Lipids and Cellular Senescence

Taken From: Lipid (per) oxidation in mitochondria: an emerging target in the ageing process?

Mitochondrial dysfunction, increased lipid peroxidation and altered catabolism will affect the cellular lipidome during cell senescence. Alterations in lipid metabolism and the generation of oxidised lipids may be beneficial for the senescent program during the early stages of senescence induction, possibly through modulating inflammatory and immune responses (Lawrence et al. 2002; van Diepen et al. 2013; Yaqoob 2003). However, if senescent cells persist in tissues, changes in lipid composition can result in cell dysfunction, altered rates of fatty acid oxidation that can induce inflammation and increased lipid peroxidation that can promote damage to neighbouring cells. These factors may contribute to ageing and age-related diseases. Research focused on altered lipid metabolism during cellular senescence, particularly regarding mitochondrial lipids, is in its infancy. However, in recent years, several studies have made progress in evaluating the senescent lipidome of fibroblasts.

One group investigated the alterations in a number of metabolites associated with the extracellular metabolome of fibroblasts induced to senesce via proliferative exhaustion or via γ-irradiation (James et al. 2015). They reported that a number of fatty acids and their precursors such as eicosapentaenoate, malonate, 7-alpha-hydroxy-3-oxo-4-cholestenoate and 1-stearoylglycerophosphoinositol were elevated during fibroblast senescence when compared with proliferating and quiescent cells, whereas linoleate, dihomo-linoleate, 10-heptadecenoate were depleted. Also amongst the secretory lipidome from senescent fibroblasts was an accumulation of monohydroxy fatty acids (2-hydroxypalmitate, 2-hydroxystearate, 3-hydroxydecanoate, 3-hydroxyoctanoate) and a phospholipid catabolite (glycerophosphorylcholine). It was suggested that whilst some of these changes may be due to oxidative stress, other observed increases may be a response to increased biomass commonly observed amongst senescent cells.

Maeda et al. (2009) investigated the regulation of fatty acid synthesis and ∆9-desaturation during cell senescence in human fibroblasts (Maeda et al. 2009). They found that the levels of fatty acid synthase and stearoyl-CoA desaturase-1 were decreased in senescent fibroblasts compared to proliferating fibroblasts, consequently leading to a decrease in monounsaturated fatty acids. In addition, reduced de novo synthesis of phospholipids with an associated increase in the formation of cholesterol in senescent cells was also observed and exogenous fatty acids were shown to be preferentially incorporated into the triacylglycerol pool of senescent cells.

In another study, the metabolic alterations associated with oncogene-induced senescence (OIS), using Ras-induced senescent human fibroblasts as a model were investigated (Quijano et al. 2012). Through the profiling of ~300 different intracellular metabolites, these authors showed that cells that have undergone OIS develop a metabolic signature which is distinct from cells which have undergone replicative senescence in response to extended in vitro cell culture. In the latter, a switch towards glycolysis has been observed that precedes the onset of senescence (Bittles and Harper 1984). In OIS, an increase in certain intracellular long chain fatty acids, including eicosanoate, dihomo-linoleate, mead acid and docosadienoate were observed. This altered metabolome was shown to associate with a decline in lipid synthesis and increases in fatty acid oxidation. Interestingly, the pro-inflammatory activity of the senescent secretome was reduced by inhibition of carnitine palmitoyltransferase 1, the rate limiting step in mitochondrial fatty acid oxidation, suggesting that alterations in lipid metabolism during OIS may play a role in regulating the pro-inflammatory senescent secretome. Although the mechanism underlying the increase in fatty acid levels during OIS were not fully explored, it may be due to promyelocytic leukemia (PML) activation of the fatty acid oxidation pathway through PPAR signalling (Aird and Zhang 2014). The differences between replicative senescence and OIS are intriguing; they may relate to the physiological need in preventing cancer to switch away from glycolysis as a rapid source of energy that is harnessed by cancer cells to enable them to proliferate rapidly versus the increasing insulin resistance that is seen in ageing and which associates with impaired oxidative metabolism (Burkart et al. 2016). However, while this and other studies have indicated an increase in glucose uptake during OIS, a number of other studies have observed either no change or a significant decrease in glucose uptake. This may relate to the timing of senescence induction, the cell type or the oncogene responsible.

A further study compared global lipid profiles and associated mRNA levels of proliferating and replicative senescent BJ fibroblasts; 19 specific polyunsaturated triacylglycerol species were identified as undergoing significant changes in lipid composition during cell senescence (Lizardo et al. 2017). In addition, significant changes in the expression of genes involved in specific lipid-related pathways, including glycerolipid metabolism, glycerophospholipid metabolism, unsaturated fatty acid synthesis and sphingolipid metabolism were observed during cell senescence. Based on these lipidomic and transcriptomic analysis, the authors postulated that activation of CD36-mediated fatty acid uptake and alteration to glycerolipid biosynthesis may contribute to the accumulation of triacylglycerols during cell senescence. It was suggested that these changes may be a mechanism to prevent lipotoxicity during elevated oxidative stress conditions during cell senescence.

In addition to an altered lipidome during cellular senescence, elevated ROS, likely from uncoupled mitochondria, can promote lipid peroxidation which potentiates cellular damage at distant sites. For example, stable aldehydes can diffuse from their site of generation and form adducts at distant locations, thereby propagating the responses and injury initiated by ROS (Ramana et al. 2013), including the induction of cell senescence in neighbouring cells. Flor and Kron observed an accumulation of lipid-derived aldehydes such as 4-hydroxy-2-nonenal (4-HNE) during accelerated senescence (Flor and Kron 2016). Whereas, the treatment of cells with either 4-HNE or low (5 Gy) γ-irradiation only generated low levels of cell senescence, combining both 4-HNE and 5 Gy γ-irradiation significantly elevated the senescence response. Furthermore, the use of the aldehyde-sequestering drug hydralazine blocked cell senescence induction by 25 Gy and etoposide treatment, demonstrating the potential importance of lipid peroxidation during therapy-induced senescence (Flor et al. 2016). Despite the highly damaging and pro-ageing potential of senescence-derived lipid peroxidation, little research has been conducted in this area and this requires further study.

Research on cell senescence has primarily been undertaken on fibroblasts and more research is required to explore whether the same phenomena are observed in cell-types linked to age-related disease such as in senescent adipocytes, pancreatic beta cells, renal proximal tubular epithelial cells and vascular endothelial cells. Whilst different types of senescent cells may share similarities in lipid metabolism, there may also be differences that are cell type-dependent or due to the mechanism of senescence induction and these require further study to better assess the role of altered lipid metabolism during ageing and disease. Finally, an important question to contemplate is whether the alterations in ROS, lipid metabolism and mitochondrial lipids observed during ageing and diseases are due solely to the presence of senescent cells or whether lipidomic changes can occur in absence of senescent cell accumulation.

Post-mitotic tissue

Damage to proteins

Damage from ROS is also thought to have an age-related impact on proteins. Such damage is thought to result in amino acid modifications, fragmentation of peptide chains, altered electrical charge and protein aggregation (Davies, 1987). Since the structure of proteins is pivotal for performing its functions then any structural changes would therefore result in an impairment of function. As discussed previously, the turn-over rate of proteins is an important factor in determining whether any protein damage is going to have a significant affect. Since the majority of intra-cellular proteins have high turn-over rates, damage proteins are not going to persist long enough to cause any problems. Therefore, long-lived proteins which may be affected by AGE formation may also be affected by oxidative damage. Examination of extracellular tissues in the lens, skin collagen and articular cartilage (low turnover proteins) of humans ranging in age from infancy to 80 years showed an increase in oxidative markers with age (Linton et al, 2001). The same study also looked at intracellular proteins with high turnover rates and found no evidence to suggest that intracellular proteins accumulate oxidative damage with age.

Damage to Lipids

Lipids which make up the membranes of cells are also potential targets for ROS which may consequently result in biologically significant alterations to membrane proteins. ROS are thought to attack membrane phospholipids and act on unsaturated fatty acids to produce lipid peroxidation. The consequent of this may be alterations in membrane fluidity, increased permeability and loss of membrane integrity. Experimental evidence suggests that altered membrane fluidity might affect permeability, transport systems, receptor functions or enzyme activities (Stark, 2005). The functionality of proteins in the membrane is critically dependant on membrane fluidity, especially when proteins have to collide with other molecules to exert their effects (such as G-proteins). This is seen in many receptor mediated pathways. For example, cardiac membranes from rats with cirrhotic cardiomyopathy are rigid and associated with diminished cAMP production. When the fluidity of these membranes are restored to control values cAMP production was significantly increased (Ma et al, 1997).

Biological impact of ROS

The above discussion on ageing of post-mitotic cells reviews the mechanisms thought to result in an overall increase in cellular damage by ROS, but little evidence is provided for the age-related biological consequence of such damage. If accumulative damage from ROS is an ageing mechanism of post-mitotic tissue, then such damage may result in three possible outcomes:

1) Post-mitotic cells become damaged and are subsequently removed. Post-mitotic-cells cannot be replaced, thereby resulting in a decrease in overall cell number and a decrease in tissue function.

2) Damage to post-mitotic cells results in an impairment of cellular function but such damage is not extensive enough for the removal of cells and so they persist.

3) Cells are removed due to damage and are subsequently replaced (i.e. by mitotic cells). However, the ageing mechanism in this case may therefore be a decline in the capacity to replace cells due to an ageing mechanism specific for mitotic tissues which will be discussed later. This may result in an observation similar to point 1).

The literature was reviewed to identify whether these three possible outcomes are present in ageing of post-mitotic tissues. One study observed a substantial increase in the amount of DNA single strand breaks in hippocampal pyramidal and granule cells as well as cerebellar granule cells but not in cerebellar Purkinje cells in the brain of ageing rats (Rutten et al, 2007). However, a reverse pattern was found for age-related reductions in total numbers of neurons. Cerebellar Purkinje cells were found to be significantly reduced during ageing (point 1 above) whereas the total number of hippocampal pyramidal and granule cells as well as cerebellar granule cells were not (points 2 or 3 above). This may seem as a confusing result since those cells which have undergone the most DNA damage may be expected to decline in numbers. This result may be explained if there are processes in place that replace damaged pyramidal and granule cells but do not, or cannot, replace damaged Purkinje cells. Cells that may be more prone to damage may have a higher potential capacity to be replaced than cells that are less prone to damage. It is also possible that the damage inflicted on hippocampal pyramidal and granule cells is not severe enough to warrant their removal. These findings suggest that if ROS does cause detrimental damage leading to an ageing phenotype, such damage is not tissue specific damage, but rather, cell type specific. All post-mitotic cells cannot therefore be treated equally.

A decline in cell numbers may not be related to damage from ROS but instead due to the loss or dysfunction of mitotic cells such as glial cells and astrocytes which are known to provide support and protection for neurons. Discussed later in more detail in ageing of mitotic tissue, mitotic cells can undergo an irreversible cell cycle growth-arrest known as replicative senescence. Senescent cells display a radically altered phenotype with potentially detrimental consequences on other cells if they accumulate in tissue. For example, a recent study has demonstrated that alterations in astrocyte function with ageing may not only affect its neuroprotective capacity, but may also contribute to neuronal injury in age-related neurodegenerative processes (Pertusa et al, 2007). Also, microglial, cells involved in immunological surveillance and neuroprotection have been shown to be subject to replicative senescence and it has been suggested that the dysfunction of these cells may contribute to the development of neurodegenerative disease by diminishing glial neuroprotection (Streit, 2006).

Myocytes, also known as muscle fibres are post-mitotic cells found in skeletal, smooth and cardiac muscle. Studies have shown that ageing muscle results in a decline in myocyte numbers, signs of atrophy and increased susceptibility to contraction-induced injury (Alnaqeeb and Goldspink 1986, Musaro and Rosenthal, 1999, McArdle et al 2002,). When myocytes become damaged and need to be repaired, mitotic satellite cells are able to differentiate and fuse to augment existing muscle fibres and to form new fibres. Decline in myocyte numbers and increases in damage with age may therefore, at least in part, be due to a reduction in number or the impairment of satellite cells (point 3). Such a reduction in satellite cell numbers would result in a decrease in cellular maintenance and an increase in cellular damage, possibly leading to the removal of the cll altogether. One study has found that the abundance of satellite cells does appear to decline with age, however, the myogenic potential of these cells does not diminish with age (Shefer et al, 2006).