Cancer cachexia has many symptoms but only one cause: anoxia

Lack of appetite: reduced food intake

Although this may seem to be the main cause, it is not. Pharmacologically improving the appetite or increasing food intake does not solve the problem. Usually, it does not stop progressive weight loss. Anorexia is frequent, but there are many patients who lose weight without a manifest loss of appetite. Important contributory factors to anorexia are depression^9,10, and cancer treatments themselves, such as chemotherapy and radiotherapy^11,12. However, this treatment-related weight loss does not seem directly associated with cancer cachexia.

Inflammatory and catabolic mediators

These mediators were found increased in cachexia, such as tumor necrosis factor alpha (TNFα)^13–15, ZAG (Zinc-α2-glycoprotein also known as lipid mobilizing factor)^16,17, interleukin (IL)-1¹⁸, IL-6¹⁹, IL-15²⁰, proteolysis inducing factor (PIF)²¹, myostatin²², and transforming growth factor-β (TGF-β)²³, among others. Levels of glucocorticoids are also increased. These mediators seem to be part of the problem but not the cause. For example, TNFα and IL-6 produce loss of appetite by interacting with hypothalamic receptors that regulate food intake²⁴. However, steep elevation of IL-6 is found mainly in very advanced stages of CC²⁵. IL-6 is also involved in an autophagy inducing activity found in serum of patients with CC. Blocking IL-6, this activity disappears²⁶.

Increased energy expenditure (higher basal metabolic rate)

A previous study showed that 143 out of 297 (48%) unselected cancer patients exhibited increased resting energy expenditure²⁷. Several authors have confirmed increased energy expenditure as a cause of weight loss in cancer patients^28–33. This probably is the consequence of increased uncoupling at the electron transport chain and increased Cori cycle activity. In the liver, the excess lactic acid produced by the tumor can be converted into glucose (Cori cycle) consuming ATP. The Cori cycle is considered an important culprit in cachexia. The Cori cycle (also known as the lactic acid cycle), converts lactate produced by anaerobic glycolysis in muscles to glucose in the liver. In cancer, instead of muscle, the tumor is the provider of lactic acid. While glycolysis produces a positive balance of two molecules of ATPs, the Cori cycle uses up six molecules of ATPs. Each turn of the cycle represents a net loss of four ATP molecules³⁴. The Cori cycle has been held responsible for energy loss in cancer by many authors^35–43. An intracellular Cori cycle cannot be ruled out as the main cause of energy loss (Figure 1).

Figure 1. The Cori cycle and glycolysis.

The figure shows that six ATP molecules are necessary to convert two molecules of lactate into glucose (left panel). On the other hand, the glycolytic pathway (right panel) produces two ATP molecules by degrading glucose to lactate. If two molecules of lactate produced through glycolysis are reconverted into glucose, there is a net loss of four ATP molecules. If this is established as a permanent circuit glucose-lactate-glucose circuit, each turn in the circuit loses four ATP molecules.

Interestingly, the Cori cycle is activated during fasting⁴⁴ where it contributes to generating glucose. Patients with advanced metastatic cancer show an increased Cori cycle, particularly those patients with high glycolytic flux^35,36,45,46.

According to our criteria, the Cori cycle by itself, is the main responsible actor in CC, but not the originating cause. Figure 2 and Figure 3 show a proposed mechanism of how the Cori cycle develops in cancer and the energy imbalance it drives. For each molecule of glucose produced through the Cori cycle, six molecules of ATP are used. For each molecule of glucose degraded to lactic acid only two molecules of ATP are produced. Therefore, if a cycle is established in which one molecule of glucose produces two ATP molecules, and the lactic acid thus formed is used to regenerate glucose, four molecules of ATP are lost in each complete cycle.

Figure 2. Energy consuming effects of the Cori cycle.

Left panel shows condition 1 (slightly hypoxic environment) while the center panel shows condition 2 (extreme hypoxia or anoxia). Right panel shows the effects of anoxia on the energy balance.

Figure 3. Energy consuming effects compounding in each Cori cycle.

The figure shows how a further turn of the cycle increases glycolytic flux and at the same time increases energy loss.

Loss of adipose tissue

Loss of adipose tissue due to increased lipolysis⁴⁷ seems to be activated by protein kinase A⁴⁸. Hepatic nuclear factor-4 (HNF4) mRNA was downregulated in adipose tissue of patients with CC⁴⁹. Degradation of triglycerides also has a role in the loss of adipose tissue⁵⁰ with the intervention of adipose triglyceride lipase⁵¹. ZAG (lipid mobilizing factor), which decreases lipids from adipocytes⁵², also increases the expression of uncoupling proteins in adipose tissue and skeletal muscle and therefore produce potential energy loss^53,54. Interestingly, the expression of ZAG is increased with hypoxia and induces insulin resistance⁵⁵.

Loss of skeletal muscle

Loss of skeletal muscle occurs due to increased proteolysis and decreased protein synthesis.

Tumor stage

Tumor stage seems to be a predictive factor of cancer cachexia⁵⁶ and is probably related to tumor mass.

Insulin resistance

Asp et al.⁵⁷ found that CC bearing mice had a significantly decreased glucose response to insulin. Rosiglitazone improved insulin sensitivity. Muscle wasting seems to be also related to insulin resistance⁵⁸. HIF-1α can induce insulin resistance⁵⁹. Intermittent hypoxia has the same effect^60–63.

Condition 1

This condition is characterized by a pO₂ higher than 0.5%. In this situation the oxidative phosphorylation would remain active and 30% or more glucose would be metabolized by mitochondria (Krebs cycle and electron transport chain) to CO₂ and H₂O. The rest, 70% of glucose would undergo glycolysis to lactic acid. (Warburg effect: preferential glycolytic metabolism in aerobiosis). The 140 molecules of lactate thus formed are expelled to the extracellular space by the monocarboxylate transporters. Oxidative phosphorylation remains operative during the Warburg effect. In a famous debate with Warburg⁶⁴ in 1956, Sidney Weinhouse stated that Warburg’s concept about tumor cells being unable to oxidize glucose was wrong. Furthermore, Weinhouse showed that glucose can be oxidized to CO₂ in cancer at a rate similar to normal cells⁶⁵. This concept has since been validated by many authors^66–73. The amount of oxidative phosphorylation that continues working after the metabolic shift varies considerably among different tumors and duration of hypoxia⁷⁴.

What must be kept in mind is that the Warburg effect is not the shutdown of the oxidative metabolism. It is the predominance of glycolytic metabolism over oxidative metabolism, but oxidative metabolism continues working. Oxidative metabolism may be decreased, equal to or greater than in normal counterparts. The presence of oxygen increases oxidative metabolism in normal and cancer cells; however, this increase is much lower in the latter. Oxidative metabolism is present even in highly glycolytic cells but operating at a lower capacity⁷⁵.

As a conclusion, the Warburg effect is not about mitochondrial metabolism impairment (as Warburg thought) but about increased glucose uptake and glycolytic flux as postulated by Weinhouse. There is high metabolic variability among cancer types and also inside a tumor. This means that glycolytic cells may conserve variable degrees of mitochondrial metabolism.

Condition 2

In this condition, the pO₂ has decreased below 0.2%. Mitochondrial activity is practically downregulated by such a low level of oxygen. Therefore, nearly 100% of glucose is degraded to lactic acid, generating 200 molecules of lactate. Such a high lactate load can easily surpass monocarboxylate extruding capacity and a certain amount of lactate would remain inside the cell creating intracellular lactic acidosis that would endanger cancer cell survival. Activation of the Cori cycle comes to solve this situation by reconverting part of the lactate to glucose. This creates a vicious cycle in which the more glucose is degraded, the more the Cori cycle “works”, consuming four ATPs in each turn of the cycle. Each turn of the cycle increases the glycolytic flux by the generation of more glucose. This creates a vicious cycle. Figure 2 and Figure 3.

Figure 2 and Figure 3 are not based on real calculations. They are only used to illustrate reasons why the Cori cycle is activated and how this cycle consumes progressively higher amounts of energy. As we can see the main culprit in this scheme is extreme hypoxia.

The following concepts were taken into consideration in the construction of Figure 2 and Figure 3.

Hypoxia is an activator of gluconeogenesis (Cori cycle)

Hypobaric hypoxia has been shown to produce weight loss through diverse mechanisms including loss of appetite and activation of the Cori cycle⁷⁶. HIF-1α and HIF-2α are strongly increased with ascent above 4,000 meters of altitude⁷⁷. HIF-1α is a transcriptional activator of phosphoenolpyruvate carboxykinase (PEPCK) which is the rate-limiting enzyme for gluconeogenesis⁷⁸. Experimental downregulation of HIF-2α decreased gluconeogenesis in hepatoma cells (HepG2) and decreased tumor size⁷⁹.

The Cori cycle is a defense mechanism against lactic acidosis

Suhara et al.⁸⁰ have shown that gluconeogenesis (Cori cycle) is a mechanism that defends against lactic acidosis. Why does the cancer cell need the Cori cycle? The need stems from the fact that the monocarboxylate transporter (MCT) system is saturable. The complete or almost complete abrogation of the mitochondrial metabolism plus the increased glycolytic flux represent such a burden that the MCT capacity is surpassed. In muscle, the half-maximal rate of lactate transport is achieved with a lactate concentration between 13 and 40 mM⁸¹. If the maximal rate is achieved (about 20 nmol/min per μl of intracellular volume at 25°C)⁸², the excess would remain inside the cell. Therefore, by transforming lactate into glucose or pyruvate The Cori cycle prevents intracellular lactic acidosis which would induce acidic stress and kill the cell. The velocity of lactate extrusion by MCTs is also dependent on intracellular and extracellular pH⁶⁰. Decreased intracellular pH increases extrusion velocity, while it is lowered by extracellular acidity⁵⁹. Anoxic areas of tumors have a very acidic extracellular substance, and this may slow down lactate extrusion.

The main source of glucose formed by gluconeogenesis is lactate

Koloyianni et al.⁸³ and Ludholm et al.³⁸ found that in normal cells, 60% of glucose generated by gluconeogenesis used lactate as the source molecule. Glutamine and alanine contributed 10% each and glycerol 3%. The rest came from serine, glycine, and threonine.

Complex I inhibitors also inhibit gluconeogenesis

Phenformin inhibits gluconeogenesis⁸⁴. This is a paradoxical result in the scheme, because Complex I inhibition decreases oxidative phosphorylation. However, this anti-gluconeogenesis activity of Complex I has been tested in cells that were still performing oxidative phosphorylation. It is possible that under full anaerobiosis Complex I inhibition would have no effect on gluconeogenesis.

A proof of this last concept is that pharmacological inhibition of HIF-1α reduces cancer cachexia⁸⁵. The authors used emodin and rhein (from Rheum palmatum) to decrease HIF-1α expression. Interestingly, emodin and rhein have been shown many other anti-cancer effects^86–89.

Mitochondrial oxidative defect produces lactic acidemia

Mitochondrial diseases that decrease mitochondrial activity can produce lactic acidemia⁹⁰. The same happens with excessive anaerobic exercise^91,92. However, in the case of exercise, even though the ability of MCTs to expel lactate is not exceeded, there is no intracellular lactic acidosis.

Metabolic modifier: DCA

DCA is an investigational drug for lactic acidosis, pulmonary hypertension and now for cancer. In some African countries, its use is unofficially accepted for the treatment of lactic acidosis in children with malaria²³⁶. It is a small molecule with the following formula HOOC-CH₂-Cl₂. DCA is an orally available molecule that is almost completely and quickly absorbed at the digestive system^237,238. It is metabolized by GSTZ1 (a glutathione transferase isoform)²³⁹.

When DCA was administered to 16 healthy individuals, in 1 to 50 mg/kg IV infusions, it lowered plasma glucose, lactate and alanine concentrations. Plasma levels linearly followed those of the administered dose up to 30 mg/kg. A dose of 35 mg/kg was considered most effective regarding lactic acid, which fell to 75% below baseline concentrations within 2 hours of the infusion²⁴⁰. Blood glucose was not affected in these healthy individuals but was reduced in diabetic patients through stimulation of peripheral glucose utilization and inhibition of gluconeogenesis. DCA also inhibits lipogenesis and cholesterol synthesis. Daily oral administration of 50 mg/kg DCA to diabetic patients with slightly elevated lactic acid concentration, reduced alanine and lactic acid in plasma²⁴¹.

The experimental administration of a single dose of DCA may be misleading upon dose and metabolism, because as Gonzalez-Leon et al.²⁴² have shown in rats with repeated administration of DCA, this drug has a surprising feature: it inhibits its own metabolism. This means that the chronic administration of DCA differs from single doses in its plasma concentration, toxicology and metabolism. Figure 7.

Figure 7. Differences in DCA clearance between first and subsequent administration.

This is an important issue to be considered for the clinical use of DCA.

In humans, the maximum plasma concentration achieved with an IV infusion of 10 mg/kg was between 19.9 μg/ml and 24.7 μg/ml with a half life of only 20 minutes. When the infused dose was increased to 20 mg/kg the plasma concentration was between 57.3 and 74.9 μg/ml with a half life of 36 minutes. In dogs and rats the half life was much longer (the half life of a 100 mg/kg dose was 4 hours and 24 hours in dogs and rats respectively²⁴³. These marked differences among species cast doubts about the possibility of translating the findings in other mammals to humans.

Confirming these inter-species differences, Maissenbacher et al. have found that dogs present enhanced inhibition of DCA degradation and slower clearance than humans and rats due to increased inhibition of GSTZ1²⁴⁴. The DCA metabolic pathway starts with dehalogenation to monochloroacetate and glyoxylate and then it continues to glycine, and the final products are oxalate and carbon dioxide²⁴⁵. The first dose is cleared from plasma faster than subsequent doses, as Gonzalez-Leon et al. have shown; this is probably due to GSTZ1 inhibition by DCA in subsequent doses. This is similar in all species. The decrease in DCA clearance by multiple doses is not a minor issue, because the initial clearance may be reduced to less than 25% of the initial one in successive doses²⁴⁶.

The main mechanism of action of DCA is the inhibition of pyruvate dehydrogenase kinase (PDK) and its isoforms. This inhibition increases the flux of pyruvate into the mitochondria, promoting glucose oxidation instead of glycolysis²⁴⁷. PDK inactivates the pyruvate dehydrogenase enzyme complex through phosphorylation. By downregulating the activity of this complex, PDK decreases the oxidation of pyruvate in mitochondria and increases the conversion of pyruvate to lactate in the cytosol.

Other mechanisms of action of DCA in cancer can also be found in the medical literature. Stockwin et al.²⁴⁸ described that cytotoxicity of DCA is only achieved in those cells that suffered mitochondrial DNA mutations that “condemn” them exclusively to the glycolytic pathway. Therefore, DCA has features that make presume it will reduce glycolysis, lactic acid production and gluconeogenesis in anaerobic malignant cells or even cause their death.

Why DCA?

Increasing tumor oxygenation and inhibiting TNFα: Pentoxifylline, nitroglycerin and thalidomide

Pentoxifylline reduces the expression of TNFα in cancer cells. It has been tested in CC in five patients with good results in three of them²⁷⁵. Many other reports have confirmed the inhibitory actions of pentoxifylline on TNFα^276–279, not only in cancer, but also in other pathologies^280,281. Pentoxifylline has other very important actions in CC: it is a hemorheological agent that increases red blood cell deformability, reduces blood viscosity and decreases platelet aggregation improving microcirculation²⁸². Therefore, a better oxygenation of the tumor is expectable. However, a study by Goldberg et al.²⁸³ with pentoxifylline as a stand-alone drug did not show any improvement in patients with CC. There is evidence that pentoxifylline increases tumor oxygenation^284,285.

Nitroglycerin (NTG) is another drug that deserves consideration. It is a well-known vasodilator and oxygenation improver. In addition to these effects NTG also reduces HIF-1α levels in hypoxic tumors^286,287, because it acts as a nitrous oxide donor. Unfortunately, NTG has both pro and anti-tumor effects²⁸⁸. However, used on a short-term basis as an adjunct to pentoxifylline, an important improvement of tumor oxygenation can be expected. For a review of nitroglycerin’s anti-tumoral activity, read Sukhatme et al.²⁸⁹.

Thalidomide is a TNFα downregulator²⁹⁰ and has been tested in CC with encouraging perspectives²⁹¹. It had similar results in wasting syndromes of other origins^292,293. Thalidomide has also other anti-cancer effects²⁹⁴, such as anti-angiogenesis and T-cell stimulation²⁹⁵, which will not be considered here because they go beyond the scope of this manuscript. It has been established as part of multiple myeloma treatment protocols. For a review of thalidomide, read Luzzio²⁹⁶.

Downregulation/inhibition of HIF-1α

As mentioned above, emodin can be used for this purpose. Emodin down-regulates HIF-1α expression²⁹⁷. It also inhibits pro-inflammatory responses²⁹⁸ that play a role in CC. Furthermore, hepatic cancer cells (HepG2) treated with emodin showed a significant decrease of lactate. This decrease is a signal of glycolytic inhibition which has been further confirmed because emodin decreased mRNA levels of hexokinase II (HKII), pyruvate kinase isoform M2 (PKM2), and lactate dehydrogenase-A (LDHA) in a concentration-dependent manner²⁹⁹. Emodin also inhibits TNFα, NF-κB and IL-6, all mediators of CC³⁰⁰. Emodin has many other anti-cancer effects, such as:

For a review on emodin pharmacology see Dong et al.³¹⁴.

The treatment proposed here includes the association of pentoxifylline, emodin and DCA as an added scheme to classical treatments such as high calorie nutritional supplements and anabolics like megestrol (Figure 8). Adding these drugs to the conventional treatment would decrease anoxia, HIF-1α expression, and the glycolytic pathway restoring oxidative phosphorylation and reducing TNFα expression. This type of treatments targets the etiology of CC rather than the symptoms.

Figure 8. A diagram showing the specific site of action of each drug with anti-CC potential.

Inhibition of PEPCK to block gluconeogenesis

PEPCK is the rate-limiting enzyme for gluconeogenesis. Therefore, its inhibition should block the Cori cycle. Many drugs have been identified with the ability to inhibit PEPCK, such as metformin³¹⁵, troglitazone³¹⁶, berberine³¹⁷, among others. None of these drugs have been tested in CC. Berberine should be considered a particularly interesting drug because it inhibits PEPCK but also downregulates HIF-1α³¹⁸. Berberine also has many other anti-cancer effects³¹⁹, such as down-regulation of COX2³²⁰, increased apoptosis in cancer cells without affecting the normal ones^321,322, reduced migration and invasion, among others. For a review of other anti-cancer effects of berberine, read Kaboli et al.³²³

Other potential drugs

3-bromopyruvate (3BP) is a protein alkylating agent that has shown many anti-cancer effects. We included it in this list of possible drugs for treating CC because it is a potent inhibitor of aerobic glycolysis³²⁴. Inhibition of glycolysis should decrease lactate production, thus decreasing the Cori cycle. The exact mechanism of action is not fully known, but there is some evidence pointing to inhibition of glycolytic enzymes³²⁵. It has important cytotoxic effects on highly glycolytic tumor cells³²⁶.

Tocilizumab (a humanized anti-IL6 receptor antibody)³²⁷ may produce some benefits^328,329. It requires further testing. The association of tocilizumab with gemcitabine for the treatment of advanced pancreatic cancer failed to show clear clinical benefit in a phaseI/II clinical Trial³³⁰.

Insulin, besides its known actions (inhibitor of gluconeogenesis), exerts inhibitory activity on PGC1α expression³³¹.

Anamorelin is a small molecule ghrelin receptor agonist that has shown favorable effects on appetite, food intake and weight gain in patients with CC^332–334. However, its approval was rejected twice by the European Medicines Agency³³⁵.