Mitochondria in the Tumourigenic Phenotype
By Bertrand C. Liang
Copyright 2011 Bertrand C. Liang
Smashwords Edition
Smashwords Edition, License Notes
Thank you for downloading this free ebook. Although this is a free book, it remains the copyrighted property of the author, and may not be reproduced, copied and distributed for commercial or non-commercial purposes without permission. Please visit smashwords.com to discover other works by this author, and thank you for your support.
***
Table of Contents
Introduction: the Tumourigenic Phenotype
Discussion: Mitochondria and Cancer
A Model of the Tumourigenic Phenotype
***
Acknowledgments
As is the usual case, the generous support received for this work cannot be adequately conveyed in a short page of gratitude. Mentors, colleagues, and friends, both new and old, in the scientific, clinical and industrial worlds, are far too numerous to list individually, but of whom I hope understand my significant indebtedness to them, as well as my most heartfelt thanks.
My sincere appreciation to Professor Martin Grootveld, Dr. Richard Iles, Dr. Neil Morgan and Dr. Subhash Anand MBE, who took time out of very busy schedules to be part of this work; grateful thanks.
My most significant note of thanks is to my supportive family –Christopher A. Liang, Kate M. A. Liang, Bryan A. Liang, Anita C. Liang and Anna R. Winn – and particularly, my loving and understanding wife, Diane M. Liang - who appreciate not only the value of education, but the rationale of life-long learning, both as example and beneficiary. Finally, to my father, Charles C. Liang, PhD, a victim of cancer at too young an age – a mentor, and in later life, a dear friend; I would have very much liked for him to both understand this work and see this day, for I miss him and his wise scientific insight and counsel very, very much.
***
TO MY FATHER
“It does not matter how slowly you go, so long as you do not stop.”
***
List of Publications
BC Liang. Evidence for association of mitochondrial DNA sequence amplification and nuclear localization in human low grade gliomas. Mutation Research, 354:27-33, 1996.
BC Liang, Hays L. Mitochondrial DNA copy number changes in human gliomas. Cancer Letters, 105:167-173, 1996.
L Cavalli, Varella-Garcia M, Liang BC. Diminshed tumorgenic phenotype achieved by depletion of mitochondrial DNA. Cell Growth and Differentiation, 8:1189-1198, 1997.
BC Liang, Ullyatt E. Chemosensitization of glioblastoma cells to bis-dichloroethyl-nitrosourea with tyrphostin AG17. Clinical Cancer Research, 4:773-781, 1998.
LR Cavalli, Liang BC. Mutagenesis, Tumorigenicity, and Apoptosis: Are the Mitochondria involved? Mutation Research (Fundamental and Molecular Mechanisms of Mutagenesis), 398:19-26, 1998.
BC Liang, Ullyatt E. Increased sensitivity to cis-diamminedichloroplatinum-induced apoptosis with mitochondrial DNA depletion. Cell Death and Differentiation, 5:694-701, 1998.
E Ullyatt, Liang BC. 2’3’-dideoxycytidine is a potent inducer of apoptosis in glioblastoma cells. Anticancer Research, 18:1859-64, 1998.
BC Liang, Miller L, A Weller. Ethyl-nitrosourea transformed astrocytes exhibit mitochondrial membrane hyperpolarization and constrained apoptosis. Apoptosis, 4:89-97, 1999.
***
List of Figures
Figure 1. Sequence of extracted cDNA clone.
Figure 3. Ethidium bromide/acridine orange assay for chromatin condensation.
Figure 5. Effects of oligomycin on U937 ATP levels.
***
Monograph Abstract
Cancer arises from the accumulation of nuclear and cytoplasmic abnormalities, a phenomenon which allows for the expression of the tumourigenic phenotype. The studies documented herein have evaluated the role of mitochondria in this phenotype. An increased copy number of mitochondrial DNA (mtDNA) was observed, along with changes in the expression of cancer cell lines and primary tumour tissue. Such amplification was associated with mtDNA within the nucleus, and in low grade (ontologically-early) tumours. Tumour cell lines were depleted of mtDNA, creating “rho-” cells, with minimal mtDNA/functional mitochondria. While rho- cells retained immortality, there was a loss of anchorage-independence and xenograft growth; moreover, these cells became more sensitive to the effects of cytotoxic agents. When normal mitochondria were replaced into the rho- cells, the cells reverted to the parental phenotype, including amplification of mtDNA. To further evaluate the early onset of mitochondrial alterations, normal human astrocytes were exposed to ethyl-nitrosourea, a process which created transformed cells. Hyperpolarization of the mitochondrial membrane was noted as an early manifestation of transformation, with corresponding decreases in mitochondrial mass. The inhibition of the glycolytic enzyme phosphofructokinase diminished the development of the tumourigenic phenotype in the transformed cells. Tumour cells were treated with mitochondrial DNA-polymerase gamma inhibitor 2’,3’-dideoxycytidine (ddC) and tyrphostin AG17; both compounds demonstrated an ability to effectively inhibit mitochondrial function and thereby either directly, or in combination with other cytotoxic drugs (1,3-bis-dichloroethyl-nitrosourea, BCNU), effect apoptosis. Hence, the mitochondrion serves as an important element in the tumourigenic phenotype, and clinical approaches targeting this organelle have potential for the development of effective treatment regimens for patients with cancer.
***
Introduction: the Tumourigenic Phenotype
Conceptually, understanding the behavior of tumor cells by analysis of the phenotype is an important aspect required to develop preventative measures, in addition to treatment modalities which are both effective and safe for patients with cancer. This “tumourigenic phenotype” has, in the past, been typically evaluated using approaches primarily addressing nuclear encoded genes. However, it is clear that the influence of cytoplasmic components of the cell also can play a role in the manifestations of malignancy therein. As a component of the cytoplasm, the mitochondrion has, in the past, been minimally investigated as a source of influence on the tumourigenic phenotype.
In order to further investigate the role of the mitochondria in this specific phenotype, an evaluation of copy number and fidelity of mitochondrial DNA was performed, as well as a determination of the role of functional mitochondria (manifest by removing mitochondrial DNA and evaluation of mitochondrial membrane potential) and phenotypic alterations based on sensitivity to programmed (or other mechanisms) cell death.
***
Investigation of copy number alterations of mitochondrial DNA in human gliomas, including investigations of mitochondrial DNA as a transposable element
Cancer arises from the accumulation of DNA and cytoplasmic abnormalities, which allows for the expression of a resultant phenotype. This phenotype – the tumourigenic phenotype – is manifested by certain elements, such as lack of terminal differentiation, immortality, anchorage-independent growth and resistance to cellular death from cytotoxic and further inducing stimuli (Cavalli and Liang, 1998). Gene amplification (of nuclear genes) is an important aspect of this multistep pathway to the neoplastic phenotype, and occurs in most, if not all, tumour types (Liang et. al., 1995). Indeed, cytogenetic and molecular cytogenetic data suggest that this genetic finding is an intermediate step in the progression of cells in the multistep pathway noted above. Within this context, several techniques have been developed to isolate variably expressed and amplified genes in human cancer, using brain tumors (gliomas) as a model system (correlating with the clinical interest in this tumor type). Using a molecular subtractive hybridization approach (Liang, 1995), an amplified cDNA was isolated which revealed mitochondrial (cf. nuclear) sequence from a glioblastoma cell line (PFAT-MT), an unexpected finding. Whilst mtDNA had been observed to be structurally abnormal in the past (Firkin and Clark-Walker, 1979; Gianni et. al., 1980), no studies had noted significant changes in copy number. Consequently, further studies were focused on attempts to understand the role of mitochondria, and mitochondrial DNA (mtDNA), in the tumourigenic phenotype, since the differential expression screen suggested a level of amplification and/or expression which differed from that of corresponding normal tissues.
Of especial note is that the mitochondrial genome is the only known extranuclear genome in mammalian cells, and consists of 16,569 base pairs arranged in a circular array, with each strand expressed as a single polycistronic message. There are effectively no introns in this genome, and all mRNA derived from this genome are transcribed and translated within the mitochondria (Scheffler, 1999).
The first study conducted after the identification of this cDNA fragment (Genbank Accession MDL U25123, or “U25123”) (Figure 1) involved an evaluation of further tumour cell lines in addition to primary tumours, in order to determine the frequency of differential expression and/or amplification.
Both cell lines and primary tumours were evaluated to determine whether this was an in vitro phenomenon only, or whether this was pathologically important and therefore detectable at the primary tissue level; moreover, the investigation of different stages of glioma (low grade – pilocytic tumors, astrocytomas; intermediate grade – anaplastic astrocytomas; high grade – glioblastomas) was performed to determine the relative ontogeny of any findings. Indeed, evaluations of the genomic DNA of these tissues revealed frequent amplification of not only the cell lines, but also the primary tumours; of much significance was the finding that lower grade tumours also harboured the amplification of this U25123 sequence, an observation suggesting that this was an early event in the multistep pathway towards malignancy, in distinct contrast to nuclear DNA amplification in cancer, which is generally more considered to be an intermediate step. Correlated with this was expression data in a subset of tumours in which RNA was available; these revealed an increased level of expression of the cDNA in both lower grade and cell line samples.
To further investigate the amplification of mtDNA in the glioma tumors and cell lines, probes from each segment of the mitochondrial genome were created. Interestingly, most of the mtDNA in the tumours and cell lines revealed an amplification process, although a segment which included components of the electron transport chain (including ND1 and ND2, viz. amino acid positions 4121 - 5275) had minimal increases (if any) in copy number. Indeed, in some cases the minimal hybridization on Southern blot suggested a deletion of that area of the mitochondrial genome.
These findings – i.e. amplification of the mitochondrial sequence in gliomas, the ontologically-early identification of amplification and heterogeneous areas of mtDNA amplification (suggesting deletions in the mitochondrial genome) – extended early findings of abnormal mtDNA structure in cancer (Clayton and Smith, 1975; Clayton and Vinograd, 1967) which were previously unreported in the literature. Moreover, the association with petit mutants in yeast (Ragnini and Fukuhara, 1989; Traven et. al., 2001), where truncated mtDNA is associated with a dysfunction of the mitochondrion and high degrees of amplification of the remaining DNA, was a unique insight, as was the hypothesis that a miscommunication between the nuclear and mitochondrial genomes was occurring, observations which would have implications in the later studies conducted (see below).
As a result of the discovery of mtDNA amplification, and the hypothesis that the pathobiology could recapitulate findings in yeast, the next set of experiments evaluated tumours with regard to “mtDNA escape”. In yeast, mtDNA escape occurs as a manifestation of the communication between the nuclear and mitochondrial genome (Campbell and Thorsness, 1998; Thorsness and Weber, 1996), and has been postulated to be related to both genetic and environmental control. mtDNA has been found transposed to the nucleus in cultured cells (HeLa) as well as carcinogen-induced hepatic carcinoma cells (Shay and Werbin, 1992; Corral et. al., 1989). Indeed, it has been hypothesized that high copy numbers of mtDNA may be required to transfer mitochondrial sequences from mitochondrial sources to the nucleus in cancer, a phenomenon similar to that in yeast (Netter and Robineau, 1989; Liang, 1996).
Given the above finding of mtDNA amplification, an evaluation of the presence of mtDNA in the nucleus of tumour cells was performed, using interphase nuclei with fluorescent DNA probes (interphase fluorescent in situ hybridization, or “interphase FISH”) (Figure 2). Moreover, lower grade tumours were employed in order to evaluate the early stage findings noted in previous experiments.
Utilizing the U25123 probe, mtDNA showed increases in copy number in all the lower grade tumours studied, but not the normal brain specimens. When evaluated by the interphase FISH technique using fluorescent probes of the entire mitochondrial genome, at least 69% of the nuclei revealed mitochondrial sequence hybridization in the tumour specimens. In comparison, when evaluating two normal brain samples, as well as a glioblastoma specimen (all of which did not reveal increases in mtDNA copy number), < 8% of nuclei revealed hybridization with the mtDNA fluorescent probes. This observation suggested that there was an association with the amplification of mtDNA and the finding of “promiscuous” mtDNA sequences in the nuclear compartment. Indeed, this was the first description of nuclear localization of mtDNA found to be associated with mtDNA amplification, and as an ontologically early event; furthermore, it was the first description of such an event found in primary tumour tissues, rather than cell culture experiments.
Hence, in this specific aim, the evaluation of copy number, sequence, promiscuity and diversity of tumour grade was performed, and the experiments involved revealed previously unreported results which suggested the importance of mitochondria and its DNA in the tumourigenic phenotype.
Data noted in alterations of the mitochondrial genome were followed by investigations by others, i.e. those regarding its importance in various other tumour types. Penta et. al. (2001) noted that in addition to gliomas, breast, liver and colon cancers showed increases in mtDNA/expression, with deletions in other tumours including those of the colon, stomach, bladder, head/neck and lung. Carew and Huang (2002) also identified mtDNA as being altered in these tumour types, and additionally, described findings in ovarian, esophageal, pancreatic, renal cell and prostate cancer, as well as thyroid cancers and hematologic malignancies. In particular, these authors noted gliomas to be the best-characterised tumours with respect to mtDNA alterations, noting the studies documented in this dissertation, as well as the study by Kirches et. al. (1999). Beheshti et. al. (2003) found mtDNA amplification in neuroblastoma, and Haugen et. al. (2003) found an increased expression of mtDNA encoded genes in papillary thyroid carcinomas. Jia et. al. (1999) showed increases in mtDNA levels with the mdr-1 genotype, and Tarantul et. al. (2000) found in SIV-infected primate non-Hodgkin lymphoma (NHL) evidence for amplified mitochondrial sequences; similarly, Nikolaev et. al. (2001) showed mtDNA increases in simian NHL, and Nenasheva et. al. (2004) found such sequences increased in SV40-transformed fibroblasts, observations suggesting that these modifications serve as early markers for the development and/or pathogenesis of malignancy. However, Dani et. al. (2004) evaluated breast, colorectal, gastric and head/neck tumours for the mtDNA4977 deletion (“common deletion”), and concluded that such tumour cells are essentially free of this mutation; Kiebish and Seyfried (2005) also did not find significant mtDNA mutations in mouse brain tumours, and concluded these did not play a role in either chemically-induced or spontaneous mouse brain tumours. However, this has not been confirmed by others; indeed, primary tumour cells in these tumour types did show alterations in the mtDNA4977 region, amongst other mutations (Kirches et. al., 2001; Kamalidehghan et. al., 2006; Poetsch et. al., 2004; Chung et. al., 2000; Polyak et. al., 1998; Dai et. al., 2006). Furthermore, the D-loop region (where mtDNA initiates replication, and is non-coding) has been frequently associated with alterations in copy number as well as mutation in osteosarcoma, uterine serous carcinoma, gastric carcinoma, and ovarian, endometrial and breast cancer (Zhu et. al., 2005; Guo and Guo, 2006; Pejovic et. al., 2004; Maximo et. al., 2001; Liu et. al., 2001; Wang et. al., 2005; Tan et. al., 2002; Lievre et. al., 2006); indeed, Duncan et. al. (2000) identified a novel human mitochondrial D-loop RNA species which was up-regulated subsequent to cellular immortalization, supporting data herein that mtDNA alterations represent an early modification in generation of the tumourigenic phenotype. Moreover, associations of D-loop mutations have been noted in the background of mutation of p53, associated with changes in expression of mitochondrial proteins (Zhou et. al., 2003; Gochhait et. al., 2008). Of note, others have found a distribution of mutations throughout the mitochondrial genome (Jakupciak et. al. 2005; Ivanova et. al., 1998). These studies support a role for mtDNA in the tumourigenic phenotype, and directly support the findings first reported herein of alterations in copy number (particularly increases) as an ontologically-early aspect of the pathway to malignant behaviour.
Czarnecka et. al. (2007) describe cancer as a “mitochondriopathy”, and note the alterations of mtDNA in various tumors, as well as insertion of mtDNA sequences into the nuclear genome, with findings in hepatoma and meningioma (Bianchi et. al., 2001; Yamauchi, 2005; Sorenson and Fleisher, 1996). Interestingly, in both human myeloid leukemia HL60 and colorectal carcinoma COLO 320DM cells, amplification of mtDNA was found to be associated with mitochondrial sequences identified in the nucleus; in normal fibroblasts and peripheral blood T-cells, no promiscuous mtDNA sequence was noted (Hirano, 1999). Similarly, in mouse and rat model systems, mtDNA-like inserts were found at higher frequencies in tumour tissue nuclei when expressed relative to those of normal cells (Hadler et. al., 1998). Finally, Ricchetti et. al. (2004) found mtDNA sequences interrupting intronic regions of MADH2, a tumour suppressor gene in colorectal cancer (Eppert et. al., 1996), and Borensztajn et. al. (2002) and Turner et. al. (2003) showed promiscuous mtDNA in the nuclear genome in both the intron and exon regions, which were associated with clinical phenotypes of disease (Factor VII deficiency and Pallister-Hall Syndrome, respectively). These studies directly confirm results documented in this dissertation regarding human gliomas; such results suggest that this may represent an important mechanism in the generation of the tumourigenic phenotype. These mtDNA changes could have significant implications regarding both the developmental mechanisms and detection of malignant cells in patients with cancer. Indeed, Copeland, (2002), Modica-Napolitano and Singh, (2002), Saffroy et. al. (2004), Chatterjee et. al. (2006), and Paul and Mukhopadhyay, (2007), have noted the relevance of mtDNA mutations and their potential consequences regarding both the diagnosis and treatment of patients with malignancy; in fact, both diversity of changes throughout the mitochondrial genome (Ray et. al., 2000; Cai et. al., 2004) in addition to specific polymorphisms (G10398A: Darvishi et. al., 2007; Canter et. al., 2005) have been noted to be associated with risk for cancer. Jiang et. al. (2005) found that there was an increased mtDNA content in saliva associated with head/neck cancer patients; treatment with apoptosis-inducing therapy (but not surgery alone) was associated with a decline in mtDNA levels (Jiang et. al., 2006). Hence, alterations in mtDNA may serve as a diagnostic measure reflecting response to therapeutic intervention with modalities such as radiation therapy. Indeed, Polyak et. al. (2002) have filed a patent to utilize the presence or absence of mtDNA mutations to detect tumour cells as a diagnostic approach in cancer patients.
***
Evaluation of the biology of mitochondrial alterations in the neoplastic phenotype, including the effect of removing mitochondrial DNA (creation of rho- tumor cells) and evaluation of the ontogeny of mitochondrial membrane potential changes in a model of neoplastic transformation
As an extension of the findings noted of the importance of mtDNA in neoplastic tissue from Specific Aim 1, a further understanding of the pathobiology involved was evaluated at the level of the phenotypic effects of decreasing the amplification and mitochondrial function, which employed techniques to decrease the levels of mtDNA (“rho-“ cells). In this manner, the direct effect of mtDNA, and mitochondrially-encoded biomolecules, would be evaluable, and the concepts potentially translated into the treatment of malignant disease.
Neoplastic cells (MCF-7, a well-known breast cancer cell line, and DBTRG-O5MG, derived from a glioblastoma), were treated with low concentrations of ethidium bromide (EtBr). EtBr is an inhibitor of mtDNA replication (Nass, 1972); hence, during cell growth and passage, the initial number of mtDNA molecules is decreased by a factor of two on their treatment with this agent. These cells were passaged for at least 30 doublings, and thus, < 1% of the original number of mtDNA molecules (viz. 1/230) were present. With such treatment, the cells became auxotrophic on uridine and pyruvate. The requirement for uridine relates to that for dihydrooratate dehydrogenase (an enzyme of pyrimidine biosynthesis, localized to the inner mitochondrial membrane) to have an active electron transport chain for normal function. The auxotrophic dependence of rho- cells on pyruvate is not definitively known or well described, but most likely relates to the inability of the cell to oxidize cytoplasmically-produced NADH. Cells were also noted to have minimal or no hybridization on dot blot to mtDNA probes, a monitoring system indicative of phenotypic rho- cells.
Further characterization was performed of the rho- vs. rho+ (parent) cells. Evaluations of the genes bcl-2, bax, mdr1, mrp and O6-alkyltransferase were conducted by reverse northern blots. These genes have been associated with changes in responsiveness to cytotoxic chemotherapy, and believed to be important in maintenance of the tumourigenic phenotype (Hickman, 1996; Susin et. al., 1996). Moreover, both BCL2 and BAX can be located at the inner mitochondrial membrane (Susin et. al., 1996) after being translated cytoplasmically. In all cases (bcl-2, bax, mdr1, mrp and O6-alkyltransferase), using reverse northern blotting, although each gene was expressed, there were no changes in expression observed with these genes in the rho- cells when compared to their rho+ parents.
A key aspect of the tumourigenic phenotype is the ability of such cells to grow in an anchorage-independent fashion (Nikiforov et. al., 1996). Both rho+ parent and rho- cells were assessed with regard to their abilities to form colonies in soft agar, an indication of anchorage-independence. Whilst both types of cells were able to attach to plastic tissue culture flasks, formation of colonies only occurred in the rho+ parent cells that harboured mtDNA, with minimal colony formation found in either MCF-7 or DBTRG-O5MG rho- cells. Indeed, even prolonged incubation of the rho- cells (over double the length of time of that for rho+ cells) in soft agar did not result in additional colonies being formed. Employment of a cell-conditioned medium from the rho+ parent cells into the rho- cell culture experiments did not result in any additional colonies being formed; furthermore, cell-conditioned medium from the rho- cells did not inhibit colony formation in the rho+ cells. When evaluated in xenograft experiments, only rho+ cells showed growth in the flanks of nude mice; no growth was found with the rho- cells. It has been noted that altered expression of bcl-2 and mutant p53 are potential mediators of anchorage independence (Nikiforov et. al., 1996). In these neoplastic cells, the p53 gene was wild type (Kruse et. al., 1992; Engel and Young, 1978), and no changes between rho+ and rho- cells in the expression of bcl-2 was noted. These data strongly suggested that changes in mtDNA, and mitochondria, were playing a role in anchorage independence, a manifestation of the tumourigenic phenotype.
Further studies were performed on the rho+ parents, and their derived rho- cells, with respect to sensitivity to apoptosis effected by cytotoxic chemotherapeutic agents. DBTRG-O5MG cells were treated with 1,3-bis-dichloroethyl-nitrosourea (BCNU), a common drug utilised in the therapy of patients with glioblastoma; MCF-7 cells were exposed to cis-diamino-dichloro platinum(II) (CDDP), one of the most commonly employed chemotherapy drugs in cancer treatment. In both cases, rho- cells were distinctly and significantly more sensitive to the respective agents than their parent rho+ cells, a previously unreported observation. Interestingly, despite a lack of mtDNA, both rho+ and rho- cells underwent apoptosis as a mechanism of cell death, as noted by morphologic criteria and chromatin condensation after exposure to these cytotoxic agents. This was observed at all doses of BCNU and CDDP investigated and in both MCF-7 and DBTRG-O5MG cells, i.e. an increased dosing level did not revert cells to a necrotic mechanism of cell death. Whilst it had been hypothesized that the putative drug resistance/apoptosis resistance genes bcl-2, bax, mdr1, mrp and O6-alkyltransferase would be involved in the phenotype changes, no changes were noted in expression. Hence, mitochondria must play additional role(s) in the drug resistance phenotype, at least in the manifestation of resistance to apoptosis. Of note, however, is that removal of functional mitochondria via depletion of mtDNA failed to prevent apoptotic cell death, an observation indicating that actively-translated, mitochodrially-encoded molecules are not, per se, required for apoptosis induction by cytotoxic chemotherapeutic agents, a clearly novel finding.
In order to further evaluate the potential function of the mitochondria, “cybrids” of the rho- cells were created. Cybrids are fusion cells of the rho- ones and normal cells which harbour mitochondria, in order to replace this organelle in order to further study mitochondrial function, and is a technique utilised significantly in the evaluation of mitochondrial encephalomyelopathies. In this case, normal platelets were employed, rather than either DBTRG-O5MG or MCF-7 de novo enucleated cells and cytoplast transformation, in order to avoid nuclear and cytoplasmic component influences of the tumour cells on the cybrid phenotype, i.e. entirely normal mitochondria with a minimal level of influence of nuclear-encoded genes; platelets are, by definition, fragmented aspects of cells, and thus anuclear. Subsequent to fusion, the cells were selected in uridine- and pyruvate-free medium, and assessed to determine mtDNA identity of the donor, copy number, anchorage independence, and sensitivity to apoptosis via cytotoxic agents.
Interestingly, a novel finding was that cybrids created from the rho- cells and platelet mitochondria showed mtDNA levels which were similar to that of the rho+ ones, viz. amplified; of particular note is that this level was higher than that of peripheral blood platelets and lymphocytes, an observation suggesting that a cellular influence to maintain increased mtDNA levels is still acting through the rho- state, and subsequent to platelet fusion. Moreover, the morphology of the cybrids differed from the rho- cells, and more resembled the rho+ cells, a potential reflection of the role of mitochondria on cell morphology (Felty and Roy, 2005).
When evaluating the cybrids for anchorage-independent growth, both MCF-7 and DBTRG-O5MG showed a return of this aspect of the tumourigenic phenotype. Indeed, both neoplastic cybrids essentially demonstrated a return to the rho+ cell levels of colonies formed in soft agar, with colony sizes and densities almost if not identical in nature. The growth as xenografts in nude mice also returned in both cell lines after mitochondrial transfer. Furthermore, the relative sensitivity of cells to the effects of cytotoxic chemotherapy – viz. induction of apoptosis – also changed from rho- levels to those of the rho+ cells. MCF-7 and DBTRG-O5MG cells returned to their levels of sensitivity to CDDP and BCNU, respectively, which was observed prior to removal of mtDNA amplification. This phenotypic reversion toward a more apoptosis-resistant state (at least with respect to chemotherapeutic agents) showed no difference with regard to the overall mechanism of cell death when compared to those of rho+ and rho- cells (Figure 3).
Hence, these data were the first to suggest that the nuclear influences controlling mtDNA copy numbers are maintained through the rho- state, and that mitochondria are at least participating in the maintenance of the tumourigenic phenotype, manifested as anchorage-independence and relative resistance to the cytotoxic effects of apoptosis-inducing agents CDDP and BCNU. Because cybrids were constructed from normal mitochondria, tumour mitochondria/mtDNA per se is not responsible for the observed phenotypic change of the immortal but diminished tumourigenic rho- cells, which revert to the de novo phenotype. Thus, mitochondria play a direct role in neoplastic maintenance (with nuclear encoded genes, which support mtDNA amplification), and yet may not be strictly required for apoptotic cell death.
To further understand the pathological significance of the role of mitochondria in malignancy, the mitochondrial membrane potential was evaluated in a model system of normal cell transformation, again to determine the relative ontogeny of mitochondrial modifications associated with progression to the tumourigenic phenotype. This is of particular interest, since further studies have shown, in chemotherapy resistant cell lines, that mitochondrial dysfunction is manifested by changes in mitochondrial membrane potential, dissociated from the energy-generating capacity of mitochondria (Fantin et. al., 2002). Moreover, evaluation of shunting cells towards oxidative phosphorylation by the inhibition of glycolysis was also evaluated to determine the potential value of using the aforementioned data in a therapeutic manner.
Normal human astrocytes (NHA) were treated with ethyl-nitrosourea (ENU) in order to derive transformed astrocytes. Cells were initially observed to become immortal, continuing to divide after untreated NHA reached terminal division; experiments were performed subsequent to the achievement of immortality (anchorage independence, low density growth). In addition, phosphofructokinase (PFK) is the major regulatory enzyme in glycolysis, and is allosterically-inhibited by citrate; cells were also treated with this glycolysis inhibitor to determine the effects of metabolic manipulation on the expressed phenotype.
NHA cells were unable to divide past approximately 20 doublings, and, moreover, could not grow in an anchorage-independent manner or under low density conditions. In contrast, cells exposed to ENU became immortal, and at approximately 26 doublings, achieved the ability to grow in soft agar and in low density. Interestingly, when NHA cells were treated with ENU and pre- and co-treated with the PFK inhibitor citrate, they still became immortal, dividing well past terminal differentiation. However, despite achieving the ability to avoid replicative senescence, these dually-exposed cells did not gain anchorage independence, nor the ability to grow in a low density. When ENU-transformed clones were subsequently treated with citrate (cf. pre- and co-treatment) and evaluated for growth under low density conditions, at and above an added level of 100 M of this agent, there was a decrease in colony formation. Hence, although ENU treatment resulted in immortal and anchorage independent cells, pre- and co-treatment of cells with citrate with subsequent ENU exposure diminished the morphological changes and transformed characteristics of the cells; use of citrate to “treat” malignant cells also decreased the transformed phenotype, albeit at higher concentrations. This was the first report of changes in cellular phenotype by the modification of glycolytic enzymes in chemically-transformed cells.
When evaluating mitochondrial membrane potential, additional alterations were found. In NHA cells, mitochondrial membrane potential was found to progressively decrease with increasing doublings, until terminal division. Citrate did not exert an effect on the NHA cells with respect to mitochondrial membrane potential. In contrast, when ENU was administered, after cells underwent 3 – 5 doublings, the ENU cells exhibited a significantly higher polarized mitochondrial membrane potential, when compared to that of NHA cells. With subsequent doublings, the ENU cells did decrease in mitochondrial membrane potential, although values were, in general, higher than NHA (or citrate exposed cells), with maintenance of an elevated membrane potential after the development of immortality, an observation which was previously unreported. Interestingly, in parallel with the lack of anchorage-independence and growth in low density, pre- and co-exposure of cells to citrate did not result in the increased mitochondrial membrane potential observed treated with ENU alone. No difference was noted between the ENU/citrate-exposed cells, and the NHA or NHA citrate-exposed cells with respect to mitochondrial membrane potential.
Furthermore, mitochondrial mass was also influenced by exposure to ENU and the development of a transformed phenotype. The NHA and NHA/citrate cells showed a very gradual, modest decline of mitochondrial mass, with successive doublings. With ENU administration, however, after ca. 10 doublings, there was a substantial drop in mitochondrial mass, to about 25% of the starting mass measured. Such mass changes stabilized after about 16 doublings, and were maintained after the achievement of immortality, and also during the achievement of anchorage independence. Citrate prevented the dramatic drop in mitochondrial mass; even with immortality achieved, the mitochondrial mass of ENU/citrate-treated cells was not different from NHA cells which were terminally differentiated. When calculating the mitochondrial membrane potential/mitochondrial mass ratios, such data revealed a hyperpolarized state of the mitochondrial membrane of the ENU-transformed cells, which was prevented by citrate exposure both prior to and concomitantly (Figure 4). These data suggested that hyperpolarization of the mitochondrial membrane represents is an early manifestation of the transformed/tumourigenic phenotype, associated with anchorage independence, and may be mitigated by use of the PFK glycolytic inhibitor citrate.
Finally, cells were evaluated to determine their levels of spontaneous apoptosis. NHA cells progressively underwent a higher level of apoptosis with each successive doubling, as did NHA cells treated with citrate. However, ENU cells only maintained a low level of apoptosis, which did not attain the same levels as the NHA- or NHA/citrate-treated cells at any cell culture doubling. In contrast, cells with ENU, both pre- and co-treated with citrate, had an increased degree of spontaneous apoptosis, which was, however, lower, than that of the NHA- and NHA/citrate-treated cells at each doubling. Hence, the treatment of NHA cells with ENU was associated with a decreased extent of programmed cell death and the achievement of immortality; citrate increased the level of spontaneous apoptosis in ENU cells (although not to the same degree as that of the NHA cells).
Thus, data derived from this specific aim indicate that mitochondria and/or its genome play a significant role in human cancer, and that alterations of this organelle can effect basic changes in the phenotypic representation of malignancy (anchorage independence; xenograft growth; relative resistance to apoptosis via cytotoxic chemotherapeutic agents). Furthermore, these modifications are found early in the ontogeny of the tumourigenic phenotype, since the development of transformed astrocytes was found to involve early changes in mitochondrial membrane potential, hyperpolarization of the mitochondrial membrane, and constraints on apoptosis, all of which could be modified by inhibition of the glycolytic enzyme PFK.
These data directly implicated the mitochondrion in the tumourigenic phenotype; these observations were subsequently confirmed and extended to further systems. Karthikeyan and Resnick (2005) noted the consequences of mitochondrial dysfunction on genome stability, which included an identification of the importance of heteroplasmy (differences in mtDNA types within the same cell and/or organ), as well as effects of diminished mitochondrial gene expression which influence both the invasive as well as the transformative phenotype. Moreover, Xue et. al. (2001) using Boc.Aspartyl(O-methyl)CH2F, removed mitochondria from neurons and HeLa leukemia cells; such removal did not damage other cellular components, but essentially irreversibly committed these cells to death via an apoptotic mechanism. Hence, in cells with mitochondria inhibited or removed, apoptosis is more readily effected when expressed relative to that occurring in control cells. Dong et. al. (2002) showed that mitochondrial transcription factor A (mTfam) (a key regulator of DNA transcription) in Morris (mouse) hepatoma cells was significantly over-expressed (over 10-fold when compared to normal liver), with augmentation of downstream mitochondrial gene expression. Moreover, mTfam protein was also found in the hepatoma cell nucleus, suggesting a role in nuclear–mitochondrial communication in this neoplastic lesion. Indeed, these findings indicated that the mitochondrial influence on the observed tumourigenic phenotype is complex, and involves a myriad of different proteins interacting on both upstream and downstream pathways. Grandemange et. al. (2005) performed an elegant study, in which human dermal fibroblasts acquired transformed properties (morphologic changes, anchorage-independent growth, growth in xenografts) with stimulation of mitochondrial activity, using p43. Moreover, such transformation was associated with an increase in c-Jun and c-Fos expression, and virtual extinction of tumour suppressor genes p53, p21WAF1, and Rb. When mitochondria were then inhibited using chloramphenicol, a loss of the tumourigenic phenotype was observed (although restoration of the tumour suppressor genes did not occur). This study thus identified p43, at least in dermal fibroblasts, as a potential target mechanism by which the early involvement of mitochondria in the tumourigenic phenotype may be manifested. Similarly, additional studies in differing model systems have identified proteins which may be responsible for observation of the importance of mitochondria in the malignant phenotype. In a clinically-based study, Ambrosone et. al. (1999) found an association of manganese superoxide dismutase (a protein transported to mitochondria) genotypes (with limitations on transport into mitochondria) with breast cancer risk; Roy et. al. (2007) noted estrogen-related processes in the nuclear–mitochondrial communication, via estrogen-induced mitochondrial oxidants. Ohnami et. al. (1999) showed that K-ras impacts the expression of mitochondrial genes; when utilizing an anti-sense approach to pancreatic cancer cells, they identified 11 over-expressed genes, which were all mitochondrial in origin. This continues to suggest the importance of the nuclear–mitochondrial communication pathway, where K-ras represents an important mediator of mitochondrial gene expression and changes in phenotype. Mills et. al. (1999) identified CD14 as a potential mediator of apoptosis in HL60 cells; they documented that inhibition of mitochondria using rotenone, antimycin A, and oligomycin all resulted in an increased expression of CD14 in these leukemic cells, followed by loss of viability and apoptosis. Magda et. al. (2008) studied A549 rho- cells (lung cancer) induced by exposure to EtBr (although auxotrophy was not reported in the study). This group showed a less proliferative ability in the rho- cells, but a greater level of expression of intermediate filaments, suggesting a more invasive phenotype. This observation is interesting in light of the hypothesis by Felty and Roy (2005) that malignant cell spreading may be ascribable, in part, to the contraction and expansion of mitochondria, rather than only the maintenance of tension on the cytoskeleton via binding to the extracellular matrix. Moreover, these A549 rho- cells, whilst able to grow in mouse xenografts, required longer induction periods with smaller tumours than their rho+ parents. This suggests a definite difference in phenotype, perhaps manifest by differences in nuclear gene expression (Li et. al., 2008; Ismail and Bateman, 2009). Interestingly, rho- cells increased the expression of MHC-1 in addition to glucuronidation genes, relevant in immune surveillance and detection function; this indicates the relevance of mitochondria to the extrinsic manifestations of apoptosis induction, an observation similar to that made by Mills et. al. (1999) with the expression of CD14, and our studies with CD95 (Fas) (Liang, Depletion of Mitochondrial DNA from Glioblastoma Cells results in Increased Fas Expression with Increased Sensitivity to Apoptosis, manuscript in preparation).
As noted, the communication between the nucleus and mitchondria as hypothesized from these studies (maintenance of amplified mtDNA/increased expression and reversion of phenotype of cybrids; alterations in mitochondrial membrane potential during the transformation of NHA cells) represents an important aspect of the maintenance of the tumourigenic phenotype. In particular, gradual reductions of mtDNA levels in rhabdomyoblasts (transformed muscle cells) using EtBr effected changes in mitochondrial membrane potential and elevated free Ca2+ concentrations in the cytosol (Biswas et. al., 2005, Amuthan et. al., 2001); these changes resulted in alterations of genes related to Ca2+ transport and storage, which could be reverted once cells were allowed to grow without EtBr present. Interestingly, as mtDNA levels were declining (but not absent), there was an increase in the ability of the rhabdomyoblasts to move through a reconstituted Matrigel membrane, an observation suggesting a migratory phenotype which was lost once mtDNA levels were allowed to return to their control values. These data suggested that mtDNA levels are affected not only early in the tumourigenic phenotype, but also in ontologically-subsequent manifestations of malignancy, and could serve as a marker which identifies a pathway by which the migratory phenotype is effected. Hence, using feedback to the mitochondria and mtDNA levels, the nuclear–mitochondrial communication could effect the expression of nuclear genes in response to the microenvironment milieu; this occurs in yeast at a variety of rho- settings (Butow and Avadhani, 2002; Giudice et. al., 2005).
The maintenance of the tumourigenic phenotype was also found to be related to mitochondrial membrane potential, the work contained within this dissertation being the earliest study conducted with regard to the identification of the early stages of chemical transformation in mitochondrial membrane potential. Interestingly, in neu-initiated mammary epithelial cell tumours with increased mitochondrial membrane potential values, inhibition of glycolytic pathways using short hairpin RNAs of LDH-A decreased mitochondrial membrane potential and reduced the ability of these cells to proliferate in both de novo and hypoxic environments (Fantin et. al., 2006), supporting this researcher’s data acquired with the use of citrate as a glycolytic inhibitor of PFK, with associated changes in membrane potential reflective of the overall tumorigenic phenotype (i.e. resistance to apoptosis and growth in xenografts). Other researchers have hypothesized similar approaches (Bucay, 2007). Matarrese et. al. (2001) showed in a drug resistant human T-lymphoblastoid CEM cell line VBL100 variant (with amplified mtDNA), which was sensitive to TNF-, so that the induction of apoptosis was attributable to the depolarization of mitochondrial membrane potential per se. Indeed, these authors showed a direct correlation between susceptibility to apoptosis (by various agents) and the state of the mitochondria (mitochondrial membrane potential); prevention of the decrease in the membrane potential, and thus mitochondrial function, using bongokreic acid (which inhibits the decrease in mitochondrial membrane potential by antagonising the adenine-nucleoside transporter, which, in turn, is responsible for allowing mitochondrial-membrane permeability transition and thus a run down of the potential), the apoptosis-inducing effects could be mitigated by the various agents which effected the intrinsic (viz. mitochondrial) pathway of apoptosis (but not the extrinsic pathway).
***
Evaluation of the mechanism and potential of inhibition of mitochondrial function in effecting tumour cell death as a transition to approaches to the treatment of patients with cancer
With data obtained as noted, interest focused on a translational evaluation of understandings of the mechanisms of programmed cell death using chemotherapeutic agents, combined with the mitochondrion as a target for treatment.
Further mechanistic work was performed with CDDP. Indeed, this is one of the most commonly utilised chemotherapeutic agents, and recent data has suggested that nucelotide excision repair (NER) genes may play a significant role in the resistance of cancer cells to this drug. A further goal was to determine in further tumour cell types (e.g. hematologic malignancies) whether the findings of the impact of mitochondria on the tumourigenic phenotype could be extended. Using the U937 cell line (derived from a histiocytic lymphoma), rho- cells were created using EtBr, as were cybrids, using normal human platelet mitochondria in order to avoid potential nuclear-encoded cytoplasmic components which could exert an impact on results. Expression of the NER genes ercc1 and ercc2 was evaluated, since these NER genes have been found to be important in platinum(II) drug-resistance in ovarian carcinoma patients; moreover, the BCL-2 family genes bcl-2 and bax, as well as the drug resistance genes mdr and mrp, were also evaluated.
Rho- U937 cells were found to have an increased susceptibility to cell death induced by CDDP when compared to the rho+ parent, that being observed ca. twice as sensitive to the drug. When U937 cybrids were assessed for relative drug resistance, no difference was noted between the cybrids and the rho+ cells at any CDDP dose level tested, an observation similar to that of previous results obtained with MCF-7 breast cancer and DBTRG-O5MG glioma cultured cells. Moreover, an increased copy number was again noted in the U937 cybrids, similar to the rho+ cells, and this observation is similar to the findings previously noted in breast and glioma cancer cybrids. mRNA expression levels were evaluated in the NER, BCL-2 family member, and drug resistance genes; whilst there were relative differences in expression of the different genes evaluated, no changes were noted between the rho+ and rho- cells, nor in the cybrids. Hence, depletion of mtDNA from parental cells did not alter the expression of putative genes potentially responsible for overall CDDP resistance. Moreover, as previously noted, cells continued to die an apoptotic death, irrespective of whether they were rho+, rho- or cybrid in nature.
Furthermore, in order to determine whether changes (i.e. decreases) in mitochondrially-derived ATP levels accounted for the increases in sensitivity to CDDP-induced apoptosis in the rho- cells, the mitochondrial F0F1-ATPase was inhibited with oligomycin in the rho+, rho- and cybrid cells. Interestingly, for all cell types, no evidence for differences in ATP levels between all samples collected before and after treatment with oligomycin were found (Figure 5).
Moreover, no differences were noted between the ATP concentrations of each cell type. This observation suggests the generation of ATP by these cells is less dependent on oxidative phosphorylation, but rather glycolysis, in accordance with further earlier studies (Warburg, 1926); hence, mitochondrial energy production per se (or lack thereof) is not responsible for the increased sensitivity of rho- cells to CDDP-induced cell death, a previously unreported finding.
Consequently, the mitochondria may be an appropriate mechanistic target to sensitize cells to apoptosis using cytotoxic agents. Therefore, the penultimate goal is to determine an approach which would allow for mitigation of the tumourigenic phenotype manifested by an increased sensitivity to programmed cell death by cytotoxic chemotherapeutic agents, but one which would have minimal (if any) toxicity toward normal tissues.
Hence, further studies evaluated agents which were noted to have anti-mitochondrial activity. 2’,3’-dideoxycytidine (ddC) is an anti-retroviral agent, and is a potent inhibitor of DNA polymerase gamma, and consequently a distinct anti-mitochondrial activity (Bouayadi et. al., 1997). It is currently approved for patients with acquired immune deficiency syndrome (AIDS) as part of the multi-agent cocktail treatment regimen. Two tumour cell lines (glioma), DBTRG-O5MG and DK-MG, were exposed to ddC at doses ranging from 10 nM to 100 M. At all doses, ddC effected morphological changes in cells, consisting of larger cell bodies with more prominent processes and a ground glass appearance. Such changes occurred at relatively early stages, i.e. within 2-3 days following the commencement of ddC treatment, regardless of dose. Indeed, at the onset of the morphologic changes noted, alterations of mitochondrial membrane potential were also noted, with a decrease in polarized mitochondria at all doses evaluated; up to 31% diminished in DTBRG-O5MG and 84% in DK-MG cells treated with ddC. At all doses of ddC studied, cell death ensued eventually in both glioma cell lines, which was apoplectic in nature; cells became non-viable in < 6 hours after first being noted to detach from the tissue culture flask. A threshold between 100 nM and 1 M was noted in both DBTRG-O5MG and DK-MG cell lines, where the time between exposure to ddC and subsequent cell death decreased from approximately 30 days to <10 days. This observation suggested that the mechanism of ddC-induced programmed cell death may be a saturable phenomenon (e.g. enzyme- or receptor-mediated). Interestingly, at some doses of ddC (between 10 nM and 15 M), a delay (but not total rescue) in the death of both cell lines could be obtained with supplementation of the medium with glucose, uridine and pyruvate, suggesting the cells lacked a fully functional respiratory chain; however, cells treated with higher doses of ddC were unresponsive to such supplementation. This suggests that mitochondria (in a functional rho- state) are involved in at least one mechanism of the observed ddC-induced cellular death, although since rescue was only partially possible with glucose, uridine and pyruvate, and the cells eventually died, that the overall mechanisms effecting apoptosis are likely to be more complex and multifactorial. Again, in all cases the mechanism of death was apoptosis. These data are suggestive that mitochondria may per se serve as a target for therapy, but the therapeutic index may represent an issue (particularly as most regimens contain multiple drugs). Whilst the toxicity of ddC clinically can be managed (peripheral neuropathy, neutropenia), the type of dosing that might be required (chronic dosing, with other cytotoxic agents) could result in some not insignificant levels of organ toxicity. Hence, agents which have a wider therapeutic index and more selective towards mitochondria, particularly within combination regimens, might be more appropriate for patient therapy.
Hence, a search for compounds which could target the mitochondrion ensued; in particular, compounds which had lipophillic (i.e. diffusible into the tumour cell and past mitochondrial membranes), and cationic (allowing attraction and retention in the anionic environment of the mitochondrion) characteristics were sought. A compound which fit these criteria was tyrphostin AG17 [NSC 242557, (3,5-di-tert-butyl-4-hydroxybenzylidene) malononitrile] (Burger et. al., 1995). This agent also has tyrosine kinase inhibitory activity, which is manifest as a growth retardation mechanism in other studies (Ara and Teicher, 1996). AG17 was used at non-growth inhibitory concentrations (0.25 M) in order to avoid normal cell toxicity, and to evaluate the potential anti-mitochondrial contribution to any cytotoxicity noted in combination treatments.
Treatment of tumour cell lines DBTRG-O5MG and DK-MG, as well as NHA cells, with AG17 in culture at 0.25 M did not inhibit growth. In order to determine whether the agent was affecting mitochondria in either tumour or normal cells, evaluations of auxotrophy on uridine and pyruvate was assessed after 30 days of exposure. On removal of such supplements, glioma cells stopped proliferating and gradually became detached from the tissue culture flasks over periods of 7 to 15 days; staining with EtBr and acridine orange showed chromatin condensation, indicative of apoptotic cell death. In contrast, in the NHA cells, removal of uridine and pyruvate supplementation did not result in either detachment nor increased apoptosis as noted in the tumour cell lines. Of note was the observation that NHA cells did undergo replicative senesence after about 20 doublings, but this was not affected by the treatment of cells with AG17. Consequently, the AG17 cell treatment effected a auxotrophic phenotype in tumor (but not normal) cells, indicative of diminished respiratory chain function, which did not occur in NHA. This suggested a higher selectivity of AG17 towards tumour mitochondria over those of normal cells at these lower doses, and the avoidance of non-targeted and potential toxic effects might be possible with this agent.
In order to further define the putative mitochondrial mechanism, mitochondrial membrane potential was determined with the treatment of cells with AG17. In tumour cells, treatment with the agent was associated with a decline in the prevalence of polarized mitochondria, from 32 – 65% (as noted by cytofluorimetric analysis), with the highest level of depolarization observed with the lengthiest exposure to this agent. Of note was the observation that mitochondrial mass was found to be diminished in these glioma cells, reflecting a higher overall relative polarization of the mitochondria, and thus an even more impressive decrease in overall degree of depolarization of mitochondrial membrane potential. Moreover, the mitochondrial mass was also found to decrease with prolonged exposure to AG17, suggesting that the decrease in mitochondrial membrane potential was at least in part ascribable to reductions of the overall mitochondrial mass in the tumour cells. In contrast, NHA cells displayed only slight reductions in mitochondrial mass, and approximately similar mitochondrial membrane potential values over the study period. With a more prolonged evaluation (> 45 – 50 days), reductions in mitochondrial membrane potential were noted, both in the AG17-treated and untreated cells (data not shown).
The evaluations of AG17 treated cells were then focused on their susceptibilities to cytotoxic agent-induced apoptosis (Figure 6). In both tumour cell lines, treatment with BCNU (at the highest dose level,144 g/ml) was associated with significant increases in apoptotic cell death after AG17 exposure of 2 days in DBTRG-O5MG, and 3 days in DK-MG glioma cell lines. Indeed, exposure to AG17 after 2 or 3 days was associated with increased sensitivity to lower doses of BCNU as well. Maximally, after 30 days of exposure to AG17, treatment with BCNU reduced the number of viable cells by over 99%. A linear relationship was noted between survival at a BCNU dose of 144 g/ml and mitochondrial membrane potential in the glioma cell lines. In contrast, NHA cells sensitivity to BCNU was not modified (LD50 of 40 g/ml for both AG17 treated and untreated cells). In all cases, the mechanism of BCNU-induced cell death was not altered when mitochondrial membrane potential and mitochondrial mass are diminished by AG17 treatment, and thus decreases in the prevalence of both polarized and total mitochondria, with increases in sensitivity to cytotoxic agents, still gives rise to an apoptotic death.
These data suggest that the use of AG17 as a chemosensitizer acts, at least in part, at a mitochondrial level. Since this compound is largely a lipophillic cationic species, it and related compounds should be valuable for treatment of tumours where an increased anionic mitochondrial membrane potential is present, as noted in these glioma cell lines as well as those of other cancers (Modica-Napolitano et. al., 1996). Indeed, in view of the ability to diffuse into and be retained within the mitochondrial space, this may serve as the mechanism involved in studies which revealed cellular chemosensitisation by AG17. Within this specific aim, the finding that targeting mitochondria to enhance the potential sensitivity of tumour cells to cytotoxic agents is of particular importance with respect to the design of new chemotherapeutic and other approaches with available as well as newly-developed novel agents. Indeed, such increases in sensitivity of tumours to chemotherapy could allow the disease process to be more accessible to therapeutic intervention, with limited toxic effects to normal tissues.
With the findings of changes in the tumourigenic phenotype related to mitochondria, a great deal of interest has been focused on targeting organelle-related functions for the treatment of human malignancy (Ngyuyen and Hussain, 2007). Yen et. al. (2005) confirmed data noted previously in U937 cells, in rho- osteosarcoma (143b cell line) treated with CDDP; that particular group showed increases in sensitivity to CDDP, which were related to significantly increased caspase 3 activation. Indeed, these 143b rho- cells were also found to have an increased release of cytochrome c into cytoplasm with exposure to CDDP, an observation suggesting that caspase activation, in addition to increases in the magnitude of response to CDDP, play a role in rho- apoptotic sensitivity. Moreover, Cummings and Schnellmann (2002) also showed at least partial (50%) reliance of CDDP-induced apoptosis on caspase 3 activation in normal renal proximal tubular cells, without changes in ATP concentrations; further potential mechanisms included those speculated to involve either BAX or BID proteins, but most likely are not ascribable to caspase 8 or 9, which were inhibited in these studies. Similarly, Schwerdt et. al. (2005), studying normal collecting duct-derived cells, showed a similar increased sensitivity to CDDP when incubated with mitochondrial inhibitors; again, no apparent modifications in ATP levels were observed. In U937 lymphoma cells, Troyano et. al. (2001) also showed that ATP levels were maintained in cells where mitochondria were inhibited with oligomycin, a result further emphasising that data contained herein demonstrated that these cells are not dependent on mitochondria for energy production. However, of much interest was the additional finding that depletion of glutathione (GSH)(a reactive oxygen species scavenger, transported into the mitochondria) increased the sensitivity of cells to CDDP-induced apoptosis when mitochondria were inhibited, an observation similar to those of Kharbangar et. al. (2000) and our studies conducted on rho- U937 cells. Moreover, with increases in duration of CDDP exposure, a necrotic phenotype was noted, along with a decline in mitochondrial membrane potential. This suggested GSH may serve as a candidate for moderating sensitivity, as well as being involved in the mechanism of mitochondrially-mediated cell death, and that mitochondria, at the extreme, may also play a role in necrotic cell death. In contrast, Yang et. al. (2006) found that in head and neck cancer cell lines which were enucleated, there was no difference in sensitivity to CDDP when compared to parental cells harbouring nuclei; however, rho- cells were more resistant to the effects of CDDP, i.e., up to five times more resistant to the effects of the drug. Similarly, Lee et. al. (2004) found, when evaluating TRAIL-induced apoptosis in SK-Hep1 hepatoma cells, that rho- cells were more resistant than parental cells to the effects of this cytokine; a different model (HeLa cells treated with photodynamic therapy or adriamycin) also showed resistance to apoptosis in rho- cells (Singh et. al., 1999). However, these data were in contrast to the observations by Mizutani et. al. (2009), which revealed cybrids derived from tumour mitochondria were more resistant to CDDP than rho- cells, an observation which agreed with previously cited results and results herein, and Mizumachi et. al. (2008) who documented that head/neck cancer cells which were anthracycline-resistant had significantly higher mtDNA content. Furthermore, in alternative model systems, photodynamic approaches have also been shown to reveal increases in cell death with mtDNA depletion (Kessel and Sun, 1999; Morgan et. al., 2000); indeed, in L1210 murine leukemia cells, there was a significant increase in their sensitivity to mitochondrially-related photodamage in rho- cells, a process giving rise to an apoptotic death; when made resistant to ddC, CEM lymphoblast cells showed increases in their mtDNA content and expression (Bjerke et. al., 2008). These differences may arise from the various cell sensitivities to apoptosis (based on grade and stage, and thus nuclear gene expression, e.g. bax expression), the effects of CDDP directly on the mitochondria, and/or alterations in mitochondria during the development of CDDP-resistance, and hence its susceptibility to the platinum(II) ion-based chemotherapy. This hypothesis has been proposed by a number of research groups (Liang and Ullyatt, 1998; Lee et. al., 2004c; Yang et. al., 2006; Hirama et. al., 2006; Cullen et. al., 2007). Nonetheless, sensitivity to apoptotic-inducing stimuli, such as CDDP, can be manifested by the state of the mitochondria, which has direct therapeutic implications with regard to both targeting and combination therapies.
Download this book for your ebook reader.
(Pages 1-25 show above.)