Apoptosis is the most predominant form of genetically controlled physiological cell death serving to coordinate the elimination of excess, hazardous or damaged cells. Severe homeostatic disturbances of any particular cell population or lineage can cause major pathologies in multicellular organisms. Furthermore, alterations in the regulatory mechanisms of cell death/survival contribute to the pathogenesis of many human diseases, including cancer and neurodegenerative diseases, thus highlighting the importance of maintaining tight regulation of the apoptotic machinery.
Two major pathways of apoptosis (extrinsic and intrinsic) have been delineated. The extrinsic pathway is triggered upon ligand-dependant activation of Death Receptors, such as Fas/CD95/APO-1. The intrinsic pathway involves changes in mitochondrial integrity that leads to the release of pro-apoptotic molecules (e.g. Cytochrome c) from the mitochondrial outer membrane into the cytoplasm. Cytochrome c release induces oligomerization of the adaptor protein APAF-1 and the assembly of a large protein complex, the apoptosome, which promotes the recruitment and activation of apoptosis-specific proteases, in particular Caspase-9 (Casp-9) and Casp-3.
Several apoptotic stimuli, including cancer treatment involving radiation and chemotherapy, ultimately result in activation of the intrinsic mitochondrial pathway. Mutational events occurring "downstream" of Cytochrome c release, which inhibit mechanisms of apoptosis, would function to protect tumor cells from cell death, regardless of exposure to various apoptotic stimuli. Therefore, it is important to identify and characterize additional proteins that either inhibit or enhance Casp-9 activation in the apoptosome.
We developed a functional yeast survival assay in order to isolate mammalian proteins that inhibit cell death at the level of the apoptosome. We are presently investigating the molecular mechanism of this inhibition for several candidate proteins isolated from previous screens, and we are interested in their physiological functions. In addition, we are analyzing the tumor suppressor function of a novel pro-apoptotic APAF-1 binding protein, which has been termed CABY/FAM96A.
We always encourage motivated students to inquire about open positions and practical courses.
In a functional survival screen with CED-4 as an inducer of yeast cell death, a cDNA library cloned from several pooled human mamma carcinoma was used to identify new anti-apoptotic oncogenes. Several candidate oncogenes were identified including a cDNA coding for a C-terminal fragment of FUSE Binding Protein 1 (FUBP1). FUBP1 has previously been described as an important transcriptional regulator of the proto-oncogene c-myc (1).
As FUBP1 activates expression of c-myc while protecting cells from apoptosis, we hypothesized that FUBP1 itself might be involved in tumorigenesis. Therefore, in collaboration with Prof. Dr. Peter Lichter (DKFZ, Heidelberg), we have analyzed the expression of FUBP1 in various tumor entities by immunohistochemistry. We found that hepatocellular carcinoma (HCC) samples exhibited high levels of FUBP1 in more than 80% of the analyzed tumor samples.
To confirm that FUBP1 is important for HCC formation, we analyzed tumor growth by utilizing the HCC cell line Hep3B in a mouse xenograft transplantation model. In this experiment, FUBP1 expression was stably downregulated by transducing Hep3B cells with a lentivirus containing FUBP1 shRNA. Transduced cells were subcutaneously injected into the flanks of immunocompromised mice, and tumor expansion was compared to that observed in mice that received control virus-transduced cells. We found that knockdown of FUBP1 strongly reduced tumor growth in this system, confirming the importance of FUBP1 expression in hepatocellular carcinoma maintenance and/or formation.
Closer investigation of HCC FUBP1 knockdown cells in cell culture revealed a specific phenotype characterized by increased sensitivity to apoptotic stimuli as well as a reduction in cell proliferation. To further elucidate the mechanism by which FUBP1 regulates these cellular processes, we investigated the expression of selected cell cycle- and apoptosis-related genes by quantitative PCR. Here, we could show that as a consequence of FUBP1 knockdown, cell cycle inhibitors, such as p21 and p15, as well as certain pro-apoptotic genes were upregulated. In collaboration with Dr. David Levens (NCI, Bethesda), we demonstrated that the p21 promoter is directly regulated by FUBP1.
Both, the inhibition of apoptosis and repression of cell cycle inhibitors contribute to the oncogenic potential of FUBP1 in liver tumors. We published these results in the journal Hepatology, where our observations highlight the role of FUBP1 in HCC development (2).
To further elucidate the function of FUBP1 in Hepatocellular Carcinoma, we performed microRNA (miRNA) expression analysis following FUBP1 knockdown in Hep3B and HuH7 cells. Several miRNAs were found to be deregulated in the absence of FUBP1 and are now being tested as potential direct FUBP1 targets.
Currently, we are also investigating the potential of FUBP1 as a therapeutic target for HCC treatment. For this purpose, we aim to assess whether FUBP1 inhibition confers protection against malignant growth in established tumors in mouse models of HCC.
Two strategies are being persued in the lab: One approach is the inhibition of FUBP1 by small molecule inhibitors. For this purpose, recombinant human FUBP1 was expressed and purified from E. coli, and 14,440 small molecules of a chemical library as well as 2,500 FDA-approved drugs were screened by using the AlphaScreen technology (collaboration with Dr. Ricardo Biondi, Medizinische Klinik I, University Clinic Frankfurt, and Prof. Eugen Proschak, Biochemistry Department, University of Frankfurt). Several compounds were isolated that were able to prevent or disrupt the binding of FUBP1 to its DNA binding site FUSE. The “hits” were bioinformatically evaluated, and their binding characteristics and inhibitory potentials were confirmed in further in vitro studies like surface plasmon resonance (SPR; collaboration with Prof. Robert Tampé, Biochemistry Department, University of Frankfurt) and microscale thermophoresis (MST) measurements. A subset of these hit compounds were further tested in cell culture experiments with HCC cell lines to evaluate their impact on apoptosis, cell proliferation, cell expansion and FUBP1 target gene expression.
In a second approach, we will target FUBP1 with adeno-associated virus (AAV) constructs containing a FUBP1 shRNA sequence. AAV represents a promising delivery system for gene therapy (3), and following the work scheme presented in Fig. 2, we will apply an HCC-specific AAV virus (that is being developed by our collaboration partner, Prof. Hildegard Büning, Zentrum für Molekulare Medizin, Köln) to downregulate FUBP1 specifically in HCC tumor cells. An endogenous HCC mouse model will be used to monitor the efficacy of our AAV therapy.
Fig. 2: Working scheme for the planned AAV-based therapy.
A. Vector maps of the cloned plasmids (psc hU6 FUBP1 and psc Alb e/p FUBP1), which will be packaged into AAV2 serotype and tested in cell culture.
B. Development of a HCC-specific AAV serotype which will be used as serotype for the inhibition of FUBP1 in an endogenous HCC mouse model (C).
Prof. Kinzler’s lab (Ludwig Center for Cancer Genetics and Howard Hughes Medical Institutions, USA) found that “mutations in CIC and FUBP1 contribute to human oligodendroglioma” (4). This study motivated us to investigate the expression levels of FUBP1 in different gliomas. While we detected low expression of FUBP1 in astrocytes, our immunohistochemistry analysis (in collaboration with Prof. Michel Mittelbronn, Edinger Institute, University Hospital Frankfurt, Germany) revealed elevated levels of FUBP1 in glioblastoma, the most common malignant brain tumor in humans. Glioblastomas are usually highly malignant because the cells expand quickly and they are supported by a large network of blood vessels. One of the challenges in the treatment of glioblastoma is the frequency of recurrence, the aggressiveness and the infiltrative behavior of the tumor. In contrast to oligodendroglioma, glioblastoma express significant FUBP1 levels.
The glioblastoma cell lines LNT229, U87 and U373 are being used to explore the potential oncogenic function of FUBP1 in glioblastoma. Several functional studies revealed less proliferation and enhanced apoptosis sensitivity in LNT229 cells upon knockdown of FUBP1 mRNA, a result in line with our HCC data. In contrast to LNT229, the U87 and U373 glioblastoma cell lines showed more proliferation and increased apoptosis resistance upon FUBP1 knockdown. When we performed xenograft experiments and injected NOD/SCID mice with the glioblastoma cell lines, FUBP1-deficient U87 cells displayed enhanced xenograft tumor growth compared to parental U87 cells. In contrast, LNT229 cells lacking FUBP1 showed a delayed tumor expansion compared to LNT229 cells with wildtype FUBP1 levels (Fig. 3). The decreased tumorigenicity of FUBP1 knockdown LNT229 cells and the increased tumorigenicity of FUBP1-deficient U87 cells correspond to the data obtained in the above-described cell culture experiments.
Fig. 3: Differential role of FUBP1 in glioblastoma cell lines.
LNT229 and U87 cells were stably transduced with sh-FUBP1 lentivirus or control shRNA constructs. 5x106 cells were injected subcutaneously into the flanks of immunodeficient NOD/SCID mice. A. Tumor volumes were measured over a period of 8 weeks using calipers (n = 8 for each cell type). Tumors arising from FUBP1-deficient LNT229 cells expanded significantly slower than control tumors. B. Tumors arising from U87 cells lacking FUBP1 expanded significantly faster than control tumors (n = 8 for each cell type).
To explore the differential role of FUBP1 in different glioblastoma cell lines, we performed Affymetrix array studies in collaboration with Dr. Radlwimmer at the DKFZ in Heidelberg. The results revealed significant upregulation of pro-apoptotic genes in FUBP1 knockdown LNT229 cells in contrast to FUBP1 knockdown U87 and U373 cells in which pro-proliferative and angiogenic genes were significantly upregulated. To understand the role of FUBP1 in different glioblastoma cell lines and in different gliomas is the focus of our further investigation.
To study the physiological role of FUBP1, we established an Fubp1 gene trap mouse model (5) in collaboration with Dr. Franz Vauti (TU Braunschweig; Fig. 4).
Fig. 4: Gene trap vector insertion in mouse ES cells results in expression of an FUBP1-b-geo fusion protein.
In the Fubp1 gene trap mouse model, a ß-geo gene trap vector is inserted into the Fubp1 locus. This leads to expression of a fusion protein consisting of a truncated FUBP1 protein lacking the C-terminal 62 amino acid residues, together with ß-Galactosidase and Neomycin Phosphotransferase.
In the homozygous gene trap animals, the functional impairment of FUBP1 led to embryonic lethality at around day E15.5. FUBP1-deficient mice displayed growth retardation and an anemic phenotype. Further analyses revealed a significant reduction in total fetal liver cell numbers (Fig. 5). As the fetal liver is the main site of embryonic hematopoiesis at this time point of development, we are currently investigating the role of FUBP1 in the hematopoietic system. Competitive transplantation of E15.5 fetal liver cells into lethally irradiated mice resulted in a significantly lower engraftment of cells derived from homozygous Fubp1 gene trap mice compared to cells from wild type and heterozygous animals. Ex vivo gene knockdown studies with adult hematopoietic stem cells (HSCs) indicated a pro-proliferative and anti-apoptotic function for FUBP1, establishing the protein as an important regulator of HSC self renewal.
Fig. 5: Inactivation of FUBP1 leads to embryonic lethality and reduced liver cell numbers.
A. Homozygous Fubp1-/- gene trap mice display growth retardation and an anemic phenotype.
B. The reduced total fetal cell number at embryonic day E15.5 indicates a hematopoietic problem in the absence of FUBP1 (Mean±SD).
FUBP1 is a known transcriptional regulator of the proto-oncogene c-myc and the tumor suppressor gene p21. For a better understanding of how FUBP1 is implicated in the transcriptional network that regulates the development and the homeostasis of the hematopoietic system, ongoing projects in our group aim at the identification of novel FUBP1 target genes as well as transcriptional regulators that are involved in the upstream signaling of FUBP1. Transcriptome profiling of hematopoietic Fubp1 knockdown cells revealed potential FUBP1 target genes that might account for its pro-proliferative and anti-apoptotic function. In silico analyses of the Fubp1 promoter region predicted putative binding sites for several transcription factors with important functions in hematopoiesis, some of which could already be approved for their binding to the Fubp1 promoter.
1 Brooks TA, Hurley LH (2009). The role of supercoiling in transcriptional control of MYC and its importance in molecular therapeutics. Nat Rev Cancer 9(12): 849-61.
2 Rabenhorst U, Beinoraviciute-Kellner R, Brezniceanu M-L, Joos S, Devens F, Lichter P, Rieker RJ, Trojan J, Chung H-J, Levens DL, Zörnig M (2009). Overexpression of the Far Upstream Element Binding Protein FUBP1 in Hepatocellular Carcinoma is required for Tumor Growth. Hepatology 50(4), 1121-9.
3 Bartel MA, Weinstein JR, Schaffer DV (2012) Directed evolution of novel adeno-associated viruses for therapeutic gene delivery. Gene Ther 19: 694-700.
4 Bettegowda C, Agrawal N, Jiao Y, Sausen M, Wood LD, et al. (2011) Mutations in CIC and FUBP1 contribute to human oligodendroglioma. Science 333: 1453-1455.
5 Stanford WL, Cohn JB, Cordes SP (2001). Gene-trap mutagenesis: past, present and beyond. Nat Rev Genet. 10, 756-68.
The conventional cancer treatments, chemotherapy and radiation, primarily cause mitochondrial membrane disruption and activation of the mitochondrial (intrinsic) apoptosis pathway, thereby triggering cell death in tumor cells. Consequently, defects in the mitochondrial apoptosis pathway contribute to carcinogenesis and also provide the molecular basis for resistance against standard cancer therapies (1, 2). Therefore, the identification of new anti-apoptotic target proteins to overcome cancer therapy resistance and to restore the mitochondrial apoptosis pathway in tumors has become an important clinical issue.
To identify novel anti-apoptotic oncoproteins, which are able to inhibit the apoptotic pathway downstream of Cytochrome c release from the mitochondria, we performed a functional yeast survival screen using a human breast tumor cDNA library. One of the isolated cDNAs encoded for the C-terminal half of the AVEN protein (?N-AVEN). During our initial characterization of this artificial ?N-AVEN, the group of Prof. J.M. Hardwick (Johns Hopkins University, Baltimore, USA) published the initial paper describing AVEN as an anti-apoptotic protein binding to BCL-xL and APAF-1 (3). Our own cell death assays in the human colon carcinoma cell line RKO revealed an anti-apoptotic behavior of ?N-AVEN, but not of full length AVEN. These results could be confirmed in Casp-3 activity assays with transfected HEK 293T cells, suggesting that full-length AVEN needs to become activated in vivo, e.g. by proteolytical cleavage that generates an active anti-apoptotic protein product.
To test whether an endogenous molecular correlate for the artificial ?N-AVEN mutant exists in the cell, single- and double-tagged full-length AVEN constructs were cloned. Western Blot analysis of overexpressed and endogenous AVEN protein could indeed demonstrate the presence of smaller AVEN immunoreactive bands. Western Blot analysis performed with MCF-7 breast adenocarcinoma cells revealed an endogenous 30 kDa C-terminal AVEN fragment corresponding in size to our artificial ?N-AVEN. In subsequent studies, we identified Cathepsin D as the protease responsible for the N-terminal AVEN cleavage and activation (4).
To investigate the role of AVEN in the healthy cell/organism, we generated an Aven knockout/knockin mouse model. The coding sequence within the first exon of the wild-type Aven locus was replaced by an in frame GFP cDNA which is terminated by a stop codon and a polyadenylation signal. This animal model allows us to study both the consequences of AVEN deficiency and the activity of the endogenous Aven promoter in various organs and tissues by flow cytometry analysis of GFP fluorescence.
Homozygous deletion of Aven leads to embryonic lethality, and Aven-deficient embryos show significantly decreased organ and tissue sizes at all embryonic stages analyzed (Fig. 1). The few surviving Aven-/- knockout mice are smaller and weigh less at birth than their wild-type counterparts. Eventually, these animals catch up in weight and appear normal. We are currently trying to pinpoint the exact reason for the observed embryonic lethality in most of the homozygous Aven knockout mice. Preliminary observations support the published functions of AVEN in apoptosis protection and DNA damage repair (5).
Fig. 1: Aven deficiency leads to embryonic lethality in mice.
Right panel: Wild type E18 embryo compared with an Aven-/- knockout littermate embryo.
Left panel: Statistic analysis of viable wildtype, heterozygous and homozygous Aven knockout mice. Less than 10% of the expected number of Aven-/- animals (C57Bl/6 background) were born and viable, indicating that homozygous knockout of Aven leads to embryonic lethality.
Published data suggest a correlation between childhood ALL and AVEN mRNA expression levels, implying an oncogenic role for AVEN (6,7). We therefore established an lck AVEN-transgenic mouse line in which the full-length human AVEN cDNA is specifically expressed in T cells. lck AVEN-transgene expression alone is not sufficient for leukemic development in our mouse model. However, analysis of Ick AVEN transgenic mice on a p53+/- heterozygous background demonstrated that AVEN overexpression cooperates with the loss of the tumor suppressor p53 in tumor development. Furthermore, we could show that depletion of AVEN in T-ALL and AML cell lines using lentivirus-mediated shRNA knockdown leads to significantly reduced tumor growth in NOD/SCID mouse xenograft models (Fig. 2) (8). In an ongoing study, we aim to elucidate a potential oncogenic role of AVEN in breast cancer development and progression.
Fig. 2: AVEN overexpression is causally involved with onset of leukemia (T-ALL) and malignant progression.
A. AVEN is overexpressed in T-ALL malignant tissues compared to healthy lymph nodes.
B. T-cell-specific AVEN overexpression in mice leads to accelerated onset of leukemia in cooperation with loss of p53 heterozygocity.
C and D. Lentiviral knockdown of AVEN expression in T-ALL (MOLT-4) and AML (Kasumi-1) cell lines significantly decreases tumor growth in murine xenograft models.
1 Debatin, KM. (2004). Apoptotis pathways in cancer and cancer therapy. Cancer Immunol Immunother 53, 153-159.
2 Hajra, KM. and Liu, JR. (2004). Apoptosome dysfunction in human cancer. Apoptosis 9, 691-704.
3 Chau, B. N., Cheng, E. H., Kerr, D. A., and Hardwick, J. M. (2000). Aven, a novel inhibitor of caspase activation, binds Bcl-xL and Apaf-1. Mol Cell 6, 31-40.
4. Melzer IM, Fernandez SB, Bosser S, Lohrig K, Lewandrowski U, et al. (2012) The Apaf-1-binding protein Aven is cleaved by Cathepsin D to unleash its anti-apoptotic potential. Cell Death Differ 19: 1435-1445.
5. Guo JY, Yamada A, Kajino T, Wu JQ, Tang W, et al. (2008) Aven-dependent activation of ATM following DNA damage. Curr Biol 18: 933-942.
6. Choi J, Hwang YK, Sung KW, Kim DH, Yoo KH, et al. (2006) Aven overexpression: association with poor prognosis in childhood acute lymphoblastic leukemia. Leuk Res 30: 1019-1025.
7. Paydas S, Tanriverdi K, Yavuz S, Disel U, Sahin B, et al. (2003) Survivin and aven: two distinct antiapoptotic signals in acute leukemias. Ann Oncol 14: 1045-1050.
8. Eissmann M, Melzer IM, Fernandez SB, Michel G, Hrabe de Angelis M, et al. (2012) Overexpression of the anti-apoptotic protein AVEN contributes to increased malignancy in hematopoietic neoplasms. Oncogene.
The mitochondrial (intrinsic) pathway of apoptosis is triggered by various cellular stress signals, such as DNA damage. Its activation is critical for the successful treatment of cancer by chemotherapy and radiation. The intrinsic signaling cascade involves the release of the respiratory chain protein Cytochrome c (Cyt c) from the mitochondrial inner membrane space into the cytosol. Cyt c interacts with apoptotic protease activating factor-1 (APAF-1) which upon simultaneous binding of dATP undergoes drastic conformational changes and forms a heptameric complex. Subsequently, pro-Casp-9 is recruited to this so-called "apoptosome" complex and forms homodimers that activate each other due to their close proximity (1). CASP-9 further activates downstream effector caspases that promote apoptosis of the target cell.
We performed yeast-two-hybrid screenings to identify proteins that bind to APAF-1 and that are involved in the regulation of the intrinsic apoptosis pathway. A 21 kDa protein was discovered that co-localizes and co-precipitates with APAF-1 (see Fig. 1). We named this novel protein CABY (for Ced4-like APAF-1 binding protein discovered in yeast two hybrid), which is now officially termed FAM96A. The protein is highly conserved among species, most likely plays an important role in cytosolic iron-sulfur cluster biogenesis, and its overexpression leads to increased sensitivity to stimuli of the mitochondrial apoptosis pathway. Complementary, knockdown of FAM96A in murine fibroblasts and human colon carcinoma cells inhibited cell killing following UV radiation.
Comparative genomic hybridization (CGH) data reveal loss of the FAM96A locus in gastrointestinal stromal tumors (GISTs), and we demonstrated diminished FAM96A expression in GIST biopsies. The tumor-suppressive potential of FAM96A was demonstrated in the GIST cell line GIST882 (Fig. 2), and we could show that the protein is lost during malignant transformation of mouse GIST stem cells. Loss of FAM96A therefore resembles a marker for GIST tumorigenesis, and preliminary data suggest a broad tumor-suppressive potential of the protein.
Fig. 2: FAM96A functions as a tumor suppressor in GIST.
Left panel: Tumor growth was significantly impaired in FAM96A-overexpressing GIST882 cells (hFAM96A) compared with empty vector-transduced controls (LeGoiG2) in a subcutaneous xenograft experiment. Nine NOD/SCID mice per group were injected into the right flank with GIST882 cells expressing low or elevated FAM96A protein levels; tumor growth was significantly reduced in the hFAM96A group compared with the control. The data are presented as the mean ± SEM. Right panel: Tumor weight of hFAM96A (n=5) and control tumors (n=6) at day 83 post-injection; *p ? 0.05; unpaired student´s t-test. Data are presented as the mean ± SEM.
1. Boatright KM, Renatus M, Scott FL, Sperandio S, Shin H, et al. (2003) A unified model for apical caspase activation. Mol Cell 11: 529-541.