Altered autocrine and paracrine signaling controls drug effects
Intra- and inter tumor heterogeneity are key factors affecting drug efficacy in individual patients. Mechanisms helping tumor cells persist drug treatment require immediate adaptation while long-term drug exposure establishes and fixes resistance states. We research on both, short-time drug effects and on long time resistance development, and there uncover molecular mechanisms underlying tumor cell survival.
Understanding failure of neoadjuvant chemotherapy in TNBC
The tumor microenvironment (TME) or tumor stroma comprises all cell types and extracellular matrix (ECM) that surrounds the tumor cells, jointly forming the tumor mass. The stromal compartment is comprised of immune cells from both, the innate and the adaptive systems, vascular cells, mesenchymal stem cells (MSC), cancer associated fibroblasts (CAF), and several other cell types. The TME affects tumor aggressiveness and the way tumor cells respond to therapies.
Fibroblasts modulate drug effects in breast cancer. (Maia et al. 2021 Mol Onc, 15:1308-29 - https://pubmed.ncbi.nlm.nih.gov/33476079/)
© dkfz.de
We have shown that primary patient-derived cancer associated fibroblasts (Berdiel-Acer et al. 2021 Oncogene, 40:2651-66) support tumor cell recovery from the impact of clinically applied chemotherapeutic drugs in breast cancer model systems (Maia et al. 2021 Mol Onc, 15:1308-29). This process involves interferon beta signaling, an antiviral response, and expression of interferon stimulated genes.
Figure 1: Treatment with high doses of chemotherapeutic agents (CTX) induces the expression of IFNb1 by cancer cells and its subsequent secretion into the tumour microenvironment. Stromal fibroblasts binding to IFNb1 through IFNAR1 get activated and acquire an anti-viral, pro-inflammatory signature associated with the expression of several ISGs. Through this process, stromal fibroblasts are reprogrammed to a pro-tumorigenic state leading to cancer cell recovery after CTX treatment and cancer relapse in patients with breast cancer (Maia et al. 2021 Mol Onc, 15:1308-29).
Expression of the ISG protein OAS1 significantly correlated with residual disease (i.e., non-pCR) in the TNBC subtype of breast cancer after neoadjuvant chemotherapy (NACT), indicating clinical relevance of our findings (Bauer et al. 2022 Cancer Res, 82:P1-08-15).
Mechanisms of endocrine therapy resistance
Using in vitro models, we induced resistance to the ER-modulator tamoxifen as well as to long-term estrogen deprivation to mimic clinical aromatase inhibition, which are applied in pre- and postmenopausal patients, respectively. These models have since been utilized aimed at characterizing driver mechanisms of resistance (Borgoni et al. 2020 Cancers, 12(10):2918) and to understand the involvement of epigenetic mechanisms (Soleimani Dodaran et al. 2020 BMC Cancer, 20:676).
Evolution of primary and secondary drug resistance, captured with barcoded cell line models
© dkfz.de
We currently work to uncover clonal events triggering tumor cell persistence and resistance, and to characterize a candidate protein that could be a link to epigenetic reprogramming upon resistance acquisition.
Figure 2: Tumor progression is driven by clonal evolution. To mirror and assess tumor evolution at the individual clone level, MCF7 and T47D cells were genetically tagged with a barcode library. Five replicates of cells were then kept for eight months in media containing estrogen (+E2) or were treated with either estrogen deprivation (-E2) or 4-hydroxy tamoxifen (4-OHT), yielding long-term estrogen deprived (LTED) and tamoxifen-resistant (TAMR) cell pools, respectively. Barcode sequencing revealed clonal selection of de novo resistant populations as well as adaptation, indicating secondary resistance.
Modulations of EGFR-signaling via growth factors and therapeutic drugs
Receptor tryrosine kinase signaling via the EGF-receptor family of RTKs is a central signaling path also in breast cancer. Amplification of receptors (ERBB2) and mutations in signaling (like PIK3CA, RAS/RAF) are key events in different subtypes. We use targeted therapeutics (i.e., inhibitors and therapeutic antibodies) to better understand the wiring and rewiring in disease conditions. For relative quantification of protein activation states, we employed a targeted proteomics approach using Reverse Phase Protein Array (RPPA) technology (e.g., Sonntag et al. 2014 Transl Prot, 2:52-9, Bernhardt et al. 2017 Breast Cancer Res, 19(1):112, Bernhardt et al. 2019 J Prot Res, 18(3):1352-62; Byron et al. 2020 Sci Rep, 10(1):21985).
Wiring of ERBB downstream signaling in breast cancer cell lines. (Kemmer et al. 2022 Cancers, 14(10):2379)
© dkfz.de
With funding by the BMBF (e:Med) we performed time-course analysis of cellular responses to combinations of different activators and inhibitors of the EGFR signaling network and developed a universal mathematical model that can predict the effects growth factors and inhibitors have in breast cancer subtypes.
Figure 3: Wiring of ERBB and downstream signaling is mostly regulated by receptor levels, mutation states and growth factors in breast cancer subtypes. Top: ERBB receptors and their interactions with specific ligands and antagonistic inhibitors investigated in the study. Bottom: Dominating signaling fluxes in three breast cancer cell lines without any ligand or drug treatment, as supported by the data. PI3K signaling forms the main signaling branch in MCF7, T47D and SKBR3 (highlighted in green), while MAPK signaling dominates in MDA-MB-231 cells, including a negative feedback loop on c-RAF and RAS (highlighted in blue). Asterisks indicate activating mutations in the respective components (note: PIK3CA, component of the the PI3K complex, is mutated in MCF7 and T47D). (Kemmer et al. 2022 Cancers, 14(10):2379)
In another time-course study, we found that glutamate ammonia ligase (GLUL) expression was negatively affected by hypoxia, and that this was associated with aggressive phenotypes in breast cancer in vitro, in vivo, and in patients (Bernhardt et al. 2019 J Proteome Res, 18(3):1352-62).
miRNAs and isomiRs are non-coding determinants of cancer biology
The division has a long history for making discoveries in the field of miRNA-tumor interactions (e.g., Uhlmann et al., 2010 Oncogene, 30:4297-306, Uhlmann et al., 2012 Mol Syst Biol, 8:570, Körner et al., 2013 JBC, 288:8750-61; Keklikoglou et al., 2015 Oncogene, 34:4867-78; Breunig et al., 2018 Mol Oncol, 12(8):1447-63).
The singnaling network regulated by miR-193b. (Giacomelli et al, 2021, BMC Cancer, 21(1):1296)
© dkfz.de
More recently, we determined previously unknown roles of miRNAs using a targeted proteomic screening approach and identified miR-193b to coordinately regulate Wnt/b-catenin, c-Met, and integrin signaling in aggressive triple-negative breast cancer (Giacomelli et al, 2021, BMC Cancer, 21(1):1296).
Figure 4: miR-193 is an oncogenic miRNA regulating Wnt/b-catenin, c-MET, and integrin signaling pathways. KEGG maps with downstream phenotypes with miRNA effects indicated. Target proteins probed in the study are shaded in lilac or pale yellow when they are repressors or activators of the pathways, respectively. miR-193b repressive or activating effect on the pathways is represented by a box around the proteins significantly regulated, of red or green colour, respectively. The chemical inhibitors' activities are highlighted in blue (iCRT-14), purple (Capmatinib), and green (Erlotinib). (Giacomelli et al, 2021, BMC Cancer, 21(1):1296)
Expanding the scope of our studies also to the role of 5'isomiRs, we uncovered a negative feedback loop between a specific 5'isomiR of miR-183-5p and E2F1 (Li et al, 2022, J Exp Clin Cancer Res, 41(1):190). Clinical relevance of these findings was established with the help of data from The Cancer Genome Atlas (TCGA). Along these lines, a batch-correction strategy for reliable analysis of that data was developed (Ibing et al, 2021, NAR Cancer, 3(1):zcab007). Within a DFG-funded project, we currently investigate the mechanisms and the functional relevance of aberrant 5'isomiR processing in cancer.
Collaborations
Signaling: Yosef Yarden (Rehovot), Moshe Oren & Yael Aylon (Rehovot), Pernette Verschure (Amsterdam), Luca Magnani (London), Niels de Jonge, Diana Peckys (Saarbrücken)
Breast cancer organoids: André Koch (Tübingen)
Autophagy: Silvia Vega Rubin de Celis (Essen)
Bioinformatics: Lars Feuerbach, Benedikt Brors (DKFZ), Tim Beissbarth (Göttingen)
Mathematical modeling: Jens Timmer (Freiburg)
Clinics: Martina Vetter, Eva Kantelhardt, Christoph Thomssen (Halle/Saale)
Proteomics: Dominic Helm (DKFZ)