we will investigate whether epigenetic modifications controlled by this enzyme are involved in the phenotypes observed.
Regarding the second strategy, we custom ordered a pooled doxycycline inducible shRNA library targeting approximately 200 metabolic regulators. This library was used to knockdown the selected metabolic genes in a primary murine mammary cancer cell line generated in our lab harboring the PIK3CA H1047R mutation. These cells were cultured in suspension and serum free conditions for 10 to 13 weeks. Under these conditions, cells form monoclonal spheres that can be propagated over several generations and are enriched in cells with high self-renewal and tumor initiation properties. Along passages, a part of the samples was collected, the DNA extracted, and the bar codes sequenced. shRNAs that target genes important for self-renewal are expected to be depleted, while shRNAs that target genes detrimental for self-renewal are expected to be enriched. We repeated the same experiment three independent times and we identified two isoforms of the same enzyme that are depleted along passaging of the spheres. We are currently validating these results in breast cancer models of PI3K pathway activation.
In the third strategy, we used GFP-labelled MDA-MB-231 cells which we injected orthotopically in immunocompromised mice. After primary tumor resection, mice were allowed to develop metastases. All samples derived from the primary tumor and respective metastases (lungs, liver and lymph nodes) were digested and sorted. We then extracted the metabolites from the cells and submitted them for mass spectrometry- based metabolic analysis. We detected that the metabolic profile of primary tumors was very consistent among different mice. At the metastasis levels, and within the same mouse, different organs had metastatic cells with distinct metabolic profiles, although those differences were more subtle thanthose found when the metastases were compared to the respective primary tumors they originated from. Particular enrichments in the metastases compared with primary tumors were found in arachidonic acid metabolism and steroidogenesis.
Overall, this project had led to the identification of new metabolic biomarkers and/or targets for metastatic breast cancer, underlying the still underestimated importance of metabolism in this particular type of cancer.
In this project, three strategies were used:
1) Analysis of gene expression databases to identify mutant PIK3CA -associated metabolic genes which are highly expressed in normal and neoplastic breast cell populations with stem-cell activity.
2) An inducible pooled shRNA screen to identify additional regulators of self-renewal and identified two candidate enzymes whose expression is lost in cells with high self-renewal ability (tumor suppressor activity).
3) Through a collaboration with an external lab with expertise in metabolomics analysis, I characterized the metabolome of primary breast tumors and respective organotropic metastasis using MS/LC methods and identified a metabolic enzyme that is upregulated and highly active in metastases compared with primary tumors, which I am currently validating.
To pursue Aim 1, we pre-selected a list of metabolic regulators that we found to be highly expressed in mutant PI3K breast cancer and overlapped it with datasets of genes whose expression is particularly high in normal breast cell populations enriched for stem cell activity, as well as in breast cancer cells with self-renewal activity. By overlapping our mutant PI3K gene list with three different datasets, we selected one candidate gene, which I later found to be specifically expressed in basal breast cancer and absent in luminal, estrogen receptor alpha (ERa) positive breast cancer. I used loss and gain of function approaches to dissect the role of this candidate. Accordingly, I knockdown its expression in basal human breast cancer cell lines. Interestingly, we found that knockdown of this enzyme leads to a decrease in the cancer stem cell- like population as assessed by ALDH activity, stem cell surface markers and growth in serum-free, ultra low attachment conditions (suspension). Furthermore, we confirmed a decreased tumorsphere forming potential and a concomitant decrease in tumor initiation capability using limiting dilution experiments in vivo. We also verified that the overexpression of this enzyme in luminal breast cancer cells accelerates tumor growth. Interestingly, we noticed that such overexpression caused a dampened response to estrogen treatment in MCF7 cells. We then investigated whether the dowregulation of this enzyme in the basal subtype could cause any change in the expression of cell differentiation markers and that was the case. Not only we noticed a loss in the expression for basal/mesenchymal markers, such as CK14, TGFb and FN1 but we also observed an upregulation of the luminal markers such as CK18 and several ERa- downstream target, such as GATA3, GREB1 and ERa itself. Such effect was also observed in primary normal human breast epithelial cells, where the downregulation of this enzyme in basal cells generated cells with pure luminal differentiation.
Altogether, our data indicate that we may have found a novel metabolic regulator of the stem cell population and of cell differentiation in the normal and neoplastic mammary gland.
We have also studied the molecular mechanisms underlying the alterations observed above and have identified a series of gene expression changes that are likely associated with a role for this enzyme in epigenetic regulation. Since cell fate can be regulated by epigenetic modifications of key linage-determining genes, we will investigate whether epigenetic modifications controlled by this enzyme are involved in the phenotypes observed.
Regarding the second strategy, we custom ordered a pooled doxycycline inducible shRNA library targeting approximately 200 metabolic regulators. This library was used to knockdown the selected metabolic genes in a primary murine mammary cancer cell line generated in our lab harboring the PIK3CA H1047R mutation. These cells were cultured in suspension and serum free conditions for 10 to 13 weeks. Under these conditions, cells form monoclonal spheres that can be propagated over several generations and are enriched in cells with high self-renewal and tumor initiation properties. Along passages, a part of the samples was collected, the DNA extracted, and the bar codes sequenced. shRNAs that target genes important for self-renewal are expected to be depleted, while shRNAs that target genes detrimental for self-renewal are expected to be enriched. We repeated the same experiment three independent times and we identified two isoforms of the same enzyme that are depleted along passaging of the spheres. We are currently validating these results in breast cancer models of PI3K pathway activation.
In the third strategy, we used GFP-labelled MDA-MB-231 cells which we injected orthotopically in immunocompromised mice. After primary tumor resection, mice were allowed to develop metastases. All samples derived from the primary tumor and respective metastases (lungs, liver and lymph nodes) were digested and sorted. We then extracted the metabolites from the cells and submitted them for mass spectrometry- based metabolic analysis. We detected that the metabolic profile of primary tumors was very consistent among different mice. At the metastasis levels, and within the same mouse, different organs had metastatic cells with distinct metabolic profiles, although those differences were more subtle thanthose found when the metastases were compared to the respective primary tumors they originated from. Particular enrichments in the metastases compared with primary tumors were found in arachidonic acid metabolism and steroidogenesis.
Overall, this project had led to the identification of new metabolic biomarkers and/or targets for metastatic breast cancer, underlying the still underestimated importance of metabolism in this particular type of cancer.