Mitochondria and cancer: is there a morphological connection?


Mitochondria are key players in several cellular functions including growth, division, energy metabolism, and apoptosis. The mitochondrial network undergoes constant remodelling and these morphological changes are of direct relevance for the role of this organelle in cell physiology. Mitochondrial dysfunction contributes to a number of human disorders and may aid cancer progression. Here, we summarize the recent contributions made in the field of mitochondrial dynamics and discuss their impact on our understanding of cell function and tumorigenesis.


Mitochondria play essential and diverse roles in the physiology of eukaryotic cells. Not only do they provide energy, but they also participate in numerous metabolic reactions and play central roles in apoptosis (Desagher and Martinou, 2000). Impairments of mitochondrial functions have been implicated in a wide variety of human pathologies, among which cancer and age-related diseases (Wallace, 2005).

In the past years, a number of groups showed that mitochondria are dynamic structures that undergo fusion and fission events continuously throughout the life of a cell (Okamoto and Shaw, 2005). Mounting evidence indicate that mitochondrial dynamics have roles beyond maintenance of morphology, and impact on both cell death and cell metabolism (Chen and Chan, 2005). In mammals, mitochondrial fusion is driven by two GTPases of the outer membrane named Mitofusins 1 and 2 (Mfn1, Mfn2) (Santel and Fuller, 2001; Chen et al., 2003; Eura et al., 2003). The Mitofusins appear to play distinct roles in fusion and the Mfn1 presumably works together with Opa1, a GTPase of the inner mitochondrial membrane (IMM) to promote fusion (Cipolat et al., 2004). While its exact function remains unclear, Opa1 seems also involved in further intramitochondrial remodelling (Olichon et al., 2003). Mitofilin is another protein that has recently been involved in cristae morphology (John et al., 2005). Although models for mitochondrial fusion are still incomplete, the mechanism proceeds through three steps: docking, fusion of the outer mitochondrial membrane (OMM) and fusion of the IMM. All three steps involve GTP hydrolysis, and the IMM fusion requires an intact membrane potential (Mattenberger et al., 2003). Other players have been identified in yeast but no mammalian homologs to them have been described so far (Tieu and Nunnari, 2000).

The mitochondrial fission pathway involves at least two proteins: dynamin related protein 1 (Drp1) and Fis1. Drp1 is a cytosolic GTPase, which has been proposed to couple GTP hydrolysis to membrane constriction and fission (Smirnova et al., 2001). Fis1 is an outer membrane protein evenly distributed on the surface of mitochondria (James et al., 2003). It is thought to recruit Drp1 to punctuate structures on mitochondria and is thus considered as the limiting factor in the fission process (Stojanovski et al., 2004). The mechanism of constriction is not understood but the current model postulates that Drp1 homo-oligomerizes and forms a ring around the mitochondrial tubule (Okamoto and Shaw, 2005). Drp1 complexes might generate mechanical force via conformational changes, leading to membrane constriction, similar to dynamin (Danino et al., 2004). Modification of lipid composition is likely to be involved, and additional proteins may participate in mitochondrial fission. For example, endophilin B1, a putative fatty acyl transferase, has been shown to act downstream of Drp1 in the control of mitochondrial morphology (Karbowski et al., 2004) and MTP18 is a novel mitochondrial protein that is able to trigger mitochondrial fragmentation in a Drp1-dependent fashion (Tondera et al., 2004). GDPAP1, which is expressed in Schwann cells and neurons, has also been reported to mediate mitochondrial fission (Niemann et al., 2005).

Membrane fusion and fission engage rupture and deformation of lipid bilayers, and involve destabilized monolayer intermediates (Chernomordik and Kozlov, 2003). The two membranes are first brought in close contact, then some hydrophobic monolayer discontinuity appears that enables the rupture of the bilayer and the formation of a fusion stalk intermediate. Successive steps include the opening of a pore and its expansion, connecting the two aqueous volumes (Kozlovsky et al., 2002). These processes involve specific proteins that establish local intermembrane contacts and generate local membrane stresses. Absence or excess of one of these proteins could in theory result in failure of specific steps and accumulation of intermediates.

As summarized in Table 1, disruption of the fission/fusion machinery strongly alters organelle function affecting both programmed cell death and energy production. Several lines of evidence suggest that mitochondria contribute to neoplastic transformation by changing cellular energy capacities, increasing mitochondrial oxidative stress and modulating apoptosis. A shift in the rate of mitochondrial fission or fusion may provide new mechanistic explanation for the aetiology of cancer, and may offer additional strategies for therapeutic intervention.

Table 1 Proteins involved in mitochondrial fusion and fission and their effect on apoptosis and energy metabolism

Here, we discuss a possible role played by mitochondria in oncogenesis and explore how misregulation of fission and fusion machinery may influence apoptosis, cell metabolism and tumour formation.

Mitochondrial apoptosis and its subversion in oncogenesis

The possibility that apoptosis is implicated in cancer formation was first raised in 1972 when Kerr et al. (1972) described massive apoptosis in tumoural cell populations. In the following years, the antiapoptotic protein Bcl2 was shown to be a tumour-specific marker which expression was enhanced in many lymphoma (Tsujimoto et al., 1984; Bakhshi et al., 1985; Crescenzi et al., 1988; Vaux et al., 1988; Nunez et al., 1990), opening the investigation of apoptosis in cancer at the molecular level. Suppression of apoptosis is among the minimal requirements for a cell to become cancerous, and a hallmark of most, if not all, types of cancer (Hanahan and Weinberg, 2000; Green and Evan, 2002). The inactivation of the p53 protein is the most common strategy developed by tumour cells to evade apoptosis. p53 is central in the DNA damage response and triggers cell death by both transcription-dependent and -independent mechanisms (Slee et al., 2004; Yee and Vousden, 2005). Other examples of apoptosis evasion include disruption of Fas death receptor pathway, upregulation of the widespread antiapoptotic family of inhibitors of apoptosis proteins (IAPs) or downregulation of caspases (Muschen et al., 2000; Teitz et al., 2000; Yang et al., 2003). The PI3K/Akt/ PTEN pathway was also shown to play a role in abrogating apoptosis. Akt/PKB is activated in a wide variety of cancers and this activation leads to the phosphorylation of a number of key apoptotic proteins thereby inactivating them (Testa and Bellacosa, 2001).

Mitochondrial membrane permeabilization

Many death signals converge to mitochondria which respond by releasing numerous proteins in the cytosol (Patterson et al., 2000), among which several have apoptogenic properties (e.g. cytochrome c, SMAC/DIABLO, AIF, Endo G and Omi/HtrA2). In many cell types mitochondrial membrane permeabilization (MMP) is considered as the point of nonreturn in the cell suicide (Ferri and Kroemer, 2001). However, in sympathetic neurons for example, this event appears to be reversible and death can be circumvented downstream of cytochromic c release (Martinou et al., 1999; Deshmukh et al., 2000; Tolkovsky et al., 2002). Whether it is also the case in cancer cells remains to be determined. Upon release, cytochrome c triggers the apoptosome formation and initiates the caspase cascade (Nicholson, 1999). These killer proteases target many vital proteins, committing the cell to death (Fischer et al., 2003).

Mitochondrial membrane permeabilization is regulated by members of the Bcl2 family (Lucken-Ardjomande and Martinou, 2005b). These proteins contain at least one of the four conserved regions called Bcl2-Homology domains (BH1-BH4). Antiapoptotic members (e.g. Bcl2, Bcl-xL, Bcl-xW, Mcl-1 and A1/bfl-1) display all four BH domains, proapoptotic members (Bax, Bak and Bok/Mtd) lack BH4, and BH3-only proapoptotic members (Bim/Bod, Bid, Bad, Bmf, Bik/Nbk, Blk, Noxa, Puma/Bbc3 and Hrk/DP5) harbour only BH3. Most of them also display a hydrophobic C terminus that allow them to associate with membranes (Schinzel et al., 2004). Heterodimerization of the Bcl2 family members seems crucial to the regulation of MMP, and structural data has shown that antiapoptotic proteins display an hydrophobic groove on their surface, providing a binding site for the BH3 domain of the proapoptotic members (Petros et al., 2004).

The exact mechanism leading to MMP is still unclear, although several models have been proposed for this process (Figure 1) (Martinou and Green, 2001; Crompton et al., 2002; Waterhouse et al., 2002). According to the first model, the opening of a high-conductance channel, the permeability transtition pore (PTP), would cause loss of mitochondrial membrane potential (Δψm), swelling of the organelle and rupture of the OMM as solute and water enter the mitochondrial matrix (Halestrap et al., 2002). It should be noted that Δψm loss is not sufficient to prove PTP involvement, as many other events can provoke it. Although the composition of the PTP remains to be resolved, the prevalent hypothesis is that it consists of three core proteins, namely the voltage dependent anion channel (VDAC), the adenine nucleotide translocase (ANT) and cyclophilin D (Halestrap et al., 2002). The PTP is thought to open in response to calcium and oxidative stress and Bcl2 family members also seem to regulate its activity (Zamzami and Kroemer, 2001). However, recent studies of mitochondrial permeabilization in ANT and Cyclophilin D deficient backgrounds tend to show that these proteins are not obligatory components of the channel (Kokoszka et al., 2004; Baines et al., 2005; Basso et al., 2005; Nakagawa et al., 2005). Both proteins appear to be regulators of the PTP opening as they influence its sensitivity to Ca2+.

Figure 1

Proposed mechanisms for mitochondrial membrane permeabilization (MMP). (a) Classic models involve either the formation of channels by proapoptotic Bcl2-family members (e.g. Bax or Bak) or opening of the permeabilization transition pore (PTP). Upon apoptosis induction, Bax/Bak undergo conformational changes and oligomerize to form large channels. Apoptotic factors are then released from the mitochondrial inter membrane space (IMS) and cristae compartment into the cytosol where they trigger caspase activation and cell death. Alternatively, death stimuli activate pathway that trigger PTP opening such as Ca2+ or ceramide, allowing water and solutes to enter the mitochondrial matrix. Subsequent swelling of the organelle provokes the rupture of outer mitochondrial membrane (OMM) and release of apoptogenic factors into the cytosol. (b) An alternative scenario could be that during apoptosis, a membrane-perturbing agent (either tBid, Bax or a protein that controls mitochondria morphology) translocates to the mitochondria and destabilizes the lipid bilayer. The recruitment of enzymes involved in lipids modification, such as lysolipid transferases (e.g. tBid or Endophilin B1) would enable the formation of a lipid pore and further intramitochondrial remodelling by mitochondrial shaping proteins (e.g. Opa1 or Mitofilin), triggering the release of apoptogenic factors. (c) Proposed model for mitochondrial division process and fission induced MMP. Under normal conditions the fission machinery (Fis1, Drp1 and putative proteins implicated in membrane perturbation and lipids modifications) is recruited and assembled on local foci on the OMM. IMM and OMM are brought in close apposition to form a stalk intermediate. Under physiological division, Drp1 assembles into a ring and constricts the mitochondria until it divides. Alternatively failure of mitochondrial fission machinery could result in bilayer destabilization and lipid pore formation as previously shown. Apoptogenic factors would then be released in the cytosol.

An alternative model for MMP proposes that proapoptotic Bcl2 family members oligomerize and form channels within the OMM (Saito et al., 2000; Epand et al., 2002b). This view is supported by the similarity between these proteins and some pore-forming bacterial toxins. Following a death signal, the proapoptotic proteins Bax and Bak undergo a conformational change, and their N terminus becomes exposed. Bax, which is usually in the cytosol or loosely associated with mitochondria, translocates to the mitochondria, oligomerizes and inserts in the OMM causing MMP (Hsu et al., 1997; Desagher et al., 1999; Antonsson et al., 2000). These events can be promoted by tBid, the activated form of Bid, and inhibited by Bcl2 and Bcl-xL (Desagher et al., 1999; Eskes et al., 2000; Cheng et al., 2001).

Mitochondrial membrane permeabilization generally occurs in a concerted and rapid fashion, affecting most, if not all, mitochondria in the dying cell. The classic models mentioned above are far from being complete, and theories on the molecular mechanism are formulated, debated and modified on a regular basis (Lucken-Ardjomande and Martinou, 2005a). Recent investigations revealed that during apoptosis, the mitochondrial membrane properties are profoundly modified and the mitochondrial network undergoes extensive fragmentation (Desagher and Martinou, 2000; Cristea and Degli Esposti, 2004). Mitochondrial membrane permeabilization requires drastic membrane remodelling, and both lipids and components of the fission/fusion machinery may participate in this process.

Importance of lipids and membrane topology

The first insight into the role of membrane lipids in cell death came from the observation that phosphatidylserine (PS) became exposed at the cell surface during apoptosis (Fadok et al., 1992). Subsequently several studies revealed that apoptosis correlated with a progressive proliferation of intracellular membranes leading to the formation of apoptotic blebs. This was in contrast to the process of necrosis, as apoptotic cells disintegrate without leaking their intracellular contents out in the medium, preventing the inflammatory reactions that are induced by necrotic cells (Cristea and Degli Esposti, 2004). Modifications of lipid composition and bilayer curvature of the mitochondrial membranes seem to be crucial for permeabilization. Recent work has highlighted the importance of tBid, which is able to insert into specific lysolipids of the OMM (Esposti et al., 2001; Goonesinghe et al., 2005) and to promote a negative membrane curvature of synthetic lipidic membranes (Epand et al., 2002a). These changes presumably prepare the ground for insertion and oligomerization of Bax and Bak. tBid is also able to modify IMM topology, as incubation of mouse liver mitochondria with tBid induces a profound remodelling of the cristae compartment (Scorrano et al., 2002). Tomographic reconstructions revealed that the cristae become more interconnected and that their junctions widen concurring with enhanced mobilization of cytochrome c. The molecular events leading to cristae remodelling are not known. tBid could interact with proteins of the IMM or could directly affect lipids to modulate IMM topology. Indeed, tBid has been shown to bind to cardiolipin, a negatively charged mitochondrial phospholipid present in the IMM (Lutter et al., 2000; Liu et al., 2005). Recently, Kim et al. (2004) reported that binding to cardiolipin is required for tBid induced cristae remodelling and cytochrome c release. Yet these rearrangements might not always be required given the fact that Bid knockout mice do not display particular deficits in cell death and that Bid is not expressed in all cells in vivo (Krajewska et al., 2002). Other reports show that Bax destabilizes lipid bilayers (Basanez et al., 1999; Epand et al., 2003) and interacts with endophilin B1, a potential lysophosphatic acid acyltransferase (Modregger et al., 2003). However, a recent study suggests that the effect of endophilin B1 effect on membrane curvature could be an experimental artefact (Gallop et al., 2005).

Lipid bilayer bending and tilt of hydrocarbon chains can provide sufficient lateral tension to drive pore opening in a nonbilayer intermediate (Chernomordik and Kozlov, 2003). This implies that membrane permeabilization is not necessarily mediated by a proteinaceous pore, but rather results from the concerted actions of proteins and particular lipids. As illustrated in Figure 1, during apoptosis the formation of discontinuity in the OM can lead to nonspecific lipid pores. Thus Bax pore forming capacity could be ascribed to membrane perturbation rather than discrete channel activity (Cristea and Degli Esposti, 2004).

These observations shed a new light on the process of mitochondrial membrane permeabilization, suggesting that membrane topology is of direct relevance for apoptosis. In addition, several investigations in the field of mitochondrial dynamics further strengthened the intimate relationship that exists between organelle morphology and cell death.

Role of fission and fusion in apoptosis

As mentioned earlier, mitochondria undergo frequent fission and fusion events that regulate their morphology. During early stages of apoptosis, mitochondria undergo massive fragmentation and it has been proposed that changes in mitochondrial shape contribute to the mitochondrial membrane permeabilization, perhaps by setting free inter membrane space (IMS) proteins for release (Frank et al., 2001; Karbowski et al., 2002; Reed and Green, 2002; Lee et al., 2004). Both Drp1 and Fis1 have been implicated in this pathway (Perfettini et al., 2005). Fis1 overexpression not only induces fragmentation of the mitochondrial network, but also leads to the release of cytochrome c from the mitochondria caspase activation and cell death (James et al., 2003). Silencing of either Drp1 or Fis1 confers resistance to several apoptotic stimuli and impairs cytochrome c release suggesting that fission molecules are required for MMP to occur (Frank et al., 2001; Karbowski et al., 2002; Breckenridge et al., 2003; Lee et al., 2004; Germain et al., 2005). However, translocation of Bax to the mitochondria remains unaffected by fission inhibition, suggesting that fission molecules act downstream of Bax translocation. Mitochondrial fusion proteins also seem to protect cells from apoptosis, as their overexpression reduces cell death induced by several stimuli (Sugioka et al., 2004), and their knocking down sensitizes cells to apoptosis (Lee et al., 2004).

Several models exist to explain how mitochondrial fission could participate in MMP. Karbowski et al. (2004) reported that in addition to its role in apoptosis, endophilin B1 regulates mitochondria morphology. They also showed that upon apoptosis induction, Bax, Drp1 and endophilin B1 translocate to Mfn2-containing foci on the mitochondria (Karbowski et al., 2002, 2004). The model they propose is the following: rather than forming a ring around mitochondria, Drp1 would mediate vesicle scission from focal regions around the mitochondrial fission sites. This would allow the removal of membrane by shearing away lipid vesicles from OMM. One possible explanation for cytochrome c release would be that Bax inserts preferentially in these narrow curvature vesicles and permeabilize them. Endophilin B1 could promote Bax insertion into membranes through its membrane deformation activity (Youle and Karbowski, 2005). However, the study from Gallop et al. (2005) challenges this model as it shows that endophilins are devoid of lysophosphatidic acid acyl transferase activity.

Another model for the role of mitochondrial morphology in apoptosis postulates that mitochondrial shaping proteins could regulate cristae remodelling and favour MMP (Germain et al., 2005; Scorrano, 2005). As mentioned earlier, rearrangements of the cristae compartments is thought to increase the susceptibility to apoptosis by enhancing the ability of mitochondria to release cytochrome c (Scorrano et al., 2002). Indeed, it seems that two pools of cytochrome c exist: a minor soluble pool, which is present in the IMS and a major pool confined in the mitochondrial cristae and which binds cardiolipin (Bernardi and Azzone, 1981; Ott et al., 2002; Scorrano et al., 2002). Mitochondrial fission is thought to be necessary for the release of the membrane-attached pool of cytochrome c.

A role for fission and fusion molecules in this process is supported by two observations. Firstly, loss of Opa1 expression has been reported to trigger apoptosis by inducing such intramitochondrial remodelling (Olichon et al., 2003). Secondly, Bik, a BH3-only protein present on the ER, has been shown to induce Drp1-dependent remodelling of the cristae compartment (Germain et al., 2005).

However, there is still some controversy about the role of mitochondrial fission in apoptosis and recent work suggests that depending on the stimuli, fission can also interfere with MMP (Szabadkai et al., 2004). The biochemical and biophysical relationships between Bcl2 family members and the fission machinery remain to be clarified. The observation that these two classes of proteins are responsible for membrane perturbation and/or the release of proapoptotic factors does not mean that they act in a concerted fashion. Moreover, colocalization cannot account for functional interaction, as local lipid bilayer properties directly affect the recruitment, insertion and activity of proteins. It is, therefore, crucial to determine the contribution of lipids rearrangement to mitochondrial fragmentation both in healthy and apoptotic cells. In addition, further investigations are needed to determine what are the molecular differences between physiological and death-associated fission. Addressing these issues should help us understand how mitochondrial fragmentation participates in programmed cell death and possibly correlates with tumorigenesis.

Therapeutic perspectives

As impaired apoptosis plays a central role in oncogenesis, the search is accelerating for novel agents that engage the cell death machinery (Nicholson, 2000; Johnstone et al., 2002). Impaired apoptosis is also a significant impediment to cytotoxic therapy. The mutations that favour tumour development stifle the response to chemotherapy and radiation, and treatment might select more refractory clones. Nevertheless, most tumour cells still remain sensitive to some apoptotic stimuli, and most cytotoxic anticancer drugs currently in clinical use induce apoptosis of malignant cells (Don and Hogg, 2004). Therapeutic approaches that enhance apoptosis include targeting of Bcl-2 or the caspase inhibitors IAPs, or engaging the death receptor pathway, by, for example, ligating the receptors for TRAIL (Ashkenazi, 2002). Among the recent strategies, the ones attempting to target mitochondria and trigger MMP are very promising. With much debate surrounding the components of the PTP and its possible manipulation in cancer treatment, many chemotherapeutics efforts have focused on targeting the Bcl2 family (Oltersdorf et al., 2005). Other cytotoxic drugs acting on mitochondrial lipids have been shown to induce MMP in vitro and some are being tested in preclinical mouse models (Don and Hogg, 2004). Recent findings on the role of mitochondrial fission machinery in the MMP process may provide additional possibility for the management of neoplastic diseases.

Many scientists in the field of cancer research may have considered mitochondria only as a reservoir for stocking harmful molecules such as ROS and apoptotic proteins. However, several recent findings demonstrate that mitochondrial participation in oncogenesis is far from being restricted to apoptosis, and that mitochondrial energy production affects both cell proliferation and tumour progression.

Mitochondrial physiology in cell proliferation and tumour progression

Mitochondria produce ATP through oxidative phosphorylation

In many cell types, cellular ATP is produced primarily by the oxidation of glucose (Figure 2). This process can be divided into two major steps. The first, which takes place in the cytosol, is glycolysis. In the absence of oxygen, the end product of glycolysis, pyruvate, is converted to lactic acid through cytosolic fermentation. In the presence of oxygen, pyruvate is utilized by the second step, which takes place in the mitochondria, and includes the tricarboxylic acid (TCA) cycle and oxidative phosphorylation (OXPHOS). Further oxidation of pyruvate in the mitochondria leads to the production of reducing equivalents (NADH and FADH2). These compounds fuel the electron transport chain (ETC) located in the IMM. The ETC is composed of four multimeric complexes and couples the electron transport to the extrusion of protons across the IMM. This movement of protons creates an electrochemical gradient (Δψpψm+ΔpH) across the IMM. Most of the gradient is in the form of an electrical component, Δψm. The charge separation across the IMM generates free energy, which is utilized by the ATP synthase to synthesize ATP. As a toxic by-product, OXPHOS generates reactive oxygen species (ROS). ROS can act both as tumour initiators, by inducing mutations in proto-oncogenes and tumour-supressors, and tumour promoters through enhancement of cell proliferation (reviewed in Behrend et al., 2003; Wallace, 2005). Impairment of the ETC very often results in excessive ROS production and subsequent oxidative stress. It is, therefore, crucial for the cell to keep watch over mitochondrial respiration as it can turn into a hazardous weapon when it fails.

Figure 2

Mitochondrial shape changes according to energy state. (a) Schematic representation of glucose metabolism. In the cytosol, glucose is converted through glycolysis to pyruvate. Pyruvate is then either transformed to lactate via lactic fermentation, or transported into the mitochondrial matrix where it enters the tricarboxylic acid (TCA) cycle. Electrons enter the electron transport chain (ETC) via the NADH produced by the TCA cycle. Complexes I, III and IV couple electron transport to protons pumping. The proton gradient generated is then used by ATP synthase to form ATP. (b) Different metabolic strategies are associated with different mitochondrial morphology. In the case of the respirative phenotype, glucose is converted into pyruvate and further oxidized by mitochondrial respiration (TCA+ETC). The mitochondrial network appears interconnected and cristae compartment is enlarged. In the case of glycolytic phenotype, the majority of cytosolic glucose is converted into lactate, leading to an acidification of the extracellular environment. Mitochondria appear fragmented and undergo matrix expansion.

Mitochondria modulate their shape according to their bioenergetic activity

Substrate availability, and energetic state both modulate the mitochondrial network (Figure 2). In the 1960s Hackenbrock (1966, 1968a, 1968b) reported that depending on their respiration activity, the mitochondrial morphology changed from an orthodox to a condensed conformation. The condensed conformation, in which mitochondria display large intracristae spaces, was associated with state III respiration, when ADP is high and OXPHOS activated. The orthodox conformation, corresponding to an expanded matrix and a reduced intracristae space, was observed under state IV. This corresponds to low ADP and minimal O2 consumption. Further evidence that mitochondria remodel according to changes in OXPHOS activity came from the work of Rossignol et al. (2004) who showed that the mitochondrial network extended within cells and became more ramified and interconnected in the presence of galactose.

As previously mentioned, mitochondria shaping proteins seem to affect energy production. A direct connection between bioenergetics and mitochondrial fusion machinery has recently been established. When the balance is artificially tipped towards fission, an energy defect emerges. Disruption of Mitofusins and/or Opa1 results in severe cellular defects, including poor cell growth, widespread heterogeneity of mitochondrial membrane potential, and decreased cellular respiration (Chen et al., 2005). Importantly, the respiration defect turns out to be reversible upon reintroduction of the wild-type fusion proteins. Another recent study reports that downregulation of Mfn2 in L6E9 myotubes reduces both the activity of the TCA and the ETC, while it enhances glucose transport, glycolysis and lactic fermentation (Pich et al., 2005). Conversely, the overexpression of Mfn2 stimulates OXPHOS activity, suggesting that its role is not limited to the control of organelle shape. Mutations in Mfn2 have been ascribed to Charcot–Marie–Tooth type 2A neuropathy, and altered glucose oxidation could provide a mechanistic explanation for the aetiology of the disease (Zuchner et al., 2004). Silencing of Mitofilin also affects metabolic fluxes (John et al., 2005).

Thus, it appears that mitochondrial morphology is crucially linked to energy metabolism. Enhanced respiration correlates with an interconnected network and enlarged cristae compartment, whereas low OXPHOS activity and high glycolysis correlates with smaller mitochondria displaying reduced intracristae space. In an extreme pathological situation, overfragmented mitochondrial fail to run the ETC correctly, potentially contributing to oxidative stress. On the contrary, promoting fusion stimulates respiration. It is therefore fundamental to maintain balanced mitochondrial fission and fusion as its disruption could increase the susceptibility to develop metabolic defects and consequent diseases.

Altered energy metabolism is a hallmark of many types of cancer

A curious but common property of invasive cancers is altered glucose metabolism. Under aerobic conditions, glycolysis is inhibited (the so-called Pasteur effect) (Krebs, 1972) and normal mammalian cells rely mainly on the mitochondrial OXPHOS for their energy supply. However, cancer cells display a significant increase in glycolysis and lactate production even in the presence of oxygen. This phenomenon, termed aerobic glycolysis was first described in 1956 by the German biochemist and Nobel Laureate Otto Warburg (Warburg, 1956). He postulated that cancer resulted from impaired mitochondrial metabolism. Although it turned out that mitochondrial dysfunction is not the fundamental cause of cancer, aerobic glycolysis certainly confers a growth advantage to tumour cells (Gatenby and Gillies, 2004). Firstly, it enables cancer cells to adapt to hypoxic conditions as the premalignant lesion grows progressively further from the blood supply. Secondly, the glycolytic phenotype contributes to the acidification of tumour microenvironment, which facilitates tumour invasion (Gatenby and Gawlinski, 1996; Schornack and Gillies, 2003). This metabolic transition towards glycolysis is also seen in human cancer cell lines such as Hela cells, in which minor amounts of glucose are prone to the TCA cycle, the majority being converted to lactate (Reitzer et al., 1979).

Mitochondria certainly have a role to play in this metabolic shift, as their physiology is inextricably linked to energy metabolism. Several groups reported that tumour aggressiveness correlates with a low-mitochondrial respiratory chain activity, and that enhancement of OXPHOS can reduce tumour growth (Hervouet et al., 2005; Schulz et al., 2006). Furthermore, a number of mitochondrial enzyme deficiencies are linked to inherited neoplasia (Gottlieb and Tomlinson, 2005). These include succinate deshdrogenase (SDH) genes SDHB, SDHC, SDHD and fumarate hydratase (FH), which all participate in the TCA cycle. In addition SDH is a functional member (complex II) of the ETC. The mechanism linking failure of the TCA cycle to tumour development is not clear but two explanations have been suggested so far. The first hypothesis postulates that SDH and FH deficiencies increase the production of ROS (Messner and Imlay, 2002; Yankovskaya et al., 2003). An alternative model proposes that succinate, which accumulates due to TCA cycle impairment, acts as a signalling molecule and triggers the activation of hypoxia inducible factor 1α (HIF1α) (Selak et al., 2005). Hypoxia inducible factor 1α promotes adaptation of cells to low-oxygen consumption and activates the transcription of glycolytic genes (Firth et al., 1995; Semenza et al., 1996), stimulating aerobic glycolysis.

It seems thus that mitochondrial dysfunction favours the emergence of the glycolytic phenotype, and that after dwindling in importance for decades, Warburg's hypothesis is enjoying a resurrection. Although further studies are required and warranted, induction of mitochondrial respiration might be worth considering as a mechanism to restrict tumour growth. The energetic conversion towards aerobic glycolysis most probably requires remodelling of mitochondria. As OXPHOS activation seems to perturb malignant transformation, it would be interesting to see whether tumour metabolism is affected by the promotion of mitochondrial fusion.

Mitochondrial ATP production and cell cycle progression

Another common property of cancers is a high proliferation rate, which results both from cellular growth and cell division. Although the energetic requirements for cell cycle progression have not been investigated thoroughly, it is highly probable that a metabolic checkpoint exists. The status of cellular energy stores is sensed by the AMP-activated protein kinase (AMPK), which is activated upon ATP depletion (Hardie, 2005). AMP-activated protein kinase has been shown to inhibit mTOR and to coordinate G1-S transition with carbon source availability (Kimura et al., 2003; Jones et al., 2005). In mammals numerous studies demonstrated that the G1-phase is associated with an overall enhancement of mitochondrial function (Van den Bogert et al., 1988; Leprat et al., 1990) and that inhibiting mitochondrial function results in G1 arrest (van den Bogert et al., 1986, 1992; Heerdt et al., 1997; King and Radicchi-Mastroianni, 2002; Gemin et al., 2005). A recent study in Drosophila links a defect in mitochondrial energy production to the activation of AMPK (Mandal et al., 2005). Taken together this data suggests the existence of an energy checkpoint at G1–S transition controlled by the mitochondrial energy production.

Interestingly the G1–S transition also correlates with rearrangements of the mitochondrial network. Studies in both fibroblasts and osteosarcoma cells show that the mitochondria become filamentous and highly interconnected during the G1-phase while they appear fragmented in the S-phase (Barni et al., 1996; Margineantu et al., 2002). This is consistent with the observation that mitochondrial shape change according to energy state. During the G1-phase the cell displays elongated mitochondria correlating with increased OXPHOS activity. This presumably enables efficient energy production in the expectation of S-phase, which requires substantial amounts of ATP. At the end of the G1-phase the cell evaluates its energy stocks, and takes the decision either to pause and exit the cell cycle, or to proceed toward the S-phase. However, maintaining a high OXPHOS activity during DNA replication might be dangerous, because of the deleterious effects of ROS production. This could explain the decreased respiration activity and the fragmented mitochondrial morphology observed during the S-phase.

The proposal that according to energy requirements cells modify their metabolism by remodelling the mitochondrial network is challenging but still needs experimental proof. Preliminary results show that knockingdown components of the fission and fusion machinery not only affects ATP production but also impedes cell cycle progression ((Chen et al., 2005) and Parone P and Martinou JC unpublished data). Along with apoptosis resistance, hyper proliferation is a common property shared by many cancers, and cell cycle checkpoints are privileged targets of chemotherapy. Future work will certainly address the importance of mitochondrial network remodelling in energy production, and establish whether these morphological changes play a role in cell proliferation.

Conclusion and future perspectives

Over the past decades cancer research has generated a rich and complex body of knowledge, revealing a multitude of parameters that either cause or favour malignant transformation. Several features, however, emerge as essential requisites of tumourigenesis: apoptosis evasion, limitless cell proliferation, sustained angiogenesis and tissue invasion are crucial properties of cancers. Growing evidence show that mitochondria are profoundly altered in transformed cells and participate in these processes. The study of cancer-associated mitochondrial dysfunctions such as OXHOS impairment and apoptosis evasion has to take into account their structural context, including the influence of membrane topology on internal diffusion and compartimentation. Mitochondrial dynamics are gaining increasing interest as it becomes indubitable that organelle fission and fusion play a role in diverse biological processes. Here, we have discussed the importance of mitochondrial membrane remodelling in cell death and energy metabolism, and its potential effect on tumorigenesis. Elucidating the molecular events that control mitochondria dynamics will certainly give new insight into organelle function in both normal and malignant cells and will help to determine whether the changes in organelle shape are meaningful for mitochondrial contribution to oncogenesis.


  1. Antonsson B, Montessuit S, Lauper S, Eskes R, Martinou JC . (2000). Biochem J 345 (Part 2): 271–278.

  2. Ashkenazi A . (2002). Nat Rev Cancer 2: 420–430.

  3. Baines CP, Kaiser RA, Purcell NH, Blair NS, Osinska H, Hambleton MA et al. (2005). Nature 434: 658–662.

  4. Bakhshi A, Jensen JP, Goldman P, Wright JJ, McBride OW, Epstein AL et al. (1985). Cell 41: 899–906.

  5. Barni S, Sciola L, Spano A, Pippia P . (1996). Biotech Histochem 71: 66–70.

  6. Basanez G, Nechushtan A, Drozhinin O, Chanturiya A, Choe E, Tutt S et al. (1999). Proc Natl Acad Sci USA 96: 5492–5497.

  7. Basso E, Fante L, Fowlkes J, Petronilli V, Forte MA, Bernardi P . (2005). J Biol Chem 280: 18558–18561.

  8. Behrend L, Henderson G, Zwacka RM . (2003). Biochem Soc Trans 31: 1441–1444.

  9. Bernardi P, Azzone GF . (1981). J Biol Chem 256: 7187–7192.

  10. Breckenridge DG, Stojanovic M, Marcellus RC, Shore GC . (2003). J Cell Biol 160: 1115–1127.

  11. Chen H, Chan DC . (2005). Hum Mol Genet 14 (Spec No. 2): R283–R289.

  12. Chen H, Chomyn A, Chan DC . (2005). J Biol Chem 280: 26185–26192.

  13. Chen H, Detmer SA, Ewald AJ, Griffin EE, Fraser SE, Chan DC . (2003). J Cell Biol 160: 189–200.

  14. Cheng EH, Wei MC, Weiler S, Flavell RA, Mak TW, Lindsten T et al. (2001). Mol Cell 8: 705–711.

  15. Chernomordik LV, Kozlov MM . (2003). Annu Rev Biochem 72: 175–207.

  16. Cipolat S, Martins de Brito O, Dal Zilio B, Scorrano L . (2004). Proc Natl Acad Sci USA 101: 15927–15932.

  17. Crescenzi M, Seto M, Herzig GP, Weiss PD, Griffith RC, Korsmeyer SJ . (1988). Proc Natl Acad Sci USA 85: 4869–4873.

  18. Cristea IM, Degli Esposti M . (2004). Chem Phys Lipids 129: 133–160.

  19. Crompton M, Barksby E, Johnson N, Capano M . (2002). Biochimie 84: 143–152.

  20. Danino D, Moon KH, Hinshaw JE . (2004). J Struct Biol 147: 259–267.

  21. Desagher S, Martinou JC . (2000). Trends Cell Biol 10: 369–377.

  22. Desagher S, Osen-Sand A, Nichols A, Eskes R, Montessuit S, Lauper S et al. (1999). J Cell Biol 144: 891–901.

  23. Deshmukh M, Kuida K, Johnson Jr EM . (2000). J Cell Biol 150: 131–143.

  24. Don AS, Hogg PJ . (2004). Trends Mol Med 10: 372–378.

  25. Epand RF, Martinou JC, Fornallaz-Mulhauser M, Hughes DW, Epand RM . (2002a). J Biol Chem 277: 32632–32639.

  26. Epand RF, Martinou JC, Montessuit S, Epand RM . (2003). Biochemistry 42: 14576–14582.

  27. Epand RF, Martinou JC, Montessuit S, Epand RM, Yip CM . (2002b). Biochem Biophys Res Commun 298: 744–749.

  28. Eskes R, Desagher S, Antonsson B, Martinou JC . (2000). Mol Cell Biol 20: 929–935.

  29. Esposti MD, Erler JT, Hickman JA, Dive C . (2001). Mol Cell Biol 21: 7268–7276.

  30. Eura Y, Ishihara N, Yokota S, Mihara K . (2003). J Biochem (Tokyo) 134: 333–344.

  31. Fadok VA, Voelker DR, Campbell PA, Cohen JJ, Bratton DL, Henson PM . (1992). J Immunol 148: 2207–2216.

  32. Ferri KF, Kroemer G . (2001). Nat Cell Biol 3: E255–E263.

  33. Firth JD, Ebert BL, Ratcliffe PJ . (1995). J Biol Chem 270: 21021–21027.

  34. Fischer U, Janicke RU, Schulze-Osthoff K . (2003). Cell Death Differ 10: 76–100.

  35. Frank S, Gaume B, Bergmann-Leitner ES, Leitner WW, Robert EG, Catez F et al. (2001). Dev Cell 1: 515–525.

  36. Gallop JL, Butler PJ, McMahon HT . (2005). Nature 438: 675–678.

  37. Gatenby RA, Gawlinski ET . (1996). Cancer Res 56: 5745–5753.

  38. Gatenby RA, Gillies RJ . (2004). Nat Rev Cancer 4: 891–899.

  39. Gemin A, Sweet S, Preston TJ, Singh G . (2005). Biochem Biophys Res Commun 332: 1122–1132.

  40. Germain M, Mathai JP, McBride HM, Shore GC . (2005). Embo J 24: 1546–1556.

  41. Goonesinghe A, Mundy ES, Smith M, Khosravi-Far R, Martinou JC, Esposti MD . (2005). Biochem J 387: 109–118.

  42. Gottlieb E, Tomlinson IP . (2005). Nat Rev Cancer 5: 857–866.

  43. Green DR, Evan GI . (2002). Cancer Cell 1: 19–30.

  44. Griparic L, van der Wel NN, Orozco IJ, Peters PJ, van der Bliek AM . (2004). J Biol Chem 279: 18792–18798.

  45. Hackenbrock CR . (1966). J Cell Biol 30: 269–297.

  46. Hackenbrock CR . (1968a). Proc Natl Acad Sci USA 61: 598–605.

  47. Hackenbrock CR . (1968b). J Cell Biol 37: 345–369.

  48. Halestrap AP, McStay GP, Clarke SJ . (2002). Biochimie 84: 153–166.

  49. Hanahan D, Weinberg RA . (2000). Cell 100: 57–70.

  50. Hardie DG . (2005). Curr Opin Cell Biol 17: 167–173.

  51. Heerdt BG, Houston MA, Augenlicht LH . (1997). Cell Growth Differ 8: 523–532.

  52. Hervouet E, Demont J, Pecina P, Vojtiskova A, Houstek J, Simonnet H et al. (2005). Carcinogenesis 26: 531–539.

  53. Hsu YT, Wolter KG, Youle RJ . (1997). Proc Natl Acad Sci USA 94: 3668–3672.

  54. James DI, Parone PA, Mattenberger Y, Martinou JC . (2003). J Biol Chem 278: 36373–36379.

  55. John GB, Shang Y, Li L, Renken C, Mannella CA, Selker JM et al. (2005). Mol Biol Cell 16: 1543–1554.

  56. Johnstone RW, Ruefli AA, Lowe SW . (2002). Cell 108: 153–164.

  57. Jones RG, Plas DR, Kubek S, Buzzai M, Mu J, Xu Y et al. (2005). Mol Cell 18: 283–293.

  58. Karbowski M, Jeong SY, Youle RJ . (2004). J Cell Biol 166: 1027–1039.

  59. Karbowski M, Lee YJ, Gaume B, Jeong SY, Frank S, Nechushtan A et al. (2002). J Cell Biol 159: 931–938.

  60. Kerr JF, Wyllie AH, Currie AR . (1972). Br J Cancer 26: 239–257.

  61. Kim TH, Zhao Y, Ding WX, Shin JN, He X, Seo YW et al. (2004). Mol Biol Cell 15: 3061–3072.

  62. Kimura N, Tokunaga C, Dalal S, Richardson C, Yoshino K, Hara K et al. (2003). Genes Cells 8: 65–79.

  63. King MA, Radicchi-Mastroianni MA . (2002). Cytometry 49: 106–112.

  64. Kokoszka JE, Waymire KG, Levy SE, Sligh JE, Cai J, Jones DP et al. (2004). Nature 427: 461–465.

  65. Kozlovsky Y, Chernomordik LV, Kozlov MM . (2002). Biophys J 83: 2634–2651.

  66. Krajewska M, Mai JK, Zapata JM, Ashwell KW, Schendel SL, Reed JC et al. (2002). Cell Death Differ 9: 145–157.

  67. Krebs HA . (1972). Essays Biochem 8: 1–34.

  68. Lee YJ, Jeong SY, Karbowski M, Smith CL, Youle RJ . (2004). Mol Biol Cell 15: 5001–5011.

  69. Leprat P, Ratinaud MH, Maftah A, Petit JM, Julien R . (1990). Exp Cell Res 186: 130–137.

  70. Liu J, Durrant D, Yang HS, He Y, Whitby FG, Myszka DG et al. (2005). Biochem Biophys Res Commun 330: 865–870.

  71. Lucken-Ardjomande S, Martinou JC . (2005a). J Cell Sci 118: 473–483.

  72. Lucken-Ardjomande S, Martinou JC . (2005b). C R Biol 328: 616–631.

  73. Lutter M, Fang M, Luo X, Nishijima M, Xie X, Wang X . (2000). Nat Cell Biol 2: 754–761.

  74. Mandal S, Guptan P, Owusu-Ansah E, Banerjee U . (2005). Dev Cell 9: 843–854.

  75. Margineantu DH, Gregory Cox W, Sundell L, Sherwood SW, Beechem JM, Capaldi RA . (2002). Mitochondrion 1: 425–435.

  76. Martinou I, Desagher S, Eskes R, Antonsson B, Andre E, Fakan S et al. (1999). J Cell Biol 144: 883–889.

  77. Martinou JC, Green DR . (2001). Nat Rev Mol Cell Biol 2: 63–67.

  78. Mattenberger Y, James DI, Martinou JC . (2003). FEBS Lett 538: 53–59.

  79. Messner KR, Imlay JA . (2002). J Biol Chem 277: 42563–42571.

  80. Modregger J, Schmidt AA, Ritter B, Huttner WB, Plomann M . (2003). J Biol Chem 278: 4160–4167.

  81. Muschen M, Moers C, Warskulat U, Even J, Niederacher D, Beckmann MW . (2000). Immunology 99: 69–77.

  82. Nakagawa T, Shimizu S, Watanabe T, Yamaguchi O, Otsu K, Yamagata H et al. (2005). Nature 434: 652–658.

  83. Nicholson DW . (1999). Cell Death Differ 6: 1028–1042.

  84. Nicholson DW . (2000). Nature 407: 810–816.

  85. Niemann A, Ruegg M, La Padula V, Schenone A, Suter U . (2005). J Cell Biol 170: 1067–1078.

  86. Nunez G, London L, Hockenbery D, Alexander M, McKearn JP, Korsmeyer SJ . (1990). J Immunol 144: 3602–3610.

  87. Okamoto K, Shaw JM . (2005). Annu Rev Genet 39: 503–536.

  88. Olichon A, Baricault L, Gas N, Guillou E, Valette A, Belenguer P et al. (2003). J Biol Chem 278: 7743–7746.

  89. Oltersdorf T, Elmore SW, Shoemaker AR, Armstrong RC, Augeri DJ, Belli BA et al. (2005). Nature 435: 677–681.

  90. Ott M, Robertson JD, Gogvadze V, Zhivotovsky B, Orrenius S . (2002). Proc Natl Acad Sci USA 99: 1259–1263.

  91. Patterson SD, Spahr CS, Daugas E, Susin SA, Irinopoulou T, Koehler C et al. (2000). Cell Death Differ 7: 137–144.

  92. Perfettini JL, Roumier T, Kroemer G . (2005). Trends Cell Biol 15: 179–183.

  93. Petros AM, Olejniczak ET, Fesik SW . (2004). Biochim Biophys Acta 1644: 83–94.

  94. Pich S, Bach D, Briones P, Liesa M, Camps M, Testar X et al. (2005). Hum Mol Genet 14: 1405–1415.

  95. Reed JC, Green DR . (2002). Mol Cell 9: 1–3.

  96. Reitzer LJ, Wice BM, Kennell D . (1979). J Biol Chem 254: 2669–2676.

  97. Rossignol R, Gilkerson R, Aggeler R, Yamagata K, Remington SJ, Capaldi RA . (2004). Cancer Res 64: 985–993.

  98. Saito M, Korsmeyer SJ, Schlesinger PH . (2000). Nat Cell Biol 2: 553–555.

  99. Santel A, Fuller MT . (2001). J Cell Sci 114: 867–874.

  100. Schinzel A, Kaufmann T, Borner C . (2004). Biochim Biophys Acta 1644: 95–105.

  101. Schornack PA, Gillies RJ . (2003). Neoplasia 5: 135–145.

  102. Schulz TJ, Thierbach R, Voigt A, Drewes G, Mietzner B, Steinberg P et al. (2006). J Biol Chem 281: 977–981.

  103. Scorrano L . (2005). J Bioenerg Biomembr 37: 165–170.

  104. Scorrano L, Ashiya M, Buttle K, Weiler S, Oakes SA, Mannella CA et al. (2002). Dev Cell 2: 55–67.

  105. Selak MA, Armour SM, MacKenzie ED, Boulahbel H, Watson DG, Mansfield KD et al. (2005). Cancer Cell 7: 77–85.

  106. Semenza GL, Jiang BH, Leung SW, Passantino R, Concordet JP, Maire P et al. (1996). J Biol Chem 271: 32529–32537.

  107. Slee EA, O'Connor DJ, Lu X . (2004). Oncogene 23: 2809–2818.

  108. Smirnova E, Griparic L, Shurland D-L, van der Bliek AM . (2001). Mol Biol Cell 12: 2245–2256.

  109. Stojanovski D, Koutsopoulos OS, Okamoto K, Ryan MT . (2004). J Cell Sci 117: 1201–1210.

  110. Sugioka R, Shimizu S, Tsujimoto Y . (2004). J Biol Chem 279: 52726–52734.

  111. Szabadkai G, Simoni AM, Chami M, Wieckowski MR, Youle RJ, Rizzuto R . (2004). Mol Cell 16: 59–68.

  112. Teitz T, Wei T, Valentine MB, Vanin EF, Grenet J, Valentine VA et al. (2000). Nat Med 6: 529–535.

  113. Testa JR, Bellacosa A . (2001). Proc Natl Acad Sci USA 98: 10983–10985.

  114. Tieu Q, Nunnari J . (2000). J Cell Biol 151: 353–366.

  115. Tolkovsky AM, Xue L, Fletcher GC, Borutaite V . (2002). Biochimie 84: 233–240.

  116. Tondera D, Santel A, Schwarzer R, Dames S, Giese K, Klippel A et al. (2004). J Biol Chem 279: 31544–31555.

  117. Tsujimoto Y, Finger LR, Yunis J, Nowell PC, Croce CM . (1984). Science 226: 1097–1099.

  118. Van den Bogert C, Muus P, Haanen C, Pennings A, Melis TE, Kroon AM . (1988). Exp Cell Res 178: 143–153.

  119. van den Bogert C, Spelbrink JN, Dekker HL . (1992). J Cell Physiol 152: 632–638.

  120. van den Bogert C, van Kernebeek G, de Leij L, Kroon AM . (1986). Cancer Lett 32: 41–51.

  121. Vaux DL, Cory S, Adams JM . (1988). Nature 335: 440–442.

  122. Wallace DC . (2005). Annu Rev Genet 39: 359–407.

  123. Warburg O . (1956). Science 123: 309–314.

  124. Waterhouse NJ, Ricci JE, Green DR . (2002). Biochimie 84: 113–121.

  125. Yang L, Cao Z, Yan H, Wood WC . (2003). Cancer Res 63: 6815–6824.

  126. Yankovskaya V, Horsefield R, Tornroth S, Luna-Chavez C, Miyoshi H, Leger C et al. (2003). Science 299: 700–704.

  127. Yee KS, Vousden KH . (2005). Carcinogenesis 26: 1317–1322.

  128. Youle RJ, Karbowski M . (2005). Nat Rev Mol Cell Biol 6: 657–663.

  129. Yu T, Fox RJ, Burwell LS, Yoon Y . (2005). J Cell Sci 118: 4141–4151.

  130. Zamzami N, Kroemer G . (2001). Nat Rev Mol Cell Biol 2: 67–71.

  131. Zuchner S, Mersiyanova IV, Muglia M, Bissar-Tadmouri N, Rochelle J, Dadali EL et al. (2004). Nat Genet 36: 449–451.

Download references


We thank D James for careful reading of the manuscript and helpful comments and C Frezza for helping with figures. We thank the Swiss National Foundation (Grant 3100A0-109419/1) for support. EA is currently supported by a fellowship from MFSP CDA 0025/04 awarded to Luca Scorrano, VIMM, Italy.

Author information

Correspondence to J C Martinou.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Alirol, E., Martinou, J. Mitochondria and cancer: is there a morphological connection?. Oncogene 25, 4706–4716 (2006).

Download citation


  • mitochondrial morphology
  • apoptosis
  • energy metabolism
  • cancer

Further reading