CD99 Engagement and Apoptosis in Ewing’s Sarcoma: The Role of Mitochondria and Mitochondrial DNA Mutations

By Michela Rugolo, PhD and Katia Scotlandi, PhD

Introduction

CD99 is a 32kDa glycosylated transmembrane protein encoded by the MIC2 gene, highly expressed in all Ewing’s sarcomas (EWS). CD99 has been involved in many physiological processes, such as cell adhesion, apoptosis and differentiation and intracellular trafficking. Under pathological conditions, CD99 is able to affect cell death, migration, metastasis and differentiation. Noticeably, reduced expression of CD99 is sufficient to abrogate oncogenic transformation and to commit EWS cells toward neural differentiation.¹ Anti-CD99 monoclonal antibodies induce massive apoptosis and reduce malignant potential of EWS cells, through a mechanism whose molecular features are still elusive.² Most recently expression of CD99 was found to modulate migration and metastatic ability of EWS.¹ In addition, the engagement of CD99 increases sensitivity to conventional chemotherapeutic agents and induces tumor regression and significant delay and reduction in metastatic potential.³

Although CD99 engagement induces massive cell death in EWS cells, there is no consensus on the molecular mechanisms involved. In particular, mitochondria have been involved in CD99-induced cell death, given that a reduction of mitochondrial membrane potential (Dy_m) was observed 2 hours after CD99 engagement in EWS cells.⁴ It is well known that mitochondria are key players in the activation of the apoptotic process, being specific target of many Bcl-2 family members and also enclosing within the outer membrane several “apoptogenic factors”, among them cytochrome c, that once released from mitochondria into the cytoplasm, drive the assembly of the apoptosome, which in turn triggers activation of caspase-9 and downstream executioner caspases in the “classical” caspase-dependent apoptosis.⁵ According to Sohn et al, 1998, caspases are activated during CD99-stimulated cell death, whereas we and others showed that the process is completely caspase-independent.^2,6

The first aim of this study has been therefore to disclose the molecular mechanisms underlying CD99 induced cell death, namely to evaluate the involvement of mitochondria and then to identify other crucial players of the process.

Another important aspect we have afforded in this study has been the evaluation of the role of mitochondrial DNA (mtDNA) mutations in EWS.

Indeed, it has been previously reported that the mtDNA D-loop region exhibits a high frequency of somatic mutations in patients with EWS, suggesting that mtDNA variations may play a role in its development.⁸ However, only the D-loop fragment was analyzed in this study, adding little to the knowledge that D-loop instability is a common somatic event in cancer. We have previously reported that mtDNA mutations affecting complex I subunits and destroying both its activity and assembly, make cells resistant to apoptosis.⁹ Moreover, other studies indicate that homoplasmic (100% mutant load) truncating mtDNA mutation confers a less pronounced resistance to oxidative stress-induced cell death compared with heteroplasmic state.¹⁰ Hence, a link between the oxidative phosphorylation impairment caused by mtDNA mutations and apoptosis may be envisaged, since the mitochondrion is central to both processes. We have therefore sequenced the whole mitochondrial genome to reveal the presence of mutations in EWS cells and tumors, to attempt a correlation between mtDNA genetic lesions, mutation load, and possibly apoptosis resistance and tumorigenicity.

Results

Cell death triggered by CD99 does not involve mitochondria and caspases.

The effect of 0662 antibody is very fast, being annexin exposure, a typical hallmark of cell death, already detected after 5min. At the very same time, the levels of the pro-apoptotic Bcl-2 family protein Bax and of phosphorylated p53 were markedly increased. We also observed that increased mRNA Bax levels were associated with decreased amount of the anti-apoptotic protein Bcl-2, in agreement with the opposite effects of these Bcl2-family proteins on the apoptotic cell death. However, the mitochondrial membrane potential (Dy_m) was maintained and decreased only at longer times, after 15min, i. e. subsequently to Bax and p53 increase. It is noteworthy that, despite the massive and rapid increase in Bax levels and subsequent Dy_m depolarization, the apoptogenic factor cytochrome c was not released into the cytosol even after a prolonged treatment. Accordingly, electron microscopy analysis of 0662 Mab-treated cells failed to show evidence for nuclear chromatin condensation/fragmentation typical of classical apoptosis. Therefore, as illustrated in Figure 2, CD99 triggering induces a rapid cell death pathway in EWS that clearly involves Bax elevation and p53 signaling, but without release of cytochrome c, thus resulting caspase-independent, in agreement with our previous reports.² These findings also indicate that the loss of mitochondrial energetic function is a late event and that mitochondria are not directly involved in the triggering of the death process.

Autophagy-associated cell death is elicited by CD99 engagement.

Autophagy-associated cell death is the most widely studied form of non-apoptotic, caspase-independent death. Autophagy is the process of catabolism of cellular components. It degrades cytosolic contents, including proteins and many types of organelles. During autophagy, organelles and long-lived cargo proteins are encapsulated by a double-membrane structure, known as the autophagosome, that then fuses with lysosomes, generating the autolysosome (see figure 3). The degradation of intracellular components recycles essential biomolecules to provide energy. The different steps of autophagy are mediated/regulated by autophagy-related (Atg) proteins. It is well known that autophagy plays a dual role in cells: it can enhance survival in response to nutrient deprivation, organelle damage and stress, or result in death when sustained for prolonged time.¹¹ This type of cell death occurs without DNA fragmentation, nuclear collapse or cytoplasmic shrinkage and death is the result of excessive engulfment and subsequent self-digestion. Given that autophagy has recently been implicated in cell death mediated by p53-dependent mechanisms, we asked whether CD99 engagement in EWS cells by 0662 Mab might exhibit typical features of autophagy. Combined analysis by electron microscopy, immunofluorescence and western blotting allowed us to clearly demonstrate that CD99 engagement in EWS cells results in a caspase-independent cell death mechanism, with some features of autophagy.

mtDNA sequencing in EWS cell lines

The whole sequencing of mtDNA of 11 cell lines generated at the Rizzoli Institute from EWS and 15 osteosarcoma cell lines was carried out, in order to reveal occurrence of pathogenic mutations.

We detected homoplasmic genetic variants in the MT-ND1, MT-CYTB and MT-COIII genes in one EWS line, which were not predicted per se to be pathogenic and yet an overall haplotype role in determining a bioenergetic impairment may not be completely ruled out until functional studies. Interestingly, a heteroplasmic mutation in MT-ND5 gene was also detected in the U2OS cell line (osteosarcoma). To assess whether these mutations affected the mitochondrial energetic function, we measured the ability of these cells to grow in a glucose-free medium containing galactose, i.e. where the rate of glycolysis is strongly reduced and cells are forced to rely on oxidative phosphorylation only for energy production. Under these conditions, cells with defective oxidative phosphorylation are unable to survive in galactose medium. We observed that both the two cells lines bearing the mtDNA mutations did grow nicely under these conditions. Furthermore, the efficiency of oxidative phosphorylation was not significantly different with control cell lines. There results indicate that the genetic variants in mtDNA identified in the two cell lines did not cause a functional dysfunction. The surely pathogenic mutation in U2OS did not have any effect on OXPHOS efficiency likely because of its heteroplasmic status. It has been in fact shown that disruptive mutations in complex I such as this one ought to reach a specific mutation threshold in order to display a pathogenic phenotype, since they ablate the subunit synthesis and may not act as dominant negative. The mutation we detected in U2OS was in fact below the predicted threshold for a phenotypic effect and it remains interesting to know whether such a mutation was also present in the tumour from which the cell line originated, and at what heteroplasmy degree. Unfortunately this was not a feasible task, although future studies may shed light on the occurrence of certain pathogenic variants in osteosarcomas.

mtDNA sequencing in EWS specimens

High frequency of somatic mutations and germline polymorphisms in the mtDNA D-loop region in EWS tumors has been previously reported,⁸ suggesting that these mtDNA variations may play a role in the development of EWS. Moreover, our finding in U2OS cells especially, prompted us to shift to in vivo analysis of the mtDNA genotype. We have therefore sequenced the full length mtDNA of 4 EWS specimens available at the Rizzoli Institute. This number was chosen as the minimum for a preliminary analysis, given that these biopsies are very precious, due to the known difficulties to obtain specimens from these lesions in patients. No pathogenic  mutations were found in these samples, neither in hetero- nor in homoplasmy. Therefore even if mutations could be found in the non-coding D-loop, no evidence for mutations occurring in the coding regions was obtained in EWS. Recently, Yu et al. reported the occurrence of somatic mutations in the coding regions of the mitochondrial genome, with a total number of 11 mutations over 31 patients with osteosarcomas, most of them (55% ) were within the D-loop region.¹² Some of these mutations were predicted to be functionally relevant mutations in tRNA and polypeptide-coding  genes with potential detrimental impacts on the mitochondrial OXPHOS system. However all these mutations were heteroplasmic (ranging between 35-80%), likely below the threshold for dysfunctional phenotype and their pathogenic role in tumor development is questionable. 

In conclusion, the analysis performed both in both cell lines and specimens tends to rule out a significant role for mtDNA mutations in the pathogenetic mechanisms of EWS development.

V10N4 ESUN Copyright © 2013 Liddy Shriver Sarcoma Initiative.

CD99 Engagement and Apoptosis in Ewing’s Sarcoma: The Role of Mitochondria and Mitochondrial DNA Mutations

By Michela Rugolo, PhD and Katia Scotlandi, PhD

Abstract

CD99 is a cell surface transmembrane protein highly expressed in all Ewing’s sarcomas (EWS). Engagement by anti-CD99 monoclonal antibodies induces massive apoptosis and reduces malignant potential of EWS cells, through a caspase-independent mechanism whose molecular features are still elusive.¹ Most recently expression of CD99 was found to modulate migration and metastatic ability of EWS.² In addition, the engagement of CD99 increases sensitivity to conventional chemotherapeutic agents and induces tumor regression and significant delay and reduction in metastatization.³ As a consequence understanding the molecular mechanisms underlying CD99 induced cell death is mandatory, since it may reveal new molecular therapeutic targets of general interest. In addition a better understanding of the apoptotic mechanisms may be helpful for intensifying effects of conventional therapy limiting high doses of chemotherapy and sparing toxicity.

Although CD99 engagement induces massive apoptotic cell death in EWS cells, there is no consensus on the molecular mechanisms involved.^1,4,5

Preliminary Results

We observed that the cell death process is very rapid, being annexin V staining already apparent after 5 min incubation of cells with the anti-CD99 antibody. At the same time, a significant increase in the intracellular level of the pro-apoptotic Bcl-2 family protein Bax was detected, suggesting that mitochondria may be involved in CD99-induced cell death. These organelle are known to be key players in the activation of apoptosis, being specific targets of many Bcl-2 family members and also enclosing within the outer membrane several "apoptogenic factors."⁶

Among them, cytochrome c can be released from mitochondria into the cytoplasm driving the assembly of the apoptosome, which in turn triggers activation of caspase-9 and downstream executioner caspases in the "classical" caspase-dependent apoptosis. Apoptosis inducing factor (AIF) and Omi/HtrA2 are other mitochondrial apoptogenic factors involved in caspase-independent apoptotic pathways.^7,8 Given that caspases seem not to be primarily involved in CD99-triggered cell death, further studies are clearly required to identify the apoptogenic factors involved in the process.

On the other hand, other intracellular events can regulate the ability of CD99 engagement to induce apoptosis, and the strictly associated reduced malignant potential, whose identification may help to disclose novel mechanisms potentially useful for therapy. In this regard, mitochondrial DNA (mtDNA) mutations have been shown to protect cells from apoptosis, thus promoting tumor growth.⁹

Furthermore, the cell fate was demonstrated to depend on the type of mitochondrial dysfunction, i.e. abolition of electron flux confers resistance to apoptosis, whereas a partial reduction makes cells more prone to cell death.¹⁰ Accordingly, we previously reported that missense mutations in ND subunits of complex I affecting the redox activity but not complex I assembly, predispose cells to death,¹¹ whereas our preliminary experiments clearly indicate that truncative mutations affecting both complex I activity and assembly, make cells resistant to apoptosis (Porcelli et al., submitted). Moreover, other studies indicate that homoplasmy (100% mutant load) of a truncating mtDNA mutation confers a less pronounced resistance to oxidative stress-induced cell death compared with heteroplasmic state.¹² Hence, a link between the oxidative phosphorylation impairment caused by mtDNA mutations and apoptosis may be envisaged, since the mitochondrion is central to both processes. An increasing number of clearly pathogenic mtDNA mutations is being detected in a variety of cancers.^13,14,15

These mutations are often homoplasmic, suggesting the existence of a positive selection of mutant mtDNA copies during tumor progression, a process that has yet to find explanation. Interestingly, Yu et al., 2009, reported that the mtDNA D-loop region exhibits a high frequency of somatic mutations in patients with EWS, suggesting that mtDNA variations may play a role in development of EWS.¹⁶ In this study, only the D-loop fragment was analyzed, adding little to the knowledge that D-loop instability is a common somatic event in cancer. This issue deserves further studies focused on whole mitochondrial genome sequencing, heteroplasmy mutations level evaluation and assessment of their pathogenic potential, to better understand the functional implications of mtDNA instability in EWS.

Purposes

Two are the main aims of the present project:

1. To investigate the molecular pathways triggered by the anti-CD99 0662 antibody leading to cell death, in particular the involvement of mitochondria
2. To sequence the whole mitochondrial genome to reveal the presence of mutations in EWS cells and tumors, to attempt a correlation between mtDNA genetic lesions, mutation load, apoptosis resistance and tumorigenicity

Research Plan and Experimental Design

1. Assessment of mitochondrial involvement in anti-CD99 antibody triggered apoptosis

To define whether early activation of apoptotic cell death by monoclonal 0662 antibody involves mitochondria, we will evaluate ΔΨm, that provides the most direct indication of the energetic status of these organelles in a panel of 22 EWS cell lines, including patient-derived cultures and cell lines widely used in the different labs. Sohn et al. previously reported a significant ΔΨm reduction after 2 hours treatment with the anti-CD99 antibody.⁴ We will utilize tetramethylrhodamine (TMRM), a fluorescent cationic probe that accumulates in polarized mitochondria and is released when ΔΨm decreases.¹⁷ The agonistic antibody 0662 will be added to cells loaded with TMRM and the fluorescence changes will be determined by a fluorescence microscope as a function of time. If the fluorescence signal decreases, it means that ΔΨm is reduced by CD99 engagement. Given that pre-treatment with cyclosporine A (CsA) was reported to protect RD-ES cells from CD99-induced apoptosis, possibly through inhibition of mitochondrial permeability transition pore opening,⁴ we will directly evaluate the effect of CsA on ΔΨm changes induced by 0662 antibody addition. The molecular mechanism underlying PTP opening will also be analyzed, evaluating the role of ROS overproduction and calcium deregulation, by preincubation with antioxidants and with the intracellular Ca2+-chelator BAPTA/AM.¹⁷ Finally the effect of 0662 on ΔΨm as a function of CD99 levels will also be determined, using a collection of cell lines already available, exhibiting different CD99 content. Studies will also be performed in normal mesenchymal stem cells, which likely are the cell of origin of EWS.¹⁸

Given that Bax levels have been shown by us to increase within the first minutes of 0662 incubation, we will determine its activation by using a specific antibody for activated Bax, by IF analysis and western blotting. Cleavage of the BH3-only protein Bid will also be tested in cellular lysates by western blotting, given that caspase-8 activation has been recently reported after ligation of D4, the mouse homologue of CD99.¹⁹ Finally to elucidate the mechanism of caspase-independent apoptotic cell death we will measure the release of AIF and Omi/Htr2A, as well as of cytochrome c, from mitochondria to cytosol, as previously described.²⁰ Silencing of AIF and incubation with the Omi/Htr2A inhibitor will be used to verify whether these proteins are required for CD99-activation triggered cell death.

If experiments rule out the involvement of mitochondria in CD99-induced apoptosis, we will investigate the role of lysosomal membranes permeabilization (LMP), resulting in the release of lysosomal proteases into the cytosol, by using specific inhibitors of the different cathepsins. In fact it is well known that these lysosomal proteases can also mediate a caspase-independent cell death pathway. It is noteworthy that LMP has been reported to trigger mitochondrial membrane permeabilization through Bax activation. In this model system Bax activation relies on reactive oxygen species overproduction.²¹

2. mtDNA sequencing of EWS tumors and cell lines

Mitochondrial genome sequencing, evaluation of mutations heteroplasmy and of their pathogenic potential will be performed in EWs cell lines and tumors, as previously described.²² Briefly, DNA will be extracted from tumor tissues and cells with commercial kits (Qiagen). The MitoAll re-sequencing kit (Applied Biosystems) will be used to screen for mtDNA mutations upon assembly of the 46 amplicons through the SeqScape software optimized to this purpose (Applied Biosystems). A thorough analysis of the entire mtDNA sequence obtained will be performed in terms of reconstruction of the mitochondrial haplogroup/haplotype, in silico prediction of variants pathogenicity as well as variability data on the rare variants detected, via the implementation of HmtDB algorithms, available to the proponent through a collaboration with the creators of the HmtDB database, as previously described.²³ Heteroplasmy evaluation and quantification will be performed with optimized methods available at the Unit of Medical Genetics such as DHPLC and fluorescent PCR for small insertions and deletions. A panel of 22 cell lines (including all the most commonly used EWS cell lines and primary cell cultures established in our lab) will be screened for the presence of mtDNA mutations by sequencing, as described above. Data will be compared with those obtained in clinical samples (see below). It is noteworthy that preliminary experiments revealed mtDNA mutations in 2 out of the 8 cell lines analyzed so far.

Pivotal analysis will also be performed on 20 biopsy and 20 tumor speciments after preoperative chemotherapy, to verify the presence of mitochondrial mutations in untreated tumors and after exposure to standard chemotherapy treatments. A tissue array for EWS samples including more than 200 cases is already available at the Rizzoli Institute and may be used for complementary studies to mtDNA sequencing. In selected samples, IHC analysis will be carried out using antibodies for both mitochondria- and nucleus-coded subunits of the five oxidative phosphorylation complexes,²³ to gain insight on their assembly status. Moreover, in the very same samples IHC analysis will be performed also for HIF1α. Whenever possible, IHC findings will be correlated with the tumor specific mitochondrial genotype. If encouraging results are obtained we will analyze all the samples available in the Rizzoli biobank to allow associations with patient prognosis.

Impact

Definition of the molecular mechanisms of CD99-induced apoptosis will identify more common therapeutic targets, for which drugs may be developed and become available. In particular, a more precise and detailed knowledge of the apoptotic mechanisms will help combined therapies with targeted drugs and conventional chemotherapy. Furthermore, exploring mDNA mutations will provide novel findings that may be used as predictors of prognosis and help stratifying patients at diagnosis based on risks, also in relationship with HIF-1α stabilization. Functional in vitro studies will validate correlative clinical analysis and provide indications on molecular mechanisms, thus facilitating the design of more efficient and personalized therapies.

V8N2 ESUN Copyright © 2011 Liddy Shriver Sarcoma Initiative.

Grant Funding

The Liddy Shriver Sarcoma Initiative funded this $50,000 research grant in April 2011. The study was made possible by generous donations made in memory of Michael Lio (Strike out Sarcoma), Peter Skelton (Peter Skelton Sarcoma Research Foundation), Ryan Glenny, Paul Roth, Mike Homan, Craig Goris, Richard Pollak, and Joe Rivas, who lost their lives to Ewing's sarcoma; and by generous donations made to honor Jeremy Weingrod and Alex Franke, who are fighting the disease.

The University of Bologna and the Rizzoli Institute published announcements about this grant.