Review

Dietary modifications for enhanced cancer

therapy

https://doi.org/10.1038/s41586-020-2124-0 Naama Kanarek1,2,3 ✉, Boryana Petrova1 & David M. Sabatini3,4,5,6

Received: 27 February 2019

Accepted: 27 January 2020 Tumours depend on nutrients supplied by the host for their growth and survival.
Published online: 25 March 2020 Modifications to the host’s diet can change nutrient availability in the tumour
microenvironment, which might represent a promising strategy for inhibiting tumour
Check for updates
growth. Dietary modifications can limit tumour-specific nutritional requirements,
alter certain nutrients that target the metabolic vulnerabilities of the tumour, or
enhance the cytotoxicity of anti-cancer drugs. Recent reports have suggested that
modification of several nutrients in the diet can alter the efficacy of cancer therapies,
and some of the newest developments in this quickly expanding field are reviewed
here. The results discussed indicate that the dietary habits and nutritional state of a
patient must be taken into account during cancer research and therapy.

The use of dietary modification to supplement conventional cancer

therapy is an eminently practical approach that is receiving growing Dietary restriction of nutrients
attention. Dietary composition dictates nutrient availability in the Pre-clinical studies have suggested that some nutrients are good tar-
plasma and therefore in the microenvironment of cells in the body, gets for dietary restriction in order to enhance the efficacy of cancer
including cancer cells1. Manipulation of the metabolic environment therapies. In addition, some dietary regimes that restrict whole groups
of cancer cells markedly changes their metabolic activity, producing of nutrients appear to have a beneficial effect on inhibition of cancer
shifts in drug sensitivity2, proliferation rate and metabolic require- progression.
ments3,4. The diet also dictates signal transduction through nutri-
ent-sensing pathways that are strongly associated with oncogenic Fasting
signalling5. Several fasting practices have been shown to prevent cancer in mice9,10.
As research tools for studying the metabolic dependencies of Intermittent fasting prevents lymphoma in aged mice11, and delays the
tumours in vivo have improved over the past few years, there has been development of sarcoma and lymphoma in Trp53+/− mice, even when
a surge in reports demonstrating the effects of modifying the avail- implemented at a late stage in life (7–10 months old)12. The combination
ability of specific metabolites on cancer progression and response to of prolonged fasting cycles (48–60 h) and chemotherapy significantly
therapy. These pre-clinical studies are distinct from epidemiological improved the response to therapy in mouse xenograft models of breast
studies that have described associations between lifestyle, dietary cancer, melanoma, glioma and neuroblastoma13. It is likely that the
habits, overeating and cancer and are limited to retrospective cor- main mechanism by which fasting inhibits the progression of cancer
relation with risk factors. The recent in vivo research points towards is through a reduction in systemic levels of insulin-like growth factor 1
prospective strategies for dietary manipulation that might enhance (IGF1)14, but genetic or pharmacological perturbations are still required
responses to treatment following a cancer diagnosis, and indicates to verify the molecular factors that contribute to the anti-cancer effect
that specific types of tumour might be targets for these manipulations. of fasting.
Hypothetically, dietary modifications can enhance cancer therapy An alternative fasting-mimetic diet that consists of a periodic low-
by a number of mechanisms (Fig. 1). We refer to other reviews for exten- calorie and low-protein regimen resulted in a similar anti-cancer effect to
sive discussion of dietary modifications and their major and complex that of prolonged fasting15. As patients with cancer are often frailer than
effects on the immune system6 and on tumour immunity7,8, which are healthy individuals, this diet might be superior to the drastic prolonged
outside the scope of this review. fasting protocol, because it is easier to comply with and would probably
Here we discuss in detail the dietary restriction of a number of nutri- not impair the patient’s fitness as much as would prolonged fasting.
ents, including glucose, fructose and amino acids, and we describe Below, we discuss how less harsh diets, such as the restriction of
the effects of each of these nutrients on cellular metabolism and their glucose, fructose or some amino acids can influence cancer develop-
interplay with signalling pathways. We then outline how other dietary ment or responses to therapy, and introduce a more feasible diet for
interventions, such as nutrient supplementation or pharmacological appropriate cases of cancer.
depletion of nutrients can be harnessed for cancer therapy. We also
explore other potential targets. Finally, we present an outlook of our Glucose restriction
view on the potential of future combinations of diets and therapies Glucose is a central nutrient that is used in several metabolic pathways,
that may be tailored to specific patients. and is consumed by tumours at high levels to support their growth16,17.

Department of Pathology, Boston Children’s Hospital, Boston, MA, USA. 2Harvard Medical School, Boston, MA, USA. 3Broad Institute of Harvard and Massachusetts Institute of Technology,
1

Cambridge, MA, USA. 4Whitehead Institute for Biomedical Research, Cambridge, MA, USA. 5Department of Biology, Howard Hughes Medical Institute, Massachusetts Institute of Technology,
Cambridge, MA, USA. 6Koch Institute for Integrative Cancer Research, Cambridge, MA, USA. ✉e-mail: naama.kanarek@childrens.harvard.edu

Nature | Vol 579 | 26 March 2020 | 507

Review
Food

a b Enhancement of drug-based therapy c Anticancer immune response activation

Drug-enhancing Tumour Sugar CD8 T-cell

Metabolites metabolite

Drug
Dying area in
a tumour
PD-L1

d Cancer-specific toxicity e Tumour starvation

Inert metabolite
Starved

Well-nourished

Tumour-specific
toxic metabolite

Fig. 1 | Mechanisms of tumour inhibition by dietary modifications. a, immune inhibitory ligands such as PDL-1138 (orange). An additional example
Tumours are nourished by the patient’s diet. Nutrients consumed by patients (not shown) is that arginine supplementation enhances T cell activation and
with cancer become available to tumours. Dietary modifications of specific improves T cell-mediated antitumour responses and survival in tumour-
nutrients can have an anti-tumorigenic effect through several mechanisms, bearing mice139. Other examples are reviewed elsewhere6–8. d, Cancer-specific
conceptually shown here. b, Enhancement of drug-based therapy. It is toxicity. A nutrient that is inert for healthy tissues but can be toxic for cancer
hypothetically possible to increase the efficacy of metabolic anti-cancer drugs cells owing to their unique metabolism can inhibit tumour progression if
by nutritional changes through several mechanisms. For example, dietary supplemented through the diet. For example, mannose inhibits the
histidine supplementation enhances the response to methotrexate in a mouse progression of pancreatic cancer xenografts when given alongside
model of leukaemia81. In an additional example (not shown), methionine chemotherapy83. e, Tumour starvation. As tumours rely on nutrients provided
restriction improves cancer inhibition by chemotherapy and radiation through by the host for their growth and survival, dietary restriction of nutrients that
reduction of one-carbon metabolism and nucleotide synthesis51. c, Anticancer are essential for the tumour but not for other organs can inhibit tumour
immune response activation. Cancer immunotherapy is used to treat various growth. There are several examples of such nutrients, including glucose,
cancers and is being tested for additional types of cancer. There is strong glutamine, glutamate, asparagine, aspartate, methionine, serine and folate.
association between diet and the tumour-relevant immune response. For However, dietary restriction of some of these nutrients does not result in
instance, hypocaloric or ketogenic diets are associated with an increase in immediate depletion of the nutrient at the tumour site, compromising the anti-
tumour-infiltrating CD8+ T cells136,137 (purple) and a reduction in expression of cancer benefit of the dietary restriction.

Glucose has many pro-tumorigenic roles (Box 1); for example, glucose rather than for energy production is known as the Warburg effect19. Al-
is a source for energy production and for the synthesis of biomolecules though this effect may vary in its extent among different types of tu-
that sustain the high proliferation rate of cancer cells. In addition, mours in vivo20, it is a good example of how tumour cells can rewire their
dietary consumption of glucose results in the secretion of insulin, a metabolic programs to fuel their unique metabolic requirements16,21.
well-characterized oncogenic signalling factor5. Glucose contributes
to cancer progression, resistance to therapy, and possibly cancer initia- Glucose induces pro-oncogenic signalling. In vivo, blood glucose levels
tion18. Unlike other nutrients, such as essential amino acids, systemic are buffered effectively by the actions of insulin. Dietary intake of sugar
levels of glucose can be reduced by relatively simple dietary changes: increases blood glucose levels; this is sensed by pancreatic beta cells,
therefore, glucose consumption is an ideal target for enhancing the which respond immediately by secreting insulin. Insulin reduces glucose
efficacy of cancer therapies through dietary modification. Below, we levels in the blood primarily by increasing the uptake of glucose by skel-
describe features of glucose metabolism and signalling, and how they etal muscle and the liver. However, insulin is also sensed by tumour cells,
can be therapeutically exploited. which express insulin receptors that activate the downstream phospho-
inositide 3-kinase (PI3K) signalling pathway. Aberrant PI3K activation is a
Glucose is a source for energy and biomass production. Glucose hallmark of cancer that is found in most cancer cells and promotes cancer
feeds several anabolic processes that are enhanced in cancer cells and growth through positive regulation of the cell cycle, survival, anabolic
are vital for their high proliferation rate (Fig. 2). The majority of glucose metabolism and more22. The PI3K pathway is driven by activating muta-
that enters cancer cells is channelled to glycolysis and pyruvate synthesis. tions in PIK3CA and the AKT kinase group of genes and inhibition of the
A surprisingly small fraction of pyruvate is used through acetyl-CoA to tumour suppressor PTEN22. Activation of PI3K signalling by insulin or by
the tricarboxylic acid (TCA) cycle for ATP production in tumours, but, mutations activates mTORC1, which decreases autophagy and increases
however small, this aspect of glucose metabolism is essential for tumours anabolic processes, proliferation and biomass buildup5,22,23. Thus, the mo-
in vivo4. The preference of cancer cells to use glucose for biomass buildup lecular links between glucose consumption, blood insulin and oncogenic

508 | Nature | Vol 579 | 26 March 2020

 Box 1

Glucose and glutamine

Cancer cells consume both glucose and glutamine in amounts that pentose-phosphate pathway, and NADPH with glutamine form the
substantially exceed those of normal cells. This characteristic can be antioxidant GSH105,133.
used for live imaging of tumours17,144 and indicates the key role these Lipid synthesis
two metabolites play in the unique metabolic state of transformed cells. Both glucose and glutamine contribute to lipid synthesis; glucose
Perhaps not surprisingly, these metabolites share several metabolic is the default contributor to cytoplasmic acetyl-CoA, which is
fates and each can sometimes compensate for the lack of the other. converted to lipids, and glutamine becomes the main carbon source
The TCA cycle for lipids under hypoxic conditions, which are common in tumours
Both glucose and glutamine contribute to the TCA cycle: glucose in vivo106.
through pyruvate and acetyl-CoA, and glutamine through It has been suggested that glucose and glutamine have metabolic
α-ketoglutarate. The TCA cycle produces ATP and NADH, which roles beyond their contributions to biomass buildup in cancer
allows more ATP production through oxidative phosphorylation104 cells, as their consumption by cancer cells exceeds their
and is therefore important for the energetic balance of cells132. incorporation into macromolecules but is essential for
proliferation134. Some of these potential roles have been discussed
Nucleotide synthesis previously16, but future experimental work is needed to clarify and
Both glucose and glutamine are required for nucleotide synthesis. validate them.
In this pathway, their roles are complementary and therefore Unfortunately, the strong dependency of cancer cells on glucose
deprivation of either will result in insufficient nucleotide synthesis. and glutamine is not easily translated into targetable vulnerability
Glutamine contributes one nitrogen atom to the pyrimidine ring in vivo, as discussed above. As glutamine is the most abundant
and two nitrogen atoms to the purine ring, whereas glucose is amino acid and glucose is the most abundant carbohydrate
necessary for synthesis of PRPP from ribose-5-phosphate (the in human serum135, it is very challenging to effectively reduce
product of the pentose phosphate pathway). PRPP is a cofactor of both of them through dietary adjustments, and pharmacologic
the pyrimidine synthesis enzyme UMPS, and is also the base for interventions are currently too toxic for clinical use. Future
both de novo and salvage purine synthesis. improvements in the pharmacologic reduction of glutamine levels
Cellular redox balance in the blood, combined with dietary restriction of glucose, might
Glucose and glutamine work together to maintain cellular redox be effective in preventing tumour growth by denying access of
balance; glucose is essential for NADPH production through the transformed cells to their bread and butter.

signalling have been well characterized. In addition, chronically high various degrees and do result in reduced blood glucose (for example, in
levels of insulin, such as those found in individuals with obesity and/or humans, a 15% reduction in blood glucose was achieved following a 50%
diabetes, are correlated with a higher risk of cancer18. Although additional reduction in total caloric intake24), but also have far broader metabolic
factors beyond insulin and glucose are likely to play a pro-tumorigenic effects. Caloric or dietary restriction, although likely to be beneficial
role in individuals with obesity and diabetes, insulin, and, by proxy, dietary for patients with cancer in terms of depriving tumours of glucose25, are
glucose, contributes to the initiation of cancer, and therefore might also not optimal as a recommended diet for patients as they are difficult
be targeted for cancer prevention. to comply with, and might compromise the patients’ general physical
condition. Perhaps a better strategy is a diet that is low in glucose but
Dietary glucose restriction has benefits for cancer therapy. Many has normal caloric value (an isocaloric diet) such as the ketogenic diet.
laboratories have investigated whether reductions in dietary glucose The ketogenic diet is rich in fat but low in carbohydrates, and can indeed
can inhibit cancer progression. The most commonly tested glucose- inhibit cancer progression26,27. The ketogenic diet reduces glucose levels
reducing dietary restriction scheme is calorie restriction, sometimes in the blood (a 30% reduction in blood glucose was achieved in humans
termed dietary restriction. These diets restrict total caloric intake to adhering to an isocaloric diet in which carbohydrates provided only 8%

Nature | Vol 579 | 26 March 2020 | 509

Review
Glucose A combination of dietary and pharmacological inhibition of insu-
Pentose phosphate pathway lin is likely to be a more effective way to inhibit tumour growth. This
Glucose-6-phosphate NADPH Ribose approach was tested in mice, and dietary restriction of glucose by a
Oxidative stress reduction ketogenic diet was found to prevent hyperglycaemia and compensa-
Glycolysis tory insulin secretion following pharmacologic inhibition of PI3K. The
De novo serine synthesis combined approach resulted in reduced pro-oncogenic signalling
3-Phosphoglycerate Serine Glycine through mTOR, enhanced tumour inhibition and improved survival
Pyruvate Biomass of tumour-bearing mice in several cancer models33. These results hold
great promise for patients being treated with PI3K inhibitors, but their
Lactate Alanine implications go far beyond this group of patients and should motivate
Energy Aspartate One-carbon metabolism a thorough consideration of dietary recommendations to patients
Asparagine Nucleotides with cancer in general.
Acetyl-CoA Biomass DNA replication
Although all tumours feed on glucose, and the vast majority of
Gene expression tumours proliferate optimally in the presence of excess glucose, a few
tumours progress more rapidly when supplied with fat as their energy
Fatty acid TCA cycle source34. In these few cases, patients will benefit from diets in which
Glutamate
synthesis
ATP
carbohydrates are the main source of calories. Therefore, it is important
Proline
Fatty acids Energy to study the metabolic preferences of each type of tumour and to com-
Arginine
plete pre-clinical evaluation of the outcome of changes in the dietary
Biomass Biomass
energy source before adopting any recommendations for patients.
Fig. 2 | Glucose metabolism in cancer cells. The majority of glucose that enters Several clinical studies have tested the potential benefits of a
cancer cells is channelled to glycolysis and pyruvate synthesis. Pyruvate is used ketogenic diet for patients with cancer35. These studies included only
for the synthesis of the amino acids alanine, aspartate, and asparagine, but a small number of participating patients, making it hard to conclude
mostly for lactate synthesis. Lactate is taken up by many organs and functions whether the ketogenic diet holds significant benefits. In addition, the
as a substrate for the TCA cycle through conversion back to pyruvate140,141. lack of a standard protocol for this diet, as well as high variability in
Pyruvate can enter the mitochondria and serve as a substrate for acetyl-CoA the compliance of patients to the diet in the different studies, makes it
biosynthesis. Acetyl-CoA feeds the TCA cycle and contributes to the synthesis difficult to compare their results. Nonetheless, some of these studies
of additional amino acids, fatty acids and cholesterol. The remaining glucose
have provided promising indications that dietary restriction of sug-
that is not converted into pyruvate is used for the pentose phosphate pathway,
ars by a ketogenic diet improves outcomes in cancer patients, is safe,
which produces NADPH and ribose-5-phosphate, and some is used for de novo
and is feasible in patients with several types of advanced cancer28,36,37.
synthesis of serine. NADPH is essential for decreasing oxidative stress, and
ribose-5-phosphate is used for nucleotide synthesis. Serine can be used for
Although some of these studies didn’t show that glucose withdrawal
glycine synthesis and is important for one-carbon metabolism and nucleotide induced significant tumour inhibition, they set the stage for further
synthesis. studies that should emphasize a standard, clear protocol for a ketogenic
diet and include careful monitoring of compliance. It is also likely that
testing the benefit of the ketogenic diet at earlier stages of cancer pro-
of total caloric intake28) and elevates ketone bodies, which are not con- gression will yield better results. In the meantime, reducing glucose
sumed by most cancer cells but provide energy to the brain and other consumption, monitoring blood glucose levels and insulin secretion
tissues. The ketogenic diet is probably most effective when the diet is in patients, and assisting patients in maintaining low-carbohydrate
portion-controlled and food intake is not unlimited, to allow for nor- diets will probably improve the survival of many patients with cancer.
mal, and not higher than normal, caloric consumption by the patient29.
Mechanistically, the ketogenic diet reduces circulating insulin, thereby Potential benefits of fructose restriction
lessening activity in the pro-proliferative PI3K and mTORC1 signalling The Western diet is very rich in fructose, a monosaccharide that has
pathways29,30, and may also increase oxidative stress in cancer cells27. causal roles in many pathologies, including cancer38,39. Until recently,
Data from studies using mouse models of cancer suggest that the only epidemiological studies had provided evidence for a harmful, pro-
role of glucose as an inducer of pro-survival signalling pathways is vital tumorigenic effect of fructose consumption40, but now, with new data
for its pro-tumorigenic effects. Carbohydrate restriction is less effec- regarding the fate of fructose consumed through the diet41 and its role
tive in inhibiting the growth of cancer cells that harbour mutations in in cancer progression, it is becoming clear that fructose consumption
the PI3K pathway31,32. Even though these cells still require glucose as by patients with cancer should be carefully considered.
an energy source, the cells are refractory to the signalling effects of Cancer cells can use fructose as energy source, and many types of
glucose restriction because signalling downstream to PI3K is consti- cancer import large amounts of fructose by upregulating the fructose-
tutively active. These findings highlight the important role of systemic specific transporter GLUT542. The dependency of tumours on fructose
glucose as the initiator of the PI3K pathway through insulin, and imply as an energy source is underlined by the inhibitory effect of GLUT5
that inhibition of this pathway will circumvent the pro-tumorigenic role silencing on cancer progression in vitro and in animal models43–45.
of glucose and effectively inhibit tumour growth. This approach holds Although fructose is mostly absorbed and metabolized by the intes-
promise for patients with tumours bearing mutations in PIK3C genes, tine, if consumed in high doses that saturate the intestinal capacity for
owing to the tumour’s addiction to insulin signalling. However, given fructose absorption, it spills over to the liver41. Fructose that reaches the
that most tumours consume high amounts of glucose and thrive on liver induces de novo lipid synthesis, fatty liver46, increased secretion
the proliferative cues downstream of insulin and PI3K22, this strategy of triglycerides to the blood, type two diabetes and obesity47. However,
might be as effective for inhibiting tumours with unmutated PI3K. chronic consumption of fructose even in moderate doses (equivalent
Unfortunately, pharmacological inhibition of PI3K is surprisingly inef- to one can of soda per day) has been shown to increase the incidence of
fective, as compensatory insulin secretion confers resistance to this colorectal cancer in an Apc−/− mouse model48. The fact that this moder-
strategy22. In mice, blood insulin levels increase within hours of inhibi- ate dose is sufficient to activate glycolysis in intestinal tumours and
tion of the insulin–PI3K–AKT–mTOR signalling pathways, which allows induce fatty acid synthesis and tumour growth challenges the com-
increased glucose uptake and activation of oncogenic PI3K signalling mon assumption that only a large excess of fructose can be causal for
by tumours33, circumventing the anti-tumorigenic effects of the drug. cancer, at least in the intestine.

510 | Nature | Vol 579 | 26 March 2020

 Serine synthesis pathway
Exogenous 3-PG, 3-phosphoglycerate
Glucose PHP, 3-phosphohydroxypyruvate
P-serine, 3-phosphoserine

Glycolysis PHGDH PSAT PSPH

3-PG PHP P-serine Serine
THF

5,10-methylene 5-methyl
Phosphatidylserine Cystathionine Serine
SHMT THF THF
synthase synthase dehydratase
Phosphatidylserine Cysteine Pyruvate Glycine

Lipid synthesis GSH TCA cycle One-carbon Methylation cycle

Mitochondrial Oxidative stress Energy metabolism SAM/
function and management Nucleotide methylation
membrane synthesis reactions
maintenance

Fig. 3 | Serine metabolism in cancer cells. Serine is an essential metabolite for important for glutathione (GSH) synthesis and the oxidative stress response,
many cancer cells, even though it can be synthesized in mammalian cells. or for pyruvate, a TCA cycle substrate that is important for ATP production.
Serine can be synthesized from glucose through the serine synthesis pathway, Serine can also be converted into glycine during the conversion of THF to
or from glycine by the enzyme SHMT. Serine has been shown to contribute to 5,10-methylene THF, which is critical for one-carbon metabolism and
ceramide synthesis142, which is important for mitochondrial function; in nucleotide biosynthesis, as well as for the methylation cycle and thus
addition, through its metabolism to phosphatidylserine, serine participates in methylation of DNA, RNA and proteins143.
membrane synthesis. Serine can serve as a source for cysteine, which is

Amino acid deprivation one week only is sufficient to yield significant metabolic changes in
Although amino acids are traditionally classified as ‘essential’ or ‘non- mice66, and methionine deprivation for short time periods is more
essential’, reflecting their relative dietary requirements for organismal likely to be complied with by patients, future studies should probably
fitness and growth, these categories are less relevant for cancer cells. focus on the restriction of dietary methionine for periods shorter than
Most ‘non-essential’ amino acids are required in large amounts by several weeks.
cancer cells for various anabolic processes, and the deprivation of Recently, it was shown that methionine restriction in humans results
specific amino acids often severely compromises the fitness of can- in similar metabolic changes to those observed in methionine-deprived
cer cells, regardless of their ability to synthesize these amino acids. tumour-bearing mice, including inhibition of one-carbon metabolism and
When discussing dietary restriction of amino acids it is important to nucleotide synthesis51. These encouraging results imply that methionine
consider the buffering of their levels in the blood by muscle atrophy, restriction in patients with cancer is likely to inhibit tumour progression
which occurs in order to allow the supply of essential amino acids to similarly to its well-documented anti-tumorigenic effect in mouse models.
essential organs49. Owing to muscle atrophy, dietary deprivation of There are drawbacks to methionine deprivation. For example, the
essential amino acids is likely to cause an increase in the  levels of all restriction of both methionine and cysteine has a pro-angiogenic
non-deprived amino acids in the serum, and even full withdrawal of all effect67 that should be carefully considered when recommending
dietary protein is ineffective at reducing the levels of most amino acids dietary changes for patients with solid tumours. Specifically, secre-
in the blood50. Therefore, it is not trivial to deplete amino acid levels tion of VEGF by endothelial cells following restriction of methionine
in the tumour milieu without causing a severe whole-body reduction and cysteine might support tumour growth, and must be taken into
of these amino acids that is likely to be toxic. From a technical per- account in cases of tumours that overexpress VEGF and rely on it for
spective, specific amino acid restriction can be achieved in humans their growth68.
by defined protein beverages, supplemented with fruits, vegetables Dietary restriction of methionine shows particular promise as a
and some grains51. cancer-preventive regime: it has been shown to extend the lifespan
Here we do not consider the common requirement for all amino acids of rodents69,70, and to result in several metabolic benefits that might
as building blocks for protein synthesis, but rather focus on the unique improve healthspan as well70–72. These studies suggest that long-term
metabolic functions of specific amino acids in cancer metabolism, and methionine restriction is a safe and beneficial dietary strategy that is
the potential benefits of their dietary depletion. worth considering by healthy individuals to prevent cancer.

Methionine. Cancer cells require high levels of methionine for their Serine. Serine contributes to several metabolic processes, including
growth, and many cancer cell lines are methionine auxotrophs52. In ad- nucleotide synthesis, the response to oxidative stress, and the TCA
dition to its key requirement for translation, methionine is a substrate cycle. Serine is considered a non-essential amino acid because cells can
for S-adenosylmethionine (SAM) metabolism and is thus important for synthesize it de novo from glucose or glycine (Fig. 3). However, cancer
SAM-dependent downstream methylation reactions, including DNA cells depend on exogenous serine to support their high proliferation
methylation, which regulates gene expression53. Adequate levels of SAM rate73,74. This positions serine as a cancer-specific essential nutrient.
are also necessary for the activation of mTORC154, which is the effector Dietary deprivation of serine might enhance serine synthesis from
kinase of a major pro-proliferative oncogenic signalling pathway5,23. glucose and/or glycine. While de novo synthesis of serine from glu-
Finally, dietary methionine reduces cellular redox stress55, which is of cose by the serine synthesis pathway is an essential process in some
particular importance for cancer cells. cancer cells and can promote tumour growth in vivo, in these cases it
The potential of dietary methionine restriction to enhance cancer is uncoupled from serine availability and its activation depends mostly
treatment was first suggested in the early 1990s, and has been dem- on genomic variations in the first enzyme in the serine synthesis path-
onstrated in various mouse models and types of cancer, including way—phosphoglycerate dehydrogenase (PHGDH)75. PHGDH catalyses
sarcoma51,56, glioma57,58, prostate cancer59,60, colorectal cancer51,61, the committed step in serine synthesis (3-phospho-d-glycerate + NAD+
breast cancer62–64, and melanoma65. As restriction of methionine for ↔ 3-phosphonooxypyruvate + NADH). In many cases, serine synthesis

Nature | Vol 579 | 26 March 2020 | 511

Review
is harmful for cancer cells because it steers the glycolytic intermediate (THF), an enzymatic cofactor that is essential for nucleotide synthesis,
3-phosphoglycerate (3-PG) away from completion of glycolysis and which is targeted by methotrexate and therefore is a limiting metabo-
therefore reduces energy production76,77. Indeed, overexpression of lite in methotrexate-treated cells. The histidine degradation pathway
PHGDH in vivo is beneficial to tumours only when serine is limiting78. competes with the essential nucleotide-synthesis enzymes for limiting
Serine synthesis from glycine, which is induced by low levels of cel- amounts of THF; therefore, a diet-mediated increase in flux through
lular serine, is harmful for cancer cells because it involves reversal the histidine degradation pathway reduces THF availability, de novo
of the directionality of the one-carbon metabolism enzyme serine nucleotide synthesis and cancer cell survival81.
hydroxymethyltransferase (SHMT). The conversion of glycine to serine This is an example of a dietary modification that can enhance the
by SHMT prevents efficient nucleotide synthesis and compromises efficacy of an established cancer therapy (Fig. 1). In this case the dietary
the proliferation rate of serine-deprived cells73,79. Together, these find- modification is highly feasible, as it involves supplementation with an
ings suggest that serine deprivation is a promising candidate as an amino acid that, unlike deprivation or restriction, is easy for patients
anti-cancer strategy with the potential to slow tumour progression to comply with. Moreover, the proposed enhancement involves meth-
owing to the induction of serine synthesis, which is harmful for rapidly otrexate—a widely used therapy, for which improvements in efficacy
proliferating cells. will benefit a substantial number of patients.
There have been several reports that serine deprivation has anti-
tumour effects in vivo, in various genetic backgrounds and types of Mannose
tumours. For instance, the loss of Trp53, which is mutated in many Mannose is a monosaccharide that can enter cells through glucose trans-
tumours, renders colon cancer cells sensitive to serine deprivation porters but, instead of contributing to energy production, its downstream
because they fail to induce the necessary G1 arrest. In addition, these metabolite—mannose-6-phosphate—accumulates intracellularly. Because
cells suffer from oxidative stress because they fail to channel the resid- mannose is metabolized by the same enzymes that metabolize glucose,
ual serine for glutathione production, and instead use it for nucleotide it interferes with glucose metabolism and inhibits cancer cell growth. In
synthesis80. Similarly, serine deprivation effectively inhibits the growth pancreatic cancer-derived xenografts83, the disrupted glucose metabolism
of autochthonous tumours driven by either Apc loss (in intestinal that results from dietary mannose supplementation renders cancer cells
tumours) or Myc activation (in lymphoma)77. more sensitive to chemotherapy-induced apoptosis, as shown for cispl-
Serine deprivation has also been shown to enhance cancer therapy atin and doxorubicin. The cancer-specific inhibitory effect stems from
by the mitochondrial complex I inhibitor metformin (or phenformin). the high expression of glucose transporters on tumour cells. Mannose
Normally, cells respond to the metformin-mediated reduction of oxida- supplementation therefore offers a promising strategy for enhancing the
tive phosphorylation by increasing glucose consumption and glycolysis. response to cancer therapy that is relatively safe and easy to administer.
However, serine-deprived cells fail to induce this compensatory increase
in glycolytic flux and therefore fail to produce the necessary energy
for their survival. When tested in vivo, either phenformin treatment or Pharmacological depletion of nutrients
serine deprivation alone did not inhibit the growth of colon adenocar- Cancer cells are often auxotrophs for nutrients that are non-essential
cinoma allografts in mice, but their combination resulted in significant for most healthy cells. This presents a vulnerability of the cancer cells
inhibition76. Unfortunately, phenformin coupled with serine deprivation that can be targeted not only by dietary interventions, but also by
is toxic, and the effect achieved by phenformin was not observed with pharmacological approaches. For some of the nutrients discussed
low-dose metformin for the treatment of intestinal tumours in mice77. here, the literature provides information only on pharmacological,
We conclude that dietary deprivation of serine for inhibition of and not dietary, depletion, even though dietary depletion of some of
tumour progression is promising and should be tested in combination these nutrients might be superior. A combination of dietary restriction
with cancer therapies and in various genetic backgrounds. However, and pharmacological targeting of these selectively essential nutrients,
because experiments done in cell culture suggest that serine depriva- such as amino acids and the vitamin folate, is likely to result in improved
tion alone is more effective than simultaneous depletion of serine and depletion of these nutrients and better responses to cancer therapy.
glycine73, future studies should use serine deprivation alone, and not
the combined serine–glycine deprivation that was used in the studies Asparagine
cited above. Reduced levels of asparagine in the plasma of patients with leukae-
mia have been shown to improve survival84. The reduction of aspara-
gine by l-asparaginase demonstrates the key role of asparagine in
Supplementation of nutrients tumour growth and the potential therapeutic benefit of its dietary
Unlike the deprivation of essential nutrients, such as amino acids, sup- depletion. Asparagine is important for several anabolic pathways
plementation of the diet with vital nutrients, such as histidine and because it acts as an amino acid exchange factor (mainly for serine,
mannose, is a feasible dietary modification that can be easily applied as arginine and histidine)85 and as a metastasis-promoting metabolite86.
a complement to established therapies. The challenge here is to iden- Unfortunately, despite the frequent toxicity that accompanies phar-
tify candidate nutrients whose supplementation will inhibit tumour macologic reduction of asparagine84, dietary depletion of asparagine
growth, as these must either be selectively toxic to transformed cells has not been thoroughly studied. However, we can infer from published
or synergize with anti-cancer therapies. work that it is possible to reduce asparagine levels in the serum of
mice fed an asparagine-depleted diet, and that this dietary-mediated
Histidine depletion is sufficient to reduce metastasis in a mouse model of breast-
The flux through the histidine degradation pathway can modify the cancer86. More studies should be done to test the anti-tumour thera-
response of cancer cells to the commonly used chemotherapeutic drug peutic potential of dietary-mediated asparagine depletion, especially
methotrexate81. Low flux through the histidine degradation pathway, for leukaemia, where the transformed cells are frequently asparagine
whether achieved genetically or as the result of intrinsically low expres- auxotrophic84.
sion of the rate-limiting enzyme of the pathway, renders cells refrac-
tory to methotrexate treatment81,82. Conversely, high flux through the Arginine
pathway can be attained by supplementation with histidine, and this Quiescent cells can synthesize arginine from citrulline (as part of the
sensitizes leukaemia tumours to methotrexate therapy81. Mechanisti- urea cycle), but some melanomas and hepatocellular or prostate car-
cally, the histidine degradation pathway consumes tetrahydrofolate cinomas are arginine auxotrophs owing to the silencing of the enzyme

512 | Nature | Vol 579 | 26 March 2020

argininosuccinate synthetase 1 (ASS1)87. This arginine-dependency of
cancer cells88 can be targeted by dietary deprivation of arginine89 or Box 2
by treatment with the arginine-degrading enzyme arginine deiminase
(ADI). Depletion of arginine by ADI has been shown to be effective in
preclinical models of melanoma and hepatocellular carcinoma90, and
The seesaw of glutamine and
in a clinical trial involving patients with hepatocellular carcinoma91. asparagine
Notably, another way to inhibit tumour progression by a reduction in
systemic levels of arginine is secretion of the enzyme arginase 1 (ARG1) Glutamine and asparagine can compensate for the depletion
by hepatocytes, which occurs after inhibition of autophagy in the liver of each other. Asparagine is sufficient to rescue cancer cells
of melanoma-bearing mice92. from glutamine depletion-induced apoptosis, although it was
Arginine is the substrate of the enzyme nitric oxide (NO) synthase, not sufficient to restore proliferation in glutamine-deprived
and therefore influences systemic NO levels93. NO plays a role in cells145. Notably, asparagine rescues cancer cells from glutamine
immunity and has a carcinogenic effect in conditions of chronic deprivation through its role in pro-anabolic signalling85, and not
inflammation94, but it is unknown whether dietary arginine depriva- through its catabolism. In the absence of glutamine, exogenous
tion can reduce NO levels and thereby circumvent this effect. Also, asparagine allows glutamine to be synthesized by glutamine
if the role of NO in immune cell activation is important for tumour synthetase (GLUL)146, which partially restores glutamine-
immunity, NO or its precursor arginine would not be ideal targets dependent metabolic pathways.
for cancer therapy. On the other side of this seesaw, glutamine can compensate
for extracellular depletion of asparagine as long as the cells can
Cystine synthesize asparagine through asparagine synthase147. Together,
Cystine has an important role in maintaining cellular redox balance these findings suggest that dual depletion of both glutamine and
because it is a precursor of cysteine, a component of the anti-oxidant asparagine will result in better inhibition of tumour growth than
glutathione (GSH). However, unlike cysteine, cystine is abundant in depletion of either of these alone.
the plasma and is imported to cells through the cysteine–glutamate
antiporter xCT (SLC7A11). Cystine deprivation inhibits tumour growth
and has been shown to improve survival in three tumour-bearing mouse however, it is unclear whether dietary restriction of these nutrients
models: EGFR-mutant non-small-cell lung cancer (NSCLC) xenografts95; is feasible or safe, or whether it will inhibit tumour growth. These
prostate xenografts; and a genetic leukaemia (TCL1-Tg:Trp53−/−) nutrients and their potential as targets for dietary modifications are
model96. Although these studies used pharmacological depletion of discussed below.
cystine by enzymatic degradation, a similar result might be achieved
through dietary restriction in terms of systemic depletion of cystine. Glutamine and glutamate
Mechanistically, cystine depletion deprives cells of cysteine, and as a Glutamine and its immediate downstream metabolite glutamate con-
consequence also of the antioxidant GSH. This compromises the ability tribute to cancer cell proliferation through several metabolic pathways
of cells to respond to oxidative damage, which is specifically high in (Box 1). They fuel the TCA cycle for production of ATP and NADH104.
cancer cells, and the cells die of ferroptosis. The tumour-specific effect Glutamine is a key component in nucleotide biosynthesis (as the donor
is underlined by the lack of toxicity in healthy tissues upon long-term of two nitrogen atoms in the purine ring and one in the pyrimidine ring)
cystine depletion in mice (up to 50 days)96. and is also essential for glutathione synthesis and maintenance of the
redox balance in cells105. Finally, under hypoxic conditions, glutamine
Folate and glutamate contribute to lipid synthesis106. Glutaminase, the first
Folate is a vitamin (B9) that is necessary for nucleotide synthesis and enzyme in glutamine catabolism, converts glutamine to glutamate;
is therefore essential for all cell types. Fast-proliferating cells, such as inhibition of glutaminase has been shown to reduce tumour growth
enterocytes, haematopoietic cells, and tumour cells, consume high in pre-clinical and clinical trials, but not when glutamate is supplied
levels of folate to fulfill their need for newly synthesized nucleotides for in excess to bypass the inhibition107. Therefore, dietary restriction of
DNA replication and gene expression. Dietary folic acid supplementa- glutamine without restriction of glutamate is not very likely to result
tion was shown to accelerate growth of acute lymphoblastic leukaemia in significant metabolic consequences in cancer cells, and asparagine
(ALL) more than 70 years ago97. This finding enabled the development of can also override glutamine depletion in some cases (Box 2).
anti-folate cancer therapy98, and emphasizes the critical role of folate in There is ample evidence that glutamine is essential for cancer cells
tumour growth99,100. Today the antifolate methotrexate is the standard in culture, but it is less clear whether tumours depend on glutamine
of care for paediatric leukaemia, and is used for the treatment of other in vivo105. Glutamine is considered to be conditionally essential for both
cancer types as well, mostly blood tumours101. Antifolates target one- transformed and non-transformed rapidly proliferating cells108. There-
carbon metabolism and inhibit the synthesis of nucleotides, that are fore, it might be beneficial to focus on the functions of glutamine that
in high demand in fast-proliferating cancer cells102. are unique to cancer cells. For instance, cancer cells are more sensitive
Anti-folate cancer therapies are commonly pharmacologic, so it to oxidative stress than other cells109, so the combination of glutamine
is hard to estimate what dietary folate deprivation would contribute deprivation with a diminished anti-oxidant defence might be specifi-
to tumour inhibition. Folate deprivation combined with dietary uri- cally toxic to cancer cells110,111. Other ways to increase the specificity of
dine supplementation inhibited the initiation and growth of intestinal the inhibitory effect of glutamine deprivation include the inhibition
tumours in Apcmin/+ mice103. However, a comparison of tumour burden of mTOR112 or aspartate transporters113, which would enhance tumour
between normal and folate-free diets is yet to be done and is crucial inhibition by preventing the metabolic rewiring that allows glutamine-
for any conclusions regarding the potential anti-tumorigenic role of deprived cancer cells to survive.
dietary folate deprivation alone. When considering dietary glutamine deprivation as means to
enhance therapy, it is important to consider the levels of glutamine in
the tumour environment. For example, when glutamine was reduced in
Other potential nutritional targets the blood but not the cerebellum of glioblastoma-bearing mice, it had
There are further nutrients, including other amino acids and fatty no significant effect on tumour burden114, but when dietary glutamine
acids, that are essential for the high proliferation rate of cancer cells; deprivation resulted in a systemic reduction in glutamine, including a

Nature | Vol 579 | 26 March 2020 | 513

Review
significant reduction at the cerebellum and in the cerebrospinal fluid non-essential amino acids such as alanine, it is unlikely that a dietary
of medulloblastoma-bearing mice, the treatment resulted in improved approach will be beneficial in this case.
survival115. As it is particularly challenging to reduce glutamine levels
in the brain tumour microenvironment, because stromal cells secrete Fatty acids
glutamine114,116, glutamine measurements at the tumour site are key for More than twice as many ATP molecules can be made from fatty acid
accurate interpretation of in vivo experiments that involve glutamine oxidation (FAO) as from glucose or amino acid catabolism. To under-
depletion in these tumour models. stand whether fatty acid deprivation might affect cancer growth, it
In proliferating mammary epithelial cells117 and in cancer cells is important to establish whether cancer cells can use fatty acids for
that express the oncogene KRAS110, conversion of glutamate to this purpose when glucose is limited, or even as a preferred substrate
α-ketoglutarate by the enzyme glutamate dehydrogenase (GLUD1) regardless of glucose availability.
is inhibited, and glutamate is channelled to aspartate anabolism by In a subset of tumours, FAO is the preferred process for ATP pro-
the cytoplasmic enzyme aspartate transaminase (GOT1) (l-aspartate duction. Some breast cancer tumours activate FAO under metabolic
+ 2-oxoglutarate ↔ oxaloacetate + l-glutamate). This improves redox stress123,124, or in the context of overexpression of MYC125 or carnitine pal-
balance in human pancreatic ductal adenocarcinoma (PDAC) cells and mitoyltransferase (CPT1C)126. A subtype of diffuse large B cell lymphoma
contributes to their proliferation110. In lung adenocarcinoma, the role (DLBCL), OxPhos-DLBCL, also has the ability to oxidize fatty acids as
of glutamate in maintaining intracellular antioxidant levels includes its main energy source, and is sensitive to pharmacologic perturba-
both a direct contribution to glutathione and an indirect contribution tion of this pathway127. A low-fat diet improves the survival of obese
to glutathione biosynthesis through cystine import by the glutamate– leukaemia-bearing mice, emphasizing the essentiality of FAO for the
cystine antiporter xCT118. Glutamate is also secreted through the xCT growth of these tumours and implying that adipocytes are unlikely to
antiporter by glioma cells, which benefit from the antioxidant effect compensate for the shortage in fatty acids128.
of the imported cystine. This glioma-dependent secretion causes a A high-fat diet increases the number of adenomas and carcinomas in
significant increase in glutamate levels in the peritumoral fluid, which the intestine by increasing the number and function of intestinal stem
is toxic to the healthy neurons in the tumour environment and is the cells. The high-fat diet allows progenitor cells to initiate aberrant prolif-
cause of seizures in patients with glioma116. Although targeting the eration in vivo and to form colonies in culture, similar to intestinal stem
xCT transporter showed clinical success, and glutamate was shown to cells, through activation of PPARδ129. However, it has not yet been tested
contribute to glioma growth114,116, it remains unclear whether a reduc- whether restriction of fatty acids will yield the opposite results—that is,
tion in glutamate in the peritumoral fluid can be achieved by dietary reduce the number of spontaneous adenomas or carcinomas in the gut.
restriction of glutamate consumption. FAO also supports cancer cells through NADPH supply. NADPH is
a growth-limiting metabolite and a key mediator of redox balance.
Aspartate Inhibition of FAO causes the accumulation of reactive oxygen species
Aspartate is required for protein synthesis, the urea cycle, the malate– (ROS) and halts the proliferation of glioma cells130, and induction of
aspartate shuttle and nucleic acid synthesis. Low levels of aspartate can FAO increases NADPH production and cancer cell survival through
be growth-limiting under hypoxic conditions, which are common in AMPK signalling131.
tumours, because aspartate synthesis depends on cellular respiration. Together, these findings suggest that there might be therapeutic
Indeed, aspartate levels are low in hypoxic cells119,120 and aspartate is potential in depriving patients of nutritional fatty acids in order to
a limiting metabolite in tumours in vivo119. Inhibitors of the electron starve specific types of cancer34. These findings also raise the possibil-
transport chain block aspartate synthesis through the respiration- ity that a high-fat ketogenic diet might have adverse, pro-tumorigenic
dependent mitochondrial enzyme GOT2. Under these conditions, effects in a subset of cancers. Therefore, it is not clear what cancer types
cells use the cytosolic enzyme GOT1 to synthesize aspartate in order or which patients (with various dietary habits and BMI) will benefit the
to enable protein and nucleic acid biosynthesis121. Increasing cellular most from either a low-fat or a high-fat diet.
levels of aspartate by overexpression of an exogenous asparaginase,
which converts asparagine to aspartate119, or SLC1A3, the aspartate
transporter120, enhances the survival of cancer cells in vivo under Outlook
hypoxic conditions. Notably, under hypoxic conditions, aspartate is Although dietary manipulation of many of the metabolites reviewed here
not a major contributor to the TCA cycle flux; rather, it contributes has shown a clear pre-clinical benefit, and some of them have shown prom-
to pyrimidine and purine biosynthesis, as underlined by the rescue ise in clinical trials, there are still no clear guidelines or recommended
of low de novo aspartate synthesis in hypoxic cells by nucleoside sup- regimes of dietary modification for patients with cancer. More clinical
plementation120. work and preclinical research focusing on understanding how dietary
Aspartate restriction can be considered as a supplementary treat- modifications inhibit cancer in vivo must be completed before dietary
ment for solid tumours, which are poorly vascularized and often interventions become a common approach to cancer therapy.
hypoxic. The hypoxic conditions cause several cellular vulnerabilities It is unlikely that there will be a single recommendation or fit-for-
that can be targeted in a relatively specific manner, including the limited all dietary composition for the prevention or treatment of cancer.
ability to synthesize aspartate. As aspartate is essential for nucleotide Different types of cancer vary in their metabolic activity, preferred
synthesis in hypoxic tumours, it is worth considering a combination of energy source, and nutritional dependencies. Likewise, drugs may have
aspartate restriction with therapies that target nucleotide biosynthesis, different effects when combined with different dietary adjustments.
such as methotrexate and fluorouracil. Furthermore, changes in the diet will not alter nutrient levels equally
in all tissues and tumour microenvironments, so the tumour location
Alanine has an important influence on the efficacy of dietary modifications.
Although alanine has a unique metabolic role in pancreatic ductal ade- This means that, similar to new combinations of drugs, the combina-
nocarcinoma (PDAC), and may substitute for glucose and glutamine as tion of a pharmacologic therapy and a dietary modification must be
a major TCA cycle substrate122, it is unlikely to be a nutrient that can be clinically tested and tailored to each cancer and its type, site and grade.
targeted through the diet. PDAC cells use alanine secreted by stromal In light of the recent promising pre-clinical and clinical findings, we
cells, which reduces their dependency on nutrients supplied by the conclude that some dietary interventions should be tested in the clinic
bloodstream. As dietary modifications shift nutrient availability in the for validation of their therapeutic value sooner rather than later. These
blood stream but do not change the ability of stromal cells to synthesize include reduced glucose consumption, which is likely to be beneficial

514 | Nature | Vol 579 | 26 March 2020

to a substantial number of patients with cancer, and possibly supple- 27. Allen, B. G. et al. Ketogenic diets as an adjuvant cancer therapy: history and potential
mechanism. Redox Biol. 2, 963–970 (2014).
mentation of specific nutrients, as these are feasible and safe dietary
28. Champ, C. E. et al. Targeting metabolism with a ketogenic diet during the treatment of
modifications that can easily be assessed clinically with regard to their glioblastoma multiforme. J. Neurooncol. 117, 125–131 (2014).
benefit in specific cases. 29. Seyfried, T. N., Sanderson, T. M., El-Abbadi, M. M., McGowan, R. & Mukherjee, P. Role of
glucose and ketone bodies in the metabolic control of experimental brain cancer.
We hope that more attention will be given in the future to the dietary
Br. J. Cancer 89, 1375–1382 (2003).
recommendations provided to patients with cancer, and that the motiva- 30. Martuscello, R. T. et al. A supplemented high-fat low-carbohydrate diet for the treatment
tion to investigate the potential to enhance the response to cancer thera- of glioblastoma. Clin. Cancer Res. 22, 2482–2495 (2016).
31. Curry, N. L. et al. Pten-null tumors cohabiting the same lung display differential AKT
pies by modifying the diet will increase. Further work in this direction
activation and sensitivity to dietary restriction. Cancer Discov. 3, 908–921 (2013).
could lead to a paradigm shift in the treatment of patients with cancer. 32. Kalaany, N. Y. & Sabatini, D. M. Tumours with PI3K activation are resistant to dietary
restriction. Nature 458, 725–731 (2009).
This study emphasizes the key role played by glucose as an upstream activator of PI3K
signalling in cancer cell proliferation, by showing in vivo that tumours with constitutive
Online content PI3K activation can survive glucose restriction.
Any methods, additional references, Nature Research reporting sum- 33. Hopkins, B. D. et al. Suppression of insulin feedback enhances the efficacy of PI3K
inhibitors. Nature 560, 499–503 (2018).
maries, source data, extended data, supplementary information,
This paper demonstrates that a combination of dietary and pharmacological insulin
acknowledgements, peer review information; details of author con- inhibition results in reduced pro-oncogenic signalling through mTOR, enhanced
tributions and competing interests; and statements of data and code tumour inhibition and improved survival of tumour-bearing mice in several cancer
models.
availability are available at https://doi.org/10.1038/s41586-020-2124-0. 34. Carracedo, A., Cantley, L. C. & Pandolfi, P. P. Cancer metabolism: fatty acid oxidation in
the limelight. Nat. Rev. Cancer 13, 227–232 (2013).
1. Sullivan, M. R. et al. Quantification of microenvironmental metabolites in murine cancers 35. Erickson, N., Boscheri, A., Linke, B. & Huebner, J. Systematic review: isocaloric ketogenic
reveals determinants of tumor nutrient availability. eLife 8, e44235 (2019). dietary regimes for cancer patients. Med. Oncol. 34, 72 (2017).
2. Cantor, J. R. et al. Physiologic medium rewires cellular metabolism and reveals uric acid This systematic review summarizes reports that have used a ketogenic diet in patients
as an endogenous inhibitor of UMP synthase. Cell 169, 258–272.e217 (2017). with cancer.
3. Muir, A. et al. Environmental cystine drives glutamine anaplerosis and sensitizes cancer 36. Fine, E. J. et al. Targeting insulin inhibition as a metabolic therapy in advanced cancer: a
cells to glutaminase inhibition. eLife 6, e27713 (2017). pilot safety and feasibility dietary trial in 10 patients. Nutrition 28, 1028–1035 (2012).
4. Davidson, S. M. et al. Environment impacts the metabolic dependencies of Ras-driven 37. Jansen, N. & Walach, H. The development of tumours under a ketogenic diet in
non-small cell lung cancer. Cell Metab. 23, 517–528 (2016). association with the novel tumour marker TKTL1: a case series in general practice. Oncol.
5. Torrence, M. E. & Manning, B. D. Nutrient sensing in cancer. Annu. Rev. Cancer Biol. 2, Lett. 11, 584–592 (2016).
251–269 (2018). 38. Maino Vieytes, C. A., Taha, H. M., Burton-Obanla, A. A., Douglas, K. G. & Arthur, A. E.
This review addresses a topic (not discussed here) that is key for understanding the Carbohydrate nutrition and the risk of cancer. Curr. Nutr. Rep. 8, 230–239 (2019).
consequences of dietary modifications that change nutrient availability at the tumour 39. Hannou, A. A. et al. Fructose metabolism and metabolic disease. J. Clin. Invest. 128,
microenvironment, such as the dietary modifications listed above. 545–555 (2018).
6. Hotamisligil, G. S. Inflammation, metaflammation and immunometabolic disorders. 40. Makarem, N., Bandera, E. V., Nicholson, J. M. & Parekh, N. Consumption of sugars, sugary
Nature 542, 177–185 (2017). foods, and sugary beverages in relation to cancer risk: a systematic review of longitudinal
7. Zitvogel, L., Pietrocola, F. & Kroemer, G. Nutrition, inflammation and cancer. studies. Annu. Rev. Nutr. 38, 17–39 (2018).
Nat. Immunol. 18, 843–850 (2017). 41. Jang, C. et al. The small intestine converts dietary fructose into glucose and organic
8. Soldati, L. et al. The influence of diet on anti-cancer immune responsiveness. J. Transl. acids. Cell Metab. 27, 351–361.e353 (2018).
Med. 16, 75 (2018). 42. Douard, V. & Ferraris, R. P. The role of fructose transporters in diseases linked to excessive
9. Nencioni, A., Caffa, I., Cortellino, S. & Longo, V. D. Fasting and cancer: molecular fructose intake. J. Physiol. (Lond.) 591, 401–414 (2013).
mechanisms and clinical application. Nat. Rev. Cancer 18, 707–719 (2018). 43. Jin, C., Gong, X. & Shang, Y. GLUT5 increases fructose utilization in ovarian cancer.
10. Di Francesco, A., Di Germanio, C., Bernier, M. & de Cabo, R. A time to fast. Science 362, OncoTargets Ther. 12, 5425–5436 (2019).
770–775 (2018). 44. Weng, Y., Zhu, J., Chen, Z., Fu, J. & Zhang, F. Fructose fuels lung adenocarcinoma through
11. Descamps, O., Riondel, J., Ducros, V. & Roussel, A. M. Mitochondrial production of GLUT5. Cell Death Dis. 9, 557 (2018).
reactive oxygen species and incidence of age-associated lymphoma in OF1 mice: effect 45. Su, C., Li, H. & Gao, W. GLUT5 increases fructose utilization and promotes tumor
of alternate-day fasting. Mech. Ageing Dev. 126, 1185–1191 (2005). progression in glioma. Biochem. Biophys. Res. Commun. 500, 462–469 (2018).
12. Berrigan, D., Perkins, S. N., Haines, D. C. & Hursting, S. D. Adult-onset calorie restriction 46. Softic, S., Cohen, D. E. & Kahn, C. R. Role of dietary fructose and hepatic de novo
and fasting delay spontaneous tumorigenesis in p53-deficient mice. Carcinogenesis 23, lipogenesis in fatty liver disease. Dig. Dis. Sci. 61, 1282–1293 (2016).
817–822 (2002). 47. Mai, B. H. & Yan, L. J. The negative and detrimental effects of high fructose on the liver, with
13. Lee, C. et al. Fasting cycles retard growth of tumors and sensitize a range of cancer cell special reference to metabolic disorders. Diabetes Metab. Syndr. Obes. 12, 821–826 (2019).
types to chemotherapy. Sci. Transl. Med. 4, 124ra27 (2012). 48. Goncalves, M. D. et al. High-fructose corn syrup enhances intestinal tumor growth in
14. Lee, C. et al. Reduced levels of IGF-I mediate differential protection of normal and cancer mice. Science 363, 1345–1349 (2019).
cells in response to fasting and improve chemotherapeutic index. Cancer Res. 70, 49. Kamata, S. et al. Dietary deprivation of each essential amino acid induces differential
1564–1572 (2010). systemic adaptive responses in mice. Mol. Nutr. Food Res. 58, 1309–1321 (2014).
15. Brandhorst, S. et al. A periodic diet that mimics fasting promotes multi-system This study demonstrates that deprivation of essential amino acids results in a systemic
regeneration, enhanced cognitive performance, and healthspan. Cell Metab. 22, 86–99 adaptive response that circumvents the systemic drop of the deprived amino acid as
(2015). measured in the plasma.
16. Locasale, J. W. The consequences of enhanced cell-autonomous glucose metabolism. 50. Adibi, S. A., Modesto, T. A., Morse, E. L. & Amin, P. M. Amino acid levels in plasma, liver,
Trends Endocrinol. Metab. 23, 545–551 (2012). and skeletal muscle during protein deprivation. Am. J. Physiol. 225, 408–414 (1973).
This paper describes in depth the metabolic alterations that occur in cells that process 51. Gao, X. et al. Dietary methionine influences therapy in mouse cancer models and alters
abnormally high amount of glucose, such as transformed cells. human metabolism. Nature 572, 397–401 (2019).
17. Belhocine, T. et al. 18FDG PET in oncology: the best and the worst (Review). Int. J. Oncol. This study compared the metabolic consequences of methionine restriction between
28, 1249–1261 (2006). patients with cancer and tumour-bearing mice, and found them to be similar. In
18. Gallagher, E. J. & LeRoith, D. Minireview: IGF, insulin, and cancer. Endocrinology 152, addition, methionine restriction was shown to influence anti-cancer therapy.
2546–2551 (2011). 52. Hoffman, R. M. & Erbe, R. W. High in vivo rates of methionine biosynthesis in transformed
19. Warburg, O. On the origin of cancer cells. Science 123, 309–314 (1956). human and malignant rat cells auxotrophic for methionine. Proc. Natl Acad. Sci. USA 73,
20. Courtney, K. D. et al. Isotope tracing of human clear cell renal cell carcinomas 1523–1527 (1976).
demonstrates suppressed glucose oxidation in vivo. Cell Metab. 28, 793–800.e2 The work described here provides experimental proof that methionine auxotrophy in
(2018). transformed cells stems not from an inability of these cells to synthesize methionine,
21. Vander Heiden, M. G., Cantley, L. C. & Thompson, C. B. Understanding the Warburg but rather from their high demand for methionine.
effect: the metabolic requirements of cell proliferation. Science 324, 1029–1033 53. Mentch, S. J. et al. Histone methylation dynamics and gene regulation occur through the
(2009). sensing of one-carbon metabolism. Cell Metab. 22, 861–873 (2015).
22. Fruman, D. A. et al. The PI3K pathway in human disease. Cell 170, 605–635 (2017). 54. Gu, X. et al. SAMTOR is an S-adenosylmethionine sensor for the mTORC1 pathway.
23. Saxton, R. A. & Sabatini, D. M. mTOR signaling in growth, metabolism, and disease. Cell Science 358, 813–818 (2017).
169, 361–371 (2017). 55. Eriksson, S., Prigge, J. R., Talago, E. A., Arnér, E. S. & Schmidt, E. E. Dietary methionine can
24. Fontana, L., Meyer, T. E., Klein, S. & Holloszy, J. O. Long-term calorie restriction is highly sustain cytosolic redox homeostasis in the mouse liver. Nat. Commun. 6, 6479 (2015).
effective in reducing the risk for atherosclerosis in humans. Proc. Natl Acad. Sci. USA 101, 56. Guo, H. et al. Therapeutic tumor-specific cell cycle block induced by methionine
6659–6663 (2004). starvation in vivo. Cancer Res. 53, 5676–5679 (1993).
25. Brandhorst, S. & Longo, V. D. Fasting and caloric restriction in cancer prevention and 57. Poirson-Bichat, F. et al. Methionine deprivation and methionine analogs inhibit cell
treatment. Recent Results Cancer Res. 207, 241–266 (2016). proliferation and growth of human xenografted gliomas. Life Sci. 60, 919–931 (1997).
26. Weber, D. D., Aminazdeh-Gohari, S. & Kofler, B. Ketogenic diet in cancer therapy. Aging 58. Liu, H. et al. Methionine and cystine double deprivation stress suppresses glioma
(Albany NY) 10, 164–165 (2018). proliferation via inducing ROS/autophagy. Toxicol. Lett. 232, 349–355 (2015).

Nature | Vol 579 | 26 March 2020 | 515

Review
59. Poirson-Bichat, F., Gonfalone, G., Bras-Gonçalves, R. A., Dutrillaux, B. & Poupon, M. F. 94. Ohshima, H. & Bartsch, H. Chronic infections and inflammatory processes as cancer
Growth of methionine-dependent human prostate cancer (PC-3) is inhibited by ethionine risk factors: possible role of nitric oxide in carcinogenesis. Mutat. Res. 305, 253–264
combined with methionine starvation. Br. J. Cancer 75, 1605–1612 (1997). (1994).
60. Sinha, R. et al. Dietary methionine restriction inhibits prostatic intraepithelial neoplasia in 95. Poursaitidis, I. et al. Oncogene-selective sensitivity to synchronous cell death following
TRAMP mice. Prostate 74, 1663–1673 (2014). modulation of the amino acid nutrient cystine. Cell Rep. 18, 2547–2556 (2017).
61. Komninou, D., Leutzinger, Y., Reddy, B. S. & Richie, J. P. Jr Methionine restriction inhibits 96. Cramer, S. L. et al. Systemic depletion of l-cyst(e)ine with cyst(e)inase increases reactive
colon carcinogenesis. Nutr. Cancer 54, 202–208 (2006). oxygen species and suppresses tumor growth. Nat. Med. 23, 120–127 (2017).
62. Strekalova, E., Malin, D., Good, D. M. & Cryns, V. L. Methionine deprivation induces a 97. Farber, S. et al. The action of pteroylglutamic conjugates on man. Science 106, 619–621
targetable vulnerability in triple-negative breast cancer cells by enhancing TRAIL (1947).
receptor-2 expression. Clin. Cancer Res. 21, 2780–2791 (2015). 98. Farber, S. & Diamond, L. K. Temporary remissions in acute leukemia in children
63. Jeon, H. et al. Methionine deprivation suppresses triple-negative breast cancer produced by folic acid antagonist, 4-aminopteroyl-glutamic acid. N. Engl. J. Med. 238,
metastasis in vitro and in vivo. Oncotarget 7, 67223–67234 (2016). 787–793 (1948).
64. Hens, J. R. et al. Methionine-restricted diet inhibits growth of MCF10AT1-derived 99. Wilson, P. M., Danenberg, P. V., Johnston, P. G., Lenz, H. J. & Ladner, R. D. Standing the test
mammary tumors by increasing cell cycle inhibitors in athymic nude mice. BMC Cancer of time: targeting thymidylate biosynthesis in cancer therapy. Nat. Rev. Clin. Oncol. 11,
16, 349 (2016). 282–298 (2014).
65. Miousse, I. R. et al. Modulation of dietary methionine intake elicits potent, yet distinct, 100. Newman, A. C. & Maddocks, O. D. K. One-carbon metabolism in cancer. Br. J. Cancer 116,
anticancer effects on primary versus metastatic tumors. Carcinogenesis 39, 1117–1126 (2018). 1499–1504 (2017).
66. Yu, D. et al. Short-term methionine deprivation improves metabolic health via sexually 101. Pui, C. H. & Evans, W. E. Treatment of acute lymphoblastic leukemia. N. Engl. J. Med. 354,
dimorphic, mTORC1-independent mechanisms. FASEB J. 32, 3471–3482 (2018). 166–178 (2006).
67. Longchamp, A. et al. Amino acid restriction triggers angiogenesis via GCN2/ATF4 102. Suvalov, O. et al. One-carbon metabolism and nucleotide biosynthesis as attractive
regulation of VEGF and H2S production. Cell 173, 117–129.e114 (2018). targets for anticancer therapy. Oncotarget 8, 23955–23977 (2017).
68. Carmeliet, P. VEGF as a key mediator of angiogenesis in cancer. Oncology 69 (Suppl. 3), 103. Field, M. S., Lan, X., Stover, D. M. & Stover, P. J. Dietary uridine decreases tumorigenesis in
4–10 (2005). the ApcMin/+ model of intestinal cancer. Curr. Dev. Nutr. 2, nzy013 (2018).
69. Orentreich, N., Matias, J. R., DeFelice, A. & Zimmerman, J. A. Low methionine ingestion by 104. DeBerardinis, R. J. et al. Beyond aerobic glycolysis: transformed cells can engage in
rats extends life span. J. Nutr. 123, 269–274 (1993). glutamine metabolism that exceeds the requirement for protein and nucleotide
70. Miller, R. A. et al. Methionine-deficient diet extends mouse lifespan, slows immune and synthesis. Proc. Natl Acad. Sci. USA 104, 19345–19350 (2007).
lens aging, alters glucose, T4, IGF-I and insulin levels, and increases hepatocyte MIF 105. Cluntun, A. A., Lukey, M. J., Cerione, R. A. & Locasale, J. W. Glutamine metabolism in
levels and stress resistance. Aging Cell 4, 119–125 (2005). cancer: understanding the heterogeneity. Trends Cancer 3, 169–180 (2017).
71. Ables, G. P., Perrone, C. E., Orentreich, D. & Orentreich, N. Methionine-restricted C57BL/6J 106. Metallo, C. M. et al. Reductive glutamine metabolism by IDH1 mediates lipogenesis under
mice are resistant to diet-induced obesity and insulin resistance but have low bone hypoxia. Nature 481, 380–384 (2011).
density. PLoS One 7, e51357 (2012). 107. Biancur, D. E. et al. Compensatory metabolic networks in pancreatic cancers upon
72. Malloy, V. L. et al. Methionine restriction decreases visceral fat mass and preserves insulin perturbation of glutamine metabolism. Nat. Commun. 8, 15965 (2017).
action in aging male Fischer 344 rats independent of energy restriction. Aging Cell 5, 108. Lacey, J. M. & Wilmore, D. W. Is glutamine a conditionally essential amino acid? Nutr. Rev.
305–314 (2006). 48, 297–309 (1990).
73. Labuschagne, C. F., van den Broek, N. J., Mackay, G. M., Vousden, K. H. & Maddocks, O. D. 109. Moloney, J. N. & Cotter, T. G. ROS signalling in the biology of cancer. Semin. Cell Dev. Biol.
Serine, but not glycine, supports one-carbon metabolism and proliferation of cancer 80, 50–64 (2018).
cells. Cell Rep. 7, 1248–1258 (2014). 110. Son, J. et al. Glutamine supports pancreatic cancer growth through a KRAS-regulated
74. DeNicola, G. M. et al. NRF2 regulates serine biosynthesis in non-small cell lung cancer. metabolic pathway. Nature 496, 101–105 (2013).
Nat. Genet. 47, 1475–1481 (2015). 111. Chakrabarti, G. et al. Targeting glutamine metabolism sensitizes pancreatic cancer to
75. Possemato, R. et al. Functional genomics reveal that the serine synthesis pathway is PARP-driven metabolic catastrophe induced by ß-lapachone. Cancer Metab. 3, 12 (2015).
essential in breast cancer. Nature 476, 346–350 (2011). 112. Momcilovic, M. et al. The GSK3 signaling axis regulates adaptive glutamine metabolism
76. Gravel, S. P. et al. Serine deprivation enhances antineoplastic activity of biguanides. in lung squamous cell carcinoma. Cancer Cell 33, 905–921.e905 (2018).
Cancer Res. 74, 7521–7533 (2014). 113. Tajan, M. et al. A role for p53 in the adaptation to glutamine starvation through the
77. Maddocks, O. D. K. et al. Modulating the therapeutic response of tumours to dietary expression of SLC1A3. Cell Metab. 28, 721–736.e6 (2018).
serine and glycine starvation. Nature 544, 372–376 (2017). 114. Tardito, S. et al. Glutamine synthetase activity fuels nucleotide biosynthesis and
78. Sullivan, M. R. et al. Increased serine synthesis provides an advantage for tumors arising supports growth of glutamine-restricted glioblastoma. Nat. Cell Biol. 17, 1556–1568
in tissues where serine levels are limiting. Cell Metab. 29, 1410–1421.e1414 (2019). (2015).
79. Pacold, M. E. et al. A PHGDH inhibitor reveals coordination of serine synthesis and one- This paper emphasizes the importance of assessment of nutrients manipulated by dietary
carbon unit fate. Nat. Chem. Biol. 12, 452–458 (2016). modifications at the tumour microenvironment and not elsewhere, such as the plasma.
80. Maddocks, O. D. et al. Serine starvation induces stress and p53-dependent metabolic 115. Niklison-Chirou, M. V. et al. TAp73 is a marker of glutamine addiction in medulloblastoma.
remodelling in cancer cells. Nature 493, 542–546 (2013). Genes Dev. 31, 1738–1753 (2017).
The study provides evidence that serine starvation in vivo is harmful for p53-deficient 116. Yang, L. et al. Targeting stromal glutamine synthetase in tumors disrupts tumor
tumours, and discusses the role of serine in oxidative stress management.  microenvironment-regulated cancer cell growth. Cell Metab. 24, 685–700 (2016).
81. Kanarek, N. et al. Histidine catabolism is a major determinant of methotrexate sensitivity. 117. Coloff, J. L. et al. Differential glutamate metabolism in proliferating and quiescent
Nature 559, 632–636 (2018). mammary epithelial cells. Cell Metab. 23, 867–880 (2016).
This paper describes a dietary supplement that enhances the efficacy of a common 118. Sayin, V. I. et al. Activation of the NRF2 antioxidant program generates an imbalance in
chemotherapy drug in vivo. central carbon metabolism in cancer. eLife 6, e28083 (2017).
82. Newman, E. M., Nierenberg, D. W. & Santi, D. V. Selective killing of transformed cells by 119. Sullivan, L. B. et al. Aspartate is an endogenous metabolic limitation for tumour growth.
methotrexate with histidine deprivation or with alpha-amino alcohols. Cancer Res. 43, Nat. Cell Biol. 20, 782–788 (2018).
4703–4708 (1983). 120. Garcia-Bermudez, J. et al. Aspartate is a limiting metabolite for cancer cell proliferation
83. Gonzalez, P. S. et al. Mannose impairs tumour growth and enhances chemotherapy. under hypoxia and in tumours. Nat. Cell Biol. 20, 775–781 (2018).
Nature 563, 719–723 (2018). This study studied the dependency of tumours on aspartate under hypoxic conditions,
This research describes a rare example of a nutrient that is toxic for cancer cells but and it provides an interesting example of potential specificity of tumour targeting by
otherwise harmless, and that has the potential to be supplemented in the diet as an dietary deprivation of nutrients that might be achieved by tumour-specific
anti-cancer treatment. vulnerability, in this case owing to the hypoxic conditions at the tumour site.
84. Hijiya, N. & van der Sluis, I. M. Asparaginase-associated toxicity in children with acute 121. Birsoy, K. et al. An essential role of the mitochondrial electron transport chain in cell
lymphoblastic leukemia. Leuk. Lymphoma 57, 748–757 (2016). proliferation is to enable aspartate synthesis. Cell 162, 540–551 (2015).
85. Krall, A. S., Xu, S., Graeber, T. G., Braas, D. & Christofk, H. R. Asparagine promotes cancer cell 122. Sousa, C. M. et al. Pancreatic stellate cells support tumour metabolism through
proliferation through use as an amino acid exchange factor. Nat. Commun. 7, 11457 (2016). autophagic alanine secretion. Nature 536, 479–483 (2016).
86. Knott, S. R. V. et al. Asparagine bioavailability governs metastasis in a model of breast This study shows that although alanine has a unique metabolic role in pancreatic
cancer. Nature 554, 378–381 (2018). ductal adenocarcinoma, it is not likely to be a good target for dietary modifications
87. Dillon, B. J. et al. Incidence and distribution of argininosuccinate synthetase deficiency in because it is supplied to the tumour by surrounding stroma cells, illustrating the
human cancers: a method for identifying cancers sensitive to arginine deprivation. importance of understanding a tumour’s sources of nutrition.
Cancer 100, 826–833 (2004). 123. Carracedo, A. et al. A metabolic prosurvival role for PML in breast cancer. J. Clin. Invest.
88. Lind, D. S. Arginine and cancer. J. Nutr. 134 (Suppl.), 2837S–2841S (2004). 122, 3088–3100 (2012).
89. Kremer, J. C. et al. Arginine deprivation inhibits the Warburg effect and upregulates 124. Schafer, Z. T. et al. Antioxidant and oncogene rescue of metabolic defects caused by loss
glutamine anaplerosis and serine biosynthesis in ASS1-deficient cancers. Cell Rep. 18, of matrix attachment. Nature 461, 109–113 (2009).
991–1004 (2017). 125. Camarda, R. et al. Inhibition of fatty acid oxidation as a therapy for MYC-overexpressing
90. Ensor, C. M., Holtsberg, F. W., Bomalaski, J. S. & Clark, M. A. Pegylated arginine deiminase triple-negative breast cancer. Nat. Med. 22, 427–432 (2016).
(ADI-SS PEG20,000 mw) inhibits human melanomas and hepatocellular carcinomas 126. Zaugg, K. et al. Carnitine palmitoyltransferase 1C promotes cell survival and tumor
in vitro and in vivo. Cancer Res. 62, 5443–5450 (2002). growth under conditions of metabolic stress. Genes Dev. 25, 1041–1051 (2011).
91. Izzo, F. et al. Pegylated arginine deiminase treatment of patients with unresectable 127. Caro, P. et al. Metabolic signatures uncover distinct targets in molecular subsets of
hepatocellular carcinoma: results from phase I/II studies. J. Clin. Oncol. 22, 1815–1822 (2004). diffuse large B cell lymphoma. Cancer Cell 22, 547–560 (2012).
92. Poillet-Perez, L. et al. Autophagy maintains tumour growth through circulating arginine. 128. Tucci, J. et al. Switch to low-fat diet improves outcome of acute lymphoblastic leukemia
Nature 563, 569–573 (2018). in obese mice. Cancer Metab. 6, 15 (2018).
93. Palmer, R. M., Ashton, D. S. & Moncada, S. Vascular endothelial cells synthesize nitric 129. Beyaz, S. et al. High-fat diet enhances stemness and tumorigenicity of intestinal
oxide from l-arginine. Nature 333, 664–666 (1988). progenitors. Nature 531, 53–58 (2016).

516 | Nature | Vol 579 | 26 March 2020

130. Pike, L. S., Smift, A. L., Croteau, N. J., Ferrick, D. A. & Wu, M. Inhibition of fatty acid 146. Pavlova, N. N. et al. As extracellular glutamine levels decline, asparagine becomes an
oxidation by etomoxir impairs NADPH production and increases reactive oxygen species essential amino acid. Cell Metab. 27, 428–438.e425 (2018).
resulting in ATP depletion and cell death in human glioblastoma cells. Biochim. Biophys. 147. Chan, W. K. et al. The glutaminase activity of l-asparaginase is not required for anticancer
Acta 1807, 726–734 (2011). activity against ASNS-negative cells. Blood 123, 3596–3606 (2014).
131. Jeon, S. M., Chandel, N. S. & Hay, N. AMPK regulates NADPH homeostasis to promote
tumour cell survival during energy stress. Nature 485, 661–665 (2012).
132. Yang, C. et al. Glioblastoma cells require glutamate dehydrogenase to survive Acknowledgements We thank all members of the Sabatini lab, and members of the Pathology
impairments of glucose metabolism or Akt signaling. Cancer Res. 69, 7986–7993 Department at Boston Children’s Hospital for providing the research environment that inspired
(2009). the writing of this review; K. Hixon, L. A. Shinefeld, L. D. Schweitzer, L. Pernas,
133. Hensley, C. T., Wasti, A. T. & DeBerardinis, R. J. Glutamine and cancer: cell biology, L. Chantranupong and B. D. Hopkins for helpful comments on the manuscript; L. Nip for the
physiology, and clinical opportunities. J. Clin. Invest. 123, 3678–3684 (2013). graphic design; and D. A. Guertin for his mentorship. This work was supported by grants from
134. Hosios, A. M. et al. Amino acids rather than glucose account for the majority of cell mass the NIH to D.M.S. (R01 CA103866, R01 CA129105 and R37 AI047389), and from the Lustgarten
in proliferating mammalian cells. Dev. Cell 36, 540–549 (2016). Foundation. Fellowship support was provided by the European Molecular Biology Organization
135. Psychogios, N. et al. The human serum metabolome. PLoS ONE 6, e16957 (2011). (EMBO) (Long-Term Fellowship ALTF 350-2012) and the American Association for Cancer
136. Di Biase, S. et al. Fasting-Mimicking Diet Reduces HO-1 to promote T cell-mediated tumor Research (16-40-38-KANA) to N.K., with additional support from by the Women in Science/
cytotoxicity. Cancer Cell 30, 136–146 (2016). Revson Foundation Award (Weizmann Institute) and The Advancement of Women in Science
137. Pietrocola, F. et al. Caloric restriction mimetics enhance anticancer immunosurveillance. Award (The Hebrew University). N.K. and B.P. are supported by Boston Children’s Hospital and
Cancer Cell 30, 147–160 (2016). the Children’s Hospital Pathology Foundation. D.M.S. is an investigator of the Howard Hughes
138. Lussier, D. M. et al. Enhanced immunity in a mouse model of malignant glioma is Medical Institute and an American Cancer Society Research Professor.
mediated by a therapeutic ketogenic diet. BMC Cancer 16, 310 (2016).
139. Geiger, R. et al. l-Arginine modulates T cell metabolism and enhances survival and anti- Author contributions N.K. wrote the first draft. B.P. and N.K. wrote and edited the final
tumor activity. Cell 167, 829–842.e813 (2016). manuscript. D.M.S. provided substantial edits and suggestions for each of the drafts. All
140. Hui, S. et al. Glucose feeds the TCA cycle via circulating lactate. Nature 551, 115–118 authors edited and approved the final manuscript.
(2017).
141. Faubert, B. et al. Lactate metabolism in human lung tumors. Cell 171, 358–371.e359
Competing interests The authors declare no competing interests.
(2017).
142. Gao, X. et al. Serine availability influences mitochondrial dynamics and function through
lipid metabolism. Cell Rep. 22, 3507–3520 (2018). Additional information
143. Mattaini, K. R., Sullivan, M. R. & Vander Heiden, M. G. The importance of serine Correspondence and requests for materials should be addressed to N.K.
metabolism in cancer. J. Cell Biol. 214, 249–257 (2016). Reprints and permissions information is available at http://www.nature.com/reprints.
144. Rajagopalan, K. N. & DeBerardinis, R. J. Role of glutamine in cancer: therapeutic and Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
imaging implications. J. Nuclear Med. 52, 1005–1008 (2011). published maps and institutional affiliations.
145. Zhang, J. et al. Asparagine plays a critical role in regulating cellular adaptation to
glutamine depletion. Mol. Cell 56, 205–218 (2014). © Springer Nature Limited 2020

Nature | Vol 579 | 26 March 2020 | 517