Bioenergetics of Human Cancer Cells and Normal Cells 2 during Proliferation and Differentiation

6 Cancer cells are known to have different metabolic properties than normal cells, particularly their tendency to 7 undergo glycolysis even under aerobic favoring conditions. This has created interest in how mitochondrial 8 function in tumor cells may differfrom that in normal cells. Using human malignant cells(SW-620, PC-3, HT9 1080, SK-MEL, HL-60, K-562 and MOLT-3), human fibroblast (CCL-153) and human T Cells,we 10 investigated three key parameters that have been typicallyto describe mitochondrial function: cellular ATP 11 production, mitochondrial potential and cellular cardiolipin levels. On average, tumor cancer cells had more 12 ATP production and greater mitochondrial potentials. For example, ATP levels in malignant cells ranged 13 from 20 to 69 μmole/106 cells, with a cancer cell average of 40 ± 18 μmole/106 cells. For normal cells, the 14 ATP level range went from 9 to 24 μmole/106 cells, for an average of 15 ± 11 μmole/106 cells. 15 Mitochondrial potentials tended to be three times higher in cancer cells, perhaps because overall 16 mitochondrial mass (as measured by relative cardiolipin levels) were twice as high in cancer cells. Higher 17 mitochondrial masses are consistent with proliferation. Proliferating cells in general showed higher 18 mitochondrial function compared to quiescent cells (confluent monolayers), and HL-60 cells showed 19 reductions in all three mitochondrial parameters measured here when the cells were exposed to the 20 differentiating agent TPA. The effects of ATP production inhibitors CCCP and oligomycin on mitochondrial 21 function in normal and cancer cells were also compared. In general, in these experiments, cancer cell 22 mitochondrial inhibition with these agents produced a decrease ATP levels by 30-40% while in normal cells 23 ATP production was reduced by 60%. These results provideevidence of a mitochondrial dysfunction in 24 cancer cells. Cancer cells appear to better withstand interference with ATP synthesis in mitochondria since 25 they rely mainly on glycolysis as an energy producing mechanism. 26


Introduction 30
Research into the energy metabolism of cancer cells began in the early 20 th century withOtto Warburg,31 whoobserved that tumor tissues appear defective in respiration and have abnormally high rates of aerobic 32 glycolysis [1]. This led Warburg to propose that cancer arouse as a result of mitochondrial injury [2]. Since 33 then several cancer cell metabolism and mitochondrial function has been subject to extensive study. Two of 34 the most well-known and acceptedfeatures of tumor cell metabolism are the "Crabtree effect" [3] and the 35 "Pasteur effect" [4]. The former refers to inhibition of cancer cell respiration by elevated glucose 36 concentrations, while the latter refers to inhibition of glycolysis by elevated oxygen concentration. 37 Presumably, the Crabtree effect arises due to competition between glycolysis and oxidative phosphorylation 38 for Pi and ADP [5]. It has also been observed in some cancer cells that the consumption rateof one nutrient 39 (oxygen or glucose) increases when concentrationof the other nutrient is reduced, suggesting an ability of 40 cancer cells to adjust their metabolism based on micro-environment [6].These observations showthat cancer 41 has a very relevant metabolic component.

43
A variety of abnormalities in cancer cell mitochondrial structure and function have been reported [15][16][17][18][19][20][21][22][23][24][25][26][27]. 44 These peculiarities in glucose metabolism may be linked to differences between the mitochondria of cancer 45 cells and those of normal cells . These include increases in, and alteration of, consumption of metabolic intermediates toward anabolic reactions, and concomitantly less conversion of 56 pyruvate to oxaloacetate, leaning toward an augmented formation of lactic acid thus indicating an increase 57 dependency on glycolysis as energy mechanism. 58 59 In the present study, we examine differences between cancer cells and normal cells in three parameters related 60 to mitochondria: ATP production, cardiolipin concentration, and mitochondrial potentials. Moreover, we 61 examine how, in cancer cells and normal cells, these parameters are affected by ATP synthesis inhibitors 62 and,in case of one cancer cell type,chemicallyinduced differentiation. Our results support the idea that 63 mitochondria in cancer cellsrely more in glycolysis as their main energy production mechanism. 64

110
Three key variables were assessed in this study, mitochondrial potential, mitochondrial mass and cellular ATP 111 concentration. Results for several tumor and normal cell types are given in Table 1. There was some variation  112  in values with cell type, but there was a general pattern of the cancer cells having higher ATP levels, greater  113 mitochondrial mass levels (as determined by cardiolipin levels) and higher mitochondrial potentials. This 114 confirms the hypothesis that cancer cell mitochondria have different properties than normal cell 115 mitochondria, consistent with the differences in cancer cell metabolism described in the introduction. Some, 116 but not all, of the variation in mitochondrial potential and ATP production can be explained by differences in 117 mitochondrial mass, as shown in Figure 1(a). According to these data, levels of ATP show a statistically 118 significant correlation with cardiolipin levels in normal cells (r=0.8) and relation with ATP in cancer cells with 119 lower measured ATP for higher levels of cardiolipin. When mitochondrial potential or ATP levels are 120 normalized with mitochondrial mass (dividing by cardiolipin), they are roughly thirty percent higher in cancer 121 cells than in normal cells. This suggests that the larger mitochondrial mass in cancer cells may account in part 122 for their increased ATP production and mitochondrial potentials, although it should be noted that cardiolipin 123 levels may be an imperfect corollary to mitochondrial mass if cardiolipin concentrations vary significantly 124 from one cell type to another. As expected, ATP production rates are highly correlated with mitochondrial 125 potentials ( Figure 1b)  We next examined how tumor cells and normal cells are affected by mitochondrial inhibitors oligomycin and 155 carbonyl cyanide m-chlorophenylhydrazine (CCCP). Fluorescence emission spectra for K-562 leukemia cells 156 loaded with JC-1 dye, as shown in Figure 2, yield two peaks, one at roughly 535 nm and another at 595 nm. 157 Exposure of cells to oligomycin, an ATP-synthase inhibitor that increases mitochondrial potential via buildup 158 of hydrogen ions, increases emission at 595 nm. In contrast, exposure to CCCP, an ionophore that reduces 159 mitochondrial potential, decreasesemission at 595 nm. Note that neither inhibitor prevents the electron 160 transport chain from operating; they simply undo its potential building work (CCCP) or prevent the proton 161 gradients generated from being used to produce ATP (oligomycin). We examined the effects of one or both 162 inhibitors on the mitochondrial parameters described above. with there is a much greater effect on the fibroblasts (normal cells) than on the melanoma (malignant) cells. 181 We were able to confirm this trend for a variety of tumor and normal cell types. The results for T cells and 182 two leukemia cell lines are shown in Figure 3 For T-lymphocytes, ATP levels reduced to 60% of their control levels, while the corresponding ATP level 190 reductions for HL-60, andMOLT tumor cells were 40% of control and 30 % of control.It appears that cancer 191 cells are able to do a better job of maintaining ATP production in the midst of mitochondrial damage since 192 they mainly rely on glycolysis.

193
The HL-60 leukemia model allows us to examine the effects of differentiation on mitochondrial metrics.

194
This leukemia cell line normally grows in suspension in vitro. However, in the presence of 8 nM to 64 nM of 195 12-O-tetradecanoylphorbol 13-acetate (TPA), these cells will attach to plastic substrate, stop dividing, and 196 undergo morphological changes. They also show increased expression of CD11a, consistent with 197 differentiation to a monocyte or macrophage phenotype. Figure 4 shows the effects of TPA induced 198 differentiation on mitochondrial potential (A) and ATP production (B). Undifferentiated HL-60 cells 199 (Control) had significantly higher mitochondria potentials and cellular ATP levels. Again, part of this can be 200 explained by changes in mitochondrial masses, which in transformed HL-60 cells were twice those in TPA 201 differentiated HL-60 cells. We also examined the effects of oligomycin and CCCP on these cells, but the 202 inhibitors were not significant compared to the effects of TPA. are also increased in proliferating cells, relative to confluent monolayers and in undifferentiated cancer cells, 252 relative to differentiated cells of the same type. The increased mitochondrial mass and activity in cancer cells 253 is also manifest in their ability to more closely maintain normal ATP levels in the face of the ATP synthase 254 inhibitor oligomycin and the ionophore CCCP.A correlation analysis between ATP production and 255 cardiolipin between normal and malignant cells, reveals that there is a high correlation (R=0.8) between ATP 256 in and cardiolipin in the normal cells, while this correlation in malignant cells is relativelypoor 257 (R=0.33) (Figure 1a).

259
Discussion 260 For about a century it has been known that one of the most common properties of cancer cells is their ability 261 to utilize and catabolize glucose at high rate. This metabolic alteration of cancer cells called the ''Warburg pure uncoupler that acts as ionophore, completely dissipating the chemiosmotic gradient, but leaving the 279 electrotransport system uninhibited, resulted in a decrease mitochondrial potential. The addition of CCCP 280 with the antibiotic oligomycin or oligomycin alone, that acts by binding ATP synthase in such a way that 281 blocks the proton channel, the ETC runs but no ATP synthesis occurs, the inhibition of ATP production was 282 60%-70% in normal cells in comparison 30%-40% in cancer cells showing that energy production in cancer 283 cells is mainly from glycolysis. 284 The same results were found for transformed and differentiated leukemia cells HL-60. To find if the shift to a glycolytic pathway is increased in cancer cells because glycolysis is required for 291 cellular proliferation, we compared the levels of ATP and mitochondrial potential for normal fibroblast cells 292 (CCD153) at the stages of confluence and proliferation. According to our data there was increase in the 293 mitochondrial potential, mass and energy production in proliferative cells in comparison to confluent cells. It 294 is interesting that the ratio of ATP and mitochondrial potential in confluent cells as compared to proliferative 295 cells was the approximately the same as for cancer and normal cells. We also showed that the growth of 296 tumorigenic and non-tumorigenic cells in typical cell culture media increase cardiolipin, the signature 297 phospholipid of the inner mitochondrial membrane. According to our data the amount of these proteins 298 correlated with mitochondrial potential (R=0.5). 299 Comparison of the levels of total mitochondrial potential (accumulation of JC-1 dye) and mass in 300 transformed and normal cells showed the increased levels of these parameters in transformed cells that can 301 be, probably, explained by proliferation of the cells. However, our data showed that the ratio of 302 mitochondrial potential to mass was not statistically significant in malignant cells (leukemia cells and 303 transformed fibroblasts) different from normal cells (T cells and fibroblasts CCD-153). 304 We suggest that increased cardiolipin and the increased level of accumulation of JC-1aggregates in cancer cells 305 can be explained by the increased number of aberrantmitochondria in proliferative cells (42).Also in Fig1(b) 306 in whichATP production rates highly correlated with mitochondrial potentials with higher measures of ATP 307 in normal cells in comparison with cancer cells at the same value of mitochondrial potential supports the 308 same principle. 309 Several studies rule out the possibility that aerobic glycolysis is unique to cancer cells or that the Warburg 310 effect only develops when oxidative capacity is damaged [43]. Indeed, many highly proliferative cancer cell 311 lines that have been carefully studied do not seem to have defects in oxidative metabolism [43][44][45]. An 312 explanation to these findings is that these studies have not considered mitochondrial substrate level 313 phosphorylation that could give an impression that respiration is active when is not. It could be considered a 314 form of "pseudo-respiration". This "pseudo-respiration" can also be achieved by Tri-carboxylic acid (TCA ) 315 cycle metabolism of Glutamine, which occurs in the mitochondria and not in the cytoplasm giving the false 316 impression of an active oxidative metabolism (17). 317 According to the review [41], to synthesize lipids, proteins, and nucleic acids, cells use precursors derived 318 from TCA cycle intermediates and a key role of the TCA cycle in proliferating cells is to act as a hub for 319 biosynthesis. Nevertheless these can be provided by glutamine metabolism(substrate level phosphorylation) 320 and not necessarily from oxidative respiration. This is an important difference related to the metabolism of 321 transformed and normal cells. Authors of the review [41] examined the idea that several fluxes, including 322 aerobic glycolysis, de novo lipid biosynthesis, and glutamine-dependent anaplerosis, support proliferation of 323 diverse cell types. The regulation of these fluxes by cellular mediators of signal transduction and gene 324 expression, includes the phosphatidylinositol 3-kinase (PI3K)/Akt/mTOR system, hypoxia-inducible factor 1 325 (HIF-1), and Myc, during physiologic cell proliferation and tumorigenesis. 326 In particular, HIF-1 induces expression of pyruvate dehydrogenase kinase 1 (PDK1), which phosphorylates 327 and inhibits the PDH complex [46,47]. This limits entry of glycolytic carbon into the TCA cycle and increases 328 conversion of pyruvate to lactate. This adaptation may be important for cell survival during hypoxia. 329 However, authors suggest that while HIF-1 stimulates glycolysis, and actively represses mitochondrial 330 function and oxygen consumption, HIF-dependent mitochondrial changes are mainly functional. 331 The other suggestion is that the mechanisms that integrate signal transduction and cell metabolism are largely 332 conserved between normal cells and cancer cells. The major difference is that in normal cells, initiation of 333 In the study [9] data shows that the defective mitochondrial system described in cancer cells can be 350 dramatically improved by solely changing substrate availability and that HeLa cells can adapt their 351 mitochondrial network structurally and functionally to derive energy by glutaminolysis only. This could also 352 provide an explanation for the enhancement of oxidative phosphorylation capacity observed after tumor 353 regression or removal. This work demonstrates that the pleomorphic, highly dynamic structure of the 354 mitochondrion can be remodeled to accommodate a change in oxidative phosphorylation activity. 355

Conclusion: 356
The cancer cell lines we examined tend to have higher mitochondrial potentials, cardiolipin levels, and ATP 357 levels (on a per-cell basis) than the normal cell lines, with increased mitochondrial mass (as indicated by 358 increased cardiolipin levels) being a major factor in elevating levels of the other two variables in cancer cells.

359
This suggests that cancer cells seem to make up for the apparent insufficiency of aerobic respiration (in terms 360 of ATP production) by increasing the glycolytic rate and, probably, by utilizing glutamine fermentation as an 361 energy source. 362 363 364 365