Inverse correlation between heme synthesis and the Warburg effect in cancer cells

Cancer cells show a bias toward the glycolytic system over the conventional mitochondrial electron transfer system for obtaining energy. This biased metabolic adaptation is called the Warburg effect. Cancer cells also exhibit a characteristic metabolism, a decreased heme synthesizing ability. Here we show that heme synthesis and the Warburg effect are inversely correlated. We used human gastric cancer cell lines to investigate glycolytic metabolism and electron transfer system toward promotion/inhibition of heme synthesis. Under hypoxic conditions, heme synthesis was suppressed and the glycolytic system was enhanced. Addition of a heme precursor for the promotion of heme synthesis led to an enhanced electron transfer system and inhibited the glycolytic system and vice versa. Enhanced heme synthesis leads to suppression of cancer cell proliferation by increasing intracellular reactive oxygen species levels. Collectively, the promotion of heme synthesis in cancer cells eliminated the Warburg effect by shifting energy metabolism from glycolysis to oxidative phosphorylation.


23
Introduction 24 Cancer is caused by accumulation of various genetic mutations. A common feature of most 25 cancer cells is suppression of mitochondrial aerobic respiration and an enhanced glycolytic ATP synthesis 26 for supporting abnormal cellular proliferation and metastasis. Therefore, cancer cells require and utilize 27 abundant glucose and produce excessive lactic acid by accelerated glycolysis, resulting in lactic acidosis

28
[1]. This concept, first advocated by Nobel laureate Otto Warburg in 1924, is widely recognized as the

29
Warburg effect and has pioneered research toward analysis of tumor metabolism.

30
HIF-1 (hypoxia inducible factor 1) is activated owing to low oxygen concentration in cancer 31 cells within the tumor tissue [2][3][4]. HIF-1 induces pyruvate dehydrogenase kinase-1 (PDK-1), which 32 inactivates pyruvate dehydrogenase (PDH) [5]. PDH is a key player that converts pyruvic acid to acetyl and easily oxidized. These mixtures were centrifuged to remove proteins, and the supernatants were 129 incubated for a day at room temperature in the dark. HPLC analysis was performed as previously described

156
The homogenate was centrifuged at 600 × g for 5 minutes and the supernatant was centrifuged further at 157 11,000 × g for 10 minutes. The pellet was suspended in storage buffer and used as the mitochondrial 158 fraction. Protein concentrations were determined by the Bradford assay (Bio-Rad Laboratories, CA). COX 159 activity was measured using a Cytochrome c Oxidase Assay Kit (Sigma-Aldrich). Briefly, 100 μg of the 160 mitochondrial fraction was diluted with enzyme dilution buffer containing 1 mM n-dodecyl β-d-maltoside.

161
Ferrocytochrome c (reduced cytochrome c with dithiothreitol) was added to the sample, and COX activity

Detection of reactive oxygen species
189 ROS detection assay was performed using the cell-permeable fluorogenic probe DCFH-DA.

190
Briefly, DCFH-DA diffuses into cells and is deacetylated by cellular esterase to DCFH, which is rapidly 191 oxidized to highly fluorescent DCF by ROS. The fluorescence intensity of DCF can be assessed as an

192
indicator of cellular ROS levels.

194
After washing with PBS, the media was changed to serum-free medium supplemented with 10 µM DCFH-

195
DA. After incubating for 30 minutes, the medium was discarded and cells were collected by a cell scraper

217
The cell lines were cultured for 24 hours under normoxic (21% O 2 ) or hypoxic (1% O 2 ) 218 conditions. On performing expression analysis, in both cell lines, the expression of GLUT1 was found to 219 be enhanced by HIF-1α under hypoxic conditions ( Fig 1A). In correlation with up-regulated GLUT1, the 220 uptake of 2-NBDG was significantly increased under hypoxic conditions ( Fig 1B). Moreover,

221
concentrations of lactic acid concentration ( Fig 1C) and H + concentration ( Fig 1D) were also significantly 222 increased. Thus, hypoxic conditions accelerated glycolysis in gastric cancer cell lines. These experiments 223 also served to confirm their application in evaluating the glycolytic system.  (Fig 2A),

240
whereas extracellular coproporphyrin III (CPIII) was remarkably increased (Fig 2B). Conversely, no other 241 type of porphyrin was detected intra-or extra-cellularly. Thus, under hypoxic conditions, the porphyrin synthesis pathway is thought to be inhibited at the CPgenIII step resulting in excretion of the synthesized 243 CPgenIII and decrease in PpIX. The amount of intracellular heme under hypoxia was also measured by 244 HPLC after culture for 24 hours under normoxic (21% O 2 ) or hypoxic (1% O 2 ) conditions ( Fig 2C). A 245 significantly reduced intracellular heme concentration indicated inhibitory effect of hypoxia on heme 246 synthesis. This also supported the hypothesis that the porphyrin synthesis pathway is inhibited at the 247 CPgenIII step under hypoxia. Thus, we can postulate a negative correlation between heme synthesis and 248 the glycolytic system.

256
Inverse correlation between heme biosynthesis and glycolysis 257 Next, we examined the effect of heme synthesis on glycolysis in cancer cell lines using ALA,

258
SFC as iron ion, and SA. The addition of ALA and iron ion promotes heme synthesis because the rate-limiting step in PpIX production is ALA synthesis and heme formation by coordination of divalent iron 260 ions to PpIX [11]. Conversely, SA suppresses heme synthesis by inhibiting ALA dehydratase, which 261 functions in condensing two molecules of ALA.

262
The gastric cancer cell lines were cultured for 24 hours in media containing 1 mM ALA, 0.5 263 mM SFC, or 0.5 mM SA, followed by measurement of intracellular heme by HPLC. The addition of ALA 264 alone or ALA+SFC induced an increase in intracellular heme concentration, whereas SA led to its 265 reduction ( Fig 3A). Thus, heme synthesis can be regulated by the addition of ALA, SFC, or SA.

266
Next, we tested the effect of heme synthesis on glycolysis by analyzing the uptake of 2-NBDG, 267 lactic acid concentration, and H + concentration after 24 hours of cell culture with 1 mM ALA, 0.5 mM 268 SFC, or 0.5 mM SA. Under conditions that enhanced heme synthesis (i.e., supplementation with ALA or 269 ALA+SFC), the uptake of 2-NBDG was similar or reduced compared with control ( Fig 3B). Conversely, 270 the uptake of 2-NBDG was significantly higher under conditions that suppressed heme synthesis (SA).

271
Lactic acid concentration in MKN45 cells was down-regulated by supplementation of ALA or ALA+SFC 272 and up-regulated by addition of SA ( Fig 3C). However, no such difference was observed in KatoIII cells.

273
Furthermore, both cell lines showed a tendency toward decrease in H + concentration upon addition of 274 ALA+SFC (Fig 3D). H + concentration in KatoIII cells was increased after supplementation with SA (Fig   275  3D). Thus, heme synthesis and glycolysis are inversely related in gastric cancer cell lines.  with ALA, SFC, and SA ( Fig 4A). COX expression increased in both cell lines when heme synthesis was 20 293 enhanced by ALA or ALA+SFC. In contrast, COX expression was reduced in MKN45 cells when heme 294 synthesis was suppressed by SA. Analysis of COX activity in MKN45 cells revealed significant elevation 295 upon ALA and ALA+SFC supplementation and reduction after SA supplementation (Fig 4B). These 296 results indicate that the expression ( Fig 4A) and enzyme activity of COX increased when heme synthesis 297 was promoted and decreased when it was suppressed. Enhanced heme synthesis after addition of 298 ALA+SFC led to up-regulation of COX, most probably because it is a hemoprotein. It can thus be 299 predicted that SA-mediated suppression of heme synthesis would probably reduce COX expression and 300 activity.

301
Second, we measured the copy number of mitochondrial DNA (mtDNA) (Fig 4C). Genomic

309
Expression of COX IV in KatoIII and MKN45 cells was detected by Western blotting after incubation